Estimation and Selection of Fuel Consumption Forecasting Model for MF399 Tractor (Case study: Pars Abad Region of Moghan)

Considering the high consumption of diesel fuel in the agricultural sector, it is necessary to find solutions to reduce its consumption, and it will be feasible to have a convenient mathematical model more easily and transparently.Fuel and lubrication costs range from at least 16% to more than 45% of total machine costs, depending on the type of fuel and the amount of time that the tractor or the machine carries out for agricultural operations. Therefore, the fuel consumption index has a significant role in the selection and management of tractors and agricultural equipment. Most budgeting models also use a simple method to estimate the consumption of diesel fuel, but it is needed a model that describes the real conditions of agronomic operations used to compare agricultural machinery management policies.

This case study was conducted in Parsabad city of Moghan, the northernmost province of Ardabil province. The main agricultural products in Pars-Abad Moghan include wheat, maize, maize, canola and sugar beet. The product of this study was irrigated wheat with a crop area of 18042 hectares.In this study, in order to create homogeneous conditions in the study of diesel fuel consumption and the ineffectiveness of the type and model of tractor in it, only diesel fuel consumption was considered by the tractor MF-399. Selection of sample farmers was also carried out among owners of this type of tractor. Selection of owners of tractor MF-399 in Pars-Abad Moghan city was done by random sampling method. For this purpose, Cochran formula was used. Two-way flexible and non-flexible models have been used to predict the diesel fuel consumption. The model used includes the Cobb-Douglas function and transcendental function. Statistical calculations in this study were performed using Excel software and SPSS16 software.

For comparing the best form of the fuel function, the test formulas for the comparison of the form of functions such as bounded least squared F, LR test, White test, Breusch-Godfrey test and Rigorous test were used. Diagnostic statisticians (well-fitting coefficient), the normal distribution of distorted sentences, and the heterogeneity of variance showed that both forms were acceptable. Based on the LR statistic, zero statistics did not rule out the discrepancy between the two coherent models (Cobb-Douglas) and non-dominant (transcendent), but the coherent model was preferable to be the transcendental model because of its simplicity and power of explanation.According to the estimated model, the duration of soil tillage operations had a positive stretch in diesel fuel consumption and, among other variables, had the highest elongation. It should be noted that the average time required for tillage operations was 387.6 min ha-1, which will save 0.31 L ha-1, if one percent of this time (3.9 minutes) is reduced. Thus, the value of the amount of gasoline saved will be about 990 Rials per hectare and equal to 7.7 percent of the value of one kilogram of wheat. Therefore, if the operating time is reduced at the macro level of the country, a significant amount of cost will be saved. Therefore, it is imperative that farm managers take time management in serious soil tillage operations and try to reduce this time.So that, in exchange for an increase of 1% over the duration of the tillage, a fuel consumption of 0.6% would be increased. It is also clear that an increase of 0.6% in fuel consumption for tillage operations is significant, indicating the fact that farm managers have made the need for time management, especially in the tillage operations, to reduce this time. According to the estimated model, the duration of the planting operation also had a positive stretch in the consumption of diesel fuel. So that, in exchange for an increase of 1% over the duration of the planting operation, a fuel consumption of 0.04% would be increased.

Use of the Cobb-Douglas model with five sentences and four independent variables including cropping area, soil tillage operation time, planting time and weeding operation time in order to predict the amount of diesel fuel used to produce wheat, had acceptable results and as a predictive model with low complexity but with high precision, can be easily used in annual budgeting for the production of wheat.