Global and Local Feature Extraction Using a Convolutional Autoencoder and Neural Networks for Diagnosing Centrifugal Pump Mechanical Faults

Alexander E Prosvirin,Jong-Myon Kim,Zahoor Ahmad

doi:10.1109/access.2021.3076571

Alexander E Prosvirin, Jong-Myon Kim + Show 1 more

Open Access

https://doi.org/10.1109/access.2021.3076571

Copy DOI

Abstract

Centrifugal pumps are important types of electro-mechanical machines used for fluid and energy conveyance. Mechanical faults in centrifugal pumps lead to abnormal impacts in the vibration signal of the system. Those impacts induce nonstationarity in vibration signals and hence complex time-frequency domain signal analysis techniques are required to investigate the mechanical fault features of centrifugal pumps. In this paper, an end-to-end pipeline for diagnosing faults in centrifugal pumps is proposed. To create a two-dimensional representation of the transients that appear in the vibration signals due to centrifugal pump operating conditions, first, a 1/3-binary tree fast kurtogram is computed. Next, a convolutional autoencoder and convolutional neural network are trained to autonomously extract global and local features from the kurtograms. Then, global, and local features are merged to form a joined feature vector that contains different visual features that are extracted using convolutional deep architectures using their specific loss functions during the training. Finally, this feature vector is propagated to a shallow-structured artificial neural network to accomplish fault identification. The proposed framework has been validated by the dataset collected from a real industrial centrifugal pump test rig. The results obtained during the series of experimental trials demonstrated that the introduced method achieved high classification accuracies when diagnosing faults based on signals collected under 3.0 and 4.0 bars of pressure.

Highlights

Centrifugal pumps (CPs) are an important electro-mechanical energy conversion machine and have become an important part of everyday business
We propose a combined deep learning model for extracting features from the representations of the vibration signals collected from the CP and for decision making
The tiny convolutional neural networks (CNNs) and CONVOLUTIONAL AUTOENCODER (CAE) are used in parallel, where the CNN is used to extract the local fault features from the kurtogram padding of the signal, and the CAE is utilized to detect the global features of the pattern using the data compression property of autoencoders

Summary

Introduction

Centrifugal pumps (CPs) are an important electro-mechanical energy conversion machine and have become an important part of everyday business. CPs are versatile, cheap, simple in construction, and reliable in operation Their unexpected failure may lead to severe consequences which include economic losses, energy losses, costly repairs, threats to the safety of operating staff, and long downtimes. Several health management strategies (HMS), such as reactive maintenance, preventive maintenance, and predictive maintenance have been developed in past research [2]. CBM maximizes the running time of the machine with minimum cost [3]. Considering these advantages, this study uses the CBM strategy for centrifugal pump fault diagnosis

Methods

Results

Conclusion