Abstract

The determination of reliable shear-wave velocity models using Multichannel Analysis of Surface Waves (MASW) has increased importantly for site characterisation studies due to their use in geotechnical studies and regulations. The standard MASW approach is commonly based on the analysis of vertical component of Rayleigh waves, which can result in inaccurate and potentially erroneous interpretations by personal bias. Thus, we present the joint analysis of different and independent multi-component data based on Rayleigh and Love waves to obtain 2D Vs sections for site characterization. Those seismic data were recorded using a landstreamer consisting of 8 triaxial 4.5Hz geophones. To generate Rayleigh waves, the blows were given vertically on a plate, and for the Love waves the blows were given laterally on a horizontal wooden beam. A joint analysis of Rayleigh and Love waves data was conducted on seismic data recorded from the metropolitan area of Granada city (Spain) to generate their dispersion curves. This new approach enabled a proper identification of fundamental- and higher-mode surface waves facilitating the reliable reconstruction of subsurface Vs profiles through a robust joint inversion process. The MASW 1D Vs versus depth models were corroborated at several test sites by the information obtained from boreholes. Thus, the main geological formations could be inferred from MASW 2D Vs sections down to a depth of 30 meters, as well as the Vs30 parameter to perform a reliable seismic microzonation of the study area. This methodology provides a very well constrained inversion procedure capable of providing a robust subsurface Vs model for site characterization.

Abstract

Secondary compression can be an important source of settlement in artificial fills, even when these fills are well constructed. In some cases, especially when the fill thickness is greater than about 15 m, the resulting long-term settlements can adversely impact the performance of structures and infrastructure, and thus may necessitate special preventive design provisions. Yet, this source of settlement is often mistakenly overlooked. Secondary compression can be even more problematic when the fill is poorly constructed. Backfills of former open-pit mines are examples of practical projects where assessments of long-term secondary compression settlement are necessary, especially when these backfills are deep and/or not properly engineered. Laboratory assessments of secondary compression in these materials are inherently problematic and become impossible when the fill contains large particles or has other complicating characteristics. However, this problem is an excellent opportunity to apply the observational method where the coefficient of secondary compression, Cαε, is assessed in-situ using settlement monument data. This Cαε value is then used to forecast future settlements, which typically continue for decades, and thus provides essential information for the site-specific design of structures and infrastructure. However, the experimental and analytical processes for conducting these evaluations are more difficult than might be expected, and missteps can lead to significant errors in the computed future settlements. Some of these difficulties are due to limitations in our knowledge of the underlying physical processes and in the analytical models used to describe them. Methods of collecting the required field data and conducting these settlement evaluations are discussed based on experience with deep fills in California as well as published data from elsewhere.

Abstract

Capturing the spatial variability in soil is crucial for ground response analyses in the context of seismic hazard mitigation. The lateral variability in thickness and properties of the different soil layers is one of the main factors that determines the variability of the ground motion spectrum from one location to another. The absence of such lateral variability information in the subsoil in between the locations of Cone Penetration Tests (CPTs) may be compensated by the use of more densely sampled seismic data. In this research we aim to derive a shear-wave velocity field through seismic full-waveform inversion that yields a model resolution approaching that of high-resolution seismic CPT surveys. Following this, a datadriven correlation between geophysical and geotechnical information is attempted through the application of new machine-learning-based approaches.

Abstract

Knowledge of seabed properties away from investigated locations is often required, for example, when geotechnical surveys are sparse or when the field layout changes between investigation and construction phases. In such cases, design lines that appropriately incorporate the uncertainty of the seabed properties must be defined to ensure reliable (yet not overly costly) design. This paper explores how two different approaches, traditional engineering judgement and advanced statistical methods, fare at quantifying the uncertainty of a real offshore site. This is achieved by ‘hiding’ different data and ‘scoring’ the predictive performance of the methods against a range of criteria. The work reveals that complex geological sites are significantly more challenging to represent than the stationary random fields often examined in research and suggests that more advanced approaches incorporating broader data sets are required to reduce uncertainty.

Abstract

Non-linear soil behaviour adds complexity in accurate parameter selection for numerical modelling. One of these parameters is the small-strain shear stiffness. This parameter depends strongly on the soil mass density and the shear wave velocity; the latter can be determined through in-situ tests or laboratory tests. The paper focuses on training various machine learning models to predict shear wave velocity estimates based on raw data from cone penetration test soundings. Three decision tree algorithms are considered for the analysis: XGBRegressor, HistGradientRegressor, and RandomForest. Various data preprocessing approaches are investigated, including noise removal and outlier identification, to assess their impact on the model performance. The results indicate that different data preprocessing approaches yield significant differences in the model performances. When applied to unseen raw data from a sand site of the Norwegian GeoTest Site, the model demonstrates promising predictive capabilities and is in a good agreement with well-known correlations. This study underlines the importance of data quality and preprocessing for reliable machine learning models. To enhance transparency and reproducibility, a GitHub repository with all the used files is made available online.

Abstract

Soil boundary delineation is an important task in geotechnical site characterization. It can be achieved by either extracting borehole samples, conducting laboratory tests, and classifying them according to a soil classification system such as the Unified Soil Classification System (USCS) or utilizing multiple cone penetration test (CPT) soundings, and identifying soil boundaries at the soundings from the Ic (soil behavior type index) profiles. However, most soil-layer delineation methods can only take a single type of test result as the input. For instance, the well-known Markov random field (MRF) method can only take soil-type data such as sand, silt, or clay at boreholes as the input. Recognizing that soil classifications and soil properties are correlated, this paper proposes a novel coupled MRF-Bayesian framework to infer the spatial variation of USCS classifications (e.g., sand, silt, and clay) as well as soil properties by integrating both CPT and borehole data. This integrated approach leverages both CPT and borehole data to address some main challenges e.g., uncertainties and multivariate soil data input in underground stratification problems by simultaneous sampling of soil properties and soil types. The new unified framework can accommodate multivariate data, hence the new framework is compatible with the geotechnical engineering practice. The uncertainties for the spatial variation of USCS classification at sounding locations are quantified through a “layer-specific” Bayesian updating i.e., updating posterior cross-correlation behaviors for different layers (such as sand, silt, and clay), independently. In this Bayesian updating, soil-type data can provide some information about the soil properties according to the unified soil classification system. Further, the soil boundaries can be identified across the entire domain by the realization of conditional random fields of soil properties once the spatial variation of USCS classification is inferred at sounding locations, followed by a 3-dimensional Markov random field process.

Abstract

In cone penetration testing (CPT) an electronic penetrometer is pushed at a constant rate into penetrable soils and cone bearing (qc), sleeve friction (fc) and dynamic pore pressure (u) are recorded with depth. The measured qc, fs and u values are utilized to estimate soil type and associated properties. Cone tips have areas which vary from 5cm2 to 40 cm2. The larger tips allow for the penetration of gravely soils while small cone tips are utilized for shallow soil investigations. The measured cone bearing and sleeve friction values are blurred or averaged. The measurements are also susceptible to anomalous peaks and troughs due to the relatively small diameter cone tip penetrating sandy, silty and gravelly soils. The cones with relatively smaller cone tips are significantly more susceptible to the anomalous peaks and troughs while the cones with larger cone tips are more susceptible to the smoothing of the cone tip and sleeve friction measurements. Baziw Consulting Engineers (BCE) has invested considerable resources in addressing the qc and fs measurements distortions. This paper outlines the techniques developed by BCE and integrates them so that optimal soil properties can be obtained from CPT data sets. Particular focus is put on relatively larger cone tips because they can penetrate soils with high resistance and are less susceptible to the additive measurement noise of anomalous peaks and troughs. The anomalous peaks and troughs are more challenging to remove or minimize than the qc and fs blurring effects. It is of paramount importance to first implement newly developed signal processing and optimal estimation algorithms on extensive test bed simulations prior to processing real data sets. This paper also outlines the results from processing a challenging test bed simulation of a 40 cm2 cone tip data set with BCE’s newly developed algorithms.

Abstract

We present a novel method using four artificial intelligence (AI) algorithms to anticipate the cumulative degree of soil compaction (CDSC) after dynamic compaction (DC). Four AI algorithms adopted in this study include support vector regression SVR, artificial neural network (ANN), random forest (RF), and gradient boosting machine (GBM). Input variables for AI algorithms involve the average SPT N-value before dynamic compaction, cumulative applied energy normalized with a cross-sectional area of tamper, and the number of the tamper drops. Apart from cross-validation with a testing set, additional in situ test data compiled from a different section within the studied site are used to estimate the generalized capacity of the AI models. In addition, we conduct out-of-distribution analyses for the four AI algorithms in view of parametric studies. The CDSC prediction performance for the four AI models results in high prediction metrics of accuracy with the r2 higher than 0.9 for the testing scenario while the r2 of the other AI models is more than 0.9 when out-of-sample data are considered except for the GBM. The ANN seems to be the best model as the parametric study considers out-of-distribution data and suggests a strong relationship between input variables and CDSC that is more coherent with engineering principles for DC. Finally, the ANN model can be utilized to develop a mathematical model for CDSC prediction.

Abstract

Soil boundary delineation is an important task in geotechnical site characterization. It can be achieved by either extracting borehole samples, conducting laboratory tests, and classifying them according to a soil classification system such as the Unified Soil Classification System (USCS) or utilizing multiple cone penetration test (CPT) soundings, and identifying soil boundaries at the soundings from the Ic (soil behavior type index) profiles. However, most soil-layer delineation methods can only take a single type of test result as the input. For instance, the well-known Markov random field (MRF) method can only take soil-type data such as sand, silt, or clay at boreholes as the input. Recognizing that soil classifications and soil properties are correlated, this paper proposes a novel coupled MRF-Bayesian framework to infer the spatial variation of USCS classifications (e.g., sand, silt, and clay) as well as soil properties by integrating both CPT and borehole data. This integrated approach leverages both CPT and borehole data to address some main challenges e.g., uncertainties and multivariate soil data input in underground stratification problems by simultaneous sampling of soil properties and soil types. The new unified framework can accommodate multivariate data, hence the new framework is compatible with the geotechnical engineering practice. The uncertainties for the spatial variation of USCS classification at sounding locations are quantified through a “layer-specific” Bayesian updating i.e., updating posterior cross-correlation behaviors for different layers (such as sand, silt, and clay), independently. In this Bayesian updating, soil-type data can provide some information about the soil properties according to the unified soil classification system. Further, the soil boundaries can be identified across the entire domain by the realization of conditional random fields of soil properties once the spatial variation of USCS classification is inferred at sounding locations, followed by a 3-dimensional Markov random field process.

Abstract

In cone penetration testing (CPT) an electronic penetrometer is pushed at a constant rate into penetrable soils and cone bearing (qc), sleeve friction (fc) and dynamic pore pressure (u) are recorded with depth. The measured qc, fs and u values are utilized to estimate soil type and associated properties. Cone tips have areas which vary from 5cm2 to 40 cm2. The larger tips allow for the penetration of gravely soils while small cone tips are utilized for shallow soil investigations. The measured cone bearing and sleeve friction values are blurred or averaged. The measurements are also susceptible to anomalous peaks and troughs due to the relatively small diameter cone tip penetrating sandy, silty and gravelly soils. The cones with relatively smaller cone tips are significantly more susceptible to the anomalous peaks and troughs while the cones with larger cone tips are more susceptible to the smoothing of the cone tip and sleeve friction measurements. Baziw Consulting Engineers (BCE) has invested considerable resources in addressing the qc and fs measurements distortions. This paper outlines the techniques developed by BCE and integrates them so that optimal soil properties can be obtained from CPT data sets. Particular focus is put on relatively larger cone tips because they can penetrate soils with high resistance and are less susceptible to the additive measurement noise of anomalous peaks and troughs. The anomalous peaks and troughs are more challenging to remove or minimize than the qc and fs blurring effects. It is of paramount importance to first implement newly developed signal processing and optimal estimation algorithms on extensive test bed simulations prior to processing real data sets. This paper also outlines the results from processing a challenging test bed simulation of a 40 cm2 cone tip data set with BCE’s newly developed algorithms.