DATA-DRIVEN FAULT PREDICTION FRAMEWORK FOR PLC-BASED AUTOMATIC STORAGE/ RETRIEVAL SYSTEM USING ADVANCED MACHINE LEARNING TECHNIQUES

Abstract

PLC-based Automatic Storage and Retrieval Systems (AS/RS) are widely used in modern industrial automation to enhance warehouse efficiency and accuracy; however, these systems often suffer from unexpected faults due to mechanical wear, sensor failures, and control system issues, leading to downtime and reduced productivity. Traditional maintenance strategies are not effective for early fault detection, creating a need for a data-driven predictive approach. This study proposes a machine learning-based fault prediction framework using operational data collected from PLC sensors, including parameters such as motor current, position signals, load variations, and error logs. The collected data is pre-processed using techniques like normalization, noise filtering, and feature extraction to ensure data quality. Advanced machine learning algorithms such as Random Forest, Support Vector Machine (SVM), Artificial Neural Networks (ANN), and Gradient Boosting are implemented and evaluated using performance metrics like accuracy, precision, recall, and F1-score. The proposed system aims to detect faults at an early stage, thereby reducing downtime and improving system reliability and efficiency. Validation is carried out using real-time or simulated AS/RS data, demonstrating the effectiveness of the framework in predictive maintenance. Overall, the study contributes to the development of intelligent, data-driven solutions for industrial automation, supporting smart manufacturing and Industry 4.0 initiatives by enabling proactive maintenance and optimized operational performance.