Designing sustainable concrete mixture by developing a new machine learning technique

Concrete is the most used materials in the construction industry known as an environmental pollutant, posing enormous challenges for sustainability regarding resource depletion, energy consumption, and greenhouse gas emission. Therefore, efforts need to be focused on the reduction of environmental impacts of concrete to boost its sustainability. To develop eco-friendly concrete mixtures, this study aimed to investigate the mixture design of sustainable concrete. To this end, six machine learning techniques, including water cycle algorithm, soccer league competition algorithm, genetic algorithm, artificial neural network, support vector machine, and regression, are applied in order to predict the compressive strength of concrete. The accuracy of these methods is compared based upon performance indicators (e.g., mean absolute error), and the equation generated by the most precision model is utilized for mixture proportioning. Consequently, compressive strength, cost, environmental impacts, including embodied CO₂ emission, and energy and resource consumptions are taken into account as sustainability criteria. To integrate these criteria, six types of objective functions are defined and applied and the most efficient sustainable objective function is used to estimate the mixture design of sustainable concrete. Ultimately, the estimated mixtures are compared based on the sustainability index defined by previous studies. The results indicate that water cycle algorithm is the most accurate model with the mean absolute error of 2.86 MPa. Besides, the quadratic distance to the ideal level is the most effective sustainable objective function. Furthermore, increasing the content of cement and super-plasticizer in mixture design deteriorate the sustainability index. Eventually, 16 sustainable mixture proportions are designed, and the most sustainable, the most economical, the eco-friendliest, and the least material-consuming mixtures are presented and compared based on their sustainability indices.