Fault diagnosis on slipper abrasion of axial piston pump based on Extreme Learning Machine

Hydraulic pump constitutes the key component in hydraulic power system and its fault diagnosis is of great importance. The structure of the pump is very complex and the relationship among parameters of pump is highly nonlinear. As a result, data-driven diagnosis method is commonly used for the pump. In this article, a new intelligent fault diagnosis scheme based on many signal processing techniques, such as Wavelet Packet Transform (WPT), Local Tangent Space Alignment (LTSA), Empirical Mode Decomposition (EMD) and Local Mean Decomposition (LMD), and recognition technique Extreme Learning Machine (ELM) is proposed to manage the fault detection on slipper abrasion of axial piston pump. Three feature extraction methods were utilized to seek for sensitive fault features. A novel classifier, ELM, was introduced in this study to diagnose the slipper abrasion fault. ELM has been proved to be extremely fast and can provide good generalization performance on many pattern recognition cases. The performance of ELM classifiers has been compared with the performance of other state-of-art techniques. The empirical results showed that ELM classifiers are really fast in learning and perform well on slipper abrasion fault detection.