Evaluation of the Reliability of Empirical Mathematical Models of Dynamic Systems Using Input Signal Shift Method

An algorithm is proposed that makes it possible to effectively detect the presence of correlated interference in experimental data used for parametric identification. The main problem leading to such interference is structural inconsistencies between models and real objects. Establishing the presence of correlated interference opens up opportunities for expanding identification tools. In this paper, the identification of correlated interference using the parametric identification method was carried out, while using a modified Newton method to minimize the objective functional. The results obtained from the numerical example demonstrate the effectiveness of the proposed method in detecting correlated interference, and also indicate that the values of the criteria contain information about the magnitude of the estimation errors. The results obtained in this scientific research work can be useful for further developments in the field of parametric identification and signal processing.