Machine learning-based solar irradiance forecasting model using GPS

Tang, Sin Yee (2024) Machine learning-based solar irradiance forecasting model using GPS. Project Report. Universiti Teknikal Malaysia Melaka, Melaka, Malaysia. (Submitted)

Text (Full text) Machine learning-based solar irradiance forecasting model using GPS.pdf - Submitted Version Download (3MB)

Abstract

Integration of large-scale solar energy into an existing or future energy supply framework is becoming essential for the near future global energy supplies. This integration requires accurate forecasting of solar system output power, which is essential for the efficient functioning of the power grid and optimal utilization of energy fluxes within the solar system. In fact, this output power forecasting relies on solar irradiance forecasting. Accurate solar irradiance forecasting is essential for efficient integration of solar energy into the grid, as it enables grid operators and solar power plant operators to plan and manage energy production and consumption. Thus, this project proposed a Machine Learning-based solar irradiance forecasting model using GPS. Specifically, the model processed two types of input data, namely water vapor data (IWV) and total electron content (TEC) data obtained from RINEX GPS data. An Artificial Neural Network (ANN) machine learning algorithm has been used to process the input data and predict solar irradiance. Subsequently, the predicted results were displayed and validated using MATLAB GUI software. At the end of this project, the Bayesian Regularization algorithm with 10-training-layer size was identified as the best model among several algorithms that were tested, with a mean square error (MSE) of 20882.4233 and a correlation coefficient (R) of 0.86138.

Actions (login required)

View Item