Solar Power Prediction Models: Classification Based on Time Horizon, Input, Output and Application

Abstract

Large number of solar power prediction models such as Artificial Neural Networks (ANN), Numerical Weather Prediction (NWP), Support Vector Regression (SVR), Bayesian approaches and hybrids are exist in literature. Different models are used for different type of applications such as system operation, market operation based on their time horizon. Also, different models use different types of inputs. There are less literature available for the classification of these models based on application, time horizon, input types and predicted outputs. Therefore, this paper classify the solar power prediction models based on time horizon, input types, application and predicted output type and also gives most suitable model for each class. It concludes that position of sun and earth is same corresponding to previous year similar day, only difference is cloud cover and cloud movement, wind speed and temperature. Compared to wind speed and temperature cloud movement and cloud speed have high co relation with solar power. Image processing based on sky imagery data for real time cloud cover and cloud movement along with previous years similar day sky data will give accurate predictions.