A data-driven fault detection approach for unknown large-scale systems based on GA-SVM

Abstract

This paper is concerned with the data-driven fault detection (FD) problem for large-scale systems with unknown system matrices and interconnection signals. Due to the existence of the unknown system matrices and interconnection terms, the FD problem of large-scale systems is hardly solved by the existing model-based methods. To tackle this problem, the interconnection terms are firstly estimated by the subspace intersection technique. Then, based on the estimated interconnection terms and input-output data, the support vector machine (SVM) is constructed to design fault detector for each subsystem. Furthermore, genetic algorithm (GA) is introduced to optimize the parameters of SVM such that FD performance is improved. Finally, a numerical example is adopted to illustrate the effectiveness of the developed FD approach.