Nonlinear observer-based fault detection technique for electro-hydraulic servo-positioning systems

The purpose of this paper is to present an experimental evaluation of a model-based technique for faults in electro-hydraulic servo-positioning systems using a nonlinear observer. In general, a fault is defined as an undesirable deviation of a characteristic property, which prevents a desired purpose to be achieved, and which leads to an anomaly in the overall system performance. A nonlinear observer-based principle, combined with Wald's sequential test is proposed for detecting faults. The fault detection process is facilitated by the control signal and the actuator's linear velocity, which are the measurable input and output variables, respectively. Simulation studies are conducted for both observer performance and the fault detection technique. System and sensor faults are simulated for changes in the bulk modulus. The validity of the nonlinear observer was confirmed by simulations and experiments for faults due to incorrect supply pressures as well as sensor faults. Results reveal that faults can be quickly detected. Overall, they show the efficacy and the promise of the proposed scheme.