Data Augmentation Using Generative Adversarial Network for Automatic Machine Fault Detection Based on Vibration Signals

Van Bui,Tung Lam Pham,Yeong Min Jang,Huy Nguyen

doi:10.3390/app11052166

Abstract

In the last decade, predictive maintenance has attracted a lot of attention in industrial factories because of its wide use of the Internet of Things and artificial intelligence algorithms for data management. However, in the early phases where the abnormal and faulty machines rarely appeared in factories, there were limited sets of machine fault samples. With limited fault samples, it is difficult to perform a training process for fault classification due to the imbalance of input data. Therefore, data augmentation was required to increase the accuracy of the learning model. However, there were limited methods to generate and evaluate the data applied for data analysis. In this paper, we introduce a method of using the generative adversarial network as the fault signal augmentation method to enrich the dataset. The enhanced data set could increase the accuracy of the machine fault detection model in the training process. We also performed fault detection using a variety of preprocessing approaches and classified the models to evaluate the similarities between the generated data and authentic data. The generated fault data has high similarity with the original data and it significantly improves the accuracy of the model. The accuracy of fault machine detection reaches 99.41% with 20% original fault machine data set and 93.1% with 0% original fault machine data set (only use generate data only). Based on this, we concluded that the generated data could be used to mix with original data and improve the model performance.

Highlights

Machine maintenance is one of the most important fields in the industrial environment
We briefly review the characteristic of the vibration signal data with different approaches in fault detection applications
This study proposed a novel method to generate the fault machine vibration signal data, enhancing the model performance in the case of a limited fault dataset for training

Summary

Introduction

Machine maintenance is one of the most important fields in the industrial environment. Vibration data from Spectra Quest’s Gearbox Prognostics Simulator (GPS) is tested using various fault detection approaches for both limited and unlimited input data Another data source that can be considered is the real-scenario data, such as Reference [21]. The main contribution of this paper is the different approaches used to evaluate the generation data and to guarantee similarity with the original data These approaches include different preprocessing processes and a variety of machine learning techniques in pattern recognition. The advantage of FFT is that we can process more significant features in the frequency domain classification between the normal and broken machine signals Another advantage of the FFT transform is that the generated signal is evaluated indirectly, which leads to better performance analysis. We will discuss different approaches and AI models that can be applied to the fault diagnosis results

Fault Diagnosis with Original Data

Statistical Analysis

Data Generation in Machine Fault Detection

Data Generating

Findings

Conclusions