Optimum IMFs Selection Based Envelope Analysis of Bearing Fault Diagnosis in Plunger Pump

Wenliao Du,Xiaoyun Gong,Liangwen Wang,Zhiyang Wang,Guofu Luo

doi:10.1155/2016/1248626

Wenliao Du, Xiaoyun Gong + Show 3 more

Open Access

https://doi.org/10.1155/2016/1248626

Copy DOI

Abstract

As the plunger pump always works in a complicated environment and the hydraulic cycle has an intrinsic fluid-structure interaction character, the fault information is submerged in the noise and the disturbance impact signals. For the fault diagnosis of the bearings in plunger pump, an optimum intrinsic mode functions (IMFs) selection based envelope analysis was proposed. Firstly, the Wigner-Ville distribution was calculated for the acquired vibration signals, and the resonance frequency brought on by fault was obtained. Secondly, the empirical mode decomposition (EMD) was employed for the vibration signal, and the optimum IMFs and the filter bandwidth were selected according to the Wigner-Ville distribution. Finally, the envelope analysis was utilized for the selected IMFs filtered by the band pass filter, and the fault type was recognized by compared with the bearing character frequencies. For the two modes, inner race fault and compound fault in the inner race and roller of rolling element bearing in plunger pump, the experiments show that a promising result is achieved.

Highlights

The plunger pump constitutes the key part in hydraulic cycle, which is widely used in the huge construction machinery, marine engine, military equipment, and so forth
As reported by Jiang et al [5], the time-frequency resolution ratio is invariant for short time Fourier transform (STFT), which is a limitation for the signal with strong noise level
Though the WVD usually suffers from the disturbance of cross terms, the pseudo Wigner-Ville distribution (PWVD)

Summary

Introduction

The plunger pump constitutes the key part in hydraulic cycle, which is widely used in the huge construction machinery, marine engine, military equipment, and so forth. As the vibration signal of bearing in plunger pump shows a typical nonstationary character, the Fourier transform is not appropriate in this situation. As the pump always works in a high speed and heavy load situation and the pump self is a typical mechanical-hydraulic coupling system, the collected vibration signal shows strong nonstationary and nonlinear characteristic [13], and the bearing character signals are always submerged in the noise and the disturbance impact signals coming from plungers and swash plate, which makes it more difficult to extract the effective diagnosis information. The EMD procedure is performed to the vibration signal, and the appropriate IMFs are selected based on the resonance frequency band, and the filter bandwidth is determined. The envelope analysis is utilized to the filtered signal to implement the bearing fault diagnosis in plunger pump

Optimum IMFs Selection Based Procedure to Detect Bearing Faults

Fault Diagnosis Experiment of Bearing in Plunger Pump

Conclusions