Fault diagnosis and tolerance control for accelerator in maglev system based on signal processing

A new style tracking-differentiator based on border characteristics is applied in the Fault Diagnosis and Active Fault Tolerance problem for accelerator in maglev system. The integral of accelerator signal is compared with the differential of gap signal, and the difference is used to obtain the decision variable and threshold according to Bayes´s decision theory. When fault occurs, integral of accelerator signal is switched by differential of gap signal as the feedback. Simulation result shows that the diagnosis is fast enough, the switch of feedback signal has no influence on the stability of the system and the fluctuation during the switch process is small.