Artificial neural network–based internal leakage fault detection for hydraulic actuators: An experimental investigation

Abstract

Internal leakage is a typical fault in the hydraulic systems, which may be caused by seal damage, and result in deteriorated performance of the system. To study this issue, this article carries out an experimental investigation of artificial neural network-based detection method for internal leakage fault. A period of pressure signal at one chamber of the actuator was taken in response to sinusoidal-like inputs for the closed-loop controlled system as a basic signal unit, and totally, 1000 periodic signal units are obtained from the experiments. The above experimental measurements are repetitively implemented with 11 different active exerted internal leakage levels, that is, totally 11,000 basic signal units are obtained. For signal processing, the pressure signal in the operation condition without active exerted leakage is chosen to generate a baseline with suitable pre-proceed, and the relative values of the other basic signal units (D-value between the baseline and other original signals) act as the global samples of the following artificial neural networks, traditional back propagation neural network, deep neural network, convolution neural network and auto-encoder neural network, separately; 8800 samples by random extraction as train samples to train the above neural networks and the other samples different from the train samples act as test samples to examine the detection accuracy of the proposed method. It is shown that the deep neural network with five layers can obtain a best detection accuracy (92.23%) of the above-mentioned neural networks. In addition, the methods based on wavelet transform and Hilbert-Huang transform are also applied, and a comparison of these methods is provided at last. From the comparison, it is shown that the proposed detection method obtains a good result without a need to model the internal leakage or a complicated signal processing.

A Novel 2-D Current Signal-Based Residual Learning With Optimized Softmax to Identify Faults in Ball Screw Actuators

“…For these systems, actuator data should be monitored after installing in aircraft to ensure system reliability related to acquired data [5] to be as close as compared with conventional systems (hydraulics). To be specific, ball-screw drive actuators are the most common electric actuators explored in many experimental and simulation research studies; not only in comparison to hydraulic actuators [6,7] as well as by installing it on actual aircrafts.…”

Graphical temporal semi-supervised deep learning–based principal fault localization in wind turbine systems

“…4,5 So, the fault diagnosis for wind turbines, as a branch of the fault diagnosis, have been studied through cooperative efforts between the industry and academia, and many results have been carried out, which can mainly be categorized into four classes: 6 signal-based fault diagnosis which uses different kinds of signals for the different sub-components diagnosis, 7–9 model-based fault diagnosis which uses the online input–output data to build a physical model or find out the mechanism for fault diagnosis, 10–12 knowledge-based fault diagnosis which is a data-driven method applying data analysis, learning and training approaches with a large number of historical data and the aid of various computational intelligence techniques for the fault diagnosis, 13–20 and hybrid fault diagnosis which is the combination of the above three methods. 21–23 By the rapid growth of information technology and machine learning algorithms, the knowledge-based fault diagnosis has become an attractive and successfully improved the accuracy. Some machine learning algorithms, such as Gaussian mixture model (GMM), artificial neural network (ANN), support vector machine (SVM), and extreme learning machine (ELM), are introduced into the knowledge-based fault diagnosis.…”

A study on global optimization and deep neural network modeling method in performance-seeking control

“…The increase in the hidden layers would greatly enhance the nonlinear fitting capacity and improve the model precision. Therefore, a new aeroengine model modeling method, deep neural network (DNN), [24][25][26][27] which has deeper network structure and stronger expressive ability than the conventional NN, is proposed to improve model precision.…”