CAJ | 학술논문

本文基于人工神经网络代理模型对某大功率机车用轴流冷却风机叶轮进行优化设计。采用S-A湍流模型和求解三维雷诺平均N-S方程分析叶轮内部流动，以叶片中弧线进口角、出口角和叶片的安装角为设计变量，优化目标函数选择效率和静压升。设计点采用随机离散层取样方式，在几何参数的设计范围内生成样本并进行三维流动分析，以得到目标函数的模拟值；取不同自由参数可变范围，多次优化。优化设计结果与原始模型相比提高效率1.5%，静压升提高87Pa，其非设计点性能也均有所提高，满足设计需要。
본문기우인공신경망락대리모형대모대공솔궤차용축류냉각풍궤협륜진행우화설계。채용S-A단류모형화구해삼유뢰낙평균N-S방정분석협륜내부류동，이협편중호선진구각、출구각화협편적안장각위설계변량，우화목표함수선택효솔화정압승。설계점채용수궤리산층취양방식，재궤하삼수적설계범위내생성양본병진행삼유류동분석，이득도목표함수적모의치；취불동자유삼수가변범위，다차우화。우화설계결과여원시모형상비제고효솔1.5%，정압승제고87Pa，기비설계점성능야균유소제고，만족설계수요。
This paper presents an optimization procedure based on a artificial neural network surrogate model for design of for axial-flow cooling fan impeller. Numerical analysis of air-flow in the impeller has been carried out by solving three-dimensional Reynolds-averaged Navier-Stokes equations with the Spalar-Allmaras turbulence model. The optimization processes has been conducted with three design variables defining the inlet angle, the outlet angle of medial camber line of blade and the setting angle of blade. The efficiency and the static pressure rise as aerodynamic performance parameters have been selected as the objective function for optimizations. The objective function values have been assessed through three-dimensional flow analysis at design points sampled by Random among Discrete Levels sampling in the design space. The optimization processes have been performed many times with the different ranges of design variables. Compared with the original model, the optimization design result shows that the efficiency has improved 1.5% and the static pressure rises 87 Pa respectively. The off-design performance has been also improved in all of the optimum shapes, which meets design requirements.