摘要
Abstract
In response to the integrated construction of subsidence monitoring and prediction in mining areas,this study pro-posed a surface subsidence prediction model combining the Newton-Raphson optimization(NRBO)algorithm,variational mode decomposition(VMD)algorithm,and support vector machine(SVM)algorithm.Taking a coal mine area in Inner Mongolia as a case study,small baseline subset interferometric synthetic aperture radar(SBAS-InSAR)data was used to obtain the subsidence time series of the study area.Based on the cumulative subsidence values during the study period,char-acteristic points were selected,and the NRBO-VMD-SVM combined model was used to decompose,predict,and recon-struct the subsidence time series at these characteristic points,yielding corresponding subsidence prediction data.The predic-tion results were then compared with the SBAS-InSAR monitoring results for analysis.The research results show that the NRBO-VMD algorithm effectively decomposes and predicts subsidence time series,proving reasonable and efficient perfor-mance.Compared to other conventional prediction models such as the back propagation(BP)neural network and long short-term memory(LSTM)networks,the NRBO-VMD-SVM combined model reduces the root mean square error(RMSE)by more than 24.7%and the mean absolute error by more than 32.3%.The coefficient of determination fluctuates between 0.78 and 0.98,which indicates a significant improvement in prediction accuracy and stability.This validates the feasibility of using the NRBO-VMD-SVM combined model for surface subsidence prediction in mining areas.关键词
沉降监测/差分干涉测量短基线集时序分析技术(SBAS-InSAR)/沉降预测/牛顿-拉夫逊优化的变分模态(NRBO-VMD)算法/支持向量机(SVM)算法Key words
subsidence monitoring/small baseline subset interferometric synthetic aperture radar(SBAS-InSAR)/subsidence prediction/Newton-Raphson based optimization variational mode decomposition(NRBO-VMD/support vector machine(SVM)分类
计算机与自动化
