Scalable edge computing architecture for multi-site photovoltaic systems monitoring: A modular microservice-based approach

Abstract

The large-scale deployment of photovoltaic installations at geographically remote sites underscores the need for effective and reliable monitoring systems. The monitoring tools should be capable of managing numerous devices that generate large amounts of data. One of the major challenges is to ensure interoperability with equipment and guarantee the reliability of photovoltaic system operations. The present work aims to propose monitoring architecture that can handle remote photovoltaic plants and integrate operation and maintenance strategies. The architecture embeds a distributed data flow, an anomaly detection pipeline in addition to performance indicator calculation in accordance with IEC 61724-1, and a real-time notification system. The architecture is based on a microservice architecture that integrates edge computing for local data acquisition and preprocessing. System validation was carried out to assess latency, data completeness, and data consistency. The results show stable latency with an average of 3.4 s with data completeness of 100% and 99.57% consistency. These findings demonstrate the ability of architecture to monitor remote plants, highlighting its suitability for large-scale deployment and the integration of advanced analysis. • A scalable edge computing and microservice architecture for PV monitoring. • Ensure multi-vendor interoperability via protocol adapters and data standardization. • Provides KPIs compliant with IEC 61724-1, anomaly detection pipeline and predictive maintenance insights. • Real-time data visualization for multi-site PV systems and notifications.