Interval Approach to Identification of Catalytic Process Parameters

The paper deals with applications of interval analysis methods to a special practical problem of parameter identification of a real chemical catalytic process. Several possible models of the process are given. The measured data of the process are corrupted by both usual fluctuation errors and chaotic disturbances. Samples of the input information are short. Any characteristics (including probabilistic ones) of corruptions are unknown, but the total corruption in each measurement is assumed to be bounded. As the result, the identification of model parameters under the uncertainty is performed by interval methods. The output results are represented in the form of the information sets (membership sets) of admissible values of the parameters.