A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
Prediction of the Soil Water Retention Curve from Basic Geotechnical Parameters by Machine Learning Techniques
Abstract The determination of the soil water retention curve (SWRC) is a perquisite to solving numerous geotechnical and geo-environmental problems. Laboratory drying or wetting tests are the most efficient way to identify the SWRC while they remain time-consuming, costly and somewhat uncertain. In the recent developments, artificial intelligence and machine learning techniques have arisen as a promising means in various scientific domains including some applications in geotechnical engineering. The UNSODA unsaturated soil hydraulic database (Leij et al. 1996) is here used to provide measured water retention data on a wide range of soils. This data is used to train some of the most common machine learning prediction algorithms (Regression Tree, Support Vector Machine, Gaussian Process Regression, Artificial Neural Network). The training of the models was satisfactory with the drying tests data while higher errors were obtained with wetting tests dataset. Artificial Neural Network (ANN) method was found to give the most accurate predictions. The consistency of the ANN predicted curve for the drying path is finally checked trough a parametric study.
Prediction of the Soil Water Retention Curve from Basic Geotechnical Parameters by Machine Learning Techniques
Abstract The determination of the soil water retention curve (SWRC) is a perquisite to solving numerous geotechnical and geo-environmental problems. Laboratory drying or wetting tests are the most efficient way to identify the SWRC while they remain time-consuming, costly and somewhat uncertain. In the recent developments, artificial intelligence and machine learning techniques have arisen as a promising means in various scientific domains including some applications in geotechnical engineering. The UNSODA unsaturated soil hydraulic database (Leij et al. 1996) is here used to provide measured water retention data on a wide range of soils. This data is used to train some of the most common machine learning prediction algorithms (Regression Tree, Support Vector Machine, Gaussian Process Regression, Artificial Neural Network). The training of the models was satisfactory with the drying tests data while higher errors were obtained with wetting tests dataset. Artificial Neural Network (ANN) method was found to give the most accurate predictions. The consistency of the ANN predicted curve for the drying path is finally checked trough a parametric study.
Prediction of the Soil Water Retention Curve from Basic Geotechnical Parameters by Machine Learning Techniques
Abdallah, Adel (author)
2019-09-25
10 pages
Article/Chapter (Book)
Electronic Resource
English
A model to predict the water retention curve from basic geotechnical properties
| British Library Online Contents | 2003
A model to predict the water retention curve from basic geotechnical properties
| Online Contents | 2003
Hysteresis and Uncertainty in Soil Water-Retention Curve Parameters
| British Library Online Contents | 2014