Design of a Smart Metering Device with Edge Computing for Monitoring Silicon Photovoltaic Panels

Abstract

Optimizing power generation with photovoltaic panels seeks to achieve the best performance and efficiency, both of single panels and grids comprising several of them. The outcomes depend on ambient conditions, such as irradiance or temperature, but also on external causes (e.g., soiling) that may lead to critical power losses. This paper proposes a smart metering device containing sensors to monitor these panels continuously by collecting I-V curves, a Maximum Power Point Tracker (MPPT) to maintain the optimal operation, as well as security elements to prevent external accesses and protect the generated data. The objective is to create a predictive maintenance system with artificial intelligence algorithms on the edge. Developing those algorithms requires a characterization of the selected panels to get optimal results, detecting failures and troubles swiftly, both of the panel and the smart device. Moreover, periodically or in case of noteworthy events, the device reports the information to a cloud platform by using LoRa 2.4GHz communication protocol. The device must ensure the proper operation of panels, allowing to take sustainable corrective measures by preventing an unnecessary waste of resources (electricity or water, among others). The aim is to avoid irreparable failures that imply high costs and reach high efficiency of the monitoring system and the energy generation process.