A multi-sensor fault detection strategy for axial piston pump using the Walsh transform method

Qiang Gao,Yongteng Zhong,He-Sheng Tang,Jia-Wei Xiang

doi:10.1177/1550147718772531

Qiang Gao, Yongteng Zhong + Show 2 more

Open Access

https://doi.org/10.1177/1550147718772531

Copy DOI

Abstract

The axial piston pump is a key component of the industrial hydraulic system, and the failure of pump can result in costly downtime. Efficient fault detection is very important for improving reliability and performance of axial piston pumps. Most existing diagnosis methods only use one kind of the discharge pressure, vibration, or acoustic signal. However, the hydraulic pump is a typical mechanism–hydraulics coupling system, all of the pressure, vibration, and acoustic signals contain useful information. Therefore, a novel multi-sensor fault detection strategy is developed to realize more effective diagnosis of axial piston pump. The presence of periodical impulses in these signals usually indicates the occurrence of faults in pump. Unfortunately, in the working condition, detecting the faults is a difficult job because they are rather weak and often interfered by heavy noise. Therefore, noise suppression is one of the most important procedures to detect the faults. In this article, a new denoising method based on the Walsh transform is proposed, and the innovation is that we use the median absolute deviation to estimate the noise threshold adaptively. Numerical simulations and experimental multi-sensor data collected from normal and faulty pumps are used to illustrate the feasibility of the proposed approach.

Highlights

Axial piston pump, one of the most widespread components in the hydraulic systems, has been widely used in industrial machines
To solve the shortcoming of the conventional fault diagnosis method for axial piston pumps, a new multi-sensor fault detection strategy for axial piston pump using the Walsh transform method is proposed in this article
The working principle of axial piston pump, the discharge pressure, vibration, and acoustic responses to the pit fault are introduced in this article

Summary

Introduction

One of the most widespread components in the hydraulic systems, has been widely used in industrial machines. EMD is one of the most popular methods that have been recently developed in faulty diagnosis of axial piston pumps.[15,16] A complicated signal can be decomposed into a number of intrinsic mode functions (IMFs) followed by the Hilbert spectrum analysis to calculate the corresponding instantaneous amplitude and instantaneous frequency hidden in the noisy signals. There are many advantages in the EMD, for instance, self-adaptive ability, can be applied to non-linear and non-stationary process, and well engineering application.[17,18] there are some limitations in this method, such as the distortion of the faulty impulses, the end effect, and the physical interpretations of IMFs. What’s more, EMD sometimes cannot extract fault features because of the mode mixing phenomena.[19,20,21] To solve these problems, EEMD is developed by Wu and Huang, which can be applied to the spectral kurtosis optimized band-pass filtered signal and obtain a number of IMFs.[22,23] the problems of reducing the narrow band impulses and huge running time are still need to be solved.[24,25]. It is easy to see that the basis vectors in a Walsh transform kernel only consist of the value of ‘‘ + 1’’ or ‘‘21.’’ For example, the Walsh matrix of order three is given by function (6)

À1 À1 1 À1 1 1

Findings

Conclusion