Probabilistic outlier detection for sparse multivariate geotechnical site investigation data using Bayesian learning

Various uncertainties arising during acquisition process of geoscience data may result in anomalous data instances(i.e.,outliers)that do not conform with the expected pattern of regular data instances.With sparse multivariate data obtained from geotechnical site investigation,it is impossible to identify outliers with certainty due to the distortion of statistics of geotechnical parameters caused by outliers and their associated statistical uncertainty resulted from data sparsity.This paper develops a probabilistic outlier detection method for sparse multivariate data obtained from geotechnical site investigation.The proposed approach quantifies the outlying probability of each data instance based on Mahalanobis distance and determines outliers as those data instances with outlying probabilities greater than 0.5.It tackles the distortion issue of statistics estimated from the dataset with outliers by a re-sampling technique and accounts,rationally,for the statistical uncertainty by Bayesian machine learning.Moreover,the proposed approach also suggests an exclusive method to determine outlying components of each outlier.The proposed approach is illustrated and verified using simulated and real-life dataset.It showed that the proposed approach properly identifies outliers among sparse multivariate data and their corresponding outlying components in a probabilistic manner.It can significantly reduce the masking effect(i.e.,missing some actual outliers due to the distortion of statistics by the outliers and statistical uncertainty).It also found that outliers among sparse multivariate data instances affect significantly the construction of multivariate distribution of geotechnical parameters for uncertainty quantification.This emphasizes the necessity of data cleaning process(e.g.,outlier detection)for uncertainty quantification based on geoscience data.

Assessing site investigation program for design of shield tunnels

In the design of shield tunnels,it is expected that a design can be more economical as more site investigation data are available.Nev-ertheless,the cost of site investigation will also increase as more site investigation data are required.It is thus important to assess the potential benefit of a site investigation program before it is conducted considering the uncertainty in the soil properties and the site inves-tigation outcomes.In this paper,a probabilistic framework is suggested to assess the effectiveness of a site investigation program for the design of shield tunnels through the random field theory.An efficient method based on the generalized extreme value distribution is used to calculate the failure probability of the tunnel,through which the expected benefit from a site investigation program can be estimated conveniently.The result shows that a site investigation program is more valuable when a greater target design reliability index is needed.The expected benefit from a site investigation program increases as the borehole intensity increases,and it also increases as the scale of fluctuation of the soil properties increases.The method suggested in this paper provides a useful tool for planning borehole layout for the design of shield tunnels,which is promising for the optimization of site investigation.

Targeted montoring to surgical site infection and investigation of standard application of antibiotics during perioperative period

Objective To investigate current surgical site infection and perioperative antibiotics in inpatients and explore the controlling aim and methods. Methods The infection rates of surgical sites of 287 operated cases from May to Dec 31,2007 were studied and compared with

Effect of trigger system on experimental dispersion characteristics of active surface wave testing

A trigger system is typically employed in active seismic testing to trigger and synchronize multichannel surface wave data acquisition.The effect of the trigger system on the dispersion image of surface waves is empirically known to be negligible,however,theoretical explanation regarding the effect of the trigger system is insufficient.This study systematically examines the theory for surface wave dispersion analysis and proves that the effect of the trigger system on a dispersion image is negligible via a solid theoretical explanation.Subsequently,based on the new theoretical explanation,an alternative method that uses only the relative phase difference between sensors to extract dispersion characteristics with better conceptual clarity is proposed.Two active surface wave testing cases are considered to validate the theory and method.The results indicate that(1)an accurate trigger system is not necessary for surface wave data acquisition,and(2)it is unnecessary to assume that the impact point is the generation point of the surface waves for the experimental dispersion analysis.

Reducing risks in the investigation, design and construction of large concrete dams

An overview of the GeoSafe 2016 Symposium topic is provided using the example of large concrete dams for purposes of illustration.It is essential that the risks associated with large dams be evaluated rigorously and managed proactively at all stages of their lives so that the risk of failure remains As Low As Reasonably Practicable(ALARP).Rock engineering features of large concrete dams that require particular attention,assessment and monitoring during the investigation,design,construction,initial filling,inservice operation,and subsequent repair and upgrade stages of the lives of concrete dams are identified and illustrated by examples from recorded experiences.A number of major concrete dam failures,including that of the St.Francis dam,California,U.S.A.,in 1928,have led to significant developments in rock mechanics and rock engineering knowledge and techniques,as well as in dam design and review processes.More recent advances include a range of analytical,numerical modelling,probabilistic,reliability,failure mode and risk assessment approaches.

Seasonal influence on cone penetration test: An unsaturated soil site example

Interpretation of electric cone penetration test(CPT) based pore water pressure measurement(CPTu) is well established for soils with behavior that follows classical soil mechanics. The literature on the interpretation of these tests performed on unsaturated tropical soils is limited, and little is known about the influence of soil suction on in situ test data. In this context, the CPT data are presented and discussed to illustrate the seasonal variability in an unsaturated tropical soil site. The test data show that soil suction significantly influenced CPT data up to a depth of 4 m at the study site. It shows the importance of considering seasonal variability in unsaturated soil sites caused by soil suction, which was related to water content through a soil-water retention curve(SWRC). It is also important to consider this aspect in the interpretation of CPT data from these soils.

A generic framework for geotechnical subsurface modeling with machine learning

This study introduces a generic framework for geotechnical subsurface modeling, which accounts for spatial autocorrelation with local mapping machine learning(ML) methods. Instead of using XY coordinate fields directly as model input, a series of autocorrelated geotechnical distance fields(GDFs) is designed to enable the ML models to infer the spatial relationship between the sampled locations and unknown locations. The whole framework using GDF with ML methods is named GDF-ML. This framework is purely data-driven which avoids the tedious work in the scale of fluctuations(SOFs)estimating and data detrending in the conventional spatial interpolation methods. Six local mapping ML methods(extra trees(ETs), gradient boosting(GB), extreme gradient boosting(XGBoost), random forest(RF), general regression neural network(GRNN) and k-nearest neighbors(KNN)) are compared in the GDF-ML framework. The results show that the GDFs are better than the conventional XY coordinate fields based ML methods in both accuracy and spatial continuity. GDF-ML is flexible which can be applied to high-dimensional, multi-variable and incomplete datasets. Among these six methods, GDF with ET method(GDF-ET) clearly shows the best accuracy and best spatial continuity. The proposed GDF-ET method can provide a fast and accurate interpretation of the soil property profile. Sensitivity analysis shows that this method is applicable to very small training dataset size. The associated statistical uncertainty can also be quantified so that the reliability of the subsurface modeling results can be estimated objectively and explicitly. The uncertainty results clearly show that the prediction becomes more accurate when more sampled data are available.

Practical configured microtremor array measurements(MAMs)for the geological investigation of underground space

Underground space utilization is becoming increasingly essential for modern metropolitan cities such as Singapore.Mapping a soil/rock interface using traditional borehole investigation methods is expensive and difficult,owing to the numerous physical constraints within a built-up city.Boreholes are often far apart,resulting in many unforeseen ground conditions during subsequent excavation.Geophysical methods are sometimes employed as possible alternatives for fast,economical,and efficient bedrock surveys.The goal of this study is to investigate the practical details of applying microtremor array measurement(MAM)as a non-invasive surface wave survey for mapping soil/rock interfaces in Singapore.Critical configurations in field data acquisition are examined,and practical recommendations for array construction are provided.In addition,30 in situ MAM tests are carried out for two major geological formations in Singapore.From the results,a standard shear wave velocity(V_(s))of 500 m/s is found to be suitable for interpreting the soil/rock interface,for the Bukit Timah Granite and Jurong formations.However,the method does not predict well when soft Kallang formation deposits are present.Other limitations are also discussed in the later parts of this paper.Conclusions and practical recommendations are discussed,providing constructive guidance to the industry.The proposed Vs-based method and associated guidelines and limitations can be used to create a digital geological database and are especially useful for rock profiling in an urban environment.

Uncertainty quantification and reduction in the characterization of subsurface stratigraphy using limited geotechnical investigation data

Subsurface stratigraphy is critical to the design,construction,and subsequent performance of geotechnical structures.However,in practice it is impossible to identify the stratigraphy of a subsurface geological domain with absolute certainty,due to the limitations imposed by geotechnical investigation techniques and project budgets.This paper presents a subsurface stratigraphic modeling and uncertainty quantification approach,which is established based on an improved and extended geological modeling technique previously established by the author and others,for simulating the stratigraphy of both two-dimensional(2D)and three-dimensional(3D)cases with more complex geological features.Furthermore,this approach provides quantitative evaluation of the amount of stratigraphic uncertainty in the current interpretation and enables the systematic reduction of stratigraphic uncertainty through the investigation of additional targeted borehole drilling locations.Illustrative examples,including artificial cases as well as two real cases from existing geotechnical projects,are presented in this study to demonstrate the use of the proposed analysis approach.

50 years change in urban land use and ecological planning globally in the era of design with nature

Ian McHarg’s 1969 book,Design with Nature,was grounded in earlier good practice,especially the European vision of fitting cities into the regional landscape and carrying thorough surveys before planning commenced.The book emerged in the wake of publications that worried about the ability of the earth to support expanding populations and their growing impact on the environment.McHarg argued that ecological design should follow planning based on detailed site information.Although McHarg advised many urban authorities in the USA and other countries,especially in Africa,Asia and Europe,and his ideas had influenced some British new towns,they had little attention in the planning of Milton Keynes,England’s largest new urban settlement,where the initial geological surveys were inadequate,but a successful large tree-planting programhas created a“city in the forest.”McHarg was however closely involved in planning the Nigerian Capital city,Abuja,but his sound master plan was ill-equipped to provide for a nation undergoing post-conflict rebuilding and a local population more concerned with access to land than ecological planning.Nonetheless,McHarg helped us to promote,protect,and preserve urban green spaces.Although teams from different disciplines now work well together,the integrated thinking advocated by McHarg remains sometimes in short supply in local practice.

Methods for evaluating overconsolidation ratio from piezocone sounding results

Methods for estimating the overconsolidation ratio(OCR)of soil deposits from piezocone sounding results have been investigated.Three existing methods have been briefly reviewed and a new method has been proposed.The proposed method can be applied to both normally consolidated/overconsolidated and underconsolidated deposits.Furthermore,existing methods have been modified so that they can be applied to underconsolidated deposits.Then,all the methods have been applied to 12 case histories collected from 6 different countries.The estimated values of OCR have been compared with the measured data.It has been shown that if the parameters(soil properties and empirical parameters)can be determined appropriately,then all the methods can achieve a reasonable prediction.It has also been shown that the proposed method exhibits a relatively better performance and results in less scattered data than the other methods.