This paper analyzes the relationship between the rates of change of elastic strain energy, the strength during the concrete failure process, and proposes that the increased dynamic strength of concrete was caused by the hysteresis effect of energy release—according to the basic principle of energy conversion. Dynamic Brazilian disc tests were carried out on concrete specimens, with diameter of 100 mm, by using the split Hopkinson pressure bar. Test results were obtained through using a gas gun, with an impact pressure of 0.15 MPa, 0.20 MPa and 0.25 MPa, respectively. The dynamic failure process of concrete is then reproduced by numerical calculation methods. Finally, the energy characteristics during the concrete failure process at different strain rates are studied, and the enhancement mechanism of the dynamic strength of concrete is verified. The results showed that the dynamic tensile strength of concrete increased by 9.79% when the strain rate increased by 61% from 60.25 s−1; and when the strain rate increased by 92.8% from 60.25 s−1, the dynamic tensile strength of the concrete rose by 46.28%. The rates of change of both input energy and dissipated energy meet at the peak stress of the material. The increases in rates of change for the two kinds of energy were not synchronized, so excess input energy could be stored as concrete strength increased. As a result, the extra energy stored after peak stress led to a higher degree of concrete fragmentation and greater kinetic energy of the fragment. These results offer research directions for improving the dynamic strength of concrete.

The applicability of mesoscopic models plays an important role in studying the mesoscopic mechanical properties of concrete. In this study, the computerized tomography (CT) test of concrete under uniaxial compression conditions is conducted using a portable dynamic loading equipment developed by Xi’an University of Technology and a medical Marconi M8000 spiral CT scanner. On the basis of damage partition theory, a probabilistic statistical method for determining threshold values is proposed, and a CT test images is obtained and divided into aggregate, hardened cement and hole-crack areas. A ‘structural random numerical concrete model’ is also established on the basis of the coordinates of each pixel unit in CT images. Uniaxial static compression and tensile numerical simulation tests are conducted. Results show that the structural random numerical concrete model can not only reflect the microscopic composition of concrete but also the interfacial transition zone (ITZ) between aggregate and mortar. The ITZ thickness is approximately 0.04 mm, which is close to the real concrete sample ITZ thickness (approximately 10–50 μm). In the two tests, the specimen damage starts from the initial defects, and the damage crack expands through the weaker ITZ around the aggregate. No matter under the action of static tension or compression load, the damage cracks of the sample almost never pass through the aggregate. Most of the many cracks in uniaxial compression are shear cracks. However, many cracks form at the beginning of uniaxial tension, and only one main crack, which is roughly perpendicular to the loading direction, exists in the end.

In this study, the portable power loading device and medical Marconi M8000 spiral CT scanner are used to conduct the uniaxial compression CT scanning test of the concrete. The concrete porosity, hardened cement rate, and aggregate rate are defined, and the variation law of the concrete is analyzed in the uniaxial compression CT test. The proposed method is considered to utilize the value of each CT number, to realize the quantitative partition of the various components of concrete, and to quantitatively evaluate the damage evolution law of the concrete and strain localization during stress. It is shown that damage at the middle section increased from inside to the outside in the uniaxial compression CT test of the concrete.