本文選題：MDOE　切入點：非參數(shù)模型　出處：《空氣動力學(xué)學(xué)報》2016年01期

【摘要】：MDOE風(fēng)洞實驗方法能夠用相對于傳統(tǒng)實驗方法更少的吹風(fēng)次數(shù),獲得更高精準(zhǔn)度的數(shù)據(jù)。為了解決現(xiàn)有基于參數(shù)模型的MDOE方法獲取較強(qiáng)非線性氣動規(guī)律能力的不足,需要發(fā)展基于非參數(shù)模型的MDOE方法。本文對基于非參數(shù)模型的MDOE的形式實驗設(shè)計方法進(jìn)行研究。通過"虛擬"風(fēng)洞實驗的方法,對兩種常用的"空間填充設(shè)計"——拉丁超立方設(shè)計和均勻設(shè)計應(yīng)用于風(fēng)洞實驗的適用性進(jìn)行對比,并在此基礎(chǔ)上發(fā)展了幾種對均勻設(shè)計的優(yōu)化改進(jìn)方法,可以進(jìn)一步提高樣本點設(shè)計質(zhì)量,使其滿足風(fēng)洞實驗的要求。研究表明:均勻設(shè)計較拉丁超立方設(shè)計更為穩(wěn)健、均勻,更適合基于非參數(shù)模型的風(fēng)洞MDOE方法;在均勻設(shè)計方法基礎(chǔ)上,根據(jù)風(fēng)洞實驗的特點發(fā)展了優(yōu)化方法,包括邊界點補(bǔ)充、樣本點密度調(diào)整和重復(fù)點設(shè)計,能夠?qū)⒁延械?先驗信息"應(yīng)用于實驗設(shè)計中;所發(fā)展的形式實驗設(shè)計方法所需的測量點要少于OFAT方法的測量點(如示例中所用的測量點數(shù)僅為OFAT方法的66.7%),且能夠充分和準(zhǔn)確地對較劇烈的非線性變化規(guī)律進(jìn)行采樣。本文對風(fēng)洞MDOE的形式實驗設(shè)計方法的研究結(jié)果,為后續(xù)發(fā)展基于非參數(shù)模型的風(fēng)洞MDOE方法奠定了基礎(chǔ)。
[Abstract]:The MDOE wind tunnel experimental method can obtain more accurate data with less blowing times than the traditional experimental method. In order to solve the shortage of the existing MDOE method based on parameter model to obtain strong nonlinear aerodynamic laws. It is necessary to develop MDOE method based on nonparametric model. In this paper, the formal experimental design method of MDOE based on nonparametric model is studied. This paper compares the applicability of two common "space filling design"-Latin hypercube design and uniform design in wind tunnel experiments, and develops several optimization and improvement methods for uniform design. The design quality of sample points can be further improved to meet the requirements of wind tunnel experiments. The results show that the uniform design is more robust and uniform than the Latin hypercube design and is more suitable for the MDOE method based on non-parametric model. On the basis of uniform design method, the optimization method is developed according to the characteristics of wind tunnel experiment, including boundary point supplement, sample point density adjustment and repeated point design, which can be applied to the experimental design. The developed formal experimental design method requires fewer measuring points than the OFAT method (for example, the measurement points used in the example are only 66.7 points of the OFAT method, and can be fully and accurately applied to the more severe nonlinear variation laws.). Sampling. The research results of the formal experimental design method of wind tunnel MDOE in this paper, It lays a foundation for the further development of wind tunnel MDOE method based on nonparametric model.
【作者單位】： 中國航天空氣動力技術(shù)研究院;
【分類號】：V211.74
，

本文編號：1685941