In real practice, to assess concrete strength in structures, engineers usually use non-destructive tests (NDT) (e.g. rebound hammer or ultrasonic pulse velocity) in addition to destructive tests (DT) that are carried out on cores extracted from the structure. The results of these tests (NDT and DT) are used to identify a relationship (a conversion model) between the non-destructive measured features and the concrete strength. This model can be then used to assess the strength at any location within the structure under consideration, as well as the mean strength and the strength standard deviation (strength variability). In fact, the assessment of concrete strength variability is as important as the mean strength since the mean strength alone cannot provide a clear picture about the concrete under investigation. However, due to the presence of many uncertainties, the reliability of the values estimated by NDTs need to be improved. In the present study, a wide range of concretes having mean strengths of 10–50 MPa and concrete strength variabilities of 10–30% (defined by the coefficient of variation) is analysed. The main target of this paper is to analyse how would the reliability of assessing the concrete strength variability changes if the bi-objective method is used as a model identification approach and the conditional coring concept is applied for selecting core locations. Results are analysed in terms of the minimum number of cores NC that corresponds to a specific uncertainty level, concrete characteristics, and quality of NDT measurements. The results show an important improvement in the reliability of assessing the strength variability when both the bi-objective method and conditional coring are applied together. © 2021, RILEM.

cited By 0