Evaluation and Improvement of Power Forecasting Indicators for Renewable Energy Stations

As an important guidance for power forecasting in renewable energy stations, evaluation indicators play crucial roles. By thoroughly studying currently used power forecasting indications, limitations are found for indicators based on Root Mean Square Error (RMSE), Mean Absolute Error (MAE) and other indicators, especially in power stations with temporal generation characteristics, exemplified by photovoltaic power stations where distortion is obviously found. In this paper, basic conditions of power forecasting indicators are firstly studied, with which the advantages and disadvantages are compared. Based on this, an improved indicator of Weighted Root Mean Square Error (WRMSE) is suggested. It is verified by the result that the new evaluation indicator has evident advantages in all aspects, and is more adaptable as an indicator for power forecasting level in renewable energy stations.