Optimized Neural Network of Predictive Maintenance for Air Booster Compressor (ABC) Motor Failure

Recently, predictive maintenance of complex machines is becoming a vital issue in industry 4.0. This is because of inexplicably occur in high voltage rotating machine can give prominent impact towards the industrial economy. Therefore, an accurate predictive maintenance is needed in reducing the degree of damage to the industrial instruments. the current method was not sufficient to indicate satisfactory result to large industrial plants. Additional auxiliary techniques are required to develop predictive maintenance. This paper suggests some methods to optimize the model of Neural Network (ANN) using Particle Swarm Optimization (PSO) and Spiral Dynamics Algorithm (SDA). The proposed predictive maintenance algorithm has been verified with actual data of ABC Motor. The ANN model was subsequently modified using PSO and SDA to provide an accurate and optimal network architecture with improved accuracy results. The result demonstrated in this paper by comparing the conventional ANN with Optimized ANN with the value of Mean Square Error (MSE) and Root Mean Square Error (RMSE).