ABSTRACT Induction Motors (IM) are ideal for a wide variety of industrial applications because they require less maintenance and run with strength and durability. Fault detection in IM is critical for ensuring industrial system dependability and preventing unexpected downtime and expensive repairs. Traditional fault detection approaches frequently struggle to capture the complex geographical and temporal patterns inherent in motor fault signals, which limit their efficiency. To address these challenges, this manuscript proposes a novel approach that integrates Complex‐Value Spatio‐Temporal Graph Convolutional Neural Networks (CVSTGCNN) with High‐level Target Navigation Pigeon Inspired Optimization (HTNPIO) for optimized fault detection and classification, named CVSTGCNN‐HTNPIO. The CVSTGCNN effectively models the intricate Spatio‐temporal dependencies in induction motor fault data, while HTNPIO optimizes the network parameters to enhance detection accuracy and efficiency. The major goal of the proposed technique is to differentiate between electrical issues that arise in IM under both faulty and healthy circumstances. When tested on a benchmark induction motor fault dataset, the proposed CVSTGCNN‐HTNPIO method achieves a high classification accuracy of 99.8% and reduces the root mean square error (RMSE) to 0.02%, outperforming existing techniques such as Spatial–Temporal Recurrent Graph Neural Networks (STRGNN), Multi‐Parallel Graph Convolutional Network (MGCN), Particle Swarm Optimization (PSO), Back Propagation Neural Networks (BPNN), and Artificial Neural Networks (ANN). These findings illustrate the method's enhanced capacity to identify various fault types with more precision, allowing for more reliable and timely motor fault diagnostics. This development has the potential to greatly improve motor operational safety, lower maintenance costs, and increase equipment longevity.
Saranya et al. (Tue,) studied this question.