skid resistance - transportation research...

NATIONAL COOPERATIVE HIGHWAY RESEARCH PROGRAM SYNTHESIS OF HIGHWAY PRACTICE

SKID RESISTANCE

AHOThANSORTATION DEPARTMENT

L WESEARCHLIBRARY

HIGHWAY RESEARCH BOARD NATIONAL RESEARCH COUNCIL

NATIONAL ACADEMY OF SCIENCES-NATIONAL ACADEMY OF ENGINEERING

HIGHWAY RESEARCH BOARD 1972

Officers ALAN M. VOORHEES, Chairman WILLIAM L. GARRISON, First Vice Chairman JAY W. BROWN, Second Vice Chairman W. N. CAREY, JR., Executive Director

Executive Committee HENRIK E. STAFSETH, Executive Director, American Association of State Highway Officials (cx officio) RALPH R. BARTELSMEYER, Federal Highway Administrator (Acting), U.S. Department of Transportation (cx officio)

CARLOS C. VILLARREAL, Urban Mass Transportation Administrator, U.S. Department of Transportation (cx officio) ERNST WEBER, Chairman, Division of Engineering, National Research Council (cx officio) D. GRANT MICKLE, President, Highway Users Federation for Safety and Mobility (cx officio, Past Chairman 1970)

CHARLES E. SHUMATE, Executive Director-Chief Engineer, Colorado Department of Highways (cx officio, Past Chairman 1971)

HENDRIK W. BODE, Professor of Systems Engineering, Harvard University JAY W. BROWN, Director of Road Operations, Florida Department of Transportation W. J. BURMEISTER, Executive Director, Wisconsin Asphalt Pavemnent Association HOWARD A. COLEMAN, Consultant, Missouri Portland Ce,nent Company

DOUGLAS B. FUGATE, Commissioner, Virginia Department of Highways WILLIAM L. GARRISON, Professor of Environmental Engineering, University of Pittsburgh

ROGER H. OILMAN, Director of Planning and Development, The Port Authority of New York and New Jersey

GEORGE E. HOLBROOK, E. I. du Pont de Ne,nours and Company GEORGE KRAMBLES, Superintendent of Research and Planning, Chicago Transit Authority

A. SCHEFFER LANG, Office of the President, Association of American Railroads JOHN A. LEGARRA, Deputy State Highway Engineer, California Division of Highways

WILLIAM A. McCONNELL, Director, Product Test Operations Office, Ford Motor Company JOHN J. McKETTA, Department of Chemical Engineering, University of Texas JOHN T. MIDDLETON, Deputy Assistant Administrator, Office of Air Programs, Environmental Protection Agency

ELLIOTT W. MONTROLL, Professor of Physics, University of Rochester

R. L. PEYTON, Assistant State Highway Director, State Highway Conunission of Kansas

MILTON PIKARSKY, Commissioner of Public Works, Chicago DAVID H. STEVENS, Chairman, Maine State Highway Commission ALAN M. VOORHEES, President, Alan M. Voorhees and Associates ROBERT N. YOUNG, Executive Director, Regional Planning Council, Baltimore, Maryland

NATIONAL COOPERATIVE HIGHWAY RESEARCH PROGRAM

Advisory Committee ALAN M. VOORHEES, Alan M. Voorhees and Associates (Chairman)

WILLIAM L. GARRISON, University of Pittsburgh J. W. BROWN, Florida Department of Transportation HENRIK E. STAFSETH, American Association of State Highway Officials RALPH R. BARTELSMEYER, U.S. Department of Transportation ERNST WEBER, National Research Council D. GRANT MICKLE, Highway Users Federation for Safety and Mobility CHARLES E. SHUMATE, Colorado Department of Highways W. N. CAREY, JR., Highway Research Board

Advisory Committee on Project 20-5

J. N. CLARY, Virginia Department of Highways (Chairman)

JACK F. ANDREWS, New Jersey Department of Transportation

W. S. EKERN, Minnesota Department of Highways

WILLIAM P. HOFMANN, New York State Department of Transportation

JOHN W. HOSSACK, Village of Hoffman Estates, Ill.

FRANK E. LEGG, JR., University of Michigan

ALGER F. MALO, City of Detroit

GEORGE W. McALPIN, New York State Department of Transportation

JOHN K. MLADINOV, New York State Department of Transportation

T. F. MORF, Consulting Engineer

CARL F. IZZARD, Federal Highway Administration

ROY C. EDGERTON, Highway Research Board

Topic Advisory Panel on Skid Resistance:

R. CLARKE BENNETT, Federal Highway Administration WILLIAM GARTNER JR., Florida Department of

Transportation DAVID C. MAHONE, Virginia Council of Highway

Investigation and Research ROLANDS L. RIZENBERGS, Kentucky Department of

Highways RICHARD K. SHAFFER, Pennsylvania Department of

Transportation GEORGE SHERMAN, California Division of Highways WILLIAM P. WALKER, Federal Highway Administration

Consultants to Topic Panel: W. E. MEYER, The Pennsylvania State University WILLIAM A. GOODWIN, The University of Tennessee

Program Staff K. W. HENDERSON, JR., Program Director LOUIS M. MAcGREGOR, Administrative Engineer HARRY A. SMITH, Projects Engineer

JOHN E. BURKE, Projects Engineer WILLIAM L. WILLIAMS, Projects Engineer

GEORGE E. FRANGOS, Projects Engineer HERBERT P. ORLAND, Editor

ROBERT J. REILLY, Projects Engineer ROSEMARY M. MALONEY, Editor

NATIONAL COOPERATIVE HIGHWAY RESEARCH PROGRAM 1 4 SYNTHESIS OF HIGHWAY PRACTICE

SKID RESISTANCE

RESEARCH SPONSORED BY THE AMERICAN ASSOCIATION

OF STATE HIGHWAY OFFICIALS IN COOPERATION

WITH THE FEDERAL HIGHWAY ADMINISTRATION

AREAS OF INTERESTS

PAVEMENT DESIGN

PAVEMENT PERFORMANCE

BITUMINOUS MATERIALS AND MIXES

CONSTRUCTION

MAINTENANCE, GENERAL

HIGHWAY SAFETY

HIGHWAY RESEARCH BOARD DIVISION OF ENGINEERING NATIONAL RESEARCH COUNCIL

NATIONAL ACADEMY OF SCIENCES- NATIONAL ACADEMY OF ENGINEERING 1972

N


Systematic, well-designed research provides the most ef-fective approach to the solution of many problems facing highway administrators and engineers. Often, highway problems are of local interest and can best be studied by highway departments individually or in cooperation with their state universities and others. However, the accelerat-ing growth of highway transportation develops increasingly complex problems of wide interest to highway authorities. These problems are best studied through a coordinated program of cooperative research.

In recognition of these needs, the highway administrators of the American Association of State Highway Officials initiated in 1962 an objective national highway research program employing modern scientific techniques. This program is supported on a continuing basis by funds from participating member states of the Association and it re-ceives the full cooperation and support of the Federal Highway Administration, • United States Department of Transportation.

The Highway Research Board of the National Academy of Sciences-National Research Council was requested by the Association to administer the research program because of the Board's recognized objectivity and understanding of modern research practices. The Board is uniquely suited for this purpose as: it maintains an extensive committee structure from which authorities on any highway transpor-tation subject may be drawn; it possesses avenues of com-munications and cooperation with federal, state, and local governmental agencies, universities, and industry; its rela-tionship to its parent organization, thë National Academy of Sciences, a private, nonprofit institution, is an insurance of objectvity; it maintains a full-time research correlation staff of specialists in highway transportation matters to bring the findings of research directly to those who are in a position to use them.

The program is developed on the basis of research needs identified by chief administrators of the highway depart-ments and by committees of AASHO. Each year, specific areas of research needs to be included in the program are proposed to the Academy and the Board by the American Association of State Highway Officials. Research projects to fulfill these needs are defined by the Board, and qualified research agencies are selected from those that have sub-mitted proposals. Administration and surveillance of re-search contracts are responsibilities of the Academy and its Highway Research Board.

The needs for highway research are many, and the National Cooperative Highway Research Program can make significant contributions to the solution of highway transportation problems of mutual concern to many re-sponsible groups. The program, however, is intended to complement rather than to substitute for or duplicate other highway research programs.

NCHRP Synthesis 14

Project 20-5 FY '70 ISBN 0-309-02024-7 L. C. Card No. 72-13941

Price: $400

This report is one of a series of reports issued from a continuing research program conducted under a three-way agreement entered into in June 1962 by and among the National Academy of Sciences-National Research Council, the American Association of State High-way Officials, and the Federal Highway Administration. Individual fiscal agreements are executed annually by the Academy-Research Council, the Federal Highway Administration, and participating state highway departments, members of the American Association of State Highway Officials.

The study reported herein was undertaken under the aegis of the National Academy of Sciences—National Research Council. The National Cooperative Highway Research Program, under which this study was made, is conducted by the Highway Research Board with the express approval of the Governing Board of the NRC. Such ap-proval indicated that the Board considered that the problems studied in this program are of national significance; that solution of the problems requires scientific or technical competence, and that the resources of NRC are particularly suitable for the oversight of these studies. The institutional responsibilities of the NRC are discharged in the following manner: each specific problem, before it is ac-cepted for study in the program, is approved as appropriate for the NRC by the NCHRP Program Advisory Committee and the Chair-man of the Division of Engineering of the National Research Council.

Topics for synthesis are selected and defined by an advisory com-mittee that monitors the work and reviews the final report. Mem-bers of the advisory committees are appointed by the Chairman of the Division of Engineering of the National Research Council. They are selected for their individual scholarly competence and judgment, with due consideration for the balance and breadth of disciplines. Responsibility for the definition of this study and for the publication of this report rests with the advisory committee.

Although reports in this category are not submitted for approval to the Academy membership nor to the Council, each report is reviewed and processed according to procedures established and monitored by the Academy's Report Review Committee. Such reviews are intended to determine inter alia, whether the major questions and relevant points of view have been addressed, and whether the reported find-ings, conclusions and recommendations arose from the available data and information. Distribution of the report is permitted only after satisfactory completion of this review process.

Published reports of the


are available from:

Highway Research Board National Academy of Sciences 2101 Constitution Avenue Washington, D.C. 20418

(See last pages for list of published titles and prices)

PREFACE There exists a vast storehouse of information relating to nearly every subject of concern' to highway administrators and engineers. Much of it resulted from research and much from successful application of the engineering ideas of men faced with problems in their day-to-day work. Because there has been a lack of systematic means for bringing such useful information together and making it available to the entire highway fraternity, the American Association of State Highway Officials' has, through the mechanism of the National Cooperative Highway Research Program, authorized the Highway Research Board to undertake a continuing project to search out and synthesize the useful knowledge from all possible sources and to' prepare documented reports on current practices in the subject areas of concern.

This synthesis series attempts to report on the various practices without in fact making specific recommendations as would be found in handbooks or design manuals. Nonetheless, these documents can serve similar purposes, for each is a compendium of the best knowledge available concerning those measures found to be the most successful in resolving specific problems. The extent to which they are utilized in this fashion will quite logically be tempered by the breadth of the user's knowledge in the particular problem area.

FOREWORD This report should be of special interest to highway engineers responsible for design, construction, maintenance, materials testing and specificatiohs, research into pave-ment surface characteristics, and control of traffic operations. The report offers

Highway Research Board information on skid resistance requirements and measurement, pavement surface characteristics, control measures, and survey procedures.

Administrators, engineers, and researchers are faced continually with many highway problems on which much information already exists either in documented form or in terms of undocumented experience and practice. Unfortunately, this information is often fragmented, scattered, and unevaluated. As a consequence, full information on What has been learned about a problem is frequently not assembled in seeking a solution. Costly research findings may go unused, valuable experience may be overlooked, and due consideration may not be given to recommended practices for solving or alleviating the problem. In an effort to resolve this situation, a continuing NCHRP project, carried out by the Highway Research Board as the research agency, has the objective of synthesizing and reporting on common high-way problems—a synthesis being identified as a composition or combination of separate parts. Reports,from this endeavor constitute an NCHRP report series that collects and assembles the various forms of information into single concise docu-ments pertaining to specific highway problems or sets of closely related problems. This is the fourteenth report in the series.

Skid resistance is the force developed when a tire is prevented from rotating and slides along the pavement surface. The wet-skid resistance of a pavement can be a crucial factor in the "pre-event phase" of a traffic accident in determining potential damage to property and human life The rationale for level of wet-skid resistance requirements involves inputs that range from physical properties of materials through driver behavior to vehicle performance under normal conditions to emergency maneuvers that may lead to accidents.

The highly variable characteristics of pavement surfaces make wet-skid resist-ance difficult to predict because of the impacts of polishing by traffic, studded snow tires, climatic conditions, and the aging of pavement surfaces. Because friction force (the resistance measured or experienced when one body in contact with another is being moved or is to be moved) is dependent on contact between the tire and the pavement surface, any factor affecting either macro- or microtexture becomes sig-nificant. The Highway Research Board has attempted in this project to identify and describe those factors found to be most significantly related to pavement skid resistance. The report discusses these factors from the standpoint of the design and construction of skid-resistant surfaces, maintenance of pavement skid resistance, and measurement of skid resistance.

To develop this synthesis in a comprehensive manner and to ensure inclusion of significant knowledge, the Board analyzed available information, assembled from many highway departments and agencies responsible for highway planning, design, construction, operations, and maintenance. A topic advisory panel of experts in the subject area was established to guide the researchers in organizing and evaluating the collected data, and to review the final synthesis report.'

As a follow-up, the Board will attempt to evaluate the effectiveness of this synthesis after it has been in the hands of its users for a period of time. Meanwhile, the search for better methods is a continuing activity and should not be diminished. An updating of this document is ultimately intended so as to reflect improvements that may be discovered through research and practice.

CONTENTS

1 SUMMARY

PART I

2 CHAPTER ONE What Skid Resistance Is

Definitions: Skid Resistance and Friction Force Skid Resistance and the Driving Task The Wet Pavement State of the Art

5 CHAPTER TWO Skid Resistance Requirements

Skid Resistance Demands of Traffic Accident Frequency and Skid Resistance Driver Behavior and Frictional Needs of Traffic Vehicle Design Factors and Frictional Needs of Traffic Frictional Needs of Traffic and Highway Geometrics Definition of Minimum Skid Resistance Requirements Future Developments Spot Improvement - Sliperiness and Hydroplaning Cost-Effectiveness of Skid Resistance Control Variations of Skid Resistance Legal Aspects of Skid Resistance

13 CHAPTER THREE Methods of Skid Resistance Measurement

Locked-Wheel Trailer Methods Automobile Methods Portable Field Testers

-: Laboratory Testers Methods Employing Tires in Other Than the Locked-Wheel

Mode Correlation Between Testers

-25 CHAPTER FOUR Characteristiç of Skid-Resistant Surfaces

Texture Microtexture General Surface Properties Grooving and Porous Layers -

31 CHAPTER FIVE Deterioration of Skid Resistance

Wear and the Loss of Skid Resistance Polishing -

- - Bleeding of Asphaltic Pavements Compaction, Rutting, Particle Loss Studded Tires Contamination

Roughness - Cyclic Effects - Runoff Changes

38 CHAPTER SIX Design and Construction of Skid-Resistant

Surfaces

Design of Pavement Surfaces

Surface Aggregates

Aggregate Characteristics

Pavement Mixtures Bituminous Concrete Pavements Portland Cement Concrete Pavements

49 CHAPTER SEVEN Maintenance of Pavement Skid Resistance

Pavement Surface Inventory Modification of Existing Surface

Application of a New Surface

Other Corrective Measures

Summary

54 CHAPTER EIGHT Skid Resistance Control Principles

Measurement Comparison with Requirements and Standards

Estimates of Deterioration

Choices of Action to Be Taken

57 CHAPTER NINE Skid Resistance Surveys

Mandates

Survey Objectives and Scope

Survey Methods and Procedures

Data Acquisition Data Evaluation

61 CHAPTER TEN Lowering Skid Resistance Requirements

Speed Restrictions and Warning Signs

Education and Enforcement

Smoothing Traffic Flow

Removal of Hazards Preventing Water Accumulation

62 REFERENCES

PART II

66 APPENDIX A Summary of an Updated Inventory of Existing

Practices and Solutions to Slippery Pavements, 1971

ACKNOWLEDGMENTS

This synthesis was completed by the Highway Research Board under the supervision of Paul E. Irick, Assistant Director for Special Projects. The Principal Investigators responsible for conduct of the synthesis were Thomas L. Copas and Herbert A. Pennock, Special Projects Engineers.

Special appreciation is expressed to W. E. Meyer, Professor of Mechanical Engineering, The Pennsylvania State University, and William A. Goodwin, Dean for Research, The Graduate School, The University of Tennessee, who, as special consul-tants to the Advisory Panel, were responsible for the collection of data and the preparation of the report.

Valuable assistance in the preparation of this synthesis was provided by the Topic Advisory Panel, consisting of R. Clarke Bennett, Chief, Technical Development and Standards Division, Office of Highway Safety, Federal Highway Administration; William Gartner, Jr., Deputy State Highway Engineer (Opera-tions), Florida Department of Transportation; David C. Ma-

hone, Highway Research Analyst, Virginia Council of Highway Investigation and Research; Rolands L. Rizenbergs, Research Engineer Chief, Division of Research, Kentucky Department of Highways; Richard K. Shaffer, Research Coordinator, Pennsyl-vania Department of Transportation; George Sherman, As-sistant Materials and Research Engineer (Pavement), Materials and Research Department, California Division of Highways; and William P. Walker, Chief, Geometric Standards Branch, Highway Design Division, Office of Engineering, Federal High-way Administration.

H. A. Smith, Projects Engineer, and L. F. Spaine, Engineer of Design, both of the Highway Research Board, assisted the Special Projects staff and the Advisory Panel.

Information on current practice was provided by many high-way agencies. Their cooperation and assistance were most helpful.

SKID RESISTANCE

SUMMARY Skid resistance is the force developed when a tire that is prevented from rotating slides along the pavement surface. Usually, however, skid resistance is thought of as a pavement property. The question is what level of wet skid resistance should be provided, because on dry pavements the skid resistance is always high. There are various possible approaches to a rationale on skid resistance requirements. Inputs range from driver behavior and vehicle performance under normal condi-tions to emergency maneuvers that may lead to accidents. Accident records are highly useful for identifying roadway sections at which increased skid resistance can pay large dividends. Setting standards for skid resistance levels, aimed at elimination of wet-weather skidding accidents, is a complex task. Success or failure of any measures taken, however, ultimately can be judged only by the reduction in skidding accidents. This points up the importance of good and complete accident records.

Hydroplaning is a phenomenon that occurs when there is an abnormally thick water layer on the pavement, or the speed is high, or both. Once it does occur, pavement surface characteristics lose significance, but the onset of hydro-planing can be moved to higher speeds by providing heavily textured surfaces, by grooving, and by using porous surface courses.

Several means are available for measuring skid resistance. The locked-wheel trailer method in accordance with ASTM Method E 274 is the most generally accepted method. It does not directly predict vehicle performance. It does rank pavements in the same order as they would, be found to be slippery by traffic, particularly if measurements are made at prevailing traffic speeds. Other methods of measuring skid resistance, although individually having some particular ad-vantages, do not necessarily correlate with the E 274 method; but when used for inventory work and analysis of high-accident locations, they can provide valuable inputs in evaluating pavement slipperiness. However, caution is advised in the use of portable friction testers, because these cannot usually be employed to appraise skid resistance for traffic on high-speed facilities.

The basic characteristics of pavement surfaces that make them skid resistant are discussed in detail. A distinction is made between macro- and microtexture. The, former controls the escape of water from under the tire, and, thereby, the loss of skid resistance with increased speed; the latter controls the actual contact between tire and pavement by penetrating the thin ifim of water that 'is not removed by the tire. Microtexture controls the level of skid resistance. Means and methods for measuring both types of texture are des'cribëd. None of the available techniques, although useful, are fully satisfactory; otherwise it' would be possible to predict skid resistance from them and friction measurements wouLd be required for verification only. Contamination, pavement grooves, and surface course porosity are also discussed. Under certain circumstances, these can blot out the effects of macro- and microtexture.

Skid resistance is a highly variable characteristic of pavement surfaes. Of most concern is the reduction of microtexture by the polishing, action of traffic. Macrotexture also changes with traffic exposure. It is reduced or increased

2

by attrition, wear, compaction, and other effects. Laboratory tests are available for predicting polishing and wear of aggregate and pavement specimens. Although these, techniques can provide valuable guidance, they cannot fully duplicate the processes that take place in the field. Therefore, field studies must remain an integral part of any skid resistance management program.

Among the factors that can lead to deterioration of available skid resistance are studded tires. Studs cause excessive pavement wear and influence the skid resistance available to traffic in several ways. It is difficult to balance the safety improvements that studded tires provide on ice against their detrimental effects.

The design and construction of skid-resistant surfaces requires a thorough knowledge of the factors that influence skid resistance of pavement surfaces in service. In the design stage, consideration must be given to the role of natural and synthetic aggregated and other pavement constituents. During construction, mix design and construction techniques—and their control—must be considered.

It is not possible, or at least not economically feasible, under most circum-stances to build in adequate skid resistance for the design life of the total pave-ment section. Therefore, numerous maintenance techniques have been developed.. These range from modification of existing surfaces to the application of new surface courses. The same basic requirements exist for these techniques as for new construction, but there is a considerably greater choice of options. No one technique, however, is best for all pavements and conditions, particularly when

economics is considered. Skid resistance control principles involve measurement, comparison with'

requirements, and corrective measures. There are certain federal mandates that include the necessity of investigating whether the skid resistance at high-accident locations is adequate. These also include systematic and periodic surveys of the entire highway system of a state to prevent the development of dangerously slippery conditions. Such skid resistance inventories also can be powerful tools in the planning phase of skid resistance management programs.

If traffic flow were perfectly smooth and all drivers would and could adjust their speed to the limitations of the roadway, skid resistance requirements would be rather modest. Improvements in geômetrics, signing, enforcement, education, etc., can ease the requirements.

CHAPTER ONE

WHAT SKID RESISTANCE IS

In the "pre-event phase" of a traffic accident, various factors determine whether or not potentially damaging energy exchanges will actually take place. It would be folly to expect that human behavior can be so modified that potential dangers or hazards are always anticipated soon enough not to require the utmost in performance of the vehicle-road system.

A simple example is the emergency stop to avoid a collision. It would, with rare exceptions, not be an emer-gency stop if the driver had anticipated the impending collision earlier. When he finally acts, he more likely than not applies the brakes with all his might. He is now at the mercy of the vehicle brakes and of the ability of the tire-pavement interface to accept the retarding forces developed

by the brakes. When the pavement is wet, even a poorly maintained braking system is likely to be capable of developing more braking force than is needed. Thus, in an emergency stop the wheels lock up, the vehicle skids along, and its deceleration becomes entirely a function of the friction between the tires and the pavement.

During an emergency stop the last chance for avoiding an accident or reducing the speed at impact resides with the tire-pavement friction. This is true whether or not the pavement is dry, wet, or covered with ice. On dry pave-ments available friction is normally high. The urgent problem is that of doing something about wet pavements. In the eastern and central United States pavements are wet at least 15 percent of the time and the wet-pavement accidentrate (that is, the number of accidents per million vehicle-miles) is more than twice the dry-pavement rate. The difference is not entirely chargeable to the reduced skid resistance of the wet pavements; For instance, at night both sight distances and visibility are reduced, but, conservatively, about one-half of the wet-pavement acci-dents can be assumed to involve skidding as either the primary or the secondary cause.

Skidding accidents are not only those in which, as in the example of the emergency stop, a vehicle slides along the pavement with all wheels locked, but also those in which only the rear wheels lock, causing the car to spin around or a tractor-trailer to jackknife, or when a vehicle "breaks away" in a curve taken. too fast. Some of these mishaps can be prevented or softened by design changes on the vehicles and their components, but in the final analysis it is always tire-pavement friction that is the limiting factor.

As the term "tire-pavement friction" indicates, the fric-tion that can be developed in a given case involves two components—the tire and the pavement—and in the critical case, a third one—water. Improvements are possible by improved control of all three, although there are limits for all and compromises must be made, if not for tech-nological reasons, then because of cost considerations.

Rain cannot be prevented, but runoff from the pavement can be facilitated. Tire friction can be improved, but tires must be designed to satisfy numerous other requirements. They must withstand impact, resist cuts, run in desert temperatures at sustained high speeds, act as springs, pro-vide precise control in cornering, not wear too fast, and do many other things besides. Pavements must be struc-turally capable of withstanding heavy truck traffic and seasonal stresses, they must not wear too fast, and yet they must also maintain good frictional properties in con-tact with a wide variety of tires and in the presence of water.

Because present wet tire-pavement friction is often less than satisfactory, both tires and pavements should be im-proved. Changes in tires are probably coming on the market as fast as advancing technology permits because they are being manufactured by one of the most competi-tive industries. Additionally, tire manufacturers are being prodded into improvements by the automobile industry, and lately by the National Traffic Safety, Administration.

Tire improvements are available rather fast because the average life of a tire is probably not more than two years.

Pavements, on the other hand, are expected to last 15 to 20 years before major structural repairs are necessary. Pavements are in the public domain and therefore are subject to the restraints of and competitive demands for public funds. Next to congestion, pavement slipperiness is the greatest sin a public agency can, in the eyes of the public, permit to befall highways and streets. Indeed, when skidding accidents occur, the public is more likely to blame the pavement than anything else. For these reasons, con-trol of pavement slipperiness has extremely high priority in the continuing campaign to reduce traffic accidents.

Nomenclature

The symbols used herein are defined where they first appear in the text and are also listed here for the convenience of the reader:

D = degree of curvature (= 5729.6/R, with R in ft); d = deceleration; e = superelevation of curve, in ft/ft; F = force in the tire-pavement interface plane; f = friction factor;

Id = friction factor during deceleration; fr = coefficient of rolling resistance; Is = side friction factor; g = acceleration of gravity; L = load normal to the tire-pavement interface plane

(or axle load); p = grade, in percent; R = radius of curve (or curvature of vehicle path); s = slip, in percent; v = speed;

W = gross vehicle weight; a = angle of grade (tan a = p, with p in percent);

= coefficient of friction; = angular wheel speed; = angular wheel speed in free rolling;

BPN = British Portable Tester No., per ASTM E 303; and

SN = Skid No., per ASTM E 274.

DEFINITIONS: SKID RESISTANCE AND FRICTON FORCE

Skid resistance is the force developed when a tire that is prevented from rotating slides along the pavement surface. More commonly, skid resistance is thought of as a pavement property; it is the antonym of slipperiness. It is in this sense that skid resistance is used in this report. The term, therefore, does not have a precise meaning, and permits one to describe a pavement surface in the most general way.

Friction force is the resistance measured or experienced when one body in contact with another is being moved or is to be moved. It is dependent on the contact area and is, therefore, not suited for describing the character of a friction pairing. In mechanics, "coefficient of friction" (,z) is used, or

(1)

4

in which F is the frictional resistance to motion in the plane of the interface and L is the load perpendicular to the interface. The coefficient is a useful term when all conditions can be precisely defined.

When a tire rolls, slips, or slides on a pavement, how-ever, various conditions influence the amount of friction developed. Most are difficult to describe and measure. This is particularly so when water is present at the inter-face. In this case the preferred term, in place of coefficient of friction, is "friction factor" (1)

f=F/L (2)

It is wrong to say that a pavement has a certain friction factor (or coefficient of friction), because friction always involves two bodies. It is even imprecise to say that a particular tire on a given pavement produces a certain fric-tion factor, unless forward or sliding speed, inflation pres-sure, load, temperature, water-film thickness, and other details are specified. To overcome the resulting communi-cation problem, standards have been developed that pre-scribe all variables that influence the friction factor.

The prime example for such a standard is ASTM Method E 274. Measurements made in accordance with it are reported as skid numbers (SN):

SN= 100 1=100 FIL (3)

in which F is obtained in a strictly defined manner by sliding a locked, standardized tire at a constant speed (usually 40 mph) along an artificially wetted pavement. The term skid number (or SN) should not be used in connection with any other skid resistance measurements except those made at the same speed in accordance with ASTM E 274.

SKID RESISTANCE AND THE DRIVING TASK

Skid resistance was defined earlier as a means of charac-terizing a pavement surface. The numbers used for this characterization (for instance, skid number) do not tell directly what friction is available to traffic using the pave-ment: Although skid numbers refer to a locked tire sliding on a wet pavement, they are measured with a special tire that resembles, but is not like any tire in actual use. Even if it were, it still would not represent all tires on the road.

Although tires in a particular group may not differ greatly from each other when new, they change as they wear. On dry pavements they develop more friction after the tread is worn away because more rubber comes into contact with the pavement. When the pavement is wet, the friction diminishes with tread wear because it is more diffi-cult for the tire to expel the water from the contact area. The point is that the same friction factor is not available to all traffic participants at all times, even if the skid resistance of the pavement is constant.

Skid resistance, measured in accordance with ASTM Method E 274, is an imitation of an emergency stop, some-thing that does not happen very often (although skid resistance is likely to be crucial when it does). The driver more often needs friction to carry out ordinary maneuvers in which he accelerates, decelerates, or steers a curved path. Often he accelerates or decelerates while not traveling a

straight course, as when passing or lane changing. He does all these maneuvers while the wheels are turning.

Can skid resistance obtained in the locked-wheel mode tell anything about the friction factors available for other than stopping maneuvers? Strictly speaking, no; but in a qualitative manner, yes, though not with very good pre-cision. If the skid number of one pavement is higher than that of another one, the friction factors available for vari-ous maneuvers are almost certainly also higher on the first pavement than on the second. The percentage of the difference is not likely to be the same for all modes, however.

If the friction mechanism of pneumatic rubber tires on pavement surfaces were completely understood and if all variables that influence friction could be measured, or their magnitude predicted, it would be possible to predict from measured skid numbers the performance of different tires operating in the various modes. This is, however, not yet possible.

THE WET PAVEMENT

It has been pointed out that even a single tire has variable properties that affect its frictional behavior. The same is true for pavements: they change with time, though rather slowly. The most notable change is the reduction of the skid resistance by the polishing action of traffic. There are other processes that make pavements more slippery: more of the binder may work to the surface or texture may be diminished, thus robbing the water trapped under the tire of escape paths.

On the other hand, some processes work toward an in-crease in skid resistance. Freezing and thawing, the action of deicing chemicals, of antiskid materials and tire chains, and other influences, tend to roughen pavements to varying degrees and in varying ways. On the whole and over the long haul, however, skid resistance decreases, eventually reaching a mean minimum level.

The important point is that, at a given time in the life of a pavement surface, the skid resistance has a momentary value that may be higher or lower than one measured some months, or even weeks or days, earlier or later. Because traffic-generated effects are the most potent ones, it also follows that a pavement that has been in service will not have uniform skid resistance. Where traffic is well chan-neled the wheel paths may show considerably lower skid resistance than the areas near the shoulders or between the wheel paths.

Wear of a pavement surface does not necessarily signify a lowering of the skid resistance. Although individual aggregate particles may wear as they are polished (or vice versa), this is not necessarily so for the pavement surface as a whole. Brittle aggregates may, as fragments are re-moved by traffic and other stress-producing agents, expose fresh surfaces to the passing tires. Similarly, wear may open new water escape passages in the surface. Similar processes may take place as some binders become brittle with age and exposure.

Most of these processes have little effect on the per-formance of dry pavements, but they can bring about

significant .changes in skid resistance when the same pave-ments are wet. Thus, wet skid resistance is likely to be variable in time as well as with location.

STATE OF THE ART

Skid resistance of pavement surfaces in the locked-wheel, straight forward mode can be measured today with con-siderable accuracy by several methods. If such measure-ments are made at more than one speed, with different water-film thicknesses, in and out of the wheel path, etc., a fairly complete and objective description of the skid resistance characteristics of surfaces can be obtained. Skid resistance so measured can be related empirically to the needs of traffic and drivers.

Therefore, such measurements have become an im-portant tool in the management of pavement surfaces. They can serve to identify excessive slipperiness, to plan maintenance schedules, to monitor construction practices, to explore new types of surface construction and materials, and to serve other purposes. The technical knowledge is

available to insure proper interpretation of results of skid resistance tests. For many purposes even simplified tests and test schedules can give valuable guidance.

The knowledge also is at hand for constructing skid resistant pavements, overlays, and surface treatments. In this area, even more than in the field of measurements, empirical information must still be heavily relied upon, particularly because nearly all surfaces lose skid resistance as time and traffic pass, so that the final performance of a particular surface design becomes evident only after several years.

It is customary to use locally available aggregates for surface courses because this keeps construction costs down. As the importance of maintaining adequate skid resistance is more commonly recognized, local materials are not necessarily the most economical ones when they make resurfacing a frequent necessity. Excessively slippery pave-ments are no longer acceptable or defensible, particularly because the means for correcting or preventing slipperiness are now available.

CHAPTER TWO

0 0

SKID RESISTANCE REQUIREMENTS

A crucial question for the highway engineer is the level of skid resistance that should or must be maintained. There is no simple answer. No federal, AASHO, ASTM or other standards exist as yet; nor is there general agreement as to how to arrive at numbers for specific situations.

The problem is being attacked in different ways. One can either determine what skid resistance is needed to permit normal traffic to move safely and then bring the highway system to the required levels, or, if available funds do not permit this, one can provide the highest practical skid resistance level at the most critical locations first. In either case, however, it is necessary to have an objective scale with which to measure failure or success of the chosen approach.

SKID RESISTANCE DEMANDS OF TRAFFIC

For a vehicle to follow a prescribed or desired course the tires must be capable of developing definite friction factors. These differ, and vary with speed, direction, vehicle and tire type, etc., and, most importantly, with the maneuvers that drivers impose on their vehicles (changes in direction, rate of acceleration, etc.). Only in rare instances does the vehicle as such set the limits (one example: a truck as-cending a steep grade).

Thus, one must know what drivers do. Only a few attempts have been made to determine this in terms appli-

cable to the skid resistance problem. In part this is so because it is not clear how driver requirements are to be defined. Is one concerned only with maneuvers in smoothly flowing traffic, or also with emergency maneuvers?. Where is the borderline between the two areas? What is an emer-gency maneuver?

No rigorous answers to these questions are possible and arbitrary delineation is necessary. The problem is more easily handled if the minimum skid resistance requirement is first defined for normal, smoothly flowing traffic. This means that there is, or should be, a minimum value below which skid resistance must not drop—anywhere, at any time. It does not mean that certain sites may not require a higher, perhaps considerably higher, skid resistance to accommodate unusual or emergency maneuvers if these occur frequently at these sites.

ACCIDENT FREQUENCY AND SKID RESISTANCE

This may seem a roundabout way of arriving at reasonable guidelines when the objective is to provide sufficient skid resistance so that skidding accidents are prevented. Hence, the, question is what information can be obtained from accident records. The highway accident problem is of truly alarming proportions; however, statistically, accidents are rare events. In addition; because the possible causes are quite numerous, it is difficult to determine the contri-

6

bution of any one cause unless a representative sample of accidents is subjected to detailed analysis and reporting is accurate and uniform.

These requirements have not been met in the past, at least not in combination. A few isolated studies have been made that permit conclusions to be drawn, but their general validity remains in doubt. One such example, represented in Figure 1, is from a special survey of 150 miles of rural highway in Great Britain (1). Whether the regression line is straight or flattens out toward the right is still uncertain. Neither is it certain that the 'figure is applicable to all roads in Great Britain, or to any in the United States. (Note that Fig. 1 uses "sideway friction factor," not skid resistance. The numerical values are not directly comparable.)

Actually, Figure 1 is not very useful. It only shows that a higher friction factor (or skid resistance) is better. It gives no clue as to what a reasonable minimum value might be. In addition to the skid number, one should have in-formation on the speeds at Which the accidents Occurred, whether the speeds were reasonable or not, visibility data, road geometry, etc. Consequently, attempts have been made to glean information from the ratio of the number of all accidents on wet pavements to the total number of acci-dents in a given district (for this purpose accidents on ice and snow should always be excluded from the total). Figure 2 (2), from German studies, shows this ratio ex-pressed as a percentage against the mean friction factor of the sections that the accident reports permitted to be iden-tified. Here the mean wet accideát rate tends to level out at a constant value above a certain friction factor (some-what less than 0.4). this may be taken to mean that above a friction factor of 0.4 pavement slipperiness is not a

significant accident cause. The scatter of the data points in this region is, however, noteworthy. It indicates the possibility that even at a friction factor of 0.55 slipperiness may still be excessive under certain circumstances.

Thus, accident statistics would not seem to lend them-selves readily to the derivation of a minimum permissible skid resistance value. The inadequacy of general accident data is illustrated by the finding of McCullough and Hankins (3). In their correlation with skid resistance it made little difference whether they used all accidents, wet road accidents, or skidding accidents. However, accident rates for specific sites in conjunction with skid resistance data can be a useful means for determining whether or not low skid resistance is a major contributory cause. If of two similar sites the one with the higher accident rate has lower skid resistance, that value is obviously too low.

Many skids occur that do not result in accidents. If the total number of skids could be obtained, this would provide a much better measure of the adequacy of available skid resistance than does the number of accidents, because the causes of no-accident skids are usually much less obscured by extraneous effects than are those of skids deduced from accidents.. Conversely, the fact that a skid did not occur, or was not reported, does not mean that the skid resistance was adequate. For instance, with automatic brake control systems wheels do not lock; consequently, 'the vehicle leaves little evidence at an accident site that deceleration has been inadequate due to low skid resistance. It would appear, therefore, that other methods must be resorted to for deriving the lower permissible limit for, pavement skid resistance, even though the adequacy of any skid resistance management methodology will eventually become evident in the number of accidents that do or do not occur.

SIDEWAY FRICTiON FACTOR

Figure 1. Decrease of wet-skidding accidents with improved friction factor.

z w

z 0 U)

z Li

C, C)

LA. 0 Li (9.

z Li 0 (Li

NUMBER OF ACCIDENTS • 20

LOCKED WHEEL FRICTION FACTOR Figure 2. Wet-pavement accident rate as a function of friction factor.

DRIVER BEHAVIOR AND FRICTIONAL NEEDS OF TRAFFIC

Observation of vehicles approaching a STOP sign shows that the deceleration used by individual drivers varies over a certain range. From the observed decelerations the friction factors required to produce these decelerations (f,) may be obtained by means of the (simplified) relationship:

f=d/g (4)

in which d is the vehicle deceleration in ft/ sec2 and g is the acceleration of gravity (32.2 ft/see2 ). The required friction factor would be found to vary with distance from the STOP sign, approach speed, vehicle, driver, and other factors.

Obviously, a driver approaching a STOP sign at high speed in a situation in which the sign can be seen only from a limited distance, will make considerable demand on the friction level of the pavement. In fact, this driver may have to make what amounts to an emergency stop. This example illustrates the difficulty of separating normal from emergency maneuvers. The approach speed does not have a definable maximum nor has the distance at which the driver perceives the STOP sign a definable minimum. However, "normal traffic" might be defined as that for which the observed decelerations are those chosen by the driver at his discretion, rather than those dictated by events or agencies not under his control. Drivers tend to select the deceleration that is comfortable or tolerable to them. According to Tignor (4), about 0.25g is considered com-fortable longitudinal deceleration, whereas 0.35g is ac-cepted as tolerable by passengers, but used by the driver only reluctantly. Table 11 of NCHRP Report 37 (5) gives another set of drivers' preference ratings, which largely agree with Tignor's values.

In negotiating a curve the required side friction factor (18) is

v2 / R /8 = ( 5)

in which v is the tangential velocity, R is the radius of the path, and g is the acceleration of gravity (all in compatible units) with the sum of the friction forces at the four wheels equalling the centrifugal force acting on the center of gravity of the vehicle. According to the same table, drivers tend to prefer, for curve travel, lower radial or lateral acceleration than they do in straight-line deceleration: anything above 0.30g laterally is thought to be excessive (0.30g corresponds to about 10 ft/see2 ).

One can readily appreciate that such values are highly subjective and vary with vehicle and seat design, as well as with external factors. More objective information is ob-tained from observations of traffic. There are several pub-lications that deal with driver performance on curves (Taragin (6), Tignor (4), and others). Curves have at-tracted more attention than situations calling for straight-line deceleration, because the information is needed for the development of design rules. The maximum normal lateral acceleration used by all but a few vehicles is re-ported by several investigators to be approximately 0.3 5g.

Such data permit estimating general minimum friction

factor levels of skid resistance by means of the cited equa-tions. Usually the reported accelerations are based on 95th percentile samples. These may be used, but it might be argued that, ideally, roads should accommodate safely all drivers and all vehicles. Yet it may constitute an unrea-sonable burden to do so. To settle this question, more and better data are needed than are presently available. The results of NCHRP Project 1-12 (7), which has been completed recently, should do much to close this gap. The project concerns the development and use of methods for determining the skid resistance requirements of traffic un-der a variety of typical conditions and deriving, from the data, recommendations for minimum skid resistance re-quirements.

It should be recognized, however, that data of this type reflect human behavior and that human behavior is variable, among individuals, among regions, and in time. For instance, mean speed of traffic on straight, open high-ways is thought to go up 1 mph per year (8). The extent to which this change is reflected in higher speeds on curves and other maneuvers is not well documented. One set of observations is presented as Figure 3 (from Statens Väginstitut (9)). There is little question that all vehicle, road, and traffic improvements tend to increase all speeds. Therefore, if traffic observations are used to derive guide-lines and standards for the skid resistance of pavements, the basis must be not only thoroughly documented once, but also at least spot-checked periodically thereafter.

There is no evidence that average speeds differ signifi-cantly whether the pavements are dry or wet. It is, how-ever, likely that drivers who know a particular stretch of road to be excessively slippery when wet drive more care-fully by allowing greater headway, cornering more slowly, etc. Whether this has sufficient influence on accident rates to demand different treatment of similar roads carrying commuter as opposed to long-distance traffic is not known.

a E

0 w w Q. (I) z U

RADIUS OF.cURVATURE , ft Figure 3. Mean speed of passenger cars as a function of curva-ture of two-lane roads, for two different years.

8

For deriving general skid resistance requirements such variations may not be important, provided a large and diverse enough data base is used.

VEHICLE DESIGN FACTORS AND FRICTIONAL NEEDS OF TRAFFIC

Vehicles equipped with fully automatic brake control cannot go into a skid in straight-line deceleration, no matter how slippery the pavement is. One U.S. automobile is now being offered with a device that prevents wheel spin during acceleration. No device has yet been developed that prevents breakout in cornering. Driving aids, however, do not eliminate the need for providing pavement surfaces with some minimum skid resistance. Automatic brake control may prevent skids and maintain steerabiity of a vehicle, but it does not change stopping distance by a significant amount whether the pavement is slippery or not. Such devices may actually place a greater burden on the skid resistance level of pavements than were apparent before.

In an emergency application of brakes on a wet pave-ment, a vehicle without automatic brake control is essen-tially uncontrollable. The driver, in panic, is likely to apply the brakes so hard that all wheels lock. With wheels locked, the vehicle will simply continue in the direction in which it was originally moving or it will spin around if any moment acts on it. Such a moment may result from differences in the sliding friction of the tires, local differ-ences of the pavement skid resistance, differences of the instants at which different wheels lock, etc. When, how-ever, an automatic brake control compensates for vehicle deficiencies, most of the burden of accident avoidance shifts to the pavement. Because the vehicle remains con-trollable, the pavement will be expected to be capable of permitting evasive maneuvers, not just straight-line de-celeration.

Because automatic controls for vehicle performance and handling are still far from common, it will suffice here to consider only vehicles without them. NCHRP Report 37 (5) has done so in some detail. It examined, for instance, the friction factors that would be required to accommodate the maximum acceleration of various types of automobiles. This is not of much practical value

'because the horse-

power-to-weight ratios of many of today's automobiles are such that it is possible to induce wheel spin even on dry pavements. More to the point would be an examination of the capability of certain truck configurations to travel, or to start from a standstill, on steep upgrades when the pavement is wet.

Braking performance of vehicles presents a similar pic-ture. Brakes must be designed to meet federal safety standards, with the result that brakes are capable of being locked under most conditions, even on dry pavements. Thus, to accommodate available vehicle braking capabili-ties, wet pavements would ideally be as skid resistant as dry ones.

On most present-day vehicles the brake torques devel-oped at the individual axles have a fixed relationship to each other. The designer decides what the distribution should be, but in service this distribution is not always

optimum (that is, all wheels do not lock at the same de-celeration), if only because payload and payload distribu-tion change. But even with constant payload distribution the dynamic axle load distribution changes with decelera-tion. For current passenger cars the brake system is so laid out that the front-rear distribution of brake force is equal at a deceleration of 0.35 to 0.40g. At less than this deceleration the front wheels lock first (this happens on ice); above it, the rear wheels do. It is interesting that designers choose as the design point the deceleration that requires a skid resistance of 35 to 40 SN (see Eq. 4 for the relation between friction factor and deceleration). The same skid resistance range is most frequently identified as the value needed or used by normal traffic.

If the skid resistance permits higher deceleration (as on dry pavements), the rear wheels of current automobiles lock first or only. With severe brake applications this is of little concern, because front and rear wheels lock up at essentially the same moment. With less severe braking on pavements with skid resistance slightly above that for optimum braking effort distribution, the danger of rear-wheel lockup is greatest. This, however, is no argument against high wet-skid resistance, but merely an indication that all safety improvements cannot come from the highway side. Systems now available on some passenger cars prevent further buildup of the braking pressure to the rear wheels after a preset deceleration has been reached, thereby pre-venting the rear wheels from locking before the front wheels do.

In cornering and travel through curves, somewhat simi-lar considerations apply. Cornering capability is limited either by the breakdown of the frictional coupling between tire and pavement or by overturning. Either limit is a function of speed and consequently under control of the driver. Thus, vehicle design per se is not a useful criterion for establishing minimum skid resistance requirements.

Tires can be considered a vehicle component and the same reasoning applies to them as does to the vehicle as a whole. Obviously, of two identical vehicles, but one having new, the other bald tires, the one with the bald tires is more likely to go into a skid. Vehicles with bald tires no doubt add a disproportionate share to the number of skidding accidents, but this fact is of little use in the derivation of standards for the skid resistance of pave-ments. The laws of many states prescribe minimum tread depths to which tires may be worn, but because this minimum (MG in.) is met at inspection time, does not mean that in practice tires with the tread completely worn off will not be encountered. More stringent tread depth re-quirements and stricter enforcement of these and other vehicle safety requirements are highly desirable develop-ments. Unless such improvements significantly change the limits of safe vehicle performance, it is not likely that the demands on pavement skid resistance are affected notice-ably.

Vehicle improvements are likely to result in higher speeds. The trend toward higher mean traffic speeds supports this contention. Only when comfort considerations or other factors become overriding, will this trend abate. There are documented instances in which resurfacing has

led to higher wet-weather accident rates despite improved skid resistance. Resurfacing caused other improvements of the roadway, resulting in higher traffic speeds. But other hazards, such as intersections, were not removed and acci-dent rates increased. It is also important to remember that accidents at higher speeds tend to be more severe, whether the speed increase is the result of improvements to the highway or to the vehicle.

FRICTIONAL NEEDS OF TRAFFIC AND

HIGHWAY GEOMETRICS

The torque transmitted by the drive wheels to the pavement is the sum of that needed for overcoming the grade, the rolling resistance, and the air drag (5). For accommodat-ing available driving torque, upgrades demand higher fric-tion factors than level tangents, but in most cases the demands are quite low even if the grades are steep. How-ever, when only one or two of several axles are driving (as, for instance, in the case of a tractor-semitrailer com-bination with trailer) the available friction at the driving wheels in wet weather may well become limiting on almost any grade if the engine power is high and a multispeed transmission is used.

Because such vehicles ascend the steeper grade only at low speeds, air resistance may be neglected. The required friction factor is then

f = W (f,+p)/L (6)

in which W is the gross vehicle weight, L is the load acting on the drive axle(s), Jr is the coefficient of rolling resist-ance (assumed to be 0.01), and p is the grade (percent). Figure 4 shows .the friction factor that must be available for a combination consisting of a high-gross-weight tractor-semitrailer-trailer combination for climbing various grades. Such combinations can be in difficulties on wet days on steep grades. Engine power limits the steepest grade that the combination can negotiate fully loaded, but before this limit is reached the drive wheels may begin to slip because

C.,

z 0

C.)

U.

GRADE, pct

Figure 4. Effect of load distribution on the friction factor needed at the drive wheels of a truck-semitrailer-trailer com-bination (3424) for climbing various grades.

the load on them is too small, as when the semitrailer is being run empty (a prudent driver would not drive with such a load distribution, but not alIdrivers are prudent). Generalizations are difficult to make because legal limits for sizes, gross weights, and axle loads vary greatly, as do actual dimensions and loads. From Figure 4 it is seen that for an 8-percent grade a friction factor of 0.6 must be available if trailers are to be permitted, but minimum axle loadings are not prescribed or enforced.

Downgrades require a friction factor greater than that necessary on the level for obtaining a given deceleration. The deceleration is (10)

d=(f cos a— sin a) p (7)

in which a is the angle of the grade (p = tan a, with s in percent). This works out to requiring an increase of about 0.01 in the friction factor for every 1 percent of downgrade. The equation does not take into account the change in. wheel load caused by the grade. With vehicles having a low center of gravity (passenger cars) this is not significant, but it can be so with other vehicles, particularly when the rear axle is loaded lightly to begin with.

Grades thus do have higher skid resistance requirements than do level tangents, but the needed step-up is difficult to define because it is primarily determined by nonaverage vehicle types or load distributions.

Most states design horizontal curves in accordance with AASHO policy (11) or a close approximation thereof. Figure 5 shows the minimum radius as a function of design speed for the extremes of superelevation for which the policy gives data. The curves represent

R=v2/15 (e+f 8 ) (8a)

DESIGN SPEED, mph

Figure 5. Minimum radii of horizontal curves as a function of design speed (11).

10

in which

e = superelevation, in ft/It; = side friction factor;

v = design speed, in mph; and R = radius of the vehicle path, in.ft

brake on curves. Unless there is a frictional reserve, di-saster is inescapable. Also, despite the fact that drivers tend to flatten out curves by making the radius of travel larger than the geometric one, on occasion they need or choose smaller radii, as in passing or avoidance maneuvers. Design equations must take these eventualities into account.

The side friction factor in Eq. 8 is specified as changing linearly from 0.16 at 30 mph to 0.11 at 80 mph. The radius of the vehicle path and the radius of the center line of the innermost lane are assumed to be identical. The equation is based on the assumption that all points of the vehicle experience the same lateral acceleration.

The intent of Eq. 8 is to serve as a guide to the design of curves providing lateral accelerations that drivers con-sider comfortable. Thus, I in the equation is not to be taken as a suggested safe value. The side friction factor in Eq. 8 is based on constant speed and no change of radius while negotiating a curve. This, of course, assumes vehicle and driver characteristics that are far from realistic.

Furthermore, traffic speeds are not necessarily equal to or lower than the design speeds. On the contrary, in keep-ing with general trends, traffic speeds are likely to be higher than the design speeds. Rewriting Eq. 8a gives

18= (v2/15R) —e (8b)

which indicates that for a given curve the required side friction factor increases with the square of the speed.

The conservative side friction factor of the AASHO design equation gives a larger radius for a given speed than if a larger factor were used. Thus, even if the curve is traveled at a speed somewhat in excess of the design speed, the required side friction factor is not likely to ex-ceed the available friction, as long as the pavement surface has acceptable skid resistance. For design purposes, the assumption must be made that skid resistance may, as a consequence of pavement polishing (and in northern lati-tudes because of ice), drop well below what is considered normal.

It is important to recognize that superimposed on the frictional demands dictated by road geometry are those made by driving maneuvers. Drivers will, and have to,

TABLE 1

TENTATIVE INTERIM SKID RESISTANCE REQUIREMENTS FOR MAIN RURAL HIGHWAYS

RECOMMENDED MINIMUM SN l, TRAFFIC

SPEED

MEASURED AT MEASURED AT

(MPH)

TRAFFIC SPEED 40 MPH

30 36 31 40 33 33 50 32 37

60 31 41 70 31 46

'From Table 18, NCHRP Report 37. These values are recommended for main rural two-lane highways. For limited-access highways lower values may be sufficient, whereas certain sites may require higher values.

SN = skid number, measured according to ASTM Method E 274.

DEFINITION OF MINIMUM SKID RESISTANCE REQUIREMENTS

NCHRP Report 37 (5) discusses in detail the problem of defining minimum skid resistance requirements once it has been established what friction factors the traffic needs dictate. It must first be decided how skid resistance is to be measured. At present the customary method employed is ASTM E 274, which prescribes measurement of the friction of a locked wheel equipped with a specified tire.

Because, however, locked-wheel pavement skid resistance measurements vary significantly with speed, it is necessary to specify the speed at which the measurement is to be made. ASTM Method E 274 prescribes 40 mph as the standard speed. Yet the friction factors most frequently demanded by traffic also change with speed.

From these and other considerations NCHRP Repprt 37 derives SN =37 as the minimum permissible for standard imain rural highways. It is impotnt iiote that this

to measurement at 40 mph, although it is assumed that the mean traffic spei5mph. When traffic speed is higher, the SN (measured at. 40 mph) should be higher (see Table 1). When methods other than ASTM E 274 are used for determining skid resistance, appropriate correlation equations must be applied.

The values suggested by NCHRP Report 37 have so far remained what they are—recommendations. Although they have been derived rationally on the basis of available data and information, various jurisdictions adopted or pre-scribed, earlier or since, minimum skid resistance values (specified as SN or in other terms) based on their own considerations, frequently on intuitively condensed experi-ence, and sometimes with provisions for higher values at specified locations. Interestingly enough, such standards and guidelines agree, on the average, with the recommen-dations of NCHRP Report 37. The Federal Government has not yet issued a standard, but recommends NCHRP Report 37 as a guide (13, 110).

FUTURE DEVELOPMENTS

It is more than likely that the documented needs of traffic will be the basis for future standards. Therefore, the re-sults of the current NCHRP Project 1-12 will be of major importance. Analysis of accident statistics must, however, play at least an auxiliary role, not only as a means of testing the validity of the recommended minimum values but also to answer the question as to what extent more severe standards may be required for special situations or locations. (In addition, accident analysis will continue to serve as a surveillance tool for spotting locations requiring remedial steps whether or not their skid resistance meets any present or future standard.)

Determinations of the needs of traffic are generally based

11

on data obtained in dry weather, when the available skid resistance is well above that needed by normal traffic. This is justified and necessary because traffic characteristics appear not to change significantly with the change from wet to dry weather (14). Therefore, the ultimate aim (whether attainable or not) of pavement surface design must be to make it insensitive to the presence of water.

It is, however, debatable whether standards should be based exclusively on the needs of normal traffic, or to what extent emergency maneuvers should be considered in their development. An "emergency maneuver" is any situation in which the driver is forced to make a maneuver not of his choice. It may range from mild to severe. The problem is to define the maneuvers that a pavement should be expected to accommodate.

It should be recognized that high pavement skid resist-ance imposes its own restraints. Articulated vehicles will no longer jackknife; but loads may shift, or on a curve the vehicle may overturn instead of spinning out. For these reasons, accident reports must remain a continued concern of all those responsible for developing standards or for controlling pavement skid resistance.

Accidents are, however, just that—accidents. For a skidding accident to occur, only one ofmany factors needs to exceed its critical value (in the plus or minus direction): speed, acceleration, deceleration, steering angle, tire condi-tion, pavement condition, driver judgment, driver reaction and action, etc. Yet even though a skid may occur, it does not necessarily lead to an accident, perhaps because the skidding vehicle did not meet an obstacle in its path. This means that sufficieitly large samples must be available for valid conclusions to be drawn. This is especially true when specific sites are under consideration. Accident rates at the same site vary considerably from year to year in the absence of any detectable changes. Even for high-accident locations, rates have been found to vary by a ratio of 1 to 4 from year to year (15).

The only alternate source of information is in-depth studies of traffic behavior in selected locations. The studies made under NCHRP Project 1-12 provide this type of information and the methods developed by the project can be applied to locations that present unusual problems. Because skid resistance control has as its main objective the prevention of accidents, accident data will remain the chief tool for validating standards and for identifying locations requiring skid resistance upgrading, whether or not they conform to an applicable standard.

The latter use of accident data will become progressively more important as more and more pavements conform to or exceed a prescribed minimum for skid resistance. Analy-sis of localized accident data is virtually the only method by which it is possible to identify objectively locations requiring remedial measures; that is, by either increasing the available skid resistance, or lowering the skid resistance requirements by realignment, traffic signs, etc.

SPOT IMPROVEMENT

Tackling the skid resistance control problem by spot im-provements is prescribed by the Federal Highway Safety Program Manual (13). Identification of high skidding

accident sites can often be done subjectively by inspection of pin maps and reference to the accident reports for the location. Eventually, however, a more objective approach is necessary (16). It involves comparing the skidding accident rate at the site under consideration to the same type of accident rate at similar sites or for the highway system of the district or region as a whole.

According to National Safety Council figures (17), the national average is 14.0 reported accidents per million vehicle-miles. The wet-pavement accident rate is estimated to be about 50 percent higher than the mean rate. Of the wet-pavement accidents, an estimated one-third involve skidding, so that the approximate national wet-skidding rate is 7 per million vehicle-miles. This figure, of course, varies appreciably from state to state, because speeds, geometrics, precipitation patterns, reporting accuracy, etc., vary across the United States.

Accident rates at particularly skid-prone sites can be 100 or more times as high as the statewide average (18). Such sites are not difficult to identify. High accuracy of the accident data here is not necessary. As the worst sites are corrected or eliminated, the need for accuracy increases. This also requires that the sites are carefully defined and that counted accidents have actually occurred within the limits of the site. If there is doubt concerning the reli-ability of the identification of wet-skidding accidents, wet-pavement accidents can be used instead (19).

SLIPPERINESS AND HYDROPLANING

If many wet-skidding accidents are found to occur at a site, the cause can be either low pavement skid resistance or some other feature that demands unusually high skid resistance (e.g., alignment, sight distance, superelevation). Normally, it is not difficult to distinguish between the two causes, particularly if the skid resistance of the pavement has been determined.

There are, however, situations in which the skid resist-ance may be adequate and the normal skid resistance requirement low, and yet many skidding accidents occur during rainy periods. This situation is usually found at sites where speeds are high. The accidents are then very likely caused by hydroplaning, which occurs when the combination of water depth on the pavement and vehicle speed is such that-the water cannot completely escape from the tire-pavement interface. Tire and pavement are com-pletely separated from each other by a continuous water film. Friction is zero and the driver is completely without control. The obvious remedies are reduction of (a) the waterfilm thickness on the pavement or (b) speed, or both. Pavement surface characteristics can raise the speed at which hydroplaning will first occur for a given water film thickness (or water input, by rain or drainage, Onto the pavement), but there is as yet no accepted method for identifying the hydroplaning potential of a site.

Skid resistance measurements can qualitatively distin-guish between pavements that are more or less conducive to hydroplaning because wet skid resistance drops with speed. The sharper the drop, the sooner skid resistance will reach zero. At present there is no way by which to

12

anticipate with any degree of accuracy if, when, and where hydroplaning may occur. However, when vehicles hydro-plane, accidents are almost certain to occur. Hence, acci-dent records should be carefully scrutinized for evidence of this hazard. Such accidents will almost invariably occur at only clearly defined locations, comprising short sections of highway.

COST-EFFECTIVENESS OF SKID RESISTANCE CONTROL

The National Safety Council gives the total cost of all motor vehicle accidents in 1969 as $12.2 billion (17). This converts to $0.0115 per vehicle-mile, or $790 per accident. This is the reported economic cost and includes the generally less destructive urban accidents. For rural highway accidents, $2,000 per accident is probably a more realistic estimate (15). State accident records should be used to obtain data that permit more precise consideration of reporting procedures and other variables.

The number of accidents that can be avoided by counter-measures can be estimated if accident rates and skid re-sistance data by sites or road sections are available for a sizeable sample throughout the state or other maintenance jurisdiction. A generalized relationship between wet-pave-ment or wet-skidding accidents per vehicle-mile and skid resistance will show how many accidents would be pre-vented if the skid resistance were increased by a certain amount. Figure 1 is an example of such a plot.

It must, of course, be known what skid resistance can be attained with available remedial measures. This infor-mation should be obtainable from prior practice. Experi-ence will also indicate for how many years the improvement can be expected to last. When novel measures are being considered, estimates of the degree and life expectancy of the improvement must be made.

With this information in hand, the number of accidents prevented over the useful life of the improved surface can be obtained and, multiplied by the cost per accident, the amount is obtained against which the cost of installing a new surface may be measured. It is, of course, true that no direct benefit accrues to a maintenance budget from prevented accidents, but being able to show the possible reduction in social cost may well help to secure an increase in the budget. The greatest usefulness of a benefit/cost analysis lies, however, in its ability to help establish priority ratings for maintenance programs. The analysis can also aid in choosing among two or more remedial treatments when the more effective ones are more costly.

More formal approaches can be taken by developing general benefit and cost functions (5), but determination of the constants requires a large amount of historical data. Such data are not available, except in isolated instances. Therefore, the simplified analysis outlined here is likely to be more practical.

The cost of installing countermeasures can vary con-siderably. The cost of aggregate and binder, as long as conventional materials are used, does not vary much, but if aggregate providing high skid resistance is not available locally, transportation cost becomes a major factor. When synthetic aggregates or binders are to be used, their cost

can, of course, alter the situation significantly, but the cost/benefit ratio may not be affected to the same extent if the desired skid resistance level can be maintained over a significantly longer period.

System-wide cost/benefit studies are necessary before minimum skid resistance requirements are mandated. If the cost of meeting a standard is excessive it may not be met. Hence, policy decisions are necessary and these do require the availability of cost/benefit ratios.

VARIATIONS OF SKID RESISTANCE

As important as the absolute skid resistance level of a pavement is its uniformity. When the wheels on one side of a vehicle encounter a different skid resistance from those on the other side, a turning moment develops during braking that can spin the vehicle around or at least force it out of its original path (or cause jackknifing in the case of articulated vehicles). This happens quite frequently on snow- and ice-covered roads, because skid resistance is not only low but also far from uniform.

Dry pavements usually have a skid resistance.adequate for most maneuvers, but nonuniformities on them can also convert a normally safe maneuver into a dangerous one because the wheels on one side of the vehicle may lock while on the other side they do not. However, limits for permissible nonuniformity across a pavement have not been established.

Nonuniformity in the longitudinal direction can also be dangerous. As a driver travels along a highway, he sub-consciously gets to know it and adjusts his driving to its characteristics. If he encounters a sudden change, either he may not become aware of it, perhaps because traffic or other things divert his attention, or he may at first not know how to allow correctly for the change. He requires some experience; he must relearn, so to speak. If he does not have time to do so before a more-severe-than-normal maneuver is required of him, he may end up in trouble. The magnitude of this problem is not yet known.

The learning process may also affect the wet-pavement accident rate generally; i.e., the rate is not constant with time. Immediately after the start of a rain, the skid resist-ance is thought to be lower than after the rain has con-tinued for some time. This has never been conclusively documented. In effect, efforts to do so have failed (20, 21). It is, however, likely that drivers simply require time to adjust to the new condition; they have to relearn with every new rain that a wet pavement is more slippery than a dry one. Whatever the cause, it has been shown in at least one instance (22) that there may be a closer rela-tionship between yearly accident rates and the number of days with pre9itation than between accident rates and the number of hours during which pavements were wet, for the reason that the number of times the pavement changes from dry to wet is probably more significant than the total duration of wetness.

There are, significant seasonal changes in skid resistance. The cause of these changes is not important here, but the fact that they exist is. When compliance with a standard is to be shown, the day on which the tests are made may

13

make the difference between meeting the standard or not meeting it.

In setting standards it is important to recognize that traffic encounters a constantly changing skid resistance, even though it is still unknown if these changes have a significant bearing on skidding accident rates.

LEGAL ASPECTS OF SKID RESISTANCE

There is no easy answer to the degree of skid resistance that an agency must maintain to be free of liability. It goes without saying that the highest degree of skid resist-ance possible should be maintained.

Basically, however, laws state that an agency has an obligation to provide for users of any transportation me-dium a roadway that is reasonably safe for the purposes of travel. The agency involved, no matter at what level of

government it is, is not an insurer of the roadway. In practical language, the courts have interpreted these prin-ciples just about as widely as can be imagined. But the important point to remember here is that all of this inter-pretation appeared without any guidelines, programs, or official recommendations.

If past practice is any aid, it can be predicted that, with the appearance and common use of recommended guide-lines for minimum standards, the courts will rely on these standards as establishing reasonable safety. In this event, it is advisable that agency programs are, in the very least, set up to meet recommended minimum factors as promul-gated in these uniform guidelines (115). The fact that an isolated state court or two does not recognize these guide-lines will in no way alter the recognition that the majority of courts will extend to them.

CHAPTER THREE

METHODS OF SKID RESISTANCE MEASUREMENT

The term "skid resistance," as commonly used, refers to the characteristics of pavement surfaces that inhibit skid-ding; that is, the sliding of a tire or a vehicle in an uncon-trolled manner, with the lack of control being the result of the tire having ceased to rotate. In keeping with this interpretation the basic method of determining skid resist-ance consists in measuring the force required to drag over the wetted pavement under test a tire that is pre-vented from rotating.

There are other methods for characterizing pavement skid resistance that are technically just as valid as the locked-wheel (or locked-tire) method, but the latter has come to be considered the basic method in the United States. Other methods give different numerical results and respond differently to such variables as speed or pavement texture than the locked-wheel method does. Although the pavement can normally be considered perfectly rigid, the tire or rubber is not only elastic, but even viscoelastic in nature. For this reason its interaction with the pavement is extremely complex.

Thus, it is best for skid resistance measurement not to speak of "coefficient of friction," because this implies a fairly simple basic interaction between two bodies sliding on each other. The term "friction factor" conveys a more general meaning and is used when one does not wish to or cannot define the nature of the frictional interaction (char-acteristics of tire and vehicle and its operation have not been isolated). For the measurement of skid resistance, as a criterion of pavement surface quality, certain terms, such as "skid number," have come into use because, by spelling out their meaning in standards, it is precisely understood what is meant by them. Such terminology may

seem arbitrary, but adherence to it will prevent errors and misunderstandings and will aid communication in a field in which much can be gained through interchange of data and findings.

It cannot be overemphasized how important it is, even in routine work, to adhere to standardized procedures for determining skid resistance. Although pavements, even the so-called flexible ones, are essentially rigid as far as skid resistance is concerned, their surface characteristics change with time and are frequently far from uniform. Unless repeat measurements are made in the identical manner, changes may be thought to have occurred that actually have not, and the magnitude of real changes will be dis-torted.

ASTM Committee E 17 on Skid Resistance deals with the development of standard methods of measuring avail-able skid resistance. These methods should be used, unless there is valid reason for using others. Because ASTM standards are under constant review and new ones are being added from time to time, the latest ASTM Book of Standards should be consulted. This synthesis attempts only to present the salient points of such standards as are discussed.

LOCKED-WHEEL TRAILER METHODS

By using a tire representative of those most commonly used on vehicles, it is thought that the results of skid resistance tests can be directly applied to the performance of vehicles in traffic. This is only approximately true for reasons that will be dealt with elsewhere. However, the difference in

14

performance of the variety of the tires available for one vehicle model makes such a generalization hazardous (see Fig. 6, which compares tires of identical cross section). These difficulties are not overcome, but comparison of the skid resistance of different pavements !ecomes possible, if a standard tire is used. In the United States the standard-ized pavement test tire is that defined by ASTM Standard E 249. It is a bias-ply 7.50 X 14 tire with five circum-ferential grooves. The standard rigidly prescribes the rub-ber composition. Thus, the tire type and design are elimi-nated as variables in the measurement of pavement skid resistance.

ASTM Method E 274 describes how this tire is to be used when measuring skid resistance of highway pavement (or airport runway) surfaces by the locked-wheel method. It presupposes that the test tire is installed on the wheel or wheels of a single- or two-wheel trailer. For a measure-ment, the trailer is towed at a speed of 40 mph over the dry pavement and water is applied in front of the test wheel (all skid resistance measurements refer to wetted pave-ments *); the ASTM method specifies the method of apply-ing the water and other details. The test wheel is locked up by a suitable brake, and after it has been sliding along the pavement for a certain distance (to permit the tempera-ture in the contact patch to stabilize) the force that the friction in the tire contact patch produces or the resulting torque on the test wheel is measured by suitable means

* Exceptions are NHTSA regulations for certain vehicle tests •for which dry surfaces are specified, but which nevertheless are to be checked for skid resistance according to ASTM Method E 274.

6.50 x 15 BIAS PLY 24psi INFL PRESSURE

ULL STANDARD TREAD HIGH STYRENE RUBBER

RIBBED TIRE SYNTHE11C RUBBER

SMOOTH liRE—' NATURAL RUBBER

30 40

SPEED, mph

Figure 6. Differences in locked-wheel performance on inter-changeable tires on the same pavement surface (fine cold as-

p/mIt) (29).

and recorded for a specified length of time. The result of such a test is reported as skid number (SN).

Many states and some universities and private organiza-tions now have skid trailers. Although most conform to ASTM Method E 274, only a few have been built according to identical plans. Therefore, some differences exist between them. Almost every one of them is (more or less) permanently paired with a towing vehicle that carries a water supply, instrumentation, and other associated equip-ment. These units might be referred to as road friction testers. Figure 7 shows typical road friction testers with two- and one-wheel trailers.

Skid resistance measurements are not as unequivocal as this description might suggest. For instance, even though the standard tire has been found to be very uniform, friction of rubber is temperature dependent, and so are tire characteristics. For a single temperature, it would not be difficult to standardize measured skid numbers. In reality, there is a complicated interplay that involves air, pavement, and water temperatures, and the heating of the tire structure by flexing, and of the tread by friction. Thus, no universally accepted method of correcting skid number for temperature exists as yet.

There are other potential error sources, some of which affect the repeatability of tests made with the same tester, whereas others cause different testers to give different re-sults on the same surfac& under the same conditions. The latter aspect has most recently been documented in the "Florida Correlation Study" (23). Comparison with earlier, similar studies shows that much progress has been made, but that, particularly for tests at high speed, im-provement is still needed. The current NCHRP Project 1-12 (7) is endeavoring to resolve these problems. Major points to be considered in the design and operation of locked-wheel testers were discussed by Kummer and Meyer (24).

These imperfections, however, do not diminish the use-fulness of ASTM Method E 274 for most purposes, such as determining whether or not pavement slipperiness is a governing or controlling factor at high-accident locations, making surveys of a road system for the purpose of iden-tifying slippery sections and setting priorities for remedial programs, and determining the relative in-service perform-ance of different types of construction, aggregates, and surface treatments. Difficulties arise only when high pre-cision is needed, as in determining if a particular surface does or does not meet a mandated level of skid resistance.

AUTOMOBILE METHODS

The most natural method of determining the skid resistance of a pavement is to drive an automobile on it, lock up the wheels and find out how far it slides until it comes to a full stop. There are both advantages and disadvantages to this method. Its main advantages are low capital invest-ment and ready availability. A major disadvantage is that it is potentially hazardous, particularly if high-speed tests are to be made.

The hazards can be reduced by not braking all wheels. When all four wheels are locked, the vehicle becomes unstable and any small disturbances (such as a small differ-

15

ho AUTI

N C STAff PWY CON

ROAD IESTI

NN-OW, I WMW17 pXTur,4 t AI"Ih

Figure 7. Typical road friction testers.

ence in the friction of the wheels on either side) can cause it to veer from its intended path. By braking the front wheels only, or a diagonal pair of wheels, the risk of spin-out is significantly reduced (25).

It can be argued that not braking all wheels is unrealistic because it does not show how far an automobile would slide in an emergency. This is correct, but even if all four wheels are locked up for the test, it is still not feasible to deduce from the results the stopping distance of vehicles in traffic. The locked-wheel stopping performance among vehicles differs significantly because of differences in vehicle design, suspension design, tire types, vehicle condition (particularly the condition of shock absorbers), tire infla-tion pressure and tread wear, vehicle payload and load distribution, etc. For this reason it also is impossible, with any degree of accuracy, to estimate from tire tracks the speed a vehicle was traveling prior to an accident.

Nevertheless, if they are made carefully locked-wheel automobile tests can serve well for characterizing and comparing pavement surfaces. To eliminate tires as a variable, such tests are generally made with the ASTM standard pavement test tire (E 249), if only because it is

difficult to procure commercial tires of adequate uniformity over long periods of time.

A disadvantage of locked-wheel automobile tests for assessing the skid resistance of vet pavements is that one must depend either on rain or on it sprinkler truck. Neither assures the controlled water film thickness that is obtained with a road friction tester. The cost of a sprinkler truck adds significantly to the apparently low unit cost of auto-mobile tests. For a comparison of the cost of various methods of measuring skid resistance see Kummer and Meyer (24).

The tests can be carried out in various ways. One can measure the distance the vehicle travels from the point at which the wheels cease to rotate or from the point at which the vehicle, with wheels locked, passes through a specified speed to the point at which the vehicle stops. One can measure the distance or the time the vehicle requires for decelerating from one speed to another, say from 30 to 20 mph. Deceleration can be measured directly with a suitable instrument and either the deceleration at a certain point in the test cycle or the mean deceleration over a stated portion of the cycle can be used to charac-terize the pavement tested.

- -:-..• -

16

In comparing the results of any automobile method with those from a locked-wheel skid tester, one must recognize that the automobile decelerates during the test but the road friction tester does not. However, if they are used correctly the automobile methods tend to give more consistent results than are obtained with road friction testers. This is largely accounted for by the fact that on the automobile a single reading averages the performance of four (or at least two) tires, whereas on locked-wheel testers, even those with two-wheel test trailers, normally only one wheel is locked up (with both wheels locked the trailer tends to become too unstable for safety and water consumption is doubled).

At stops from high speeds the vehicle is partially re-tarded by air drag, in addition to the resistance at the tire-pavement interface. In still air and with an aero-dynamically clean car this is of minor importance, espe-cially in direct comparison tests, but when there is wind, the magnitude of the error becomes elusive. Load transfer to the front axle during a stop is not important if front and rear wheels are locked. For this reason, the diagonal braking method is preferred if not all wheels are to be braked. Because the friction factor is slightly load depen-dent, braking only one axle may introduce a systematic error (especially if the vehicle has a high center of gravity) and a variable error (if the static load distribution is per-mitted to vary).

ASTM Committee E 17 has developed E 445-71T as a recommended method for automobile tests. Information on the details of various methods may be found in, among others, Home and Sparks (25), Dillard and Mahone (26),

and Rizenbergs and Ward (27). A thorough discussion of the dynamics of automobile locked-wheel stops is presented byBrach (10).

PORTABLE FIELD TESTERS

The cost of road friction testers and the hazards inherent in all types of high-speed tests have repeatedly led to the development of so-called "portable" testers, machines that are not operated by or as a moving vehicle, but can be moved readily from location to location. They are not laboratory devices that are used on the highway, but units specifically designed for highway use. Two examples are described and shown.

The California Skid Tester (Fig. 8) operates on the principle of spinning up a rubber-tired wheel while it is off the ground, lowering it to the pavement, and noting the distance it travels against the resistance of a spring before it stops. The device is attached to the rear of a suitable vehicle, which is stationary during a test, per-mitting it to be easily placed (for details see Beaton et al. (48). It is worth noting that this tester is normally operated with glycerine instead of water as the pavement lubricant. This is done because glycerine insures a longer lasting, more uniform film.

A hand-carried device is the Drag Tester (Fig. 9) de-veloped by The Pennsylvania State University (24) and marketed as Keystone Tester. It employs a rubber shoe that slides along the pavement as the operator "walks" the tester. The frictional resistance experienced by the shoe is converted to hydraulic pressure and displayed on a gauge. Water must be applied to the pavement ahead of the tester by suitable means.

These two examples show how completely different prin-ciples can be employed in friction testers. The numbers obtained with them may be similar to. or may differ greatly from, those obtained by other methods. Approximate agreement among differing types of testers can be achieved by suitable dimensioning of the operating elements or

Eigurc 8. California portable skid tester.

17

Figure 9. Keystone Mark IV skid resistance tester.

choice of the readout scale. Perfect correlation between testers must not be expected if the operating principles and/or modes differ. The correlation problem is discussed in detail in a later section. The advantage of portable testers is that they can be inexpensive, therefore a large number of operating units can be equipped with means for measuring friction. They also permit friction to be mea-sured in locations where a road friction tester cannot be operated. They permit testing pavement markings and checking the transverse change in the friction factor across a highway or lane. The truly portable types permit the tests to be made without the need to interrupt traffic flow.

Portable testers, because they do not employ or ade-quately simulate locked-wheel testers, cannot substitute for the latter, but they can serve as useful supplements.

LABORATORY TESTERS

Researchers have devised innumerable devices and ma-chines for measuring friction in the laboratory. Such test-ers are of little interest here because special objectives and conditions govern their design and limit their usability for purposes other than those they were intended for. How-ever, one design that is extremely versatile in its applica-bility to many test situations and, in consequence, has attained wide use, is the British Portable Tester, developed by the British Road Research Laboratory. Its operating

principle, performance, and use are described in detail by Giles et al. (28). As shown in Figure 10, it consists of a pendulum to which a spring-loaded rubber shoe is attached. By letting the pendulum drop the shoe is made to slide over the surface to be tested. The attenuation of the re-bound serves as a measure of the friction. Because the details of the test must be carefully controlled, ASTM Method E 303 has been developed. The results are re-ported as British pendulum numbers (BPN) to emphasize that they are specific to this tester and not directly equiva-lent to those obtained by other methods or testers, includ-ing other pendulum testers (such as the Leroux).

One advantage of the British Portable Tester is that it can be used not only in the laboratory, but also on the highway. It was used for this purpose rather extensively in the past, before trailer-type road friction testers became widely available. The ability to be used in the laboratory on cores and other specimens, as well as on the highway, is a decided advantage of the pendulum tester. Weighing against its use on the highway is that setting it up and getting it ready for measurement is a rather lengthy process, and traffic must be stopped or diverted. All this makes the cost per test quite high (24). In addition, because it is a low-speed device (the rubber shoe contacts the pavement at a speed of about 7 mph) and because only a narrow rubber edge contacts the pavement, it does not correlate well with locked-wheel trailers that operate at 40 or more mph.

18

METHODS EMPLOYING TIRES IN OTHER THAN THE LOCKED-WHEEL MODE

If the interaction of tires with pavement surfaces could be modeled perfectly, one would be able to predict from the results of tests with a tire in one mode its performance in all other modes. The present state of knowledge does not permit this. Consequently, there is a certain validity to the argument that locked-wheel tests are not generating all information necessary to judge how well pavement surfaces assure the safe movement of traffic using them.

The question is whether one wishes to rate a pavement in terms of its ability to provide minimum stopping dis-tance in an emergency or of the maximum longitudinal or lateral accelerations that the pavement will accommodate. In the absence of sufficient data this is at present a matter of subjective choice. The following discussion is therefore

designed to show what information alternates to the locked-wheel method yield and what problems their implementa-tioti and use pieseiit.

The Slip Mode

If the brake of a wheel running straight ahead is applied with gradually increasing force, the wheel develops increas-ing slip (Fig. 11). At first the friction factor increases steadily; eventually it reaches a maximum at what is re-ferred to as "critical slip." If the braking force is increased further, the wheel locks up. Above the critical slip the friction factor drops off; therefore, there is an excess of available braking force.

Slip as used here is defined as

s=100 (9) ()Q

Figure 10. British Road Research Laboratory's pendulum friction tester (British Portable Tester).

'19

7,

FRICTION FACTOR '

SLIP

Figure 11. Friction factor as a function of slip (wheel moving in the direction of the wheel plane while being braked).

in which s is the slip, in percent, co is the angular wheel speed at the time of measUrement, w0 is the angular wheel speed in free rolling. When braking, < co, and the slip is positive. It is 100 percent when the wheel locks up (w '= 0). When driving force is transmitted, the slip be-comes negative because co > w0.

Two points are worth noting: (1) the friction factor at critical slip is higher than that at lock-up, so that greatest deceleration of a vehicle is not obtained when the wheels are locked, and (2) control is lost when the wheels lock. For these reasons automatic brake control systems have been developed whose task it is to keep the wheels operat-ing in the vicinity of the critical slip.

Thus, knowledge of the maximum friction factor is of real and practical interest. It cannot be computed from the locked-wheel factor (Fig. 12). The illustration repre-señts only one set of data that is not universally valid. For instance, Giles (30) reports much smaller ratios for some fine-textured surfaces, as well as significant increases of the ratio with speed. Consequently, because at present the ratios of the two friction factors can at best be explained qualitatively, the maximum factor can be obtained only by an appropriate measurement.

This would be a relatively simple problem if the maxi-mum friction factor would always occur at the same slip; that is, if the critical slip were constant. Unfortunately, this is not the case; as Figure 13 shows, the critical slip changes with surface texture. But it also changes with friction force. The reason for the latter phenomenon is that the slip of the wheel (measured at the hub) is larger than the mean slip in the contact patch because of "tire wind-up" (32). Furthermore, the critical slip changes with temperature. For the reasons just stated, a tester measur-ing friction at constant slip will not always measure the maximum value of the friction factor, as may beseen from Figure 14. Because, however, most friction factor vs slip curves have a smaller slope to the right of the critical slip (Fig. 14) than to the left of it, one can mini-mize the possible error by choosing for the tester a fairly

ock at40MPH

Figure 12. Ratio of maximum and locked-wheel friction factors at 40 mph on various wet surfaces (29).

large slip (about 15 percent), with, the best value depend-ing on the test tire chosen and with due consideration for the pavement types and operation conditions most fre-quently encountered.

0 a. a--J CO

-J 0

C-)

GRIT SIZE Figure 13. Change of the critical slip with surface tex-ture for approximately constant maximum friction fac-tor (31).

LOCKED WHEEL FRIC11ON FACTOR

/ U.

—CRITICAL SLIP

/ / WHEEL LOCKED

I .3 0 E..

0 100%

20

Such testers can provide useful data if their limitations are recognized.' An example of a runway trailer is shown in Figure 15. The slip is determined by the difference in the radii of the drive tires and the test tires, and the trans-mission ratio of the belt drives between the two sets of wheels. A simpler version of this "Skidometer," developed by the-Swedish Road Research Institute, omits the belt drive and uses only a single test wheel that is connected via clutches to both drive wheels. On both models the test wheel(s) can also be locked up.

By providing an infinitely variable transmission between drive and test wheel, the indicated limitations can be cir-cumvented. By changing the slip slowly, any error caused by the changing moment of inertia of the drive train can be made negligible. Once the critical slip has been found on a given surface, only small excursions are necessary' to insure that the true maximum friction factor continues to be recorded. The tester shown in Figure 16 is of this type. It differs in one major respect from the machines just described in that the drive train of the towing vehicle serves to force the test wheel to rotate at a reduced rate. The two wheel systems are connected by a hydraulic trans-mission (34). Several other testers of this general type have been built, usually for research rather than for rou-tine use, and therefore often incorporating additional capa-

CRITICAL SLIP

ISLIP AT WHICH TESTER OPERATES

SLIP

Figure 14. Constant-slip road friction tester on two difterent surfaces (A and B).

3 and 11. °Rolling" or drive wheels Loading weight and 7. Test wheels Wheel for speed and distance measurement Drive 'shaft, Brake Driven sheave Drive belt Clutch

'Force transducer and suspension spring

Figure 15. Plan view of model BV9 constant-slip friction trailer of the Swedish Road Research Institute (33).

21

Fiç'ure 16. Run stay friction tester with adjustable slip (PA A

bilities. One of the latest and most elaborate machines of this class is Swiss (35).

Less mechanical complexity is necessary if the tests are made with transient slip. By gradually applying the brake of an undriven test wheel, or retarding it by other means, its slip is gradually increased. As soon as the critical slip (that is, maximum friction) is exceeded, the wheel goes rapidly into lock-up. By judicious timing of the brake release action, lock-up can be avoided if desired. Cycling the test wheel in this manner can simulate the action of the automatic antilock braking systems used on aircraft and automobiles. Obtaining the friction force requires that any effect of the changing moment of inertia of the test wheel is eliminated from the signal. Systems for doing this are described by Domandl and Meyer (36) and by Goodcnow et al. (37). A typical record is shown in Figure 17. If the tester speed is held constant, the slip can be computed from the test wheel angular velocity. This per-mits construction of friction factor vs slip curves in the

BRAKE FULLY BRAKE APPLICATION

INSTANTS OF PEAK FRICTION FORC E

DEVELOPMENT 0

RIION FORCE

o-J WHEEL

LOCKED TEST WHEEL

ANGULAR VELOCITY

-TIME

Figure 17. Osci/lot,'ra,n of a friction test in which the test st/seel brake is s/on/v applied and re/eased (tester speed Constant).

manner of Figure II. Usually, however, only the maximum and locked-wheel coefficients are of interest. By recording the friction force and the test wheel speed on tape, the data evaluation can be performed by a computer.

An alternate way is to use an automobile and install a recording decelerometer as close to the center of gravity of the vehicle as possible (29, 38). A typical record of a test in which only the front-wheel brakes were applied (gradually) is reproduced in Figure 18. The peak friction factor can be obtained from the deceleration curve, as can the locked-wheel factor (using Eq. 4). The vehicle speed is also recorded. It can be seen that a number of tests may be needed to obtain the peak coefficient at a specific speed and that at least two tests are necessary to get peak and locked-wheel friction factors at the same speed.

z 0

LU —J

Iw 50.

Lcs

cPeak Deceleration

El

- 30s 1.0.4

uj —

Deceleration Wheels Lock

to Rolling Resistance and Air Drag

TIME, sec

Figure /8. Speed and deceleration record of vehicle duruzg s/ott' brake application (18).

22

YAW ANGLE, deg

Figure 19. Typical sideways friction factor vs yaw angle rela-tionships for -two wet pavements (A and B).

The Yaw Mode

To cause a vehicle to change course, the front wheels must be turned so that they make an angle with their original path. When doing so, the tires develop a transverse force that causes the vehicle to change direction. The same force lets the wheels support the centrifugal forces acting on the tire-pavement interface as the vehicle travels through a curve. (The angle of the wheel plane with the direction of motion is the slip or yaw angle, and the force is the cornering or side force.)

Without yaw the tire develops no side force (Fig. 19). This is analogous to what applies to straight ahead motion —without slip the tire develops no braking (or accelera-tion) force (Fig. 11). The analogy goes further: the side force (or the side friction factor) peaks at some yaw angle, just as the braking force (or the longitudinal friction factor) does. Indeed the two peak values are almost uniquely and linearly related to each other (39, 40).

In consequence, measuring pavement friction with a wheel (unbraked) set at a constant yaw angle, presents in principle the same problem as arises in connection with testers operating with constant, longitudinal slip: what yaw angle to use. The answer is not clear; many of the published data go only up to relatively small angles because they usually are intended for vehicle dynamics studies and in normal service 8° or 100 is about the maximum that is of interest. Besides, the side force peaks are usually fairly flat, so that detailed knowledge in this region is not vital.

The critical yaw angle, however, is subject to the same variations as the critical slip. Therefore, for pavement evaluation it is desirable to work with a yaw angle that is relatively insensitive to variations in surface characteristics and operating conditions. According to Giles (21), 15° is a good choice. Accordingly, the side force measuring tester ("SCRIM machine") of the British Road Research

Laboratory operates. with this angle. There are a number of testers in existence, other than those of the Road Re-search Laboratory (1), that permit operation with variable yaw angle (35, 40), so that answers to some of the pend-ing questions should be obtained in the near future.

A relatively simple, commercially available trailer (Fig. 20) uses yawed wheels with smooth tires and measures the side force developed by both because both are yawed at equal, but opposite, angles so that the trailer will travel in a straight line without requiring a restraining mechanism.

In first approximation for pavement work, maximum sideways friction factor and maximum braking friction factor of the same tire on the same surface can be con-sidered quantitatively equal. Consequently, the informa-tion that can be obtained by either method about surface properties is nearly the same and the choice between the two methods is a matter of how tester detail problems are resolved and which method proves to be more practical and convenient in day-to-day use. Whether either method provides better or more information about the safety of pavement surfaces remains to be shown.

CORRELATION BETWEEN TESTERS

it is clear from what has been said so far that each type of tester measures a different aspect of the friction devel-oped on a pavement surface from the other types. Even when the same tire or slider is used, speeds or the modes of operation differ, the water film control may not be the same, and other details may vary. Therefore, it should not be expected that there will be a 1: 1 correlation between the results obtained with different types of testers.

This fact has not been recognized in the past by all who have tried to obtain correlations. This has led to con-demnation of certain tester types. It can be stated cate-gorically that any tester that measures the force (or a parameter uniquely related to it) acting at the interface of the tire or slider and the contacting surface in the plane of the surface, produces a valid friction measurement. However, measurements with two different testers must be made under comparable conditions to permit correla-tion without some type of correction. Any correction would require knowledge of at least one additional variable, and usually more than one.

The reasons for this difficulty lie in the complexities of the frictional behavior of rubber and tires. Figure 21 shows how the friction of a sphere sliding on rubber changes with speed. The total friction is the sum of two components—adhesion or surface friction, and hysteresis or deformation losses. The relative contribution of the two components changes with the micro- and macroge-ometry of the surface (here that of the sphere; in case of a tire, with that of the pavement). Microroughness (that roughness that is diminished when the surface is being polished) influences the adhesion component most strongly. Macroroughness, or texture (as given by the shape and arrangement of the aggregate particles embedded in the matrix of the pavement surface course), has a first-order effect on the hysteresis component. The adhesion com-ponent can disappear if the surface is completely covered

59-19-174

59 _.11 JEW- VII

23

Figure 20. "Mu-Meter" of the Utah Slate Highway Department.

by a lubricant, whereas the hysteresis losses can disappear on a perfectly smooth surface.

Because the independent change of the two components with speed also is affected by the conditions of and at the surface, it is easy to sec not only that two testers operating at different speeds are likely to measure different total friction factors on the same surface, but also that the ratio of the components of the two friction factors is different. Consequently, the measured friction factors for two different surfaces do not have a simple relationship to each other.

The matter is further complicated by the fact that the water film under a tire and any type of rubber slider have entirely different dependence on changes of speed, micro-and niacroroughness, and other factors. Figure 6 shows how much a change in tread pattern and rubber composi-tion alone can change the friction characteristics. Obvi-ously still greater differences are possible when sliding or slipping tires, or tires and sliders are compared.

lemperature introduces its own complications. Basi-cally, increases in temperature shift the curves of Figure 21 to the right. Energy expended in overcoming friction reappears as heat. Some of it is removed by the water used in friction tests, some goes into the pavement, and

the remainder goes into the tire tread. The effect of the heat going into the tire (or tester shoe) is obviously a complex one.

Without discussing further possible obstacles to simple correlation between different types of testers, it should be clear that general correlations are, at least in a practical sense, not possible and that when correlations are found, it is either because the surfaces on which they were obtained included only a limited range of types, or the testers do not differ significantly in operating principles, or the expected precision of the correlation is low. In short, when a correlation is found this should be con-sidered fortuitous, rather than as fulfillment of a justified expectation.

ASTM Method E 274 vs Automobile Methods

For stopping distance or deceleration methods, for which the test vehicle is equipped with the ASTM Standard E 249 tires, comparison with ASTM Method E 274 is quite feasible because the most powerful factor, the rubber slider in contact with the pavement (in this case the tire), is eliminated as a variable. However, if true stopping dis-tance tests are made, the friction factors obtained with them will be numerically higher than those represented by

24

U. iL

-ADHESION U. ------ .11 --

- —'

U.

z;STE;E;

SLIDING SPEED

Figure 21. Generalized representation of the coefficient of fric-tion between a steel sphere and rubber as a function of sliding speed.

the skid numbers obtained according to ASTM E 274. The reason is that wet pavement friction factors increase with decreasing speed and that in stopping distance tests the car decelerates from 40 to 0 mph (rarely from a some-what higher speed), whereas the standard skid number is determined at 40 mph.

Rather than computing a friction factor (or a skid number) from the stopping distance, empirical correlations should be used. One such correlation was obtained by the Virginia Council on Highway Research in the late 1950's (41). It is referred to in NCHRP Report 37 (5). Newer studies (26, 42) give a somewhat different correlation (Fig. 22).

ASTM Method E 274 vs British Portable Tester

The British Portable Tester not only measures friction at relatively low speed (7 mph or less), but it also brings

a E 0

0 U- 0.

U)

SKID NUMBER (40 mph)

Figure 22. Correlation between stopping distance friction factor and skid number (43).

TRAILER AC. ASTM E274, SN

Figure 23. Correlation between Mu-Meter and two-wheel trailer conforming to ASTM Method E 274 (ten pavements).

the edge of a rubber shoe (instead of a tire) into contact with the pavement (or pavement sample). Consequently, any correlation with Method E 274 would be purely fortuitous. NCHRP Report 37 (5) gives a correlation that is based on Dillard and Mahone (26), but cautions that the correlation "is not very satisfactory".

Giles et al. (28) give a correlation between the friction factors as measured with the pendulum tester and a side-ways friction faôtor at 30 mph. On "rough-looking" and medium-textured surfaces the correlation is good; but "on smooth-looking surfaces there is less evidence of correla-tion, the differences being sometimes as much as 0.3 in terms of coefficient values." The reference also gives a suggested temperature correction of the order of 3 BPN per 10°C (18°F). It should be noted that in Europe natural rubber is used for the slider, although ASTM Method E 233 prescribes a synthetic rubber, the same rubber used in the ASTM E 249 standard pavement test tire.

ASTM Method E 274 vs Mu-Meter

Statistical correlation between the Mu-Meter and a tester conforming to ASTM Method E 274 has been shown to be good (44) when both use tires without tread (which are standard for the Mu-Meter) and both operate with the pavement wetted by sprinkler truck (standard practice for the Mu-Meter) (see Fig. 23). The correlation is not quite so good when the locked-wheel trailer is operated strictly in conformance with Method E 274. In either case, the average maximum deviation from the correlation line is ± 5 SN at 40 mph for tests on the same pavement. A \

'possible explanation is the fact that on the Mu-Meter the side forces acting on the two test wheels are averaged. Because their track width is less than that of most motor)

25

(vehicles, only one of the test wheels can be made to run in the "most traveled wheel path" where locked-wheel

ç tests according to Method E 274 are to be made. There-fore, the correlation will be affected by the transverse skid resistance profile of a pavement.

Drag Tester vs Other Testers

The Drag Tester (24) uses the same slider as the British Portable Tester, but it is normally operated at still lower speed. However, Kummer (.45) reports good correlation when using a slider made from ASTM E 249 rubber.

Correlation with the California Tester (Fig. 8) was found to be poor (46). This is not surprising in view of the totally different method of operation of the two testers.

Burchett and Rizenbergs (47) compared the results of the Drag Tester with their automobile method that obtains a friction factor from the time or distance a passenger car equipped with ASTM E 249 tires takes to decelerate from 30 to 20 mph with all wheels locked. They found acceptable correlation between the two methods on Ken-tucky hot-mix bituminous pavements.

CHAPTER FOUR

CHARACTERISTICS OF SKID-RESISTANT SURFACES

The ideal pavement surface has the following characteris-tics, which, however, are not necessarily all compatible with one another:

High skid resistance—ideally the skid resistance when wet would be as high as that of the dry pavement. Little or no decrease of the skid resistance with in-creasing speed—the skid resistance of dry pavements is nearly independent of speed, but this is not so on wet pavements. No reduction of skid resistance with time, as from polishing or other causes. Resistance to wear—by abrasion of aggregate, attri-tion of binder or mortar, loss of particles, etc., even when exposed to traffic using studded tires. Structural durability—resistance to compaction, ravel-ing, breakup, etc. Low noise generation. Low cost—not necessarily low first cost, but cost per year of service with acceptable skid resistance. Low tire wear and rolling resistance.

How an acceptable mix of these characteristics can be obtained is discussed in Chapter Six. The skid resistance characteristics are determined by physical features that must be obtained by appropriate materials, and appropriate design and construction methods.

TEXTURE

Texture is the "roughness" that, in a bituminous surface, is most significantly influenced by aggregate size and in a portland cement surface by the finishing method (burlap drag, brush finish, etc.) Texture generates resistance to sliding via the hysteresis effects in the tread rubber and facilitates the expulsion of water from the tire-pavement interface.

As was pointed out in Chapter Three, the friction that develops between tire and pavement has two components —adhesion and hysteresis. The latter reflects the energy loss that occurs as the rubber is alternately compressed and expanded. (The "lost" energy appears as heat.) Thus, as the tire slides over the irregularities of a textured sur-face, resistance develops even if the surface is perfectly lubricated.

On a skid-resistant pavement the contribution of the hysteresis component to the total friction is usually small, but when the pavement is slippery it may represent an appreciable percentage of the total. A block of rubber sliding on a wet sheet of glass develops almost no friction; but if the glass had a pebbled, but still smooth surface, a significant resistance to sliding would be noted (the rubber block must of course carry some load, as a tire does, to obtain some deformation around the irregularities). The hysteresis contribution usually is fairly independent of speed in the range in which highway tires are likely to slide. Because the adhesive friction component on wet surfaces tends to decrease with speed (Fig. 21), the hys-teresis component gains in importance at the higher speeds.

Thus, texture is beneficial to the generation of friction, but its more important function is to provide channels by which the water can escape from under the tire so that the tread rubber can make contact with the pavement (Fig. 24). A tire sliding (or rolling) on a flooded surface buckles in slightly under the impact on the water layer that covers the pavement and a water wedge forms under the tire so that over some distance (A in the figure, which greatly exaggerates the thickness of water layer) the tire is completely separated from the pavement.

Some of the water is splashed out of the way, but more escapes through whatever channels may be available. These may be the grooves and slots in the tire tread or they may be. in the pavement surface. In Zone B the tire

Figure 24. Tire sliding on a flooded, textured surface.

begins to touch the irregularities, but intimacy of contact is still controlled by the water escape channels. In Zone C the channels do not play a role.

Because impact and flow are involved, it will be appre-ciated that the relative lengthof the three zones is a func-tion of the sliding speed of the tire. At very low speeds, Zones A and B are nonexistent; but as the speed increases, first Zone B, then Zone A develops. Zone A grows with

To S

Weigi

Figure 25. Outflow meter (48).

increasing speed, moving Zone B farther back into the footprint. This process is the cause of the decrease of the skid resistance with speed. The growth of Zone A will be slower the more ample the water escape channels are; hence the importance of pavement surface texture.

In reality, the situation is much more complex: tire grooves may be closed off by the distortion in the contact patch, and, at least with bias-ply tires, Zone A tends to grow faster in the center of the footprint than at the edges, etc. Yet, as for the pavement, texture is the governing variable, the absence or insufficiency of which causes the skid resistance to deteriorate with speed. (Grooves or other artificial channels in the pavement surface are discussed separately later, even though they represent simply an extreme form of texture.)

How can texture be described quantitatively? Because its main function is to permit the escape of water from under the tire, it is logical to simulate this process. This has been done by the so-called Outflow Meter. One form (Fig. 25) consists simply of a tube with a flange at the bottom. The flange rests on the pavement, with a rubber ring interposed between the two. The assembly is weighted down, to press the ring against the irregularities of the surface with about the same pressure as exists between tire and pavement. This leaves the channels in the surface open and water in the tube will flow out through them. The time required for the water level to drop a measured distance is taken to be a measure of the texture.

Another method uses a measured quantity of fine sand that is leveled off above the pavement in some prescribed manner (Fig. 26); the area covered will change with the amount of texturing of the surface. Instead of sand, grease can be used (49) or putty (Fig. 27). Texture depth is the mean thickness of the layer of sand or putty that fills the cavities and channels in the pavement surface, if spread evenly on a smooth surface over the covered area.

Correlations of such texture characterizations with the change of the skid resistance with speed ("skid resistance gradient", SN/mph) are not perfect (Figs. 28 and 29), but are quite adequate for many purposes. Certainly, such "texture depth" measurements are vastly superior to such descriptions as "fine-textured", or "open-textured".

Various attempts have been made to obtain more pre-cise information about texture. These include profile trac-ings (53), from which various characteristic numbers can be deduced (average irregularity height, mean void width, roOt mean square of the texture depth, etc.) Figure 30 shows one such profile tracer. The number of possibilities is almost endless and varibus researchers have attempted to make cases for the one or other device or parameter; but when they are applied to a wide variety of pavements the correlations with the skid resistance gradient usually turn out not much better than those with "texture depth" from outflow or patch measurements.

One of the problems is that all these surface characteriza-tions usually represent only a very small sampling of the area over which the locked tire slides in a skid number determination. Somewhat better in this respect are stereo—photographs that are analyzed by standard topographical methods (54), but the improvement does not seem worth

27

the extra cost. Simpler methods that permit quick measure-ment of a large number of samples seem to be preferable and nearly as good. For instance, the Texturenieter (Fig. 31), consisting of a number of pins held in a guide and dropping into the channels and cavities of the surface, permits rapid readout by measuring the free end of a string that passes through holes in all the pins. Sonic detail is lost with this device, but the numbers obtained with it are preferable to word descriptions.

The fundamental difficulty consists in that texture can-not be described by a single characteristic. It is not only the escape of water that controls the skid number-speed gradient, but also the nature of the contact between the tire tread and the surface. Figure 32 shows this schemati-cally. The surfaces in the two sketches differ only in that the upper edges of the channels in A are sharp whereas those in B are rounded. It can be seen that deformation of the rubber as it slides over the edges is much more severe in Case A than in Case B. Furthermore, in Case B small water wedges may form that influence the intimacy of contact between rubber and ridges.

Until a method of texture characterization has been found that takes account of these latter effects, one will have to remain satisfied with correlations such as that of Figure 28.

MICROTEXTURE

Microtexture is what makes a pebble or other aggregate particle feel smooth or rough to the touch. It is difficult to measure and quantify even in the laboratory, but par-

ticularly on pavement specimens. The most commonly used and probably best method is a low-speed friction test. The British Portable Tester is a useful instrument for this purpose (see Chapter Three).

Microtexture. on wet pavements and specimens, governs the adhesion component prominently because it controls the intimacy of contact between rubber and surface by breaking through the thin water film that remains even after the bulk of the water has been displaced. The manner in which microtexture is effective is complex because it affects the molecular and electric interaction between the contacting surfaces (55).

Experiments with artificially roughened surfaces have shown that (low-speed) friction, which is essentially zero on a water-wetted and perfectly smooth surface, increases with increasing microroughness until a peak-to-peak dis-tance of about 40 microns (1 micron= 0.001 mm or 0.00004 in.) is reached and then remains fairly constant.

This explains why the effects of microroughness, and hence the adhesion component of friction, can be quite variable. The dust on the pavement, as well as small par-ticles embedded in the tire, control the finish that is put on the aggregate particles as traffic passes over them. Even with a rolling tire, some motion takes place between tread rubber and pavement as the tire deforms in the contact patch. This, particularly in the presence of loose particles of whatever origin, leads to wear and polishing of the pavement. The pavement, and particularly exposed aggre-gate, may, at any given time, become smoother or rougher than it was before, depending on the size and character of the particles. These most frequently originate frdni the

Figure 26. The .candpatch method of texture ?neasure/nent (The Pen,,svhania State University)

28

pavement itself. but may also be blown in from the shoulders or elsewhere, may be brought in by the tires, etc.

The abrasives can be quite variable. Rain and wind change them, and so do the seasons, particularly if antiskid and deicing materials are applied during the winter. Tire chains and studs, and freezing and thawing, also influence microtexture, even if they otherwise do not affect the surface course in any unusual manner. These processes are complex and subtle, so that one is pretty much limited to surmising them.

Their consequences are, however, observable. For in-

stance, the seasonal fluctuations have been documented repeatedly (Fig. 33), but even shorter-range changes can be observed (as when, for instance, rain follows a long dry period) (Fig. 34). These changes differ in magnitude with the type of aggregate, but obviously the softer ma-terials, such as limestones, respond more rapidly and strongly to extraneous influences.

GENERAL SURFACE PROPERTIES

There is no sharp demarcation between micro- and macro-texture. Hence, ideally the same method would serve to measure both. Schonfeld (56) suggests doing this by means of stereophotographs. He has attempted to relate empiri-

E

.....:.: .41 .---. :.....-...

-i.. -. - . .'. 0 •. .- - .. -. - 2

0

LL

S

al S •

• S. ••

ILl . S

0 S z ILl

01 :

U-

0 10 20 30 40

TEXTURE DEPTH, 0.001 in.

Figure 28. Relation of skid resistance decrease it'ith speed

and texture dept/i (5 1).

TEXTURE DEPTH, QOOl In

Figure 29. Relation of skid resistance-speed gradient and texture

(IC pt/i on variously finished portland ce,nent concrete surfaces

Figure 27. The putt)' method of texture mcasure,nent (50)

(52).

29

I/gore 30. l'i•ofilograp/i.

: .:,";.:.•, p..

-. '?

Figure 3/. Te.viuremeter.

cally seven visually determined texture parameters to skid resistance and skid resistance gradient by means of multi- pie regression techniques. This approach shows promise, although it depends to a considerable degree on operator skill.

If this or some other method or combination of methods can be perfected, it may eventually become possible to dispense with skid resistance measurements altogether, provided the cost/benefit ratio of the substitute techniques is favorable. Aside from Schonfeld, Gillespie (53) and Yandell (57) have made at least partially successful at-tempts in this direction.

However, there are some additional parameters that, at least for sonic pavements or under some circumstances. may have to be considered as inputs into any future skid resistance prediction equation. For instance, the surface course may be porous enough that a significant amount of water can escape from the tire footprint through inter-nal channels. Porosity is discussed in the next section.

Wettability of the surface can play a part in the frictional performance of a pavement. If the surface does not wet readily the water is more easily expelled from under the tire. This occurs most readily in consequence of contami-nation, usually by the oil drippings of traffic. Oil is usually thought to promote slipperiness; it will do so if the oil layer has any significant thickness. The oil, rather than the water, then becomes the separating agent in the contact area and because of its higher viscosity is more difficult to displace and for the irregularities to penetrate. But when the oil film is very thin, this effect is negligible and the oil reduces the wettability of the surface. (The oil even-tually emulsifies in the water.)

Some aggregates have a significant capacity to absorb water. Thus, during the beginning of a rain they become gradually wetter. If their microroughness is minimal, this may produce a measurable lowering of the skid resistance. However, all of these effects are rather small and con-sequently are not overly significant in the design and maintenance of surfaces.

Eventually, when all these factors can be related quanti-tatively to skid resistance, universal specifications for the

characteristics of materials and surface course design can be written, provided, however, that the effect of exposure to traffic. weather, etc., can also be predicted.

Figure 32. Eflect of sharpness of channel edges of pave-meat surface on contact u/f/i the lire.

30

w

I Jan July Jan July Jan

1959 1960

Figure 33. Seasonal change of skid resistance (28).

10

NUMBER OF DAYS

Figure 34. Change of skid resistance of a limestone pavement surface with precipitation (20).

0.4

(0 20 30

50 60 SPEED, mph

Figure 35. Hydroplaning occurs when water depth is excessive (40).

GROOVING AND POROUS LAYERS

Hydroplaning occurs when a tire moves so fast over a fairly thick layer of water on the pavement that it "rides up" on the water and loses all contact with the surface underneath. As either speed or water laye- thickness increases, Zone A in Figure 24 eventually grows so much that Zones B and C disappear.

Hydroplaning was recognized first with aircraft. Landing speeds steadily increased and, because runways are usually level or nearly so, are wide and have little cross slope, the combination of water layer thickness and speed began to prove deadly for more and more aircraft lnding in heavy rain or on slush. Airplanes can literally be blown off the runway if the tires are hydroplaning and there is a cross-wind. Thus, it is not surprising that NASA (58) and the Royal Aircraft Establishment (RAE) have done a great deal of work on hydroplaning and on remedies for it.

On highways hydroplaning is not as readily an apparent accident cause as with aircraft. Most vehicles are driven by nonprofessionals who have difficulty recognizing or reconstructing what led to a particular accident. Police are not well trained in looking for the underlying accident causes and most accident reports are slanted toward fixing blame rather than pinpointing causes. In iddition, hydro-planing accidents really began to occur only after highways began to be improved systematically. This permitted traffic to travel at higher speeds than heretofore and with more and wider lanes the runoff from the pavements was reduced. Simultaneously, because of the improvements in construction practices, it became feasible to reduce lateral pitch. Hence, fairly suddenly, hydroplaning became a new accident cause.

Figure 35 illustrates a hydroplaning case. As long as the water film on the pavement is fairly thin, the friction factor remains reasonably good even at high speed. But when the water is 1/4 in. deep, the friction factor suddenly drops sharply at -45 mph and reaches practically zero at 50 mph. Although the figure shows a cornering friction factor, curves for locked-wheel hydroplaning look quite similar. The speed at which hydroplaning begins depends on tire design, inflation pressure, water depth, and other variables. Fortunately that speed is not always as low as in the present example.

Both tread condition and pavement texture control the hydroplaning speed. The more readily either permits water to escape from the tire-pavement interface, the higher will be the speed at which hydroplaning begins. Numerical data on the relation between texture and the onset of hydroplaning are virtually nonexistent. One difficulty lies in defining water depth. It is usually given as the thickness of the water layer above the peaks of the surface irregularities. This, however, has little practical value, except in cases where puddles or pools occur. When rainfall rate is the independent variable, the water depth as defined varies with texture, pitch of the surface, wind, runoff distance, even when the rate of rainfall is constant. Gallaway et al. (50) provide some data for estimating water film thickness, but more research on this subject is needed before pre-dictive equations for field conditions will be acceptable.

Strongly textured surfaces, however, are definitely de-

31

sirable for the control of hydroplaning. Experiments (50) have shown that. coarse texture is beneficial in raising the lowest speed at which hydroplaning is apt to occur, even though texture slows the runoff speed (59). The govern-ing factor is that the water must be able to escape from the tire footprint.

Grooving has been used with excellent success on port-land cement concrete pavements that have been polished by traffic or because of insufficient runoff have a low hydroplaning limit. The size, spacing, and orientation of the grooves require still further experimentation before firm design rules can be established (60), but the reduction in wet-weather accidents can be spectacular (61), what-ever the design of the grooves.

Longitudinal grooves are preferred for highways because they are most easily and economically cut. With a 0.1-in. width of cut, longitudinal grooves do not seem to be objectionable, even for light motorcycles (62). Because of the uncertainty as to which locations might produce

hydroplaning, one is dependent on accident data to identify them, particularly in areas in which heavy rains or melt-water problems occur only infrequently.

Another way to prevent hydroplaning or reduce the drop of skid resistance with speed is to let the water drain through the pavement by making the surface course po-rous. This can be acomplished by using an open gradation so that the mixture contains 10. to .20 percent air voids. Very little water will then accumulate on the surface, and what there is does not have to be displaced by the tire along the surface, but can escape downward and out through the pavement. A bonus is that spray and splashing are reduced substantially. The technology for such courses is not yet fully developed and the life of open-graded mixes, especially under heavy traffic and in northern climates, is still uncertain. Binder modifications and the use, of highly angular or vesicular aggregates may be necessary to assure that these and other necessary properties of such wearing courses are maintained over sufficiently long periods.

CHAPTER FIVE

DETERIORATION OF SKID RESISTANCE

As traffic rolls over a pavement various things happen. The pavement is compressed, and unless it fully rebounds it will rut and eventually fail. In flexible pavements aggre-gate particles can become dislocated or reoriented and binder material can migrate to the surface. These effects are more prevalent on roads that carry heavier vehicles than those that do not.

The tires scrub and squirm. Even when a tire merely rolls, at least some of the tread elements in the contact patch move relative to the pavement surface. This occurs only after the prevailing local friction has been overcome: shear forces develop in the contact patch and they can dislocate particles or at least help to reorient them. Once a tread element moves it tends to cause abrasion and changes in the surface characteristics of the aggregate particles. These phenomena are amplified when torque is applied to the tire, whether, it is to maintain constant speed, or to accelerate and decelerate.

These are complex phenomena, all interacting with one another. Moreover, they have recently been further compli-cated by the widespread use of studded tires. Such tires add another component that is destructive to surfaces: as a stud is moved into contact by the tire it impacts on the surface with considerable speed and force. Tread elements do so too, but because they are resilient this is not serious, although tread design contributes to the noise that tires generate.

The phenomena vary with a long list of factors. In this chapter they are discussed under several headings, chosen

more for convenience than in accordance with some rigor-ous classification.

It is important to keep in mind that the lasting effects that traffic can have on a pavement fall into three broad categories—changes in the surface structure, wear, and polishing. The first of these categories does not involve loss of material but includes the reorientation of aggregate particles, binder migration, etc. The second category com-prises everything that involves permanent loss of material; the third refers to changes in the surface microtexture, principally of the aggregate particles.

Polishing does not, by definition, involve loss of material but refers only to the smoothing of a surface. This could be the result of plastic flow of the material being polished, but mechanisms involving loss of material are not ex-cluded; indeed, they are more commonly operative in the case of pavement materials. Wear, on the other hand, refers specifically to the loss of material, however gradual. It is therefore measured as weight or volume lost, decrease in thickness, etc. The degree of wear says nothing about the change in surface smoothness or polishing. Indeed, certain types of wear reduce the smoothness of a surface.

WEAR AND THE LOSS OF SKID RESISTANCE

Any kind of pavement wear is undesirable, unless accom-panied by desirable side effects of sufficient magnitude, because it means that the original surface, whether an overlay, the top course, or the top layer of any kind of

pavement, is being removed or altered. Because wear is hardly ever uniform, it is merely a question of whether wear makes a road unserviceable because of the uneven-ness of the surface or because of the complete disappear-ance of the surface course.

Wear either can be gradual or it can occur in a stepwise fashion, as by breaking out of aggregate particles. After asphalt has oxidized and is eroded, the shear stresses be-tween tire and surface may be sufficient to break out particles. This type of attrition is promoted by studded tires, and can become catastrophic on surface treatments, because the shear forces are augmented by localized impact and other forces. Brittle materials are particularly sus-ceptible to this type of attack.

Generally, though, wear is of the gradual, continuous type, at least as far as individual aggregate particles are concerned. This wear is greatly promoted by loose mate-rial, whether generated by the pavement itself or blown or brought in from elsewhere. The loose particles serving as abrasives may be the size of fine sand or smaller. After ice and snow are gone, the "antiskid materials" used for ice control may provide, temporarily, much larger par-ticles, but these are usually quickly broken up by traffic or thrown off the pavement. It is largely the fine material that serves as the abrasive.

If the surface aggregate and the filler consist of homo-geneous, amorphous stone, wear will be comparatively slow, unless an outside source furnishes more-abrasive material. Skid-resistant aggregates are rarely homogeneous and amorphous. The debris from them will then contain particles that are harder than the matrix in which they were originally imbedded. These hard particles help to wear the matrix away, releasing more hard particles to replace those that are ground down and blown off the pavement.

Wear is very much a function of the aggregate charac-teristics. It is not fully understood in all its ramifications and complications that exist on a highway surface. At-tempts have been made to relate aggregate characteristics to probable wear rates, but little success has been had in generalizing the findings, much less relating them to high-way experience. Stuffier (63) concluded from his research on abrasive wear that road stones wear by scratching and pitting and that rapidly wearing materials exhibited evi-dences of both types of attrition, whereas slower wearing ones showed mostly scratching. He established that the hardness of the mineral relative to that of the abrasive is decisive for the rapidity of wear: if the abrasive is harder than the mineral, wear is fast. The wear rates he found correlated well with the hardness of the mineral of the road stones, with some exceptions, however. The exceptions were those minerals that showed little pitting. Stuffier found no adequate general model for relating the rate of wear to known properties of the minerals, but he did develop some qualitative relations.

Aggregate wear is not identical with pavement wear. This is particularly true for portland cement surfaces. The relative hardness of mortar and aggregate can make the wear rate drastically different from what it would be for either mortar or aggregate alone.

From the skid resistance standpoint, differences in the wear rates of components of surface courses are desirable. This applies equally to rigid and flexible pavements, as well as to all overlays. The different wear rates enhance surface texture, thus providing channels through which water can escape from the contact patch. The surface irregularities also increase the hysteresis component of friction.

Portland cement pavements are given increasingly heavier surface texturing during construction. Wear eventually obliterates this type of texture. Because it is essentially formed in the mortar, the latter must be made hard and wear-resistant if the original finish is not to be lost rapidly.

Wear is promoted by the presence of abrasive material. Obviously, the finer the particles constituting this material are, the smoother the resulting wear surface will be; but the wear rate will also be slower. Consequently, for similar materials, slower wear means more polishing, hence lower skid resistance. Because, however, there is no practical way of controlling the type and quantity of abrasive mate-rial on the pavement (except in some instances where foreign material is being brought onto the pavement), there is little that can be done to influence this process, except by judicious design of the surface course.

POLISHING

Just as the wear mechanism of aggregates and pavements is still largely a mystery, so is the polishing mechanism. Chiefly this is so because for minerals the two phenomena are not strictly separable. Minerals do not deform plasti-cally to any significant extent and therefore polishing re-quires the removal of material, albeit in very small incre-ments.

Polishing is reduction of microtexture. Consequently, it is difficult to measure and this further complicates research on its mechanism. The researcher can try to obtain micro-profiles, can try to quantify ordinary or stereophotographs of the surfaces, or can use other methods that researchers have employed from time to time. Aside from cost and effort, the chief difficulty is that only very small samples are obtained and that the methods are difficult or im-possible to use in the field.

Therefore, one uses low-speed friction measurements instead. The British Portable Tester is employed most frequently. At low speeds the adhesion component of fric-tion is dominant; it is primarily a function of the micro-texture of the contacted surface. Such measurements are, of course, made with the surface wetted, but the rubber shoe of the pendulum tester displaces enough water so that the hydrodynamic effects (which influence friction at higher speeds and which are controlled by the coarser features of the surface) are practically absent. The rubber responds essentially only to microtexture.

Unfortunately, the results of laboratory experiments on the polishing characteristics of aggregates and pavement specimens are not readily related to the skid resistance as measured on the highway. The latter is obtained and rated at 40 mph, although in the laboratory such speeds cannot be used. This means that the highway measurements con-tain the effects of macrotexture although those made in the

33

laboratory do not. Thus, the correlation between labora-tory and field data is not too good; but, in general, the ranking is about the same, so that laboratory polishing has become a more and more frequently used adjunct to surface course design.

Polishing and wear are not only greatly accelerated by the presence of an abrasive, but the degree of polish de-pends directly on the size of the abrasive particles, as shown in Figure 36 (64): the finer the abrasive is, the lower the friction factor is. Although there is unquestion-ably always dust on a roadway, its nature is a matter of conjecture. A major proportion is normally probably debris of the aggregate. Other material is brought in by the tires, or drops from vehicles, or is blown in from the shoulders and adjoining land.

The effects of such a melange of materials, sizes, and shapes is difficult to assess and cannot be duplicated in the laboratory. Hence, for laboratory experiments arbitrary selections of polishing agents are made. These usually consist of fine gradations of materials much harder than any pavement components, if only because hard powders accelerate the polishing process.

On the road, most of the polishing probably takes place while the pavement is dry, because pavements are dry most of the time. In laboratory experiments, polishing is fre-quently being done wet, partially to eliminate the dust nuisance, but more importantly to cool the rubber pad or tire so as to prevent the rubber from smearing. If the interface were not cooled by water, much lower polishing rates would have to be employed and this would defeat the objective of accelerating the process compared with that taking place in the field. It is, of course, the number of pases, whether of a pad or a tire, that governs the polishing rate; but on the average it takes two or more years until a pavement in the field has been polished to its final state. (The skid resistance may, however, continue to drop because wear continues and therefore the macro-texture continues to be reduced, causing the conditions for clearing the water from under the tire to become poorer.) In the laboratory, aggregates and pavement specimens are usually polished to the equilibrium state in a matter of hours or minutes.

There has as yet not been any standardization in labora-tory polishing procedures. A review of various methods in use is given by Goodwin (65). No one method has proven itself as clearly superior to others. It may well turn out that the simplest one is the most practical and useful one. The simplest method for aggregate polishing is clearly tumbling in a jar mill (66), although it is still necessary to mount the tumbled stones in an epoxy matrix for determin-ing their friction characteristics.

However, full confidence in the validity of all laboratory methods is missing, at least when relatively small differ-ences in the resistance to polishing are of importance, as when specifications are to be drawn up. The difficulty lies in the fact that, in the field, skid resistance is highly variable, both seasonally (Fig. 33) and over short periods (Fig. 34). It is therefore difficult to determine accurately what the equilibrium skid resistance of a pavement is. The situation is further complicated by the fact that the amount

Grade

POLISHING DURATION Figure 36. Effect of abrasive size on the polish obtained in the laboratory on a limestone aggregate (64).

of traffic carried by a pavement plays a part in determining what the mean equilibrium skid resistance is.

Figure 37. illustrates this point. It shows the skid resist-ance of one type of surface course that was installed at the same time in six different locations. The data points were obtained three years after installation. The skid re-sistance had stabilized at all locations after two years. The daily traffic is the average over the three-year period. Clearly, traffic is the governing factor, not time. Interest-ingly, the exposure to trucks gives a better correlation than that to all vehicles (one truck is equivalent to about three

060

ç. \t

2 4 6 10 20 40 60 100 AVERAGE DAILY TRAFFIC, thousands

Figure 37. Deterioration of skid resistance with exposure to traffic (67)

BRITISH PORTABLE NUMBER, BPN BLEEDING OF ASPHALTIC PAVEMENTS

34

30

Unee I

45 60

Gravel

Granite Gnels Granite

Slate Expanded Slate

Sandetone

Figure 38. Ranking of various aggregates according to final fric-lion factor after polishing in a laboratory machine (68).

passenger cars; however, equivalencies up to eleven to one have been used) 7

Of late, studded tires have introduced an additional in-fluence. Any polishing effect by studs has not been docu-mented and is probably unimportant. Studs seriously stress the surface over a large area, either cutting macrogrooves or fracturing and dislocating aggregate particles. The new surfaces so created are likely to be less smooth. This method of removing polish is very costly in pavement surface wear and has led to the barring or limiting of studs on tires.

The most important surface components in determining the mean microtexture of a pavement are generally the exposed faces of the aggregate particles. The mineralogy and structure of the aggregate determine how easily and how much it polishes. Figure 38 shows that aggregate type determines approximately the ranking that a particular one will take with respect to other types. But there can be rather broad ranges for any one type. For instance, lime: stones will polish more or less depending on how much siliceous material, and in what size, they contain (69). Shupe and Lounsbury (70) have shown that any material harder than calcite will improve the skid resistance of lime-stones. Grain shape, size, and size distribution also play a part in the polishing susceptibility, as do other petrographic properties.

Much still needs to be learned about how and why aggre-gates polish; but some generalizations are possible. Dahir and Mullen (68) give this summarization: The highest permanent skid resistance is obtained by aggregates whose sacrificial surfaces are continuously renewed by traffic action. The content of sand-sized hard particles embedded in a softer matrix has greater beneficial effect on skid resistance than the total amount of "impurities." When mineral composition of an aggregate is uniform, its skid resistance will be higher the more angular and the larger the grain is. One might add that vesicular materials can also have good skid resistance even if the material per se

tends to take on a high polish. Hardness is not a virtue in a uniform aggregate, such as quartz pebbles.

The same basic considerations apply to the pavement as a whole, whether it is of bituminous or portland cement concrete. A pavement will retain good microroughness if its irregularities are sand-size and sharp-edged, and if the surface wears fast enough to release the hard particles that form the irregularities before they have worn off or become too well rounded.

Texture, whether macro or micro, can be lost by plastic smoothing of the surface by bleeding asphalt. How "bleed-ing" is caused is immaterial here and should be of historical interest only, because modern asphaltic paving technology knows how to work with such low asphalt contents that bleeding will no longer occur. However, there are cases in which the equivalent of bleeding can occur. For instance, as a surface treatment wears off, the sealant can give the now exposed original surface a smoothness equivalent to that of a bleeding asphaltic surface.

The pores of vesicular materials or the surfaces of other irregularly shaped materials can be filled in and smoothed by binder material. This most frequently occurs during the manufacture of the mix and explains why some newly placed pavements show an increase in skid resistance over a sometimes fairly long initial period: the asphalt in the pores must be worn away or eroded. In hot weather new asphalt may be deposited in the pores of mixes having a high asphalt and a low surface void content.

COMPACTION, RUUI NG, PARTICLE LOSS

Traffic tends to compact any flexible pavement in the wheel tracks. Mix design adequate to the existing or anticipated traffic loads minimizes this effect. However, it is useful to review the consequences of compaction on skid resist-ance.

Compaction tends to bring more binder material to the surface, and hence into contact with the tires. This in-variably decreases the over-all microtexture of the surface. Compaction also reduces macrotexture; thus, it is detri-mental to good skid resistance at all speeds.

Another aspect of compaction is that it permits ruts to form if traffic is strictly channeled. The ruts fill with water to a greater depth than would be present otherwise. This can lead to hydroplaning at lower speeds and increases in splash and spray presenting, certainly on heavily traveled roads, an added traffic hazard. (Ruts can also be formed by excessive wear rates, for which studded tire use is the most likely cause.)

In pavements in which compaction occurs, the aggregate particles are not necessarily just pushed downward while the binder moves upward. The larger aggregate particles may also be reoriented. This causes their large faces to become aligned with the surface. If the aggregate is sus-ceptible to polishing this can lead to rapid degrading of the skid resistance. Polish-susceptible aggregate can give acceptable skid resistance in flexible pavements if the particles are held securely in random orientation. This is particularly true if the aggregate has preferred cleavage planes. The stresses imposed by traffic can produce new fracture faces and therefore expose sharp edges and good texture to traffic. Thus flexibility in surface courses should be held to the minimum dictated by structural considera-tions.

Various considerations discussed up to this point have called for low asphalt content. However, it is important that the content not be so low, especially in coarse-graded mixes, that the exposed aggregate particles are not firmly

al

anchored. If they are not, they will eventually be torn out. Particle loss is accelerated by studded tires, which may, in addition, also fracture and split the aggregate particles. This type of attrition can be very rapid and aids in the formation of ruts. Purely from the skid resistance stand-point such wear is not objectionable, because it brings new, unpolished particles to the surface and maintains channels for water escape. However, this does not make the forma-tion of ruts acceptable, if only because whatever is gained in skid resistance in light rain is lost by the lowering of the hydroplaning limit when the ruts fill with water.

STUDDED TIRES

There is no longer any argument that studded tires cause excessive pavement wear (71). Several states and provinces now prohibit or limit their use.

Numerous investigations have been made on the effec-tiveness of studded tires (72, 73, 74). In loose snow and on sanded or cindered surface, studs offer no advantage. On ice they are most effective just below the freezing point and the effectiveness diminishes as the ice temperature gets lower. On ice at 32° F, they can reduce stopping distance by about 20 percent for each axle with studs in the tires and, with studs on all four wheels, a curve might be taken safely at 40 mph when without them 30 mph is apt to cause spinout (74). The number of studs per tire does not bring proportional improvement: above 100 very little is gained, but having studded tires on all four wheels is defi-nitely beneficial, particularly in cornering. However, pave-ment wear is in first approximation proportional to the number of studs that come in contact with it.

Most highway departments in the United States today try to implement a bare-pavement policy. The question, therefore, is: Do studded tires reduce frictional performance on bare pavements, dry or wet? There are not as many data on this as one might wish, but the evidence points to a reduction in the available friction factors. Kullberg and Ohlssen (72) cite a reduction from 10 to 20 percent. Although the numbers may be uncertain, there is agreement that on dry or wet pavements studded tires reduce the skid resistance. This, coupled with the observation that conditions under which studded tires do offer a safety gain exist for only 1 to 2 percent of the time during the winter months, has caused several agencies to prohibit or to limit the use of studs.

The real question—whether or not over-all accident rates are improved—has not yet been answered. Norway, where winter use of studded tires is now practically universal, has not found any change in the ratio of winter to total acci-dents over the 10-year period in which stud use increased 500 percent. Either the reduction of accidents on icy pavements is balanced by their increase on bare pavements, or drivers travel just as close to the safe limit with or with-out studs.

The problem, therefore, is to decide if the added main-tenance cost to highway departments of correcting the damage done to pavements by studded tires is justified by their real and/or imagined benefits to the traveling public. The solution is not a simple one, because in addition to the discussed aspects there is the indisputable fact that studs

Cl)

TIME Figure 39. The change of skid resistance during a s/lower as perceived by drivers.

compensate for pavement polishing (Preus (71) found the skid resistance on badly stud-eroded pavements to be prac-tically identical in and between the wheel tracks.)

CONTAM INATION

The changes in surface characteristics discussed so far • were the more or less permanent ones, but there are, in addition, highly transient changes. For instance, according to a generally accepted axiom pavements are more slippery at the beginning of a light rain than they are later on during the rain. This is shown in a general way in Figure 39. The precipitous drop in skid resistance at the beginning of a rain is very transient. No good d2rnflation_forit

(exists. The dip may be real or it may be imaginary in that drivers initially continue to drive as if the pavement is dry, but then discover that it is slippery-and subconsciously

"...adjust their driving to the new condition. Skid tests made with controlled water applicationjp-

port the latter view. Figure 40 shows the generalized find-ings of tests made by several researchers when water is applied to different, previously dry sections of the same pavement at increasing rates. The skid resistance steadily drops until a certain water rate is reached. (If the water is increased beyond the range of the figure, the skid resist-ance begins to drop again, and rather rapidly, as the hydro-planing condition is approached.) This corresponds to the sudden onset of rains of varying intensity. The curve refutes the existence of the dip shown in Figure 39.

That there are time effects is illustrated by Figure 41. When the same pavement section was retested after 15 minutes with renewed water application by the self-watering skid tester, the skid resistance had dropped significantly. This can be explained by absorption of some water by the pavement and the thereby increased wettability of the surface. The dashed curve, obtained on the same pavement when dry, but with a wetting agent added, shows that wettability indeed reduces skid resistance. Eventually, plain

0 25 50 75 100 125 DRYING TIME, mm

36

Figure 40. Skid resistance change with the rate at which water is applied to a dry pavement (artificially or by na-tural rain).

water wets the pavement as well as if a wetting agent had been used. When wettability is low, as when there is a light oil or grease film on the pavement, the water present is more readily displaced by the tire than if there were perfect wetting.

When contamination is heavy, the skid resistance may be affected adversely. For instance, if oil or another substance with good lubricating properties is present in any significant amount, the tire-pavement interface will be lubricated whether or not there is water, hence friction will be low. However, if the contaminant film is as thin as it usually is

ib 20 30 40 50

SPEED, nh

in the wheel tracks, and the main contaminant is water, the previously given description of its effects applies. Thick films occur only because of accidental spills, or possibly from the accumulation of oil drippings between the wheel tracks where traffic is rigorously channeled.

Other contamination may be solid—windrows of antiskid material left after snow and ice have melted, and similar accumulations of foreign material. As long as the tires roll, such contaminants can be dangerous because the tire rolls onto them and the particles act like the balls in a bearing. When the tires are locked, usually only the smaller particles get into the tire contact patch and their effect on friction is minimal.

Locked-wheel skid resistance measurements are not the best way for appraising the extent to which contaminants are dangerous to traffic. Although emergency stopping dis-tances may not be significantly increased, the reduction in the control limits during cornering or braking with the wheels rolling may suffer more severely. Because of the lack of suitable testers these questions have not been thoroughly investigated, particularly since contamination is usually local and temporary.

One other type of contamination—that by deicing agents—is worth mentioning because it occurs frequently. After ice and snow have been melted by them, the pave-ment dries, usually slowly because of the prevailing low temperatures. The deicing agents further delay the drying process with an attendant effect on the skid resistance, as Figure 42 shows. This should not be considered an indict-ment of deicing agents, but it does serve as a warning to apply them no more generously than necessary for the job to be done. The delay in drying is a function of the amount of deicing agent present.

The term "viscous hydroplaning" is sometimes applied to the reduction of the available friction by contaminants or by thin water films, particularly when the water contains a large amount of road dust. This may be a useful con-versational term; but it really designates only a degree of

0 le U)

Water Will WetfinU Ager

Figure 41. Effect of time and of a wetting agent on the skid Figure 42. Delay in the drying of a wet pavement when deicing resistance of a smooth tire (55). agents are present (75).

37

a phenomenon that exists at all times because there is no known means of removing all water or other contaminating material from the interface between tire and pavement.

ROUGHNESS

When a road is rough (that is, when the pavement is more than slightly uneven) vehicle ride also becomes rough. This may cause a reduction in the skid resistance that the vehicle can develop, certainly when the pavement roughness is such that wheel hop occurs when the brakes are applied.

With modern vehicles on modern highways this is not likely to happen because both, even under the worst of circumstances, are sufficiently well maintained so that wheel hop will rarely be induced. Sometimes an empty trailer of a tractor-semitrailer combination will develop wheel hop during an emergency application of the brakes. This is, however, more the result of the characteristics of the trailer suspension than of pavement condition.

If pavements are so maintained that ride is reasonably good, wheel hop or significant reduction of the available skid resistance is not likely to occur. Little hard evidence is available about this, but investigation with skid trailers on rough (though still serviceable) pavements has shown that roughness should not degrade the measured skid resist-ance. The common criteria for acceptable pavement rough-ness would therefore seem to be also adequate for the prevention of degradation of skid resistance due to rough-ness.

CYCLIC EFFECTS

Cyclic changes in skid resistance have been mentioned earlier in connection with the discussion on texture. Their magnitude is generally small and they are often as much caused by the accompanying changes in temperature as by those of the texture itself. For instance, if a skid resistance measurement made immediately after a summer rain (after the pavement has dried, and with wetting by the tester) is compared with one before the rain, the former is likely to show a higher skid number than the latter. The difference is probably caused, in part, by the increase in microtexture that can occur during the rain, but also by the lower tem-peratures during the second measurement (in general, skid resistance increases as the temperature decreases; pave-ments are not affected by temperature, but tires are).

Such changes are usually not large or consistent enough to have much significance for traffic, but they may not be negligible for the interpretation of skid resistance measure-ments. Figure 43 shows why (the figure is purely illustra-tive): superimposed over the gradual change of average skid resistance of a pavement with the years are annual and shorter-term fluctuations. If the pavement is measured periodically, wrong conclusions may be drawn if the mea-surements are not so spaced as to minimize the influence of these skid resistance fluctuations. By correcting the measurements for temperature, their effect can be reduced.

Figure 43. Mean and actual skid resistance of a pavement dur-ing several years after construction (S=summer, W=winter).

RUNOFF CHANGES

The importance of providing for fast removal of the water that rain (or any other cause) deposits on the pavement has been stressed. Changes in the pavement contour usually reduce runoff and drying. Rutting and increasing pavement roughness increase the local water depth and also increase the time required for the pavement to dry completely once the rain has stopped or snow and ice have melted. In winter, drying speed can be very low.

Traffic helps remove water from the wheel tracks by splash and spray, and in heavy traffic this method of water dispersal can be more important than natural runoff, at least during a light rain. But even then, low spots will delay the drying process. In heavy rain and where travel speeds are high, the water depth may become so great that hydroplaning may be induced. Although the vehicle may move out of the puddle rapidly, hydroplaning may continue beyond it because the speed at which a hydroplaning wheel will recover is lower (or the critical water depth is less) than that at which it begins to hydroplane. Thus, the effect of the increase in pavement roughness with the life of a pavement may not noticeably change the measured skid resistance (because it is measured with prescribed constant water film thickness), but traffic may still be endangered by the increased tendency of vehicles to hydroplane.

Changes in the water depth on a pavement may also occur because water runs onto the pavement, such as when drainage ditches overflow or when snowbanks melt. Some such instances are unavoidable, but others can be prevented by design or maintenance.

38

CHAPTIR SIX

DESIGN AND CONSTRUCTION OF SKID-REStSTANT SURFACES

The design and construction of skid-resistant pavement surfaces requires a thorough knowledge of the factors that influence the skid resistance of surfaces in service. In the design stage, consideration must be given to the selection of skid-resistant aggregates and ways of maximizing their use in pavement mixtures. After the design has been estab-lished, its effectiveness can be enhanced by well-planned and well-carried-out construction procedures. Such pro-cedures are a matter of using established quality control measures.

DESIGN OF PAVEMENT SURFACES

The design of the initial pavement wearing surface is a primary factor in achieving a skid-proof highway. In addition to being skid resistant, the surface must also per-form several other important functions. It must protect the underlying layers from effects of weather, exhibit smooth riding qualities, not cause excessive tire wear and noise, and be structurally stable to withstand the forces imposed by vehicle maneuvers. Also, it should be pleasing to the eye and not exhibit undesirable qualities either for daylight or nighttime driving tinder the most adverse con-ditions. The challenge to the pavement designer is to opti-mize all of these factors in order to meet the public's desire of a highway that meets all the attributes of safety, comfort, and convenience.

The design must take appropriate cognizance of the types of materials used and the manner in which they are combined. Generally, the materials should be polish-resist-ant, have low wear characteristics, and provide a textured pavement surface when placed in the roadway. A well-designed pavement surface that has exhibited excellent skid resistance is shown in Figure 44.

Selection of the appropriate materials for a pavement surface mixture that will, when properly placed, maintain a reasonable level of skid resistance, requires knowledge of their characteristics and their past performance. Because the aggregates in either a bituminous or a portland cement concrete mixture comprise in excess of 90 percent of the total mixture, they arc the primary constituent that the pavement designer must consider.

Studies of aggregates and their performance in pavement surfaces have been approached by laboratory analysis of individual particles as well as laboratory and field testing of the aggregates when used in pavement mixtures (69, 76). Final evaluation must be predicated on field per-formance under actual traffic conditions. Characteristics of aggregates that influence skid resistance are type, size (also gradation), shape, texture (micro and macro), hardness (susceptibility to wear and resistance to polishing), and mineral composition (70). These characteristics are inter-related to such an extent that their separate evaluation would be difficult, and probably impossible.

Figure 44. Asp/ui/tic concrete palenzcn( lilt/I excellent skid resistance.

0

.6

I . 0 2 4 6 8 10 12 14

TIME INTERVALS

-MEMO I I

•

Tuj

39

SURFACE AGGREGATES

Suitable aggregates may be either natural or synthetic (77). The most commonly used natural aggregates are those obtained by quarrying and crushing rocks such as limestone, sandstone, granite, diabase (Fig. 45a). They may also consist of stream and bank gravels that are ob- tamed by dredging, washing, and screening, and usually are crushed to improve their angularity. Their shape is critical insofar as the skid resistance is concerned. A non- crushed natural aggregate obtained from river dredging is shown in Figure 45b. Note the roundness and lack of angularity. The difference in angularity of the two materials is immediately apparent upon comparison of these two natural aggregates derived from different sources. If crushed, the river gravel (Fig. 45c) has angularity that is similar to that of the crushed stone. If angularity were the only factor contributing to their skid resistance value, their performance would be similar. This is not the case, as their microtexture, chemical composition, and crystalline structure differ widely. These differences are sufficient to also make their polishing characteristics different, as shown by laboratory test results (Fig. 46).

Synthetic aggregates are obtained from processing a wide variety of raw materials that may be either natural or artificial (79). Physical characteristics of different syn-thetic aggregates vary considerably, depending on the source material and the manufacturing process. Because of these differences, they can be expected to vary widely in their performance when used in the pavement surface. In general, however, they are less susceptible to polishing than are the natural aggregates, but they may tend to abrade more rapidly.

Synthetic aggregates that are manufactured by heating natural clays and shales are called "expanded lightweight aggregates." A typical expanded shale aggregate is shown in Figure 47a. Note the strong microtexture and high angu-larity. Synthetic aggregates (slag) may also be obtained as a by-product from blast furnaces in the manufacture of iron and steel (80). Figure 47b shows an air-cooled blast furnace slag aggregate. It may be noted also that it is highly angular and has a rough, pitted, vesicular surface texture. Although calcium silicate slag is obtained as a by-product from electric furnaces in manufacturing phos-phate (81), a "foam" as well as a "hard" slag is also obtained. The hard slag results from cooling the molten material by air and water, whereas the foam slag is obtained by treating the molten' material with water, steam, and air. In areas where these aggregates are available, the hard slag is the type most frequently used in pavement surfaces. Figure 47c shows such a hard slag.

Other sources of synthetic aggregates include fly ash that is produced in the combustion of coal, and waste products from the glass, brick, tile, and other industries. Synthetic aggregates have texture and chemical characteristics de-cidedly different from those of natural aggregates such as limestone, sandstones, and gravels, and therefore have widely differing skid resistance qualities.

Figure 46. Laboratory polishing performance of different aggre-gates (78).

AGGREGATE CHARACTERISTICS

The physical, chemical, and mineralogical composition of aggregates is reflected in their surface texture, shape, and susceptibility to polish-wear. Aggregate characteristics that are most significant for skid resistance qualities are those related to their susceptibility to polishing and wear (82). The degree to which an aggregate exhibits the desirable qualities is dependent on its mineralogical composition and internal structure (68). Inherent characteristics, such as texture, chemical makeup, and physical properties, are controllable by the pavement designer only through the process of source selection. Once an aggregate source is selected, certain other characteristics may be controlled within limits by processing techniques. These include angularity, particle size, and range of sizes. In making his selection, the designer should know that certain carbonate rocks give satisfactory performance whereas others perform poorly (83). Under many conditions of use, fine-grained oolitic limestones that consist essentially of pure calcium carbonate (low silica content) are poor performers, whereas the dolomitic limestones with high silica content are among the satisfactory performers. The noncarbonate rocks, such as sandstones, granite, and diabase, are generally acceptable (69, 85). The most frequently sought after characteristics for a skid-resistant aggregate are:.

Resistance to polish-wear. Texture. Shape. Size.

Resistance to Polish-Wear

The ability of an aggregate to resist the polish-wear action of traffic has long been recognized as a most important characteristic (85). When an aggregate becomes smooth,

(b) River Gravel

Fw

& TV

(a) Limestone

liE

(c) Crushed River Gravel

Figure 45. Natural aggregates, crushed and uncruslied.

40

I

41

(a) Expanded Shale Aggregate

AN

I I j

L__ -.

3

(b) Air-Cooled Blast Furnace Slag

(rT9Tj1

r , I Iv -m 2 - 3

77T , ., 3

M.

(c) Hard Slag - Produced by Electric Furnace

Fiç'ure 47. Synthetic a,qgregates.

42

I,]

"/artzfs par

hours of

--------Th('sond obrosion

Dolomite

0 10 20 30 40 50 bU (U Time - Minutes

Figure 48. Laboratory evaluation of polish-wear of concrete (86).

it will have poor skid resistance. Also, if it wears (abrades) and polishes too rapidly, the pavement will lose its texture and become essentially a terrazzo-type surface that will be slippery under wet conditions (85). Laboratory tests have been developed to evaluate this polishwear phenomenon for pavement mixtures (65, 86). Figure 48 shows the results from one type of laboratory evaluation in which a rubber-tired wheel rubs against a pavement specimen. Water cooling is provided and the vertical load is held constant. The power consumption of the drive motor is used as a wear index. The wear index decreases with time of test as an indication of the surface becoming more slippery. A pavement surface that has exhibited this type of polish-wear is shown in Figure 49. The skid number for this surface is well below the minimum desired for a skid-resistant pavement.

The polish-wear characteristics of an aggregate are not readily predictable from its physical and chemical makeup. Studies have been directed toward developing better under-standing of the polish-wear phenomenon (63). It is known that the attrition of loosely cemented grains in an aggre-gate particle will lead to renewed surfaces, consequently less polishing occurs (87). Also, certain mineral types will polish-wear more readily than others. If aggregates of known field performance are studied by means of photo-micrographs, an indication of their susceptibility to polish-wear may be obtained. Photography has proven useful in identifying the presence of coarse grain sizes and gross differences in grain hardness. These two factors appear to combine and lead to differential wear and plucking out or shearing of grains that result in a constantly renewed abrasive surface (76). Also, comparisons of photomicro-graphs of thin sections of aggregates tested in a wear machine have led to the identification of certain minerals associated with good skid resistance qualities. For example,

this approach established the superior performance of dolomitic limestone over relatively pure carbonate lime-stone (82).

Texture

The surfaces of individual aggregate particles may be grossly described in terms of their microtexture and macro-texture. Microtexture can be used to describe surface coarseness as governed by the size of individual mineral grains and the matrix in which they are cemented. Macro-texture refers to the angularity of the aggregate particles and the voids and pits in the pavement surface (88). Con-tinuing research on the evaluation of surface texture should lead to a better understanding of the relationship of aggre-gate texture and skid resistance. Such factors as grain roundness, interlock, size, and distribution provide micro-texture information useful for correlation with skid resist-ance. As might be suspected, an aggregate with larger-than-sand sizes of hard grains, and weak cementation of the grains, will wear under traffic and expose a continually renewed nonpolished surface. On the other hand, if the matrix of the aggregate is strong, the individual grains will be tightly held and consequently may be polished by traffic (87). The rate of polishing depends on the hardness of the grains, the frequency of contact by the traffic, and the media (such as dust and grime) on the roadway surface. If the surface contaminants are abrasive and harder than the mineral grains, polishing of the aggregate will be ac-celerated. For an aggregate to exhibit satisfactory skid resistance properties, it probably should contain at least two mineral constituents of different hardness in order to wear differentially and expose new surfaces (76). As stated earlier, certain carbonate aggregates have been re-ported as being satisfactory performers if they contain more than 10 percent sand-size quartz particles distributed in a weak matrix such that differential wear may occur. Car-

43

Figure 49. Asphaltic concrete paveine,i:, showing polish and wear (78).

bonatc aggregates containing less than 10 percent sand-size insoluble residue probably will be inferior for skid resist-ance purposes (89). Figure 50 shows the curvilinear rela-tionship between the skid number derived from automobile stopping distance tests and the percent total insoluble resi-due for grain sizes greater than the 200-mesh screen (90). This type of information can be developed for different aggregate sources.

Aggregrate Shape

The shape of an aggregate particle significantly affects its skid-resistant properties. Aggregate shape depends on many of the same factors that influence its texture. These include hardness of grains, strength of the matrix, and over-all resistance of the aggregate to abrasion. Also, processing procedures will govern the shape of both natural and syn-thetic aggregates: crushed aggregates exhibit an angular shape, whereas those from stream beds exhibit a rounded shape. The angularity of an aggregate contributes to its skid resistance qualities as long as it remains angular. Angularity relates to manufacturing process, but the reten-tion of angularity depends on such characteristics as mineralogical composition and the amount of polish-wear produced by traffic. Some minerals will crush into mostly flat and elongated particles and as a consequence will be poor performers.

Aggregate Size and Gradation

It has become apparent in recent years that the size of an aggregate influences its skid resistance qualities. Size, however, must be considered in relation to pavement type

44

42

40

38

36 z U) 34

32

30

28

26 0 2 4 6 8 10 12 14 16 18 20

Totat Insoluble Iesldue, Percent Figure 50. EJ]ect of insoluble residue on polishing resistance (90).

44

and mix design. Generally, larger-size aggregates in bi-tuminous pavement mixtures have greater control over the skid resistance than do the smaller-size aggregates. For concrete pavements, however, the sand-size aggregates con-trol the performance of the pavement surface, at least for the period during which the surface has not worn to the extent of exposing large quantities of the coarse aggregate (78).

Aggregate size and gradation must be taken into account in the design of pavement mixtures. The design of skid-resistant bituminous pavement surface mixtures must follow different rules from those used for concrete pavement surfaces.

PAVEMENT MIXTURES

After the appropriate aggregates have been selected for the pavement surface, a mixture is designed to achieve the best skid resistance qualities compatible with the desired strength, durability, and riding quality. The several factors that may be varied by the pavement designer in an attempt to effect a balance among these qualities include:

Blending of aggregates. Aggregate size and gradation. Relationship between aggregates and the asphalt or cement binder. Construction methods, including finishing techniques.

These factors are managed in the mix design process to effect the desirable surface texture, pavement density, etc. The design method may be one of several, but probably most agencies will use the PCA method for concrete pave-ments and the Hveem Stabilometer or the Marshall Sta-bility method for asphaltic pavements. Design criteria em-ploying these methods may be found in state, AASHO, and ASTM specifications.

Blending of aggregates to obtain the desired qualities is resorted to when superior quality aggregates are in limited supply and/or processing costs are prohibitive. Blending is frequently accomplished by combining a natural aggre-gate with a synthetic aggregate (91). Most frequently, one of the aggregates will comprise the total amount of either the coarse or fine aggregate in the mix. In blending, the two aggregates are separately graded and then combined to meet a master gradation that has been previously selected on the basis of the desired qualities for the pavement surface.

The maximum size aggregate, as well as the mix grada-tion, may be varied by the pavement designer to produce within tolerable limits the desired surface texture and pave-ment density. For concrete pavements, this means that the surface texture will be controlled essentially by the fine aggregate (that is, exclusive of the finishing operation); and for asphaltic pavements the coarse aggregate will be the primary control aggregate.

The binder percentage depends heavily on the design criteria to be satisfied for durability and stability in the case of asphaltic surface, and durability, strength, and workability in the case of concrete pavements. This is not to say, however, the binder percentage is of no importance

insofar as skid resistance is concerned. Too much asphalt binder will bleed and become slippery, whereas too little may lead to raveling and general deterioration of the pave-ment surface. For concrete pavements, if the mortar is improperly balanced in the total mix, rapid wear and early deterioration of the concrete surfaces will occur. There-fore, the designer must review carefully the binder require-ments for the pavement surface being designed.

Construction methods for high-quality asphaltic and concrete pavements are well established and there is little control that can be effectively used to improve the place-ment of the surface, assuming that the normal quality control methods are employed. There have been attempts (92) to alter the paving train for asphaltic pavements to insure that the coarse particles do not have their fiat sides oriented in the same plane as the roadway surface, but this has been largely unsuccessful. In the case of concrete pavements, however, several techniques can be employed during the finishing phase to improve texture; these are discussed in a later section dealing with concrete pavements.

The final mixes should take into account the service to which the pavement will be subjected. For example, the mix design for heavily traveled, high-speed roadways (such as freeways) should be different from the mix design on a secondary road accommodating low-speed, light traffic. On a freeway, it may be desirable to use a coarse-textured surface that provides for good drainage of surface water under wet-weather conditions (93). Suggested guidelines have been made relative to the aggregate type and traffic volumes by at least two states (84, 90). These guidelines relate to the polish-wear characteristics of aggregate type in one case and such variables as size and gradation in the other.

Before the mix design is made final, laboratory polishing tests of the proposed mixtures should be made whenever possible. Laboratory equipment and methods of test have been described in the literature (65). Although an aggre-gate or mix may exhibit desirable qualities in laboratory tests, its performance in the field must still be monitored before its acceptability for future projects is to be assumed.

BITUMINOUS CONCRETE PAVEMENTS

The criteria for mix design of high-quality plant-mix bi-tuminous surface courses include stability, durability, and skid resistance. Once the first two criteria are established, the skid resistance quality requirements will probably be met as long as the coarse aggregates have satisfactory skid resistance characteristics. After the acceptable aggregates and their combinations have been determined, either in the laboratory or on the basis of field performance, certain mix design and construction techniques may be used to ensure additional improvement in surface traction properties. The techniques selected depend on the desired characteristics of the surface to be constructed (i.e., whether it is to be open-or fine-textured). In this regard, the pavement designer has control of the following:

Percentage and type of asphalt. Voids in the mix. Surface texture. Paving techniques.

45

The percentage and type of asphalt affect the over-all fric-tional performance of the surface in a limited way. If the percentage is too high, the asphalt will "flush" to the sur-face and, as a consequence, cover the aggregates and ob-scure the effectiveness of their skid resistance qualities. To be sure, this may be a temporary condition depending on the amount of wear that occurs after the roadway is opened for traffic. If the percentage is too low, or the asphalt too hard, raveling will occur and the surface aggregate will be lost. If internal voids in the mixture can be maintained at an acceptable level to meet stability and durability require-ments, they will also contribute to the removal of surface water due to dynamic forces imposed by the vehicle tire (94). Surface texture is a primary factor in providing the desired skid resistance qualities. The texture for high-speed, high-volume roads should be coarse and gritty as produced by a well-graded open mix with maximum size aggregate between /8 in. and ½ in. On high-volume roads, fine-textured surfaces are less acceptable to some agencies because of poor drainage of surface water during heavy rainfall.

Bituminous Mix Design

The design of a high-quality plant-mix bituminous pave-ment surface is variable rather than fixed, because there is a choice of appropriate aggregates and of their use. The variability most frequently is introduced by economic con-ditions that favor use of local aggregates. Standard bitumi-nous plant mixes for surface courses reported by the State of New York (84) are given in Table 2. It is of interest to note that the Type 1A design has a nominal maximum size aggregate of ½ in. and has a defined surface texture that is granular. This mix is designated for interurban and urban roadways. Types 1AC and 2A produce a smooth-gritty surface texture and are intended for resurfacing as well as urban use.

The aggregates to be used in the three New York designs are controlled on the basis of mineralogical content of coarse aggregates from approved sources. For example, limestone from the Onondaga formation must contain at least 20 percent chert particles; or noncarbonate crushed stone may be used as coarse aggregates. Further, the pro-ducer is given the option to upgrade a mix with non-carbonate stone from ¼ to ½ in. in the Type 1A mix, but in the final mix the total aggregate over ½ in. must have at least 20 percent noncarbonate particles. Typical skid resistance data accumulated by New York that lead to these specifications are shown in Figure 51. This figure shows a comparison of skid numbers for different traffic exposure of pavements containing low-impurity limestone and non-carbonate aggregates. Similar data were collected on pave-ment surfaces constructed with dolomite as the coarse aggregate. These data serve to illustrate how a state may inventory its pavements for the purpose of improving specifications and ultimately the level of skid resistance of its highway system.

PORTLAND CEMENT CONCRETE PAVEMENTS

The first decision in the process of providing a skid-resistant satisfactory nonskid concrete pavement is the Se-

lection of appropriate aggregates and the grading for their combination. Usually this means combining a graded coarse aggregate that has a maximum size of about 11/2 in. graded to the No. 4 sieve and a fine aggregate graded from about the 3/8j sieve through the 100-mesh screen. The combined grading will depend on type of aggregate, work-ability, etc., but usually the course aggregate is about 60 percent of the total mix. In the normal construction of concrete pavements, the fine aggregate (sand)-mortar components of the mix "floats" to the surface and, as a consequence, is the principal constituent exposed to traffic action. For this reason, the fine aggregate in the mortar has the greatest influence on pavement friction and should, therefore, be of the highest antiskid quality. ASTM C-33 contains a warning on the use of manufactured sands in pavement surfaces. Because of the dominating role of the sand-mortar on skid resistance, it is not uncommon for the coarse aggregate to be a less desirable aggregate; however, if a pavement surface is of poor quality, the surface will wear and the coarse aggregate will soon be exposed to traffic in sufficiently large amounts to overshadow the dominating role of the sand-mortar. If a quality-controlled mix is placed, the rate of wear of the sand-mortar com-ponent will be such that the coarse aggregate will not be exposed in sufficient quantity to unduly affect the resulting skid resistance. Assuming that desirable aggregates and their combination have been identified, skid resistance can be further improved during construction by controlling:

Water-cement ratio. Air content. Finishing method to form texture. Curing method.

Quality concrete is a prime consideration in the develop-ment of skid-resistant concrete pavements. Factors that lead to poor-quality concrete frequently are equally detri-

TABLE 2

BITUMINOUS PLANT MIXTURES FOR STANDARD TOP COURSES"

Type 1A

General Job

Type 1AC Type 2A

General Job General Job Limits, Mix, Limits, Mix, Limits, Mix,

S S S S S S Passing Tol. Passing Tol. Passing Tol.

Screen Size 1 in. 100 ±0 - - - - 1/2 in. 95-100 ± 5 100 0 100 0 1/4 in. 65-85 ± 5 95-100 ± 5 90-100 ± 5 1/8 in. 32-65 ± 6 45-70 ± 6 65-80 ± 4 No. 20 15-39 ± 7 8-40 ± 7 - No. 40 7-25 ± 6 - 35-70 ± 4 No. 80 30-12 ± 3 3-15 ± 3 17-40 ± 3 No. 200 2-6 ± 6 2-8 ± 2 5-12 ± 2

Asphalt Cement Percent 5.8-7.0 ± 0.4 6.0-8.0 ± 0.4 7.5-8.5 ± 0.4 Penetration 85-100 85-100 60-70

Placing Temperature 225-300 F 250-325 F 275-350 F

Texture Granular Smooth-Gritty Smooth-Gritty

Typical Uses Interurban, Urban Resurface, Urban Urban

From Ref. (84).

46

mental to achieving a skid-resistant surface. A high water- surface mortar that is deficient in fine aggregate. It is re-

cement ratio (WIC), for example, not only causes early ported from laboratory studies in which a standard-size

concrete deterioration, but also brings about excessive wear automobile tire was rotated against a simulated pavement

when subjected to traffic; also, it encourages a buildup of surface that, as the W/C ratio increases, the resistance to

A

eo A

S LIMESTONE £ NON-CARBONATE

A . a 55

A.

50

£

45

N.

N a a

CID

N

_.• 5"-- N 5••• N. N 5. S..'

S. S S •S 0 N 35 S •• • N, A

• • N .N

N N • £ N

N *S 0 •

30 ----•--------•-- N

N • • N

1•' .

N N N • N

S 2.5

N N S

- . N

20- - S N

N N

N N N

N 15 1 I I. I I 11111

2 3 4 5678910 20 30 CUMULAT!VE TRAFFIC (MILLIONS OF VEHICLES)

Figure 51. Skid resistance change with traffic for pavements containing low-impurity limestones and noncarbonates (84).

A

LA

47

turning of the wheel decreases, thereby indicating that more polishing takes place (95). Quality concrete can be designed according to readily implementable procedures that have been published by the Portland Cement Associa-tion (96).

Historically, concrete pavement finishes have exhibited a fine-to-medium texture rather than a coarse texture. This is because the sand-mortar is brought to the surface by the finishing operations while the concrete is still in a semi-fluid state. This characteristic of concrete has led to the conclusion that, in a concrete mix, the fine aggregate is more important insofar as skid resistance properties are concerned than is the coarse aggregate. The finishing method, however, governs the final texture of a concrete pavement surface and, therefore, provides an opportunity for control of texture. A number of states (Indiana and California, for example) are directing greater attention to final texturing in order to achieve improved skid resistance. Some of the more common finishing methods are:

Burlap drag. Paving broom. Belt. Wire broom (brush).

Pictorial examples of the type of texture that can be formed by these methods are shown in Figure 52. Texture depths as formed by the methods shown are tabulated in Table 3. They ranged from a low of 0.014 in. to a high of 0.075 in. as measured by the sand-patch method (95). The characteristics of a particular texture configuration are related to the type of concrete mix, the water-cement ratio, and the time elapsed before finishing begins. Texturing is also influenced by environmental conditions such as wind, humidity, and temperature (97). The variation in skid number among finishes can be quite large. For example, a transverse-broomed, medium-texture surface hada wet skid, number (SN40) of 68, whereas a burlap drag of medium texture exhibited an SN40 from a low of 38 to a high of 72. An experimental field study of concrete pave-ment texturing methods recently reported on included the following methods (52):

Heavy burlap drag. Natural-bristle paving broom. Wire broom. Fluted magnesium float.

The textures were placed perpendicular to the roadway centerline using hand tools. Texture depths were measured by the sand-patch method after the residual curing com-pound had been removed.

Skid tests were performed with a locked-wheel trailer (ASTM E 274-65T) at speeds of 40 and 55 mph. A sum-mary of the skid tests is given in Table 4. The mean texture depths based on three measurements at two test sections at each of five sites are given in Table 5. Several significant conclusions were drawn from this study, including the following:

Textures produced by the different methods varied both in texture depth and initial skid number.

TABLE 3

TEXTURE DEPTHS ON LABORATORY SPECIMENS

METHOD OF FINISH TEXTURE DEPTH (IN.)

Light belt 0.015 Light burlap 0.017 Heavy belt 0.020 Heavy burlap 0.025 Medium paving broom 0.029 Wire drag 0.036 Heavy paving broom 0.037 Stiff wire brush 0.075

From Ref. (95)

TABLE 4

SUMMARY OF INITIAL SKID NUMBERS

INITIAL SKID NO. (SN) METHOD OF FINISH MIN. HIGH MEAN

Burlap drag 38 64 52 Paving broom 46 72 58 Wire broom 51 72 61 Fluted float 40 72 61

From Ref. (52).

TABLE 5

SUMMARY OF TEXTURE DEPTHS

METHOD TEXTURE OF FINISH DEPTH (IN.)

Burlap drag 0.019 Paving broom 0.031 Wire broom 0.042 Fluted float 0.045

Textures produced by the same methods varied sig-nificantly from site to site.

The permanence of texturing varied, depending on cumulative traffic volume, initial texture depth, and wear-ing characteristics of the particular. texture.

For equivalent texture depths, the texture produced by the fluted float generated no more sound, and possibly less than those produced by conventional methods.

The study is of great significance and its findings will no doubt find wide use in the design of textures for concrete pavements.

The literature does not contain a significant amount of information about the effect of texturing in the transverse direction. If it should be found to decrease stopping dis-stance, it may be desirable to restrict this technique to areas of rapid speed changes (such as intersections, toll gates). Longitudinal texturing may be most effective on

48

-5--- ------- -5---. __-_--_s_ -

'- --..-

MEDIUM BROOM

.t j•' ••.t:y' '

"' I '''' I'' •'" ''•'' ''vV. ' Y. t 41'

4.... r

. ,ç" '. t:1j'' •. '' 4

ti

LIGHT BURLAP DRAG

-4. - ________ - .- :.

— --- -

1 — ____5_•____•

_• - 55 - èS

---_----

••— -;--- -'- __5., ._-.___.--•-_-*... ..- -.5-

--. --.___.'.'_-___•-.. •_-•:_ 5- '_

.../,.

'.-.-.. .-, h5/'-v41

WIRE DRAG

'•-- .-.._:, . -•-'-

..:'••. .-\• .;'4;4_4•

HEAVY BROOM

4,; '/_• •

;

) !

f• ,. ••?1• ',P .;1 ':'i ,1 •

,. 4. . 4 ,c_ • ., .. . I

' '. 'k'I. 4, r -'' - 1 ( ..1,

iIIi ''fl ;i (

: ',r 74

!. •'? r ,•

..'.b

:•. .-'. .,

' 1 , HEAVY BURLAP DRAG

FLEXIBLE WIRE BRUSH

Figure 52. Examples of portland cement concrete surface textures (95).

tangents and curves as a technique of offering greater di-rectional stability. Also, the level of noise as generated by texturing is not well documented. This whole area is one

of needed research. In summary, the design of concrete pavements with

adequate skid resistance can he accomplished by selecting a high-quality fine aggregate, using quality control methods

in construction, and taking great care in the finishing of the surface. To the extent desirable, texture can be "worked" into the pavement surface to obtain greater skid resistance. The long-range performance of the texture will depend on the quality of the sand-mortar, the traffic vol-umes carried by the pavement, and the type of maneuvers,

among other things.

CHAPTER SEVEN

MAINTENANCE OF PAVEMENT SKID RESISTANCE

49

In maintaining skid-resistant pavements, a continuing road inventory is necessary to identify those areas that need corrective action. The inventory should involve all ele-ments of the system, including highway geometrics, road-side hazards, pavement surfaces. The need for inventory-ing the pavement surfaces had been recognized by many states (98), but the real impetus came as the result of an Instructional Memorandum, issued by the Federal High-way Administration in 1968, that provided for federal participation in the cost of resurfacing when necessary to maintain skid resistance at a minimum level of 35 SN (99). Also, the federal safety "spot" improvement program now requires the identification and removal of factors con-tributing to a high accident rate at particular locations on federal-aid highways. Obviously, this has added to the need for inventorying pavement surfaces.

PAVEMENT SURFACE INVENTORY

The pavement surface inventory may be accomplished by one of several methods, including a towed trailer, a stop-ping distance automobile, an automobile decelerometer, or portable testing devices. A recent report indicates that 25 states conduct an annual inventory of the skid resistance of pavement surfaces located throughout the primary and Interstate systems. Twenty-three states use a skid trailer, whereas the remaining two use another method. Excerpts from this report are given in Appendix A.

The pavement surface inventories using the towed trailer most generally conform to the standard method described in ASTM E 274. The results of tests are reported as skid number (SN). These and other testers are described in Chapter Three. If a pavement has a low skid number, it should be considered suspect and the accident records for it should be consulted.

In conducting the inventory, physical features of the roadway should be appraised, as certain ones may be more responsible for the roadway being unsafe than the level of skid resistance of the pavement surface. These features include rutting in the wheel tracks, excessive wear caused by studded tires, inadequate runoff of surface water, and highway geometrics preceding and leading into the site. Aside from these features, the pavement surface charac-teristics are of prime importance because they are related to the frictional needs of the motorist in performing the necessary driving maneuvers. As an example, the pave-ment friction needed by the driver when 'he negotiates a curve will be different from that needed on a tangent.

Figure 53 shows the results from an inventory of a section of roadway tested in connection with a statewide inventory being carried out in Tennessee (100). The data were collected with a two-wheel 'trailer operating at 40 mph using a self-watering system. The measurements are taken

at approximately ½-mile or shorter intervals in the inner wheel path in both directions of the outside lane. The graphical form of presenting the data provides a continu-ous log of the changes in skid numbers as related to the log mile along the highway surface. The variation in SN's can be seen to occur between pavement lanes as well as construction sections. Along with the measurement of the level of pavement skid resistance, other information is collected related to the physical features of the roadway.

After the inventory has been completed, measures to correct deficient pavement surfaces may take one of two forms:

Modification of the existing surface. Application of a new surface.

MODIFICATION OF EXISTING SURFACE

In many instances, the solution to correcting a pavement surface of low skid resistance may be modification of the existing surface rather than the application of a new sur-face. The nature of the modification depends on the type of surface and the reason for the low skid number. For example, if a concrete surface is to be corrected, grooving may be the answer; if it is a bituminous pavement, bonding another aggregate to the surface to form the desired texture may be the best and/or most economical solution.

Concrete Pavements

For concrete pavements, surface modification may involve one of the following methods:

Surface roughening. Acid etching. Grooving. Bonding thin layer of aggregate.

Surface Roughening

Attempts have been made to roughen concrete surfaces to improve their skid resistance with varying degrees of suc-cess. One surface roughening technique reported in early 1950 involved applying steel shot to the pavement and then dragging the surface with a heavy steel plate to cause the shot to cut into the pavement surface. Use of this method is obviously limited in its application and its success de-pends on a number of variables such as the hardness of the sand-mortar and the type of concrete aggregate (101). Another method uses a hand-operated machine that cuts the surface in a manner similar to grooving. The machine has a rotating drum on which cutting teeth are mounted. As the drum rotates, it cuts the surface in an 18-in, swath and to a depth that is adjustable but usually less than ½ in.

The flailing of the concrete surface as reported in the

50

70

60

50

40

30

20

10

I I I I I II

1 1 1954 I I I I II I

1969 I 1970 I 1969 1968 I 1963 I 1966 I I 1950 I

I I I I I II I

0 Eastbound I I I I • Westbound I I I I

I I I I I II I

I I I I I II I

- I I I I 1•1

I I I I I II I

I I I II I

I I I I I II I

II I I I I ii III I

0 4 8 12 16 20 24 28 32 36

LOG MILE

Figure 53. Results of skid resistance inventory on pavement surfaces of different ages (100).

Tennessee study (101) has limited effectiveness. After the surface was altered, the skid number at 30 mph was 43, but approximately six months later it had dropped to 18. The reduction was attributed mostly to the nature of the concrete aggregates and the large volume of traffic acceler-ating at the treated intersection.

Acid Etching

The concrete surface may be roughened by acid etching. This involves the use of chemicals that etch the surface, usually concentrated hydrochloric or hydrofluoric acids that are diluted with water at the time of application. Most frequently the acids are applied with hand spray equip-ment. After the acid has reacted with the minerals of the aggregates and with the cement, the residue is flushed, thereby exposing an improved surface texture. The acid-etching procedure is most effective on concrete pavements where the surface finish is coated with laitance (cement mortar), as the acid reacts strongly with constituents in the cement, erodes away the surface, and exposes the aggregates. Not only does the chemical react with the minerals in the aggregates and cement, it will loosen the oil and dirt film and, in general, clean the surface. This technique is best suited for use at intersections or similar locations where vehicles accelerate or decelerate drastically.

The benefits from acid etching are rather short lived; usually, such treatment will be effective for less than a 6-month period. Protection is required for the workers against the acid and its fumes.

Grooving

Grooving is a technique of altering an existing pavement surface to greatly increase its texture, thereby facilitating the displacement of water by the tires. It is used most often at locations where hydroplaning or wet-skidding ac-cidents at high speeds are a problem, and on horizontal curves. Probably the most concerted effort of altering pavement surfaces by grooving was made on several air-fields in England in 1956 (102). Grooves are cut with a saw that uses a blade impregnated with diamond grit. The most frequently used types of grooves are shown in Fig-ures 54 and 55. Grooves are cut either transversely or parallel to the direction of travel. Usually, highway sur-faces are grooved longitudinally and airfield pavements in the transverse direction. The geometric features that can be varied are shown in Figure 55.

There are a number of factors that control the success of grooving. The nature of the aggregates plays an im-portant role; particularly their susceptibility to wear. Spall-ing of the grooves may occur, depending on several fac-

51

tors. For example, spalling caused by freezing of water in the grooves is little affected by groove geometry, whereas groove spacing is important: the closer the spacing the more spalling is apt to occur.

Several types of groove patterns are shown in Figure 55. A minimum depth of ½ in. is preferred; a spacing of 3/4 in. to 1 in. is satisfactory; groove width tends to be critical. It is reported (102) that grooves with 1-in, spacing and ¼-in, width reduced the directional control of light cars and motorcycles, as these vehicles tended to "track" the grooves. Rectangular grooves with a width of ½ in. have proven acceptable to traffic, as well as being less costly than bevelled or wider ones.

An example of a longitudinally grooved pavement is shown in Figure 56. Accident experience attests to the effectiveness of grooves in pavements on which water depth tends to be excessive during heavy rains. The skid num-bers, as measured by locked-wheel testers with water ap-plied at the rate prescribed by ASTM E 274, do not show a significant increase, indicating that grooving is not a remedy for inadequate surface characteristics per se. How-ever, lateral skid resistance is improved (102).

Bonding a Thin Layer of Aggregate

Another method for modifying an existing surface is the bonding of a thin layer of aggregate to the surface. The bonding medium may be an epoxy resin that is applied to the surface after it has been cleaned, usually with muriatic acid. The aggregate to be bonded should be extremely hard, highly angular, and free of fine material (minus 100 mesh). The aggregate may be crushed quartz sand or a synthetic aggregate such as calcined bauxite, car-borundum, or slag. A successful treatment applied to a concrete surface, using two different aggregates, was ap-plied in Tennessee at a high-volume intersection (98). (Similar applications have been made elsewhere.) The improvement in skid resistance after one year of service is shown in Figure 57. This type of application may not be economical as a cOrrective measure for long stretches of roadway.

Bituminous Pavements

The modification of bituminous pavement surfaces to cor-rect low skid resistance can be accomplished in a number of ways. The most common are:

Grooving. Thin overlay. Heater planer.

Grooving

The grooving techniques described for concrete pavements are also applicable to bituminous pavements and give simi-lar improvements. Under certain conditions, however, grooves in an asphaltic pavement may flow together and lose their effectiveness. This usually occurs on new pave-ments if the asphalt content is too high. On the other hand, older asphaltic pavements, with high aggregate content, will maintain the groove configuration.

PAVEMENT GROOVING

TYPES

TRANSVERSE LONGITUDINAL

II !

k DIRECTION i Ji if ii OF MOTION

if liii Ii

GEOMETRY

_ -•4.-'II•-"- wI LI

PITCH

o DEPTH

METHODS

DIAMOND SAW

FlAIL

PRECAST Figure 54. Pavement grooves, types and geometry (102).

GROOVE PATTERNS

i r• r-'-i _f1/S

70 STYLE "4" '15" STYLE ç 3,4 _ 1/8

STYLE 524 I

ir"8 i— liLly .%o./4t.1 b.O9Ljl4l STYLE • .I25O62i=TST I

1 3/8 fl/s .0",/t1 h125

f.062

STYLE '0"

Ml.jo 1STYLE "E"

Figure 55. Experimental groove patterns (102).

Thin Overlay

Unless complete resurfacing is needed to correct deficien-cies in addition to. low skid resistance, the skid resistance level of bituminous pavement surfaces may be best ac-complished by the application of thin overlays using a bituminous binder. Results reported in a Michigan study (105) and elsewhere (77) indicate that thin overlays can be successfully used on, at least, non-Interstate roadways. In these applications, an angular, hard (5.5 plus, Mohs scale), well-graded, polish-resistant aggregate is needed. The thickness of application for the aggregate can be from 25 to 80 lb/sq yd. If the treatment is too thin, it may prematurely wear away. If it is too thick, stability may be impaired and the surface will "shove."

Heater Planer

A method for correcting a surface with low resistance due to bleeding of the bituminous binder employs a heater planer. The equipment consists essentially of a unit to

1. 1::T.I iii;' •;.J:. - . 11;i 1 L:1.1 .Uil 1

71/iijiIi 11 Jjt'11j 2 ' ; rs II &V1\ \T

lip

• 1 - -

.1—S r 1jWJJ'1- 1.11 I .1 1I 1 1 1 't. .1 ' • L 1!I J't 4 I 1

El I3''J I'fJ'If I IJ I' ' I 1I i A

IJJ1fIPI tJ1 i ' 4 ' V' - 7I S -I II II '• •'& I J14 I'iI . 4* I 1 I I

4 - •_ - - -

9Y)t/Ii/Pit -:

1f,A///1JI/JI46.

i f 74;ja • ,4 1: . : 1% I

-- - t.. # ,. • •t - . - - ' ve dqk

41 -- & - .t - ., ..' ' c . .' - ' • • . ' ,_.

'!J £4 ' I"< 3 't '

/içllre 56 L',zini!i,zc,I! •L('V( (1 ( &! ZCF &l, • ( a1TflIc IU.

1.0

Lu 0

0.6 RIVER SAND

rk

LM

0.2 r TRAILER METHOD SMOOTH TREAD

WET TEST

10

53

- NATURAL SAND

9-19-61

OLD CONCRETE

913-60

I I

20 30

MILES PER HOUR Figure 57. Improve,nent of a high-accident concrete pavement by application of epoxy-bonded sands (104).

heat the pavement surface and then remove the excess material by cutting or "planing" it away. After planing, aggregate chips or sand are spread and rolled into the surface while it is still hot. This method is most effective and economical in correcting isolated spots of bleeding, such as those occurring at high-volume intersections, rather than long mileage of roadway. It is also used to correct pavement roughness due to shoving. Overheating of the surface should be avoided, as deterioration of the binder is accelerated and early loss of surface aggregate will occur.

APPLICATION OF A NEW SURFACE

The most widespread corrective measure for improving the level of skid resistance of both concrete and bituminous pavements is the application of a surface that requires the normal controls of a high-quality mix produced in a hot-mix plant. In addition to improving the level of skid resistance, the new surface can correct such undesirable features as rutting, raveling, and excessive wear. Surfacing mixes that are open-graded have proven highly satisfactory. Such mixes may have a maximum size aggregate of /8 in. or ½ in. and contain no more than 3 percent passing the No. 100 screen. The average skid number (SN40) for one

such surface on a reasonably heavily traveled Interstate highway after three months of service was 49.

Surfacing studies conducted in Louisiana (106) and involving several different types of applications show that an adequate level of skid resistance can be built into a pavement surface using the plant-mix seal method. After eleven months of service of eight different types of plant-mix seal designs, the minimum skid number at 40 mph was 41.

Resurfacing can also be accomplished by the application of a seal coat. It is a treatment that is similar to the plant-mix seal, but is applied directly in the field. First an application of the bituminous material to the existing sur-face is made, then the aggregate is spread, followed by rolling. In the case of a double seal, a second application of the bituminous material is made, followed by another layer of surface aggregate. This type of seal coat is less desirable than the plant-mix seal, as field control of the construction is difficult. If too much bituminous material is applied, bleeding will occur. If the surface is not allowed to cure properly, the surface aggregate will be easily "whipped" from the surface by the action of traffic. A properly designed and applied seal coat will increase the skid resistance level of an existing surface by providing a rough texture.

54

The plant-mix seal is preferred to the "field" seal be-cause of the closer control on its application, the uni-formity of the finished surface, the smoothness, and the low noise level (107). In order for the plant-mix seal to be effective, it must contain a skid-resistant aggregate of the types described in Chapter Six. The plant-mix seal is an open-graded mix that readily accommodates rapid removal of water at the tire-pavement interface.

A plant-mix seal coat using a slag aggregate that has been successful in Tennessee has the following master gradation:

SCREEN SIZE % PASSING

½ 100 3/8 90-100 No. 4 30-50 No. 8 5-15 No.16 0-8 No. 100 0-3

The mix uses about 7.0 percent AC-8 (Pen 85-100) asphalt and is applied to an existing surface with a track spreader. Compaction is accomplished with a steel-wheel roller.

OTHER CORRECTIVE MEASURES.

Other techniques of increasing skid resistance include a "sprinkle" method and slurry. seals. The sprinkle method being evaluated by the Virginia Department of Highways consists of precoating a polish-resistant aggregate with a medium-curing asphalt (MC-70) and applying the cold mix to the surface of a hot plant-mix immediately behind the paver. A "whirly" salt spreader is used to apply the

cold mix. It spreads the mix uniformly at rates of 2 lb/ sq yd for a lightweight expanded shale to 5 lb/sq yd for normal aggregates. Rubber-tired rollers are used following the spreading operation and a steel-wheel tandem roller is used for final rolling. This technique has good potential of providing control over surface texture (108).

The application of a slurry seal is another method for changing surface characteristics for improvement of skid resistance. The effectiveness of this method is as dependent on the quality of the aggregate and the control of applica-tion as that of methods described earlier. It is reported that the slurry seal technique has been used successfully on an Interstate road in Oklahoma (109).

SUMMARY

An acceptable level of skid resistance can be maintained with a reasonable degree of success provided polish-resistant aggregates are used and great care is exercised in applying the treatment. The treatment selected depends on the nature of the surface to be corrected. Treatments used on concrete surfaces may not be suitable for use on bituminous pavements. Short experimental sections may be an effective way of optimizing aggregate gradation and asphalt percentage in the design of mixes for resurfacing or thin overlays. Plant-mix seal seems to be the most promising resurfacing method, as it provides adequate control and the desired uniformity. Pavement grooving is receiving wide acceptance and certainly should be con-sidered among the options for modifying existing concrete pavements. The methods and techniques are available—it remains the responsibility of the pavement designer to utilize this knowledge to maintain the desirable level of skid resistance for the motoring public.

CHAPTER EIGHT

SKID RESISTANCE CONTROL PRINCIPLES

Skid resistance management is not basically different from that of any other maintenance task (see Fig. 58). Quality requirements are set or mandated and present or antici-pated conditions are compared periodically with the re-quirements; if the requirements are not met, remedial steps must be determined and executed.

MEASUREMENT

In the past, the only signal for indicating inadequate skid resistance of a roadway had been an excessive number of skidding or wet-weather accidents. This is being supple-mented or supplanted more and more by measurement of

the skid resistance. Numerous methods for this purpose are in use all over the world. They have been discussed in Chapter Three; Appendix A gives an overview of the methods currently used by various states.

A point to be reiterated here is that although some of these methods correlate with each other (the automobile stopping distance method, whether two or four wheels are locked, correlates with locked-wheel skid testers when the same tires are used in both cases), other methods do not. A British Portable Tester measures something different from a skid tester operated at 40 mph. If a correlation is found, it is because a limited variety of surfaces, having certain characteristics in common, were used in establishing

55

the correlation. Using such a correlation for other types of surfaces may lead to unpleasant surprises.

This is not to say that only certain methods of friction measurement can be used to advantage. For instance, virtually"any method may be used to determine if low skid resistance is a likely cause for the difference in the acci-dent experience of two similar highway sections. Similarly, monitoring experimental surfaces for loss of skid re-sistance by traffic exposure or collecting data for setting priorities in resurfacing programs requires no specific friction measuring technique.

/

If one wishes to compare one's own data with those from other sources, or if mandated minimum skid resistance levels must be met, the method of measurement ceases to

) be at the option of the user. In the United States, ASTM (. Method E 274 has gained virtually the status of official

standard

COMPARISON WITH REQUIREMENTS AND STANDARDS

Minimum requirements for skid resistance were discussed in detail in Chapter Two. There are still none that have attained the status of standards. What goals the individual states have adopted may be seen from Appendix A. NCHRP Report 37 (5) made certain recommendations. These have been adopted, still in the form of recommen-dations, into the Highway Safety Program Manual (110). They are reproduced as Table 1 (in Chapter Two).

One of the reasons why standards have not been either agreed upon or prescribed by the Federal Highway Ad-ministration is the absence of a clear-cut answer to the question of what values are needed in "difficult" locations. There is little disagreement that on certain curves, on exit ramps, at intersections, etc., higher values than those recommended in NCHRP Report 37 are necessary (the latter being general minimum values that should apply anywhere). Consequently, individual jurisdictions must, for the time being, be guided by experience and judgment in setting minimum values for certain locations.

ESTIMATES OF DETERIORATION

Because skid resistance is not necessarily constant (even if allowance is made for the cyclic effects discussed in Chap-ter Five), it is necessary to estimate the probable future loss in skid resistance though at the time of measurement a location or section still meets the established require-ments. If the pavement has been in service several years, chances are that the skid resistance is near the lowest value it is likely to attain. But if the surface has not yet been polished to equilibrium, one must estimate what the lowest skid resistance will be and how long it will take to reach it. Unfortunately, experience is the only reliable guide. There are as yet no methods available for predicting with any degree of certainty how fast a pavement will polish.

The only cure for this situation is to keep accurate records on all pavements that are included in periodic skid resistance surveys. The importance of such records, which must include traffic counts and other details, cannot be overstressed; they are the only reliable source from which a life prediction methodology can be developed. Test strips

Figure 58. Pavenent slipperiness as a control problem.

of conventional or novel surface courses are helpful, par-ticularly for eventually validating laboratory prediction methods; but they cannot fully reflect all the possible variations and their effects on polishing. Increased use of studded tires has introduced a new variable that, in the affected states, has degraded much of the accumulated experience.

The ability to predict the skid resistance life of pave-ments permits, or will permit, planning of maintenance operations and allocation of the 'attendant cost. For 'this reason, considerable research is being directed toward developing accelerated laboratory polishing methods, as explained in Chapters Five and Six. Some of the described methods are used with good success to establish relative ratings for aggregates and pavement designs, but so far it has not been possible to predict actual skid resistance or useful life unless there are records available on in-service performance.

A good maintenance program, therefore, includes pe-riodic surveys of the entire highway system and accelerated laboratory polishing tests on all aggregates used and on pavement cores of newly placed surface courses. Although a British polishing method has been standardized (111) and ASTM is in the process of doing the same for some American methods, it must be reiterated that no method has been related to field experience in' such a way that reliable predictions of skid resistance service are possible. Thus, each state is still forced to select the method that suits its purposes best and build up correlations with field experience.

The turmoil created by studded tires has been mentioned several times. It tends to shift the emphasis from polishing

56

to wear. In states in which studded tire use is high, surface courses tend to become unserviceable because of wear rather than because skid resistance has been lowered below a postulated minimum value. In some cases, surface course failure may have still other causes. Obviously, these pos-sibilities must be considered in planning maintenance programs.

CHOICES OF ACTION TO BE TAKEN

If skid resistance becomes inadequate, the only solution would seem to be to resurface or otherwise upgrade the surface. This may, however, not be the best solution. There is also the possibility of easing the skid resistance requirements. This possibility is indicated by the dashed lines in Figure 58.

How can the required skid resistance be lowered? It can be done by realignment (eliminating curves, etc.), improv-

ing sight distances (cutting away an embankment, etc.), bettering visibility (lighting, etc.), improving traffic control (lowering the speed limit, changing from a traffic light to STOP signs at a side road entrance, etc.), and so on. All these measures tend to lower the required skid resistance if they smooth traffic flow. Anything that will eliminate or reduce the need to accelerate or decelerate, or reduce the tendency to pass or to change lanes or the direction of travel, reduces the need for providing high skid resistance.

Obviously, the mentioned alternatives to resurfacing are in most cases not simple alternatives because they are higher in cost and, therefore, must justify themselves on grounds other than reduction of skid resistance require-ments alone. The decision in every instance requires full analysis of the hazards and of all possible remedies. If skidding or wet-weather accidents are high on a particular stretch of road or at a particular location, dry-weather

HIGHER I ACCELERATION OR LESS POLISHING

R RESISTANCE

ETARDING TORQUE

OF PAVEMENT

GREATER STEERING ANGLE HIGHER

FRICTION DEMAND

POORER VISIBILITY AND LESS SIGHT DISTANCE

MORE COMPLEX HIGHWAY

GEOMETRI CS

HIGHER VEHICLE

PAVEMENT POLISHING

LARGER TRAFFIC VOLUME

REDUCED CONTROL POTENTIAL

ANY OR MORE

LOWER WATER ON

FRICTION PAVEMENT

AVAILABILITY

LESS TIRE TREAD DEPTH

Figure 59. Factors reducing ability of driver to maintain safe level of control over his vehicle (modified from 112). .

57

accident records should also be studied to determine if some cause other than low skid resistance is the overriding one.

In making such an analysis, Figure 59 may be helpful.

It looks at the situation from the viewpoint of the driver. Because he must largely be considered as a given, it be-hooves the highway engineer to take his demands and handicaps into account.

CHAPTER NINE

SKID RESISTANCE SURVEYS

MANDATES

Federal law requires that the states maintain the highways that were constructed with federal-aid funds. The National Highway Safety Act of 1966 specifies that adequate skid resistance must be maintained, and the term "maintain" includes the correction of excessive slipperiness. Accord-ing to the provisions of the Safety Act, this applies, how-ever, not only to federal-aid highways, but to all highways and streets, including those under the jurisdiction of and constructed by local governments, even though no federal or state funds were used in their construction.

The Highway Safety Program Manual (110) requires periodic surveys for the purpose of determining the pre-vailing skid resistance. Present requirements are not spe-cific enough to serve as guides for what should be done, how, where, and how often. Therefore, it is still necessary for the individual jurisdiction to devise a survey program and select the means and methods for implementing it.

In many cases, skid resistance surveys can also serve functions other than those required by the maintenance mandate.

SURVEY OBJECTIVES AND SCOPE

Obviously, a basic objective of any skid resistance survey is to identify pavement sections that require immediate remedial action. Suspected sections can be identified from accident records: any location that shows an abnormal number of wet-weather accidents is a candidate for inclu-sion in a survey. So are sections and locations that are considered excessively slippery by the maintenance units, or about which complaints are received.

The next step is to include sections that are suspected of requiring corrective measures in the near future. Such sections are more difficult to identify, but maintenance superintendents frequently have a feel for the rate at which skid resistance deteriorates and, therefore, can identify such sections.

If the rate at which skid resistance deteriorates is known or can be estimated, the frequency with which skid mea-surements should be made can be determined. Generally, skid resistance drops rather rapidly with exposure to traffic immediately after construction and eventually levels out at

some near-equilibrium value (Fig. 60). Therefore, one might be inclined to measure each section fairly often in the beginning and less frequently thereafter. However, this is necessary only if there is reason to assume that the skid resistance may drop below the acceptable minimum. On some pavements this never happens, or does so very slowly; hence, too frequent measurement is not warranted at any time during their life.

There is still not sufficient information available to pre-dict with adequate confidence the skid resistance life of pavement types. There are too many variables: traffic volume and composition, maneuvers, climate and weather, winter driving aids and snow and ice removal, contamina-tion, and all the variables in design and composition of surface courses. Gradual accumulation of life histories of the important pavement types within a maintenance juris-diction through systematic, periodic surveys eventually permits skid resistance life predictions. Such data will also permit validation of accelerated laboratory tests on polish-ing characteristics. Systematic surveys will provide an inventory of data from which knowledge about long-range skid resistance performance of aggregates and pavements under a variety of conditions can be deduced.

Thus, each highway department must lay out a survey program that best serves its specific needs while furnishing the data required for scheduling resurfacing activities. Such programs are, of course, limited by the availability of measuring equipment and personnel. Because trailer-type testers can travel almost equally fast whether they are measuring skid resistance or not, proper routing becomes important and has a significant influence on the cost per measurement.

For estimating purposes, one may assume that, on the average, Interstate and primary highways should be tested every one or two years, whereas other roads may be tested less frequently. Critical sections or locations (critical in the sense that they may cause serious increases in wet-weather accident rates or because they are near or can be expected to approach mandated minimum, skid resistance levels rapidly if a change in traffic volume, etc., occurs) may require more frequent measurement. In such cases, skid resistance should be checked at intervals of usually not more than 500,000 vehicle passes per lane.

58

SURVEY METHODS AND PROCEDURES

The trailer method according to ASTM E 274 is by all odds the most convenient, and in the long run the most economical, method for survey. Relative costs of different methods were compared by Kummer and Meyer (24).

The cost per site tested was given by them as $3.35 for the British Portable Tester, $4.70 for the automobile stopping distance method, and $0.32 for trailer-type testers (1962 dollars).

The British Portable Tester, the Penn State Drag Tester, or other slow-speed devices are totally unsuited for sur-veying high-speed roads, but they can be useful on the approaches to a STOP sign or a traffic signal and in similar locations. The higher the traffic speed is, the more mis-leading the information obtained with such devices can be, because at high speed the available friction when wet is highly dependent on the drainage properties of the surface, which slow-speed testers cannot recognize.

Much cost can be avoided if each project is traversed only once in a survey. This is best done by operating the tester at the prevailing average travel speed. When travel speeds are not known, the legal maximum speed should be used. •NCHRP Report 37 (5) and the Highway Safety Program Manual (110) give the recommended minimum skid resistance values in terms of measurement at either 40 mph or the prevailing speed.

If, however, tests are scheduled not just for comparison with recommended or mandated values, but to obtain per-formance data and histories of specific sections or pave-ment types, measurements at 40 mph should also be made. Otherwise, comparisons between different sections or type of construction, etc., are not possible, except in those cases in which the test speeds happen to be identical. When for this or other reasons several tests at the same site are to

80

501 Ui I

z I o 4oJ

S - -

be made, it is essential that they be made one after the other. It is still not possible to correct skid resistance mea-surement reliably for temperature and for effects that may change the actual skid resistance over shorter or longer periods of time.

Any single measurement, of course, is subject to these "error" sources. This can become important when com-parisons with a mandated minimum skid resistance value are to be made and the difference between mandated and measured value is small. This also applies when the sur-vey data are being used to accumulate a history of specific sections or pavement types over several years. Work now being done under NCHRP Project 1-12 will provide in-formation on how to correct for certain variables, how to control error sources, and what the confidence limits for routine skid resistance measurements are.

Ideally, surveys are best planned as concentrated efforts during late summer and early fall. Every winter, depending on geographical location and other factors, some re-covery of the skid resistance of almost all pavements takes place. By late summer the winter effects have been ob-literated, and the skid resistance of a particular pavement is probably at its lowest for the year.

Furthermore, almost everywhere in the United States the period of late summer-early fall (mid-July to mid-October) is one of infrequent rainfall and relatively stable tempera-tures. By scheduling the major part of the skid resistance tests during this part of the year, temperature and rain effects on the collected data are minimized. As experience is accumulated individually and collectively it will be possible to correct the collection data for such effects.

Where the same pavement types or locations are to be compared from year to year, it will be advantageous to program the survey tests so that the same pavements or locations are tested not only during the same month, but also at the same time of the day or night in consecutive years. Tests are best scheduled for nighttime because traffic interference is at a minimum and temperatures are more stable than during the daylight hours. The presence or absence of solar radiation can have a significant effect on the effective test temperature and radiation effects are also much smaller during the night hours.

Pavements do not necessarily have uniform skid re-sistance, either transversely or horizontally. For compli-ance tests, suitable averaging techniques can reduce these effects to acceptable levels. When accumulating life his-tories of pavements or locations, the problem is not as

4 5. VEHICLE PASSES • millions

Figure 60. Loss of skid resistance of two pavements as a function of traffic exposure (113, 114).

Total as Passenger Car Equivalents £ Passenger Cars Only

Trucks Only

59

easily solved. Every effort should be made to ensure that the same pavement location is tested every year.

The location of the test site longitudinally along the road is not critical if the pavement project was uniform to begin with and the traffic characteristics are such that wear and polishing may be considered as uniform. When this is not the case, mileposts or landmarks must be used to ensure that successive measurements are comparable.

Lateral positioning of the tester during measurement is, however, extremely important, especially during middle age of the surface. Considerable transverse differences in skid resistance can exist, particularly on roads on which traffic is strongly channeled. When the surface is new this is no problem. When it is old the polished part of the wheel path has widened sufficiently to make proper tracking less difficult. Figure 61 explains why skid resistance life his-tories frequently show the largest scatter of data points just 'prior to reaching the equilibrium value of skid resistance (Fig. 60).

DATA ACQUISITION

During individual tests the minimal data to be recorded are those specified in ASTM E 274. To obtain maximum benefit from surveys it may be advantageous to have additional data and information.

For instance, there is some evidence that average daily traffic is not sufficient for defining pavement exposure. Figure 37 seems to indicate that the number of commercial vehicles per day correlates better with pavement polishing than does the total number of vehicles. The same data were used for plotting Figure 62 (67). Now, if total passenger car equivalents are plotted it appears as if a further increase in traffic density would not lower the skid resistance significantly. Whether Figure 62. can be gen-

l I c,

U. U. I 7 II

j U ZI I

TRANSVERSE LOCATION

Figure 61. Idealized change of transverse skid resistance pro-file of a wheel pat/i with time.

eralized or not is unknown at this time. For this reason it is desirable to obtain information not only about total vehicle passes, but also about traffic composition.

Information should be obtained about stud use, nearby construction, or other activities that may have taken place just prior to a survey to 'cause temporary contamination of the surface. Weather and winter-maintenance records should be available alongside the test logs. This informa-tion is necessary so that the causes of abnormal changes in skid resistance can be traced.

Obviously, all available information on the pavement,

O ID 20 30 40 50 60 70 80 AVERAGE DAILY TRAFFIC , thousands

Figure 62. Skid resistance of several roads with 3-year-old, identical surface courses as a function of traffic density (67).

60

its design and construction, the source and mineralogy of the aggregate, etc., must be available if the survey results are to yield general information about the relation of these variables to the equilibrium skid resistance under pre-vailing traffic and other conditions. Only when this in-formation is available can surveys produce the store of knowledge on which future improvements can be built.

Individual tests should be supported, to the extent prac-tical, by secondary data. Temperature, as an important effect on measured skid resistance, should be recorded. At a minimum, air temperature should be measured. It is only indirectly related to the temperatures at the tire-pavement interface, but is useful when examining large deviations from expected skid resistance values. Wet-pavement temperature is more representative of the several governing temperatures and can be used to correct skid nUmbers for deviations from a selected base-line tem-perature (usually 70°F). However, it should be noted that, even for a given test tire, somewhat different corrections apply to different pavement types.

To obtain wet-pavement temperature it has been neces-sary for the test crew to leave the vehicle and apply a suitable temperature sensor to the pavement. For this reason it was scarcely possible to make such measurements for every skid test, or even on every project. Cost and safety considerations, with due regard for the probable error, if only occasional pavement temperatures were taken, decided how often and where such measurements were to be made. New infrared thermometers now avail-able can measure pavement temperatures during testing.

Ideally, for every pavement, skid resistance should be known not at just one speed, but as a function of speed

over the range of speeds at which traffic moves. This ideal can be approached by measurement at two speeds, suitably spaced. This requires, in many cases, traversing the same highway section twice. If a project is long and uniform, high- and low-speed measurements can be alternated.

Instead of making all skid resistance measurements at two speeds, one may supplement single-speed measure-ments with texture measurements. Because texture is re-lated to the change of the skid resistance with speed, it is possible to infer the skid number at one speed from the measured SN at another speed (Fig. 63).

Photographs, preferably stereophotographs, both in and outside the wheel track, can be taken at the same time and will provide additional documentation that will be helpful when interpreting the results of several years. Such photo-graphs should be made with the camera mounted on a box with internal lighting to reduce apparent differences due to changes in lighting.

When pavements are grooved, groove depth should be measured, and if the pavement is permeable, permeability measurements should be made. In both cases, standard skid resistance measurements do not tell the full story because such pavements are less dangerous than their skid resistance would indicate. Measurements of hydroplaning potential would be necessary to properly rank them with respect to high-speed accidents, but as yet no practical method for making such measurements as a matter of routine has been developed.

Pavement roughness, if not so severe as to be objection-able as such, does not influence the trailer skid resistance measurements, but it can affect the behavior of vehicles during locked-wheel stops. Hence, it is desirable to make any roughness measurements that are scheduled at the locations at which skid resistance is being measured during the periodic surveys.

a E 11

Cl)

AVERAGE PEAK HEIGHT, in

Figure 63. Skid number-speed gradient vs average peak height (from profilograph) for surface treatment and seal coats (based on data from 88).

DATA EVALUATION

One problem with surveys is that they generate a large mass of data that must be analyzed and related to location, pave-ment design, weather, etc., and to the data from earlier and future surveys. Then sections needing remedial mea-sures must be identified. This work should be done promptly so that abnormal data can be checked and their causes traced before large changes have taken place or winter obliterates the evidence.

Histories should be compiled from consecutiye surveys for pavements embodying new design features. This is to provide base lines for future design changes or different volumes and mixes of traffic, and to develop a basis for long-range planning of maintenance activities.

The value of routine surveys can be greatly enhanced by proper evaluation and presentation of the results. Com-puterized data management is definitely indicated. The collated data should, of course, be made available or readily accessible to all offices and individuals concerned; that is, to maintenance, materials, pavement design, geology, traffic engineering, etc.

CHAPTER TEN

LOWERING SKID RESISTANCE REQUIREMENTS

61

Reduction of wet-weather skidding accidents can be at-tained either by increasing pavement skid resistance or by reducing the skid resistance requirements of traffic. Theo-retically it is possible, though not economically feasible, to raise wet-pavement skid resistance to that of the dry pave-ment. This would prevent all skidding accidents caused by pavement wetness. When skid resistance requirements are lowered, for instance by eliminating all intersections, acci-dents are reduced. But not all accidents attributable to pavement wetness are prevented. This is so because driv-ers still make errors and therefore severe maneuvers are still being made from time to time. When the pavement is wet, these maneuvers may cause skidding. Accidents might not have taken place if the pavement had been dry.

The fact remains that lowered skid resistance require-ments can reduce skidding accidents (provided the pave-ments meet reasonable skid resistance standards) as much as by raising pavement skid resistance. The effect of both methods is not directly comparable and must be reviewed on a total accident basis. With improved accident data, and their storage in computers, it should be possible to compare the benefits from deslicking operations with those from measures involving the lowering of skid resistance require-ments. It should thereby be possible to determine what skid resistance a highway section must have to obtain the same accident reduction by measures that reduce skid resistance requirements.

The measures to be discussed in this chapter have, to a varying degree, other beneficial effects. Implementation cost should therefore not be charged entirely to skid resistance control.

SPEED RESTRICTIONS AND WARNING SIGNS

The admonishment to "slow down when the road is slip-pery" is reasonable enough. Not only does reducing speed reduce the severity of any accident and give drivers more time for recovery from a potentially dangerous situation, but on almost all pavements wet skid resistance also in-creases with a reduction in speed. Consequently, all mea-sures that result in reduction of the average travel speed are beneficial.

The most primitive measure is to tell the driver that the pavement is slippery. However, fixed "SLIPPERY WHEN WET" signs have low effectiveness, just as do all fixed warn-ings for temporary hazards. Slight improvement may be obtained by posting an advisory speed. But unless the posted speed is mandatory, and reasonable, and strictly enforced, not much is gained.

Variable signs that display a warning and/or speed restriction only when the pavement is wet may be some-what better. In reality, drivers already have a warning: the pavement is wet. Whether a manmade supplementary warning is more effective remains to be proved. Variable

signs are probably most effective for temporary spot haz-ards. In either case, it is absolutely essential that such signs are fail-safe.

EDUCATION AND ENFORCEMENT

Like all accidents, wet skidding accidents are rare events for the individual driver. Therefore, education is prac-tically useless because in the absence of the threatened event drivers subconsciously discount the possibility of its ever occurring. The only really effective method is to sub-stitute a feared event that has a significantly greater prob-ability of occurring. This means enforcement of a man-dated reduced speed by frequent patrols and arrest of all violators. It is, however, next to impossible to enforce effectively spot speed restrictions, as on a curve. Instead, it is necessary to limit speed over a significant length of highway, or over the entire highway system, where normal speed limits are high.

Because skidding accidents are such rare events for the average driver (particularly if he lives in parts of the U.S. where snow and ice never or rarely occur), he will not be prepared to cope with the situation where his vehicle goes into the skid. Whether or not training in handling skids can be beneficial is not known. Intuitively this would seem to be the most promising step in an educational approach to skid prevention.

SMOOTHING TRAFFIC FLOW

It is an established fact that accident rates are not so much a function of speed as of speed differences in the traffic stream. Thus, anything that homogenizes traffic flow ,is beneficial and should contribute to a reduction in skidding accidents. Grade separations, multiple lanes, and elimina-tion of all features that permit differences in the hp/weight ratios of vehicles that cause speed differences to develop, help to smooth traffic flow and reduce the likelihood of skids. Elimination of grades and curves (or their reduc-tion), cross traffic, traffic signals, bottlenecks, etc., should almost always results in reduced skid resistance require-ments.

Obviously, such measures are usually justified on existing highways only when other reasons compel reconstruction or improvement. However, new construction plans should be reviewed carefully to ensure that every possible step is taken to minimize skid resistance requirements. Considera-tion should also be given to such facts as: a curve of given radius in a mountainous area is likely to be less hazardous than the same curve at the end of a long tangent in level country. Not only are the approach speeds likely to differ, but driver conditioning, or the lack of it, also will make the curve in level country more hazardous.

The last example describes a situation in which good

62

warning signs can be of benefit. Similarly, adequate super-elevation can reduce the skidding hazard for a large percentage of the traffic.

REMOVAL OF HAZARDS

Poor sight distance, lack of lighting at night, and innumer-able other factors can force drivers to execute maneuvers that can lead to skidding. Their elimination results in a reduction of skid resistance requirements and again is also beneficial to safety in dry weather.

One might include here the removal of roadside obsta-cles, the provision of adequate shoulders, etc. Because these measures reduce the consequences of skidding acci-dents, they can be counted among the measures reducing skid resistance requirements.

PREVENTING WATER ACCUMULATION

Complete loss of skid resistance by hydroplaning can be most effectively prevented by speed reduction. Speed con-trol requires early recognition of the hazard by the driver because conditions conducive to hydroplaning usually exist over only limited sections of a highway, as on a curve or an exit ramp, and then usually only for brief periods. This cannot be relied upon, however. Where tropical rainstorms deposit so much water on the pavement that hydroplaning would be possible almost anywhere, the hazard is usually minor because sight is so impaired that drivers automati-cally slow down.

Because a large, though unknown, percentage of hydro-planing incidents lead to accidents, spot accumulation of water should be prevented under all circumstances. This usually is not difficult when proven design rules are ap-

plied. Greater than normal cross slope is warranted when any conditions exist that tend to increase water depth abnormally. This can happen at junctions where more than the normal number of lanes produce unusual pave-ment width, or where a downgrade tends to cause the water to run in the direction parallel with the pavement.

Curves should have enough superelevation to prevent large water depth from developing. This is particularly important when the curve adjoins a downgrade. Exit ramps sometimes serve as paths for the runoff from the main pavement. Careful design can prevent this.

Temporary conditions can unduly increase the water depth on the pavement. Drains may plug and cause flood-ing. Melting snowbanks can cause meltwater to run across the pavement instead of down the shoulder. This happens frequently on bridges where the plows cannot push the snow far enough off the traveled portion of the roadway. Particularly on curved bridges, this can lead to dangerous conditions, even without freezing. Remedies for such situa-tions are fairly obvious, although not necessarily easily applied. These conditions are also more difficult to an-ticipate. Careful attention to newly constructed roadways is called for so that trouble spots can be avoided.

When none of these measures can be applied, grooving of the pavement surface must be considered. Traffic, par-ticularly where studded tire use is heavy, will wear away the grooves in time. Regrooving, if it can be done at all, cannot be done very often. Resurfacing with a very coarse overlay is then the only appropriate measure. Permeable surface courses might be used in the future if their development should prove successful. These pavements

'must ex-

hibit adequate permeability to pass water downward and laterally from the tire path.

REFERENCES

SABEY, B. E., "The Road Surface and Safety of Vehicles." Proc. Inst. Mech. Eng., Vol. 183, Pt. A (1968-9) pp. 6. SCHULZE, K. H., "Das Verhalten der Strassen bei Naesse" (The Characteristics of Roads when Wet). Tech. Univ. Berlin, VDI-Z, Vol. 106 (1964) pp. 1143-1 148. MCCULLOUGH, B. F., and HANKINS, K. D., "A Study of Factors Affecting the Operation of a Locked Wheel Skid Trailer." Texas Highway Dept., Res. Rep. No. 45-3 (Aug. 1966) 22 pp. Ticioi, S., "Car Deceleration Performance Changes Little from 1955 to 1968." SAE Jour., Vol. 77 (Apr. 1969) p. 65. KUMMER, H. W., and MEYER, W. E., "Tentative

Skid-Resistance Requirements for Main Rural High-ways." NCHRP Report 37 (1967) 87 pp. TARAGIN, A., "Driver Performance on Horizontal Curves." Proc. HRB, Vol. 33 (1954) pp. 446-466.

"The Determination of Pavement Friction Coeffi-cients Required for Driving Tasks." NCHRP Proj-ect 1-12 (Current Research). "Speed Trends on Main Rural Highways by Regions and Vehicle Type." Federal Highway Admin. (Apr. 16, 1969). "Annual Report of the National Swedish Road Re-search Institute for the Financial Year 1964-65." (In English) Rept. 46A, Statens Vaeginstitut, Stock-holm (1965). BRACH, R. M., "Friction and the Mechanics of Skid-ding Automobiles." Hwy. Res. Record No. 376 (1971) pp. 99-106. "A Policy on Geometric Design of Rural Highways." American Association of State Highway Officials (1965) 650 pp. "Lateral Accelerations and Lateral Tire-Pavement Forces in a Vehicle Traversing Curves Relative to

63

Available Pavement Skid-Resistance Measures." on Skidding Problems at the Road Research Labora-. NCHRP Project 20-7(4) (Current Research). tory." Hwy. Res. Record No. 46 (1964) pp. 43-59.

13. "Identification and Surveillance of Accident Loca- 31. KREMPEL, G., "Untersuchungen an Kraftfahrzeug- tions." Highway Safety Program Manual, Vol. 9, reifen" (Investigations on Automobile Tires). (In Federal Highway Admin. (1969). German) Automobiltechnische Zeitschrift (Ger-

14. STOHNER, W. R., "Speeds of Passenger Cars on Wet many), Vol. 69 (1967) pp. 1-8, 262-268. Transla- and Dry Pavements." HRB Bull. 139 (1956) pp. tion available from Nat. Transl. Center, John Crerar 49-84. Libr.

15. CAMPBELL, M. E., and TITus, R. E., "Spotting Skid- 32. KUMMER, H. W., and MEYER, W. E., "Skid or Slip Prone Sites on West Virginia Highways." Hwy. Res. Resistance?" ASTM Jour, of Materials, Vol. 1 Record No. 376 (1971) pp. 85-96. (1966) pp. 667-688.

16. BALDWIN, D. M., "Assembly and Use of Accident 33. "Annual Report of the National Swedish Road Re- Data." Hwy. Res. Record No. 376 (1971) pp. 14-17. search Institute for the Financial Year 1963-1964."

17. "Traffic Accident Facts-1972 Edition." Traffic (In English) Rept. 45A, Statens Vaeginstitut, Swe- Safety (1972) p. 40. den (1965) 44 pp.

18. BEATON, J. L., ET AL., "Reduction of Accidents by 34. MEYER, W. E., ET AL., "Design and Development of Pavement Grooving." HRB Spec. Rep. 101 (1969) an Airport Runway Surface Traction Measuring Dc- pp. 110-125. vice." Final Report, Contract FA-W-4236, ADS-55,

19. RICKER, E. R., "Use of Accident Data to Identify Federal Aviation Agency (1966) 43 pp. Wet-Pavement Locations in Pennsylvania." Hwy. 35. KOMMINOTH, V.,"Ein Geraet fuer die Bestimmung Res. Record No. 376 (1971) pp. 18-20. der Haftkraft zwischen Reifen und Strasse" (A Test

20. KUMMER, H. W., "Pavement Wetting and Skid Re- Apparatus for Investigating the Friction between sistance." Rep. J8, Joint Road Friction. Program, Pa. Tyre and Road). Proc. internat. Coll. on the inter- State Univ.-Pa. Dept. of Highways (1963) 31 pp. relation of Skidding Resistance on Wet Roads, Wil-

21. FIALA, E., "Kraftfahrzeug und Strassenoberflaeche" helm Ernst & Sohn, Berlin (1970) in German: (Motor Vehicle and Pavement Surface). Proc. In- pp. 77-80, in English: pp. 445-448. ternat. Colloquium on the Interrelation Between 36. DOMANDL, H., and MEYER, W. E., "Measuring Tire Skidding Resistance and Traffic Safety on Wet Roads, Friction Under Slip with the Penn State Road Fric- Wilhelm Ernst & Sohn, Berlin (Germany) (1970) tion Tester" Hwy. Res. Record 214 (1968) pp. pp. 215-223. Text in German and English. 34-41.

22. HUTCHINSON, J. W., ET AL., "An Evaluation of the GOODENOW, G. L., ET AL., "Tire-Road Friction Mea- Effectiveness of Televised, Locally Oriented Driver . suring System-A Second Generation." SAE Paper Reeducation." Hwy. Res. Record No. 292 (1969) . 680137 (1968) 18 pp. pp. 51-63. BAJER, J. J., "Proposal for a Procedure for Evaluat-

23. SMITH, L. L., and FULLER, S. L., "Florida Skid Cor- ing Wet Skid Resistance of a Road-Tire-Vehicle Sys- relation Study of 1967: Skid Testing with Trailers." tern." SAE Paper 690526 (1969) 16 pp. Highway Skid Resistance, ASTM Spec. Tech. Publ. 39. HOLMES, K. E., and STONE, R. D., "Tyre Forces as No. 456 (1969) pp. 4-101. Functions of Cornering and Braking Slip on Wet

24. KUMMER, H. W., and MEYER, W. E., "The Penn Road Surfaces." RRL Report LR 254, British Road State Road Friction Tester as Adapted to Routine Research Laboratory (1969) 26 pp. Measurement of Pavement Skid Resistance." Hwy. 40. VEITH, A. G., "Measurement of Wet Cornering Res. Record No. 28(1963) pp. 1-31. Traction." Rubber Chem. & Tech., Vol. 44 (1971)

25. HORNE, W. B., and SPARKS, H. C., "New Methods pp. 962-995.

for Rating, Predicting, and Alleviating the Slipperi- 41. DILLARD, J. H., "Supplemental Pavement Skidding

ness of Airport Runways." SAE Paper. 700265 . Resistance Tests in Virginia." HRB Bull. 186 (1958) (1970) 46 pp. . . pp. 46-47.

26. DILLARD, J. H., and MAHONE, D. C,, "Measuring 42. RIZENBERGS, R. L., "Florida Skid Correlation Study

Road Surface Slipperiness." ASTM Spec. Tech. of 1967-Skid Testing with Automobiles." Highway Publ. 366 (1963) pp. 1-82. Skid Resistance, ASTM Spec. Tech. Publ. 456 (1969)

27. RIZENBERGS, R. L., and WARD, H. A., "Skid Testing 43. pp. MAHONE, D. C., and RUNKLE, S. N., "Pavement with an Automobile." Hwy. Res. Record No. 189 Friction Needs." Hwy. Res. Record No. 396 (1972) (1967) pp. 115-136. pp. i-u.

28. GILES, C. G., ET AL., "Development and Perform- 44. GALLAWAY, B. M., and ROSE, J. G., "Comparison ance of the Portable Skid Resistance Tester." ASTM of Highway Friction Measurements Taken in the Spec. Tech. Publ. No. 326 (1962) pp. 50-74. Slip and Skid Modes." Hwy. Res. Record No. 376

29. MAYCOCK, G., "Studies on the Skidding Resistance 1971) pp. 107-122. of Passenger-Car Tyres on Wet Surfaces." Proc. 45. KUMMER, H. W., "Correlation Tests with the Penn Inst. Mech. Engrs., Autom, Div., Vol. 180, Pt. 2A State Drag Tester." Rept. J9, Joint Rd. Frict. Pro- (1965-66) pp. 122-141. . gram, Pa. Dept. of Highways and Pa. State Uni-

30. GILES, C. G., "Some Recent Developments in Work versity (1964) 32 pp.

64

ZUBE, E., and SKOG, J., "A Study of the Pennsylvania State Drag Tester for Measuring the Skid Resistance of Pavement Surfaces." Rept. M&R 633251, Mat. and Res. Dept., California Div. of Highways (1967)

12 pp. BURCHETT, J. L., JR., and RIZENBERGS, R. L., "Pave- ment Slipperiness Studies." Kentucky Department of Highways, Interim Res. Rep. on Project KYHPR- 64-24 (1970) 40 pp. HEGMON, R. R., and MizoGucHi, M., "Pavement Texture Measurement by the Sand Patch and Out-flow Meter Methods." Auto. Saf. Res. Program Rept. S40, The Pa. State University; Interim Rept. Agreement 38670 with Pa. Dept. of Transportation (1970) 39 pp. HORNE, W. B., and JOYNER, U. T., "Traction of Pneumatic Tires on Wet Runways." NASA Tech.

Note SP83 (1965). GALLAWAY, B. M., ET AL., "The Relative Effects of Several Factors Affecting Rainwater Depths on Pave-ment Surfaces." Hwy. Res. Record No. 396 (1972)

pp. 59-71. SABEY, B. E., "Road Surface Texture and Change in Skidding Resistance with Speed." Rep. 20, British Road Research Laboratory (1966). CHAMBERLIN, W. P., and AMSLER, D. E., "Pilot Field Study of Concrete Pavement Texturing Meth- ods." Hwy. Res. Record No. 389 (1972) pp. 5-17. GILLESPIE, T. D., "Pavement Surface Characteristics and Their Correlation with Skid Resistance." Rept. J12, Joint Road Friction Program, Pa. State Uni- versity-Pa. Dept. of Highways (1965) 100 pp. SABEY, B. E., and LUPTON, G. N., "Measurement of Road Surface Texture Using Photogrammetry." Rept. LR 57, British Road Research Laboratory (1967). KUMMER, H. W., "Unified Theory of Rubber and Tire Friction." Eng. Res. Bull. B-94, The Pennsyl-

vania State Univ. (1966) 135 pp. SCHONFELD, R., "Photo-Interpretation of Skid Re-sistance." Hwy. Res. Record No. 311. (1970) pp.

11-25. YANDELL, W. 0., "A New Theory of Hysteretic Sliding Friction." Wear, Vol. 17 (1971) pp. 229-244.

HORNE, W. B., and DREHER, R. C., "Phenomena of Pneumatic Tire Hydroplaning." NASA Tech. Note

D-2056 (1963) 52 pp. MARTINEZ, J. E., ET AL., "A Study of Variables Associated with Wheel Spin-Down and Hydro- planing." Hwy. Res. Record No. 396 (1972) pp. 33-44. SHERMAN, G. B., "Grooving Pattern Studies in Cali-fornia." Hwy. Res. Record No. 376 (1971) pp.

63-64. FARNSWORTH, E. E., "Continuing Studies of Pave-ment Grooving in California." HRB Spec. Rep. 116

(1971) pp. 134-137. BEATON, J. L., ET AL., "Effect of Pavement Groov-ing on Motorcycle Rideability." Rep. M&R 633126-6, California Div. of Highways (1969).

STIFFLER, A. K., "Relation Between Wear and Physi-cal Properties of Roadstones." HRB Spec. Rep. 101

(1969) pp. 56-58. SCHNEIDER, A. G., "First Results of a Study on Aggregate Wear and Polishing." Report J13, Pa. Dept. of Highways-Pa. State University Joint Road Friction Program (1966) 29 pp. GooDwiN, W. A., "Pre-Evaluation of Pavement Ma-terials for Skid Resistance-A Review of U.S. Tech-niques." HRB Spec. Rep. 101 (1969) pp. 69-79. MULLEN, W. G., ET AL., "Two Laboratory Methods for Evaluating Skid Resistance Properties of Aggre-gates." Hwy. Res. Record No. 376 (1971) pp. 123-135. "Road Research 1970." Annual Report of the Road Research Laboratory, Department of Environment (United Kingdom) (1971) pp. 26-27. DAHIR, S. H. M., and MULLEN, W. G., "Factors Influencing Aggregate Skid Resistance Properties." Hwy. Res. Record 376 (1971) pp. 136-148. Gity, J. E., and RENNINGER, F. A., "The Skid-Resistant Properties of Carbonate Aggregates." Hwy. Res. Record No. 120 (1965) pp. 18-34. SHUPE, J. W., and LOUNSBURY, R. W., "Polishing Characteristics of Mineral Aggregates." Proc. First Internat. Skid Prevention Conf. (publ. Univ. of Va.) (1958) pp. 509-537. PREUS, C. K., "Effect of Studded Tires on Aggregate Wear Rates and Related Effects on Skid Resistance." Hwy. Res. Record No. 341 (1971) pp. 46-48. KULLBERG, G., and OHLSSEN, E., "Dubbade Daeck, Undersoekningar 1963-1964" (Studded Tire Investi-gations, in Swedish) Spec. Rep. 38, Statens Vaegin-stitut, Sweden (1965) 57 pp. WHITEHURST, E. A., "The Cornering Capacity of Studded Tires." Highway Skid Resistance, ASTM Spec. Tech. Pubi. 456 (1969) pp. 144-158. ROSENTHAL, P., ET AL., "Evaluation of Studded Tires." NCHRP Report 61 (1969) 66 pp. MORTIMER, T. P., and LUDEMA, K. C., "The Effects of Salts on Road Drying Rates, Tire Friction, and Invisible Wetness." Hwy. Res. Record No. 396 (1972) pp. 45-58. SHERWOOD, W. C., "The Role of Aggregate Type in Pavement Slipperiness." Proc. 16th Annual High-way Geology Symposium (1965). DILLARD, J. H., "Summation of Practices of Virginia Department of Highways in Providing Skid Resistant Roads." Presentation to ASTM Committee 3-F of Committee D-4, unpublished (1969). GooDwIN, W. A., "Pre-Evaluation of Bituminous Mixes for Skid Resistance." Proc. Southeastern Assn. of State Highway Officials (1962) p. 55.

BLANK, H. R., and LEDBETTER, W. B., "Synthetic Aggregates." Texas Highways (1968). "Slag-The All Purpose Construction Aggregate." The National Slag Association (1970). AHMED, A., "The Effect of Electric Furnace Slag on the Strength of Concrete." Unpubl. thesis, Univ. of Tennessee (1956).

65

82. SHUPE, J. W., and GOETZ, W. H., "A Laboratory 99. Investigation of Pavement Slipperiness." HRB Bull. 219 (1959) pp. 56-73.

83. NICHoLS, F. P., JR., Discussion of "Predetermining the Polish Resistance of Limestone Aggregates," (Ref. 90). Addendum to Hwy. Res. Record No. 341 (1970) pp. 1-7.

84. KEARNEY, E. J., ET AL., "Development of Specifica- tions for Skid-Resistant Asphalt Concrete." Hwy. Res. Record No. 396 (1972) pp. 12-20. 102.

85. H05KING, J. R., "The Role of Aggregates in Provid- ing Skid-Resistant Roads." Symposium on the In- fluence of the Road Surface on Skidding, Univ. of 103. Salford (1968).

86. BALMER, G. G., and COLLEY, B. E., "Laboratory Studies of the Skid Resistance of Concrete." Jour. 104. of Materials, Vol. 1 (1966) pp. 536-559.

87. STUTZENBERGER, W. J., and HAVENS, J. H., "A Study of the Polishing Characteristics of Limestone and 105. Sandstone Aggregates in Regard to Pavement Slip- periness." HRB Bull. 186 (1958) pp. 58-81.

88. ROSE, J. G., ET AL., "Macro-Texture Measurements and Related Skid Resistance at Speeds from 20 to 60 mph." Hwy. Res. Record No. 341 (1970) pp. 33.45

89. BURNETT, W. C., ET AL., "Skid Resistance of Bi- tuminous Surfaces." Hwy. Res. Record No. 236 (1968) pp. 49-60. 108.

90. SHERWOOD, W. C., and MAHONE, D. C., "Pre- determining the Polish Resistance of Limestone Ag- gregates." Hwy. Res. Record No. 341 (1970) pp. 109.

1-10. 91. GALLAWAY, B. M., and HARGETT, E. R., "Blending

Lightweight Aggregates with Natural Aggregates for 110

the Production of Bituminous Concrete." Hwy. Res. Record No. 273 (1969) pp. 42-52.

92. GooDWIN, W. A., "Rockwood Test Section." Un- 111

publ. Quarterly Report, Tennessee Dept. of High- ways (1953).

112. 93. SALT, G. F., "Skidding Research in Great Britain and

Its Effect on Materials Requirements." British Gran- ite and Whinstone Fed., Vol. 9, No. 1 (Spring 1969) 8 pp.

94. HUTCHINSON, J. W., ET AL., "Pavement Dynamic Permeability Testing." Highway Skid Resistance, 113. ASTM Spec. Tech. Pubi. 456 (1969) pp. 159-176.

-95. COLLEY, B. E., ET AL., "Factors Affecting Skid Re- sistance and Safety of Concrete Pavements." HRB Spec. Rep. 101 (1969) pp. 80-99. 114.

"Design and Control of Concrete Mixtures." Port- land Cement Association, Chicago, Ill.

SPELLMAN, D. L., "Texturing of Concrete Pave- 115. ment." HRB Spec. Rep. 101 (1969) pp. 100-103.

98. "First International Skid Prevention Conference." Charlottesville, Va, (Aug. 1969).

"Construction of Pavement Surfacing to Provide Safer Courses of Skid Resistance." Instruct. Memor. 21-3 -68, U.S. Bureau of Public Roads. MOORE, A. B., "A Statewide Inventory of Pavement Skid Resistance." Quarterly Progress Report, Univ. of Tennessee, unpubl. (1972). GooDwiN, W. A., "Skid Resistance of Highway Pavements." Quarterly Report, Tennessee Highway Research Program, unpubl. (1962). "Pavement Grooving and Traction Studies." Con-ference Proc., NASA, Langley Research Center (1969). FARNSWORTH, E. E., "Pavement Grooving on High-ways." Conference Proc., NASA, Langley Research Center (1969). GOODWIN, W. A., ET AL., "Nonskid Surfacing-Rockwood, Tennessee." Tennessee Highway Re-search Program, Reprint No. 11 (1961). SERAFIN, P. J., "Michigan's Experience with Differ-ent Materials and Designs for Skid Resistance of Bituminous Pavements." Hwy. Res. Record No. 341 (1970) pp. 22-32. ADAM, Z., "Louisiana Skid Resistant Study, Asphalt." The Asphalt Institute (1970). BOLLING, D. Y., "Open-Graded Plant Mix Surface Courses in the Washington Area." Skid Conference, West Virginia Highway Department, unpubl. (1971). MAUPIN, G. W., JR., "Sprinkle Treatment Increases Highway Skid Resistance." Civil Eng., Vol. 42, No. 2 (Feb. 1972) p. 51. MORAN, W. B., Executive Director, International Slurry Seal Association. Communication, unpubl. (1971). "Highway Design, Construction and Maintenance." Highway Safety Program Manual, Vol. 12, Federal Highway Admin. (1971) 61 pp. "Methods for Sampling and Testing of Mineral Ag-gregates, Sand and Fillers." Standard BS 812, Brit. Stds. Inst. (Refer to latest edition). GLENNON, I. C., "Evaluation of Safety Criteria Em-ployed in Highway Curve Design." Final Report, Part I on Task Order 4, Reevaluation of Safe Side Friction Factors Used in Highway Design, NCHRP Project 20-7, Texas Transportation Inst. (1969) 77 pp. HANKINS, K. D., "A Laboratory and Field Evalua-tion of the Polishing Characteristics of Texas Aggre-gates.", HRP Annual Report, Texas Highway Dept. (1968): ZIJBE, E., and SKOG, J., "Skid Resistance of Screen-ings for Seal Coats." Hwy. Res. Record No. 236 (1968) pp. 29-48. OLIVER, D. C., "A Synthesis of Case Law Juris-prudence Relating to Wet-Weather Highway Con-ditions." Hwy. Res. Record No. 376 (1971) pp. 29-36.

66

APPENDIX A

SUMMARY OF AN UPDATED INVENTORY OF EXISTING PRACTICES AND SOLUTIONS TO

SLIPPERY PAVEMENTS, 1971

A task group of Highway Research Board Committee D-B4 (now A2B04) on Surface Properties-Vehicle Interaction compiled in 1969 an "Inventory of Existing Practices and Solutions to Slippery Pavements." The task group report was published in February 1970 as Highway Research Circular 106.

This inventory was updated in July 1971 by the Pave-ment Design Branch, Highway Design Division, Office of Engineering, Federal Highway Administration. This sec-ond survey employed the same format as the first, except that a few more questions were submitted to the 50 states, the District of Columbia, and Puerto Rico. A summary of the updated inventory follows.

it is interesting to note that the percentage of states that consider slippery pavements to be of major concern de-creased from 46 in 1969 to 33 in 1971. As the answers to the subsequent questions in the two surveys indicate, this does not signify an altered attitude toward an essen-tially unchanged condition, but that skid resistance is being brought under control, though slowly. The number of states taking annual skid resistance inventories increased from 17 to 25, and 25 states now specify skid-resistant materials for wearing surfaces whereas only 17 did so in 1969. Thus, progress is being made but, clearly, more effort is still required.

Broad Categories Response

Attitude . 16 states reported that they consider skid resistance a major problem 26 considered it moderate 6 considered it minor

Equipment . 33 states now have skid trailers 5 have mu meters 4 have decelerometers 4 use the stopping distance method 15 states do not own skid trailers but 6 plan to acquire them and 9 do not plan to acquire any 9 states use portable testers

Inventory . 25 states conduct an annual inventory of.pavement skid resistance 37 are doing research on skid problems 41 measure skid resistance

Criteria for 32 states use skid test data as a criterion for Resurfacing resurfacing

44 states use accident data with 37 obtaining the data from programs for detection and identification of hazardoas locations 30 analyze wet-pavement accidents separately 36 investigate other factors that might contribute to skidding 19 establish a priority for resurfacing and deslicking

Specifications . 3 states specify a minimum skid number for pavements of Minimum Skid in service and 15 contemplate a number Numbers

4 use 352 use 37

)

at 40mph

luses38) 1 uses 15 at SO mph 1 u s e s

a texture meter 2 are developing systems

9 states have policies directing resurfacing or de-clicking of pavements falling below a minimum skid number

Broad Categories Response

Skid Resistant . 7 states presently have pavement design criteria that Design Criteria specify adequate skid resistance for new pavements

3 have pavement design criteria that specify adequate skid resistance for the life of the pavement

Materials . 21 states reported that they have aggregates known to polish 14 restrict the use of the identified polishing materials Limestone, dolomite aggregate, and siliceous material were noted as coninon examples of polishing materials 10 states are conducting accelerated wear tests to evaluate the suitability of materials 24 states specify certain materials for wearing sur-faces to assure high skid resistance 4 specify different aggregate size for different design speeds

Portland Cement . 24 states reported skid resistance problems with PCC Concrete Pave- . 13 use grooving to improve PCC pavements nests . 4 report grooving to be very effective, but most of

the other states have not had time to make adequate determi nations Some states noted that studded tires may have nullified possible benefits California noted a 75% redaction in accidents and Connecticut noted a 73% redaction in accidents on grooved pavements

Surface Texture . 16 states reported that surface texture of pavements was measured

Speed Reduction . 4 states reduce speed limits during rainy weather Some states use slippery when wet signs

Published reports of the


are available from:

Highway Research Board National Academy of Sciences

2101 Constitution Avenue Washington, D.C. 20418

Rep. No. Title

- A Critical Review of Literature Treating Methods of Identifying Aggregates Subject to Destructive Volume Change When Frozen in Concrete and a Proposed Program of Research—Intermediate Report (Proj. 4-3(2)), 81p., $1.80

1 Evaluation of Methods of Replacement of Deterio- rated Concrete in Structures (Proj. 6-8), 56 p., $2.80

2 An Introduction to Guidelines for Satellite Studies of Pavement Performance (Proj. 1-1), 19 p., $1.80

2A Guidelines for Satellite Studies of Pavement Per- formance, 85 p.+9 figs., 26 tables, 4 app., $3.00

3 Improved Criteria for Traffic Signals at Individual Intersections—Interim Report (Proj. 3-5), 36 p., $1.60

4 Non-Chemical Methods of Snow and Ice Control on Highway Structures (Proj. 6-2), 74 p., $3.20

5 Effects of Different Methods of Stockpiling Aggre-gates—Interim Report (Proj. 10-3), 48 p., $2.00

6 Means of Locating and Communicating with Dis-abled Vehicles—Interim Report (Proj. 3-4), 56 p. $3.20

7 Comparison of. Different Methods of Measuring Pavement Condition—Interim Report (Proj. 1-2), 29 p., $1.80

8 Synthetic Aggregates for Highway Construction (Proj. 4-4), 13 p., $1.00

9 Traffic Surveillance and Means of Communicating with Drivers—Interim Report (Proj. 3-2), 28 p., $1.60

10 Theoretical Analysis of Structural Behavior of Road Test Flexible Pavements (Proj. 1-4), 31 p., $2.80

11 Effect of Control Devices on Traffic Operations— Interim Report (Proj. 3-6), 107 p., $5.80

12 Identification of Aggregates Causing Poor Concrete Performance When Frozen—Interim Report (Proj. 4-3(1)), 47p., $3.00

13 Running Cost of Motor Vehicles as Affected by High- way Design—Interim Report (Proj. 2-5), 43 p., $2.80

14 Density and Moisture Content Measurements by Nuclear Methods—Interim Report (Proj. 10-5), 32 p., $3.00

15 Identification of Concrete Aggregates Exhibiting Frost Susceptibility—Interim Report (Proj. 4-3(2)), 66 p., $4.00

16 Protective Coatings to Prevent Deterioration of Con- crete by Deicing Chemicals (Proj. 6-3), 21 p., $1.60

17 Development of Guidelines for Practical and Realis- tic Construction Specifications (Proj. 10-1), 109 p., $6.00

18 Community Consequences of Highway Improvement (Proj. 2-2), 37 p., $2.80

19 Economical and Effective Deicing Agents for Use on Highway Structures (Proj. 6-1), 19 p., $1.20

* Highway Research Board Special Report 80.

Rep. No. Title

20 Economic Study of Roadway Lighting (Proj. 5-4),

77 p., $3.20 21 Detecting Variations in Load-Carrying Capacity of

Flexible Pavements (Proj. 1-5), 30 p., $1.40 22 Factors Influencing Flexible Pavement Performance

(Proj. 1-3(2)), 69 p., $2.60 23 Methods for Reducing Corrosion of Reinforcing

Steel (Proj. 6-4), 22 p., $1.40 24 Urban Travel Patterns for Airports, Shopping Cen-

ters, and Industrial Plants (Proj. 7-1), 116 p., $5.20

25 Potential Uses of Sonic and Ultrasonic Devices in Highway Construction (Proj. 10-7), 48 p., $2.00

26 Development of Uniform Procedures for Establishing Construction Equipment Rental Rates (Proj. 13-1), 33 p., $1.60

27 Physical Factors Influencing Resistance of Concrete to Deicing Agents (Proj. 6-5), 41 p., $2.00

28 Surveillance Methods and Ways and Means of Com- municating with Drivers (Proj. 3-2), 66 p., $2.60

29 Digital-Computer-Controlled Traffic Signal System for a Small City (Proj. 3-2), 82 p., $4.00

30 Extension of AASHO Road Test Performance Con- cepts (Proj. 1-4(2)), 33 p., $1.60

31 A Review of Transportation Aspects of Land-Use Control (Proj. 8-5), 41 p., $2.00

32 Improved Criteria for Traffic Signals at Individual Intersections (Proj. 3-5), 134 p., $5.00

33 Values of Time Savings of Commercial Vehicles (Proj. 2-4), 74 p., $3.60

34 Evaluation of Construction Control Procedures— Interim Report (Proj. 10-2), 117 p., $5.00

35 Prediction of Flexible Pavement Deflections from Laboratory Repeated-Load Tests (Proj. 1-3(3)), 117 p., $5.00

36 Highway Guardrails—A Review of Current Practice (Proj. 15-1), 33 p., $1.60

37 Tentative Skid-Resistance Requirements for Main Rural Highways (Proj. 1-7), 80 p., $3.60

38 Evaluation of Pavement Joint and Crack Sealing Ma- terials and Practices (Proj. 9-3), 40 p., $2.00

39 Factors Involved in the Design of Asphaltic Pave- ment Surfaces (Proj. 1-8), 112 p., $5.00

40 Means of Locating Disabled or Stopped Vehicles (Proj. 3-4(1)), 40 p., $2.00

41 Effect of Control Devices on Traffic Operations (Proj. 3-6), 83 p., $3.60

42 Interstate Highway Maintenance Requirements and Unit Maintenance Expenditure Index (Proj. 14-1), 144 p., $5.60

43 Density and Moisture Content Measurements by Nuclear Methods (Proj. 10-5), 38 p., $2.00

44 Traffic Attraction of Rural Outdoor Recreational Areas (Proj. 7-2), 28 p., $1.40

45 Development of Improved Pavement Marking Ma- terials—Laboratory Phase (Proj. 5-5), 24 p., $1.40

46 Effects of Different Methods of Stockpiling and Handling Aggregates (Proj. 10-3), 102 p., $4.60

47 Accident Rates as Related to Design Elements of Rural Highways (Proj. 2-3), 173 p., $6.40

48 Factors and Trends in Trip Lengths (Proj. 7-4), 70 p., $3.20

49 National Survey of Transportation Attitudes and Behavior—Phase I Summary Report (Proj. 20-4), 71 p., $3.20

Rep. Rep. No. Title No. Title 50 Factors Influencing Safety at Highway-Rail Grade 76 Detecting Seasonal Changes in Load-Carrying Ca-

Crossings (Proj. 3-8), 113 p., $5.20 pabilities of Flexible Pavements (Proj. 1-5(2)), 51 Sensing and Communication Between Vehicles (Proj. 37 p., $2.00 3-3), 105 p., $5.00 77 Development of Design Criteria for Safer Luminaire

52 Measurement of Pavement Thickness by Rapid and Supports (Proj. 15-6), 82 p., $3.80 Nondestructive Methods (Proj. 10-6), 82 P., 78 Highway Noise—Measurement, Simulation, and

53 $3.80 Multiple Use of Lands Within Highway Rights-of- 79

Mixed Reactions (Proj. 3-7), 78 p., $3.20 Development of Improved Methods for Reduction of

Way (Proj. 7-6), 68 p., $3.20 Traffic Accidents (Proj. 17-1), 163 p., $6.40 54 Location, Selection, and Maintenance of Highway 80 Oversize-Overweight Permit Operation on State High- Guardrails and Median Barriers (Proj. 15-1(2)), ways (Proj. 2-10), 120 p., $5.20 63 p., $2.60 81 Moving Behavior and Residential Choice—A Na-

55 Research Needs in Highway Transportation (Proj. tional Survey (Proj. 8-6), 129 p., $5.60 20-2), 66 p., $2.80 82 National Survey of Transportation Attitudes and 56 Scenic Easements—Legal, Administrative, and Valua- Behavior—Phase II Analysis Report (Proj. 20-4), tion Problems and Procedures (Proj. 11-3), 174 p., 89 p., $4.00 $6.40 83 Distribution of Wheel Loads on Highway Bridges

57 Factors Influencing Modal Trip Assignment (Proj. (Proj. 12-2), 56 p., $2.80 8-2), 78 p., $3.20 84 Analysis and Projection of Research on Traffic

58 Comparative Analysis of Traffic Assignment Tech- Surveillance, Communication, and Control (Proj. niques with Actual Highway Use (Proj. 7-5), 85 p., 3-9), 48 p., $2.40 $3.60 85 Development of Formed-in-Place Wet Reflective

59 Standard Measurements for Satellite Road Test Pro- Markers (Proj. 5-5), 28 p., $1.80 gram (Proj. 1-6), 78 p., $3:20 86 Tentative Service Requirements for Bridge Rail Sys-

60 Effects of Illumination on Operating Characteristics tems (Proj. 12-8), 62 p., $3.20 of Freeways (Proj. 5-2) 148 p., $6.00 87 Rules of Discovery and Disclosure in Highway Con-

61 Evaluation of Studded Tires—Performance Data and demnation Proceedings (Proj. 11-1(5)), 28 p., Pavement Wear Measurement (Proj. 1-9), 66 p., $2.00 $3.00 88 Recognition of Benefits to Remainder Property in

62 Urban Travel Patterns for Hospitals, Universities, Highway Valuation Cases (Proj. 11-1(2)), 24 p., Office Buildings, and Capitols (Proj. 7-1), 144 p., $2.00 $5.60 89 Factors, Trends, and Guidelines Related to Trip

63 Economics of Design Standards for Low-Volume Length (Proj. 7-4), 59 p., $3.20 Rural Roads (Proj. 2-6), 93 p., $4.00 90 Protection of Steel in Prestressed Concrete Bridges

64 Motorists' Needs and Services on Interstate Highways (Proj. 12-5), 86 p., $4.00 (Proj. 7-7) 88 p. $3.60 91 Effects of Deicing Salts on Water Quality and Biota

65 One-Cycle Slow-Freeze Test for Evaluating Aggre- —Literature Review and Recommended Research gate Performance in Frozen Concrete (Proj. 4-3(1)), (Proj. 16-1), 70 p., $3.20 21 p $1 40 92 Valuation and Condemnation of Special Purpose

66 Identification of Frost-Susceptible Particles in Con- Properties (Proj. 11-1(6)), 47 p., $2.60 crete Aggregates (Proj. 4-3 (2) ) 62 p. $2.80 93 Guidelines for Medial and Marginal Access Control

67 Relation of Asphalt Rheological Properties to Pave- on Major Roadways (Proj. 3-13), 147 p., ment Durability (Proj. 9-1), 45 p., $2.20 $6.20

68 Application of Vehicle Operating Characteristics to 94 Valuation and Condemnation Problems Involving Geometric Design and Traffic Operations (Proj. 3_ Trade Fixtures (Proj. 11-1(9)), 22 p., $1.80 10), 38 p., $2.00 95 Highway Fog (Proj. 5-6), 48 p., $2.40

69 Evaluation of Construction Control Procedures— 96 Strategies for the Evaluation of Alternative Trans- Aggregate Gradation Variations and Effects (Proj. portation Plans (Proj. 8-4), 111 p., $5.40 10-2A), 58 p., $2.80 97 Analysis of Structural Behavior of AASHO Road

70 Social and Economic Factors Affecting Intercity Test Rigid Pavements (Proj. 1-4(1)A), 35 p., Travel (Proj. 8-1), 68 p., $3.00 $2.60 71 Analytical Study of Weighing Methods for Highway 98 Tests for Evaluating Degradation of Base Course

Vehicles in Motion (Proj. 7-3), 63 p., $2.80 Aggregates (Proj. 4-2), 98 P. $5.00 72 Theory and Practice in Inverse Condemnation for 99 Visual Requirements in Night Driving (Proj. 5-3),

Five Representative States (Proj. 11-2), 44 p., 38 p., $2.60 $2.20 100 Research Needs Relating to Performance of Aggre-

73 Improved Criteria for Traffic Signal Systems on gates in Highway Construction (Proj. 4-8), 68 p., Urban Arterials (Proj. 3-5/ 1), 55 p., $2.80 $3.40 74 Protective Coatings for Highway Structural Steel 101 Effect of Stress on Freeze-Thaw Durability of Con-

(Proj. 4-6), 64 p., $2.80 crete Bridge Decks (Proj. 6-9), 70 P., $3.60 74A Protective Coatings for Highway Structural Steel— 102 Effect of Weldments on the Fatigue Strength of Steel Literature Survey (Proj. 4-6), 275 p., $8.00 Beams (Proj. 12-7), 114.p., $5.40 74B Protective Coatings for Highway Structural Steel— 103 Rapid Test Methods for Field Control of Highway Current Highway Practices (Proj. 4-6), 102 p., Construction (Proj. 10-4), 89 p., $5.00

75 $4.00 Effect of Highway Landscape Development on

104 Rules of Compensability and Valuation Evidence for Highway Land Acquisition (Proj. 11-1),

Nearby Property (Proj. 2-9), 82 p., $3.60 77 p., $4.40

Rep. No Title

Dynamic Pavement Loads of Heavy Highway Vehi- cles (Proj. 15-5), 94 p., $5.00 Revibration of Retarded Concrete for Continuous Bridge Decks (Proj. 18-1), 67 p., $3.40 New Approaches to Compensation for Residential Takings (Proj. 11-1(10)), 27 p., $2.40 Tentative Design Procedure for Riprap-Lined Chan- nels (Proj. 15-2), 75 p., $4.00 Elastomeric Bearing Research (Proj. 12-9), 53 p., $3.00 Optimizing Street Operations Through Traffic Regu- lations and Control (Proj. 3-11), 100 p., $4.40 Running Costs of Motor Vehicles as Affected by Road Design and Traffic (Proj. 2-5A and 2-7),

97 p., $5.20 Junkyard Valuation—Salvage Industry Appraisal Principles Applicable to Highway Beautification (Proj. 11-3(2)), 41 p., $2.60 Optimizing Flow on Existing Street Networks (Proj. 3-14), 414p., $15.60 Effects of Proposed Highway Improvements on Prop- erty Values (Proj. 11-1(1)), 42.p., $2.60 Guardrail Performance and Design (Proj. 15-1(2)),

70 p., $3.60 Structural Analysis and Design of Pipe Culverts (Proj. 15-3), 155 p., $6.40 Highway Noise—A Design Guide for Highway En- gineers (Proj. 3-7), 79 p., $4.60 Location, Selection, and Maintenance of Highway Traffic Barriers (Proj. 15-1(2)), 96 p., $5.20 Control of Highway Advertising Signs—Some Legal Problems (Proj. 11-3(1)), 72 p., $3.60 Data Requirements for Metropolitan Transportation Planning (Proj. 8-7), 90 p., $4.80 Protection of Highway Utility (Proj. 8-5), 115 p., $5.60 Summary and Evaluation of Economic Consequences of Highway Improvements (Proj. 2-11), 324 p., $13.60 Development of Information Requirements and Transmission Techniques for Highway Users (Proj. 3-12) 239 p., $9.60 Improved Criteria for Traffic Signal Systems in Ur- ban Networks (Proj. 3-5) 86 p., $4.80 Optimization of Density and Moisture Content Mea- surements by Nuclear. Methods (Proj. 10-5A), 86 p., $4.40 Divergencies in Right-of-Way Valuation (Proj. 11- 4), 57 p., $3.00 Snow Removal and Ice Control Techniques at Inter- changes (Proj. 6-10), 90 p., $5.20 Evaluation of AASHO Interim Guides for Design of Pavement Structures (Proj. 1-11), 111 p., $5.60 Guardrail Crash Test Evaluation—New Concepts and End Designs (Proj. 15-1(2)), 89 p., $4.80 Roadway Delineation Systems (Proj. 5-7), 349 p., $14.00 Performance Budgeting System for Highway Main- tenance Management (Proj. 19-2(4)), 213 p., $8.40 Relationships Between Physiographic Units and Highway Design Factors (Proj. 1-3(1)), 161 p., $7.20 Procedures for Estimating Highway User Costs, Air Pollution, and Noise Effects (Proj. 7-8), 127 p., $5.60

Synthesis of Highway Practice

No. Title

1 Traffic Control for Freeway Maintenance (Proj. 20-5, Topic 1), 47 p., $2.20

2 Bridge Approach Design and Construction Practices (Proj. 20-5, Topic 2), 30 p., $2.00

3 Traffic-Safe and Hydraulically Efficient Drainage Practice (Proj. 20-5, Topic 4), 38 p., $2.20

4 Concrete Bridge Deck Durability (Proj. 20-5, Topic 3), 28 p., $2.20

5 Scour at Bridge Waterways (Proj. 20-5, Topic 5), 37 p., $2.40

6 Principles of Project Scheduling and Monitoring (Proj 20-5, Topic 6), 43 p., $2.40

7 Motorist Aid Systems (Proj. 20-5, Topic 3-01), 28 p., $2.40

8 Construction of Embankments (Proj. 20-5, Topic 9), 38 p., $2.40

9 Pavement Rehabilitation—Materials and Techniques (Proj. 20-5, Topic 8), 41 p., $2.80

10 Recruiting, Training, and Retaining Maintenance and Equipment Personnel (Proj. 20-5, Topic 10), 35 p., $2.80

11 Development of Management Capability (Proj. 20-5, Topic 12), SOp., $3.20

12 Telecommunications Systems for Highway Admin-istration and Operations (Proj. 20-5, Topic 3-03) 29 p., $2.80

13 Radio Spectrum Frequency Management (Proj. 20-5, Topic 3-03) 32 p., $2.80

14 Skid Resistance (Proj. 20-5, Topic 7), 66 p., $4.00

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

THE NATIONAL ACADEMY OF SCIENCES is a private, honorary organiza-lion of more than 700 scientists and engineers elected on the basis of outstanding contributions to knowledge. Established by a Congressional Act of Incorporation signed by President Abraham Lincoln on March 3, 1863, and supported by private and public funds, the Academy works to further science and its use for the general welfare by bringing together the most qualified individuals to deal with scientific and technological problems of broad significance.

Under the terms of its Congressional charter, the Academy is also called upon to act as an official—yet independent—adviser to the Federal Government in any matter of science and technology. This provision accounts for the close ties that have always existed between the Academy and the Government, although the Academy is not a governmental agency and its activities are not limited to those on behalf of the Government.

THE NATIONAL ACADEMY OF ENGINEERING was established on December 5, 1964. On that date the Council of the National Academy of Sciences, under the authority of its Act of Incorporation, adopted Articles of Organization bringing the National Academy of Engineering into being, independent and autonomous in its organization and the election of its members, and closely coordinated with the National Academy of Sciences in its advisory activities. The two Academies join in the furtherance of science and engineering and share the responsibility of advising the Federal Government, upon request, on any subject of science or technology.

THE NATIONAL RESEARCH COUNCIL was organized as an agency of the National Academy of Sciences in 1916, at the request of President Wilson, to enable the broad community of U. S. scientists and engineers to associate their efforts with the limited membership of the Academy in service to science and the nation. Its members, who receive their appointments from the President of the National Academy of Sciences, are drawn from academic, industal and government organizations throughout the country. The National Research Council serves both Academies in the discharge of their responsibilities.

Supported by private and public contributions, grants, and contracts, and volun-

*

tary contributions of time and effort by several thousand of the nation's leading scientists and engineers, the Academies and their Research Council thus work to serve the national interest, to foster the sound development of science and engineering, and to promote their effective application for the benefit of society.

THE DIVISION OF ENGINEERING is one of the eight major Divisions into which the National Research Council is organized for the conduct of its work. Its membership includes representatives of the nation's leading technical societies as well as a number of members-at-large. Its Chairman is appointed by the Council of the Academy of Sciences upon nomination by the Council of the Academy of Engineering.

THE HIGHWAY RESEARCH BOARD, organized November 11, 1920, as an agency of the Division of Engineering, is a cooperative organization of the high-way technologists of America operating under the auspices of the National Research Council and with the support of the several highway departments, the Federal Highway Administration, and many other organizations interested in the development of trans-portation. The purpose of the Board is to advance knowledge concerning the nature and performance of transportation systems, through the stimulation of research and dissemination of information derived therefrom.

HIGHWAY RESEARCH BOARD NATIONAL ACADEMY OF SCIENCES—NATIONAL RESEARCH COUNCIL 2101 Consttuflon Avenue Washington, D. C. 20418

ADDRESS CORRECTION REQUESTED

NON-PROFIT ORG.

U.S. POSTAGE

PAID

WASHINGTON, D.C.

PERMIT NO. 42970

rI

(I-

(-'4 w -'-•

ON

Lr. o. OLJ

.ow. I- -•L <0--

ow O

skid resistance - transportation research...

Documents