Soft Target Protection Analysis

Using Pedestrian Simulation

Tomas Apeltauer, Jiri Apeltauer, Petra Okrinova
Prescriptive Codes vs Realistic Simulation

2
Prescriptive Codes vs Realistic Simulation

The codes do not consider different speeds according to age or ability to move,
also assume immediate evacuation, i.e. a zero reaction time to alarm.
Pedestrians with reduced mobility are taken into account in the larger required
width of the escape route.

These significant simplifications are compensated by very restrictive basic

values, especially the walking speed. The final assessment is thus significantly
on the safe side.

Parameter Value (Czech prescriptive code)

Pre-movement time 0 s
Walking speed 0.418 - 0.583 m/s (stairs and flat surface)
3
So Why Do We Need Pedestrian Simulations?

4
Brno Football Stadium, 30,000 Pedestrians

5
Brno Football Stadium, 30,000 Pedestrians

Queue time of pedestrians during evacuation

14%

12%

10%
Frequency

8%

6%

4%

2%

0%

1020
1050
1080
210

480

870
30
60
90
120
150
180

240
270
300
330
360
390
420
450

510
540
570
600
630
660
690
720
750
780
810
840

900
930
960
990
Total queue time [s]

6
Brno Football Stadium, 30,000 Pedestrians

Distance traveled by pedestrians during evacuation

14%

12%

10%
Frequency

8%

6%

4%

2%

0%
25
50
75
100
125
150
175
200
225
250
275
300
325
350
375
400
425
450
475
500
525
550
575
600
625
650
>650
Distance [m]

7
Soft Targets and Pedestrian Simulations

The term is used by the security community for places with a high
concentration of visitors and a low level of security against violent attacks.

These objects do not necessarily have to be complex or complicated as in the

first example of stadiums, but the attack will cause a situation where the
assessment of a classical evacuation through a prescriptive code is worthless.

The reason is simple: the fire safety assessment using prescriptive codes is
designed for a case of fire, where all safety infrastructure is powered on and
working.

However, the attacker's very basic motivation is to change this situation.

8
Soft Targets and Pedestrian Simulations

For these reasons, it is necessary to perform a risk analysis, identify

possible attack scenarios, and determine the impacts of these scenarios.

It does not make sense to search for example the exact probability of a knife
attack, if it occurred in the object only exceptionally, or even not at all.

On the other hand, the same qualitative analysis is too simple to study the
impacts of these scenarios. The security manager should demand specific data
on the number of affected persons, an extension of evacuation time, reduction
of visibility due to smoke, etc.

Simulation tools are extremely suitable for these answers.

9
Prague
Main
Railway
Station

10
Risk Analysis

Risk identification (R) has the following phases:

• Asset analysis (A).

• Threat analysis (T).

• Vulnerability analysis (V).

• Determination of the resulting risk.

11
Risk Analysis

The risk assessment of asset, threat, and vulnerability values using semi-
quantitative approach.

Score Asset value (A) Threat value (T) Vulnerability value (V)
0 None or not rated Once every 10 years None
1 Low Once a year Low
2 Not very significant Once in half of year Not very significant
3 Medium Once a quarter Medium
4 High Monthly High
5 Very high Once a week Very high

R=AxHxZ
12
Quantitative risk analysis with their correlation

If possible, we should also express how the risks can interact with each
other. For this purpose, we can apply quantitative risk analysis with their
correlation.

Index 1 2 3 4
Index Risk High temperature Lightning Fall of a tree Icing
1 High temperature X 1 0 0
2 Lightning 1 X 1 0
3 Fall of a tree 0 0 X 0
4 Icing 0 0 1 X

13
Active and Passive Risks

Activity expresses the overall potential of the risk to cause additional risks.
On the contrary, passivity expresses that other risks that can cause a given
risk.

The following relations are used to calculate the given coefficients:

14
Active and Passive Risks

Activity and passivity of selected risks divided into segments

1,0
0,9 2 1
15
0,8
0,7
0,6
3 6
KpRi

0,5 5 7 14
0,4 4 1
13 2
0,3 10
0,2 9 11
12
0,1 4 3
8
0,0
0,0 0,2 0,4 0,6 0,8 1,0 1,2
KaRi

15
Prague Main Railway Station

The described analytical procedure was also applied in the case of Prague
main railway station.

The results risk analysis was used for the identification of 7 specific scenarios.
Examples include the following:

• Simple evacuation (all exits are available).

• An attack in the central part of the station (blocking of the central tunnel).

• An attack at the main exit, or the identification of an explosive device nearby.

16
Prague Main Railway Station

17
Movement Speed

Distribution of movement speed for different groups of agents

20 %
18 %
16 %
14 %
12 %
Frequency

10 %
8%
6%
4%
2%
0%
0,0 0,2 0,4 0,6 0,8 1,0 1,2 1,4 1,6 1,8 2,0 2,2 2,4 2,6 2,8 3,0
Movement speed [m/s]

Youths Adults Seniors

18
Pre-movement Time

Distribution of the pre-movement time of the agents

18 %

16 %

14 %

12 %
Frequency

10 %

8%

6%

4%

2%

0%
0 20 40 60 80 100 120 140 160 180 200
Pre-movement time [s]

19
Occupancy

20
Outputs: Total Evacuation Time

Total evacuation time for selected scenarios (Q95 quantile)

720

660
Total evacuation time [s]

600

540

480

420

360

300

240

180
1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Pedestrians

All exits opened Main entrance closed

21
Outputs: Maximum Density

22
Outputs: Maximum Time to Exit

23
Outputs: Maximum Usage

24
Outputs: 3D flyby with VR support

25
Workflow Automation

We analyzed a total of 7 different scenarios, each with a capacity of 1,000

to 10,000 people, for each occupancy tens of simulation runs with random
input parameters. This means several thousands of individual simulations.

We use R scripts to change the Pathfinder Input File to achieve the following
changes in particular: a number of agents, parameters of agents, random seed,
open or close the doors.

The results have been processed using the same approach.

Simulations were deployed on dual CPU Intel Xeon with 24 cores to use the
multi-threaded capability of Pathfinder.
26
Conclusion

27
Thank you for your attention
tomas.apeltauer@recognity.cz
tomas.apeltauer@vut.cz

28