Proportional distribution method for estimating actual grain flow under combine harvester dynamics

Wang He, Bai Xiaoping, Liang Hongbin
2017 International Journal of Agricultural and Biological Engineering  
The yield monitors use a constant delay time to match the grain flow with location. Therefore, mixing and smoothing effects on the grain flow are neglected. Although constant time delay compensates for time mismatch, actual grain flow at a combine harvester head is not equal to the grain flow measured by a sensor due to the dynamics effects. In order to eliminate the dynamics effects, a new method for estimating actual grain flow, called proportional distribution (PD), is proposed. This method
more » ... posed. This method assumes that actual grain flow is directly proportional to the feedrate. Based on this assumption, the actual grain flow results from redistributing accumulated grain mass over a certain time according to the profile of the feedrate. The PD can avoid the dynamics effects because the feedrate is measured at a combine harvester's head. Compared with constant time delay, the proposed method can effectively estimate actual grain flow and be applied to improve the accuracy of yield maps. Numerous researchers have proposed methods to solve this problem. Previous researches have focused on a simple time delay or a first-order model for the combine harvester as a whole [1] [2] [3] . These models assume that the grain enters combine harvester and goes through combine harvester components without being disturbed until the flow is measured by the yield sensor. They rely on the assumption that shifting of the flow signal suffices for determining the actual field coordinates of yield. In recent researches, in essence, the combine harvester dynamics is of a higher order. The relationship between actual grain flow and measurement grain flow is approximated by a fourth-order linear transfer function [4, 5] . However, these linear flow models fail to reconstruct actual grain flow since combine harvester dynamics is more complex than a linear model. Other interesting researches are concerned with describing the impact of combine harvester dynamics. Boydell et al. [6] applied the control convolution theory to recognize peanut yield variability in smaller regions. Their results showed that the deconvolution method was greater confidence than the simple time delay method. Deconvolution of the July, 2017 Wang H, et al. New method for estimating actual grain flow under combine harvester dynamics Vol. 10 No.4 159 sensor response and the combine's internal processes would result in finding the grain flow through the combine, which could be used to generate more accurate yield maps. Characterizing variation of grain flow inside the combine and how these variations relate to yield variation was studied [7] . It was suggested that it would not be feasible to model all processes within a combine, which mechanistically determine flow rate. They made an assumption that flow rate measured by the yield sensor as a function of time j(t), can be considered as a shift-invariant, linear transform of the flow rate at which crop enters combine head as a function of time, h(t). Then, a convolution function can be established, which relates flow rate j(t), to a characteristic function i(t). This approach is a quick and inexpensive way to characterize the behavior of different combines with different settings, speed, and crops. Whelan and McBratney [8] attempted to better Int J Agric & Biol Eng
doi:10.25165/j.ijabe.20171004.2732 fatcat:wpcfjvgea5fghho26hbihvq4si