Artificial Neural Network Application for the Prediction of Global Solar Radiation Inside a Greenhouse

Solar radiation prediction is essential for several research applications in renewable energy. In particular, solar irradiation can be considered as a target parameter related to greenhouse microclimate. This study developed a model based on an artificial neural network (ANN) for predicting incident solar radiation on a horizontal surface in an agricultural greenhouse. Standard neural networks with different architectures (6-15-1) were designed using the Neural Toolbox for MATLAB based on the meteorological data collected of semiarid region of Ghardaïa (32.36° N, 3.81° W) Algeria. It considering as input several parameters including temperature, relative humidity, wind speed, and solar radiation was addressed. Our results showed that the ANN predictions had a correlation coefficient of over 96% with the actual solar radiation, indicating that the model is highly reliable for assessing solar radiation levels inside the greenhouse. We also found that the ANN method is suitable for predicting other greenhouse climatic data and can be used for the preliminary design of agro-system.