基于深度强化学习的轴流式通风机流型优化应用研究  

Application Research of Flow Type Optimization Design of Axial Flow Fan Based on Deep Reinforcement Learning

作　　者：洪煜凯 胡国军 王灿星[1] Yu-kai Hong;Guo-jun Hu;Can-xing Wang(School of Aeronautics and Aerospace,Zhejiang University;Hangzhou Dunli Electric Appliance Co.,Ltd.)

机构地区：[1]浙江大学航空航天学院流体工程研究所 [2]杭州顿力电器有限公司

出　　处：《风机技术》2023年第3期89-93,共5页Chinese Journal of Turbomachinery

基　　金：国家自然科学基金重点项目(12132015,11872333)。

摘　　要：随着机器学习方法的深入研究,其应用范围越来越广,在风机方面的应用也越来越得到关注。本文采用深度确定性策略梯度算法建立了一种轴流式通风机流型优化设计方法,运用该方法对某轴流式通风机进行优化设计,并通过数值模拟对比优化前后的风机的气动性能。结果表明,在大部分工况下,优化后的风机效率均有明显提升,说明机器学习方法用于轴流式通风机的优化设计是可行的,这为风机的优化设计提供了一种新的途径。With the intensive study of machine learning methods,its applications are becoming more and more extensive,and its application in fans has also attracted more and more attention.In this paper,deep deterministic policy gradient algorithm is used to establish a flow type optimization design method for axial flow fan.This method is used to optimize the design of a certain type of axial fan,and the aerodynamic performance of the fan before and after the optimization is compared through numerical simulation.The results show that in most working conditions,the fan efficiency has significantly improved after optimization which indicates this method is feasible for the optimal design of axial fan,and it provides a new way for the optimal design of axial flow fan.

关 键 词：深度强化学习 深度确定性策略梯度算法 轴流式通风机 优化设计 流型设计 

分 类 号：TH432.1[机械工程—机械制造及自动化]