Historical statistic based fault-tolerant control mechanism for dissimilar redundant actuation system with parameter uncertainty

Abstract

This paper presents a historical statistic (HS)-based fault-tolerant control (FTC) mechanism for dissimilar redundant actuation system (DRAS), which is composed of hydraulic actuator (HA) and electro-hydrostatic actuator (EHA) during its long working term. Due to long-term service and severe working condition, multiple gradual faults would occur and degrade the system performance, leading the system model shift into fault state with parameter uncertainty. By setting reasonable sampling period and saving the historical statistic of the multiple gradual faults, the historical statistic can be used to amend the fault system model with parameter uncertainty. Then the FTC mechanism can be developed for the amended more precise system model, the control gain matrix can be solved using linear quadratic regulator (LQR) and updated in every end of system sampling and parameter estimation, to match the current multiple gradual faults condition. Case studies of DRAS subject to multiple faults have been accomplished to illustrate the effectiveness of the new design approach.