data-driven modeling of concrete fracture energy using linear genetic programming

The fracture energy of the concrete is an important parameter which can be used to identify the fracture process of concrete members, especially when subjected to tension and flexural loading. practices to measure this property in experiments can be expensive and time-consuming. in this study, a statistical model using linear genetic programming is introduced to predict concretes' fracture energy with three readily measured input parameters, namely, compressive strength, maximum aggregate size, and water to cemented ratio. the model was developed and trained based on a dataset of 64 measured experimental values taken from published research. the performance of the model was evaluated using statistical indices such as the coefficient of determination, root mean squared error, and mean absolute error, and compared with previously proposed empirical models. the experimental results show that the proposed lgp- based model is superior to old regression-based equations in accuracy and generalization. this model can be a useful methodology for engineers in design and analysis, minimizing the need for large amount of laboratory testing.