摘要

This article models a class of stochastic systems with faults by switching diffusion processes (SDP), and analyses the fault tolerability of such stochastic systems in the sense of input-to-state stability of overall SDP. The fault tolerability relies on the trade-off among the fault occurrence transition rate, the frequency of switching, and the decreasing rate of Lyapunov functions along the solution of the system. Our results show that it may not be necessary to design the fault tolerant controller even though the stochastic system is not separately stable in the healthy and faulty situations, the stability of the overall system process is still guaranteed under some conditions. The proposed tools are illustrated through an example of a fault-prone manufacturing system.

全文