关键词: 减基法, 参数反求, Galerkin映射, 非奇异, 信赖域

Abstract:

An effective parameter identification technique  which was based on the Galerkin mappingreduced basis method was proposed for dealing with parameterized structural problems. The investigated structure was firstly discretized by the finite element method. And then using the angle of vector subspace, the nonsingular monitoring basis matrix and corresponding whole basis matrix were constructed in the positions of monitoring points. Subsequently, the identification model of structural parameters in the reduced basis model was constructed by minimizing discrete L2 norm of vector. Finally, the parameters of original problem were identified using the corresponding trust region optimization technique. An example herein demonstrates the presented method has higher computational accuracy and total efficiency than that of the existing classical methods.

Key words: reduced basis method, parameter identification, Galerkin mapping, nonsingular, trust region