| China is currently in a period of rapid urbanization,and a large number of constructions on municipal,transportation and hydropower projects are being carried out.Site characterization,which aims to characterize the subsurface stratigraphic configuration and the associated geotechnical parameters(or geo-properties for short),is a prerequisite for the construction of a typical project.However,owing to the deposit histories,tectonic activities and human activities,both the stratum type and the geo-properties could be spatially varied.In geotechnical practice,only a few boreholes are drilled at a given site,due to the project budget and schedule constraints.As an outcome,the stratigraphic configuration and geo-properties may only be revealed at the sparse borehole locations.At other places,such subsurface information has to be inferred from that at borehole locations,which often results in a significant error(i.e.,uncertainty)due to the complexity and spatial variability of the natural deposits.The uncertainty involved in the inferred geological model may be divided into the stratigraphic uncertainty and the geo-properties uncertainty.The former refers mainly to the uncertainty of boundaries between different strata,while the latter refers primarily to the inherent variability of the geo-properties within each stratum.The geological model is the fundamental input parameter to the analysis of the performance of a geotechnical system,the geological model uncertainty would in turn lead to the uncertainty in the predicted performance of the geotechnical system.However,the existing probabilistic approaches for characterizing the geological model uncertainty are not perfect.The limitations of these probabilistic approaches could be summarized as follows:(1)the abilities of the existing stratigraphic uncertainty modeling approaches in capturing the complex geological setting(e.g.,the stratigraphic dip is spatially varied)are degraded.(2)past studies treated the stratigraphic uncertainty and the geo-properties uncertainty separately,and the coupled characterization of the stratigraphic and geo-properties uncertainties is ignored.(3)the coupled influence of the stratigraphic and geo-properties uncertainties on the efficiency of site exploration program is neglected.To address the above three problems,a random field-based approach for modeling the stratigraphic uncertainty is first presented in this thesis.Second,on the basis of the above approach,a conditional random field-based stratigraphic uncertainty modeling approach is established.Third,a conditional random field approach for the coupled characterization of the stratigraphic and geo-properties uncertainties is proposed,in which the stratigraphic configuration could be sampled using the conditional random field-based stratigraphic uncertainty modeling approach;then,the spatial correlation of the geo-properties is updated based upon the sampled stratigraphic configuration.With the generated geological models as inputs,the influence of the stratigraphic and geo-properties uncertainties on the performance of the concerned geotechnical system can be evaluated explicitly.Four,a framework for the optimization of site exploration program based on the coupled characterization of the stratigraphic and geo-properties uncertainties is presented.The implementation details and some conclusions of this study could be summarized as follows:(1)A random field-based approach for characterizing the stratigraphic uncertainty is presentedThis thesis presents a random field-based approach to overcome the limitations of the existing approaches for the characterization of the stratigraphic uncertainty.Within the framework of the proposed method,the spatial correlation of the stratum existence between two subsurface elements is characterized by an autocorrelation function,and the probability of the existence of a particular stratum in a given non-borehole element is determined according to the derived spatial correlations.With the existence probabilities calculated,Monte Carlo simulation is used for sampling the possible realizations of the stratigraphic configuration.To illustrate the effectiveness of this new approach,a set of hypothetical examples(with different stratigraphic settings)are studied;and,the advantages of this new approach over the existing stratigraphic uncertainty modelling approaches are depicted through a comparative analysis.(2)A conditional random field approach for the characterization of the stratigraphic uncertainty is presentedTo overcome the limitations of the random field-based approach,a conditional random field approach is presented for simulating the subsurface stratigraphic configuration and its uncertainty.The proposed method has three features that are significant improvements over the random field-based approach.These features include:1)the spatial correlation structure of the strata is characterized with the maximum likelihood method,2)the initial stratigraphic configurations are sampled with the conditional random field theory,and 3)the local anomalies in the initial stratigraphic configurations are removed through MCMC-based updating.With the proposed method,multiple realizations of the stratigraphic configuration can be systematically sampled from the borehole data,and the stratigraphic uncertainty can be quantified.The effectiveness of the proposed method and its advantages over the existing stratigraphic characterization methods are demonstrated through a series of comparative analyses.The versatility of the new approach in modeling the 3-D stratigraphic configuration is further revealed through a case study of a site in Western Australia.(3)A conditional random field approach for the coupled characterization of the stratigraphic and geo-properties uncertainties is presentedOn the basis of the conditional random field approach for the characterization of the stratigraphic uncertainty,a conditional random field approach for the coupled characterization of the stratigraphic and geo-properties uncertainties is presented in this thesis.The spatial correlation of the stratum existence between different subsurface elements and the spatial correlation of geo-properties are characterized by two autocorrelation functions,determined with the maximum likelihood principle.With the knowledge of the spatial correlation of the stratum existence,the stratigraphic configuration can be sampled using the conditional random field-based stratigraphic uncertainty modeling approach.Then,the spatial correlation of the geo-properties is updated based on the sampled stratigraphic configuration.With the updated spatial correlation of the geo-properties,the spatial distribution of the geo-properties can readily be simulated with the conditional random field theory.The effectiveness of the proposed approach is demonstrated through a case study of probabilistic site characterization of an offshore wind farm site in Taiwan.The versatility of the proposed approach in characterizing the 3-D geological model is further revealed through a case study of a site in Western Australia.With the generated geological models as inputs,the influence of the stratigraphic and geo-properties uncertainties on the performance of the geotechnical system of concern can be evaluated explicitly.(4)A framework for the site investigation optimization based on the coupled characterization of stratigraphic and geo-properties uncertainties is presentedThe thesis proposes a framework for optimizing the site exploration program based on coupled characterization of the stratigraphic and geo-properties uncertainties,to overcome the limitations of the existing site investigation optimization approaches.Since the site exploration optimization aims to cause a most significant reduction of the stratigraphic and geo-properties uncertainties,the additional borehole is drilled in the place,at the concerned site,with the maximum uncertainty.In the proposed framework,the spatial correlation of the existence of the stratum(between different subsurface elements),and that of the geo-properties are estimated with a Bayesian updating approach.Then,the conditional random field approach,for the coupled characterization of the stratigraphic and geo-properties uncertainties,is adopted to simulate the stratigraphic uncertainty and the geo-properties uncertainty.With the characterized stratigraphic and geo-properties uncertainties,the optimal location and depth of the additional borehole can be determined.The effectiveness and versatility of the proposed framework,and its advantages over the traditional framework,are demonstrated through two case studies,including a site with a shallow foundation site and a slope site. |
