Analysis of the accuracy of sampling procedures for coring locations in non‐destructive in situ concrete strength assessment

Abstract The structural capacity of existing buildings is frequently questioned due to various reasons, including age, exposure to harsh environments, material degradation, and changes in building codes or use. The non‐destructive assessment of in situ concrete strength, achieved by combining non‐destructive testing (NDT) with destructive coring, offers an interesting approach to addressing this concern. Conditional coring, based on information provided by previous NDT measurements, significantly improves the accuracy of non‐destructive in situ concrete strength assessment. However, limited background information on how to select coring locations is available in the scientific literature. The real challenge is to develop a practical sampling procedure for where to extract cores, with a reduced number, ensuring accurate coverage of the full range of strength variability in the structure. This minimizes errors in strength evaluation and ensures an accurate assessment of the structure's overall capacity. In this study, eight practical sampling procedures for coring locations are proposed, encompassing two partition variants (size and range), each with four selection options (random, median, mean, and extreme). These procedures aim to identify the most suitable coring locations, leading to a more accurate strength assessment. Synthetic data generated through a simulation approach, comprising rebound hammer and core strength values, served as the basis for this analysis. An extensive statistical analysis of the proposed sampling procedures, based on the root mean square error and the risk as statistical indicators of wrong strength assessment, using the Monte Carlo simulation process, highlighted the most accurate sampling procedures for core extraction to ensure better coverage of the entire range of concrete strength. The results show that the extreme selection option in size or range partition variants is highly recommended as a sampling procedure for coring locations, compared to other proposed selection options, especially when a low number of cores is used.