Improvement of nondestructive assessment of on-site concrete strength: Influence of the selection process of cores location on the assessment quality for single and combined NDT techniques

摘要

For existing constructions, there is a need to estimate the concrete mechanical strength for a more accurate evaluation of their structural capacity. The destructive evaluation of strength by coring is expensive, sometimes technically difficult and it may be impossible in some situations and the evaluation is limited to a restricted surface of the structure. Non-destructive testing (NDT) in conjunction with destructive testing by coring (DT) offers an interesting alternative for the strength assessment of concrete in existing buildings. For a reliable assessment of concrete strength, it is necessary to fully master the calibration quality of the conversion models between DT and NDT used for strength estimation. Because of technical and economic constraints, the professional practice is frequently based on a very limited number of cores. If this optimal number is not reached, the risk is high that the resulting conversion model is not reliable. This practical context, in which the common number of cores is small, forces to review the calibration procedure in order to take care of the relevance of the identified model. The procedure using conditional coring, which is based on the selection of critical locations of cores after an analysis of NDT test results, is the subject of methodological developments presented in this paper. Our research works aims at analyzing this issue on a broad investigation program by NDT (Rebound Hammer and Ultrasonic Pulse Velocity) and coring realized on the structural elements (columns and beams) of an existing building. 205 data triplets (core strengths and NDT test results) are used to carry out this study. The thorough statistical data analysis allows quantifying the potential interest of the conditional coring compared to the random coring with two measurement strategies of NDT (by using single or combined methods). (C) 2018 Elsevier Ltd. All rights reserved.