基于机器学习预测纳米有机工质导热系数研究  

Prediction of Thermal Conductivity of Nano-organic Working Medium Based on Machine Learning

作　　者：兰先圣 唐美玲[1] 盛伟[1] LAN Xiansheng;TANG Meiing;SHENG Wei(School of Energy and Power and Nuclear Technology Engineering,Shenyang Institute of Engineering,Shenyang 110136,Liaoning Province)

机构地区：[1]沈阳工程学院能源动力与核技术工程学院,辽宁沈阳110136

出　　处：《沈阳工程学院学报(自然科学版)》2025年第2期90-96,共7页Journal of Shenyang Institute of Engineering:Natural Science

摘　　要：在研究纳米有机工质导热系数问题时,为了避免仪器测量的操作烦琐和理论公式的计算误差,并能快速准确地得到纳米有机工质的导热系数,建立基于遗传算法优化的BP神经网络预测模型。将纳米有机工质的种类、温度、纳米颗粒的粒径和体积分数作为神经网络的参考变量,将纳米有机工质的导热系数作为结果,对纳米有机工质导热系数进行非线性预测,预测结果与实验数据高度吻合,证明预测误差很小。对比未优化的BP神经网络,遗传算法优化后的BP神经网络预测精度更高。When studying the thermal conductivity of nano organic working medium,in order to avoid the cumbersome operation of instrument measurement and the calculation error of theoretical formulas,and to quickly and accurately obtain the thermal conductivity of nano organic working medium,this paper establishes a BP neural network prediction model based on genetic algorithm optimization.The type,temperature,particle size,and volume fraction of nano organic working medium are used as reference variables for the neural network,and the thermal conductivity of nano organic working medium as the result,nonlinear prediction of the thermal conductivity of nano organic working medium is carried out.The predicted results are highly consistent with experimental data,which proves that the prediction error is very small.This model can be used to predict the thermal conductivity of nano organic working medium.Compared to the unoptimized BP neural network,the BP neural network optimized by genetic algorithm has higher prediction accuracy.

关 键 词：纳米有机工质 机器学习 神经网络预测 导热系数 

分 类 号：TK124[动力工程及工程热物理—工程热物理]