可靠性全局靈敏度指標(biāo)的空間分割高效方法
發(fā)布時間:2018-06-08 01:02
本文選題:可靠性全局靈敏度指標(biāo) + 連續(xù)無重疊區(qū)間上的全方差公式; 參考:《北京航空航天大學(xué)學(xué)報》2017年06期
【摘要】:可靠性全局靈敏度指標(biāo)能夠有效地分析輸入變量的不確定性對結(jié)構(gòu)系統(tǒng)失效概率的影響程度,為提高該靈敏度指標(biāo)求解數(shù)字模擬法的效率,提出了一種基于密度權(quán)重及連續(xù)無重疊區(qū)間全方差公式的空間分割高效方法。所提方法通過連續(xù)無重疊區(qū)間上的全方差公式來加快該指標(biāo)計算的收斂速度,利用密度權(quán)重法在輸入變量可能的取值區(qū)間內(nèi)進(jìn)行均勻抽樣,并以均勻樣本點的聯(lián)合概率密度函數(shù)的權(quán)重來保證計算的等價性,這使得所構(gòu)造的方法不需要尋找失效域的設(shè)計點,因此其可以有效解決非線性程度較高難以找到設(shè)計點及多設(shè)計點的問題。除此之外,應(yīng)用空間分割技術(shù),使得本文所提方法僅需重復(fù)利用一組樣本點,就可同時得到各個輸入變量的可靠性全局靈敏度指標(biāo),消除了計算量與輸入變量維數(shù)的相關(guān)性,大大地提高了樣本的利用率和計算效率。驗證算例的計算結(jié)果,說明了本文方法對計算功能函數(shù)非線性程度較高及多設(shè)計點問題的高效性。
[Abstract]:The reliability global sensitivity index can effectively analyze the influence of the uncertainty of the input variables on the failure probability of the structural system. In order to improve the efficiency of the sensitivity index to solve the digital analog method, An efficient spatial segmentation method based on density weight and continuous nonoverlapping interval total variance formula is proposed. The proposed method accelerates the convergence rate of the index calculation by the formula of total variance on the continuous non-overlapping interval. The density weight method is used to carry out uniform sampling in the range of the possible values of the input variables. The weight of the joint probability density function of the uniform sample points is used to ensure the equivalence of the calculation, which makes it unnecessary to find the design point of the failure region. Therefore, it can effectively solve the problem that it is difficult to find design points and multiple design points with high degree of nonlinearity. In addition, by using spatial segmentation technique, the reliability global sensitivity index of each input variable can be obtained by using only one set of sample points repeatedly, which eliminates the correlation between the calculation amount and the dimension of input variable. The efficiency of sample utilization and calculation is greatly improved. The results of the examples show that the proposed method is highly nonlinear and efficient for the problem of multiple design points.
【作者單位】: 西北工業(yè)大學(xué)航空學(xué)院;中國飛行試驗研究院飛機(jī)所;
【基金】:國家自然科學(xué)基金(51475370) 中央高;究蒲袠I(yè)務(wù)費(fèi)專項資金(3102015BJ(Ⅱ)CG009) 西北工業(yè)大學(xué)博士論文創(chuàng)新基金(CX201708)~~
【分類號】:TB114.3
,
本文編號:1993653
本文鏈接:http://sikaile.net/kejilunwen/yysx/1993653.html
最近更新
教材專著