AI-POWERED PREDICTIVE FAILURE ANALYSIS IN PRESSURE VESSELS USING REAL-TIME SENSOR FUSION : ENHANCING INDUSTRIAL SAFETY AND INFRASTRUCTURE RELIABILITY

Abstract

The integration of Artificial Intelligence (AI) into Structural Health Monitoring (SHM) systems has emerged as a transformative solution for predictive failure analysis in pressure systems such as pressure vessels, pipelines, and industrial reactors. This study aims to systematically examine the role of AI-powered SHM frameworks in enhancing the reliability, safety, and operational efficiency of these high-risk infrastructures. A total of 63 peer-reviewed journal articles and conference papers published between 2000 and 2023 were reviewed following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines. The selected studies were analyzed in terms of AI techniques applied, types of sensors integrated, fusion architectures, model performance metrics, validation methods, and real-world industrial applications. The review reveals that AI models—especially machine learning and deep learning algorithms—have significantly improved the early detection of faults, classification accuracy, and remaining useful life (RUL) prediction when supported by multi-sensor fusion frameworks. Models such as support vector machines (SVM), convolutional neural networks (CNN), and long short-term memory (LSTM) networks were frequently used and demonstrated strong performance, often achieving accuracy levels exceeding 90% across varied industrial scenarios. Furthermore, many of these systems have been successfully deployed in operational environments, leading to measurable improvements in maintenance scheduling, reduced downtime, and heightened safety. However, the review also identifies critical implementation challenges, including data scarcity, limited model interpretability, system integration constraints, and cybersecurity vulnerabilities. These barriers highlight the need for standardized practices, improved data governance, and interdisciplinary collaboration.