基于稳健设计优化技术的缺陷参数红外识别研究

Robust Design Optimization Algorithm Applied in Defect Parameters Identification in Infrared NDT/E

将缺陷参数的红外识别问题描述成多目标优化问题并引入了稳健设计优化(RDO)技术。传热反问题控制方程的不适定性和边界条件不确定因素的波动,造成了优化解通常不具有稳健性(robustness)。稳健设计优化技术在取得目标的最优解的同时,使该目标在可行区域内对变量的变化不敏感,使得结果不仅更为可靠,且具有很强的灵活性。最后给出了稳健设计优化技术在沉孔型缺陷参数红外识别中一个简单的应用例子。

The defect parameters identification in infrared NDT/E is described as a multi-objective problem and the robust design optimization （RDO） is introduced to solve it. The robustness of the optimization resuits is decreased by the ill-posedness of the control equations of heat transfer and the uncertainties of the boundary conditions. The RDO obtains an optimum value insensitive to variations on the variables within the feasible region. The flexibility, efficiency and robustness of the method are validated by a simple case in infrared NDT/E.