An experimental-computational evaluation of the accuracy of fracture toughness tests on concrete

This research aims to answer the question “Do the round double beam or single edge specimens and associated loading techniques and data reduction methods produce accurate values of fracture toughness for concrete for ‘reasonably’ sized specimens?” A reasonably sized specimen weighs less than approximately 75 lbs [34 kg]. Accuracy is how close the measured value is to the true value. The true value is the fracture toughness that would be obtained from an infinitely large test specimen.;Laboratory measured results from a broad-based experimental program have been combined with cohesive cracking simulations to determine the true fracture toughness of two batches of concrete. The laboratory experimental program consists of three sizes each of single edge and round double beam specimens. The two-parameter, size effect and Barker data reduction methods are used to obtain values of fracture toughness from the test data. The laboratory data produces inconsistent values of fracture toughness. Therefore, the laboratory measured results alone are not enough to determine the true fracture toughness and thus evaluate the accuracy of the various measured values.;Cohesive cracking simulations are used to reproduce the loading and unloading behavior of the laboratory test specimens. By adjusting the input parameters of the simulations, the simulated response can be made to match the laboratory measured response. The fracture toughness value input to the simulation is then the fracture toughness of the batch. This is the inverse analysis data reduction method. The fracture toughness values obtained from inverse analysis agree with the laboratory measured fracture toughness of the largest specimens. Therefore, this value is likely to be the true fracture toughness of the concrete.;The assessment of accuracy shows that the single edge and round double beam specimens, up to 12 inches [305 mm] high, with the two-parameter, size effect and Barker data reduction methods do not produce fracture toughness values within 10% of the true value. Only the inverse analysis data reduction method produces accurate values in the “reasonable” range of sizes.