Faults Feature Extraction Using Discrete Wavelet Transform and Artificial Neural Network for Induction Motor Availability Monitoring—Internet of Things Enabled Environment

Abstract

This paper presents the high contact resistance (HCR) and rotor bar faults by an extraction method for an induction motor using Discrete Wavelet Transform (DWT) and Artificial Neural Network (ANN). The root mean square (RMS) and mean features are obtained using DWT, and ANN is used for classification using activation functions. Activation provides output by assigning the specific input with respect to the transfer function according to the nature and type of the activation function. Method: The faulty conditions are induced using MATLAB by adopting the motor current signature analysis (MCSA) method to achieve current signature signals of the healthy and faulty motors. Results: The DWT technique has been applied to obtain fault-specific features of the average continuously varying signal (RMS) and an average of the data points (mean) at levels 5, 7, 8, and 9, followed by ANN to classify the faults for condition monitoring. Utility: The utility of the results is to reduce unscheduled downtime in the industry, thus saving revenue and reducing production losses. This work will help provide support to ensure early indication of faults in induction motors under operating conditions, enabling in-service engineers to take timely preventive measures as part of the availability of resources in IoT-enabled systems. Application: Resource availability and cybersecurity are becoming vital in an environment that supports the Internet of Things (IoT) as the essential components of Industry 4.0 scenarios. The novelty of this research lies in the implementation of high contact resistance and rotor bar faults using DWT and ANN with different activation functions to achieve accuracy up to 98%.

Exploring Cybersecurity Education and Training Techniques: A Comprehensive Review of Traditional, Virtual Reality, and Augmented Reality Approaches

“…We followed the SLR approach [25] due to its systematic and prevalent use in literature reviews. It involves a review of the literature using a predefined set of steps to identify, analyze, and interpret the available research related to the given research question [25,26]. The complete scenario of the SLR is depicted in Figure 1.…”

“…Searching for articles in online digital libraries is one of the critical steps in an SLR. First, we designed a search string according to the SLR recommendations [25]. For this purpose, we developed a search string consisting of the basic keywords along with necessary synonyms using different Boolean operators.…”

Solar Power Prediction Using Dual Stream CNN-LSTM Architecture

“…These techniques are broadly categorized into statistical (ST), artificial intelligence (AI), and hybrid methods (HM) [ 11 ]. In ST-based methods, several algorithms are developed, including auto-regressive [ 12 ], Bayesian [ 13 ], Kalman [ 14 ], grey models [ 15 , 16 ], and the Markov chain model [ 17 ]. Additionally, MaatAllah et al [ 18 ] and Reikard et al [ 19 ] developed ST-based models for renewable power prediction.…”

An Efficient Lightweight Hybrid Model with Attention Mechanism for Enhancer Sequence Recognition

“…Several computational prediction models to rapidly recognize enhancers in genes were developed in recent years as a result of the improvement of machine learning. These include Enhancer-LSTMAtt [7], CSI-ANN [8], EnhancerFinder [9], Chrome, GKM-SVM [10], DEEP [11], GenSVM [12], RFECS [13], EnhancerDBN [14], and BiRen [15]. Despite this, these methods merely act as a classification tool for enhancers that have been identified.…”