Due to the limitation of the experimental data or the unknown random effects in the experiment,it may lead to large errors in the parameters estimation during modeling process,and a reliable quality design can-not be obtained.Therefore,the joint confidence region is constructed to quantify the uncertainty of input parameters(model parameters and parameters to be estimated in noise variables).Secondly,according to the distribution information of noise variable and design variable tolerance,a new expected quality loss function based on the confidence region is proposed.Then from the perspective of robustness and economy,optimization object function is constructed based on the interval estimation theory,which includes the location and dispersion effects of quality loss and tolerance cost.Finally,the effectiveness of the proposed method is verified by experimental simulation and industrial example.The results show that the method simultaneously incorporates the two important uncertainty into the objective function.It not only has good robustness to the disturbance of uncertainty,but also can make a reasonable trade-off between quality loss and tolerance cost to achieve a lower total cost.
关键词
输入参数不确定/参数和容差整合设计/区间估计/经济性质量设计
Key words
input parameter uncertainty/integrated design of parameters and tolerances/interval estimation/economic quality design
维普
万方数据