Real-Time Monitoring and Diagnostics for Enhanc-ing Central Inverter Efficiency in Large-Scale Solar Power Plants

Abstract

Relevance: the large-scale deployment of solar power plants has highlighted the importance of improving the efficiency and reliability of central inverters. Central inverters represent both the most critical and the most failure-prone element of the entire system. Failures occurring in inverters account for up to 20–25% of downtime and cause energy losses of 1.5–2.5% of total generation. Therefore, the application of real-time monitoring and diagnostic (RTM&D) methods has significant scientific and practical value, as it enables higher inverter efficiency, reduced operational costs, and lower electricity production costs.

Objective: to develop and validate real-time monitoring and diagnostic (RTM&D) systems aimed at improving the efficiency, reliability, and additional energy yield of central inverters in large-scale solar power plants.

Methods: during the research, sensor-based systems were employed to continuously track electrical parameters , thermal conditions, and fault signals. Data collection technologies were based on SCADA systems. Inverter efficiency was calculated as:
Diagnostic algorithms employed artificial intelligence techniques, specifically a Support Vector Machine (SVM) classifier, alongside models for calculating thermal and conduction losses in semiconductor switches.

Results: according to the findings, the monitored central inverters achieved an efficiency increase from 94.2% to 96.8%, while average monthly downtime was reduced by 18%, and mean time to repair (MTTR) decreased by 24%. Additionally, the system enabled annual savings of several hundred MWh of electricity through increased production. The proposed approach provides advantages such as real-time operation, improved diagnostic accuracy, and more effective preventive maintenance scheduling.