ANALYSIS AND MODELING OF ENERGY INPUTS IN GREENHOUSES FOR DIFFERENT CUCUMBER VARIETIES

Cucumber is cultivated extensively in greenhouses and open fields worldwide, and is a key crop for greenhouse farmers in Iraq. This study aimed to analyze energy consumption and apply an Artificial neural network model (ANN) to predict the yield of three cucumber varieties: NEDDAL F1, SUPER HERO F1, and TOP GREEN F1, in Nineveh Governorate. The experiment was conducted in 18 greenhouses, and data collected included the types of equipment used in cucumber cultivation, working hours, fuel, number of workers, agricultural practices, fertilization methods, and pest control. The results showed that the total energy consumption for cucumber cultivation was 6251.88, 6201.88, and 6251.88 MJ house^-1, while the energy output was 4400, 5326.3272, and 5014.08 MJ house^-1 for the three varieties, respectively. Fertilizer and electricity consumption are key variables in cucumber cultivation, representing the highest energy inputs at 42% and 20%, respectively. The results showed that a neural network model is useful in predicting the energy output of three cucumber varieties, the minimum MSE and the highest R-value of model calibration were obtained by combining the LM training, the 10 and 12 of a hidden neuron, using the 25-10-1, 25-12-1, and 25-10-1 ANN architectures with LM training and a Log-sigmoid transfer function in NEDDAL F1, SUPER HERO F1, and TOP GREEN F1, respectively. The study demonstrated that the trained models achieved the lowest error rate, indicating the experimental model's ability to predict cucumber yield in Nineveh Governorate.