More Like this

1. Earthquake-Induced Liquefaction Manifestation Multiclass Prediction Utilizing Random Forests for the Canterbury Earthquake Sequence

   https://doi.org/10.1061/9780784485316.024  

   Cheng, Katherine; Busch, Pablo; Ziotopoulou, Katerina (February 2024, American Society of Civil Engineers)

   The abundance of post-earthquake data from the Canterbury, New Zealand (NZ), area can be leveraged for exploring machine learning (ML) opportunities for geotechnical earthquake engineering. Herein, random forest (RF) is chosen as the ML model to be utilized as it is a powerful non-parametric classification model that can also calculate global feature importance post-model building. The results and procedure are presented of building a multiclass liquefaction manifestation classification RF model with features engineered to preserve special relationships. The RF model hyperparameters are optimized with a two-step fivefold crossvalidation grid search to avoid overfitting. The overall model accuracy is 96% over six ordinal categories predicting over the Canterbury earthquake sequence measurements from 2010, 2011, and 2016. The resultant RF model can serve as a blueprint for incorporation of other sources of physical data such as geological maps to widen the bounds of model usability. 

   more » « less
2. EXPLAINABLE MACHINE LEARNING FOR SOILCRETE UCS PREDICTIONS

   Cheng, Katherine; Gingery, J R; Yossifov, M D; Ziotopoulou, K (November 2024, Deep Foundations Institute)

   Soil mixing is a ground improvement method that consists of mixing cementitious binders with soil in-situ to create soilcrete. A key parameter in the design and construction of this method is the Unconfined Compressive Strength (UCS) of the soilcrete after a given curing time. This paper explores the intersection of Machine Learning (ML) with geotechnical engineering and soilcrete applications. A database of soilcrete UCS and site/soil/means/methods metadata is compiled from recent projects in the western United States and leveraged to explore UCS prediction with the eXtreme Gradient Boosting (XGBoost) ML algorithm which resulted in a ML model with a R2 value of 88%. To achieve insights from the ML model, the Explainable ML model SHapley Additive exPlanations (SHAP) was then applied to the XGBoost model to explain variable importances and influences for the final UCS prediction value. From this ML application, a blueprint of how to scaffold, feature engineer, and prepare soilcrete data for ML is showcased. Furthermore, the insights obtained from the SHAP model can be further pursued in traditional geotechnical research approaches to expand soil mixing knowledge. 

   more » « less
3. Earthquake Geotechnical Engineering Reconnaissance Methods and Advances

   Bray, J.D. (July 2019, 7th International Conference on Earthquake Geotechnical Engineering)

   Earthquake geotechnical engineering is an experience-driven discipline. Field observations are particularly important, because it is difficult to replicate in the laboratory the characteristics and response of in situ soil deposits. Much of the data generated by a major earthquake is perishable, so it is critical that it is collected soon after an event occurs. Detailed mapping and surveying of damaged and undamaged areas provides the data for the well-documented case histories that drive the development of many of the earthquake geotechnical engineering design procedures. New technologies are being employed to capture earthquake-induced ground deformation, including Light Detection and Ranging, Structure-from-Motion, and Unmanned Aerial Vehicles. Post-earthquake reconnaissance has moved beyond taking photographs and field notes to taking advantage of technologies that can capture ground and structure deformations more completely and accurately. Moreover, electronic data enables effective sharing and archiving of the measurements. Unanticipated observations from major events often catalyze new research directions. Important advancements are possible through post-event research if their effects are captured and shared effectively. An overview of some of recent integrated technology deployments and their role in advancing knowledge are presented. 

   more » « less
4. On Validation of a Two-Surface Plasticity Model for Soil Liquefaction Analysis

   Manzari, Majid T.; Elghoraiby, Mohamed A. (January 2021, International Conference of the International Association for Computer Methods and Advances in Geomechanics IACMAG 2021: Challenges and Innovations in Geomechanics)

   null; null; null (Ed.)

   Constitutive modeling of granular materials such as sands, non-plastic silts, and gravels has been significantly advanced in the past three decades. Several new constitutive models have been proposed and calibrated to simulate the results of various laboratory element tests. Due to this progress and owing to the surge of interest in geotechnical engineering community to use well-documented constitutive models in major geotechnical projects, a more thorough evaluation of these models is necessary. Performance of the current models should be particularly evaluated in the simulation of boundary value problems where stress/strain paths are much more complex than the element tests performed in laboratory. Such validation efforts will be an important step towards the use of these models in practice. This paper presents the results of an extensive validation study aimed at assessing the capabilities and limitations of a two-surface plasticity model for sands in two selected boundary value problems, i.e. lateral spreading of mildly sloping liquefiable grounds. The results of a large number of centrifuge tests conducted during the course of four consecutive international projects known as Liquefaction Experiments and Analysis Project (LEAP) are used in this validation study. The capabilities and limitations of the two-surface plasticity model, initially calibrated against element tests, will be carefully assessed by comparing the numerical simulations with the results of the centrifuge tests from recent LEAP projects. 

   more » « less
5. Earthquake Geotechnical Engineering Reconnaissance Methods and Advances

   Bray, J.D. (July 2019, 7th International Conference on Earthquake Geotechnical Engineering)

   Earthquake geotechnical engineering is an experience-driven discipline. Field observations are particularly important, because it is difficult to replicate in the laboratory the character-istics and response of in situ soil deposits. Much of the data generated by a major earth-quake is perishable, so it is critical that it is collected soon after an event occurs. Detailed mapping and surveying of damaged and undamaged areas provides the data for the well-documented case histories that drive the development of many of the earthquake geotech-nical engineering design procedures. New technologies are being employed to capture earthquake-induced ground deformation, including Light Detection and Ranging, Struc-ture-from-Motion, and Unmanned Aerial Vehicles. Post-earthquake reconnaissance has moved beyond taking photographs and field notes to taking advantage of technologies that can capture ground and structure deformations more completely and accurately. Moreo-ver, electronic data enables effective sharing and archiving of the measurements. Unantic-ipated observations from major events often catalyze new research directions. Important advancements are possible through post-event research if their effects are captured and shared effectively. An overview of some of recent integrated technology deployments and their role in advancing knowledge are presented. 

   more » « less