Identification and control of the cleaning section on a combine harvester

In this study, research was directed to the impact of the machine settings (fan speed, upper and lower sieve opening) on the quality parameters of the crop processing inside the harvester (Material other than grain in the grain bin and sieve losses). Several extra sensors were placed on the combine harvester in order to provide additional on-line information about the crop processing. From a perspective of identification and control, it is important to select the most appropriate input variables from a large pool of potential inputs. By means of a genetic polynomial input selection methodology the 'best' combination of predictor variables for the quality parameters of the harvesting process has been selected. Based on these offline results an algorithm was developed to automatically tune the settings of the cleaning system on-the-go in the field. This control system was found to increase the performance of the cleaning section.