Inverse Structural Analyses on Small Punch Tests, with Model Reduction and Stochastic Approach

Abstract

Integrity and durability of metal structures at present can be economically assessed by quasi-non-destructive experiments such as Small Punch Tests already envisaged by international industrial codes. In this communication the following innovative provisions are presented and proposed in view of their contributions to accuracy and economy of diagnostic analyses of structural components: the test is simulated by a finite element computer code; parameters in a popular elastic-plastic material model are computed by inverse analysis; diverse parameters identification procedures are comparatively employed; Proper Orthogonal Decomposition is employed for preliminary model reduction in order to make parameters identification more economical for multiple engineering applications; Kalman filter approach is adopted for stochastic back-analysis in structural diagnosis.