Digital Twins for Photovoltaic Plants Performance Monitoring
Abstract
The real-time monitoring of photovoltaic (PV) systems is essential for ensuring their operational health and supporting quick decision-making, aiming to mitigate losses caused by electrical faults or sensor misreadings. Digital Twins can be an effective solution to enhance the efficiency of PV system monitoring. This article provides a comprehensive overview of a cost-efficient implementation of a Digital Twin platform for photovoltaic systems using open-source technologies. The proposed architecture adopts a microservices approach, ensuring scalability and flexibility in managing large datasets across multiple plants. It integrates real-time data acquisition, a Python-based simulation core, anomaly detection, and customizable dashboards. A web interface allows users to build customized model chains, powered by a topological sorting algorithm that organizes the graph of models according to their dependencies. The system also provides tools for customizable data input and dashboard generation in Grafana, offering a user-friendly environment for model configuration and monitoring. The platform design emphasizes interoperability and modularity, allowing seamless integration with existing supervisory systems and data sources. The proposed architecture meets the essential performance and reliability requirements of modern solar operations, supporting its potential adoption in commercial applications and contributing to more sustainable energy practices.
Keywords
digital twins, photovoltaic systems, real-time monitoring, data-driven modeling, model chain, dashboards, container architecture