您好，欢迎访问北京市农林科学院 机构知识库！
北京市农林科学院机构知识库

全部

  • 全部
  • 标题
  • 学者
  • 机构
  • 关键词

Ensemble Learning Simulation Method for Hydraulic Characteristic Parameters of Emitters Driven by Limited Data

文献类型： 外文期刊

作者： Yu, Jingxin 1 ; Zhangzhong, Lili 1 ; Lan, Renping 3 ; Zhang, Xin 4 ; Xu, Linlin 2 ; Li, Jingjing 4 ;

作者机构： 1.Natl Engn Res Ctr Intelligent Equipment Agr, Beijing 100097, Peoples R China

2.China Univ Geosci, Sch Land Sci & Technol, Beijing 100083, Peoples R China

3.Jiangsu Univ, Sch Agr Engn, Zhenjiang 212013, Peoples R China

4.Natl Engn Res Ctr Informat Technol Agr, Beijing 100097, Peoples R China

5.Minist Agr & Rural Affairs, Key Lab Qual Testing Software & Hardware Prod Agr, Beijing 100097, Peoples R China

6.Univ Waterloo, Dept Syst Design Engn, Waterloo, ON N2L 3G1, Canada

关键词： ensemble learning; irrigation; CFD; data-driven; numerical simulation

期刊名称：AGRONOMY-BASEL （ 影响因子：3.7； 五年影响因子：4.0 ）

ISSN：

年卷期： 2023 年 13 卷 4 期

页码：

收录情况： SCI

摘要： The emitter is one of the most critical components in drip irrigation. The flow path geometry parameters have a significant effect on the emitter's hydraulic performance and have a direct impact on the emitter's irrigation uniformity and lifetime. The hydraulic characteristics of the emitter are the key indicators of its performance. However, obtaining the hydraulic characteristics of the emitter is complex. Typically, only a small number of calibrations are performed for specific equipment models, making it difficult to obtain the parameter. Therefore, limited data corresponding to the morphological parameters and the flow rate were simulated using the FLUENT software, and the influence of the characteristics was analyzeanalyzed, based on which a flow rate prediction model was constructed using the ensemble learning (CatBoost) model. The extended data set was generated by stochastic simulation and parameter fitting. The flow index and flow coefficient prediction model were built and evaluated using the CatBoost model again with the augmented data set as a benchmark. The results show that the significant correlation between the geometric structure and the flow index and flow coefficient provides the basis for the correlation model. CatBoost can fit the complex nonlinear relationships between the parameters well, achieving excellent simulation accuracy for the flow rate (R-2 = 0.9987), flow index (R-2 = 0.9961), and flow coefficient (R-2 = 0.9946), where the path width has the highest importance score in the model construction for the flow index (score = 55.97) and flow coefficient (score = 45.2). Furthermore, the CatBoost models used in this study achieved the best prediction results compared to seven typical models (XGBoost, Bagging, Random Forest, Tree, Adaboost, and KNN).

  • 相关文献
作者其他论文 更多>>
置顶
购物车
反馈

© 京ICP备09089781号-29 主办单位：北京市农林科学院

学术资源： 中国农业科学院 农业专业知识服务系统 农科联盟资源共建共享平台 中国农业科技文献与信息服务平台 中国农业期刊集成服务平台