Modelling the Solar Intensity of Asaba Town in Nigeria Using Response Surface Methodology and Machine Learning Techniques

The use of solar energy devices and machines such as solar pumps, heaters, solar cars, solar air conditioners, solar refrigerators, etc. has been the trend in most developed countries in the world owing to the green nature of solar technology. Based on the vast areas of application of solar technology, the knowledge of the solar intensity of a geographical location and having an estimation model to reduce the frequent measurement of solar-dependent factors are very imperative and helpful to the designers of solar devices and machines. This concern necessitated this research that was centered on determining the optimal value of the solar intensity of Asaba town, Delta state, and developing a mathematical model for estimating the solar intensity of the region. Meteorological data of the region was collected from the Nigerian Meteorological Agency (NIMET) for ten years (2011-2020). The data obtained from NIMET were: rainfall amount, relative humidity, mean temperature, and solar intensity. A multiple linear regression (MLR) model was developed using the collected data and the artificial neural network (ANN) model was also used to estimate solar intensity of the region. Response surface methodology (RSM) tool was employed to perform numerical optimization using the collected data of the region in quest of getting an optimal value of the solar intensity of Asaba town and the combinatorial best factor levels- rainfall, relative humidity, and mean temperature that would yield the optimum value of the response variable- solar intensity. Statistical tools such as mean bias error (MBE), mean percentage error (MPE), root mean square error (RMSE), Nash-Sutcliffe equation (NSE), correlation coefficient, test- statistics, and coefficient of determination (R<sup>2</sup>) were used to evaluate the estimation performance of the Quartic polynomial model developed from the optimization process, multiple linear regression, and the artificial neural network models. From the results obtained, the optimal value of solar intensity was 759.687w/m<sup>2</sup> at the factor levels of: rainfall-194.58 mm, relative humidity-28.7989 mmHg and mean temperature of 25.7288°C. Also, the statistical validation tools applied revealed that the Quartic polynomial model had a much better performance characteristic than the other estimation models. This study would find application in the field of heating, ventilation, and air-conditioning (HVAC) solar systems by aiding the designers of such systems in knowing and estimating the value of the solar intensity of Asaba and other geographical regions that have similar climatic conditions.