Switching Markov chains for modelling the loads of a four-rotor swather under different operating modes

Abstract

Modelling and simulating loads of agricultural machinery under different variable loading conditions is an important tool for optimising the design of these machines. The developed models should be based on real conditions that the machine will face during its service life. In this paper, the load-time series from a four-rotor swather were acquired under normal working conditions, performing infield swathing and headland turning. The turning points of these loads were modelled using switching Markov chains, in order to model the switching between these two operating modes (i.e. swathing and headland turning). Based on the developed model, a total of 10,000 Monte Carlo simulations were performed, to assess the proposed methodology. The comparison between the measured and the simulated loads was performed by calculating the correlation coefficient of the power spectral density (PSD) of the two signals. The mean value and the standard deviation of the correlation coefficient when comparing the PSD of all Monte Carlo simulations with the PSD for the measured load, was 0.87 and 0.05, respectively, while the 5th and 95th percentile were 0.78 and 0.94, respectively.