Fatemeh Hosseini-Fashami - M.Sc Student, Department of Mechanics of Biosystem Engineering, Sari Agricultural Sciences and Natural Resources University, Sari, Iran
Ali Motevali - Assistant Professor, Department of Mechanics of Biosystem Engineering, Sari Agricultural Sciences and Natural Resources University, Sari, Iran
Ashkan Nabavi-Pelesaraei - PhD Graduated, Department of Agricultural Machinery Engineering, Faculty of Agricultural Engineering and Technology, University of Tehran, Karaj, Iran
Seyed Jafar Hashemi - Associated Professor, Department of Mechanics of Biosystem Engineering, Sari Agricultural Sciences and Natural Resources University, Sari, Iran

The main aim of this study was to determine modeling of energy indices of greenhouse cucumber production using artificial neural networks in Alborz province of Iran. For this purpose the initial date were collected from 30 greenhouse cucumber producers by a face-to-face questionnaire in the studied area. Total energy consumption and greenhouse cucumber yield were 972247.47 MJ ha-1 and 133227.40 kg ha-1, respectively. Diesel fuel with 83.84% had the highest share of energy use among all of the inputs. The energy indices analysis indicated that energy ratio, energy productivity, specific energy, net energy and energy intensiveness were about 0.11, 0.14 kg MJ-1, 7.30 MJ kg-1, -865665.55 MJ ha-1 and 7.55 MJ $-1, respectively. The Levenberg-Marquardt learning algorithm was trained for calculation of prediction models for energy indices based energy inputs and area. The results of the ANN model revealed the 9-13-5 structure belonged to the best topology with highest R2 and lowest RMSE and MAPE. The rate of R2, RMSE and MAPE was computed between 0.951-0.994, 0.139-0.270 and 0.005-0.019, respectively. Generally, the results of present study illustrated that ANN model can model the energy indices with high accuracy.

Artificial neural network, Cucumber, Energy, Modeling