A novel approach to prediction of rheological characteristics of jet grout cement mixtures via genetic expression programming

Abstract

Highly nonlinear flow behavior of cement-based grout mixtures has always become an important issue for experimenters during jet grouting applications. In this viewpoint, an investigation has been addressed in this paper on the applicability of a recent soft computing prediction tool, genetic expression programming (GEP), to the prediction of rheological characteristics (i.e., shear stress, viscosity) of the grout mixtures with various stabilizers (clay, sand, lime) for jet grouting purposes. The experimental data (shear stress versus shear rate with respect to stabilizer dosages) of grout mixtures obtained from rheometer tests have been collected from previous study conducted in a wide range of stabilizer dosage rates (0–100 %, by dry weight of binder). For predicting the shear stress and viscosity as the output variables during the train and testing stages, the input variables in the GEP models included shear rate and stabilizer dosage primarily. As a consequence of GEP modeling compared with measured data, this study reveals satisfactory GEP formulations in a good accuracy (R ≥ 0.86) for predictions of shear stress and viscosity regarding the stabilizer additions. The GEP formulas are also found adequate for modeling the flow behavior of the shear stress–shear rate, alternatively to conventional nonlinear regression and rheological models (Herschel–Bulkley, Robertson–Stiff). For assistance of preliminary evaluations, the derived GEP formulas could be potentially considered in practice for estimations of pumping pressure (shear stress), pumping rate (shear rate) and viscosity of jet grout mixtures.