Evacuation time analysis

All topics related to questions in Pedestrian Dynamics
marlies
Posts: 301
Joined: Monday 17 January, 2011 - 09:28

Evacuation time analysis

Post by marlies »

The attached models shows an example of how to analyze evacuation times in a Pedestrian Dynamics model.

Model explanation:
- The model contains a simple route from entry to waiting area to exit.
- At time = 30 minutes the action timer (which can be found to right in the model)s triggers ALL agents to go to the exits. At this moment the 2 emergency exits at the bottom will be available as well.
- The model contains a dummy waypoint location just below the action timer. Agent will not visit this location physically, but it will be used for output analysis. The waypoint activity location has ID 6.
- Now have a look at the settings of the action timer. Tab General > Setting Event Trigger.
- Click in this field: at the start of the evacuation there will be a logging for ALL agents in the model as if they visit at that moment the dummy waypoint (look for ID = 6) in the code. So, for all agents, no matter where they are in the model, there will be a time stamp logging at start of the evacuation.

Now let's have a look at the output:
- Go to the menu Simulate > Wizard and click Start Experiment. This will run a simulation scenario of a 35 minutes simulation.
- When the scenario is completed go to the menu Results > 2D Output.
- Click on the button Show output of activity route, which looks like this:
Capture.JPG
Capture.JPG (8 KiB) Viewed 64882 times
.
- There is one route: EvacuationToExit. Select it and click Show Graph.
- This will show an evacuation time histogram for the simulation which will look like this:
Capture3.JPG
Capture3.JPG (104.09 KiB) Viewed 64882 times
- The legenda to the right shows mean/min/max evaucation times.
- By default the graph shows the aggregate values. Please note that by selecting other rows in the table below the graph it is possible to analyze the evacuation time for each exit separately.

How to create this kind of graph:
- Close the graph window.
- In the activity route output window: select the route and click Edit.
- In the tab Route Definition. See that the following route was defined: from Waypoint (the dummy logging at start evacuation!) to all Entry_Exits.
- In the tab Route Settings. See that the setting Break Down Row is set to 2. This means that the graph will not only show the aggregate results, but will also contain all the information for each location in step 2. This means: it is possible to analyze the output for each exit separately.
Attachments
20160229_EvacuationTimes.mod
(177.82 KiB) Downloaded 637 times
ie_1989
Posts: 58
Joined: Saturday 21 November, 2015 - 06:50

Re: Evacuation time analysis

Post by ie_1989 »

Does each layer need a distinct waypoint activity location for collecting the time of entry/exits located on it? For example I have 6 layers each of which has its own entry/exit location. So do I need to put a waypoint activity location on each of them?
marlies
Posts: 301
Joined: Monday 17 January, 2011 - 09:28

Re: Evacuation time analysis

Post by marlies »

Hi,

No, just one is enough. What you use the waypoints for is logging a time stamp, this time stamp can be used to calculate the travel times. The height layers isn't important for that time stamp and you can use just 1 waypoint.

Regards, Marlies
smith
Posts: 9
Joined: Wednesday 09 March, 2016 - 12:08

Re: Evacuation time analysis

Post by smith »

Hello!

I can not open the attatched file with my pedestrian dynamics software! Some compile errors ... .

I would be really nice if somebody can write me the peace of code (4d script) from the settings of the action timer and the waypoint!
marlies
Posts: 301
Joined: Monday 17 January, 2011 - 09:28

Re: Evacuation time analysis

Post by marlies »

Hi,

That's because of the PD version. I used PD3.0 to create the example. Please find the PD2.0 model version attached to this post.

The only thing that won't be applicable compared to my previous post is the option to get the results per separate exit automatically in one output graph. You'll get the results for all exits or you'll have to select each exit separately.

The dummy logging and use of it in the outputs, works the same.

Hope this helps!

Regards, Marlies
Attachments
20160309_EvacuationTimes_PD2.mod
(88.42 KiB) Downloaded 616 times
smith
Posts: 9
Joined: Wednesday 09 March, 2016 - 12:08

Re: Evacuation time analysis

Post by smith »

Thank you!! Now it works fine! :-) Thanks again!
ie_1989
Posts: 58
Joined: Saturday 21 November, 2015 - 06:50

Re: Evacuation time analysis

Post by ie_1989 »

Hi. I have a question about the use of waypoints for collecting statistics. Is there anyway to get the statistics of a special group of agents without a way point? I have attached two files (1) and (2). In both files I have defined two activity routes in the Activity Route Output. The first one uses a waypoint to collect the statistics of All profiles after the Action_Timer starts. The second one, named "Default_Route2" is set to collect data of a special agent profiles.
The statistics of the second route is totally wrong. More specifically, the dispersion of the Distances of agents is not logical. The specified agents all follow the same path in the model. But they have a minimum distance of 120 and a maximum of 1200 m.
By the way, the difference between to attached files is that in "2" the layer (ID=88) in which my specific agents are generated is longer than that of "1". The statistics of the agents are clearly different (The number of agents becomes lower as the length of the layer increases). Why does such a thing happen?
Attachments
2.rar
(222.64 KiB) Downloaded 556 times
1.rar
(223.53 KiB) Downloaded 579 times
User avatar
Nienke Valkhoff
Posts: 20
Joined: Tuesday 11 January, 2011 - 09:23

Re: Evacuation time analysis

Post by Nienke Valkhoff »

Hi,

You can indeed get statistics of a special group of agents without using waypoints. Like you did in the model for the second output route you can indeed use a filter to select a specific profile ID.

The strange distances are caused by the way the agents with the special profile are placed in the model. After 5.5 minutes a 1000 agent are placed in the model at the same time in a very narrow hall on a small Entry\Exit Area. When you do a single run can see that agents are pushed outside of the hallway. The pushing add extra distance for the agents.

To solve this you need to make sure that the agents can be placed in the model by changing the rate such that the hallway and flight of stairs can handle the agents without them being pushed outside of the hall.

I have adjusted the generator for the special profile agents and increased the run length of the experiment to 60 minutes so that I get the results of all 1000 special profile agents. See
2 adjusted.zip
adjusted model 2
(447.61 KiB) Downloaded 606 times
For the special profile agents I used the following generator settings.
Repetitive mode: continuous
Max agents: 1000
offset time: 1.5
delay time: mins(5.5)
In the arrival list I have 1 row
Creation time 0
Nr of agents 1

This means that each 1.5 seconds one agent is placed in the model until 1000 agents are placed in the model.

Using above setting the distance output is between 130 and 165 meter.

Regards,

Nienke
ie_1989
Posts: 58
Joined: Saturday 21 November, 2015 - 06:50

Re: Evacuation time analysis

Post by ie_1989 »

Thank you very much
ie_1989
Posts: 58
Joined: Saturday 21 November, 2015 - 06:50

Re: Evacuation time analysis

Post by ie_1989 »

I run the model. Unfortunately it shows the statistics of only 330 of agents. How can I fix it?
Post Reply