Study On Cleaning Performance Of Air-and-Screen Cleaning Device For Rape Combine Harvester

In this paper, it is introduced the research status and cleaning device overview about the rape combine harvester both at home and abroad, and analyzes the problems existing in current combine harvester's cleaning device at home. In order to show it is necessary to conduct theoretical and experimental studies of the cleaning device about the rape combine harvester.On the basis of analyses over the collision theory, a motion mathematical model of the mixture on the vibration sieve is established with proper assumption. By analyzing the major factors effect the motion state and motion stability of rapeseeds the motion regularity of rapeseeds with different ejection intensity is founded. A numerical simulation for cleaning airflow field is made by using FLOTRAN model within ANSYS software to investigate the effect which caused by air factors and structure of cleaning device. We can obtain the air distribution inside the cleaning device by changing the input and output condition of the model which is established to educe the major parameters related to it.Based on the theoretical study and numerical simulation to the air flow field and make a further study over the parameters through "Experiment device for threshing, separating, cleaning simulation and control of DF-1.5". After orthogonal test, range analysis and variance analysis are both done for the test data. At last the best suitable parameters are founded for the cleaning device. Change the sieve surface with different sieve apertures in order to change the structural parameter of vibration sieve. Then compare the cleaning performance with each other to find the most suitable structural of sieve surface for rape cleaning.Based on the research of the major effecting factors relate to the performance of rape combine harvester cleaning device. Using the high non-linear mapping function of BP neural networks to create a model between airflow factors, structural & motion parameters of double vibration sieves and the cleaning results .The network model is trained and predicted by the experiment data. The results indicate the high precision of this model and it can be referred as a new and effective method for studying variations of material cleaning.