Advances in Artificial Intelligence Applicationsfor Fault Diagnosis and Optimization inEngineering Systems

Abstract

Artificial intelligence has become an indispensable tool for fault diagnosis and optimization in modern engineering systems characterized by complexity, nonlinearity, and uncertainty. Conventional model-based diagnostic techniques often fail to adapt to evolving operating conditions and large-scale
sensor data environments. This review presents an extensive analysis of artificial intelligence methodologies applied to fault diagnosis and optimization across mechanical, electrical, manufacturing, aerospace, and
energy systems. Machine learning, deep learning, and hybrid physics-informed approaches are examined in terms of accuracy, robustness, and scalability. Optimization frameworks based on evolutionary algorithms,
reinforcement learning, and digital twins are critically reviewed. Key challenges related to data imbalance, model interpretability, cybersecurity, and real-time deployment are discussed. The paper concludes by identifying future research directions toward autonomous, resilient, and trustworthy AI enabled engineering systems.