Crack mouth opening displacement (CMOD) is the most widely used parameter to evaluate the crack width changing process in concrete fracture research. However, it is difficult to accurately and comprehensively describe the multi-crack cracking characteristics of high-toughness concrete with only crack width. This paper investigates the dynamic evolution of the fractal dimension (FD) of surface cracks on polypropylene fiber-reinforced geopolymer (PFRG) during flexural process. The box counting method was used to calculate the FD of cracks in the segmentation images that were generated by deep learning. Crack FD at the peak load and the fracture failure load could be used as indicators of deflection hardening behavior, with threshold values of 0.83 and 0.86, respectively. Crack development can be divided into four stages based on load, FD, CMOD, and energy capacity with time: elastic segment, internal crack development segment, surface crack extension segment, and crack destabilization segment. FD shows an obvious exponential increase in the surface crack extension segment and crack destabilization segment, and the changing trend is basically parallel with the load in the surface crack extension segment. The FD of peak load and fracture failure load showed a trend of increasing and then decreasing as the fiber factor was increased. The FD value peaks at 585 when W/B = 0.35 and at 252 when W/B = 0.38. To achieve the best stress diffusion and crack propagation process and achieve the toughening objective, it is necessary to reasonably match the fiber factor and matrix strength.
Read full abstract