In traditional methods, analyzing the evolution laws of concrete material properties, mechanical response mechanisms, and service behavior characteristics often starts from fragmented experimental data and limited scale engineering case analysis, requiring tedious experimental design, parameter screening, feature extraction, and empirical model construction processes.
The term ‘fragmentation’ here is not due to insufficient data collection by researchers, but rather because concrete materials themselves have characteristics such as multiple components, multiple scales, strong heterogeneity, and complex nonlinearity.
At the same time, they are influenced by multiple factors such as differences in raw materials, changes in mix proportions, curing conditions, load environments, and long-term service factors, like ‘peering into the whole from a local perspective’.
The hydration process of cement, the evolution of the transition zone between aggregate and slurry interface, changes in pore structure, initiation and propagation of microcracks, and the coupling effect of environment and load still result in significant uncertainties in traditional research methods for predicting concrete strength, durability evaluation, damage evolution analysis, and life prediction.
.