Untersuchung der Prognosefähigkeit von deterministischen Brandsimulationsmodellen

Investigation of the prognosis ability from deterministic fire simulation models.

The investigation and evaluation of the prognosis ability from deterministic fire simulation models is of central interest in connection with the increasing acceptation of the engineering methods in the field of fire protection. In this report an evaluation methodology for the validation of fire simulation models called EMVANEMED is described and results given on basis of the methodology are clarified by an example. The evaluation was accomplished on basis of computed medians of the central evaluation sizes PEAK and PEACOCK. These values permit the evaluation of both local and global criteria of two complete time series. As an example data of two full-scale fire tests were consulted, in which a fire room was connected with a target room by small leakages and/or over a closed channel. Since both extensive measuring data in the fire room and in the target room were present, an evaluation of the simulation results should be possible for the transmission of fire protection-technically relevant values from one room to another room. For validation, the internationally accepted CFD fire simulation model Fire Dynamics simulator (FDS) in version 5 has been used. For simulation the energy release rate due to the results of the tests were given as an input. The default value of the energy release rate corresponds to the state of the art and permits an evaluation of the suitability of fire simulation codes by quantification of the deviation between experimental and simulation results. Remarkably well is the prediction of the oxygen concentration (O₂) and the gas temperatures outside of the plume by simulation. Also the results for the peak values of the temperatures of target objects are comparatively good. With the other gas concentrations (CO and CO₂) the medians are more largely and the fluctuations are clearly larger. For the fire room the gas concentrations CO and CO₂, the radiative heat flow densities and the temperatures inside the cables are computed lower, the other values higher than the test data. The results in the fire room are altogether better than in the target room. If an evaluation is accomplished on basis of local parameters (PEAK comparison) important information cannot be evaluated for the course of the time series. During the additional evaluation of global parameters, a tendency shows up that the heat transfer into construction units is not sufficiently solved by the sub models.