Parametric fault diagnosis for electrohydraulic cylinder drive units

A novel model-based methodology for fault diagnosis (FD) of nonlinear hydraulic drive systems is presented in this paper. Due to its linear dependence upon parameters, a second-truncated Volterra nonlinear model is first used to characterize such systems. The versatile order-recursive estimation scheme is employed to determine the values of parameters in the Volterra model. The scheme also avoids separate determination of the model order; thus, the complexity of the search process is reduced. Next, it is shown that the estimated parameters, representing different states of the system, normal as well as faulty conditions, can be used to detect and isolate system faults in a geometric domain. Very promising results are exhibited via simulations as well as laboratory experiments. It is concluded that the developed parametric FD technique has potential to provide efficient condition monitoring and/or preventive maintenance in hydraulic actuator circuits.