Study on the technology of blasting to squeeze silt and build embankment in deep muddy soft soil under complex environment DOI
Jianfeng Li,

Pengyuan An,

Ronghan Wu

et al.

Research Square (Research Square), Journal Year: 2024, Volume and Issue: unknown

Published: Oct. 3, 2024

Abstract Blasting mud displacement technology has always played a crucial role in reclamation projects. To further analyze the effectiveness and application of blasting dike construction complex thick silt soft soil layers, this study first utilized excellent linear classification prediction capabilities Support Vector Machines (SVM) to construct model for slope displacement. Additionally, traditional process was optimized by proposing full lateral technique aimed at achieving mud-rock layers. The eliminates need end blasting, instead using fixed-point widening methods accomplish task. In experimental results analysis section, performance SVM models with different kernel functions tested. indicated that RBF had best performance, mean squared error values measurement points not exceeding 0.35. By adjusting parameters sites comparing settlement four technique, it found plan is feasible, all remaining within reasonable ranges. This provides new approach layers construction.

Language: Английский

Estimation of magnetic levitation and lateral forces in MgB2 superconducting bulks with various dimensional sizes using artificial intelligence techniques DOI Creative Commons
Shahin Alipour Bonab, Yiteng Xing, G.V. Russo

et al.

Superconductor Science and Technology, Journal Year: 2024, Volume and Issue: 37(7), P. 075008 - 075008

Published: May 21, 2024

Abstract The advent of superconducting bulks, due to their compactness and performance, offers new perspectives opportunities in many applications sectors, such as magnetic field shielding, motors/generators, NMR/MRI, bearings, flywheel energy storage, Maglev trains, among others. investigation characterization bulks typically relies on time-consuming expensive experimental campaigns; hence the development effective surrogate models would considerably speed up research progress around them. In this study, we first produced an dataset containing levitation lateral forces between different MgB 2 one permanent magnet under operating conditions. Next, have exploited develop based Artificial Intelligence (AI) techniques, namely Extremely Gradient Boosting, Support Vector Regressor (SVR), Kernel Ridge Regression. After tuning hyperparameters AI models, results demonstrated that SVR is superior technique can predict with a worst-case accuracy scenario 99.86% terms goodness fit data. Moreover, response time these for estimation datapoints ultra-fast.

Language: Английский

Citations

5
A stacked deep multi-kernel learning framework for blast induced flyrock prediction DOI Creative Commons
Ruixuan Zhang, Yuefeng Li,

Yilin Gui

et al.

International Journal of Rock Mechanics and Mining Sciences, Journal Year: 2024, Volume and Issue: 177, P. 105741 - 105741

Published: April 27, 2024

Blasting operations are widely and frequently used for rock excavation in Civil Mining constructions. Flyrock is one of the most important issues induced by blasting open pit mines, therefore needs to be well predicted order identify safety zone prevent potential injuries. For this purpose, 234 sets data were collected from Sungun Copper Mine site, a stacked deep multi-kernel learning (SD-MKL) framework was proposed estimate blast flyrock with confidence accuracy. The model uses stacking-based representation (S-RL) achieve on small-scale training sets. A (MKL) as base module S-RL framework, which multi-feature fusion strategy generate multiple kernels different kernel length reduce effort tuning hyperparameters. In addition, study further enhanced predictive capability SD-MKL introducing boosting method into hence boosted model. comparison several existing machine models implemented, i.e., ridge regression (KRR), support vector (SVM), random forest (RF), gradient decision tree (GBDT), ensemble functional link (edRVFL), SD-KRR SD-SVM. Our experimental results showed that achieved best overall performance, lowest RMSE 0.21/1.73, MAE 0.08/0.78, highest VAF 99.98/99.24. • prediction. Use stacked-representation learning. learn relationship between generated feature original feature. improve prediction accuracy SD-MKL.

Language: Английский

Citations

5
Intelligent identification of force chain particles within hang-up system of caved ore and rock based on graph neural networks DOI

Zongsheng Dai,

Hao Sun,

Shenggui Zhou

et al.

Expert Systems with Applications, Journal Year: 2025, Volume and Issue: unknown, P. 127990 - 127990

Published: May 1, 2025

Language: Английский

Citations

0
Study on the technology of blasting to squeeze silt and build embankment in deep muddy soft soil under complex environment DOI
Jianfeng Li,

Pengyuan An,

Ronghan Wu

et al.

Research Square (Research Square), Journal Year: 2024, Volume and Issue: unknown

Published: Oct. 3, 2024

Abstract Blasting mud displacement technology has always played a crucial role in reclamation projects. To further analyze the effectiveness and application of blasting dike construction complex thick silt soft soil layers, this study first utilized excellent linear classification prediction capabilities Support Vector Machines (SVM) to construct model for slope displacement. Additionally, traditional process was optimized by proposing full lateral technique aimed at achieving mud-rock layers. The eliminates need end blasting, instead using fixed-point widening methods accomplish task. In experimental results analysis section, performance SVM models with different kernel functions tested. indicated that RBF had best performance, mean squared error values measurement points not exceeding 0.35. By adjusting parameters sites comparing settlement four technique, it found plan is feasible, all remaining within reasonable ranges. This provides new approach layers construction.

Language: Английский

Citations

0