Evaluation of an intrinsic error estimator for the data fusion of NDT techniques used to identify the material and damage properties of concrete structures

D. Martini (1) , V. Garnier (1) , M.A. Ploix (1)

1 O&I - Ondes et Imagerie

In this paper, we propose an intrinsic error estimator for the data fusion of NDT techniques used to identify the material and damage properties of concrete structures. This error estimator is chosen based on the global distribution of the data fusion in the space of the identified material properties. The main idea is to evaluate the accuracy of the result in estimating the gap between the most and the worst likely solutions. This error estimator is applied to synthetic data depending on the parameters of the data fusion such as the regression laws that linked the material properties to the NDT measurements. This work is part of the C2D2-ACDC project that aims at methodology transfer from five research laboratories, LMA, LMDC, IFFSTAR, GhyMaC and IEMN to industrial partners, EDF and SETRA.

Domaines: Acoustique [physics.class-ph]; Acoustique [physics.class-ph]

D. Martini, V. Garnier, M.A. Ploix. Evaluation of an intrinsic error estimator for the data fusion of NDT techniques used to identify the material and damage properties of concrete structures. 6th European Workshop on Structural Health Monitoring, 2012, Dresde, Germany. ⟨hal-01300009⟩