risk-based selection of constitutive models fot geotechnical analysis – RISMOGEO

Soil, as a typical natural material, has very complex behavior which is very difficult to be accurately modeled. In geotechnical engineering, about half of the accidents are caused by insufficient designs. The soil’s constitutive model is the fundamental issue for modeling and analyzing soil-related engineering systems. Many constitutive models have been proposed in the literature. Yet, very limited researches have been conducted to examine the applicability of these models. The problem of how to select the appropriate constitutive model for geotechnical analysis and modeling is seldom addressed. As misuse of the constitutive model may mislead engineering decisions, lack of research on constitutive model selection has been one of the sources for introducing risk analysis into geotechnical analysis design. The objective of this research is to develop innovative theories, tools, as well as model application examples for the selection of constitutive models.
Selection of constitutive model should consider factors like difficulties in determining the parameters of the model, magnitude of model error, and the consequence resulting from using the constitutive model. While simplified constitutive models may have parameters easier to determine, it may have larger model error than more sophisticated models. Also, the consequence of using non conservative models may be less acceptable than using conservative models. In this research, we will develop a risk model to calculate the risk associated with using a specific constitutive model or using a set of constitutive models, where the aforementioned important factors on model selection can be explicitly considered. Then extensive laboratory tests will be conducted to determine the variability of parameters of different constitutive models. Past projects with quality monitoring data will be reanalyzed to learn knowledge about the error associated with the use of a specific constitutive model or the use of a set of constitutive models. The laboratory test results as well as case study will provide valuable information to similar projects. To incorporate the site-specific data into the risk assessment, a method to update the risk model will also be developed based on Bayer’s theorem. To facilitate practical application, software interfaced with existing geotechnical numerical programs will also be developed. To illustrate the benefits of the proposed research, model application examples will be worked out by applying the theories and software developed in this research to ongoing geotechnical projects.
How to select constitutive models is an important but largely ignored problem in geotechnical practice. The research here will provide a scientific basis for the profession for ranking existing constitutive models for their specific application. The research results will be very helpful for controlling and reducing geotechnical risk from the design stage, particularly for those important projects involving huge investments and serious consequences.