Prognostic maintenance and smart monitoring of solar panel using IoT devices and hybrid machine learning models

The profusion, flexibility, and recurrently declining installation costs of solar photovoltaic (PV) technology have made it a protuberant renewable energy source. However, environmental unpredictability, component deterioration, power-electronic failures, and aging effects continue to have a major impact on the long-term performance and reliability of PV systems. There are some drawbacks to the conventional maintenance techniques employed in PV installations, such as routine manual inspections and corrective actions. In order to improve the performance and operational efficiency of solar PV systems, this work presents a prognostic maintenance framework driven by machine learning techniques and enabled by IoT device. A low cost hardware prototype of real time solar data acquisition and monitoring system is developed using LoRa module which collects data from the solar station and in the transmitter station the data is pushed to cloud and displayed in dashboard. Different machine learning models linear regression, decision tree, random forest and XGBoost has been implemented for predictive maintenance of solar panels. A hybrid machine learning model based on Bayesian optimization of XGBoost is proposed and from the comparative assessment of the machine learning models it is validated that the proposed hybrid model showed superior performance in terms of highest accuracy and lowest computation time.