Internal Leakage Detection in Hydraulic Actuators Using Empirical Mode Decomposition and Hilbert Spectrum

The applicability of Hilbert-Huang transform (HHT) for internal leakage detection in valve-controlled hydraulic actuators is investigated in this paper. First, the actuator response to structured (periodic step) inputs directly applied to the control valve is analyzed. This procedure is a representative of an offline diagnosis scheme. Next, the capability of the approach toward online applications, whereby the actuator tracks unstructured (pseudorandom) position reference inputs in a closed-loop control scheme against a load, is examined. The pressure signal at one side of the actuator is decomposed into oscillatory functions called intrinsic mode functions (IMFs), and Hilbert transform is applied to each IMF to obtain the instantaneous amplitude. It is shown that the root mean square of the instantaneous amplitude associated with the first IMF establishes feature patterns that can be effectively used to detect internal leakage and its severity. Experimental tests show the effectiveness of the approach in detecting internal leakage values as low as 0.124 L/min (representing a reduction of approximately 2.6% of the available flow rate to move the actuator) during offline diagnosis and as low as 0.23 L/min (representing a reduction of approximately 5% of the available flow rate to move the actuator) when the actuator tracks reference position inputs online. This is done without having prior knowledge about the model of the actuator or leakage.