A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
A Data Mining Approach for Predicting Jet Grouting Geomechanical Parameters
Data Mining (DM) techniques is a useful tool to explore complex relations between data with implicit information. It is than a potential tool to apply when a huge data is available and also to discover knowledgement. In this case DM techniques were applied to predict geomechanical properties of laboratory formulations of soil-cement mixtures used in Jet Grouting (JG) geotechnical works for the improvement of ground, mainly soft soils. These properties (uniaxial compressive strength (qu) and Elastic Young Modulus (E0)) are essential to design geotechnical structures against ultimate limit state (ULS) and the serviceability limit state (SLS). In this paper, three DM models, i.e. Artificial Neural Networks (ANN), Support Vector Machines (SVM) and Functional Networks (FN), are applied for qu and E0 predictions and the results compared with Eurocode 2 predictive formula. In order to understand how DM models work, a sensitive analysis procedure was applied to quantify the effect of the key parameters. Furthermore, several experiments were held, by applying of DM techniques in order to estimate E0 normalized by qu over time. The obtained results give a new contribution to understand the behavior of JG material that improves the construction control process of JG columns and reduces the costs of laboratory formulations.
A Data Mining Approach for Predicting Jet Grouting Geomechanical Parameters
Data Mining (DM) techniques is a useful tool to explore complex relations between data with implicit information. It is than a potential tool to apply when a huge data is available and also to discover knowledgement. In this case DM techniques were applied to predict geomechanical properties of laboratory formulations of soil-cement mixtures used in Jet Grouting (JG) geotechnical works for the improvement of ground, mainly soft soils. These properties (uniaxial compressive strength (qu) and Elastic Young Modulus (E0)) are essential to design geotechnical structures against ultimate limit state (ULS) and the serviceability limit state (SLS). In this paper, three DM models, i.e. Artificial Neural Networks (ANN), Support Vector Machines (SVM) and Functional Networks (FN), are applied for qu and E0 predictions and the results compared with Eurocode 2 predictive formula. In order to understand how DM models work, a sensitive analysis procedure was applied to quantify the effect of the key parameters. Furthermore, several experiments were held, by applying of DM techniques in order to estimate E0 normalized by qu over time. The obtained results give a new contribution to understand the behavior of JG material that improves the construction control process of JG columns and reduces the costs of laboratory formulations.
A Data Mining Approach for Predicting Jet Grouting Geomechanical Parameters
Tinoco, J. (author) / Correia, A. Gomes (author) / Cortez, P. (author)
GeoHunan International Conference 2011 ; 2011 ; Hunan, China
2011-05-23
Conference paper
Electronic Resource
English
Mining Grouting: A Rational Approach
| British Library Conference Proceedings | 2003
Mining Grouting: A Rational Approach
| ASCE | 2003
| British Library Online Contents | 2014
Grouting knowledge discovery based on data mining
| Elsevier | 2019