Сельскохозяйственные машины и технологии (Aug 2017)

Artificial neural network applying for justification of tractors undercarriages parameters

  • V. A. Kuz’Min,
  • R. S. Fedotkin,
  • V. A. Kryuchkov

DOI
https://doi.org/10.22314/2073-7599-2017-4-24-30
Journal volume & issue
Vol. 0, no. 4
pp. 24 – 30

Abstract

Read online

One of the most important properties that determine undercarriage layout on design stage is the soil compaction effect. Existing domestic standards of undercarriages impact to soil do not meet modern agricultural requirements completely. The authors justify the need for analysis of traction and transportation machines travel systems and recommendations for these parameters applied to machines that are on design or modernization stage. The database of crawler agricultural tractors particularly in such parameters as traction class and basic operational weight, engine power rating, average ground pressure, square of track basic branch surface area was modeled. Meanwhile the considered machines were divided into two groups by producing countries: Europe/North America and Russian Federation/CIS. The main graphical dependences for every group of machines are plotted, and the conforming analytical dependences within the ranges with greatest concentration of machines are generated. To make the procedure of obtaining parameters of the soil panning by tractors easier it is expedient to use the program tool - artificial neural network (or perceptron). It is necessary to apply to the solution of this task multilayered perceptron - neutron network of direct distribution of signals (without feedback). To carry out the analysis of parameters of running systems taking into account parameters of the soil panning by them and to recommend the choice of these parameters for newly created machines. The program code of artificial neural network is developed. On the basis of the created base of tractors the artificial neural network was created and tested. Accumulated error was not more than 5 percent. These data indicate the results accuracy and tool reliability. It is possible by operating initial design-data base and using the designed artificial neural network to define missing parameters.

Keywords