Hybrid AI-Driven Monitoring and Reliability Enhancement Framework for Centralized Inverters in Large-Scale Solar Power Plants

Abstract

Relevance: Reliability degradation of centralized inverters significantly affects operational stability and efficien
cy of large-scale photovoltaic power plants.
Objective: This study develops a hybrid AI-driven monitoring framework integrating SCADA, thermal sensing,
partial discharge analysis, and Digital Twin technologies to improve reliability assessment, predictive mainte
nance, and operational efficiency of centralized inverters in utility-scale solar power plants.
Methods: Comparative reliability analysis, sensor-based diagnostics, AI/ML prediction, and Digital Twin-based
monitoring architectures were investigated.
Results: Hybrid monitoring improved early fault detection accuracy by 34,5 % and reduced unexpected inverter
downtime by 27,8 %.