The prediction of monthly average solar radiation with TDNN and ARIMA

• Main Authors: Wu, Ji., Chan, C. K.
• Other Authors: School of Electrical and Electronic Engineering
• Format: Conference or Workshop Item
• Language: English
• Published: 2013
• Subjects: DRNTU::Engineering::Electrical and electronic engineering
• Online Access: https://hdl.handle.net/10356/102836; http://hdl.handle.net/10220/16877
• Institution: Nanyang Technological University

In this paper, two well-known algorithms: ARIMA and TDNN (Time Delay Neural Network) are applied to conduct the short term prediction of solar radiation. For the daily solar radiation series is non-stable due to the fast weather changing, monthly average solar radiation is adopted as the data source. As ARIMA model requires the time series to be stationary, first order difference is performed on the monthly solar radiation to obtain a stationary series. AIC (Akaike's Information Criterion) is used to identify the optimal prediction model. TDNN is also used to do prediction of the monthly average solar radiation and LM (Levenberg -- Marquard) is chosen as the training algorithm. The performance of these two prediction models are compared with each other.