A low-cost digital image correlation based constitutive sensor

Abstract

In this paper, a low-cost digital image correlation-based constitutive sensor with a novel identification algorithm that is deployable and scalable in the field is proposed. The term 'constitutive sensor' is coined herein to describe a sensor that is capable of determining the target material constitutive parameters. The proposed method is different from existing identification methods in that it does not need to solve boundary value problems of the target materials using updated parameters. Since the development of the digital image correlation (DIC) technique in the 1980s, the DIC technique has been broadly evaluated and improved for measuring full-field displacements of test specimens, mainly in laboratory settings. Although its potential in damage and mechanical identification is immense, the high cost of current commercial DIC systems makes it difficult to apply the DIC technique to in-field health monitoring of structures. To realize a first ever application of DIC in the field, a prototypical low-cost sensing unit consisting of a high performance embedded microprocessor board, a low-cost web camera, and a communication module is suggested. In the proposed constitutive sensor, DIC displacement fields considered as true values are used in computing stress fields satisfying the equilibrium condition and strain fields using finite element concepts. The unknown constitutive law is initially assumed to be fully anisotropic and linear elastic. A steady state genetic algorithm is utilized to search for the material parameters by minimizing a cost function that measures energy residuals. The main features that allow the sensor to be deployable in the field are introduced, and a validation of the proposed constitutive sensor concept using synthetic data is presented.