Thunderhead Engineering Forum

Please login or register.

Login with username, password and session length
Advanced search  


Forum moved to

Author Topic: Simulation mesh, how does it work?  (Read 1550 times)


  • New Member
  • *
  • Posts: 1
    • View Profile
Simulation mesh, how does it work?
« on: September 05, 2016, 04:27:35 am »


We are carrying out some evacuation simulations in a theatre and it comes out the question, how does it work the mesh generator and how sensitive is the model to the grid?

I understand that the Snap Grid is just a tool to make it easier the modelling process but it doesn't have any implication on the calculation algorithm. I am working with a 10 cm grid and therefore I assume that if I am modelling a 175 cm door the programme is capable of understanding the object and it wont just reduce or increase the object to a fitting one (that is either a 170 cm or a 180 cm door)
Is my assumption right?

I understood (hopefully right) that the factor controlling the grid generation are found in the Tab Simulation-->Paths, but I have no idea how sensitive is the model and how should I define these factors. I understand what the factor "Max Agent Radius Trim error" means and I could using some common-sense define this parameter, but I have no clue how should I set the values for "Max Edge Length" and "Min Angle" and how sensitive could be the model. I am currently working with the default value.

Are there rules to define these factors like in FDS to define the proper mesh edge? Does this selection affect the model elements definition (for example the previous door example)

Thanks for your help


Charlie Thornton

  • Thunderhead
  • *****
  • Posts: 851
    • View Profile
Re: Simulation mesh, how does it work?
« Reply #1 on: September 20, 2016, 09:28:29 am »

You are correct about the snap grid - it impacts only the drawing tools. You can change the resolution of the grid back and forth between 10cm and 5cm (and any other convenient dimension) as needed to help you quickly add geometry to the model. Ultimately, the created objects have their own geometry that is shown when the object is selected and this is what is passed on to the simulation.

The max trim error defines the bin size we use when calculating cached door to door paths in situations where there are different sizes of occupants. In these situations, there may be places where little occupants can pass through an opening and big ones cannot. Since the distance to goal beyond the current door target is re-usable information, Pathfinder caches the result for different sizes of occupants. If the trim error is 2.54 cm, then an occupant whose radius is 40 cm might not use a path through a 41 cm opening because it's sharing the cached data used by occupants with a shoulder width of 42 cm.

The constrain edge length options are there for users that want to micro manage the simulation mesh. I've actually never seen an advantage to doing this since our algorithms work well with long edges and more triangles make pathfinding that much slower. If you find a situation where it helps, I'd love to hear about it!