[IEEE 2019 22nd International Conference on Electrical Machines and Systems (ICEMS) - Harbin, China (2019.8.11-2019.8.14)] 2019 22nd International Conference on Electrical Machines and Systems (ICEMS) - Power Forecasting of Photovoltaic Generation Based on Multiple Linear Regression Method with Real-time Correction Term

摘要： This paper proposes a photovoltaic power generation forecasting model which improves Multiple Linear Regression method (MLRM) with real-time correction term traditional day-ahead, hourly power (RCT). Firstly, a generation prediction model is developed by MLRM based on qualitative variables (hour, month, weather type), quantitative variable (solar radiation intensity) and physical characteristics of interactions between the variables. Secondly, an improved is presented which adds a model named MLRM+RCT correction term based on shorter real-time measured power data to MLRM to reduce the hourly prediction errors of MLRM. MLRM+RCT is tested based on power generation data released by IEEE Energy Forecasting Group in 2014. The results show that the performance of MLRM+RCT is better than that of MLRM and a benchmark method called exponential smoothing method.