Transformation of Maintenance Processes at a Hydropower Plant Through the Implementation of a Machine Learning-Based Predictive Model

This paper analyzes the current (“As-Is”) and target (“To-Be”) states of maintenance processes at a hydropower plant (HPP), focusing on the implementation of a predictive maintenance system. The study begins by analyzing the existing As-Is model, identifying key limitations of traditional approaches: high reactivity, reliance on manual labor, and low process transparency. Based on BPMN diagrams and interviews with personnel, a target To-Be model is developed, incorporating IIoT sensors, machine learning, and integration with EAM/CMMS systems. Special attention is paid to the methodological justification (system analysis, Gap analysis, BPMN modeling) and the concept of applying machine learning. The findings confirm that overcoming technological and organizational barriers will make the transition to predictive maintenance a strategically sound step, enabling high reliability, efficiency, and sustainability of HPP operations.