Study On Stratification And Spatial Variability Of Soils Based On Limited Boreholes

Soils are sediments which are formed under complex geological processes and loading histories,which makes the physical and mechanical properties of soils at different locations vary in different degrees.Soils are irregular layerd distributed in space and geotechnical parameters within one soil layer possess the property of spatial variability.Therefore,site-investigation is necessary before the design of geotechnical engineering structure.However,soil properties at unsampled locations cannot be determined with certainty due to the limited site-investigation.In order to provide sufficient and reliable basis for engineering design,the study of the stratification and spatial variability of soils are quite necessary.In the studies of the stratification of soils,the spatial interpolation method cannot be used in the simulation of complex strata.Although the variogram simulation method is capable to characterize complex strata,the continuity of the simulated strata is poor.Alternatively,the coupled Markov chain model is a viable way to overcome the problems mentioned above.However,this model cannot simulate the interfaces of strata with different dip directions.It also cannot accurately simulate the prism in strata.This study improved the coupled Markov chain model.The program of improved coupled Markov chain model was programed,and the correctness of the program is verified.The improved coupled Markov chain model can well simulate the interfaces of strata with different dip directions and the soil material variation within one stratum(e.g.prism in stratum).Based on the coupled Markov chain model,the influence of borehole layout schemes on the simulation results of strata is studied.The locations of the boreholes may have a great influence on the simulation results of the stratification of soils.The probability of soil material variation within one stratum will increase with the increasing of borehole spacing.This paper also analyzes the number of boreholes required to obtain the stability simulation results of several common stratification of soils.Spatial variability within one soil layer can be studied by random field theory.However,a random field model will overestimate the spatial variability due to the data at the measured locations are still randomly generated.Therefore,in this paper,the three-dimensional conditioned random field,which can make full use of the measured data in the field and effectively reduce the spatial variability of geotechnical parameters around the measured locations,was studied.The program of conditioned three-dimensional random field was programed,and the correctness of the program is verified.Based on the three-dimensional conditioned random field,the influence of the locations of boreholes and the number of boreholes on the spatial variability of geotechnical parameters was also studied.When the correlation structure of geotechnical parameters can be expressed by the squared exponential correlation function,the main constraint distance of site-investigation data in conditioned random field is approximately half of the scale of fluctuation.Moreover,the prediction precision of conditioned random field is affected by the number of boreholes around the prediction locations and the coefficient of variation of geotechnical parameters.