Verification and Validation of Viswalk for Building Evacuation Modelling

Abstract: This thesis is evaluating the pedestrian modelling software Viswalk for the use as a building evacuation model, by verifying and validating the model. In the verification, a procedure from the National Institute of Standards and Technology (NIST) is used as a basis to assess Viswalk’s ability to represent pre-evacuation time, movement and navigation, exit usage, route availability and flow constraints. Seven tests are excluded due to delimitations of the thesis or limitations of the current version of the model. The verification tests show that Viswalk is able to represent the main core components of evacuation models that are under consideration. The model yields results that correspond with the expected results for all 10 verification tests that are performed. However, non-conservative flow rates can be obtained if the default input settings are used. In the validation, results from Viswalk are compared to four real life experiments including a corridor, a classroom, a theatre lobby and a stair, followed by an uncertainty analysis. With adjusted input settings the movement times deviate with 2-16 % from the experiments and with default input settings the movement times deviate with 12-95 %. The walking speed is an important parameter in the validation tests, even with substantial congestion, with up to 46 % increased movement times when the walking speeds are decreased with 25 %. In the validation it is also noted that the occupant densities in front of openings can differ with up to 45 % between the simulations and the experiments. Despite the aspects described above, results that are close to experimental results can be obtained if the user has a good estimation of the occupant demographics and is aware of the limitations of the model.