Estimation of Solar PV Plant Output using LSTM-CNN Algorithm

Maintaining a Solar Photovoltaic plant incorporates difficulties such as periodic panel sanitizing, inverter maintenance, monitoring and testing, weather-related damage, battery maintenance (if necessary), and insect control. While solar facilities require less maintenance than traditional power plants, appropriate servicing is necessary for optimal performance and durability. To overcome these problems, we are presenting a precise forecast of solar plant electricity production since it is vital for effective grid management, optimal scheduling of maintenance of the plant, and cost-effective energy trading. A Long Short-Term Memory - Convolutional Neural Network based approach is presented in this study for solar plant production forecasting. This method has exhibited promising results in recent years for time-series forecasting. The proposed model combines Long Short-Term Memory and Convolutional Neural Network strengths to capture the input data’s long-term dependencies and spatial features. The proposed algorithm is tested on available dataset [2] and a root mean square error of 3.3% is obtained. This error is lower as compared to 5.5432% as mentioned in [16]; thus, our method produces better results.