首页|基于机器学习的齿轮泵瞬态换热特性研究

基于机器学习的齿轮泵瞬态换热特性研究

Study on Transient Heat Transfer Characteristics of Gear Pump Based on Machine Learning

扫码查看
为研究齿轮泵的瞬态换热特性,采用多元线性回归算法构建了齿轮泵的瞬态换热模型,并提出了相应的温度预测模式,避免了复杂的建模过程.对特定时间间隔下齿轮泵出口温度的阶跃响应和余弦响应进行了预测,结果表明仿真与实验的平均相对误差仅为0.74%~1.67%.进行了温度-流量耦合变化下齿轮泵换热特性的预测,相应的误差均不大于2.68%.对不同时间间隔下温度的阶跃响应进行了预测,结果表明在一定范围内,模型可以较好地预测各个时刻齿轮泵的出口温度.本文提出的方法可以提高热管理系统中驱动部件换热的计算效率,为实际的工程应用提供一定的技术支撑.
In order to study the transient heat transfer characteristics of gear pump,a transient heat transfer model of gear pump was developed by multiple linear regression algorithm and the corresponding temperature prediction model was proposed to avoid the complicated modeling pro-cess.The step and cosine responses of the pump outlet temperature with a specific time interval were predicted.The results indicated that the average relative error between the simulation results and the experimental results is only 0.74%~1.67%.The heat transfer characteristics of the pump under the coupling change of temperature and flow were predicted and the error is less than 2.68%.The temperature step responses at different time intervals were predicted.The results showed that the model could predict the pump outlet temperature at each time well within a certain range.The method presented in this paper can improve the computational efficiency of heat transfer of drive components in thermal management system and provide some technical support for the actual engineering applications.

machine learninggear pumptransient heat transferheat capacity

杨世宇、林远方、徐向华、梁新刚

展开 >

清华大学航天航空学院热科学与动力工程教育部重点实验室,北京 100084

机器学习 齿轮泵 瞬态换热 热容

国家科技重大专项

2019-Ⅲ-0001-0044

2024

工程热物理学报
中国工程热物理学会 中国科学院工程热物理研究所

工程热物理学报

CSTPCD北大核心
影响因子:0.4
ISSN:0253-231X
年,卷(期):2024.45(9)