From Reactive to Proactive: Enhancing industrial machine Maintenance through intelligent fault detection and Diagnosis
Year: 2024
Type: article
Abstract: The research on enhancing maintenance of industrial machines through intelligent fault detection and diagnosis aims to identify the use of machine learning algorithms and data analysis techniques to develop accurate and reliable fault detection and diagnosis models for industrial machines. Determine the effective implementation of fault detection and diagnosis systems in industrial machinery maintenance to move from reactive to proactive maintenance practices. Identify the most important challenges and obstacles facing organizations in adopting smart systems to detect and diagnose faults to enhance the maintenance of industrial machines using artificial intelligence and machine learning algorithms. The methodology for enhancing industrial machinery maintenance through Intelligent Fault Detection and Diagnosis (IFDD) involves obtaining relevant data from various sources, including sensors, machine control systems or historians and historical maintenance records. This data collection process is critical to implementing effective IFDD techniques and proactive maintenance strategies. 4,200 samples were split 70/30 for training and testing, with comparisons performed using 4 deep learning models and 2 machine learning models with manual feature extraction. Deep learning models showed superior accuracy, achieving up to 100% accuracy, while machine learning models were less accurate in the 94-95% range. This confirms the effectiveness of deep learning in automatically extracting meaningful features, eliminating the need for manual feature engineering.
Author: Ahmed H. Ali
Institution: National Institute of Standards
Language: English
Subfield: Control and Systems Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W7216689837
