加入正则技巧的非线性反演方法的研究

Nonlinear Inverse Analysis of Adding Regularization Technique

对非线性参数用梯度正则化方法反演做了进一步研究。借鉴模型信赖域的思想 ,给出了确定正则因子的具体办法 ,并建立了算法的迭代过程。以识别四边简支矩形板荷载的作用点为例 ,阐明了该方法的应用。算例结果表明迭代过程收敛速度较快 ,精度高 ,且稳定性好 ,具有较强的抗噪音能力。

It has first made the nonlinear function relationship between the identifying parameters and known observed value linear Then,based on the least squar method,an objective function has been found and an iteration process formed It has applied the regularization technique to avoid sensitive matrix singular and has given a way to choose regularization factor An example of identifying the location of a load on a rectangular plate with four border hinged has described the algorithm in detail The results show that the algorithm not only possesses good capability of noise resistance,but also illustrate its higher convergence speed and better precision