Self-sensing and adaptive composite fault-tolerant control strategy for magnetic-liquid double suspension bearings...

To address suspension instability and rotor drop caused by displacement sensor faults in the Magnetic-Liquid Double Suspension Bearing (MLDSB), this paper studies anti-drop control under sensor fault conditions. A two-degree-of-freedom dynamic model of the radial support unit is established to describe the coupling among electromagnetic suspension force, hydrostatic supporting force, and rotor displacement. Differential displacement sensor fault models are then developed for open-circuit faults, fixed-bias faults, impact faults, and periodic interference faults. For abrupt measurement abnormalities, a self-sensing displacement reconstruction method based on the force-balance principle is proposed to replace faulty feedback through fault detection and signal switching. To suppress the residual oscillation that remains under periodic interference faults, a composite fault-tolerant control strategy combining adaptive back-propagation neural network proportional-integral-derivative (PID) and incremental PI is further introduced. Simulation results show that conventional constant-ratio control is highly sensitive to sensor feedback and cannot maintain stable suspension under the four fault conditions. Self-sensing constant-ratio control effectively suppresses instability caused by open-circuit faults, fixed-bias faults, and impact faults and achieves stable post-fault suspension. Under periodic interference faults, the proposed composite control strategy further reduces sustained oscillation, drives the rotor back to the reference position within a short time, and restores the control voltage, current, and supporting force to normal levels. The results show that the proposed hierarchical composite fault-tolerant control framework improves the anti-drop capability and operational reliability of the MLDSB under displacement sensor faults.