Evaluating and forecasting the initial and final setting times of self-compacting concretes containing mineral admixtures by neural network

Abstract

An experimental study was conducted to investigate the effects of using binary, ternary, and quaternary cementitious blends of portland cement (PC), fly ash (FA), ground granulated blast furnace slag (GBS), silica fume (SF), and metakaolin (MK) on initial and final setting times of self-compacting concretes (SCCs). For this purpose, a total of 65 SCC mixtures were prepared at two different water binder ratios. Furthermore, based on the experimental results, neural network (NN) model-based explicit formulations were developed to predict the initial and final setting times of SCCs in terms of the amount of concrete constituents, namely mixing water, PC, FA, GBS, SF, MK, fine (fa) and coarse (ca) aggregates, and superplasticizer (SP). The test results have revealed that the mineral admixtures were very effective on the initial and final setting times of SCCs. Besides, it was found that the model developed by using NN seemed to have a high prediction capability of initial and final setting times of SCCs.