摘要
拖拉机的传动系统结构较为复杂,是拖拉机的重要组成部分之一。拖拉机在农业作业过程中的轴扭矩(AT)实时数据是实现变速器优化的重要依据之一。为此,以拖拉机参数(发动机扭矩、发动机转速、燃油消耗率、行驶速度、耕作深度和滑移率)及土壤理化性质参数(SMC和CI)为输入,基于人工神经网络(ANN)估计拖拉机的轴扭矩(AT),并与传统的多元线性回归模型(MLP)进行对比分析。田间试验结果表明:基于ANN的模型在预测拖拉机AT数据时表现出更好的性能,可为提升拖拉机发动机管理系统提供技术参考与借鉴。
The transmission system of tractor has a more complex structure and is one of the important components of tractor.The real-time data of axle torque(AT)of the tractor during agricultural operation is one of the important bases to achieve transmission optimization.In this study,the tractor parameters(engine torque,engine speed,fuel consumption rate,travel speed,tillage depth and slip rate)and soil physical and chemical property parameters(SMC and CI)were used as inputs to estimate the tractor axle torque(AT)based on artificial neural network(ANN)and compared with the traditional multiple linear regression model(MLP)for analysis,and the results of field trials showed that the The field test results showed that the ANN-based model exhibited better performance in predicting tractor AT data,and the results of the study provide technical references and references for improving tractor engine management system.
作者
姚鹏飞
王丹丹
王瑞红
Yao Pengfei;Wang Dandan;Wang Ruihong(Huanghe Jiaotong University,Wuzhi 454950,China)
出处
《农机化研究》
北大核心
2024年第5期240-246,共7页
Journal of Agricultural Mechanization Research
基金
河南省高校国家级大学生创新创业训练计划项目(202013498009S9)。
