Optimization Via Simulation

Marker / Edwards

Facility Evacuation Industrial Engineers

"Optimization Via Simulation"

812 South 2nd Street
Canon City, CO  81212
(719) 275-2558
mheinen_1@msn.com
www.markeredwards.com

  A facility evacuation example

          
Consider the evacuation of the small  facility (a school) shown  to the right having 24 rooms.  Suppose a scenario    exists where a threat  exists to the Northwest of the building requiring an evacuation to the Southeast.  An evacuation plan is suggested which utilizes the three exits  to the East and South (color coded evacuation routes A, B, and C).   
Note: a series of doors (or other obstacles such as turns, stairways, etc.) impede the flow of the evacuation.

An administrator/manager needs to determine if this plan is feasible.   Specific answers are needed for:

  1. Evacuation time for each route
  2. Effect of bottlenecks along each route (caused by obstacles such as doors)
  3. People density along each route as the evacuation progresses.
  4. Should rooms be evacuated along different routes? 
  5. Which rooms should be delayed to prevent hallway crowding?

A Marker / Edward facility evacuation simulation analysis will answer these questions and more!

Simulation Runs

For the facility shown above, a deterministic analysis will be conducted for Evacuation Route A (only) consisting of four simulation runs (1 - 4) shown below. 

Initial (Run 1) facility data for the simulation is shown to the right.              

Listed in the table are all entries into the Evacuation Route  A,  beginning  with the exit doors (obstacle OB1).  Rooms and sub  branches  (SB1 and  SB2) are included in ascending distance from the  exit.   Rooms on the  sub branches are listed at the end of the table.

Pertinent data include people in each room, their exit velocity,  delay exiting  the room, and obstacle flow rates.  Not shown in the table are parameters  input directly to SIM-VAC.  These include passageway widths, maximum  passageway velocity, and maximum "packed" density of people as they move  in the passage ways.

The table shows a deterministic analysis (constant parameters)  but a stochastic analysis is available.  A stochastic analysis induces randomness and has applicable parameters (mostly velocities and times) whose values are derived from given distributions (uniform, normal, exponential, weibull, etc.)  The result of a stochastic evacuation route analysis will be a probability distribution function.  A manager/administrator can use this function to compute the  probability  the evacuation will be complete in a given time.

Simulation runs for Evacuation Route A

All simulation runs were conducted at:

Maximum hallway exit velocity = 6 feet/second
Maximum "packed" density = 4 ft^2/person
All passageway widths = 10 feet

Simulation Run 1     

Red   = Maximum packed density of 1 person / 4 square feet                                                       
Parameters: No room exits are delayed
Analysis:  Evacuation time = 202 seconds. 

  The SIM-VAC contour plot to for Run 1 indicates:                 

a 20 foot deep crowd at the exit doors beginning at approximately 25 seconds after the evacuation begins and lasting the entire evacuation.
the hallway juncture between the 2nd and 3rd set of doors is crowded from 5 seconds until 135 seconds
the hallway upstream from the 3rd set of doors is backed up approximately 25 feet beginning 20 seconds after the evacuation until 90 seconds after the evacuation.

Conclusions:

The doors at the exit permit 2 persons per second to exit.  Reduce crowding upstream of theses doors by reducing over all flow.  The 2nd and 3rd set of doors (with Sub branches 1 and 2 exiting onto the Main branch between the doors) is causing a major delay upstream of this vicinity.

Recommendations:

Assume the same rooms stay on this evacuation route.  Try and make this evacuation scenario work.
Delay the exiting of Room 104 on Sub Branch #1 120 seconds.
Delay the exiting of Room 107 on the Main Branch 60 seconds.
Rerun the simulation with Rooms 104 and 107 delayed.

Simulation Run 2                                              

Red   = Maximum packed density of 1 person / 4 square feet

 Parameters: Room 104 delayed 120 seconds, Room 107 delayed 60 seconds

Analysis:  Evacuation time = 202 seconds (no change from Run 1).

  The SIM-VAC contour plot for Run 2 indicates:      

a 20 foot deep crowd at the exit doors beginning at approximately 25 seconds after the evacuation begins and lasting the entire evacuation.
the hallway juncture between the 2nd and 3rd set of doors is crowded from 10 seconds until 75 seconds
the hallway upstream from the 3rd set of doors is backed up approximately 20 feet beginning 25 seconds after the evacuation until 65 seconds after the evacuation.

Conclusions:

  The doors at the exit permit 2 persons per second to exit.  Reduce crowding upstream of theses doors by reducing over all flow.  The 2nd and 3rd set of doors (with Sub branches 1 and 2 exiting onto the Main branch between the doors) is causing a minor delay upstream of this vicinity.

Recommendations:

Assume the same rooms stay on this evacuation route.  Try and make this evacuation scenario work.
Reduce overall flow at the exit doors by delaying Room 102 180 seconds.
Delay the exiting of Room 104 on Sub Branch #1 120 seconds.
Delay the exiting of Room 107 on the Main Branch 60 seconds.
Rerun the simulation with Rooms 102, 104 and 107 delayed.

Simulation Run 3                                               

Red = Maximum packed density of 1 person / 4 square feet

Parameters: Room 102 delayed 180 seconds,

Room 104 delayed 120 seconds, Room 107 delayed 60 seconds

Analysis:  Evacuation time = 239 seconds (37 sec. increase from Run 2).

 The SIM-VAC contour plot for Run 3 indicates:        

a 5 foot deep crowd at the exit doors beginning at approximately 30 seconds after the evacuation begins and lasting until approximately 135 seconds.
Beside the empty period from 20 to 25 seconds, 10 and a 20 second empty periods exist at the exit doors
the hallway juncture between the 2nd and 3rd set of doors is crowded from 10 seconds until 75 seconds
the hallway upstream from the 3rd set of doors is backed up approximately 20 feet beginning 25 seconds after the evacuation until 65 seconds after the evacuation.

Conclusions:

The doors at the exit have had flow reduced too much as indicated by the 10 second and 20 second empty periods.  The 2nd and 3rd set of doors (with Sub branches 1 and 2 exiting onto the Main branch between the doors) is causing a minor delay upstream of this location.

Recommendations:

Assume the same rooms stay on this evacuation route.  Try and make this evacuation scenario work.
Change overall flow at the exit doors by delaying Room 102 a reduced 150 seconds (this will shift the flow 30 seconds to the left).
Delay the exiting of Room 104 on Sub Branch #1 110 seconds (this will shift the flow 10 seconds to the left).
Delay the exiting of Room 107 on the Main Branch 60 seconds.
Rerun the simulation with Rooms 102, 104 and 107 delayed.

Simulation Run 4                                               

Red   = Maximum packed density of 1 person / 4 square feet 

Parameters: Room 102 delayed 150 seconds,

Room 104 delayed 110 seconds, Room 107 delayed 60 seconds

Analysis:  Evacuation time = 208 seconds (6 seconds more than

when no rooms are delayed).

 The SIM-VAC contour plot for Run 4 indicates:                

 

a 5 foot deep crowd at the exit doors beginning at approximately 30 seconds after the evacuation begins and lasting until approximately 135 seconds.
Beside the empty period from 20 to 25 seconds, only a 5 second empty period exists at the exit doors
the hallway juncture between the 2nd and 3rd set of doors remains  crowded from 10 seconds until 75 seconds
the hallway upstream from the 3rd set of doors is backed up approximately 20 feet beginning 25 seconds after the evacuation until 65 seconds after the evacuation.

Conclusions:

  The exit flow is near optimal.   No significant crowding occurs at the exit doors.  The 2nd and 3rd set of doors (with Sub branches 1 and 2 exiting onto the Main branch between the doors) is causing a minor delay upstream of this location.

Recommendations:

Assume the same rooms stay on this evacuation route.  Try and make this evacuation scenario work.
Delay the exiting of Room 102 by 150 seconds.
Delay the exiting of Room 104 on Sub Branch #1 by 110 seconds.
Delay the exiting of Room 107 on the Main Branch by 60 seconds.
If crowding is unacceptable vicinity of the 2nd and 3rd doors, reduce flow by delaying Room 106, 108, or 109 by 20 seconds.

Final recommendation for Route A is given in the recommendations for Run 4.


Back to Sim-Vac home page

See our price list