Accurate prediction of drilling efficiency is critical for developing the earth-rock excavation schedule.The single machine learning(ML)prediction models usually suffer from problems including parameter sensitivity an...Accurate prediction of drilling efficiency is critical for developing the earth-rock excavation schedule.The single machine learning(ML)prediction models usually suffer from problems including parameter sensitivity and overfitting.In addition,the influence of environmental and operational factors is often ignored.In response,a novel stacking-based ensemble learning method taking into account the combined effects of those factors is proposed.Through multiple comparison tests,four models,e Xtreme gradient boosting(XGBoost),random forest(RF),back propagation neural network(BPNN)as the base learners,and support vector regression(SVR)as the meta-learner,are selected for stacking.Furthermore,an improved cuckoo search optimization(ICSO)algorithm is developed for hyper-parameter optimization of the ensemble model.The application to a real-world project demonstrates that the proposed method outperforms the popular single ML method XGBoost and the ensemble model optimized by particle swarm optimization(PSO),with 16.43%and 4.88%improvements of mean absolute percentage error(MAPE),respectively.展开更多
Microseismic source localization plays a critical role in monitoring mining-induced dynamic disasters,assessing rock mass stability,and analyzing excavation-induced disturbances.With increasing monitoring accuracy and...Microseismic source localization plays a critical role in monitoring mining-induced dynamic disasters,assessing rock mass stability,and analyzing excavation-induced disturbances.With increasing monitoring accuracy and data volume,various localization techniques have emerged to suit different scenarios.We systematically review the development of current microseismic location methods,which can be broadly categorized into three types:(1)Pickingbased methods,such as the Geiger and double-difference algorithms,which are suitable for well-constrained velocity models;(2)Waveform stacking-based localization methods,such as the source scanning algorithm(SSA)and cross-correlation stacking,which eliminate the need for arrival-time picking.These techniques exhibit strong noise resistance and are particularly well-suited for environments with low signal-to-noise ratios(SNR);and(3)Deep learning-based automatic localization approaches,such as PhaseNet and LOCFLOW,which are suitable for large-scale,intelligent monitoring.Furthermore,key factors affecting localization accuracy,such as sensor array geometry,arrival-time picking errors,and velocity model uncertainties,are discussed,along with optimization strategies including 3D velocity tomography,non-predefined velocity inversion,and time-varying velocity modeling.Finally,we explore future directions,including multi-station collaborative deep learning models,intelligent denoising techniques,and risk prediction frameworks constrained by statistical seismology,aiming to advance microseismic localization toward higher precision and robustness.展开更多
基金supported by the Yalong River Joint Funds of the National Natural Science Foundation of China(No.U1965207)the National Natural Science Foundation of China(Nos.51839007,51779169,and 52009090)。
文摘Accurate prediction of drilling efficiency is critical for developing the earth-rock excavation schedule.The single machine learning(ML)prediction models usually suffer from problems including parameter sensitivity and overfitting.In addition,the influence of environmental and operational factors is often ignored.In response,a novel stacking-based ensemble learning method taking into account the combined effects of those factors is proposed.Through multiple comparison tests,four models,e Xtreme gradient boosting(XGBoost),random forest(RF),back propagation neural network(BPNN)as the base learners,and support vector regression(SVR)as the meta-learner,are selected for stacking.Furthermore,an improved cuckoo search optimization(ICSO)algorithm is developed for hyper-parameter optimization of the ensemble model.The application to a real-world project demonstrates that the proposed method outperforms the popular single ML method XGBoost and the ensemble model optimized by particle swarm optimization(PSO),with 16.43%and 4.88%improvements of mean absolute percentage error(MAPE),respectively.
基金funded by the Deep Earth Probe and Mineral Resources Exploration-National Science and Technology Major Project(2024ZD1004505)Gansu Provincial Joint Research Fund for the Year 2024(24JRRA803)+1 种基金Guangdong Provincial Key Laboratory of Geophysical High-resolution Imaging Technology(2022B1212010002)the National Natural Science Foundation of China(42174128).
文摘Microseismic source localization plays a critical role in monitoring mining-induced dynamic disasters,assessing rock mass stability,and analyzing excavation-induced disturbances.With increasing monitoring accuracy and data volume,various localization techniques have emerged to suit different scenarios.We systematically review the development of current microseismic location methods,which can be broadly categorized into three types:(1)Pickingbased methods,such as the Geiger and double-difference algorithms,which are suitable for well-constrained velocity models;(2)Waveform stacking-based localization methods,such as the source scanning algorithm(SSA)and cross-correlation stacking,which eliminate the need for arrival-time picking.These techniques exhibit strong noise resistance and are particularly well-suited for environments with low signal-to-noise ratios(SNR);and(3)Deep learning-based automatic localization approaches,such as PhaseNet and LOCFLOW,which are suitable for large-scale,intelligent monitoring.Furthermore,key factors affecting localization accuracy,such as sensor array geometry,arrival-time picking errors,and velocity model uncertainties,are discussed,along with optimization strategies including 3D velocity tomography,non-predefined velocity inversion,and time-varying velocity modeling.Finally,we explore future directions,including multi-station collaborative deep learning models,intelligent denoising techniques,and risk prediction frameworks constrained by statistical seismology,aiming to advance microseismic localization toward higher precision and robustness.
