Application of the distortion method of model validation using the PC-based simulation language, Berkeley Madonna

The model distortion method for model validation takes as its starting point the proposition that a mathematical model of suitably high order can be made to follow any set of recorded transients exactly by varying appropriately the nominally constant parameters as functions of time (“distorting the model”). Clearly the better the model the less will be the required distortion. The distortion method requires that a set of simulation models be run, with communication from the current submodel to the next occurring via the distorted state variables, state derivatives and parameter variations that have been calculated. In addition, the first model will need to be fed with the plant data. This implies a need for a simulation language with facilities for importing and exporting large amounts of data easily. The Berkeley Madonna simulation language affords such facilities. The language, which is PC-based, has been used to demonstrate the distortion method applied to a simple but real and important example: the testing of a back-up system to ensure shutdown of a nuclear reactor if the main hydraulic pumps failed