A mining research contract report

DECEMBER ] 985

EXPERIMENTAL STUDY OF A NOVEL JET BOOSTER PUMP

Contract J0333928

St. Anthony Falls Hydraulic Laboratory Departm~nt of Civil & Mineral Engineering University of Minnesota Mississippi River at 3rd Ave. S.E. Minneapolis, Minnesota 55414

BUREAU OF MINES

UNITED STATES DEPARTMENT OF THE INTERIOR

The views and conclusions contained in this document are those of the authors and should not be interpreted as necessarily representing the official policies or recommen­dations of the Interior Department's Bureau of Mines or of the U. S. Government.

The University of Minnesota is committed to the policy that all persons ~all have equal access to its programs, facilities, and employment Without regard to race, creed, color, sex, national origin, or handicap.

$n1'I-'!&1

REPORT DOCUMENTATION I ~ REP'Q.IIT NO. J" 3. ~pjent". A--. .....

PAGE 4. T1tl'tllCl kbtftl,

STUDY NOVEL JET BOOSTER PUMP ". "-"t (leW

XPERIMENTAL OF A Pec. 1985 I.

7. Alltttl,,151 Bintz

..... rf\lnnlq ~ ~ No. Joseph M. Wetzel, T. Jackson, D. W. P.R. 243

., Plll'fomlq Ol1lanlullon NaIM alld AcId,... UI. PnltMtlTaaklWoftl Unit He.

St. Anthony Falls Hydraulic Laboratory University of Minneosta lL eontr.et(C) ... Gr.nt<tIl No.

Mississippi River at 3rd Ave. S.E. (C) J0333928 Minneapolis, Minnesota (til

u. ~,.. OrpniDt"'" Na_ alld Add,... U. Typi ~ hPl"1 .. P'encId c-M

U.S. Bureau of Mines Final report Branch of Procurement

S4. P. O. Box 18107, Pittsburgh, PA 15236 9/13/80 - 12/31/85

lIS. $uppiMMlntary Nat ..

II. AbItnc:t (Umlt: 200 -as) ..r";

..

A new form of a low head, jet booster pump was evaluated. The jet pump con-1

(sisted of a constant diameter pipe fitted with two wall jets energizing the flow and ~

a constant area diffuser consisting of a slotted wall portion of the pipe where water· and solid fines were withdrawn. The withdrawn water was pressurized and recirculated~ to the j~ts by an external centrifugal slurry pump, thus performing a pressure 1

booster operation on the main line flow. ., ,

The jet pump was evaluated in a 3 inch pipe recirculating flow facility. Extensive tests were conducted with water flow to evaluate and optimize performance characteristics. The best efficiency obtained was about 17 percent. Addition of solid particles to the flow in sizes up to 3/4 inch resulted in plugging of the screened diffuser for concentrations above about 20 percent by weight. Co~.l par-ticles were rapidly eroded to very small sizes due to the action of the high velocity side jets. Problems associated with the screened diffuser limit the pumps' usefulness. , .... : . ·0 ~ ., ..

17. DocumetIt AM.,... L~""

b. 1cIentI"-~ Tarms

c. COSA TI FIeld/Group

lL " .... I8b111ty State_ lI, Se!:Urity ClaM (Thn ~ 2L _of ...... .. ·UNCLAS S IFIED

.. Ieaurtty c-. (Thn ..... 1 2Z. I'rIce UNCLASSIFIED

(aM AHSI-Z3t.lll ... '"8InIct ..... ell It_ WT10HAL I'ORM In {+-77) ('owmerI)' tm~51

3 Ill, .1'tIMI'It ~ ~

l

FOREWORD

This report was prepared by the St. Anthony Falls Hydraulic

Laboratory, University of Minnesota, Minneapolis, Minnesota, under USBM

Contract number J0333928. The contract was initiated under the Development

Program for a New Jet Booster Pump for Hydraulic Transport in Underground

Coal Mines. It was administered under the technical direction of the

Pittsburgh Research Center, Bureau of Mines, U. S. Department of the

Interior, with Richard C. Wang acting as Technical Project Officer. Louis

A. Summers was the contract administrator for the Bureau of Mines. This

report is a summary of the work recently completed as a part of this

contract during the period September 13, 1980 to December 31, 1985.

4

I.

II.

III.

IV.

V.

TABLE OF CONTENTS

Page No.

Foreword ••••••••• iII •••••••• ., •••••••••••••••••••••••••••••••• 4

List of Figures ." II , ......... " • II • ., • " " •• III • " II •• " II • " III • , ., , • II •• III • • 6

Introduction II •.••• , ••• , • It II •••• , II • " ..... " .. , •••• " " •• ., II •••••••••

Summary " ••..••••••••• " •••• " II •• " •••• " •• II • , ••••••••••• II " ..... III

The Test Facility •.• " .. "."."",,. II • " •• II" ..................... .

A. B.

C.

Recirculating Pipe Loop The Jet Pump Assembly

• ................... 41 " •••••••••••• . .... , ......................... . 1. Pump block unit ...................... II ••••• '. II • ,. II ......

2. Mixing chamber , •• , II • III •••••••••••••••• " •••••••••••••

3. Strainer-Diffuser ...................... ' •• " • " • II ••••••

Solids Handling ••••••••••••••••••

Discussion of Results . .................................... . A. B. C. D.

............................ Static Pressure Distribution Total Head Ratio and Efficiency . ...................... . . ........................................... . Cavitation Slurry Flows . ......................................... .

Acknowledgement . .......................................... . References .................................................

5

7

8

10

10 ]3 13 17 17 18

22

22 26 30 34

38

39

Figure 1

Figure 2

Figure :3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

LIST OF FIGURES

Assembly of Horizontal Loop.

Overall View of Facility.

Jet Pump Assembly.

Strainer Module and Suction Manifold.

Typical Static Pressure Gradient for Jet Pump Assembly.

Dimensionless Pressure Change Through Jet Pump.

Total Head Ratio for Two Nozzle Diameters, JPU1.

Typical Static Pressure Gradients for Jet Pump Assemblies.

Total Head Ratio vs. Mass Flow Ratio, JPU2.

Total Head Ratio vs. Mass Flow Ratio, JPU:3.

Effect of Jet Cavitation on Total Head Ratio.

6

I. INTRODUCTION

Pumping systems for long-line hydrotransport of solids in a pipeline

have generally been confined to two basic forms. These involve either

centrifugal pumps in which the total solids pass through the pump with the

water or systems employing centrifugal or piston pumps for pressurizing

water to force a charge of solids into the pipe via a charging chamber

downstream of the pump. The systems generally employ large high-head pumps

for overland transportation. In contrast to this method of solid

transport, the present study was directed towards development of compact

and portable low-head booster pumping units that would be closely spaced

along the haulage pipe in a coal mine. The low-head pumps consist of a

dual unit in which a conventional centrifugal slurry pump recirculates

screened water withdrawn from the haulage pipe to energize a form of

straight-through jet pump. The low pressure booster pump assembly would

permit the use of a light-weight, abrasion resistant plastic pipe system.

Multiple pumping units should aid in minimizing flow blockages and promote

start-up.

Use of conventional jet pumps for solid transport has been prevalent

in hydraulic dredging where solids of modest size are encountered (~).

Jet pumps using peripheral jets have been employed in one-pass systems (~).

However, in general they have not been satisfactory for long-line pumping

in multiple stages as each stage has introduced new energizing water. The

new water dilutes the slurry to an unacceptable condition.

Locally recirculating the energy-carrying water vehicle, by

withdrawing it from the main flow and then re-injecting it in a peripheral

jet system, overcomes the dilution problem and does away with the need for

a conventional constricted throat. Thus, the throat section is the same

size as the haulage pipe increasing the largest conveyable size of solid.

To evaluate the performance of a jet pump for this purpose, an experi­

mental study has been conducted. Results of this study are summarized in

the following sections.

7

II. SUMMARY

Experiments were conducted on a new form of a jet pump consisting of

two side jets entering at an angle of 20° to a constant diameter pipe.

The fluid for energizing the jets was withdrawn through screen elements

located downstream of the jets. This water was pressurized by a centrifu­

gal pump and recirculated to the jets. The pump was evaluated in a 3-inch

pipe loop with recirculating flow. To optimize the performance of the

unit, tests were made first with water flow. These tests were repeated

with slurry flows, with most emphasis being placed on the clogging problems

associated with the screened elements. Conclusions are summarized below:

1. The two sizes of jets tested were 1.26 in. and 0.904 in. diameter.

On the basis of tests with water flow, the larger jet resulted in

a steeper total head ratio versus mass flow ratio curve than the

smaller jet, which was considered desirable for transport of

solids.

2. The length of the mixing chamber and the number of modules

compris-ing the screened diffuser was varied. It was found that a

mixing chamber length of three pipe diameters was adequate, and

the number of modules could be reduced from the initial nine to

four.

3. The pressure rise across the jet pump can be predicted from a

semi-empirical equation developed for pipe flow.

4. The pressure rise in the screened diffuser was not very large due

to losses, which contributes to poor efficiency.

5. The maximum efficiency for the best configuration was about

17 percent.

6. Extensive cavitation in the jet region due to high jet velocities

and low suction pressures did not appear to significantly change

the total head ratio for the tested suction pressures down to

atmospheric.

8

7. Introduction of solid particles into the flow resulted in serious

difficulties. Some of these problems were associated with the

experimental facility) and were eventually resolved. The most

critical difficulty was associated with clogging of the screened

diffuser. This was found to occur for weight concentrations

greater than 20 percent in a very short period of time. Extensive

efforts to alleviate this plugging were unsuccessful. On the

basis of these operational difficulties, the usefulness of the

tested jet pump unit in field operations is not considered viable.

9

III. THE TEST FACILITY

A. Recirculating Pipe Loop

The main recirculating pipe loop was fabricated from 3 in. diameter

Schedule 40 PVC pipe. Selected sections were constructed of transparent

PVC pipe to permit flow visualization. The loop was designed so that it

could be placed in either a horizontal or vertical plane. A schematic of

the basic pipe loop oriented in the horizontal plane is shown in Fig. 1.

The overall dimensions of the loop are about 25 ft long and 5 ft wide. The

bends were long radius bends custom fabricated for this purpose. The

entire jet pump assembly, consisting of the nozzle block containing the two

nozzles, the mixing chamber, and strainer modules were made removable by

flanged connections. A 25 HP Weinman centrifugal pump was used to supply

high pressure water to the nozzles. Water was withdrawn through each of

the strainers into a manifold in the pump suction line. The discharge line

of the pump was run back to a specially fabricated bifurcation upstream of

the nozzle block. The nozzle block assembly was attached to this bifur­

cation with each of the two nozzles receiving the same flow. A valve in

the discharge line was used for flow control and a calibrated orifice meter

was used for discharge measurement. Flow recirculating through the main

3 in. diameter recirculating loop was controlled by a valve downstream of

the strainer assembly and measured with an orifice meter for the tests with

water. Details of the jet pump assembly are discussed in the later

sections.

Several auxiliary systems were added to the basic pipe loop. As the

temperature of the fluid in the closed system was expected to rise rapidly

due to friction losses, a cooling system was installed. A small quantity

of water was withdrawn from the centrifugal pump loop through a commercial

well point screen. The quantity of water withdrawn was replaced with cool

water maintaining an essentially constant temperature. In operation with

the slurry, some of the fine particles passing through the strainers caused

occasional plugging of the well screen, necessitating periodic back

flushing.

10

~

~

.......

3 '

~

Coolant Inject~on and Withdrawal Uni

Remove for Solids Injector Unit

3" Flow Manifold

I- ~I 3", sch.40, PVC

0-1

\.. 3" Orifice Meter

25' Plan View

Orifice Meter

Elevation

Fig. 1. Assembly of Horizontal Loop.

~.

5' I

sch. 40, Steel

,It:J ...... (0

Control Valve

HP

Overhead City Water Supply

Flushing Control Valve

To Variable Pressure Control

Coolant Inject{on and Withdrawal Unit

Well Screen and Point

Manifold Suction

Coolant Injection and Withdrawal Unit

24" Diameter I I

-.1_

-----; - -, Conical Cyclone Liner

-

I __ --+ __ J

Pinch Valve

-- ---I.'

t

3" PVC

Solids Withdrawal and Injector Unit

Figure 1. Inset view of assembly of horizontal loop.

12

Solids Storage

Port

I

1

I I

A system was also added to vary the surcharge pressure on the pipe

loop to determine the sensitivity of the jet pump to cavitation. It was

anticipated that the high velocities and turbulence in the mixing chamber

could produce heavy cavitation unless the ambient pressure was sufficiently

high. The teat loop was provided with a tap to which a hose was connected.

The far end of the hose was elevated and connected to a small overflow

vessel contained in a slightly larger tank. The entire tank could be

varied in elevation to about a maximum of 40 ft above the loop centerline.

Flow was adjusted so that the overflow in the inner vessel was a small,

steady flow. The water withdrawn was the water replaced with the cooling

system. This type of system for pressure control has been used for many

years in the Laboratory's water tunnels with good success. For most

tests, the reservoir was maintained at a high elevation to assure non­

cavitating conditions.

An overall view of the facility in the horizontal position is shown in

Fig. 2. The physical arrangement is slightly different from the schematic

sketch in Fig. 1, although all pertinent features are the same. The large

structure at the left and above the recirculating loop is the first design

of a cyclone separator to be used for injection and withdrawal of solids.

All piping to the separator was not completed at the time the photo was

taken.

B. The Jet Pump Assembly

The jet pump assembly consisted of three basic components: the pump

block unit, the mixing chamber, and the strainer-diffuser modules. A sche­

matic plan view of the assembly is shown in Fig. 3. Locations of pressure

taps used for reference purposes are indicated. Each of the components are

discussed below.

1. Pump block unit

The nozzles selected for study were two nozzles located diametrically

opposite each other on a horizontal plane. The angle used for the jets was

20 degrees with respect to the longitudinal axis of flow. It was felt that

the most favorable initial axial momentum component and a minimum stagnation

13

14

_ -L

3"

\.1'1

.... -------...........

---;,- -·:6D--~--. , ':. '=T

19-5/8" variable

Pump Block unit Mixing Chamber

PD

~I

1d..~:sf31t+1 H+--++-+-+

variable

Strainer­Diffuser

.. ~ --~-:',:t:"c;-- I lID ~ ~ --lID , , ,

-,---- ... ,.:%.'----.:. . .. ..-- ~ l\, ' ,. Q ....-: -~. . .~ ~ ;. D _~- 10._'

\. >:...--::.:);~./ .... -... -... -~ ... .,..,

-3" Jet Supply Pipe

PD

3"

Typical Spacing

Plan View Machined Surface 3.06" I.D.

to Match

20 0 -.------......c.. ,v \. .. ___ ... ~--L. __ _

6"

Jet Nozzle Insert

2-1.26" I.D. 2-0.904" I.D.

Fig. 3. Jet Pump Assembly.

Nozzle t.. Pipe t.

~L_~~_:~:~~:ge Pipe

!, " " •. , -* .. (jI--- .~.. -le.-.- Q • '.,\ S t _________ . ______ .D t •

Po

generated t'etut'nf1bw would be achi'evedwith a modest angle. The size

of the jet produced is dependent on the discharge and the ratio of the

entrained to jet discharge ratio. For a single phase jet pump, best

efficiencies were obtained with a ratio of jet to pipe area of about 0.17,

although some studies with dredging annular jet pumps had best efficiencies

with area ratios greater than 0.3. It was thus elected to fabricate two

sets of nozzles with diameters of 1.26 in. and 0.904 in., as standard sizes

of stainless tubing were available. The two nozzles then provided an area

ratio of 0.34 for the larger tubing, and 0.17 for the smaller tubing.

Several other factors were considered in the nozzle design:

1. As the jet velocity may be h;tgh' boundary contact in the

nozzle should be minimized.

2. Overall length of the pipe and jet combination should

be minimized.

3. The nozzle should be simple and easily replaceable.

These considerations led to the fabrication shown in Fig. 3. The

large block containing the two nozzles was machined from a solid piece of

Lucite. Sockets were bored to receive the four sections of 3 in. diameter

pipe. As the nozzle size was to be varied, the nozzles were fabricated as

inserts from stainless steel tubing to avoid machining another block for

each size. The steel inserts also were not susceptible to erosion. Two

pairs of nozzles were machined; one with 1.5 in. OD, 1.26 in. ID and one

with 1 in. OD, 0.904 in. ID. The Lucite block was bored for the 1.5 in. OD

tubing. The tubing was fastened to a plate to fit into the 3 in. opening

and held in position with screws. After a pair of nozzles was installed,

the intersection of the nozzle with the through bore was marked. The

nozzles were removed and machined so that they were flush with the inside

of the pipe. Following reassembly, the nozzles were hand honed.

16

2. Mixing chamber

Good jet pump performance requires a length of pipe following the jets

to serve as a mixing chamber. The mixing chamber permits a momentum

exchange to take place between the high velocity jets and the slower moving

entrained fluid.

For these tests, the diameter of the mixing chamber was equal to the

main pipe diameter. Various lengths of the mixing chamber were used by

replacing sections of pipe. Pressure taps were installed along the length

with spacings of one pipe diameter to measure the longitudinal pressure

gradient.

3. Strainer-Diffuser

The fluid for the two jets was withdrawn from a screened element

located downstream of the mixing chamber. The screening desired was

somewhat similar to the commercially available, spirally wound wedge-wire

screens used in well-pump installations. The wedge wires offer some

resistance to clogging, as the gap between wires increases in the direction

of flow through the screen. Thus, any material that passed through the

opening would not clog the screen.

The system adapted for use was a modification to the wedge-wire

screens. Rather than use a spirally wound screen, bars were placed

parallel to the axial direction of flow through the unit.

The screen consisted of 1/8 in. thick by 1/2 in. wide stainless steel

bars. Forty of these bars were uniformly spaced on edge around the circum­

ference of the 3 in. pipe, with an open spacing between bars of 0.113 in.

Such positioning of the bars provided an effect similar to the wedge wires,

in that the gap also increased in the flow direction through the screens.

The strainers were constructed as individual modules with an overall length

of 6.125 in. or two pipe diameters. A special jig was made to insure uni­

form spacing of the bars. The bar assembly was then concentrically cast at

both ends with epoxy resin into a 5.75 in. ID Lucite tube of 6.125 in.

length. The larger tube acted as a reservoir around the bars from which

the fluid was withdrawn through a 1-1/2 in. pipe fitting. The ends of the

17

tubes were machined with O-ring grooves for sealing of adjacent units. A

photo ofa typical strainer module is shown in Fig. 4. The two hose clamps

around the suction fitting were added to further strengthen the connection

to the Lucite tubing.

Adapter fittings were made for installation before and after the

strainer modules. The entire assembly of units was held in place with long

steel rods between the two adapter pieces. A total of nine strainer

modules was made.

Each of the strainer modules was connected to a line leading to a

manifold in the suction line of the main recirculating pump. Valves were

installed in each of the lines for control of the flow withdrawn from each

strainer module. A photo of the suction manifold is shown in Fig. 4. In

this case, four strainer modules are in the jet pump assembly.

The strainers have been referred to as strainer-diffusers, as they

also act as a diffuser during operation. Although the inside diameter is

constant in the direction of the main flow, withdrawal of fluid through the

screens results in a pressure rise due to reduced velocity. The efficiency

of the diffuser action is dependent on the headloss and the method of

withdrawal. A pressure tap was placed at the midpoint of each module to

permit measurement of the axial pressure gradient.

C. Solids Handling

Operation of a hydrotransport flow system which involves high solids

concentration carries the risk of pipe plugging"if the system velocity is

reduced below a critical value. In a closed-loop test system, this means

that the bulk of the solids must be extracted from the flow prior to shut­

down and must be progressively added to the flow upon start-up. This was

initially attempted by adding a by-pass system in a portion of the loop

which was valved to pass any desired pressurized test flow through a hopper

storage system which can either add or extract a solids load from the

system.

The initial system was based on the principle of a horizontal, coni­

cal, hydrocyclone. The cyclone was a large unit capable of processing the

18

, 1

a) Strainer Module

b) Suction Manifold

Fig. 4. Strainer Module and Suction Manifold.

19

full flow of the test loop. The cyclone was designed to remove the bulk of

all solids in a matter of a few minutes, and they were collected in an

accumulation chamber for further use. A wye fitting followed by a rubber

pinch valve was placed in the 3 in. line opposite the jet pump assembly.

This fitting was attached to the tangential inlet to the cyclone. The

clear water from the separator was returned to the main loop just

downstream of the pinch valve. In extracting solids from the flow, the

pinch valve was gradually closed, diverting the flow into the separator.

This procedure continued until all solids were removed and collected in the

accumulation chamber.

To inject solids into the flow, another pinch-valve was installed at

the bottom of the storage chamber and connected by a cross fitting into the

3 in. line. Through appropriate valve operation, the solids would gra­

dually enter the pipeline to attain the desired concentration.

A number of difficulties were experienced in actual operation of the

by-pass system. The injection process was hampered by clogging, and it was

not possible to accurately determine the amount of solids injected. The

efficiency of the removal process was also poor, and was prone to plugging

of the system. After extensive efforts to modify the primary design to a

workable unit, it was decided to completely redesign the mechanism for

injection and removal of solids.

The injection of solids into the model was accomplished through the

use of a device that came to be known as the "torpedo tube," since it

encompassed the fundamentals for the projection of torpedoes. The "tube"

chamber, mounted on the flow circuit at an angle such that the projected

solids would be traveling in the same general direction as the fluid in the

loop, was isolated from the pressurized flowing slurry by an air operated

pinch valve. At the base of the chamber was a drain valve to eliminate

water during the fill phase. Attached to the end of the chamber, not con­

nected to the circuit, was a cap with a quick-disconnect fitting to supply

pressurized water to the chamber. The fill cycle proceeded as follows:

with the fluid in the circuit flowing at the maximum velocity and the pinch

valve closed, the cap of the chamber was removed and a specific amount of

solids was poured into the chamber. When the filling was complete, the cap

20

was restored, a hose supplying pressuri~ed water (at a greater pressure

than the total head of the fluid passing by the pinch valve) was connected

to the quick disconnect, and the chamber was filled and pressuri~ed to the

solid-liquid mixture. The pinch valve was then opened and the solids were

injected into the circuit. The pinch valve was then closed, the chamber

drained down, and the fill cycle was repeated until the concentration of

solids in the system was up to the desired level of the test.

Removal of solids upon completion of the test was much more simple

than the injection technique. A gate valve was placed in the circuit in

order to segregate the system. Downstream of the gate valve, a connection

for a fire hose was supplied, and upstream of the gate valve a large portal

on the bottom of the circuit piping was installed. The firehose supplied

copious amounts of river water to the system , while the portal (referred I

to as the "dump") was positioned over a waste channel that emptied into the

river. Pressuri~ation of the firehose was supplied by the natural static

head available at St. Anthony Falls; the firehose was connected to supply

piping from the upper pool of the falls. Upon completion of a test, large

mesh baskets were placed under the dump, the portal was opened, the segre­

gating valve was closed, and the firehose supplied water and pressure to

push the solids around the circuit and out the dump, where they were caught

in baskets and the water passed through to the waste channel and back to

the river. Stilling areas in the waste channel collected fine solids that

might pass through the baskets.

21

IV. DISCUSSION OF RESULTS

A. Static Pressure Distribution

An extensive series of tests were conducted with water flow to eva­

luate the overall performance of the system, and to optimize the operating

characteristics. These tests were made for many flow rates and tests con­

figurations. Pressure distributions were obtained along the axis of the

pump assembly, including the inlet, mixing chamber, and screens. A typical

pressure distribution is shown in Fig. 5. The ordinate is the pressure

head with respect to the inlet pressure head (see Fig. 3), and the abscissa

is the distance along the pipe axis in pipe diameters. The inlet reference

pressure was measured with a calibrated precision pressure gage. Other

pressures along the longitudinal axis of the assembly were measured dif­

ferentially with respect to this reference pressure with mercury manome­

ters. The locations of the nozzle block, mixing chamber, and strainers are

indicated.

Several typical trends can be noted in the pressure distributions.

The pressure rises after the high velocity flow from the jets enters the

mixing chamber. A maximum pressure is reached followed by a decrease in

pressure. The combined flow of the pipe loop and the injected flow results

in a high velocity in this region, and after mixing has occurred, the

pressure drops due to friction losses. As the high velocity flow enters

the strainers, the pressure continues to drop until the influence of fluid

withdrawn through the screens is sufficient to result in a pressure rise.

The maximum pressure rise occurs at the end of the strainer assembly.

An analysis was first made of the pressure rise in the mixing

chamber. A semi-empirical equation for the pressure loss due to a con­

verging, double branched wye was found in the literature (!). This

equation is

22

20

Nozzle Mixing Chamber strainers

18

16

l'V !tnl >-

w

14 ~ ... I d 1.26"

QS 0.26 cfs

QD 0.87 cfs

PDfI' 126.9 ft 12 I-

PSiI' 47.8 ft

10 o 4 8 12 16 20 24 28 32

Distance from Inlet, Pipe Diameters

Fi~. 5. Typical Static Pressure Gradient for Jet Pump Assembly.

where

Q 2

2A 1+(Q2~ QC c Q \ ( - Al Qs

(l~) 1

P = upstream pressure s

1)2cos a}

P = mixing chamber pressure c

V = average velocity in mixing chamber c

Qs = inlet discharge

Qc = mixing chamber discharge

Q2,Ql = side jet discharge

A = area of mixing chamber c

Al side branch area

'a = angle of side branch to main branch

(1)

For the experimental configuration of the present tests, Q1 = Q2 and

a = 20 degrees. Furthermore, the following definitions will be adapted:

Q = Q + 2Ql cs

Thus,

~ = 1 + ( ..1L) 2 {l __ ----'2=M=-...+.,;,.......=:1 __ -:-:--_ M+l (M+l)(0.7S+0.25 M~l )2

0.94 A ___ c}

-ti Al (2)

A comparison of the experimental data and Eq. (2) is shown in Fig. 6

for the two nozzle diameters tested. It should be noted that ~ is nega­

tive, indicating a pressure rise rather than a pressure loss. The

24

~

UN p.. U I P­

U) >­p..

-7

-6

-5

-4

-3

-2

-1

o o

A ...£ - 5.90 Al -

d := 1.26"

0.1

\ \ \~Eg.2

\~Eq.3 \

\ ~ AC

\ A := 11.46 \ 1

\ d:= 0.904"

\ , \

b , \ , . , , , ,

6 ,

0.2 0.3 0.4

, , ,

0.5

Fig. 6. Dimensionless Pressure Change Through Jet Pump.

25

experimental data fall somewhat below the semi-empirical curves as the

measured pressure rise was not as large as expected.

The semi-empirical equation was modified to better fit the data for

the 1.26 in. diameter nozzle. The resulting modified equation is

and is plotted in Fig. 6 for both nozzle diameters tested. The extrapola­

tion to the smaller nozzle diameter is good.

B. Total Head Ratio and Efficiency

The experimental data were also reduced in terms of a total head

ratio, defined as

H

where Ho ' Hs ' and ~ are the total heads at the strainer exit, the

entrance to the nozzle block, and the side jets, respectively. These total

heads were determined from the pressures measured at the locations shown in

Fig. 3. The efficiency of the jet pump assembly is defined as

H - H OS) x 100

Hn - Ho MH

The experimental data for the two nozzle diameters tested are plotted

in Fig. 7 for the initial jet pump system with a mixing chamber length of

nine diameters and nhle strainer modules. The dimensionless head rise

curve is steeper for the larger nozzle and both curves are essentially

linear with M. Assuming a linear relationship, a best fit line was found

and is drawn through the data. The efficiency curve based on this line is

also plotted. Maximum efficiency for the larger nozzle was about 13 per­

cent, and about 12 percent for the smaller nozzle. Maximum efficiency

occurred at a greater value of M for the smaller nozzle than for the larger

26

----------~~-----------~-------------

0.6 .---------r---------.---------.---------,---------'24

0.5 20

0.4 16

~ .. 001 0

:>. II: II: u I I t:: o ~ 0.3 12

OJ II: tI:l "M U

·M II. tH tH II: f.iI

0.2 8

0.1 4

o ~--------~--------~--------~--------~--------~o o 0.2 0.4 0.6 0.8 1.0

Fig. 7. Total Head Ratio for Two Nozzle Diameters, JPU1

27

nozzle, similar to conventional jet pumps (i), These efficiences were

very low, and effort was directed to improve these values. As the steeper

head rise curve was desired, system modifications were made using the large

diameter nozzles.

It was previously noted that a pressure drop occurred in the mixing

chamber after maximum pressure rise was reached. Therefore, the mixing

chamber was reduced in length in the attempt to recover this loss. As the

strainers were made in modular units, some of the units were also removed.

Although the withdrawal rate through each modular unit was individually

controllable, it was found that the total pressure rise through the

strainers was not greatly affected by just closing several units.

Therefore, additional tests were conducted with some of the modules removed

from the system.

A comparison plot of pressure distributions for several modifications

is shown in Fig. 8 for M = 0.51. The basic dimensions of the assembly are

given below:

Unit Mixing Chamber Length Strainer-Diffuser Length

JPU1 9 pipe diameters 18 pipe diameters

JPU2 3 pipe diameters 14 pipe diameters

JPU3 3 pipe diameters 8 pipe diameters

The mixing chamber length was measured from the intersection of the

center lines of the two side jets to the front flange of the first strainer

module. The overall length of each strainer module was two pipe diameters,

so that nine, seven, and four modules were used.

The pressure rise through the nozzle block is essentially the same for

all three systems, in this case about 4 ft, and reaching a maximum of about

4.5 ft. For JPU1 with the longest mixing chamber, the pressure drop due to

friction is evident. The pressure continues to decrease in the first few

strainer modules and then begins to rise. The data were collected with no

withdrawal through the last two modules.

The mixing chamber was shortened in JPU2 to about three diameters, and

the last two strainer modules were removed from the system. Again the

28

___ I

N \.D

p.., (/).1 d. ?-

10

Nozzle

8

6

4

2

oV 1 0 4

Strainer Modules

_ =x:1: c~~ _ - ~ ~rai~ ~~~ -1 . .... ..... . ... ~ ~.~~~~.:: .. l!?~~ ~.:~ .... ,

JPU3

. . .. . . .. .....................

JPU2

. ' ........... . -"-­,,~ -,,'" _ JPUl

.. ,;.:.: :.:.:. . "-

I 8

.

" /' ,/'

I 12

,.,/'

"

:t 16

//

/'/

d 1.26"

Qs 0.28 cfs

QD 0.55 cfs

----- Original System JPUl

-- Modified System JPU2

......... Modi fled Sys tem JPU3

I I 20 24

Distance from Inlet, Pipe Diameters

Fig. 8. Typical Static Pressure Gradients for Jet Pump Assemblies.

28 32

pressure decreased in the first strainers followed by a rise to a final

value about the same as for JPU1. The first modules were carefully

inspected and no evidence of construction abnormalities could be found.

The units were replaced with others and no significant change in the

pressures was noted. Suction lines leading to the pump suction manifold

were also interchanged with no appreciable effect.

Maintaining the same mixing chamber length, three more strainers were

removed for JPU3. In this case the pressure drop in the first strainer

did not occur, and the total pressure rise through the system was somewhat

higher. Some tests, however, still indicated a slight pressure decrease in

this region.

The data for JPU2 and JPU3 were reduced in terms of dimensionless head

rise and efficiency as shown in Figs. 9 and 10. The head ratio curve of

Fig. 9 for JPU2 was about the same as for JPU1, and thus the efficiency was

also not changed. For JPU3 the dimensionless head was somewhat higher at a

given mass ratio, M , with an increase in efficiency of about 3 to 4 per­

cent. All further tests were conducted with this configuration.

C. Cavitation

The possibility of cavitation exists in the region of the high velo­

city jets if the ambient pressure is sufficiently low. The presence of

cavitation can result in a decrease in performance and extensive erosion

damage to the surrounding pipe material. To determine the effect of cavi­

tation on performance, or more specifically on the total head ratio, a

series of tests were conducted during which the suction or inlet pressure

to the jet pump was gradually reduced. The system pressure was reduced by

lowering the surcharge reservoir as mentioned in Section IlIA. An electro­

nic strobelight was used to visually observe the presence of cavitation.

This procedure was followed using both sizes of nozzles.

The experimental data for which cavitation was observed in the mixing

chamber are plotted in Fig. 11. For small values of M, cavitation was

observed to be the most severe. The static suction pressure was also the

lowest in this region as the discharge valve in the 3 in. line was

30

0.6 24

0.5 20

0.4 16

~ .. WIO :>-J

iIi iIi U I I

12 ~ o C\ 0.3 OJ iIi iIi -.-I

U -.-j

4-l iIi 4-l

fiI

0.2 8

0.1 4

L-________ ~ __________ ~ __________ ~ ________ ~ ________ ~ 0

0.2 0.4 0.6 0.8 1.0

M

Fig. 9. Total Head Ratio vs. Mass Flow Ratio, JPU2.

31

WIO ::q ::q I I o Q

::q II:

II

III

.--------,r--------.---------.--------~--------_,24 0.6

d 0:::

0.5 20

'Ii') .. .. ~

0.4 16

0.3 12

0.2 8

0.1 4

o L-------~--------~--------~--------~--------~o o 0.2 0.8 1.0

Fig. 10. Total Head Ratio vs. Mass Flow Ratio, JPU3.

32

~ .. :>t U s::: Q)

.r-! U

.r-! Il-l Il-l r:LI

0.5 5

QD PS/,y

cfs cfs

0 1.0 3.5 to 17 d = 1.26"

4 0.4 • 0.97 -1 to 18

t>. 0.57 -3.5 to 10

• 0.55 26

01 0.3 3 N

'-.. N

WI 0 on

:Ii :Ii :> I I ~ o 0 :Ii:Ii 0.904" F'!! p..

11 + p:: 0.2 2 W

p.,

0.1 • ~ •• • • • • 1 0 000 • 0 • 6 Il:::. 6 &. ~

8 6 &. 6 &.

o LO----L---~----~----L----L----~--~~--~----~--~O 0.2 0.4 0.6 0.8 1.0

Fig. 11. Effect of Jet Cavitation on Total Head Ratio.

33

throttled. The lines shown are taken from Fig. 7, and represent the best

fit linear line through data collected at high surcharge pressures and

non-cavitating conditions. It can be seen that cavitation did not change

the total head ratio. This may not be unusual, as visual or acoustic evi­

dence of cavitation may be noted in other systems before a decrease in a

performance characteristic can be noted. If the experimental apparatus

permitted a further reduction in static pressure, the effect should cause a

change in the head ratio similar to that found for other jet pumps (1).

The data have also been plotted in terms of the absolute suction

pressure divided by the jet velocity head as a function of M. The scale is

on the right side of Fig. 11- Here P is the measured suction pressure s

at the inlet to the jet pump, PA is the atmospheric pressure, and V. is J

the jet velocity determined from the measured QD . The plotted data

points represent definite cavitation conditions by visual observation. The

data for the smaller nozzle fall slightly lower than for the larger

nozzle; cavitation may have been more severe in this case. The data for

the larger nozzle may better represent the critical conditions.

D. Slurry Flows

The recirculating pipe loop was modified to incorporate the cyclone

solids injector as discussed in Sec. IIIe. The 25 HP Weinman centrifugal

pump was also replaced with a 50 HP Morris solids handling pump. One of

the primary concerns in the proposed type of jet pump was potential

clogging of the strainer screens with solid particles. If the system could

not continuously operate for long periods of time without clogging, the

entire concept would be of marginal value in field operations. Thus, the

main thrust of the first series of tests was to evaluate the strainer

performance.

As the use of coal particles would restrict visibility of the flow

through the strainer modules, a plastic material was used as a substitute

for the coal. The plastic used was Geon 85857 with a specific gravity of

1.47. This material was available in 1/8 in. cube pellets. The material

was considered adequate for first evaluations, as the size was about the

same as the openings between the strainer bars. The volume of the

34

recirculating 100pwas determined, and concentrations of solids by weight

up to about 27 percent were tested.

The tests were conducted with the recirculating loop in a horizontal

plane, and with the JPU3 unit consisting of four strainer modules. At low

concentrations of about 5 percent, and the system operating at maximum

velocity, some minor clogging of the last strainer element was observed.

The solids concentration was increased to about 25 percent. In this case,

severe clogging of the screens was observed. After a period of operation

of about 5 minutes, the system was shut down and the screens observed. It

was found that the last screen was about 90 percent clogged and the first

three about 25 percent clogged with particles.

While the loop was operating, observations were made of the flow in

various parts of the circuit. The slurry flow entering the jet pump unit

was in the form of a sliding bed. It was well mixed by the two side jets,

and the concentration appeared rather uniform at the exit of the strainer

module units. About halfway around the loop, the concentration was higher

at the bottom of the pipe than the top.

It was anticipated that some of the clogging would be prevented if the

mixture contained larger particles which would tear the smaller particles

from the screen by a sliding action. A mixture of particles was made con­

sisting of about 1/3 of the 1/8 in. pellets, 1/3 of a 3/8 in. size, and

about 1/3 of the 1-1/2 in. size. The larger sizes of particles were cut

from plastic sheets in random shapes. In attempting to inject a 21 percent

by weight, the larger solids jammed in the injector. After repeated

attempts, the 1-1/2 in. particles were replaced with a size in the 3/4 to

7/8 in. range. It was possible to inject about half of the desired weight

before the feeder again clogged. However, for the lower concentration

injected, observations indicated that the larger particles did in fact

dislodge the smaller particles from the screen with only slight clogging

present.

The difficulty encountered in the injection of the solids led to the

reconstruction of the feeder as discussed.in Sec. lIIC. The need for a

better system became evident for another reason. As studies continued with

the coal substitute of PVC granules, it was discovered that, due to the

35

repeated shearing that takes place as the solids are jostled and rolled

around in the circuit (which is exceptionally violent as they pass through

the jet and screening of the slurry pump), small "threads" of PVC material

were sheared off of the granules. These threads made it impossible to use

the PVC as a substitute material for coal. Because of their minute size,

they readily passed through the strainers that were meant to separate the

liquid from the waste. Frequently, large clumps of the threads could group

together and render a complete segment of strainers useless. Increasing

the ratio of large diameter solids to small diameter solids improved per­

formance by "knocking" the threads (and the inevitably impinged small

diameter solids) loose from the strainers. It was found that a minimum

ratio of large diameter solids (d>O.5 in.) to small diameter solids (d<O.25

in.) of one-to-one or better in favor of the large diameter material, was

successful in maintaining free flow through and past the strainer modules.

The unsolvable problem with the threads surrounded the finer screens that

protected the cooling water/static head system. A well point protected the

cooling water pressure regulator and the static head tank from con­

tamination from solids such as these threads. However, the size of these

threads were of sizes that would clog the well point, enter the pressure

regulator thereby causing it to malfunction, or pass up into the piping to

the static head tanks, creating clogs. When this would take place, the

static head in the circuit would rise to a level unsafe for the material

used to manufacture the model. Pressure rises occurred so fast at times

that if the operator was not looking directly at the monitoring gauge, the

first indication he would receive of trouble was a ruptured flange. To

improve the safety of the model, an electronic alarm was arrange~ that

sounded a large bell when the system static head was approaching a

dangerous level. When sounded, a quick-acting valve was opened to permit

an immediate blow-down of the system. Safety valves were not used to per­

mit discretion on the part of the operator since a blow-down destroyed any

tests in progress. Several different filtering materials were tried, all

with no success against the long and thin threads. The practice of running

a test until the system broke down, followed by a complete system clean-out

was tried. Eventually, the short time between the injection of the coal

substitute and point where the cooling/static head system was rendered use­

less was not long enough to facilitate a proper test. Facing this problem

36

l

coupled with the time consuming process of disassembling the filters and

regulating valves, the use of PVC particles was discontinued.

A supply of coal from the western regions was obtained. This coal was

seived into size groupings of 3/4, ~/2 and 1/4 inches. A mixture was made

using equal proportions of each size range. Using the new feeder system,

slurries up to various weight concentrations were injected. Here it was

observed that even after a very short period of operation, the larger

particles were completely destroyed by the jet action and all that

remained was a thick sludge. The collected material was about the size of

silt. The maximum concentration by weight that could be moved without

immediate plugging was about 48 percent.

The entire pipe loop was then rearranged in a vertical plane. This

was accomplished with a minimum of new fittings as the basic unit was

designed for either a horizontal or vertical orientation. Again, various

concentrations of the coal slurry were injected, with the same general

negative results. Although visibility was very limited even at low con­

centrations, it was observed that the larger particles moved slower than

the rest of the slurry with flow in the upper direction. This, of course,

is not unexpected. The maximum concentration that could be moved at all

was about 48 percent with 20 percent being the more likely useable value.

The extremely fine particles resulting after only short periods of opera­

tion created extensive problems in attempting to make reliable physical

measurements of the flow characteristics. In view of the poor performance

of the screened diffusers, the experimental efforts were terminated.

37

i

V. ACKNOWLEDGEMENT

The concept of the jet jump was developed by John F. Ripken, Professor

Emeritus. He was responsible for the design of the experimental facility,

and his general advice is gratefully acknowledged.

38

REFERENCES

1. Bain, A. G. and Bannington, S. T. The Hydraulic Transport of

Solids by Pipeline. Pergamon Press, 1970.

2. Wakefield, A. W. "Practical Solids Handling Jet Pumps."

Symposium on Jet Pumps and Ejectors, London, 1972.

3. Idel'chik, I. E. "Handbook of Hydraulic Resistance, Coefficients

of Local Resistance and of Friction." AEC-tr-6630, U.S. Dept. of Commerce,

1966.

4. Bonnington, S. T. "Jet Pumps." BHRA Publication S.P. 529,

British Hydrodynamics Research Assn., Harlow, Essex, England, 1956.

5. Bonnington, S. T. "The Cavitation Limits of a Liquid/Liquid

Jet Pump." BHRA Publication RR60S, British Hydrodynamics Research Assn.,

Harlow, Essex, England, 1958.

39