Abstract
The taro harvesting process is affected by a complex system composed of particle mechanics system and multi-body dynamics system. The discrete element method(DEM) can effectively solve the nonlinear problem of the interaction between harvesting components and working materials. Therefore, the discrete element model of taro tiller plants is of great importance for taro harvesting. This paper proposes a simulation method to establish a discrete element flexible plant model and dynamic clamping and pulling process of taro tiller plant. Discrete Element models of taro corm and flexible tiller petiole and leaf were established using DEM method, and the discrete element flexible model of the taro plant was established. Taro clamping and pulling force testing platform was designed and built. The single factor and Plackett-Burman experiments were used to determine the simulation parameters and optimize the taro plant model by taking the correlation coefficient of clamping force and correlation coefficient of pulling force collected from the simulation and the bench experiment as the experiment index. The parameter calibration results of discrete element model of taro plant are as follows: petiole-petiole method/tangential contact stiffness was 8.15×10(9) N·m(-3), and normal/tangential critical stress was 6.65×10(6) Pa. The contact stiffness of pseudostem- corm method was 1.22×10(9) N·m(-3), the critical stress of normal/tangential was 1.18×10(5) Pa, and the energy of soil surface was 4.15×10(6)J·m(-3). When the pulling speed is 0.1, 0.2, 0.3, 0.4 and 0.5 m·s(-1), the correlation coefficients between the simulation experiment and the bench experiment are 0.812, 0.850, 0.770, 0.697 and 0.652, respectively. The average value of correlation coefficient is 0.756, indicating that the simulated discrete element plant model is close to the real plant model. The discrete element model of taro plant established in this paper has high reliability. The final purpose of this paper is to provide a model reference for the design and optimization of taro harvester by discrete element method.