基于机器学习的结冰风洞温度场均匀性分析

Uniformity analysis of temperature field in wind tunnel based on machine learning

为提高结冰风洞的试验效率和流场品质,需建立温度快速预测方法以及提高风洞温度场均匀性。针对结冰风洞换热器出口气流温度建立了基于机器学习的预测模型,为解决真实场景下数据量小的问题,采用变分自编码器(VAE)进行数据增强,后采用遗传算法优化XGBoost方法建立最终的温度预测模型;在此基础上,利用已建立预测模型对风洞温度场均匀性及其影响因素进行了分析。结果表明:通过模型预测的换热器出口气流温度与真实值的平均绝对误差(MAE)约为0.60℃,R^(2)分数约为98.38%;试验风速、模拟高度以及回气压力等工况参数对温度场均匀性均有一定影响。

In order to improve the test efficiency and flow field quality of the icing wind tunnel,it is necessary to establish a method of fast temperature prediction and improve the uniformity characteristics of the temperature field of the wind tunnel.In this paper,a prediction model based on machine learning was established for the outlet air temperature of the icing wind tunnel heat exchanger.In order to solve the problem of small data volume in the real scene,the variational autoencoder(VAE)is used for data enhancement,and then the genetic algorithm is used to optimize XGBoost method to establish the final temperature prediction model.On this basis,the temperature uniformity of the heat exchanger airflow and its influencing factors are analyzed by the established prediction model.The results show that the mean absolute error(MAE)of the airflow temperature at the outlet of the heat exchanger between that predicted by this model and the actual value is about 0.60℃,and the R^(2) score is about 98.38%.The working parameters such as test wind speed,simulated height,and return pressure all have certain influence on the uniformity of the temperature field.