Artificial intelligence model for rheological properties of oil well cement slurries incorporating SCMs

Abstract

In this study, an artificial intelligence model has been developed to predict the rheological properties of oil well cement (OWC) slurries incorporating supplementary cementitious materials (SCM) such as metakaolin (MK), silica fume (SF), rice husk ash (RHA) or fly ash (FA). An experimental study has been carried out to create the database used for training the model. OWC slurries having a water-to-binder ratio of 0·44 along with a new-generation, polycarboxylate-based, high-range water-reducing admixture (PCH) were prepared. They had 5 to 15% partial replacement of API class-G OWC by MK, SF, RHA or FA. The rheological properties of the slurries were investigated at different temperatures in the range 23 to 60°C using an advanced shear-stress/shear-strain controlled rheometer. Experimental data thus obtained were used to develop a predictive model based on feed-forward back-propagation artificial neural networks. The developed model could effectively predict the effect of key variables such as temperature, dosage of SCM and dosage of PCH on the rheological properties of OWC slurries with an absolute error of less than 7%. The developed model could also effectively predict the rheological properties of new slurries designed within the range of input parameters of the experimental database used in the training process.