Improved prediction of slope stability using a hybrid stacking ensemble method based on finite element analysis and field data

Journal of Rock Mechanics and Geotechnical Engineering, Год журнала: 2020, Номер 13(1), С. 188 - 201

Опубликована: Ноя. 23, 2020

Slope failures lead to catastrophic consequences in numerous countries and thus the stability assessment for slopes is of high interest geotechnical geological engineering researches. A hybrid stacking ensemble approach proposed this study enhancing prediction slope stability. In approach, we used an artificial bee colony (ABC) algorithm find out best combination base classifiers (level 0) determined a suitable meta-classifier 1) from pool 11 individual optimized machine learning (OML) algorithms. Finite element analysis (FEA) was conducted order form synthetic database training stage (150 cases) model while 107 real field cases were testing stage. The results by then compared with that obtained OML methods using confusion matrix, F1-score, area under curve, i.e. AUC-score. comparisons showed significant improvement ability has been achieved (AUC = 90.4%), which 7% higher than 82.9%). Then, further comparison undertaken between method basic classifier on prediction. prominent performance over method. Finally, importance variables studied linear vector quantization (LVQ)

Язык: Английский

---

Ensemble learning framework for landslide susceptibility mapping: Different basic classifier and ensemble strategy

Geoscience Frontiers, Год журнала: 2023, Номер 14(6), С. 101645 - 101645

Опубликована: Июнь 7, 2023

The application of ensemble learning models has been continuously improved in recent landslide susceptibility research, but most studies have no unified framework. Moreover, few papers discussed the applicability model mapping at township level. This study aims defining a robust framework that can become benchmark method for future research dealing with comparison different models. For this purpose, present work focuses on three basic classifiers: decision tree (DT), support vector machine (SVM), and multi-layer perceptron neural network (MLPNN) two homogeneous such as random forest (RF) extreme gradient boosting (XGBoost). hierarchical construction deep relied leading technologies (i.e., homogeneous/heterogeneous bagging, boosting, stacking strategy) to provide more accurate effective spatial probability occurrence. selected area is Dazhou town, located Jurassic red-strata Three Gorges Reservoir Area China, which strategic economic currently characterized by widespread risk. Based long-term field investigation, inventory counting thirty-three slow-moving polygons was drawn. results show do not necessarily perform better; instance, Bagging based DT-SVM-MLPNN-XGBoost performed worse than single XGBoost model. Amongst eleven tested models, Stacking RF-XGBoost model, ensemble, showed highest capability predicting landslide-affected areas. Besides, factor behaviors DT, SVM, MLPNN, RF reflected characteristics landslides reservoir area, wherein unfavorable lithological conditions intense human engineering activities water level fluctuation, residential construction, farmland development) are proven be key triggers. presented approach could used occurrence prediction similar regions other fields.

Язык: Английский

---

Prevention/mitigation of natural disasters in urban areas

Smart Construction and Sustainable Cities, Год журнала: 2023, Номер 1(1)

Опубликована: Авг. 9, 2023

Abstract Preventing/mitigating natural disasters in urban areas can indirectly be part of the 17 sustainable economic and social development intentions according to United Nations 2015. Four types disasters—flooding, heavy rain-induced slope failures/landslides; earthquakes causing structure failure/collapse, land subsidence—are briefly considered this article. With increased frequency climate change-induced extreme weathers, numbers flooding failures/landslides has recent years. There are both engineering methods prevent their occurrence, more effectively early prediction warning systems mitigate resulting damage. However, still cannot predicted an extent that is sufficient avoid damage, developing adopting structures resilient against earthquakes, is, featuring earthquake resistance, vibration damping, seismic isolation, essential tasks for city development. Land subsidence results from human activity, mainly due excessive pumping groundwater, which a “natural” disaster caused by activity. Countermeasures include effective regional and/or national freshwater management local water recycling groundwater. Finally, perspectives risk hazard prevention through enhanced field monitoring, assessment with multi-criteria decision-making (MCDM), artificial intelligence (AI) technology.

Язык: Английский

---

Use of secondary additives in fly ash based soil stabilization for soft subgrades

Transportation Geotechnics, Год журнала: 2021, Номер 29, С. 100585 - 100585

Опубликована: Май 19, 2021

Язык: Английский

---

Modelling of municipal solid waste gasification using an optimised ensemble soft computing model

Fuel, Год журнала: 2020, Номер 289, С. 119903 - 119903

Опубликована: Дек. 19, 2020

Язык: Английский

---

Tunnel boring machine vibration-based deep learning for the ground identification of working faces

Journal of Rock Mechanics and Geotechnical Engineering, Год журнала: 2021, Номер 13(6), С. 1340 - 1357

Опубликована: Окт. 22, 2021

Tunnel boring machine (TBM) vibration induced by cutting complex ground contains essential information that can help engineers evaluate the interaction between a cutterhead and itself. In this study, deep recurrent neural networks (RNNs) convolutional (CNNs) were used for vibration-based working face identification. First, field monitoring was conducted to obtain TBM data when tunneling in changing geological conditions, including mixed-face, homogeneous, transmission ground. Next, RNNs CNNs utilized develop prediction models, which then validated using testing dataset. The accuracy of long short-term memory (LSTM) bidirectional LSTM (Bi-LSTM) models approximately 70% with raw data; however, instantaneous frequency transmission, increased 80%. Two types CNNs, GoogLeNet ResNet, trained tested time-frequency scalar diagrams from continuous wavelet transformation. CNN an greater than 96%, performed significantly better RNN models. ResNet-18, 98.28%, best. When sample length set as rotation period, achieved highest while proposed model simultaneously high feedback efficiency. could promptly identify conditions at without stopping normal process, parameters be adjusted optimized timely manner based on predicted results.

Язык: Английский

---

Distribution characteristics and utilization of shallow geothermal energy in China

Energy and Buildings, Год журнала: 2020, Номер 229, С. 110479 - 110479

Опубликована: Сен. 15, 2020

Язык: Английский

---

Real-time prediction of shield moving trajectory during tunnelling using GRU deep neural network

Acta Geotechnica, Год журнала: 2021, Номер 17(4), С. 1167 - 1182

Опубликована: Июль 30, 2021

Язык: Английский

---

Dynamic prediction of mechanized shield tunneling performance

Automation in Construction, Год журнала: 2021, Номер 132, С. 103958 - 103958

Опубликована: Сен. 15, 2021

Язык: Английский

---

Machine Learning-Based Modelling of Soil Properties for Geotechnical Design: Review, Tool Development and Comparison

Archives of Computational Methods in Engineering, Год журнала: 2021, Номер 29(2), С. 1229 - 1245

Опубликована: Июль 5, 2021

Язык: Английский

---