Forecasting of solar irradiance for solar power plants by artificial neural network

• Main Authors: Watetakarn S., Premrudeepreechacharn S.
• Format: Conference Proceeding
• Published: 2017
• Online Access: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84964944575&origin=inward; http://cmuir.cmu.ac.th/jspui/handle/6653943832/42141
• Institution: Chiang Mai University

© 2015 IEEE. This paper presents solar irradiance forecasting in Mae Sariang, Mae Hongson Province, Thailand which has a solar power plant. This solar power plan is a photovoltaic (PV) with capacity power output at 4 MW. However, the adoption of solar irradiance as a power source on a global scale has not been uniform, due to by meteorological conditions, which cause the fluctuations and inconsistencies in PV power output. This paper has applied the Artificial Neural Network by Backpropagation algorithm to forecast solar irradiance. The model uses solar irradiance and meteorological data of previous 7-day period and relevant data for the training. The forecasting results predict solar irradiance in half hour increments in present day which were not used in the modeling. Simulation results have shown that the mean absolute percentage errors in the four example days of the forecasting are less than 6%.