Intelligent optimization of FAH concrete mix ratio design considering the carbon emission allocation principles of by-products and critical transportation distances

Low-carbon fly ash (FAH) concrete is a significant focus in the concrete industry. This study proposes an intelligent generative mix design optimization model for low-carbon FAH concrete, which comprehensively considers compressive strength, carbon emissions, and costs. First, based on the allocation principle of environmental impacts for by-products, the carbon emission characteristics of FAH concrete was analyzed with proposing a rapid estimation formula for carbon emissions. Subsequently, the critical transportation distance for using FAH as a cement substitute in concrete was analyzed, taking into account the impact of different transportation methods. Furthermore, a feature selection method for predicting the compressive strength of FAH concrete was presented based on correlation and sensitivity analysis, and compressive strength prediction models for FAH concrete were established using Random Forest (RF), Support Vector Regression (SVR), and Genetic Algorithm-Optimized Artificial Neural Network (GA-ANN) algorithms. Meanwhile, a mix design multi-output regression (MOR) model was developed for FAH concrete. Finally, an intelligent generative mix design optimization model for low-carbon FAH concrete was established, integrating the MOR model and the compressive strength prediction models. The results show that the transportation phase significantly impacts the carbon emissions of FAH concrete, and both the allocation principle of environmental impacts for by-products and transportation methods substantially affect the critical transportation distance of FAH. The proposed feature selection method enhances the predictive performance of the models, with the GA-ANN model demonstrating the highest predictive performance. Optimal solutions for FAH concrete mix design under different optimization conditions were obtained.