Интеллектуальная технология определения оптимального давления воздуха в шинах колёс сельскохозяйственных тракторов

Abstract

The use of powerful tractors and wide-beam agricultural machinery, which have a strong compaction effect on the soil, is conditioned by the agricultural production profitability. The lack of a holistic systematic approach to reducing the anthropogenic compaction effect of wheel propulsors and working units of modern energy-intensive heavy machinery on the soil of agricultural landscapes requires the improvement of methods for determining the optimal air pressure in the tires of agricultural tractors. For this purpose, the authors developed an intelligent technology for determining the optimum air pressure in the various types of tires used in agricultural tractors. The problem was solved by processing “large” arrays of data on operating machine and tractor units and agrolandscapes in order to increase crop yields. Collection and analysis of primary data required for training the neural network were conducted on the fields of the Republic of Adygea during the cultivation of winter barley and winter wheat with the use of machinery equipped with Michelin AXIOBIB2 low-pressure tires. The Feed forward neural network was applied. Fourteen parameters were used as input factors to the neural network: types of soil, machinery and tires; field coordinates; presence and type of mounted equipment; season; type of tillage; granulometric composition, moisture and soil density; wheel diameters; motion speed of machines; field slope; agricultural background. The task set presumed the yield parameter as the main target function. The neural network pre-trained on a significant amount of input data calculates the optimal air pressure in tires when inputting the necessary data. Based on the designed software, the authors plan to develop a system of automatic adjustment of tire inflation depending on the incoming factors made in the offline and online modes.