Using machine learning for carbon-conditioning recycled aggregate concrete modelling

Environmental impact concerns are increasing in today's society globally. Greenhouse effect is known as a natural process that warms the Earth's surface. However, human activities in different industries are increasing the concentrations of greenhouse gases affecting the Earth's temperature. Carbon dioxide (CO2) is the most common gas emitted and rapidly increasing around the word. Australia's emissions have risen in the past three years and emissions in 2020 are projected to be 551 Mt CO2-e. The construction industry potentially can contribute to the decrease of CO2 emission by the usage of sustainable materials such as carbon-conditioning recycled aggregate concrete. Recycled concrete is an eco-friendly material however it has not been typically used for structural applications. One of the reasons is the limited prediction models to obtain concrete mechanical properties, such as compressive strength. Since Machine Learning (ML) techniques have been used to computationally model the most varied types of engineering problems, this paper presents a study of some techniques to investigate how well them can model one of the main characteristics of recycled concrete aggregate (RCA), the compressive strength. The data set characteristics will be explained and what technics can be used for parts of these information to achieve better results. The contribution to the industry is the construction of programs to facilitate further studies and the concrete mixture optimisation.