Reliability of core test – Critical assessment and proposed new approach

Abstract

Core test is commonly required in the area of concrete industry to evaluate the concrete strength and sometimes it becomes the unique tool for safety assessment of existing concrete structures. Core test is therefore introduced in most codes. An extensive literature survey on different international codes’ provisions; including the Egyptian, British, European and ACI Codes, for core analysis is presented. All studied codes’ provisions seem to be unreliable for predicting the in-situ concrete cube strength from the results of core tests. A comprehensive experimental study was undertaken to examine the factors affecting the interpretation of core test results. The program involves four concrete mixes, three concrete grades (18, 30 and 48 MPa), five core diameters (1.5, 2, 3, 4 and 6 in.), five core aspect ratios (between 1 and 2), two types of coarse aggregates (pink lime stone and gravel), two coring directions, three moisture conditions and 18 different steel arrangements. Prototypes for concrete slabs and columns were constructed. More than 500 cores were prepared and tested in addition to tremendous number of concrete cubes and cylinders. Results indicate that the core strength reduces with the increase in aspect ratio, the reduction in core diameter, the presence of reinforcing steel, the incorporation of gravel in concrete, the increase in core moisture content, the drilling perpendicular to casting direction, and the reduction in concrete strength. The Egyptian code provision for core interpretation is critically examined. Based on the experimental evidences throughout this study, statistical analysis has been performed to determine reliable strength correction factors that account for the studied variables. A simple weighted regression analysis of a model without an intercept was carried out using the “SAS Software” package as well as “Data Fit” software. A new model for interpretation of core test results is proposed considering all factors affecting core strength. The model when calibrated against large number of test data shows good agreement. The proposed model can effectively estimate the in-situ concrete cube strength from core test results.