A Quantitative Method for the Fault Diagnosability of Affine Nonlinear System

The fault diagnosability can guide the construction of diagnosis system and thereby influence the reliability of control system. But it is hard to evaluated the fault diagnosability of affine nonlinear system by using linear system theory, especially when the system noise is taken into consider. Base on the differential geometry theory, an evaluation method is proposed to quantitate the fault diagnosability of affine nonlinear system. Firstly, a qualitative judgment method for fault detectability and isolability is introduced for single fault system. Secondly, a series of quantitative indexes are designed to evaluate the complexity of fault detectability and isolability through system outputs, measuring the influence of the distribution similarity, the fault amplitude and the system states. Lastly, a validation example is presented to demonstrate the effectiveness of the proposed method.