Estimation of the Effect of Bio-Admixtures on Concrete Workability Using Linear Regression and Support Vector Machines

In this study, the slump data for cement mixtures with an added dispersing agent have been modelled. The slump measurement for cement mixtures is an important indicator for cement consistency, which affects workability. The slump data consists of five measurements taken at 15-minute intervals, starting with an initial value. Twenty such tests were analysed. When normalising all the data by the initial value, it is observed that the trajectories of the data are reasonably linear. This implies that all the data can be fitted with two parameters, the initial value, and a rate of decay. The conclusion is that if a model can be found for the initial slump value as a function of variables in the process used to produce the dispersant, then a general model to predict the slump behaviour is found. Using data from the process that produces the dispersing agent, a support vector machine approach was used to model the initial values. Reasonable fits are obtained with five inputs.