IoT and Computer Vision-based Real-time Solar PV Soiling Monitoring System

To align with decarbonization and energy electrification goals, solar PV systems are expected to gain in popularity in the coming decades. Monitoring PV systems is key to ensuring optimal performance and detecting faults and malfunctions. Faults include partial shading due to physical obstructions, soiling bands, and non-uniform soiling phenomena, which result in localized hotspot heating. The reported work presents an Internet of Things (IoT)- based solar PV monitoring system, which utilizes an ESP32 microcontroller serving as the central processing unit and capable of internet connectivity via WiFi. A range of sensors is used, which serve as an in-built weather station and optical soiling sensor. The collected data is processed and transmitted wirelessly over existing network infrastructure to a cloud-based platform, ThingSpeak, where it is stored, visualized, and analyzed. The combination of electrical and meteorological data provides accurate energy forecasting. Additionally, an independent camera system using a Raspberry Pi as a microprocessor can be used to visually inspect remote modules. This allows for prompt identification of anomalies in power production that may be caused by shading or dirt on panels. Therefore, realtime monitoring can aid in a proactive and dynamic approach to PV system maintenance, optimizing cleaning frequency and energy yield.