Modeling solar power plants with daily data using genetic programming and equivalent circuit

Abstract

Among the various methods proposed for modeling solar panels, those based on equivalent circuits have received significant attention. In these approaches, determining unknown parameters varies depending on the modeling objective. To model voltage–current characteristics, circuit analysis methods are employed to extract these unknown parameters. However, this modeling method relies on data provided by the solar panel manufacturer, leading to increased modeling error over time as coefficients change. In this article, a method independent of the manufacturer's data for modeling solar panels is presented. This method enables accurate modeling of pre-installed solar power plants. By utilizing genetic programming on a single day's worth of data from a solar panel, the proposed method can establish relationships with a high degree of fit for the open-circuit voltage, maximum power point, and short-circuit current based on weather conditions. Through these relationships, the voltage–current characteristics can be modeled with greater precision compared to traditional circuit analysis methods, and without the need for data from the solar panel manufacturer. Finally, for further evaluation, a 3 kW solar power plant is modeled, which demonstrates the effectiveness of the proposed method.