jamarat final report 11 - ksu

Jamarat Bridge Project Report Crowd Dynamics Limited

March 2003 (1/1424H) www.CrowdDynamics.com - 1 -

Jamarat Bridge Mathematical models, Computer Simulation and Hajjis Safety analysis. Crowd Dynamics Limited. March 2003 (1/1424H)

Ministry of Public Works and Housing. Saudi Arabia Authors: Professor Dr. Saad A. AlGadhi Dr. G. Keith Still Version 1.10 2003-04-04



Table of contents Executive Summary................................................................- 4 -

1.0 Objectives of the project..............................................- 4 - 2.0 Methods used in the study. ...........................................- 4 -

2.1 Sample Outputs .......................................................- 4 - 2.2 A generic model of the Jamarat Bridge and Rajm process. .- 5 -

2.2.1 Sufficient Ingress (entry) capacity. .........................- 5 - 2.2.2 Sufficient Throwing area. .....................................- 5 - 2.2.4 Sufficient Passing Area.........................................- 6 -

2.3 Model of the Design Criteria for Jamarat Bridge...............- 6 - 3.0 Testing the Jamarat Bridge (MPW&H proposal) capacity .......- 6 -

3.1 Testing for optimal shapes for the Rajm process. .............- 6 - 4.0 Results, Floor tests. ....................................................- 7 -

4.1 Level 4 ..................................................................- 7 - 4.1.1 Ingress .............................................................- 7 - 4.1.2 Circulation........................................................- 7 - 4.1.3 Egress..............................................................- 7 -

4.2 Level 3 ..................................................................- 7 - 4.2.1 Ingress. ............................................................- 8 - 4.2.1 Circulation........................................................- 8 - 4.2.2 Egress..............................................................- 8 -

4.3 Level 2. .................................................................- 8 - 4.3.1 Ingress. ............................................................- 8 - 4.3.2 Circulation........................................................- 8 - 4.3.3 Egress..............................................................- 8 -

4.4 Level 1. .................................................................- 9 - 4.4.1 Ingress. ............................................................- 9 - 4.4.2 Circulation........................................................- 9 - 4.4.3 Egress..............................................................- 9 -

4.5 Ground Floor (Level 0). .............................................- 9 - 5.0 Conclusion (Summary)................................................ - 10 - 6.0 Recommendations. (Summary) ..................................... - 10 -

Chapter 1 .......................................................................... - 11 - 1.0 Background ............................................................. - 11 - 1.1 Dr. G. Keith Still ....................................................... - 11 - 1.2 The concept of a model.............................................. - 11 - 1.4 How to build a good model. ......................................... - 13 - 1.5 Modelling risk in crowded situations .............................. - 13 - 1.6 “Too fast for us mere mortals” .................................... - 13 - 1.7 Modelling exposure to risk in crowded situations .............. - 14 - 1.8 The Highway Capacity Manual ...................................... - 14 -

1.8.1 Level of Service .................................................. - 14 - 1.8.2 Crowd Density .................................................... - 15 - 1.8.3 Density – Distribution ........................................... - 16 - 1.8.4 Analysis of sad accidents ....................................... - 17 - 1.8.5 Existing Design and location of the Sad Accidents........ - 17 - 1.8.6 Crowd Management ............................................. - 18 -

Chapter 2 .......................................................................... - 19 - 2.0 Methods used in the analysis ....................................... - 19 -



2.1.1 CAD plan of the area of interest.............................. - 19 - 2.1.2 Export/Import to Myriad ....................................... - 19 - 2.1.4 The trails model within Myriad ............................... - 20 -

2.5 Network Analysis ...................................................... - 22 - 2.6 Optimal Design Criteria. ............................................. - 23 -

Chapter 3 .......................................................................... - 23 - 3.0 Testing the Jamarat Bridge capacity .............................. - 23 - 3.1 Optimal Jamarah Perimeter ........................................ - 24 -

3.1.1 Optimal Shape for the Jamarah perimeter................. - 25 - 3.1.2 Circle. .............................................................. - 25 - 3.1.3 Evolution of the high density.................................. - 27 - 3.1.4 Ellipse. ............................................................. - 30 -

3.3 Test conditions ........................................................ - 31 - 3.4 Emergency Evacuation ............................................... - 31 - 3.5 Analysis results for the towers ..................................... - 32 - 3.6 Egress distribution .................................................... - 33 - 3.7 Further egress tests .................................................. - 33 - 3.8 Screen shot from the Egress Tool (Myriad/Simulex) ........... - 34 -

Chapter 4. ......................................................................... - 34 - 4.0 Level by level analysis against optimal design criteria........ - 34 - 4.2 Level 3................................................................... - 37 - 4.2.1 Escalators ............................................................ - 37 -

4.2.2 Al Haram ........................................................... - 37 - 4.2.3 Escalator Case Study 1.......................................... - 37 - 4.2.4 Escalator Case Study 2.......................................... - 38 - 4.2.5 Escalator Case Study 3.......................................... - 38 - 4.2.6 Proposed changes to the design .............................. - 39 -

4.3 Level 2................................................................... - 40 - 4.4 Level 1................................................................... - 41 - 4.5 Ground Level (Level 0)............................................... - 43 -

Chapter 5. ......................................................................... - 45 - 5.0 Conclusions ............................................................. - 45 -

Chapter 6. ......................................................................... - 47 - 6.0 Recommendations..................................................... - 47 - Appendix - Project Overview ............................................... - 48 -

Objectives ................................................................. - 48 - Outcomes .................................................................. - 48 - Program of Work ......................................................... - 48 - Assessment includes scenarios for ................................... - 48 - Area to be assessed...................................................... - 48 - Assessment includes..................................................... - 48 - Outputs..................................................................... - 48 - Deliverables ............................................................... - 48 - Further studies ........................................................... - 48 -

Shape tests. ....................................................................... - 49 - Results for Circle. ............................................................. - 50 - Ellipses........................................................................... - 51 - Deformed Circle ............................................................... - 52 - Maximum Throughput ........................................................ - 53 -



Jamarat Bridge – Mathematical Models, Computer Simulations and Hajjis Safety analysis. Authors: Professor Dr. Saad AlGadhi, Dr. G. Keith Still Executive Summary Due to the repeated sad accidents in the Jamarat area, where pilgrims perish during overcrowding periods in the area, the Government of the Kingdom of Saudi Arabia, represented by the Ministry of Public Works and Housing (MPW&H) in cooperation with the Custodian of the two Holy Mosques Institute for Hajj Research, has proposed redesigning the Jamarat area, by replacing the existing bridge with a 5-level structure to ease the process of performing this ritual. A conceptual design of the proposed MPW&H bridge is now completed. In order to ensure that the proposed design satisfies the required criteria of pilgrims safety, especially during overcrowding, a technical committee was setup by the authorities to review the design, which appointed Professor Dr. Saad A. AlGadhi, to oversee that an appropriate crowd simulation model be used to test the different elements of the design, and use it as a tool to evaluate and propose, as required, modifications to this conceptual design. Due to the specialized nature of this crowding event Dr. G. Keith Still, the crowd dynamics expert, was appointed by Professor AlGadhi to comprehensively model this design using his suite of crowd dynamic software tools. This report outlines their findings. 1.0 Objectives of the project The situation at the Jamarat Bridge is highly dangerous due to the vast number of Hajjis trying to stone the pillars in a relatively short period of time. Many sad accidents have occurred over the years and the area is in urgent need of development. The MPW&H proposed design offers a significant improvement to the safety of the Hajjis, replacing the one tier structure with a 4 tier structure (5 levels). The objective of modelling the MPW&H Jamarat Bridge design was to identify if any elements in the proposal may give rise to overcrowding. The goal was to determine the capacity, throughput and performance of the proposed design. The specific aim for the project is Safety first – no compromises.

2.0 Methods used in the study. Dr. Still (whose PhD is “Crowd Dynamics”) has been developing crowd models for the last 12 years. Two of his previous modelling tools (VEgAS and Legion) were award winning crowd safety/modelling tools used of such projects as the Sydney Olympics, London Underground, the Hong Kong Jockey Club and evacuation planning for Canary Wharf (London financial district). His current suite of tools, Myriad, developed specifically for modelling and testing architectural plans for crowd safety were applied to the MPW&H design to test for Hajjis safety during the stoning (Rajm) process. CAD plans were provided of the MPW&H proposal and each level was modelled. The modelling tools highlights any problem areas, this information was then discussed with the architects (dar al-handasah) and the process of testing, discussion, improvement/modification, retesting, further discussion… was repeated until ALL safety concerns highlighted by the modelling process were eliminated. Working closely with the design team, Dr’s AlGadhi and Still were able to identify all areas of concern and address these quickly to a satisfactory conclusion. 2.1 Sample Outputs The graphs below show a crowd density and individual Hajjis speeds around the immediate area of the Jamarah ring (red > 8 Hajjis per square metre). The graph on the left shows density and the graph on the right: the Hajjis speeds.



These graphs highlight the sustained high density that the Hajjis are presently exposed to during the Rajm process. We can also observe that faster movement is possible around the edges of the high density areas (illustrated by the red edges in the graph – above right). 2.2 A generic model of the Jamarat Bridge and Rajm process. To establish the design criteria for the Jamarat Bridge a generic model was first created. This sets the criteria against which the MPW&H proposal could be evaluated. The target was to evaluate if 125,000 Hajjis per hour could be accommodated, which represents the peak demand during the Rajm period (5.5 hours from noon until sunset on the 11th and 12th days of Dhu Al Hajjah). Five tests were applied to the system:

2.2.1 Sufficient Ingress (entry) capacity.

Each level was tested to establish if there was sufficient width to allow the desired number of Hajjis to enter each level and proceed to the Jamarah.

2.2.2 Sufficient Throwing area.

Each Jamarah has a perimeter (ellipse) which will allow a number of Hajjis per metre to perform the Rajm process. The terms of reference stipulated 12m as a maximum distance at which the Hajjis can throw their stones at the Jamarah. As the angle of approach determines the nature of the queueing to peform Rajm it is important to test each Jamarat in situ to establish that sufficient throwing perimeter and profile (angle of approach or effective width) is available to prevent overcrowding around the Jamarat. Dr’s AlGadhi and Still established a x3 factor such that for every metre of ingress width 3 metres of Jamarat perimeter should be available to ensure crowding around the Jamarat is minimised.

2.2.3 Sufficient Space (Density ≤ 4 Hajjis per square metre)

There should be no areas in the system where the crowd density would exceed a safety limit of 4 Hajjis per square metre. This is an internationally recognised safety standard. The present density of Hajjis in the existing design is 8-9 Hajjis per square metre. It should be noted that in areas such as entrances, exits and the area around the Jamarah the Hajjis may experience high density – but for short durations.



2.2.4 Sufficient Passing Area

Any areas where Hajjis performing Dua’a prayer must have sufficient space to allow for prayer and allow for other Hajjis to pass with ease. Passing areas also include tops of escalators where groups will reform after ascent and at exit ramps.

2.2.5 Sufficient Departure Capacity (Egress)

Egress (leaving the area) routes require two main considerations: Normal and Emergency. This is a requirement to facilitate the unimpeded movement of Hajjis off each level under normal (Rajm) and emergency situations (medical emergency for example). Emergency situations (such as a heart attack, heatstroke or other such medical emergency) will require the rapid response of the emergency services. The design has many safety exits to allow for rapid response in the event of an emergency.

2.3 Model of the Design Criteria for Jamarat Bridge An icon for the model of the Jamarat Bridge was created to illustrate the above 5 design criteria. For each Jamarah a similar 1,2,3,4,5 test was applied.

We established, through measurements and simulation, that an optimal 3 metres of Jamarah throwing perimeter should be allowed for every metre of ingress width. We should provide more Jamarah perimeter (where possible) by using a wall around each Jamarah, profiled to the approach direction. 3.0 Testing the Jamarat Bridge (MPW&H proposal) capacity 3.1 Testing for optimal shapes for the Rajm process.

Circle. The existing design has a 16 metre diameter and offers 50 metres of perimeter. The approach is 40-50 metres (via barriers used to control the Hajjis arrival rate). When the arrival (ingress) width exceed the x3 factor a queue of Hajjis will develop at an exponential rate. The perimeter around a circular Jamarah increases at a ratio of 6 metres for every 1 metre of radius. The geometry of a circle therefore does not lend itself for optimisation.

Ellipse. The ellipse has two axis (major and minor) which allow for a greater flexibility to

increase Jamarah perimeter along the major axis without increasing the width and reducing the passing area. Many elliptical shapes were tested and a scientific tool to test shapes was developed, source code was provide to the Custodian of the Two Holy Mosques Institute for Hajj Research and the MPW&H along with ½ day training in their use. As each level of the Jamarah has different approach directions (to Jamarah Al Sughrah) and due to the position and size of the well on the ground level (level 0) an ellipse was chosen to ensure that there was sufficient Jamarah perimeter



to allow the Hajjis to perform Ramj without the difficulties currently experienced in the existing structure.

Deformed Circle. Several shapes were tested but they all failed to outperform the elliptical

shape. 4.0 Results, Floor tests. As stated above the optimal geometry for providing out x3 criteria is an ellipse. The axis of this should allow for sufficient perimeter to match the arrival width. We shall now outline each floor before concluding the optimal Jamarah perimeter. There are three elements for the Rajm process at every level: Ingress (getting to the Jamarah), Capacity (sufficient space, low density, throwing and passing area) and Egress. We applied our design criteria to these three elements on every level. 4.1 Level 4 4.1.1 Ingress This level serves the Hill of Hadarim. It replaces the 30 metre descent via the stairs and is a significant improvement to Hajjis safety and crowd congestion. This level has the facility to serve latent demand (increase) of the Hajjis as it can support an integrated transportation solution. As our requirement is to facilitate 3,000,000 Hajjis in 5.5 hours we can use the flexibility of a Jamarah perimeter to match this ingress width. The Jamarah on this level should therefore be greater than 108 metres in an elliptical configuration. The use of barriers directing the Hajjis on the approach to Jamarah Al Sughrah, is easily configured to create an orderly and controlled approach. The exact location and configuration of these approach barriers will change as the design is finalised without altering the geometry of this level. 4.1.2 Circulation The area around each Jamarah, between the Jamarah’s and on ingress/egress passes criteria 4 (sufficient passing space ie: no part of the system has a reduction in width relative to the ingress width. 4.1.3 Egress There is sufficient space for Rajm and passing without increase in Hajjis density or disruption to the Dua’a. Emergency egress is provided by the exit system and 5 towers. 4.2 Level 3 This level is accessed by escalators. After careful consideration and extensive study of the Al Haram use of escalators we worked with the design team to introduce key elements to ensure Hajjis safety. These included reducing the approach width to prevent dangerous congestion around the base of the escalators. Areas for helpers (to assist Hajjis boarding and alighting each escalator) were provided and two additional emergency stairs were provided to allow for emergency situations on each level. A reduction in the number of 180 degree turns (from 6 to 2) were recommended and accepted. A ramp from Mina is recommended to reduce the risk of Hajjis and traffic mixing (at great risk) on King Fahad Road.



4.2.1 Ingress. With the 10m ramps from the Mina area, crossing King Fahad Road, and the escalators operating at 3,500 Hajjis per hour (per escalator) the ingress capacity of this level meets the design criteria and passes the required design and safety considerations after consultation and several design changes. The following conditions for level 3 are:

The escalators must be managed at all times (assisted boarding and alighting at each escalator).

Screening Hajjis (far from the escalator entrances) for luggage/rubbish to prevent any injury

on the escalators

The Mina ramp (over King Fahad Road) be included to reduce risk of injury in the Mina approach.

Further investigation for development of exit ramps for this level is also recommended.

4.2.1 Circulation. The design is similar on every level and sufficient space is provided for Dua’a and stoning. 4.2.2 Egress. Using the egress towers described above the egress capacity is sufficient to clear the areas required for emergency situations. 4.3 Level 2. Level 2 has been designed to service those Hajjis approaching from the Makkah direction. It has two ingress ramps of 20 metres width. A Jamarah perimeter of 120 metres would allow this level to operate effectively. 4.3.1 Ingress. Some crowd management/control will be required on the 12th day as this ramp has a 1km length. Hajjis may queue on this level but an operational procedure will be required. 4.3.2 Circulation. As the Hajjis approach along the length of the Jamarah Bridge a good perspective and orientation will facilitate the natural dynamics of the Hajjis during the Ramj process. Sufficient area for Rajm and Dua’a allowing for is within the design scope. 4.3.3 Egress. Egress is via 12 minor egress towers (that serve the ingress and egress ramps), 5 major egress towers (which serve every level) and the main egress ramp. It should be noted that the evacuation procedure would anticipate egress when the ingress ramps are fully loaded and therefore the minor egress towers are the means of escape/emergency access. Their location needs to be assessed as the detailed design develops.



4.4 Level 1. This level serves the Mina area (approach from Arafat/Muzdalifa). 4.4.1 Ingress. Access to this level is via 2x20m ramps which extend to the Mina area. These ramps should be as short as possible (while maintaining a slope less that 8.3% - which is the recommended safety limit for a slope) to allow for crowd management and even distribution across both ramps (to reduce congestion and balance the Hajji flow on the approach to the Jamarah Al Sughrah). 4.4.2 Circulation. The Jamarah perimeter may be limited to ensure the stones fall in the well below. This would restrict the safe capacity on this level and it is recommended that the approach ramps should be reduced to re-balance the system and prevent queueing. It is recommended that a discussion with religious scholars regarding the ground floor level and size of the well be agreed before finalising this perimeter/approach width. 4.4.3 Egress. This level has 3 egress ramps and 5 egress towers. 4.5 Ground Floor (Level 0). A discussion with the religious scholars is recommended. Two issues arose during the project to evaluate the design. Could the size of the perimeter be increased on the ground floor (various proposals would work) to increase the throughput on this level and improve Hajjis safety? As this would allow the increase of perimeter throughout the system it is a recommendation that a discussion with the religious scholars be sought on this issue. The throughput is governed by the Jamarah perimeter. Currently this is 74,000 Hajjis per hour and would not change unless the Jamarah perimeter could be increased. Crowd management is required on the ground level as it will alter the crowd dynamics.



5.0 Conclusion (Summary) The current situation at Jamarat Bridge is life threatening. The safety of the Hajjis is the primary concern in the design and management of the MPW&H & HRI proposal.

It should be highlighted that the design is conceptual but, after several sessions with the architect, fulfils the design criteria and simulation testing.

It is our recommendation that the Ground Floor should be discussed with the religious

scholars and the result of that discussion would allow an increase in the Jamarah perimeter further increasing the capacity of the system.

The lower throughput (based on the Jamarah size in the plans provided, crowd dynamics,

safety considerations and crowd management issues) has been determined as sufficient to accommodate the desired capacity. The upper limit of the MPW&H Jamarat Bridge proposal is approx. 3,900,000 Hajjis in 5.5 hours which would depend on increasing the Jamarah perimeter at level 0.

6.0 Recommendations. (Summary) The successful operation of the bridge requires integration of BOTH design and operational management. We have ensured that all possible safety measures, emergency evacuation and medical facilities are integrated in the design. It is recommended that an operations manual/procedure be developed during the detailed design phase.

A ramp to service the 3rd level and to alleviate potentially serious crowding in the Mina approach to the Jamarat Bridge was also agreed as an important element to the design.

For Level 0 (ground floor) a discussion with the religious scholars is recommended. We are

highlighting the potential for overcrowding with the existing diameter of the Jamarah. If it is possible to increase the perimeter it would significantly increase the safety of the Hajjis.

It is our recommendation that an integrated modelling and crowd dynamics plan be continuously

improved as design modifications (during the detailed planning and construction phases) are included.

It is recommended that an operational procedure and manual be developed during the design phase

working with Civil Defence and Public Security.

It is recommended that a survey of the Mina Valley consisting of questionnaires and physical measurements is required for the next Hajj, the first year of operation and the second year of operation to establish the performance of the bridge and to assess the potential impact on the crowd and traffic dynamics in the Jamarat Bridge area.

We recommend that during the detailed design phase that further tests and computer simulations

are performed. This should be a continuous process during the development and installation of the Jamarat Bridge.

We confirm that the design passes our safety considerations.



Chapter 1 1.0 Background 1.1 Dr. G. Keith Still Dr. G. Keith Still has spent the last 15 years developing methodologies and simulation techniques in his chosen field of Crowd Dynamics. The goal of his research is to improve all aspects of safety through the use of a computer simulation and appropriate methodologies. To achieve this he had to understand the problems associated with computer modelling of both the dynamics and the behaviour of people and crowds during normal and emergency egress. The objective he set for himself was to determine the critical factors involved in crowd dynamics and emergency egress in places of public assembly. Specifically, the objective was to develop computer simulation to present users with a flexible “what-if” tool to experiment with, test and educate themselves in the nature and problems associated with crowd. His two previous tools were award winners in the UK. VEgAS (Virtual Egress Analysis and Simulation) a finalist in the 1992 Sunday Times Innovation of the Year competition and his tool for large scale crowd modelling, Legion, was a named Millennium Product by the UK Prime Minister, Tony Blair. Dr. Still is a regular visiting lecturer at Easingwold the UK Cabinet Office Emergency Planning College where he runs three- day Crowd Dynamics workshops. Dr. Still was a crowd consultant to the Sydney Olympics where he presented crowd scenarios to the Minister for the Games, Michael Knight. He has created models and simulations for many international projects including Railtrack, CrossRail, Hong Kong Jockey Club, London Underground, and the Olympics Games. He now works with the Symonds Group (www.symonds-group.com), an international consultancy with over 800 staff offering a wide range of services in engineering, transportation, environment, project management and Health and Safety. Crowd Dynamics Limited also has strategic partnerships with Ove Arup (Australia) and MVA (Hong Kong). 1.2 The concept of a model. Many applications of science make use of models. The term "model" is typically used for a structure which has been built purposely to exhibit features and characteristics of some other object. Generally only some of these features and characteristics will be retained in the model depending upon the use to which it will be put. Sometimes such models are concrete, as is a model aircraft in a wind tunnel experiment. More often models are abstract or mathematical in nature. There are a number of motives for building a model:

The exercise of building a model often reveals relationships which were not apparent to many people. As a result a greater understanding is achieved of the object being modelled.

Having built a model it is usually possible to analyse it mathematically to help suggest courses

which might not otherwise be apparent.

Experimentation is possible with a model whereas it is often not possible or desirable to experiment with the object being modelled. It would clearly be politically difficult, as well as undesirable, to experiment with large numbers of people moving in and out of a stadium if there was a probability of any individual being hurt.

It is important to realize that a model is really defined by the relationships which it

incorporates. These relationships are, to an extent, independent of the data in the model. Thus the speed/density/geometry relationship is independent of the data (the specific geometry being modelled).



1.3 Misconceptions about the value of a model Many misconceptions exist about the value of a model, particularly when used for planning purposes. At one extreme there are people who deny that models have any value at all when put to such purposes. Their criticisms are often based on the impossibility of satisfactorily quantifying much of the required data. Eg: attaching a behavioural pattern to a specific individual. A less severe criticism surrounds the lack of precision of much of the data which may go into a model. Eg: if there is any doubt surrounding the coefficients in a model, how can we have any confidence in an answer it produces? The first of these criticisms is a difficult one to counter and has been tackled at much greater length by many defenders of cost-benefit analysis tools. It seems undeniable, however, that many decisions concerning unquantifiable concepts, however they are made, involve an implicit quantification which cannot be avoided. Making such quantification explicit by incorporating it in a model seems more honest as well as scientific. Thus the process of validation, explanation, demonstration and calibration of a model is an on-going concern. The more times the model is applied and produces quantifiable and correct outputs the more confidence one can have it its performance. However, this could never justify the claim of an absolute proof of the models validity. It is a scientific tool bounded by the constraints of its applicability to a scenario that would justify the motives of model building. The adage of "Garbage-in, Garbage-out" is a serious concern when building model specifically one relating to Hajjis safety. Appropriate training in both the choice of input parameters to a model and the interpretation of the outputs (results) of the model is an essential aspect of model building and should not be underestimated. The second criticism concerning accuracy of the data should be considered in relation to each specific model. Although many coefficients in a model may be inaccurate or unknown it is still possible that the structure of the model results in little inaccuracy in the solution. This is known as a parametric process. With crowd modelling we are concerned with a series of invariants which produce measure-preserving maps. The dynamic behaviour of the entities produce equivalence maps (space utilisation, flow maps, density maps) which are independent of the process used to produce them. For example, fluxing every entities speed produces the same space utilisation and density maps as a non-fluxed entity. The important point to note here is the time it takes to produce a non-fluxed map to a fluxed one. Same invariant, but developed during a much faster process using the fluxing technique. An example of this process is illustrated when we model such things as queues. We are concerned with such factors as size of queue, typical waiting time and the optimum facilities required to maximise service/revenue. If we set a target of 5 minutes for service then, so long as Bill and Joe are both served within 5 minutes, we are not concerned whether Bill gets served before Joe (or vice versa). In this type of model the order of service is not important but the overall time it takes to get served IS important. In parametric processes we are concerned with the limits of a system, such as the limits to density, flow, queuing times, constraints etc. These are best broken into separable models such that we test a system (for example, the Jamarat Bridge) for a capacity given a range of inputs. The skill in building a model is in choosing the appropriate elements of the model. At the opposite extreme to the people who utter the above criticisms are those who place an almost metaphysical faith in a model for decision making (particularly as it involves a computer). The quality of the answers the models produce obviously depends on the accuracy of the structure and data of the model builder. The definition of the project objective clearly affects the answer as well. Uncritical faith in a model is obviously unwarranted and dangerous. Such an attitude results from a total misconception of how a model should be used. To accept the first answer produced by a model without further analysis and questioning should be very rare. A model should be used as one of a number of tools for decision making. The answer such a model produces should be subject to close scrutiny. If it represents an unacceptable operating plan then



the reasons for unacceptability should be spelled out and if possible incorporated into a modified model. Should the answer from the model be acceptable, it might be wise to regard it only as one of several possible options/outputs. The specification of different input parameters would obviously create a different option. 1.4 How to build a good model. The aims of the model builder should be to construct a model which is easy to understand, easy to detect errors in the process of building a model and easy to compute a solution. To do this one has to spend more time in the analysis of the client requirements than the physical process of building the model. The process would be to break the model into specific sections that would impose constraints on the system. It is not necessary to build a complex model of all of the floors in all of their configurations at the initial pass. Indeed such a model may be impossible to build (computational intractability) and, more importantly, impossible to understand (analytically intractable). So we look to model those sections where the problems are apparent, say the top/bottom of escalators, in order to understand how this specific element affects the rest of the system. You need to break the process into logical, easy to build, easy to test, easy to understand sections. Often it is the process of the analysis of the problem that leads to a simpler model building process and, as a result, it is easier to understand the results. The skill is very much in the expertise of the model builder and not in the model itself. 1.5 Modelling risk in crowded situations Suppose we have N entities, with entity i at position (xi, yi) inside a region R of the plane R2, representing the accessible parts of the building. Each entity's path Pi through the building is constrained: first, by the entity's speed distribution, and secondly by the requirement that entity i visits certain places or sub-regions of the building in some order. Call the set of ALL these constraints on entity i's path Ci. There are also non-collision constraints Kij which assert that entities i and j cannot occupy the same position at the same time, for i not equal to j. There is a 'cost' function u(Pi) - for example, length, total time, effort... choice of these parameters lies in the skills of the model builder. For a set P of paths Pi satisfying constraints Ci and Kij, there is a total cost U(P) = u(P1)+...+u(Pn). The problem is to minimize U(P) subject to those constraints, thereby finding the set of paths (flow pattern of crowd) that requires the least total effort. You solve this optimization problem by a type of simulated annealing, iteratively starting from a path P, randomly varying it, seeing if the cost goes down, and if so choosing the cheaper path; then repeat. Stop when you can't improve solution. 1.6 “Too fast for us mere mortals” That quote came from a colleague of mine a long time ago and it made me stop and think - why do we use jargon (because that's the language of the scientist). But it is counter to the purpose of building a model - easy to build, easy to test, easy to understand. What use is the model if you cannot understand the various elements in the proposed design? The previous section is the technical, scientific description of a crowd model. Here we see a perfect example of how lack of "transparency" can befuddle a situation. What does it really mean – in plain English?



It means - model dumb people doing dumb things that could lead to personal injury or group harm. If the passive design prevents dumb people doing dumb things causing each other harm then we are testing the latent risk or exposure to risk in the system. Minimise this risk exposure and you are, by definition, making the environment safer. Over the years there have been many discussions as to whether a "dumb" person in the model is a true proxy for a participant in a crowd. Everyone related to the industry is satisfied with the assumption - people are not that smart in a large crowd. 1.7 Modelling exposure to risk in crowded situations We are really modelling the exposure to high-density and high-speed environments. The space takes on a form of an "invariant" ie: there are places in the environment that are typically high density (say street corners). These may be due to many perceived factors - but once you break a model of a street into its key components you begin to understand that some of these factors are irrelevant - corners are places where decisions are made - choice of route/stop/start. They become a "delay" in the system and delays cause congestion. This requires specific attention whereas other area may require much less attention as they are typically empty (for example, a recess in a wall). High density, high speed exposure is a major factor in assessing the dynamics of risk. The Green Guide indicates 40 people per square metre and flow rates of 109 people per metre per minute. But does NOT state how long an individual is expected to experience this high density environment. The purpose of model of the dynamics of crowds is to assess this exposure, not just for any one individual - but to the crowd in general. 1.8 The Highway Capacity Manual The Highway Capacity Manual states that the maximum rate of crowd movement is relative to the density as the following table illustrates. 1.8.1 Level of Service Paraphrased from Dr. J. J. Fruin’s “Pedestrian Planning and Design The Level of Service concept was first developed in the field of traffic engineering in recognition of the fact that capacity design was, in effect, resulting in planned congestion. The Highway capacity manual, the most authoritative reference on highway design practice, develops standards for six levels of design, based on service volume and a qualitative evaluation of convenience. Level of service standards provide a designer with a useful means of determining the environmental quality of a pedestrian space, but they are no substitute for judgement. The designer MUST examine all elements of pedway design, including such surging or platooning caused by arrivals and the economic ramifications of space utilization. Caution must be exercised in selecting design standards near maximum capacity levels, since the critical pedestrian at these levels is likely to be exceeded intermittently. When the designer is required to use a maximum capacity volume, such as Jamarat Bridge, he must examine the adequacy of the holding areas at the approaches to the critical section. In such situations, pedestrian waiting and system clearance times should form the basis for the qualitative evaluations of the design.



Recommended HCM walkway Level of Service criteria.

LOS Space Flow Rate Average Speed v/c ratio

(m2/ped) (ped/min/m) (m/s)

A > 5.6 < 16 > 1.3 0.21

B 3.7 to 5.6 16 to 23 1.27 to 1.30 0.21 to 0.31

C 2.2 to 3.7 23 to 33 1.22 to 1.27 0.31 to 0.44

D 1.4 to 2.2 33 to 49 1.14 to 1.22 0.44 to 0.65

E 0.75 to 1.4 49 to 75 0.75 to 1.14 0.65 to 1.0

F <0.75 var. < 0.75 var.

We can see from the above table that the maximum flow rate the system should encounter during normal operating use will not exceed 75 Hajjis per metre per minute. 1.8.2 Crowd Density As the following graphics illustrate, the criteria for the Jamarat Bridge is a density not exceeding 4 Hajjis per square metre.

2 Hajjis per square metre 3 Hajjis per square metre 4 Hajjis per square metre To put these graphics into perspective the current density at the Jamarat Bridge exceeds these limits by 100% (as the photographs below illustrates).

It should be noted that at these crowd densities, and in extreme heat, the risk of a collapse is significant. The enormous pressure that can be exerted on the surrounding structures is capable of bending steel as the following photographs show.



We can see from the photographs above that the force of hundreds of Hajjis is enough to cause the collapse of this steel fencing. Clearly the potential for this type of situation needs to be identified and eliminated. Analysis of a wide range of crowd disasters from around the world have some common elements such as sustained high density and crowd compression in and around complex spaces.

1.8.3 Density – Distribution The crowd density around the Jamarat Bridge is not evenly distributed. Some areas will experience very high density for limited periods as they approach the Jamarah and these areas required specific high density modelling techniques. Myriad (one of the tools used in this analysis) highlights areas of high density (high risk of injury).

Many environments (such as stadia, transport terminals, trains and theatres) experience high density during peak periods. However these are short duration and typically no greater than 4-5 people per square metre. The extreme high density surrounding the Jamarat Bridge area and during the Rajm process is a significant risk. The purpose of this study is to establish where the potential for extreme high density could exist on the proposed design and identify, reduce or eliminate them.

8-9 Hajjis per square metre – showing how little space there is for movement



1.8.4 Analysis of sad accidents Due to lack of understanding of crowd dynamics

1989 - 96 dead, 400 injured - Hillsborough, Great Britain 1996 - 83 crushed, 180 injured - Guatemala City 1999 - 51 killed, 150 injured in stampede - Kerala, India 1999 - 53 killed, 190 injured in stampede - Minsk, Belarus 2000 - 12 killed dozens injured - Harare, South Africa 2001 - 126 killed in crowd stampede - Ghana, West Africa 2001 - 7 children died, crowd trampling - Sofia, Bulgaria 2002 - 10 trampled, Mall Crowd Craze - Yokohama, Japan

And specifically around the Jamarat Bridge

1417H (1997) – North part of the eastern entrance – 24 killed. 1418H (1998) – North part of the eastern entrance – 118 killed. 1421H (2001) – North part of Great Jamarah – 21 killed. 1423H (2003) – North part of the eastern entrance – no one killed.

142 Hajjis have lost their lives at the North part of the Eastern Entrance of the Jamarat Bridge and we need to understand the nature of these sad accidents to avoid any potential risks in the new design. 1.8.5 Existing Design and location of the Sad Accidents

CAD plan of the eastern entrance of Jamarat Bridge

Site schematic of the approach routes to the eastern entrance.

This area has a high convergence (multiple routes leading into a complex geometry. This configuration will produce a high density crowd as the following photograph highlights.



The eastern entrance of the Jamarat Bridge, the bridge is to the right of the photograph above. We can see the convergence of the Hajjis in this area and note that there are patches of very high density but a few metres away there are large areas of low density. At some stage a series of barriers have been added to this entrance to reduce the flow of Hajjis towards the Jamarah. This has the effect of accelerating the convergence rates. The security forces/civil defence are aware of this convergence and the photograph below shows their contingency to prevent further sad accidents.

1.8.6 Crowd Management Elements of design that could lead to sad accidents can be managed using appropriate crowd control procedures. One of the objectives of this project is to reduce the need for crowd control and design into the Bridge optimal crowd flow onto the structure. The difference between crowd control and crowd management needs to be stressed.

Crowd Control

Holding back the crowd can lead to increased density and there is a balance between controlling the crowd to prevent harm and increasing the crowd density as the following photograph illustrates. We developed a tool which can be used to study the density around these areas that can be used to test a wide range of crowd control measures.

Crowd Management.

The combination of appropriate design criteria and minimal crowd control are the key elements of good crowd management. The design should have sufficient areas to allow the security forces to respond to crowd dynamics and emergency situations as the may arise. We have ensured that sufficient space is designed into the Bridge for the Hajjis.



No environment that would experience the vast number of people passing through a complex geometry in a short period of time would operate WITHOUT some form of crowd management. The objective is to reduce the need for crowd control and ensure that the design and the management work in harmony. Chapter 2 2.0 Methods used in the analysis The main tool used for the MPW&H Jamarat Bridge analysis is Myriad. A system was supplied to Prof. Dr. AlGadhi and demonstrations of the core code and model building content were provided to the Custodian of the two Holy Mosques Institute for Hajj Research and the Ministry of Public Works and Housing as per the Terms of Reference. Myriad is a multi-scalar modelling tool developed by Crowd Dynamics Limited. The system has been developed as a third generation tool (following the VEgAS and Legion tools previously developed by Dr. G. Keith Still). In principle the system is a geometric analysis coupled with CAD/Bitmap reading algorithms. The combination of entity based code (Simulex) and geometric analysis makes Myriad a fast and very powerful analytical system capable of meeting the timescale set by the terms of reference (model building and analysis in four weeks). 2.1 Myriad To highlight the functionality of the Myriad system, and model building process, we shall examine the area of the eastern entrance in a five step process. This will illustrate how useful the system is in determining the nature of risk in a high density environment such as Jamarat Bridge. 2.1.1 CAD plan of the area of interest. Using a CAD system we import that map of the area of interest.

2.1.2 Export/Import to Myriad This process is called “Isophotic mapping” where the CAD plan lines are converted to an area (Hajjis walking areas) in black. The buildings and structures are highlighted in grey.



2.1.3 Scaling and Origin/Destination assignment.

Using the grid function to determine the scale for the crowd flow models. The diagram on the left shows the Myriad system displaying the area of interest (eastern entrance of the existing Jamarat Bridge). 2.1.4 The trails model within Myriad This model uses the geometry (circled) to determine the crowd width from this approach. This is then superimposed across ALL the approached to highlight where the high density locations converge. 2.1.5 Superposition/Voronoi

With this process we take the superposition of ALL routes toward the eastern entrance and merge them together (within Myriad). This produces the convergence map. A process called a voronoi polyhedra is then applied to produce a relative density map (cluster convergance). We now compare this to the situation at the eastern entrance of the Jamarat Bridge. To illustrate the high density area we rotate the image to match the orientation of the photograph. The solid red areas are at maximum density. The model shows the development of high density areas.

We supplement the superposition models with a microsimulation technique to test the density of the area near the eastern entrance.



The microsimulation also highlights the high density build up at the eastern entrance. Wherever Myriad produces a superposition map (macroscopic simulation – large scale, quick analysis) that highlights a potential crushing problem we adopt one of these three additional models. 2.2 Queueing models

Queueing models can predict the rate of queueing in specific areas. There are dozens of different types of queueing mode from the very simple (Number of people in queue = arrivals – departures per time period) to the more sophisticated simulations for stochastic arrival rates, exponential services rates and loss systems (where people switch queue position).

2.3 Microscopic simulations/Myriad

Using a combination of microscopic and macroscopic techniques can assess the progression of queues such as those approaching the Jamarah (above – Simulex/Myriad, below Myriad and site photograph).



You can model large areas in a short time, modifying the design and iterating between the model and the CAD plan to understand where high density build up may exist, increase the area available and run the model again. The process has been designed to run quickly from design to model with minimum (a few mouse clicks) of effort.

2.4 Data Analysis From the microscopic models we can produce minute-by-minute cross sectional data of the areas under scrutiny – allowing us to examine the high density build up (Hajjis speed and density).

2.5 Network Analysis

This technique uses a macroscopic over view of the Bridge and couples origin and destination pairs to provide an overview of the demand at each level. Typically a project would start with an overview (network analysis) a macroscopic analysis Myriad, Walker etc) and as areas are highlighted as having the potential for problems the modeller should then adopt a higher resolution technique (microscopic simulations).



2.6 Optimal Design Criteria. All of the above are used to produce an optimal design and performance criteria. We established the following model during the early stages of the analysis of the MPW&H Jamarat Bridge Project. Jamarat Bridge – Overview model - each element of the bridge is broken down into a design component for analysis.

1. Sufficient arrival capacity 2. Sufficient throwing area 3. Sufficient space (density ≤4 Hajjis per square metre) 4. Sufficient passing area 5. Sufficient egress capacity

Chapter 3 3.0 Testing the Jamarat Bridge capacity To establish the operating parameters of the Jamarat Bridge we need to consider the key elements of the Rajm process. These are ingress (sufficient to allow 125,000 Hajjis per hour) the stoning (Jamarah throwing area/perimeter) and egress (leaving the area). Ingress capacity and the stoning area are related in a mathematical model know as a queueing system. This means that the process involves an arrival rate (a function of the entry ramps/escalators) and the service rate (the stoning process). This relationship is critical and if not balanced the queue can grow exponentially giving rise to extreme conditions such are observed in the current design.



The development of the oval shaped queue is related to the ratio of the arrival width to the throwing area. The existing design has a 45 metre ingress (eastern entrance) which opens to an 80 metre bridge width then narrows again to 45 metres. Security forces are used to control the arrival rate at the Jamarah.

The Jamarah is 16 metres in diameter and we have ratio of nearly 1:1 for the ingress width to the Jamarat perimeter (16*pi = 50m approx.). However only ½ of the circular profile is on the approach path which we can see leads to high density queueing. The lines of sight (at eye level) are such that the available perimeter is less than the total perimeter. Any ratio greater than 1:0.8 will lead to queueing due to the time it takes to throw the stones.

This problem is much clearer when we examine the arial photographs of the existing Jamarat Bridge We can clearly see that the profile of the arrivals exceeds the profile of the throwing area. This leads to a dangerous queueing system and it requires considerable crowd control to prevent serious injuries. The proposed design needs to incorporate a balanced arrival/service rate to facilitate a safer environment.

3.1 Optimal Jamarah Perimeter To determine the boundary conditions for the relative arrival rates (at 75 Hajjis per metre per minute (Highways Capacity Manual) for the Rajm process we use three different approaches and test to see if they are in agreement.



Given an arrival rate of 75 Hajjis per metre per minute we can deduce the following relationship. A ratio of 1:3 (arrival width versus available throwing perimeter) is a good approximation for the relative dimensions of the Jamarah. Thus for every metre of ingress width we need to allow at least 3 metres of throwing perimeter. Of course this type of model does NOT tell us anything about the geometry of the Jamarah with respect to the direction of the arriving Hajjis.

Our objective is to assess the design and if the design fails any of the optimal criteria to advise on suitable modifications. Using Myriad to determine if there are any areas that could give rise to high density crushing, we tested the superstructure at each level and then the ground floor (after modifications). 3.1.1 Optimal Shape for the Jamarah perimeter. We ran three different types of model to assess the optimal shapes for the Jamarah. The first set of models used a simulation and the results are show for each shape in the appendices. 3.1.2 Circle. The existing design has a 16 metre diameter and offers 50 metres of perimeter. The approach is 40-50 metres (via barriers used to control the Hajjis arrival rate). When the arrival (ingress) width exceed the x3 factor a queue of Hajjis will develop at an exponential rate. The perimeter around a circular Jamarah increases at a ratio of 6 metres for every 1 metre of radius. The geometry of a circle therefore does not lend itself for optimisation.



To evaluate the sensitivity of this shape we ran a series of tests to compare the result from the existing configuration and determine if there were some elements that may improve the existing bridge design. This also highlighted the ingress width to perimeter queueing ratio. The sensitivity of a throughput model is shown in the graph below.

Width vs Throughput

020000400006000080000

100000120000140000160000

7 10 13 17 20 23 27 30 33

Ingress Width

Haj

jis p

er h

our

EgressIngress

The throughput (graph above) is taken with a number of Hajjis already in the model (hence at low ingress width we have a slightly higher egress rate The percentage of Hajjis queueing in the model also shows the nature of the problem of the ingress to perimeter ratio. The models for the circle are shown below in a series of graphs. These illustrate the equilibrium state of the queue for a variety of different (increasing) ingress widths. This highlights the issues relating to the high density queue and we can see the evolution of the oval present in the existing system. A tool was developed to allow for experimentation of “bubbles” and further tests are planned to assess the optimal security cordons.

Width vs Queueing

0.00

5.00

10.00

15.00

20.00

25.00

30.00

0 5 10 15 20 25 30 35

Ingress Width

Que

uein

g H

ajjis



3.1.3 Evolution of the high density The graphs below show a three dimensional projection of the Hajjis density at the approach to the Large Jamarah with different ingress widths. We see that as the ingress width increases (in 3.33 metre increments) the queue grows at a much higher rate and at the level of the existing Jamarah design the queue takes the familiar oval dimensions. Red indicates > 8 Hajjis per metre.

6.7m ingress width 10m ingress width

13.3m ingress width 16.7m ingress width

20m ingress width 23.3m ingress width



26.7m ingress width 30m ingress width

33.3m ingress width 36.7m ingress width

40m ingress width to the 16m diameter Jamarah



The model was designed to test the “bubble” used by security forces and the ingress approach. A copy was provided to the Ministry of Public Works and Housing and the Custodian of the two Holy Mosques Institute for Hajj Research for further analysis. The dimensions of the queue (length and density) conform to the observations both from the video footage and from the research papers of Professor Dr. Saad AlGadhi. The simulation predicts an oval of dimensions 72 metres (without security forces) by 20 metres wide which is a close match to the existing high congestion oval. To facilitate 125,000 Hajjis per hour we would require a circle of diameter 26 metres with an approach that would allow both side to be serviced evenly. Showing this in the context of the MPW&H CAD plan (Level 2) we can see the problem is geometric in nature.

26m diameter circle in the MPW&H proposed design.

The approach to the Jamarah could be designed at a more obtuse angle and the Jamarah managed to facilitate an even distribution of Hajjis. However, this would require significant changes to the design, some of which appear to be severely constrained within the Mina Valley. To facilitate 125,000 Hajjis per hour the geometry of the approach highlights an asymmetrical perimeter would be most efficient. Further tests are planned to assess the optimal in situ design and we developed a new range of tools for this type of analysis.



3.1.4 Ellipse. The ellipse is more appropriate for the Rajm process. Further tests were performed using different angles of approach and the sensitivity of these in situ, illustrated below.

The ellipse has two axis (major and minor) which allow for a greater flexibility to increase Jamarah perimeter along the major axis without increasing the width and reducing the passing area. Many elliptical shapes were tested and a scientific tool to test shapes was developed. The programme was provide to the Custodian of the Two Holy Mosques Institute for Hajj Research and the MPW&H along with ½ day training in their use. As each level of the Jamarah has different approach directions (to Jamarah Al Sughrah) and due to the position and size of the well on the ground level (level 0) an ellipse was chosen to ensure that there was sufficient Jamarah perimeter to allow the Hajjis to perform Ramj without the difficulties currently experienced in the existing structure.

Deformed Circle. Several shapes were tested but they all failed to outperform the elliptical shape. See appendix for further information about the full range of shape tests performed in this analysis.



3.2 Working from CAD Plans We used the CAD plans supplied by Dar Al-Handasah SO175-XXX and imported each plan to the Myriad test suite. There are three elements to the testing (beyond the 5 design criteria)

Ingress The total available ingress width must be greater than 28 metres to allow 125,000 Hajjis per hour. This is a minimum requirement and provision for security forces/civil defence, bi-directional/counter flow and hesitation (Hajjis stopping to rest) on the longer ingress ramps are additional width requirements.

Circulation The Jamarat Bridge is a complex system with many levels. No area in the system should exhibit the characteristics observed in the sad accidents (see previous section). Therefore we test for convergence at a macroscopic level and for density at a microscopic level if any areas fail. This applies to all levels in the design

Egress There must be sufficient egress routes for both normal and emergency departures from each level over the whole period of use during the Hajj. 3.3 Test conditions We therefore have 5 design criteria under 3 different conditions (ingress, circulation and egress) where evacuation scenarios need to be considered for example, evacuation of the area at 11:45 on the 12th day. An emergency situation during the period 11:00 until noon on the 12th day may, for example, see the ramps on level 2 full of waiting Hajjis. If there is a need to clear this area sufficient egress capacity need to be provided. There is a great deal of flexibility to the design and the ingress ramps have provision for emergency egress towers (minor towers). The exact location and capacity of the minor towers needs to be sufficient to move the Hajjis off the structure in a reasonable time. 3.4 Emergency Evacuation For the evacuation scenarios we tested the system using the combination of Simulex and Myriad.

Simulex is a commercially available emergency egress microsimulation and the Myriad system has been designed to extract the positional information from the simulation and display these as density, speed and headway (time between successive Hajjis arriving at any point). We tested the whole system using both a network analysis and microsimulations.



Each tower is capable of evacuating a 27 square meter area per minute. Every level is served by a number of towers and the evacuation rates per level are shown below. Note that we take the worst case scenario as the benchmark for the design (ie: if it passes this test then all other scenarios should have fewer problems – naturally this requires further testing as the detailed design develops and should be part of an on-going analysis). Level 1 = 3 Egress ramps @ 75 Hajjis per metre = 6300 5 Major towers @ 734 Hajjis per minute = 675 4 Minor towers @ 734 Hajjis per minute = 540 Total Egress Capacity = 7515 Level 2 = 1 Egress ramp @ 75 Hajjis per metre = 3000 5 Major towers @ 734 Hajjis per minute = 675 12 Minor towers @ 734 Hajjis per minute = 1620 Total Egress Capacity = 5295 Level 3 = No egress ramps (as drawn) = 0 5 Major towers @ 734 Hajjis per minute = 675 6 banks escalators at 58.3 Hajjis per minute = 1400 6 banks stairs at 56 Hajjis per metre = 1344 Total Egress Capacity = 3419 Level 4 = 1 Egress ramp @ 75 Hajjis per metre = 2700 5 Major towers @ 734 Hajjis per minute = 675 Total Egress Capacity = 3375 As the structure has no fire loading (other than the Hajjis) it is unlikely that the whole structure would require simultaneous evacuation. It should also be noted that evacuating the structure to the ground floor could cause further congestion and potential problems at the discharge points. Further analysis on these scenarios may be required as the design is finalised. 3.5 Analysis results for the towers Assuming the escalators are used for vehicles and injured Hajjis we calculate the egress rates using two different approaches. Firstly the SFPE/NFPA codes (Society of Fire Prevention Engineers and National Fire Prevention Associate – USA) which again rely on the Levels of Service. Maximum flow rate in metres per minute down stairs is quoted at 56 Hajjis per metre per minute. We have 2.7 metres of stairs but these are served by a door of 1.8 metres. The calculation would therefore be at the door rate and a value of 40 Hajjis per metre per minute is typically adopted. There is a lot of debate about how egress should be calculated and a wide range of assumptions as the following demonstration illustrates. Values vary from country to country (for example, 1.5 mm per person fro buildings with more than 500 persons). Clearly this would not apply to the Jamarat Bridge so we combine a calculation and simulation to derive the appropriate values. 56 Hajjis per metre per minute (stairs) x 2.7 metres of stairs would be an egress rate of 151 Hajjis per minute per tower. Level of service at maximum flow rate for the door would be 135 Hajjis per minute (75 Hajjis per metre per minute * 1.8 metres). As the door would limit the egress capacity we take this value for egress rate. By simulation, taking the geometry of the door AND the stairs into account, we derive a mean time of 140 Hajjis per minute. This value is higher than the calculations but within 5% of the calculation level.



Tower egress rate simulation results – original design. Given the design and occupancy levels of the proposed design we used the value of 135 Hajjis per minute for the proposed design. This can be increased with some minor design modifications. Therefore the total evacuation rate per tower (Major) is 675 Hajjis per level.

3.6 Egress distribution Myriad was used to identify the distribution potential for each emergency egress (minor/major towers). This analysis is performed at a macroscopic level and the display (red) shows those areas that are not evenly serviced by egress towers.

The map (left) should show an even coverage but as these are details which have a flexible arrangement this type of analysis is best performed at a later stage of the design process when the ground floor footprint is finalised, the structural elements are added and the detailed drawings are agreed.

It should also be noted that crowd management will be required to facilitate emergency egress and these figures are best case egress rates. Assumptions about the time to start the process often miss the factors of reaction time. Furthermore, the rates of egress are dependant on the tower configuration and these could be improved by better configuration. 3.7 Further egress tests Two further designs were tested on the egress towers (see below).

The two diagrams above show the drawings of the Major Egress tower (right) and a modified tower (right) which removes the corners.

Tower Egress Rate

0

50

100

150

200

3 5 6 8 10 11 13 15 16 18 20 21 23 25 26 28

Time (minutes)

Flow

Rat

e



The main differences are the rounded ingress and alignment of the door at the top of the stairs. These produced a higher egress rate and, as such, are details that will require further testing once the design is finalised. The egress analysis takes a few minutes from CAD plans and a copy of the Myriad system was provided to Dr. AlGadhi for further work as required.

3.8 Screen shot from the Egress Tool (Myriad/Simulex)

Egress rates are a function of the initial reaction time, the travel distance and the rate of egress through the geometry. These are relatively straightforward to model and further testing as the design is finalised is recommended. Myriad provides density, speeds and headway (delay) during egress so a comprehensive analysis of the design can be implemented as required. It is a recommendation that this needs to be part of an ongoing process during the detailed design and an independent advisor appointed for continuous testing, working with the design teams. The addition of further egress towers has an impact on the ground floor, the crowd dynamics on that floor and therefore the movement of Hajjis through the system. Jamarat Bridge can be modelled as a whole system – but this has problems with model set-up, analysis and interpretation. A combination of a network model (macroscale) and a simulation (microscale) offers the simplest, most effective methodology to assess the impact key elements would have on the overall Hajjis movement and safety. Chapter 4. 4.0 Level by level analysis against optimal design criteria. We began the analysis by first testing each floor for optimal design criteria. This is a macroscopic modelling technique, the criteria for a system “failure” was based on the initial drawings. We worked with the architects to address these issues and resolved everything except the Ground Floor level which requires further discussion with the religious scholars (re: size of the Jamarah ring on the ground floor – level 0).



4.1 Level 4. Level 4 services the Hill of Hadarim and replaces the 30 metre descent via the stairs (see photograph below).

The ingress route is 36 metres wide which, at 75 Hajjis per metre per minute, would have a theoretical capacity of 162,000 Hajjis per hour (above the test limit of 125,000 Hajjis per level per hour). Using the x3 factor we derive a Jamarah perimeter of 108 metres and taking the lines of sight we can see that a minor modification to the approach route (diagram above right) will allow optimal utilisation and balanced ingress/throwing area. A similar, minor modification was suggested for the egress route, changing the angles of the departure route to facilitate move even utilisation of space and reduce congestion.

Minor changes to the angles of approach.



It should be noted that this level would benefit from an integrated transport management approach as it has the potential to service the latent demand for future expansion. Any structural changes to this level may have an impact on the ground floor (level 0). It is recommended that further analysis be performed as the detailed design progresses. The entrance and exit points may also benefit from a detailed study that would incorporate transport management. A traffic/pedestrian survey of the next Hajjis is important to understand the arrival patterns. Coupled with questionnaires this information will allow the design team to assess the impact that the new design will have on the Hajj. It is important to understand that Jamarat Bridge is one element in a complex system. By analogy, you should not put a Rolls Royce engine in Citroen 2CV. The impact that Jamarat Bridge may have on the rest of the Hajj could be significant and it is strongly recommended that a holistic model (whole view) of the Hajj is constructed to ensure that Jamarat works within context. For example, there is little point in designing a level that can accommodate 125,000 Hajjis (or more) per hour if the transport system at level 4 cannot support the arrival (and departure) rate. Therefore the limits to this level may not be the design – but the transportation system taking Hajjis to (and away from) the area.



4.2 Level 3. The third level of the MPW&H proposed design had similar ingress, lines of sight and egress problems which were quickly modified by the architect (see appendix for further details). The major issue with the third level was the use of escalators. We held discussions with all parties involved in the design and management of the Jamarat Bridge, examined video tapes from Al Haram and questioned OTIS (escalator manufacturers), London Underground Safety Engineers and police/security contacts who managed both the Sydney Olympics and the Commonwealth Games where equivalent high density crowds were exposed to mechanical systems. It was noted that the Hajjis are unique in that many are poor, elderly and come from different cultural/language backgrounds. 4.2.1 Escalators The initial design is shown below. The main concerns were related to the width of the ingress route, the number of 180 degree turns and the available area at each level.

Myriad models of the initial design

The initial design had a 17 metre wide entry gap leading to a set of escalators (4x 1Metre) and two sets of stairs (2m per stair). We tested lines of sight, rate of ingress to service rate on the escalators and high density potential. We also tested several failure modes (emergency stop on an escalator) and found that, given the escalator rates and the available space, this system had four/five minutes before crush density at any one level. The initial design had six landings where crush density could occur. After several discussions with the architect, studying the video tapes of Al Haram, and discussion with various experts around the world via the Crowd Dynamics Forum, we proposed a series of safety measures for this element of the system. 4.2.2 Al Haram In discussion with various members of the committee two different opinions developed, namely: escalators were “safe” because they had been in operation at Al Haram for many years. Conversely that they were dangerous because there have been incidents at these escalators. Dr. Osama provided some video footage of the escalators operating with problems and without problems. 4.2.3 Escalator Case Study 1. The first set of video tapes showed ingress to the escalators at Al Haram. These clearly showed that there were indeed problems with severe congestion upon arrival. However, this is an attribute of the entry system and NOT of the escalators themselves as the following photographs illustrate.



Congestion at the base of the escalators is due to the queueing arrangement. The Hajjis on the left of the photograph are waiting for the escalator to reverse to take them to the top of the Holy Mosque. This creates the congestion – a more appropriate crowd management method would be to create a queueing system in a less congested area (outside) to facilitate the normal egress of the descending Hajjis. As these escalators operate in a bi-directional manner (sometimes up, sometimes down) there

appeared to be confusion at the entry system. Queueing formed quickly and Hajjis had little room (and less choice) to manoeuvre. As more Hajjis entered the system the area became very congested and the escalators were switched off to prevent injury. It needs to be stressed that this is not a problem with the escalators themselves – but with the design of the entry area. Crowd management external to this area would prevent many potentially harmful situations. Further studies of the approach and the design should be commissioned as there appears to be some minor modifications that would create a self-regulating environment and significantly improve Hajjis safety in this environment. 4.2.4 Escalator Case Study 2.

In the second case study we observed the management of the escalators. They were switched off whenever the landing area became congested. This was to prevent the over crowding at the landing areas (180 degree turns). As we had feared, sustained high traffic on this type of system can lead to dangerous overcrowding at the landings between escalators banks. Further video footage of the 180 degree turns/landings were

observed and confirmed that improvements could be made to this area. As a managed system the escalators could be “made safe” but did require manual intervention to prevent dangerous congestion. The Jamarat design had 11 escalator towers with six landings in each. Clearly this had a high potential risk (coupled with the wide entry gates). 4.2.5 Escalator Case Study 3. The third video study we performed investigated the escalators operating very efficiently. This was during the egress (after prayers) where the movement was in ONE direction only. It was now clear why there were two opposing opinions about escalators.



Al Haram escalators working efficiently at the end of prayers where the movement is one way, the exit system is now WIDER than the escalator banks and the Hajjis are moving from the escalators AWAY from the congestion points. One area, three different modes of operation, two different opinions about the operational efficiency and safety of the system.

If you were to visit these escalators during the period where there is bi-directional movement and a managed (manual on/off) operation you would conclude that this system was chaotic. The lines of sight are poor, the area has no clear directional crowd flow and there are problems with information (which escalator do they take, are they off/broken?) If you were to visit them during egress you would observe a very efficient system of crowd movement. The Hajjis flow down and out of the area without incident. Both observations are correct. It is the operation and layout of the entry system that requires modification. In the Jamarat Bridge the escalators will operate one-way only and we proposed some modifications to their design – building in many safety features. From our research we noted that the main causes of accidents were boarding and alighting the escalators (we proposed a managed system), loose clothing/shoes getting trapped (we suggested a turning barrier as advised by OTIS engineers), safety stairs (at every level – to allow emergency access and egress) and more space at the landings to prevent crushing. This increased the size of the towers – but the safety measures are deemed necessary for Hajjis safety. 4.2.6 Proposed changes to the design

Increase landing area Managed system Balanced ingress area Fewer landings Safety stairs at every

landing Safety barriers at every

landing to prevent loose clothing getting trapped

External screening (no luggage). This needs to be done AWAY from the escalator entrances

Investigation into escalator and elevator combinations (for the mobility impaired and VIPs)



4.3 Level 2. The superstructure (support pillars) had the potential to improve the sightlines with some minor modifications and it was recommended (and approved) to change these.

Using the Myriad sightline tool (see diagram above) we can test for convergence AND lines of sight. Opening the angle on the approach provides for more streamlined crowd dynamics. The consequence of this is to distribute the Hajjis more evenly across the Jamarah perimeter and therefore reduce the high density queuing. There were no other issues with this level and it passed all the optimal design criteria.

There was a consideration for the Mosque to the East of the Makkah approach ramp. This may have the potential to be foreshortened and lead to an asymmetrical design of the levels 1 and 2. We did not model this but it is suggested that there may be additional crowd management benefits from the asymmetrical ramp configuration. It is worth additional study as there may be considerable cost benefits to this suggestion.



4.4 Level 1 There were three minor issues with this level. The Northern ingress ramp had a convergence on the approach to the Jamarah, the available perimeter was smaller than the x3 factor and the egress ramps (north/south) at the large Jamarah had the potential for failure. These issues were quickly resolved and the design modified accordingly. We advised that the central barriers be removed as these would be a potential trip hazard for the Hajjis. A split bridge was tested but it was clear that to facilitate the Rajm process (including Dua’a at the appropriate locations – right of small Jamarah etc) and prevent Hajjis migrating to areas that were not accessible, a break in the sight line would be required. It should be noted that the lies of sight are the primary direction that the Hajjis will take – if a split bridge were to be adopted it would require a barrier between the two bridges to facilitate directed circulation. This would have to be ABOVE the Hajjis heads so that pilgrims from afar would not try and move to the appropriate Dua’a areas. Our recommendations were to retain a simpler single floor, remove the barriers (but not the Jamarah ring “noses” – see chapter 3). Several suggestions relating to the ingress ramp direction of approach and angle of attack toward the small Jamarah were considered. This should be the subject of further study as the design is finalised. CAD plan S0175-A-11revC.dwg had a much shallower angle of approach considerably altering the efficiency of the Hajjis moving towards the Jamarah for the Ramj process.



A final discussion point was the angle of the Mina egress ramps (see diagram below).

Working with the architects we tested several combinations of ingress and egress layouts. With only minor modifications Level 1 passed the optimal design criteria and has the potential capacity for 125,000 Hajjis per hour.

S0175-A-11revC.dwg

In the CAD plan above DAR have altered the angle of approach to the small Jamarah and changed the angles of the exit ramps. As the design is finalised these details should be part of an on-going modelling/testing process. Further models and techniques can be applied to refine the analysis.



4.5 Ground Level (Level 0) The concern at the ground level is the potential that Hajjis may perceive a difference between the “new” structure towering above them and the “traditional” structure on the existing ground floor level. By creating a wide plaza area the ground floor will become more visible to the Hajjis.

View from the Mina approach. We ran a series of models for the ground floor indicating the directions of approach (and return path) the areas of cross over and potential crowd management point are highlighted in red.

These approach paths, allowing for distribution of arrivals evenly across the area are highlighting the bi-directional flow path – areas where Hajjis will be moving in opposite directions AND areas where the Hajjis may experience congestion. It should be noted that this area will require further modelling as the facilities, ground support and crowd management plans need to be part of a final design/development model. However, we can use the system to determine where the most appropriate location for crowd management and where the probability of conflict (potential risk) may occur. As the design is finalised this process should be repeated and included as part of the “operators manual” for the bridge. The size of the ramps, angles of approach and design of the entry and security points around the ramps and escalators require additional study as the design is finalised.



Using the distribution figures from the Hajj Research Institute we can estimate the high density potential in the new plaza area approaching the Large Jamarat. The transport terminals will affect the approach and departure paths, this should be the subject of additional studies for the ground level.

The figures (left) indicate a distribution approaching the ground floor in the order of 2/3:1/3. If we assume that we can manage the ground floor approach to be predominately the Mina approach then the queueing system approaching all three Jamarah will be manageable. This implies that policing in the areas toward Makkah should focus Hajjis to Level 2.

In the model above we have highlighted the convergence of Hajjis for the 12th day where they would depart the Large Jamarah and move toward Makkah. Only a small section in the central (some 700 square metres – shown in red above) has the potential for cross flow and conflicts. This will change depending on the layout and operation of the bus terminals, the egress routes and the modal split of Hajjis travelling back to Makkah. A further study is suggested in this area which, at present, does not give rise to concern for Hajjis safety. With the new design an environmental (Hajjis movement) impact will resultant. It is strongly suggested that monitoring of the existing Jamarat Bridge be performed both by questionnaires and physical monitoring of Hajjis and traffic be instigate for the next Hajj. We have investigated the project for monitoring this area and a proposal is being compiled for consideration. The ellipse is highlighted as an optimal shape and this would make all levels consistent in design and operation.



Chapter 5. 5.0 Conclusions The current situation at Jamarat Bridge is life threatening. The safety of the Hajjis is the primary concern in the design and management of the MPW&H & HRI proposal.

The Jamarat Bridge requires significant crowd management and gives rise to severe overcrowding across the whole area. The MPW&H design offers a significant improvement to the problems encountered every year.

We asked three questions of the study Is it safe? After a few minor changes and working closely with the design team we can conclude that the design is safe. Is it safer? The current situation is life threatening, any changes to the area that increase Hajjis safety will make the environment safer. Is it the safest? After considering the long history of the Jamarat Bridge and the many proposals that were presented we can conclude that there are many elements of a high density environment that can give rise to potential problems. The use of escalators and multiple levels may open different types of risk for example higher number of Hajjis leaving the area at the same time. This was calculated at 4x the current discharge rate of the existing bridge. Solutions toe these and many other problems need to be considered as the development of the bridge progresses. Many additional risks, for example, traffic hazards were not considered in the scope of work but have been outlined in this report. It is strongly recommended that further study be part of an on-going basis of design and development for this proposal. It should also be stressed that this would be a requirement of ANY design in a high density environment. As no alternative were presented for similar analysis we cannot conclude that this is the safest possible design.



To summarise our conclusions.

It should be highlighted that the design is conceptual but, after several sessions with the architect, fulfils the design criteria and simulation testing.

It is our recommendation that the Ground Floor should be discussed with the religious

scholars and the result of that discussion would allow an increase in the Jamarah perimeter further increasing the capacity of the system. We are running additional tests on this level to evaluate Hajjis approach, changes in crowd management and conflict with departing Hajjis. That work will be subject to a further report.

The lower throughput (based on the Jamarah size in the plans provided, crowd dynamics,

safety considerations and crowd management issues) has been determined as sufficient to accommodate the desired capacity. The upper limit of the MPW&H Jamarat Bridge proposal is approx. 3,900,000 Hajjis in 5.5 hours which would depend on increasing the Jamarah perimeter at level 0.



Chapter 6. 6.0 Recommendations The successful operation of the bridge requires integration of BOTH design and operational management. We have ensured that all possible safety measures, emergency evacuation and medical facilities are integrated in the design. It is recommended that an operations manual/procedure be developed during the detailed design phase.

A ramp to service the 3rd level and to alleviate potentially serious crowding in the Mina approach to the Jamarat Bridge was also agreed as an important element to the design.

For Level 0 (ground floor) a discussion with the religious scholars is recommended. We are

highlighting the potential for overcrowding with the existing diameter of the Jamarah. If it is possible to increase the perimeter it would significantly increase the safety of the Hajjis.

It is our recommendation that an integrated modelling and crowd dynamics plan be continuously

improved as design modifications (during the detailed planning and construction phases) are included.

It is recommended that an operational procedure and manual be developed during the design phase

working with Civil Defence and Public Security.

It is recommended that a survey of the Mina Valley consisting of questionnaires and physical measurements is required for the next Hajj, the first year of operation and the second year of operation to establish the performance of the bridge and to assess the potential impact on the crowd and traffic dynamics in the Jamarat Bridge area.

We recommend that during the detailed design phase that further tests and computer simulations

are performed. This should be a continuous process during the development and installation of the Jamarat Bridge.

We confirm that the design passes our safety considerations.



Appendix - Project Overview Objectives

Provide independent assessment of MPW&H proposed design Provide recommendations for design (and operation) to improve pilgrim safety facilitate stoning ritual to ensure throughput targets can be achieved

Outcomes

Safe design for Jamarat Bridge that achieves throughput target of 3m pilgrims during Hajj In addition the design review may lead to significant cost savings, far exceeding the cost of

the project Construction cost savings in operations costs from design that encourages the crowd to self-

regulate crowd flows and density and therefore improve Hajjis safety Program of Work

Assess MPW&H proposed design Build models/simulations of each of 5 levels Develop a holistic view and interfaces between levels Develop models/simulations of the Interfaces with the surrounding area

Assessment includes scenarios for

Maximum demand (125,000 pilgrims/hour/level) Peak within peak demand (from 1995 and 2002 measurements) Range of OD's and demand distribution between levels Sensitivity tests Jamarah shape, ingress width/queueing, stoning and prayer times 90% of pilgrims stoning and 10% not stoning (assumptions) Two evacuation scenarios from bridge only Additional scenarios/models as required.

Area to be assessed

The Bridge The immediate surrounding area

Assessment includes

CDL expertise Creation of appropriate models and simulations [they are different things] Review of existing models (Walker)

Outputs

Study will provide expected outputs as described in initial ToR Deliverables

Report describing work undertaken, analysis, findings and recommendations Models (e.g., spreadsheets) and simulations as required Presentations to the authorities

Further studies

Development and testing of designs and operations plans for areas around bridge into which pilgrims migrate during evacuations

Development and testing of designs and operations plans for other emergencies to be proposed later

Assessment of designs and plans for areas around Jamarat bridge Assessment of designs and plans for other elements of Hajj



Shape tests.

The following shapes were tested and the results show in the table below. Please note that the sensitivity of this analysis is dependant on the approach path/width/geometry. We include it here as a comparative analysis between the relative shapes. It should also be noted that the shapes are 30cm resolution geometry in a square matrix. Thus the tests is taking 1 Hajjos per pexel as the density has been recorded at 8-9 Hajjis per square metre this resolution allows the density to exceed the safety limits and hence it highlights the issues relating to Hajji safety.

16m circle in 100m wide area 20m wide circle in 100m wide area

20m circle in 100m wide area 30m wide circle in 100m wide area

As the width of the circle exceeds the ingress width we begin to experience “edge” effects at two locations. The ingress (right hand side of model) begins to split into two paths (upper and lower) of Hajjis as the queue begins to form. This encroaches on the edge of the geometry under test. We also find that the upper and lower walls of the area start to restrict the Hajjis movement. This “edge effect” is due to too big an obstacle in too small a space. The need to test geometries in situ became apparent after the tests we ran on ingress width (see main text).

36m wide circle in 100m wide area



Results for Circle. Throughput from the model is shown in the graph on the left. This indicates that the larger circles have lesser throughput – however we stress the impact of the local geometry and that this test does NOT prove that a circle has the same throughput in situ. Results from the model for the number of Hajjis in the queue ie: the number of Hajjis that are stationary. This shows us that as the circle increases in diameter the number of Hajjis queueing increases. However it should be stressed that this data includes those Hajjis navigating the edges so again has an element of the edge effect included.

Circle

0

20000

40000

60000

80000

100000

16 20 30 36

Diameter

Haj

jis p

er h

our

Circle

0.000.050.100.150.200.250.30

16 20 30 36

Diameter

% in

Que

ue



Ellipses

A wide range of ellipses were tested. The results of the same shapes tested in the Walker model are shown below. As the local geometry, edge effect and resolution on the models are different we are not surprised at the different results. In the geometry we used the 12m wide ellipse had the most queueing due to the modelling technique. As the areas and approach widths are different in each model the tests should use either a constant area or constant perimeter to be comparable. We highlight that this type of testing is not providing valid results and we focussed on working with in situ modelling techniques.

Just as testing the aerodynamics of a wing in a wind tunnel does not tell us how the wing would perform in a cross wind we highlighted that a head on approach to an ellipse as opposed to the situation at the Jamarat Bridge would be misleading. The study of performance in situ is shown in the diagrams below.

Ellipse

80000

85000

90000

95000

100000

105000

32x10 32x12 36x12 36x14 40x14

Diameter

Haj

jis p

er h

our

Ellipse

0.000.020.040.060.080.100.120.14

32x10 32x12 36x12 36x14 40x14

Diameter

% in

Que

ue



We developed a model of individual Hajjis moving toward the small Jamarah at level 1 to highlight the issues of anisotropy (directional bias). As the local geometry and crowd management in the approach paths all have a direct effect on the angle of attack we can clearly see that the approach from the two ramps have different queueing characteristics. A model of this was created in two different forms and copies of these were left with the MPW&H and HRI programmers for further exploration. We offered support and continual process testing for the duration of this project. We also highlighted the need to appoint a full time member of staff/researcher to develop these techniques for Dua’a and the Rajm process. Our work will continue on these models using a wide range of techniques and further tests are planned. Deformed Circle

We used the same geometry to test the deformed circles and the results are show below.

Again we expressed concern about isolation tests of this nature and direct comparison to the Walker model is not comparing like-for-like. The Walker model has a hexagonal 4m grid – the models we were using have a 30cm square lattice.

Deformed Ellipse

65000700007500080000850009000095000

16x8@8 20x10@15 20x10@21 24x12@22

Diameter

Haj

jis p

er h

our

Deformed Ellipse

0.00

0.05

0.10

0.15

0.20

0.25

16x8@8 20x10@15 20x10@21 24x12@22

Diameter

% in

Que

ue



Maximum Throughput There are three elements that affect the maximum throughput of the system. 1. Ingress width

This is calculated at 75 Hajjis per metre per minute as a maxima one-way flow based on 2-3 Hajjis per square metre. As all egress routes exceed the ingress width there were no ingress/egress constraints on any of the levels.

2. Rajm Process We highlighted in the main text of this report that a x3 factor of ingress width to throwing perimeter should be allowed.

3. Passing area/Dua’a The area between the Jamarah should have sufficient space for the Hajjis prayer (Rajm process). We tested the capacity and areas at the critical points for Dua’a (right, left of the small and middle Jamarah respectively). There were no further constraints in the MPW&H design.

Taking a stochastic analysis of the queuing from a wide range of models and simulations we can test the throughput of the different shapes. Taking the maxima for each shape and comparing that to the Walker results gave us the following table.

Walker % of Max AlGadhi/Still % of Max Circle 102,200 0.82 88,650 0.87 Ellipse 125,200 1.00 101,484 1.00

Deformed 110,600 0.88 90,978 0.90 It should be highlighted that these results are not comparing like for like but when you compare the relative differences the two results are in close accord. We also stress that modelling in this manner does not provide a system performance. The table above compares the Walker geometric shapes, in isolation, against similar shapes (with higher resolution) again in isolation. We adopted a stochastic model of the Rajm process with directional biases in the Dua’a. We also compared the model results with the physical measures from Al Jamarat and highlight the 69,000 field study throughput (AlGadhi “Simulation of Crowd Behaviour and Movement: Fundamental Relations and Applications – Transport Research Record 1320) is relative to a 16 metre unidirectional Hajjis loading. Using the model described in the main text – an In situ test – the throughput DROPPED to a maximum value of 82,200. It is further noted that this level is already exhibiting a saturated queueing system and therefore NOT the most efficient shape (based on the 16m circle as in the existing design). We further note that the models do NOT display the hesitation, confusion and lack of directional focus that the Hajjis exhibit in the Rajm process. It is noted from the photograph below that Hajjis lack a clear sense of direction as there is little by way of orientation once passed the Large Jamarah. This would incur additional delays in reality – so we expect our modelling results to be higher than reality. The same would apply to the Walker model which is a very good flow analysis and proxy for a rough cut capacity analysis for the basic design. As the design passes through the various stages of development is it recommend that further models are built specifically for the value engineering of the overall bridge.

jamarat final report 11 - ksu

Documents