Deep‐Learning-Based Uncertainty Analysis of Flat Plate Film Cooling With Application to Gas Turbine

Author:

Wang Yaning,Qiu Xubin,Qian Shuyang,Sun Yangqing,Wang Wen,Cui Jiahuan

Abstract

Nowadays, gas turbines intake jet air at high temperatures to improve the power output as much as possible. However, the excessive temperature typically puts the blade in the face of unpredictable damage. Film cooling is one of the prevailing methods applied in engineering scenarios, with the advantages of a simple structure and high cooling efficiency. This study aims to assess the uncertain effect that the three major film cooling parameters exert on the global and fixed-cord-averaged film cooling effectiveness under low, medium, and high blowing ratios br. The three input parameters include coolant hole diameter d, coolant tube inclination angle θ, and density ratio dr. The training dataset is obtained by Computational Fluid Dynamics (CFD). Moreover, a seven-layer artificial neural network (ANN) algorithm is applied to explore the complex non-linear mapping between the input flat film cooling parameters and the output fixed-cord-averaged film cooling effectiveness on the external turbine blade surface. The sensitivity experiment conducted using Monte Carlo (MC) simulation shows that the d and θ are the two most sensitive parameters in the low-blowing-ratio cases. The θ comes to be the only leading factor of sensitivity in larger blowing ratio cases. As the blowing ratio rises, the uncertainty of the three parameters d, θ, and dr all decrease. The combined effect of the three parameters is also dissected and shows that it has a more significant influence on the general cooling effectiveness than any single effect. The d has the widest variation of uncertainty interval at three blowing ratios, while the θ has the largest uncertain influence on the general cooling effectiveness. With the aforementioned results, the cooling effectiveness of the gas turbine can be furthermore enhanced.

Funder

National Natural Science Foundation of China

Publisher

Frontiers Media SA

Reference44 articles.

同舟云学术

1.学者识别学者识别

2.学术分析学术分析

3.人才评估人才评估

"同舟云学术"是以全球学者为主线,采集、加工和组织学术论文而形成的新型学术文献查询和分析系统,可以对全球学者进行文献检索和人才价值评估。用户可以通过关注某些学科领域的顶尖人物而持续追踪该领域的学科进展和研究前沿。经过近期的数据扩容,当前同舟云学术共收录了国内外主流学术期刊6万余种,收集的期刊论文及会议论文总量共计约1.5亿篇,并以每天添加12000余篇中外论文的速度递增。我们也可以为用户提供个性化、定制化的学者数据。欢迎来电咨询!咨询电话:010-8811{复制后删除}0370

www.globalauthorid.com

TOP

Copyright © 2019-2024 北京同舟云网络信息技术有限公司
京公网安备11010802033243号  京ICP备18003416号-3