A novel hybrid predictive maintenance model based on clustering, smote and multi-layer perceptron neural network optimised with grey wolf algorithm

Albert Buabeng,Nana Kena Frempong,Yao Yevenyo Ziggah,Anthony Simons

doi:10.1007/s42452-021-04598-1

Abstract

Considering the complexities and challenges in the classification of multiclass and imbalanced fault conditions, this study explores the systematic combination of unsupervised and supervised learning by hybridising clustering (CLUST) and optimised multi-layer perceptron neural network with grey wolf algorithm (GWO-MLP). The hybrid technique was meticulously examined on a historical hydraulic system dataset by first, extracting and selecting the most significant statistical time-domain features. The selected features were then grouped into distinct clusters allowing for reduced computational complexity through a comparative study of four different and frequently used categories of unsupervised clustering algorithms in fault classification. The Synthetic Minority Over Sampling Technique (SMOTE) was then employed to balance the classes of the training samples from the various clusters which then served as inputs for training the supervised GWO-MLP. To validate the proposed hybrid technique (CLUST-SMOTE-GWO-MLP), it was compared with its distinct modifications (variants). The superiority of CLUST-SMOTE-GWO-MLP is demonstrated by outperforming all the distinct modifications in terms of test accuracy and seven other statistical performance evaluation metrics (error rate, sensitivity, specificity, precision, F score, Mathews Correlation Coefficient and geometric mean). The overall analysis indicates that the proposed CLUST-SMOTE-GWO-MLP is efficient and can be used to classify multiclass and imbalanced fault conditions.Article HighlightsThe issue of multiclass and imbalanced class outputs is addressed for improving predictive maintenance.A multiclass fault classifier based on clustering and optimised multi-layer perceptron with grey wolf is proposed.The robustness and feasibility of the proposed technique is validated on a complex hydraulic system dataset.

Highlights

In recent years, considerable amount of resources have been dedicated to the area of classification and its applications in fault diagnosis using machine learning techniques [1, 2]
A hybrid CLUST-SMOTE-Grey Wolf Optimisation (GWO)-Multi-Layer Perceptron (MLP) technique has been proposed as a novelty for performing multiclass and imbalance classification of different fault conditions
The superiority of the proposed CLUST-SMOTE-GWO-MLP was demonstrated by comparing to its distinct modifications

Summary

Introduction

Considerable amount of resources have been dedicated to the area of classification and its applications in fault diagnosis using machine learning techniques [1, 2]. As a result, building a robust model to effectively handle such conditions is much more complicated because the complexity in the selection and Ghana. The independent implementation of supervised or unsupervised learning algorithms over the years in various fault classification task has been proven to yield some level of satisfactory results [5]. These two major forms of learning possess their strength and limitations. Among the widely used supervised algorithms for fault classification like Artificial Neural Networks (ANNs) [6,7,8], Support Vector Machine (SVM) [2, 9, 10], Linear Discriminant Analysis (LDA) [11,12,13] and Bayes classifiers [3, 14, 15] are considered superior in producing labels, but assumes that the objects classified are drawn from an independent and identical distribution, and as such does not consider their interdependencies [16]

Methods

Results

Conclusion