2021.2 Patch Release – Available Now

Posted on May 26th, 2021

One of our Reseller Partners identified an issue when dragging items in the navigation view. When this happens, the application freezes and requires a forced shutdown to proceed. This issue was introduced in the 2021.2.0512 release of both PyroSim and Pathfinder. 

We have identified the problem, and have uploaded patch releases that addresses the bug. The updated versions of PyroSim and Pathfinder are available now.

… Read More

Improvements in Pathfinder 2021.2

Posted on May 10th, 2021

Pathfinder 2021.2 introduces queueing as a new behavior which can greatly expand modeling options and capabilities. New user interface enhancements make it quicker than ever to find objects and edit properties.
… Read More

Proximity Analysis and the Thunderhead development process

Posted on April 29th, 2021

The upcoming release of Pathfinder Results version 2021.2 will include a Proximity Analysis feature based on the paper EXPOSED: An occupant exposure model for confined spaces to retrofit crowd models during a pandemic by Enrico Ronchi and Ruggiero Lovreglio. In the paper Enrico and Ruggiero lay out a method to create metrics that measure occupant exposure in microscopic scale environments, which enables evaluation of the efficacy of infection mitigation measures. In simpler terms, they supply a way to measure how effective social distancing measures are at preventing the spread of disease. This post will lay out our development process to create this and other features and summarize the proximity analysis feature as implemented in version 2021.2.

The Thunderhead development process

At Thunderhead Engineering, we have periods of time that we call “Project Phoenix” where developers can work on any project that is interesting to them, whether it is a marketable feature or just a fun research project. During our Project Phoenix iteration last June, two of our interns worked on a project involving tracking the spread of an infectious disease in a Pathfinder model. This initial project modified the Pathfinder simulator to produce a simple SIR-like model during occupant movement.

After this project was completed, our team discovered Enrico and Ruggiero’s paper. After speaking with Enrico and showing him their prior project, our interns and the rest of the development team decided to modify the project to reflect the paper’s suggestions and implement it as a new feature in our Results viewer. After 3 iterations of the project, and some help from the rest of our development team, the proximity analysis feature is ready for release.

What is Proximity Analysis?

Unlike the first iteration of the project, the Proximity Analysis feature shipping in the upcoming 2021.2 release does not modify the Pathfinder simulator. Instead, it post-processes the output of Pathfinder simulations in our Results program. The Results Viewer generates a proximity analysis report that provides various metrics regarding occupant exposure, including the amount of time that k number of occupants are within a proximity radius of another occupant, how many occupants are violating the proximity radius at any given point in time, and a global exposure time metric. Because this feature is part of Results viewer and does not modify the simulator, users will not need to re-create or re-run simulations to generate Proximity Analysis reports. They can be generated using the results from any existing Pathfinder simulation.

Example Proximity Tracing Graph
Example of one of the graphs output by the new Proximity Analysis feature of Pathfinder Results 2021.2

How would Proximity Analysis be used?

One case where proximity analysis is particularly useful is when evaluating social distancing measures. By performing a baseline/control simulation, changes made to the Pathfinder model can be evaluated by their effects on the global occupant exposure metric to determine if the changes are effective in enforcing social distancing. This can also be used to compare different social distancing measures against each other to determine which ones are the most effective.


Our team enjoyed creating this new Proximity Analysis feature, and we look forward to continuing to improve our products. You will be able to access Proximity Analysis in the upcoming 2021.2 release. If you have any suggestions for features you would like to see in our products, you can email them to our Support team at and we will consider them.


Ronchi, Enrico, and Ruggiero Lovreglio. “EXPOSED: An Occupant Exposure Model for Confined Spaces to Retrofit Crowd Models during a Pandemic.” Safety Science, vol. 130, Oct. 2020, p. 104834.,

Queues in Pathfinder 2021.2

Posted on April 23rd, 2021

Queues introduce an entirely new concept of occupant movement to Pathfinder which is restricted by pre-determined Paths and waypoints called Services. This will allow users to better simulate behaviors at airports, restaurants, amusement parks, and more. … Read More

User Experience Improvements to Pathfinder 2021.1

Posted on February 23rd, 2021

Pathfinder will also benefit from feedback from our users. This release focuses on incorporating their suggestions in to our products to improve the user experience. We have made several usability improvements to Pathfinder in the 2021.1 release, some of which are mentioned below.

Update: The 2021.1 Pathfinder Release is available now

Added copy/paste functionality to Occupant Profiles

Expanding upon the functionality that we added to copy and paste geometry in Pathfinder 2020.1, we added the ability to copy and paste Occupant Profiles between separate Pathfinder instances. This means that users will no longer have to re-define the same profile in separate models, and can instead copy the profiles from one model to another. This is especially useful if users have several different profiles that are commonly used across a large number of models.

Added as-you-go undo/redo support for the Add a Polygonal Room tool

Similar to the changes to the wall and slab tools that we mentioned in our post about User Experience Improvements to PyroSim in 2021.1, we have added as-you-go undo/redo support for the add a Polygonal Room Tool.

Now, if a user makes a mistake when drawing a complex room shape, they will no longer have to start drawing that room from the beginning. They will have the ability to undo/redo any of the points they have drawn. This is especially useful for users that commonly draw their Pathfinder models on top of a 2D floorplan. You can see a short gif of this feature below.

Enhanced Monte Carlo user manual now available under the “help” menu

The Monte Carlo utility packaged with Pathfinder can help create models with randomized occupant positions and properties. This allows users to create multiple randomized copies of a model, run them sequentially, and process the resulting data set for research or regulatory requirements. The User Manual for our Monte Carlo command line utility is now available both online and in the application help menu.

Scripted Screenshots

We recently added the ability for users to interact with the Results viewer of both Pathfinder and PyroSim to generate a series of screenshots based on views that are defined in the Results view. This is particularly for users that generate reports from a series of commonly used views, as it removes the tedium of manually creating these screenshots in the Results UI. You can read more about this feature in Section 2.16.4 of the Results User Manual.

Monte Carlo Improvements in Pathfinder 2020.5

Posted on November 13th, 2020

Since Pathfinder 2018.3, users have been able to take advantage of Monte Carlo simulations through a command line interface. In modeling, the Monte Carlo method is used to randomize inputs across many simulation instances. With a statistical analysis applied, users can create probablistic interpretations of the resulting data sets. It is often used in research and is also required by regulations such as RiMEA. … Read More

Mars Colony Biome Evacuation Analysis by Szymon Matkowski

Posted on August 11th, 2020

Our support team has been assisting Szymon Matkowski with his recent project based on a feasibility study, completed by a team at Wroclaw University of Science and Technology, of possible layouts and technical solutions for a theoretical base on Mars. The study was part of a competition sponsored by The Mars Society in which the team at Wroclaw won second place. In Matkowski’s study, he analyzes evacuation of a Mars colony biome under threat of decompression or fire.

Diagram of Twardowski Colony Biome Evacuation
An overview of Matkowski’s evacuation plan based upon the Wroclaw team’s design.

Safety of the people is a top technical priority in every building or device – in any place. All: structure, MEP systems and the layout should be focused on minimizing the risk of life loss. Design of a base, or colony on other planet also should comply with this principle. The layout of all safety egress communication (passages, corridors, staircases, lifts etc.) should not be changed or altered. The unchangeable layout of those elements is a bone system for the functional layout and all the decoration we see. The drawback of such solution is that – once designed the layout is frozen. Any change leads to re-examination of all safety measures – which is very expensive (even in Earth’s conditions). One of the most important coefficients is evacuation time. Time – which people have to save their lives. This blog entry describes my research on Evacuation movement speed on Mars and a sample evacuation time count. Depending on the threat (decompression, fire and smoke) additional means should be provided : smoke exhaust, additional air supply, sealed safety rooms, emergency teams access routes and action plan. For the purpose of the investigation I have used a fragment of Twardowsky Mars Colony. Twardowsky won 2nd prize in 2019 competition organized by Mars Society. You may see the presentation here. Also the competition entry of the Base was published in a book.

During Matkowski’s analysis, which he completed in less than two months as his first Pathfinder project, many challenges were addressed. This included approximating the speed adjustment with Mars gravity and with various space suit equipment, behavior adjustment upon triggering events such as decompression, and the simulation of airlocks with pressurization phases. The Thunderhead Engineering support team aided Matkowski throughout these challenges which led to his successful analysis.

You can read Matkowski’s full analysis for free on the website.

Social Distance Improvements in Pathfinder 2020.3

Posted on August 7th, 2020

Pathfinder 2020.3 further enhances the user’s ability to model social distance behavior as a method to address pandemic safety. Social Distance has been completely separated from Comfort Distance. Now, rather than Enable Forced Separation in the Simulation Parameters (retained in 2020.3 for legacy files), Social Distancing has its own selection field in the Advanced tab of the Edit Profiles dialog. With this recent development, occupants exhibit less “jostle” or “jitter” while smoothly and effectively maintaining distance. … Read More

New Pathfinder Data Options to Assist with Coronavirus Scenario Modeling

Posted on April 7th, 2020

* The features described in this article are now available in Pathfinder 2020.2 and later versions.

The global spread of the SARS-CoV-2 virus and associated COVID-19 sickness has profoundly changed human interactions in a matter of weeks.  It will likely be several more weeks until the first wave of illness is under control. 

Modeling how these interactions affect our environment will provide insight into what can be done to improve public safety now and in the future. Thunderhead Engineering has implemented additional analysis tools in Pathfinder to assist with evaluation of various human movement scenarios.  

Discussions with current customers that perform crowd modeling led us to focus initially on improved data output and visualization related to the minimum distance between agents in a simulation.  Maximizing the closest distance between occupants and minimizing the total number of occupants encountered has the potential to reduce overall risk when comparing alternative layouts or procedures for a space.

These changes are the beginning of rapid development in Pathfinder intended to help model and visualize important emerging scenarios:

  • Immediate changes to public spaces will need to be vetted quickly to determine the best methods of reducing transmission.
  • After the first wave, governments, factory managers, and administrators of public spaces will need tools to evaluate risk levels and exposure as these spaces are slowly reopened at reduced capacity. 
  • Increased physical distance within commercial and public spaces will likely be needed for several months. 
  • As pockets of illness reappear, space planners will want to be prepared with alternate plans that can minimize virus transmission while avoiding widespread full closures of facilities.

Pathfinder Changes

Output Data for Social and Physical Distancing

In this preview build, Pathfinder.exe supports a new command-line parameter:
(2.0 is a value in meters that can be adjusted to control some of the following features)
This parameter activates writing of the following new data files:
  • CSV-format output file modelname_sd_transient.csv  containing transient data about occupant separation. For every time and occupant in the simulation, the closest other occupant is listed along with distance.  Additionally,  the number and ID of occupants within 1m, 2m, and 3m are also listed.  The presence of the social_distance command line parameter is required for this output to be generated, however adjusting the numeric value of social_distance has no impact on this output.
  • CSV-format output file modelname>_sd_accumulated.csv containing accumulated exposure data for each occupant.  For every occupant in the simulation, the occupant who has spent the longest time within r=social_distance is listed, including the occupant ID and the total amount of time.  Additionally a count is made of all occupants who have remained within social_distance of the occupant for more than 1 minute and 5 minutes, including the occupant IDs.

Graphical Depiction of Social/Physical Radius

When Pathfinder.exe is run with the social_distance command-line flag, Results will display a disk at the base of each occupant whose radius is social_distance meters. In addition, a Debug menu will appear on the menu bar enabling the use of the Edit Debug Options… dialog where you can edit the value of Social Radius (note: set the value to zero to disable this setting).

New Occupant Contours to Quantify Exposure

We have implemented two new Occupant Contours in the Pathifnder Results viewer that may help evaluate physical distance between occupants — currently named Social Linkage and Social Usage.  
  • Social Linkage uses a given radius (R) to calculate how many other agents are within that radius and plots that as the value drawn inside a circle of radius R centered at the occupant.  Higher values display on top of lower values.  We are also considering displaying this measure as a per-occupant color for greater clarity.
  • Social Usage is similar to a density contour and is referenced to the contour mesh.  For every point on the mesh (conceptually a spot on the floor), the number of agents within radius R are calculated and assigned to that mesh point.  The mesh of all values is then contoured.  This will give the high values between the agents rather than right at the agents.

It is also possible to use the existing Density or Usage contours in Pathfinder to show different data if you consider how the calculation is done and what scale is used.  Since the Usage contour assigns a value that ranges linearly from 1 at the agent location to 0 at radius r, it can be used to draw range rings.  In this example with r=3 and three color bar segments, the green circle represents 2m from the occupant and red shows 1m.

Additionally,  all contours can be passed through filters to display average or maximum values over time.

Help with Pathfinder Preview Feedback

* The features described in this article are now available in Pathfinder 2020.2 and later versions.

These preview features are preliminary and are subject to change before an official release of Pathfinder.  However, our goal was to make these updated data options available as soon as possible for anyone evaluating procedures and space usage during the current crisis.  We encourage anyone interested to contact and try out these new features.  At the present time, we are prioritizing feature development in Pathfinder that focuses on improving simulation or data analysis for social distance modeling.  Please let us know if you have requirements or suggestions to help us build a tool that can help improve public health and reopen buildings in a responsible manner.

Vertical Ladders in Pathfinder

Posted on April 1st, 2020

We recently had a question about modeling vertical ladders in Pathfinder. This post will describe the details of how to accomplish this. This question came from Saqib Mehmood of Aalborg University in Denmark, who was modeling wind turbine evacuation. Saqib Mehmood was joined on this project by a team including Anders S. Kristensen, Francisco Manuel Dieguez Barreiro, Hana Softic, and Yannan Zhou.

Vertical ladders imported into Pathfinder to model wind turbine evacuation.


Figure 1 shows normal stairs. Each end of the stairs must connect to boundary edges of the rooms. If the stairs connect the interiors of the rooms, cutouts at the top and bottom must be used, Figure 2. The top cutout must be larger than the occupant radius and stairs cannot be truly vertical, there must be some slope. The walking surface is naturally defined as the top surface of the stairs.

Figure 1: Normal stairs connecting two rooms.
Figure 2: Cutouts required for a stairs that connects center of rooms.

Pathfinder does not have ladders, but ladders can be represented as steep stairs. Figure 3 shows ladders connecting two floors.

Figure 3: Model of ladders connecting two rooms.

In this model, the distance between floors is 20 m. The cutouts in the top and bottom floors are vertically aligned and measure 1 m by 0.5 m. Because the ladders connect to opposite sides of the holes, there is a slight slope to the ladders. The walking surface of the ladders is the face of the ladder whose normal has a small positive Z component. 


We want to specify occupant speed on a ladder. By default Pathfinder modifies the speed on stairs based on the approach described in the SFPE Engineering Guide – Human Behavior in Fire, 2003. These values are not appropriate for a vertical ladder, so we will use a stair speed modifier to obtain the desired value.

As described in the Pathfinder Technical Reference section Path Following in SFPE Mode, the occupant base speed, v_{b} is given by:

    \[v_{b}=v_{max}\star v_{f}(D)\star v_{ft}\]

where v_{max} is the occupant’s maximum speed entered in the user interface as Speed, v_{f}(D) is the density speed modifier, and v_{ft} is the terrain speed modifier.

v_{ft} is defined as 


where k=1.4 is the value for level terrain. For stairs, the value of k depends on the stair riser and tread that is input in the Pathfinder user interface, Figure 4. Note that k is not calculated using the physical stair geometry, it is calculated based on the user input. For our purposes it is sufficient to know that a stair with a rise of 19.0 cm and a tread of 25.4 cm gives a value k=1.0.

Figure 4: Stair parameters.

On steep stairs the density is low so the density modifier v_{f}(D)=1.0. For a stair with rise of 19.0 cm and tread of 25.4 cm k=1.0, so the occupant speed is given by

    \[v_{b}=v_{max}\star \frac{1}{1.4}\]

We can multiply v_{b} by a stair speed modifier to obtain the desired ladder speed

    \[v_{ladder}=modifier\star v_{b}=modifier\star \frac{v_{max}}{1.4}\]

By default Pathfinder v_{max}=1.19 m/s and if we desire a ladder speed of 0.25 m/s, then the speed modifier=0.294.


We run the example using the stair parameters shown in Figure 5. We also select some occupants, click the More button, and enable csv output of individual data.

Figure 5: Stair parameters used in the simulation to give a ladder speed of 0.25 m/s.

Figure 6 shows the initial positions of the occupants and Figure 7 shows the occupants descending. 

Figure 6: Preparing to descend.
Figure 7: Descending. Note that the occupants face in the direction in which they entered the stairs.

We can open the csv file for the ladder with the single occupant and see that the speed on the ladder is 0.25 m/s. Similar results are obtained for the ladder with multiple occupants, with some initial stop/start movement as they adjust on the ladder.

Figure 8: Speed of descending occupant.

Additional Comment: In this example, I used the stair speed modifier to control the ladder speed. An alternate approach would be to change the speed/density curve for stairs in the occupant Profile. This has the advantage of allowing a single profile to represent different speeds up and down the ladder. It has the disadvantage that it affects speeds on normal stairs also.

Download File

You can download the Pathfinder model here.

ladder post