AI-Powered Predictive Maintenance and Prognostic Health Management Using Edge-Based Predictive Algorithms for Industrial Operations

This study presents an AI-powered framework for predictive maintenance and prognostic health management (PHM) based on edge-enabled predictive algorithms to support intelligent fault diagnosis in industrial operations. The proposed framework is designed to monitor system conditions, detect early fault signatures, and anticipate degradation patterns using high-frequency operational data collected from two large industrial plants between 2024 and 2025. By leveraging edge computing, the approach enables localized anomaly detection with low latency, allowing deviations in system behavior to be identified close to the data source. The methodology integrates edge-based anomaly detection with predictive modeling techniques to estimate future system health states and fault-related risk dynamics. Anomalies identified at the edge level are aggregated and processed through forecasting models to infer degradation trends and support prognostic assessment. A health-oriented evaluation layer translates predictive outputs into actionable indicators that support maintenance planning and system recovery decisions. The framework is evaluated using standard predictive performance metrics, including MAPE, RMSE, and R², together with a health-related improvement measure reflecting system stability and recovery capability. The results demonstrate high predictive reliability, with the models explaining approximately 98.9% of the observed variability in system risk indicators and achieving measurable improvements in operational stability through early fault mitigation. This research contributes a scalable algorithmic framework that links data-driven condition monitoring, intelligent fault diagnosis, and PHM within an edge computing environment, strengthening maintenance decision accuracy in dynamic industrial settings.