Scholarlink

<p>In order to assist the development of vibrating type seed metering mechanism for nursery sowing seeder of hybrid rice, it is necessary to study a widely applicable discrete element method (DEM) modelling method for germinated hybrid rice seed. In this study, 15 seed varieties from the main production areas of hybrid rice in mainland China were selected, and three representative varieties were selected for study based on the three-axis size of the germinated seeds. The angle of repose (AOR), vibration characteristics and partial physical parameters of germinated seed were experimentally determined. Combined with the three-dimensional (3D) models of germinated seed obtained by 3D laser scanning technology, a partitioned particle filling method was proposed for constructed the DEM model of germinated seed. Through steepest ascent tests and central composite design (CCD) experiments, the value ranges and response values of calibration parameters (seed–seed static friction coefficient, seed–seed rolling friction coefficient, and seed–stainless steel static friction coefficient) were obtained. A BP neural network based on particle swarm optimisation (PSO-BP) was used for inverse optimisation of the parameters to be calibrated. It was also compared with the traditional response surface methodology (RSM). The results showed that compared with the software’s automatic particle filling method, the partitioned particle filling method proposed in this paper has higher modelling accuracy without a significant increase in the number of filled particles. The parameter calibration results obtained by PSO-BP for the three germinated seed varieties were&#xa0;(0.550, 0.115, 0.581), (0.611, 0.129, 0.665) and (0.643, 0.135, 0.758).&#xa0;The relative errors between simulated and actual AOR being 1.11%, 1.13%, and 2.29%, respectively, superior to RSM’s errors of 2.88%, 2.69%, and 3.32%. By analysing the vibration simulations and experiments with different seeds varieties and seed qualities, the vibration simulations with the parameters obtained from the PSO-BP were closer to the real situation than the RSM, and the mean relative errors of the seed quantities in the nine cells ranged from 7.51% to 13.94%. This study can provide a reference and methodology for DEM simulation modelling, parameters calibration, and simulations of vibrating type seed metering mechanism of germinated hybrid rice seed.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

DEM modelling methods and contact parameter calibration of germinated hybrid rice seed

  • 期刊论文

  • 发表日期：

    2025年11月7日

查看全文
  • Jiangtao Qi,
  • Yuchao Yang,
  • Hui Guo,
  • Panting Cheng,
  • Xv Cong

摘要

In order to assist the development of vibrating type seed metering mechanism for nursery sowing seeder of hybrid rice, it is necessary to study a widely applicable discrete element method (DEM) modelling method for germinated hybrid rice seed. In this study, 15 seed varieties from the main production areas of hybrid rice in mainland China were selected, and three representative varieties were selected for study based on the three-axis size of the germinated seeds. The angle of repose (AOR), vibration characteristics and partial physical parameters of germinated seed were experimentally determined. Combined with the three-dimensional (3D) models of germinated seed obtained by 3D laser scanning technology, a partitioned particle filling method was proposed for constructed the DEM model of germinated seed. Through steepest ascent tests and central composite design (CCD) experiments, the value ranges and response values of calibration parameters (seed–seed static friction coefficient, seed–seed rolling friction coefficient, and seed–stainless steel static friction coefficient) were obtained. A BP neural network based on particle swarm optimisation (PSO-BP) was used for inverse optimisation of the parameters to be calibrated. It was also compared with the traditional response surface methodology (RSM). The results showed that compared with the software’s automatic particle filling method, the partitioned particle filling method proposed in this paper has higher modelling accuracy without a significant increase in the number of filled particles. The parameter calibration results obtained by PSO-BP for the three germinated seed varieties were (0.550, 0.115, 0.581), (0.611, 0.129, 0.665) and (0.643, 0.135, 0.758). The relative errors between simulated and actual AOR being 1.11%, 1.13%, and 2.29%, respectively, superior to RSM’s errors of 2.88%, 2.69%, and 3.32%. By analysing the vibration simulations and experiments with different seeds varieties and seed qualities, the vibration simulations with the parameters obtained from the PSO-BP were closer to the real situation than the RSM, and the mean relative errors of the seed quantities in the nine cells ranged from 7.51% to 13.94%. This study can provide a reference and methodology for DEM simulation modelling, parameters calibration, and simulations of vibrating type seed metering mechanism of germinated hybrid rice seed.