Prediction of global solar radiation using miso ARX model

• Main Author: Mohammed Alashmoori, Abdulrahman Abdullah
• Format: Thesis
• Language: English; English; English
• Published: 2022
• Subjects: TD194-195 Environmental effects of industries and plants
• Online Access: http://eprints.uthm.edu.my/6968/1/24p%20ABDULRAHMAN%20ABDULLAH%20MOHAMMED%20ALASHMOORI.pdf; http://eprints.uthm.edu.my/6968/2/ABDULRAHMAN%20ABDULLAH%20MOHAMMED%20ALASHMOORI%20COPYRIGHT%20DECLARATION.pdf; http://eprints.uthm.edu.my/6968/3/ABDULRAHMAN%20ABDULLAH%20MOHAMMED%20ALASHMOORI%20WATERMARK.pdf; http://eprints.uthm.edu.my/6968/

The need for renewable energy sources is growing day by day because of the severe energy crisis in the world today. Renewable energy sources play a significant role in electricity generation. Several renewable energy sources (like solar, wind, geothermal, and biomass) can be used for generation of electricity and for meeting our daily energy demands. Solar energy is the most viable option for electricity generation because it is available everywhere and is free to utilize. Therefore, integration of solar energy sources has gradually become the main challenge for global energy consumption in recent decades. As a result, while predicting solar system outputs, it is essential to predict global solar radiation in a precise and efficient way. Inaccurate forecasting results in either load overestimation which leads in increased costs or failure to gather adequate supplies. However, accurate solar radiation forecasting is a challenging task because solar resources are intermittent and uncontrolled. To tackle this difficulty, several methods have been developed. This project use the system identification ARX model to predict the global solar radiation. ARX model stands for autoregressive with exogenous variables where the exogenous variables are the input terms. The project starts by collecting the meteorological data (air temperature, maximum temperature, minimum temperature, wind speed, relative humidity, solar radiation) using RETScreen software the data have been collected for a period of four years starting from 2016 to 2019 the data is then divided into two groups even and odd. The model tested for two different sets 60% of data for estimating and 40% for testing and 70% of data for estimating and 30% for testing. The project has two different ARX techniques SISO and MISO each technique has three different model with different inputs. SISO ARX model highest best fit was 72.34% when the minimum temperature set as an input. MISO ARX model shows a best fit of 89.58% when all data set as an inputs. Both SISO and MISO models gives high results when using the odd data compared to the even data.