Modeling of Energy Consumption and Environmental Indices of Production and Processing of Tea with Regression and Artificial Neural Network Models

In this research, modeling of energy consumption and environmental emissions for different process of tea production in Gilan province (Iran) were investigated using regression and artificial neural networks. Data related to agricultural and industrial activities of tea production chain were collected through questionnaire survey from tea farms of different counties in Guilan province. Also, the datatbank of agricultural organization, tea research institute and tea factories were collected. The results showed that the total energy input and output for green tea leaf production were 34343.51 and 10996.45 MJ/ha, respectively. Also, the total energy input for dry tea was obtained 64885.23 MJ/ton. The results of the life cycle assessment in tea farms revealed that the highest effect belonged to nitrogen in abiotic depletion, dcidification, eutrophication and global warming potential impacts. The survey of life cycle assessment in the tea processing factories illustrated that diesel fuel had the highest share in abiotic depletion, ozone layer depletion, human toxicity, fresh water aquatic ecotoxicity and photochemical oxidation impact categories. In acidification, eutrophication, global warming potential impacts, green leaf input had the highest share. Moreover, the highest share of marine aquatic ecotoxicity and terrestrial ecotoxicity was belonged to coated paper. The results of energy and environmental emissions modeling for green tea leaf production showed that artificial neural networks can predict the outputs and inputs with more accuracy than regression technique.