The complicated nature of concrete due to the variety of the material systems involved, makes it generally impossible to design concrete mixes with desired properties relying on a purely theoretical background and first-principle methods. Within a more utilitarian approach, one does not try to completely understand the causal physics and chemistry behind the processes having an impact on the target property of the design mix but takes advantage of gathered statistics on that property. In a view of this paradigm, industrial statistical methods and data-science frameworks play a crucial role. In this paper, several different methods to tackle the concrete mix design problem are discussed. Special attention is given to the Bayesian statistical approach, which may have a wide range of applications in the industry. In addition, a particular application of the Bayesian methods concerning the quality control problem is presented.
Subscribe to magazine