Greenhouse Energy Analysis and Neural Networks Modelling in Northern Iraq

Abstract

Abstract This study aims to analyse the energy of cucumber production in a greenhouse and examine the application of a multilayer perceptron to predict the productivity of an agricultural region in Nineveh Governorate. The research data were collected from experiments including fuel, fertilisers, pesticides, seeds, workers, electricity, and the number of hours worked in agricultural processes to produce cucumber crops. The results showed that the total energy consumption of the cucumber was 46,432.013 MJ·ha−1, while the output energy was 53,127.727 MJ·ha−1. The fungicide energy consumption, herbicide energy consumption and electricity energy consumption are considered the most critical variable in cucumber plantation procedures; its significance is the relative values of 100%, 99.7% and 93.3%. The impacts of human labour, P fertiliser, diesel fuel and N fertiliser on cucumber operation were 25,725 MJ·ha−1, 548.596 MJ·ha−1, 3,011.178 MJ·ha−1 and 7,244.545 MJ·ha−1, respectively. This research concludes that a multilayer perceptron neural network algorithm helps predict cucumber production and shows that the trained neural network produced minimal errors, indicating that the test model could predict a cucumber crop yield in Nineveh province.