基于强化学习的端壁气膜冷却孔排布优化
Optimization of Endwall Film-Cooling Hole Arrangement Based on Reinforcement Learning
李晓鹏 1陆成 1汪奇 1杨力1
作者信息
- 1. 上海交通大学机械与动力工程学院,上海 200240
- 折叠
摘要
孔排布是影响端壁气膜冷却效果的一个重要几何因素,本文提出了一种基于强化学习的气膜孔排布优化方法.通过将孔排布的设计看作是一系列开关孔的决策,将其转化为马尔科夫决策过程并利用强化学习方法进行求解.本文将涡轮叶栅抽象简化为弯折收缩通道,以模拟端壁的横向和流向压力梯度环境,并在此压力梯度环境下开展气膜冷却孔排布优化.研究结果为气膜冷却、冲击冷却、柱肋冷却等多种冷却形式的布局优化提供了新的范式.
Abstract
The arrangement of holes is an important geometric factor affecting the effectiveness of end-wall film cooling.This paper proposes a film hole arrangement optimization method based on reinforcement learning.By treating the design of hole arrangements as a series of decisions to open or close holes,it converts the problem into a Markov decision process and solves it using reinforcement learning.The paper abstracts the turbine vane cascade into a bent and contracted passage to simulate the transverse and directional pressure gradient environments of the end-wall,and carries out the optimization of film cooling hole arrangement under this pressure gradient environment.The research results provide a new paradigm for the layout optimization of various cooling forms such as film cooling,impingement cooling,and pin-fin cooling.
关键词
气膜冷却/压力梯度/强化学习/孔排布/优化Key words
air film cooling/pressure gradient/reinforcement learning/hole arrangement/opti-mization引用本文复制引用
出版年
2024
NETL
NSTL
万方数据