Back

Advanced Evacuation Simulations

Evacuation simulations are an advanced option for performance based building design enabling a deeper insight into your buildings crowd capacity: These simulations are capable of highlighting pedestrian traffic bottlenecks and ensuring safe escape routes in case of an emergency.

By Christian Kohler. This article is intended for users of such software packages as well as for beginners or regulators. I will try to give a short insight into some new features of the evacuation software Pathfinder developed by Thunderhead Engineering.

  • Path Planning (exit selection or route choice). In an evacuation simulation, much depends on the path planning. Which path is the fastest? Will this path be the most common used? The software generates the path planning automatically. How some parameters will change this Path Planning will be discussed.
  • Speed Density Profiles. Crowd densities are the most important evaluation parameters of an evacuation simulation. The tighter an area is filled with people, the more dangerous it is. With the function of speed density profiles, the walking speed in the range of higher crowd densities is adjusted automatically which gives you a more realistic result.
  • Coupled Simulations (ASET-RSET calculations): For performance-based design purposes an ASET-RSET-calculations (comparison of Available Safety Egress Time to Required Safety Egress Time) can bring you more insights than a simulation alone. The ASET-side is the fire simulation while the RSET-side represents the evacuation simulation. It is called a coupled simulation as both simulations are displayed in the same software.

Path Planning

The software Pathfinder uses a locally quickest path algorithm to solve an individual occupant’s steering behaviour [1]. It plans the route hierarchically, used on one hand local information of occupants or evacuees (hereinafter referred to as occupants) and on the other hand global information of the building [2]. It assumes that an occupant knows all the doors in the room and can identify all jam times at these doors. As a result, the occupant is free in the door choice (exit-selection) and may decide to go another route if he sees that the time to another door is shorter. This is referred to in queuing theory as jockeying. With two lines at a checkout, you will change your line, when you see that the other line will go faster.

FeuerTRUTZ International: Advanced Evacuation Simulations
Fig. 1: Results with standard settings. Maximum densities in case of an emergency. (Source: Software Version Pathfinder 2016 / Thunderhead Engineering USA, 2016)

One weakness is that the occupant receives information from the software about the shortest path in the complete building he obviously doesn’t know in reality. The occupant normally only knows the way which he came into the building. He lacks a global information of the building. The path behind an emergency exit door is not known to an occupant and can therefore in reality not be part of an exit choice selection. This also explains that it has been found that most people in an evacuation exercise use the exit which they know already. Only if they are encouraged to use the closest exit either audibly or through evacuation helper they will do so [3].
Basically, a factor known / unknown global information of the building would be a possible improvement for an evacuation simulation in the future as an occupant will tend to use the known path. In Pathfinder it is possible to approximate this theory with the Global Travel Time parameter.

FeuerTRUTZ International: Advanced Evacuation Simulations
Fig. 2: Results with customized settings for Global Travel Time, rest unchanged: Exit Time changed by 15 Seconds. Occupants also use other farther exits (Source: Software Version Pathfinder 2016 / Thunderhead Engineering USA, 2016)

This is demonstrated by the following example. In an event hall with 1,000 occupants, the emergency case has to be researched. The event hall is accessible by an L-shaped corridor (violet). There is one available emergency exit in the rear area and two additional emergency exits in the front. The total exit door widths totals just 10 m, which satisfies the requirement in Switzerland for 1000 people. Without changing the Global Travel Time parameter, most of the occupants use the front doors just because they know from the software after these doors they are outside [Fig. 1].
By changing the Global Travel Time, the occupants perform less jockeying and stay committed to a familiar exit. As a result the occupants use the exit in the L-shaped corridor more often, what seems to be more realistic (as this is the way in) and the Exit-time drops with this method [Fig. 2].

FeuerTRUTZ International: Advanced Evacuation Simulations
Fig. 3: No speed-density profile included: 150 m long 20 m wide corridor, 3200 Occupant, Start-density 4 Pers / m2: Evacuation time 3:28 min. (Source: Software Version Pathfinder 2016 / Thunderhead Engineering USA, 2016)

People Densities

The software Pathfinder allows you to enter a speed-density profile. This profile couples walking speed with the density of people (crowd density). Thus, the walking speed is not always constant, but varies depending on the crowd density which is more realistic: With a density of 3.25 persons / m2 the rate drops down to 15% of the original input value. Not only is speed-density profile an advanced option, it is also not usable in every evacuation software package.
The following example will illustrate this feature: 3200 occupants, highly crowded (density 4.0 persons / m2) walk along a long corridor (150 m long, 20 m wide). Fig. 3 shows the result with standard-settings and no speed-density profiles.

FeuerTRUTZ International: Advanced Evacuation Simulations
Fig. 4: Speed-density profile incuded Evacuation time 5:10 min. (Source: Software Version Pathfinder 2016 / Thunderhead Engineering USA, 2016)

Fig. 4 highlights the difference by using a speed-density-profile: A more realistic result with a speed-density-profiles is achieved with a significant higher evacuation-time (+ 1:50 minutes).

Coupled simulations

A coupled simulation is an evacuation simulation coupled with a fire simulation. This type of calculation is also known as an ASET-RSET calculation. ASET stands for Available Safety Egress time, RSET for Required Safety Egress Time. The ASET-side is the fire simulation side and gives a time until the tenability-criteria is critical. The RSET-side is the evacuation simulation and gives you the time an evacuation will take in a certain building. With Pathfinder, both ASET and RSET simulations can now be displayed in the same software: The tenability criteria from the fire simulation and the movement of the occupants simultaneously. As seen below [Fig. 5] a combined visualization produces a more realistic view, a better understanding of the incident itself and enables the user to easier see whether the safety objectives are fulfilled.

FeuerTRUTZ International: Advanced Evacuation Simulations
Fig. 5: Screen-shot of a coupled simulation: Fire development before the evacuation starts. (Source: Software Version Pathfinder 2016 / Thunderhead Engineering USA, 2016)

The following example [Fig. 5] of a three storey shopping-center in Switzerland demonstrates this feature. A fire simulation (fast developing fire of 100% polyurethane) was incorporated into the evacuation simulation. With this information it can be seen better how fast a person is at risk and whether an evacuation path is safe at a certain time.

Summary

The software Pathfinder can perform evacuation simulations on a very high scientific level and provide you with very useful information of a high crowded building. High crowd densities can lead to high risks in case of an emergency and should be avoided. With this software a building design analysis is possible on an early stage enables the best possible evacuation planning or analysing the weaknesses of an existing building. In the latest 2017-release of the software assisted evacuation is also included allowing for example wheelchairs displayed. This will lead to even more realistic crowd simulation enabling simulation of airports or hospitals.

Author

Christian Kohler: Fire Protection Engineer Zurich, Fire and Evacuation-Simulation Expert

Literature

[1] Kuligowski, Computer Evacuation Models for Buildings, SFPE Handbook of Fire Protection Engineering, 5th Edition 2016

[2] C. Thornton et al., New Wayfinding Techniques in Pathfinder and Supporting Research and Pedestrian and Evacuation Dynamics 2012

[3] M. Okaya and T. Takahashi, Effects of Guidance Information and Human Relations among Agents on Crowd Evacuation Behavior, Pedestrian and Evacuation Dynamics 2012

  • FeuerTRUTZ International

    This newsletter informs you about every new issue of “FeuerTRUTZ International”, our free e-magazine for fire protection engineers.

    Notes on data privacy
    Please pay attention to our general information on data privacy.