Actuator fault detection and diagnosis based on morphology-wavelet

The paper described a novel method for detecting and identifying faults that occur in the actuator of control systems with input and output signals related to the component itself. The diagnosis algorithm consists of three steps: firstly, generalized morphological filter with multi-structure elements is designed to filter the random noise and impulse noise in actuator´s input and output signals. And secondly, wavelet transform, which has been proposed for sensor´s fault diagnosis some years ago, is developed for actuator´s fault diagnosis. To effectively extract the abruptly fault characteristic, wavelet transform was used to analyze the filtered signals in this paper. By the Multi Resolution Analyzing (MRA), the faults can be detected accurately. Thirdly, according to calculating the Lipschitz Exponent (LE) at the fault´s point, the fault type can be identified. The several typical faults such as fix, gain, bias have been studied. The effectiveness of the proposed approach is demonstrated by the means of experimental results and simulations.