Context
Solar power plants generate large volumes of operational data. Using this data to continuously monitor performance and asset health can improve reliability and operational efficiency.
Challenge
Solar plants face variability in generation due to weather, limited real‑time visibility into equipment performance, and inefficient maintenance practices- leading to delayed fault detection, reduced energy yield, and higher operating costs.
AI Solution
ARMOUR+ leverages AI and machine learning to analyse real‑time and historical solar plant data to enable continuous asset health assessment and early anomaly detection. In addition, AI‑based V‑I curve analysis supports root‑cause identification of faults and performance degradation in photovoltaic modules, improving diagnostics and issue resolution.
Impact & evidence
Key metrics of success include improvements in power generation yield, reduction in unplanned downtime, and maintenance cost savings.
These impacts are assessed through analysis of historical operational and maintenance data, identifying performance losses and estimating achievable improvements.
Benefits can be realized shortly after deployment, as the solution provides immediate asset health assessment based on real-time and historical data without requiring extensive model training.
Implementation / Adoption
The solution requires real-time and historical data from solar power plant assets, including inverters, PV modules, and weather sensors, along with integration into existing monitoring systems.
Key constraints include data availability and quality, as well as differences in plant configurations and equipment across sites. Implementation typically takes 2–3 months, including data integration and system configuration. Deployment is completed within 1–2 months, covering on-site installation, validation, and user training.
Data
The solution uses real-time operational data from solar power plants (e.g., inverters, IoT sensors, SCADA, weather data such as irradiance and temperature), historical maintenance and inspection records (e.g., failure history, maintenance logs) along with asset master data and system specifications (e.g., PV module types, inverter configurations, plant design).
Future enhancements
Planned enhancements include the integration of generative AI and LLM-based capabilities to analyse unstructured data such as maintenance reports, inspection records, and image data from solar plants. These enhancements will provide deeper insights, improve asset health assessment, and enhance maintenance decision support. In addition, the solution will be expanded to support a wider range of solar assets and diverse operating environments.
