摘要
确定矿山开采沉陷边界有助于评估矿区生产活动对周围环境和基础设施的潜在影响,为制定有效的灾害防控措施提供技术手段。充分考虑到概率积分法等传统方法在划定矿山开采沉陷边界时的不足,采用在获取矿区大范围高精度地表沉降数据方面具有优势的SBAS-InSAR技术,并结合淘金算法(Gold Rush Optimizer,GRO)优化双向长短期记忆(Bidirectional Long Short Term Memory,BiLSTM)模型的预测方法,实现矿区开采沉陷边界划定。以红会煤矿为研究对象,依据SBAS-InSAR技术提取矿区沉降边缘高相干点在2018-11-29—2020-02-04时间段内共37期沉降数据,以下沉10 mm等值线划定沉陷边界,利用GRO-BiLSTM优化模型预测高相干点的地表沉降值,并将预测结果与LSTM和BiLSTM模型预测结果进行了对比分析。结果表明:GRO-BiLSTM模型在整体测试集中均方根误差为3.204mm,比LSTM和BiLSTM模型分别降低了22.16%和8.21%;平均绝对误差为2.062 mm,比LSTM和BiLSTM模型分别降低了23.96%和5.43%,表明该方法可以有效监测和预测矿区边界地区的沉陷状况。
The determination of mining subsidence boundary is helpful to assess the potential impact of production activities of mining area on surrounding environment and infrastructure,and provides technical means for formulating effective disaster prevention and control measures.Taking full account of the shortcomings of traditional methods such as probability integral method in delineating mining subsidence boundary,SBAS-InSAR technology,which has advantages in obtaining large-scale and high-precision surface settlement data in mining area is adopted.Combined with the prediction method of Bidirectional Long Short Term Memory(BiLSTM)model optimized by Gold Rush Optimizer(GRO),the mining subsidence boundary was demarcated.Taking Honghui Coal Mine as the study example,a total of 37 periods of settlement data of high coherence points on the settlement edge of the mine during the period from 2018-11-29 to 2020-02-04 were extracted based on SBAS-InSAR technology.The subsidence boundary was demarcated by the subsidence contour of 10 mm,and the surface subsidence value of high coherence points was predicted by the GRO-BiLSTM optimization model.The predicted results were compared with those of LSTM and BiLSTM models.the results show that:the root-mean-square error of GRO-BiLSTM model in the overall test set is 3.204 mm,which is 22.16%and 8.21%lower than that of LSTM and BiLSTM model,and the average absolute error is 2.062 mm,which is 23.96%and 5.43%lower than that of LSTM and BiLSTM model,respectively.The results indicated that this method can effectively monitor and predict the subsidence in the boundary area of mining area.
作者
肖海平
范永超
陈兰兰
万俊辉
陈磊
XIAO Haiping;FAN Yongchao;CHEN Lanlan;WAN Junhui;CHEN Lei(School of Civil Engineering and Surveying&Mapping Engineering,Jiangxi University of Science and Technology,Ganzhou 341000,China;School of Resources and Civil Engineering,Gannan University of Science and Technology,Ganzhou 341000,China)
出处
《金属矿山》
CAS
北大核心
2024年第10期139-144,共6页
Metal Mine
基金
国家自然科学基金项目(编号:42361012)
江西省自然科学基金项目(编号:20212BAB204030)
江西省教育厅科学技术项目(编号:GJJ2203602)。
