Compaction quality evaluation of subgrade based on soil characteristics assessment using machine learning

Transportation Geotechnics, Год журнала: 2021, Номер 32, С. 100703 - 100703

Опубликована: Дек. 13, 2021

Язык: Английский

---

Exploring the specific heat capacity of water-based hybrid nanofluids for solar energy applications: A comparative evaluation of modern ensemble machine learning techniques

Journal of Energy Storage, Год журнала: 2022, Номер 54, С. 105230 - 105230

Опубликована: Июль 12, 2022

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

---

Artificial neural network optimized by differential evolution for predicting diameters of jet grouted columns

Journal of Rock Mechanics and Geotechnical Engineering, Год журнала: 2021, Номер 13(6), С. 1500 - 1512

Опубликована: Сен. 14, 2021

A novel and effective artificial neural network (ANN) optimized using differential evolution (DE) is first introduced to provide a robust reliable forecasting of jet grouted column diameters. The proposed computational method adopts the DE algorithm tackle difficulties in training performance networks optimize four quintessential hyper-parameters (i.e. epoch size, number neurons hidden layer, layers, regularization parameter) that govern efficacy. This approach further enhanced by stochastic gradient optimization allow 'expensive' computation efforts. ANN-DE trained prepared grouting dataset, then verified compared with prevalent machine learning tools, i.e. support vector (SVM). results show that, outperforms existing methods for predicting diameter columns since it well balances efficiency model performance. Specifically, achieved root mean square error (RMSE) values 0.90603 0.92813 testing phases, respectively. corresponding were 0.8905 0.9006 ANN, then, 0.87569 0.89968 SVM, paradigm bound be useful solving various geotechnical engineering problems regardless multi-dimension nonlinearity.

Язык: Английский

---

Auto machine learning-based modelling and prediction of excavation-induced tunnel displacement

Journal of Rock Mechanics and Geotechnical Engineering, Год журнала: 2022, Номер 14(4), С. 1100 - 1114

Опубликована: Апрель 18, 2022

The influence of a deep excavation on existing shield tunnels nearby is vital issue in tunnelling engineering. Whereas, there lacks robust methods to predict excavation-induced tunnel displacements. In this study, an auto machine learning (AutoML)-based approach proposed precisely solve the issue. Seven input parameters are considered database covering two physical aspects, namely soil property, and spatial characteristics excavation. 10-fold cross-validation method employed overcome scarcity data, promote model's robustness. Six genetic algorithm (GA)-ML models established as well for comparison. results indicated that AutoML model comprehensive integrates efficiency Importance analysis reveals ratio average shear strength vertical effective stress Eur/σv′, depth H, width B most influential variables Finally, further validated by practical prediction good agreement with monitoring signifying our can be applied real projects.

Язык: Английский

---

LGBM-based modeling scenarios to compressive strength of recycled aggregate concrete with SHAP analysis

Mechanics of Advanced Materials and Structures, Год журнала: 2023, Номер 31(23), С. 5999 - 6014

Опубликована: Июнь 22, 2023

Concrete production contributes significantly to global greenhouse gas emissions, and its manufacture requires substantial natural resources. These concerns can be partly mitigated by recycling construction demolition waste as aggregates produce Recycled Aggregate (RAC). RAC has gained momentum due lower environmental impact, costs, increased sustainability. The aim of this study was advance the reasonable use recycled aggregate in concrete achieve optimal mixture ratio design. Four advanced machine learning algorithms, Support Vector Machine (SVR), Light Gradient Boosting (LGBM), Random Forest (RF), Multi-Layer Perceptron (MLP), were employed, novel optimization biogeography-based (BBO), Multi-Verse Optimizer (MVO) Gravitational Search Algorithm (GSA), integrated predict compressive strength RAC. Six potential influential factors for considered models. employed four evaluation metrics, Taylor diagrams Regression Error Characteristic plots compare model performance. result shows LGBM-based hybrid outperformed other methods, demonstrating high accuracy predicting strength. Shapley Additive Explanation (SHAP) results emphasize importance understanding interactions between various their effects on mechanical properties findings inform development more sustainable environmentally friendly building materials.

Язык: Английский

---

Extensive overview of soil constitutive relations and applications for geotechnical engineering problems

Heliyon, Год журнала: 2023, Номер 9(3), С. e14465 - e14465

Опубликована: Март 1, 2023

A state-of-the-art review has been conducted in this work on soil constitutive modeling, which emphasized on: type, ground-water conditions, loading structural behavior, relation discipline, and dimensions. By extension also, the applications were reviewed bases of: single discipline dealing with mechanical properties modeling included slope stability problems, bearing capacity, settlement of foundations, earth pressure dynamics, structure interaction, thermal hydrological conditions; bi-discipline (coupled problems) solve problems related to thermomechanical (freeze/thaw conditions), smoothed particle hydrodynamics (SPH) hydromechanical (consolidation, collapse liquefaction) conditions soils rocks multi-discipline models complex thermo-hydromechanical (THM) rocks. This shown that (HM) models, belong or coupled are better suited for geotechnical applications, generally, while solving freeze/thaw piles these proven high performance flexibility.

Язык: Английский

---

A novel physics-informed deep learning strategy with local time-updating discrete scheme for multi-dimensional forward and inverse consolidation problems

Computer Methods in Applied Mechanics and Engineering, Год журнала: 2024, Номер 421, С. 116819 - 116819

Опубликована: Фев. 1, 2024

Язык: Английский

---

A neural network approach for the reliability analysis on failure of shallow foundations on cohesive soils

International Journal of Geo-Engineering, Год журнала: 2024, Номер 15(1)

Опубликована: Июнь 18, 2024

Abstract A collection of feed forward neural networks (FNN) for estimating the limit pressure load and according displacements at state a footing settlement is presented. The training procedure through supervised learning with error loss function mean squared norm. input dataset originated from Monte Carlo simulations variety loadings stochastic uncertainty material clayey soil domain. yield Modified Cam Clay model. accuracy FNN’s in terms relative no more than $$10^{-5}$$ 10 - 5 this applies to all output variables. Furthermore, epochs required construction are found be small amount, order magnitude 90,000, leading an alleviated data cost computational expense. Karhunen Loeve random field sum appears provide most detrimental values displacement unfavorable situation result 0.05 m, that may structural collapse if they appear founded structure. These series can easy reliable estimation failure shallow foundation therefore it useful implement geotechnical engineering analysis design.

Язык: Английский

---

A super-learner machine learning model for a global prediction of compression index in clays

Applied Clay Science, Год журнала: 2024, Номер 249, С. 107239 - 107239

Опубликована: Янв. 21, 2024

Язык: Английский

---