Time Series Analysis of Solar Power Generation Based on Machine Learning for Efficient Monitoring

Abstract

ABSTRACT Solar energy, a renewable resource, is essential for the efficiency of solar photovoltaic (PV) panels. However, meteorological factors, such as solar irradiation, weather patterns, precipitation, and overall climate conditions, pose challenges to the seamless integration of energy production into the power grid. Accurate prediction of PV system power output is necessary to enhance the integration of renewable energy into the grid. The study focuses on utilizing machine learning (ML) methodologies for accurate forecasting of solar power generation, addressing challenges related to integrating renewable energy into the power grid. By analyzing power generation data and employing advanced ML models, the research aims to enhance the efficiency and predictability of solar energy systems. The significance of this study lies in its potential to optimize renewable energy production, improve grid stability, and contribute to the transition towards sustainable energy sources. This study assesses the appropriateness of ML approaches for accurately projecting solar power generation in half‐hourly cycles for the next day. The study consists of many analytical phases, including exploratory data analysis, power generation data analysis, and inverter data analysis, which are carried out on two separate power plants. The following step is to conduct comparative analyses. The data are analyzed using ML models like gradient boosting classifiers and linear regressions. The first power plant produces the best results, with an amazing 0.97% accuracy utilizing the gradient boosting classifier and linear regression classifier. Contrarily, the second power plant achieved an accuracy of 0.61% with the gradient boosting classifier and 0.62% with the linear regression models. This study's techniques and insights can help PV plant operators and electricity market stakeholders make informed decisions to optimize the use of generated PV power, minimize waste, plan for system preservation, reduce costs, and facilitate the widespread integration of PV power into the electricity grid.

Keywords

• machine learning
• photovoltaic monitoring systems
• solar energy
• solar panel