Machine Learning–Based Digital Integration of Geotechnical and Ultrahigh–Frequency Geophysical Data for Offshore Site Characterizations

Chen, Jinbo; Vissinga, M.; Shen, Yongxing; Hu, Shuang; Beal, E.; Newlin, J. A.

doi:10.1061/(asce)gt.1943-5606.0002702

Cited by 24 publications

(18 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Furthermore, Shakova et al [65] demonstrated the use of chirp sonar in combination with side-scan sonar imagery and seabed borings to detect and quantify permafrost degradation and gas migration pathways in submerged coastal Arctic environments. Chirp sonar has also been correlated to geotechnical in situ testing in addition to sediment core characterization, and thus, offers a powerful tool to interpolate and extrapolate from geotechnical point measurements in addition to offering deeper penetration depths and mapping of gas, which can have significant impacts on Arctic seabed sediments [66].…”

Section: Geophysical and Remote Sensing Opportunities In Arctic Envir...mentioning

confidence: 99%

“…Combined geotechnical and geophysical data collection and analysis is common for many engineering applications as well as for the investigation of natural processes and natural hazards [66,[92][93][94]. In the past, this has often been limited to spatial alignments and interpolations in which vertical strata and/or spatial variations are derived from the geophysical data, and are related to geotechnical data from in situ testing and core sample testing at specific locations.…”

Section: Geotechnical and Geophysical Soil Characterizationmentioning

confidence: 99%

“…Machine learning approaches, increased accessibility to quality data collection equipment, and thus, data sets, promise continuing advances in this matter [66]. This is of specific relevance for Arctic data collection, as the fusion of geoacoustic and geotechnical data will enable the reduction in costly and difficult physical data collections due to a reliable and accurate correlation to geoacoustic data that is easier and quicker to obtain.…”

Section: Geotechnical and Geophysical Soil Characterizationmentioning

confidence: 99%

See 2 more Smart Citations

Geotechnical Measurements for the Investigation and Assessment of Arctic Coastal Erosion—A Review and Outlook

Stark

Green

Brilli

et al. 2022

JMSE

View full text Add to dashboard Cite

Geotechnical data are increasingly utilized to aid investigations of coastal erosion and the development of coastal morphological models; however, measurement techniques are still challenged by environmental conditions and accessibility in coastal areas, and particularly, by nearshore conditions. These challenges are exacerbated for Arctic coastal environments. This article reviews existing and emerging data collection methods in the context of geotechnical investigations of Arctic coastal erosion and nearshore change. Specifically, the use of cone penetration testing (CPT), which can provide key data for the mapping of soil and ice layers as well as for the assessment of slope and block failures, and the use of free-fall penetrometers (FFPs) for rapid mapping of seabed surface conditions, are discussed. Because of limitations in the spatial coverage and number of available in situ point measurements by penetrometers, data fusion with geophysical and remotely sensed data is considered. Offshore and nearshore, the combination of acoustic surveying with geotechnical testing can optimize large-scale seabed characterization, while onshore most recent developments in satellite-based and unmanned-aerial-vehicle-based data collection offer new opportunities to enhance spatial coverage and collect information on bathymetry and topography, amongst others. Emphasis is given to easily deployable and rugged techniques and strategies that can offer near-term opportunities to fill current gaps in data availability. This review suggests that data fusion of geotechnical in situ testing, using CPT to provide soil information at deeper depths and even in the presence of ice and using FFPs to offer rapid and large-coverage geotechnical testing of surface sediments (i.e., in the upper tens of centimeters to meters of sediment depth), combined with acoustic seabed surveying and emerging remote sensing tools, has the potential to provide essential data to improve the prediction of Arctic coastal erosion, particularly where climate-driven changes in soil conditions may bias the use of historic observations of erosion for future prediction.

show abstract

Section: Geophysical and Remote Sensing Opportunities In Arctic Envir...mentioning

confidence: 99%

Section: Geotechnical and Geophysical Soil Characterizationmentioning

confidence: 99%

Section: Geotechnical and Geophysical Soil Characterizationmentioning

confidence: 99%

See 1 more Smart Citation

Geotechnical Measurements for the Investigation and Assessment of Arctic Coastal Erosion—A Review and Outlook

Stark

Green

Brilli

et al. 2022

JMSE

View full text Add to dashboard Cite

show abstract

“…The CPT test provides continuous and reliable soil data, making it an efficient and cost-effective method in geotechnical engineering practice. This wealth of CPT data has attracted the attention of many geotechnical researchers to further improve the prediction accuracy of V s employing machine learning (ML) algorithms [22][23][24][25][26]. ML algorithms have shown great promise in accurately predicting V s from CPT data.…”

Section: Introductionmentioning

confidence: 99%

Machine Learning Techniques for Soil Characterization Using Cone Penetration Test Data

Chala

Ray

2023

Applied Sciences

View full text Add to dashboard Cite

Seismic response assessment requires reliable information about subsurface conditions, including soil shear wave velocity (Vs). To properly assess seismic response, engineers need accurate information about Vs, an essential parameter for evaluating the propagation of seismic waves. However, measuring Vs is generally challenging due to the complex and time-consuming nature of field and laboratory tests. This study aims to predict Vs using machine learning (ML) algorithms from cone penetration test (CPT) data. The study utilized four ML algorithms, namely Random Forests (RFs), Support Vector Machine (SVM), Decision Trees (DT), and eXtreme Gradient Boosting (XGBoost), to predict Vs. These ML models were trained on 70% of the datasets, while their efficiency and generalization ability were assessed on the remaining 30%. The hyperparameters for each ML model were fine-tuned through Bayesian optimization with k-fold cross-validation techniques. The performance of each ML model was evaluated using eight different metrics, including root mean squared error (RMSE), mean absolute error (MAE), mean absolute percentage error (MAPE), coefficient of determination (R2), performance index (PI), scatter index (SI), A10−I, and U95. The results demonstrated that the RF model consistently performed well across all metrics. It achieved high accuracy and the lowest level of errors, indicating superior accuracy and precision in predicting Vs. The SVM and XGBoost models also exhibited strong performance, with slightly higher error metrics compared with the RF model. However, the DT model performed poorly, with higher error rates and uncertainty in predicting Vs. Based on these results, we can conclude that the RF model is highly effective at accurately predicting Vs using CPT data with minimal input features.

show abstract

“…While the development of empirical equations using traditional regression approaches to predict the mechanical properties of soils has facilitated geotechnical analyses to a large extent, it remains challenging to establish accurate correlations, owing to the major uncertainty in and great complexity of soil properties (Ching and Phoon, 2014). Over the past decades, the applicability of machine learning (ML) approaches, such as artificial neural networks (ANNs), random forest (RF) methods, and support vector machines (SVMs), among others, has been well-proven in terms of their ability to efficiently and accurately map highly non-linear problems in a wide variety of areas of engineering (Arditi and Pulket, 2010;Chen et al, 2021), including geotechnical engineering. Successful examples of applications include analyses of slope stability (Kardani et al, 2021;Meng et al, 2021) and deformation (Zhang et al, 2019;Zhang et al, 2020a;; pile designs (Makasis et al, 2018;Zhang et al, 2020e); prediction of the bearing capacity of strip footings (Acharyya, 2019;Sadegh et al, 2021); lateral wall deformation and basal heave stability for braced excavations (Goh et al, 1995;Zhang et al, 2020); soil constitutive relations (Najjar and Huang, 2007); liquefaction resistance of sands (Kim and Kim, 2006); lining response for tunnels (Zhang et al, 2020g); calibration of resistance factors for reliability-based load and resistance factor design (Hu and Lin, 2019); prediction of soil transparency (Wang et al, 2021); analysis of ground settlement induced by shield tunneling (Zhang et al, 2020c); reliability analysis by SVM (Pan and Dias, 2017); and mapping of groundwater potential using SVM, RF, and GA models (Naghibi et al, 2017), among others.…”

Section: Introductionmentioning

confidence: 99%

Machine learning approaches to estimation of the compressibility of soft soils

Liu

Lin

Wang³

2023

Front. Earth Sci.

View full text Add to dashboard Cite

The modulus of compression and coefficient of compressibility of soft soils are key parameters for assessing deformation of geotechnical infrastructure. However, the consolidation tests used to determine these two indices are time-consuming and the results are easily and heavily influenced by workmanship, testing apparatus, and other factors. Therefore, it is of great interest to develop a simple approach to accurately estimate these compressibility indices. This article presents the development of three machine learning (ML) models—at artificial neural network (ANN), a random forest model, and a support vector machine model—for mapping of the two compressibility indices for soft soils. A database containing 743 sets of measured physical and compression parameters of soft soils was adopted to train and validate the models. To quantify model uncertainty, the accuracies of the ML models were statistically evaluated using a bias factor defined as the ratio of the measured to the predicted compression indices. The results showed that all three ML models were accurate on average, with low dispersion in prediction accuracy. The ANN was found to be the best model, as it provides a simple analytical form and has no hidden dependency between the bias and predicted indices. Finally, the probability distribution functions of the bias factors were also determined using the fit-to-tail technique. The results of this study will be helpful in saving cost and time in geotechnical investigation of soft soils.

show abstract

Machine Learning–Based Digital Integration of Geotechnical and Ultrahigh–Frequency Geophysical Data for Offshore Site Characterizations

Cited by 24 publications

References 24 publications

Geotechnical Measurements for the Investigation and Assessment of Arctic Coastal Erosion—A Review and Outlook

Geotechnical Measurements for the Investigation and Assessment of Arctic Coastal Erosion—A Review and Outlook

Machine Learning Techniques for Soil Characterization Using Cone Penetration Test Data

Machine learning approaches to estimation of the compressibility of soft soils

Contact Info

Product

Resources

About