Prediction of dynamic toughness of UHPCC type concrete based on mechanical properties using meta-heuristic algorithms

Concrete is a fundamental material in civil engineering structures, known for its mechanical and behavioral characteristics, which often result in cracking post-construction. Understanding and predicting the fracture toughness of concrete is critical for ensuring the safety and durability of infrastructure, especially under dynamic loading conditions. In fracture mechanics, the stress intensity factor is compared with toughness rather than stress and strength, influenced by numerous geometric and physical parameters, and must be determined using experimental specimens. Compressive strength is a unique attribute of cementitious materials, universally applicable to all concrete structures and easily measured post-construction through coring tests. This study addresses the significant gap in dynamic toughness predictions based on stochastic calculations by proposing a novel modeling approach using meta-heuristic methods. The goal is to accurately estimate the fracture toughness of Ultra-High Performance Cementitious Composite (UHPCC) concrete from its compressive strength, thereby providing engineers with a reliable tool to enhance structural integrity. We collected extensive experimental data and conducted a comprehensive statistical analysis, employing 27 parameters to approximate the fracture toughness of UHPCC concrete using several advanced algorithms: genetic algorithm, artificial neural network, support vector machine, multivariate regression, gene expression algorithm, time series, and particle swarm optimization. Among these, the Gene expression programming algorithm (GEP) was identified as the most accurate model, yielding RMSE, MAE, VAF, MAPE, and R2 values of 0.41, 0.06, 98.88, 0.33, and 0.99, respectively. Furthermore, a multi-parameter sensitivity analysis based on GEP revealed that the unconfined compressive strength of the samples had the most significant impact on the dynamic toughness prediction model, with a sensitivity analysis rate of 333. This approach not only enhances the predictive accuracy but also contributes to the advancement of resilient and durable infrastructure. These findings offer a rapid and reliable method for predicting the fracture toughness of UHPCC concrete, providing a valuable tool for structural engineers to design and assess concrete structures under dynamic loading conditions.