Prediction of Grouting Efficiency by Injection of Cement Milk into Sandy Soil Using an Artificial Neural Network

Abstract

The grouting efficiency of cement milk injected into sandy soil is estimated by using an artificial neural network. In order to evaluate grouting efficiency, laboratory model tests were conducted on a dry sand bed. Based on the model test results, a neural network model was developed for use in computing normalized grout bulb diameter. A thorough sensitivity analysis was carried out to evaluate the parameters affecting grouting efficiency. Based on the weights of the developed model, a neural interpretation diagram is developed to find out whether the input parameters have direct or inverse effect on the output. A prediction model equation is established with the weights of the neural network as model parameters. The results of the experiment suggest that the developed model can predict grouting efficiency with reasonable accuracy.