Robust filtering for fault tolerant control using output PDFs of non-Gaussian systems

Motivated from some practical industrial processes, an optimal fault tolerant control (FTC) scheme is studied for stochastic continuous dynamic systems with time delays using output probability density functions (PDFs). Different from the classical FTC problems, the measured information is the PDFs of system output rather than its value, where the B-spline expansion technique is applied so that the output PDFs can be formulated in terms of the dynamic weightings. For the established weighting system with nonlinearities, uncertainties and time delays, the concerned FTC problem is investigated by using robust optimisation techniques. A linear matrix inequality (LMI) based feasible FTC method is presented to ensure that the fault can be well estimated and compensated, where the generalised H-infinity performance is optimised for the time-delayed systems with the non-zero initial condition and the truncated norms. Simulations for a model in the paper-making process are given to demonstrate the efficiency of the proposed approach.