Artificial Intelligence and Machine Learning in Mechanical Engineering: A Scientific Literature Review on Emerging Paradigms and Industrial Transformation

Abstract

Artificial Intelligence (AI) and Machine Learning (ML) represent a transformative paradigm shift in mechanical engineering (ME), moving the discipline beyond traditional analytical and computational methodologies toward data-driven and physics-informed intelligence. This comprehensive scientific literature review critically synthesizes the state-of-the-art applications of AI/ML across the core domains of ME, including design optimization, smart manufacturing, prognostic health management (PHM), dynamic system control, and computational mechanics. The analysis details how advanced techniques, such as hybrid Convolutional Neural Network–Long Short-Term Memory (CNN-LSTM) models, overcome the challenges of predictive maintenance by effectively analyzing temporal sensor data, often achieving superior performance metrics, such as Mean Absolute Percentage Error (MAPE) below 0.36%. Simultaneously, Physics-Informed Neural Networks (PINNs) revolutionize simulation speed and physical consistency in complex phenomena like fracture and fluid dynamics, providing computationally reliable surrogate models, sometimes offering speedups of four orders of magnitude in online computation. Furthermore, the review examines the crucial shift toward inverse design methodologies and the utilization of robust hybrid architectures, such as Evolutionary Reinforcement Learning (ERL), which enhances the stability and safety of autonomous control systems by mitigating issues like brittle convergence. A critical discussion addresses major integration hurdles, including model interpretability (Explainable AI, or XAI) due to the complexity of large-scale models, inherent data scarcity in industrial settings, and the necessity for establishing regulatory frameworks to govern AI deployment in safety-critical mechanical systems. The review concludes by underscoring the imperative for responsible innovation and deep interdisciplinary collaboration to responsibly harness the full potential of AI/ML for sustainable and advanced mechanical systems.