The Experimental Study And Numerical Simulation Of Air Flow Field Of Rape's Air-and-screen Cleaning

This paper introduced the researching overview about the rape's combine harvester both at home and abroad, analyzed the performance characteristics of the rape's combine harvester produced by domestic leading firms, pointed out the question of high loss rate and low cleanness of cleaning. The air speeds of five fixed positions were measured, which prepared the test data for follow-up analyses and research of cleaning air flow field. The cleanness and loss rate of cleaning test with supply were determined, which can offer basis on looking for a suitable distribution of the air flow field for rape's cleaning. Using multiple linear regress tools, the regress equations between wind-force factors and the air speeds of five fixed positions were set up. Examined by the test data, the equations seem to be effective. The equations offered simple while effective analytic expressions, to adjust the distribution of the air flow field through adjust the wind-force factors. Using neural network tools offered by MATLAB, two BP neural network models were set up, one for wind-force factors and the air speeds of five fixed positions, and the other, for the air speeds of five fixed positions and the cleaning results, include clean rate and losing rate. The prediction results of two models indicated that they are effectual and have advantage with more accurate than multiple linear regress equations, therefore can be used for predicting or controlling. Using FLUENT, a large-scale fluid project emulation software, a numerical simulation for cleaning air flow field were made, the intact equivalence line chart and vector graph of the air flow speed distribution were obtained, which provide more overall understanding to the cleaning air flow field speed distribution. Based on analyzing existing various kinds of characteristics of sieve, a improvement type of the shaky board was put forward, which can divide the materials into layer and separate seeds at the same time.This subject carried on experimental study and numerical simulation to the air flow field of rape's cleaning, set up multiple linear regress equations and BP neural network models of air flow field, which can be used in air flow field prediction and controlling of rape's cleaning effectively. They may be of certain directive significance to the cleaning device development and the research of rape's combine harvester. Introducing numerical simulation to the cleaning device research of rape'scombine harvester is an innovation, which provides a new kind of thought in research of cleaning flow field for the harvester machinery in the future.