Parameter estimation and failure detection of hydraulic system

摘要

This paper concerns estimation of system parameters in a hydraulic system and figuring out the variety of the failures. The system is assumed to be disturbed and its parameters are estimated using the Extended Kalman Filter (EKF). The hydraulic system consists of a pump, a pressure regulator, pipes and an accumulator and is modeled with nonlinear differential equations. The parameters of the model are extended to the state variables and then are estimated with EKF. Since the parameters have a correlation with the failures, the variety of failures can be figured out by observing the parameters. The proposed methodology showed its effetiveness through an experiment of internal leakage in the pump and the pipe. The leakage magnitude could be expected from the magnitude change of the estimated parameters.