Since its inception in 1997, a number of validation studies for STEPS have been carried out, both by Mott MacDonald and others, to test both the qualitative and quantitative behaviour of the model.
STEPS validation has taken three main forms:
- Fundamental verification tests which check the software behaves as it is designed to e.g. flow down a stairway at a given flow rate.
- Comparison with accepted design codes, such as the National Fire Protection Association (NFPA) 130 code. While not strictly a “validation” these comparisons give confidence that the results predicted fit within expected norms. It is not always possible to force a microsimulation to conform to the assumptions inherent in a design code (for example full utilisation of exits) however the comparisons have given very promising results.
- Comparison with full scale tests which are limited mainly by the lack of available data in the public domain. For the case of evacuation, data is gathered by monitoring routine evacuation drills for which parallel simulations can be carried out. A number of studies of this sort have been carried out by Mott MacDonald as well as the School of Fire Engineering at Leeds University in the UK and at Lund University in Sweden.
It should be recognised that validation is an on-going process and we continue to seek suitable test cases, in particular for normal operations cases.