Prediction model for strength of cement-stabilized clays for in-situ soil mixing using artificial neural network

1. Home
2. Publications
3. Online Library
4. Prediction model for strength of cement-stabilized clays for in-situ soil mixing using artificial neural network

Publication Title
Prediction model for strength of cement-stabilized clays for in-situ soil mixing using artificial neural network

Abstract
Unconfined Compressive Strength (UCS) is usually the design criterion for in-situ ground improvement projects, including deep soil mixing, mass soil stabilization, and cutter soil mixing. Unlike subgrade stabilization projects, such techniques involve cement-content as high as 50%. Thus, the paper intends to develop a prediction model for the strength of cement-stabilized clay from the laboratory test results performed on high-plastic illite clay. First, the UCS tests are performed for cement-content varying from 5% to 50% by dry weight, for two water-contents of 1.5 and 1.85 times the liquid limit (LL), and the curing time of 7, 28, and 90 days. Test results indicate that the UCS increases with the increase in curing time and cement-content while reducing with the increase in water-content. The prediction model for UCS is then developed using the artificial neural network (ANN), and the model performance is assessed based on root mean squared error (RMSE) and coefficient of determination (R2). According to the multi-layer perceptron (MLP) approach, the developed model equation can be used to forecast the performance of the UCS for the given range of water-content, cement, and curing time. Relative importance shows that cement-content is the most significant parameter on UCS.
Authors
M. Swamynaidu; Akanksha Tyagi

Published In

18th European Conference on Soil Mechanics and Geotechnical Engineering (ECSMGE2024)

Theme

D - Current and new construction methods

Download PDF