Development of a cleaning fan for a rice combine harvester using computational fluid dynamics and response surface methodology to optimise outlet airflow distribution

To reduce the variability of cleaning performance caused by uneven airflow distribution, the theory of fan design and the characteristic parameters of threshed rice outputs were used to develop a prototype of cleaning fan. Computational Fluid Dynamics (CFD) was employed to simulate the air distribution inside the fan and the airflow speed distribution at the fan outlet. Response surface methodology (RSM) with three factors and three levels was employed to optimise fan impeller geometry parameters, this included blade inlet mounting angle, blade curvature and blade inlet curvature. According to the least squares analysis of RSM, the influence of each factor on the airflow distribution at each sub-outlet was estimated. The optimal values of blade inlet mounting angle, blade curvature and blade inlet curvature were 107°, 190 mm and 135 mm, respectively. The uniformity of the outlet airflow distribution of the optimised fan was clearly improved. The number of eddy currents in the flow passage were reduced, and the transporting performance of optimised fan was improved. The maximum pressure at 5% blade height away from the central plate was reduced by 37.78%, which indicates that the service life of the cleaning fan could be increased.