Modelling the sensitivity of fresh state performance to the mix design of conventional and self-compacting concretes

Abstract

The fresh and hardened state performance of conventional and self-compacting concrete (SCC) is affected by multiple factors. Supplementary cementitious materials (SCMs), chemical admixtures, the maximum aggregate size or the grading of aggregates influence how the mix is designed to specified performance requirements. Cohesiveness and workability can influence the hardened state properties of any mix, self-compacting or not. A large dataset compiled from routine industrial testing has been analysed to develop a tool that predicts the workability from the mix design. The dataset includes SCC and conventional mixes, incorporating SCMs and water reducers or superplasticisers at different dosages. Machine learning techniques have been employed to develop a classifier which estimates the slump class of mix designs. This paper describes how the challenges posed by analysing such a broad range of mix designs has been overcome and in what ways the knowledge obtained can be applied.