摘要：在三維粘性流場的數(shù)值計算程序平臺上，利用BP神經(jīng)網(wǎng)絡(luò)和遺傳算法，通過葉片彎掠技術(shù)對一軸流風(fēng)機(jī)的轉(zhuǎn)子葉片的周向彎曲角度進(jìn)行尋優(yōu)，以使風(fēng)扇的氣動性能進(jìn)一步提高。通過對比優(yōu)化前、后的葉輪發(fā)現(xiàn)，優(yōu)化之后的葉片呈現(xiàn)明顯的周向前彎曲特征。測試結(jié)果顯示，其全壓和氣動效率分別提高了3.56％和1.27％，失速裕度大幅度拓寬36％以上，上、下端部的損失進(jìn)一步降低。
關(guān)鍵詞：周向前彎葉片，人工神經(jīng)網(wǎng)絡(luò)，遺傳算法，優(yōu)化設(shè)計
中圖分類號：TH432.1   文獻(xiàn)標(biāo)識碼：A
 
Optimization Design Based on Skewed and Swept Blade TechniqueLI Yang OU YANG-Hua DU Zhao-Hui
(Turbomachinery Laboratory of Shanghai Jiaotong University, Shanghai 200030, China)
Abstract: Based on a program for solving 3D viscous flow fields, an aerodynamic optimization design was conducted to the rotor blade (archetypal rotor blade) of anaxial flow fan with back-propagation neural network andgenetic algorithm. By skewed and swept blade technique, the optimized rotor blade having better aerodynamic performance is obtained. The results show that the optimized rotor blade is circumferential forward-skewed blade. Compared with the archetypal rotor blade, the total pressure rise and the total pressure efficiency of the optimized rotor are increased by 3.56% and 1.27% respectively. Its stall margin is greatly extended by more than 36%. At the same time, the loss in the upper and lower endwalls of the fan is reduced further.
Key words: Circumferential forward-skewed blade; Artificial Neural Network (ANN); Genetic Algorithm (GA); Optimization design
    1前言
隨著計算機(jī)技術(shù)和各種尋優(yōu)算法的不斷發(fā)展，在葉輪機(jī)械領(lǐng)域，利用某些算法進(jìn)行葉片的優(yōu)化設(shè)計來提高葉輪的性能已經(jīng)成為可能。許多學(xué)者利用遺傳算法、模擬退火法、梯度法、響應(yīng)面法等對各種葉輪的靜子和轉(zhuǎn)子葉型進(jìn)行了優(yōu)化設(shè)計，結(jié)果顯示，通過一些優(yōu)化方法可以進(jìn)一步地提高葉輪的效率，改善葉片表面的壓力分布，降低了邊界層的流動損失，同時也縮短了設(shè)計周期【1－7】。
然而，當(dāng)前的優(yōu)化設(shè)計多是針對改進(jìn)二維葉型，對于能夠減小葉輪的內(nèi)部流動損失、提高氣動效率最為重要的手段之一的彎掠葉片技術(shù)【8－10】，在葉片優(yōu)化計算中應(yīng)用的公開報告還比較少。
本文利用雷諾平均N-S方程組數(shù)值計算程序，基于人工神經(jīng)網(wǎng)絡(luò)BP算法和遺傳算法的數(shù)值優(yōu)化程序，基于葉片彎掠技術(shù)對一軸流風(fēng)扇轉(zhuǎn)子葉片進(jìn)行了優(yōu)化設(shè)計，通過尋找合適的周向彎曲角度，來獲得具有最優(yōu)氣動性能的風(fēng)扇葉片，并對優(yōu)化前、后的葉片在葉片形狀、氣動性能以及出口流場進(jìn)行了對比分析。
    2研究模型
    詳細(xì)細(xì)內(nèi)容請瀏覽附件部分。