时间: 2025-01-19 | 次数: |
顿志林,杜海广,邹友峰,等. Tanh时间函数在采空区地表沉降预测中的应用[J].河南理工大学学报(自然科学版),doi: 10.16186/j.cnki.1673-9787.2024030077.
DUN Z L,DU H G,ZOU Y F,et al. Application of Tanh time function in prediction of surface subsidence in goaf[J]. Journal of Henan Polytechnic University( Natural Science), doi: 10.16186/j.cnki.1673-9787.2024030077.
Tanh时间函数在采空区地表沉降预测中的应用(网络首发)
顿志林 1,2,杜海广 1,3,邹友峰 4,顿志元 5,任连伟 1,2,
王文唱 6,陈绍杰 7
(1. 河南理工大学 土木工程学院,河南 焦作 454000;2. 河南省采空区场地生态修复与建设技术工程研究中心,河南 焦作 454000;3. 黄河交通学院 交通工程学院,河南 焦作 454950;4. 河南理工大学 测绘与国土信息工程学院,河南 焦作 454000;5. 焦作大学 土木建筑工程学院, 河南 焦作 454000;6. 河南理工大学 安全科学与工程学院,河南 焦作 454000;7. 山东科技大学 能源与矿业工程学院, 山东 青岛 266590)
摘要: 目的 针对现有采空区地表沉降动态预测方法中存在单项时间函数模型预测效果不佳、可靠性不足和变权组合模型确定权重难度大等问题, 方法 引入双曲正切函数(Tanh),并对该函数进行优化,提出一种适用于采空区地表沉降预测的双参数Tanh时间函数,分析参数变化对Tanh时间函数曲线形态的影响规律,给出Tanh时间函数参数的辨识方法;建立采空区地表沉降预测方差倒数法等权组合模型和误差平方和倒数法变权组合模型,并应用于钱家营矿1176东工作面采空区地表实例中。 结果 结果表明,对于23号点,Tanh时间函数模型的均方根误差(Root Mean Squared Error,RMSE)、平均绝对误差(Mean Absolute Error,MAE)、平均绝对百分比误差(Mean Absolute Percentage Error,MAPE)分别为35.03,28.192,0.102,均小于常用的MMF时间函数模型的对应指标,此外,2种组合模型的均方根误差(RMSE)、平均绝对误差(MAE)均小于其组合中的单项时间函数模型,其中,误差平方和倒数法变权组合模型的RMSE,MAE,MAPE在4种模型中均最小。 结论 本文提出的Tanh时间函数模型能够实现采空区地表沉降的有效动态预测。利用误差平方和倒数法变权组合模型计算可变权重简单、快捷,而且相对于方差倒数法等权组合模型表现出较优的预测准确性和稳定性,在地表复杂变化区间的沉降预测值与实测值仍大致相同,说明应用误差平方和倒数法变权组合模型进行采空区地表沉降动态预测更具优势。
关键词: 采空区场地;地表沉降;双曲正切时间函数;误差平方和倒数法变权组合模型
中图分类号:TD325.2
doi: 10.16186/j.cnki.1673-9787.2024030077
基金项目: 国家自然科学基金资助项目(U23A20600)
收稿日期:2024-03-18
修回日期:2024-10-11
网络首发日期:2025-1-19
Application of Tanh time function in prediction of surface subsidence in goaf (Online)
DUN Zhilin1,2,DU Haiguang1,3,ZOU Youfeng4,DUN Zhiyuan5,REN Lianwei1,2,
WANG Wenchang6,CHEN Shaojie7
(1. School of Civil Engineering, Henan Polytechnic University, Jiaozuo 454000, Henan,China; 2. Henan Engineering Research Center for Ecological Restoration and Construction Technology of Goaf Sites, Jiaozuo 454000, Henan, China;3. School of Traffic Engineering, Huanghe Jiaotong University, Jiaozuo 454950, Henan,China; 4. School of Surveying and Land Information Engineering, Henan Polytechnic University, Jiaozuo 454000, Henan,China; 5. School of Civil and Architecture Engineering, Jiaozuo University, Jiaozuo 454000, Henan,China;6. College of Safety Science and Engineering, Henan Polytechnic University, Jiaozuo 454000, Henan, China; 7. College of Energy and Mining Engineering, Shandong University of Science and Technology, Qingdao 266590, Shandong, China)
Abstract: Objectives In response to the individual time function model with unsatisfactory prediction effect and insufficient reliability as well as the difficulty of calculating weights in the variable weight combination model in the existing methods for dynamic prediction of surface subsidence in the goaf site. Methods Firstly, the hyperbolic tangent function (Tanh) is introduced and optimized to derive a two-parameter Tanh time function applied to surface subsidence prediction, the parameter influence on the surface subsidence, subsidence velocity and subsidence acceleration curve morphology of Tanh time function is analyzed, and the method of identifying function parameters is given. Secondly, the variance reciprocal method equal-weight combination model and the error square sum reciprocal method variable-weight combination model are established for surface subsidence prediction analysis. Finally, the combination models are combined for prediction and used in the goaf surface example of Qianjiaying Coal Mine 1176 east working face. Results The results show that for point 23, the root mean square error (RMSE), mean absolute error (MAE), and mean absolute percentage error (MAPE) of the Tanh time function model are 35.03, 28.192, and 0.102, respectively, which are smaller than the corresponding indexes of the commonly used MMF time function model. In addition, the root mean square error (RMSE) and mean absolute error (MAE) of the 2 combination models are smaller than those of the single time function models in their combinations, in which the RMSE, MAE, and MAPE of the variable weight combination model of the squared sum of errors and the inverse method are the smallest among the 4 models. Conclusions The Tanh time function model proposed in this paper is capable of realizing effective dynamic prediction of goaf surface subsidence. The calculation of variable weights in the error square sum reciprocal method variable-weight combination model is simple and efficient, and shows better prediction accuracy and stability compared to the variance reciprocal method equal-weight combination model. The predicted values of subsidence in the complex change interval of the goaf surface are remained approximately the same as the measured values. It is indicated that the application of the combined model is more superior in the dynamic prediction of goaf surface subsidence.
Key words:goaf site; surface subsidence; hyperbolic tangent time function; the error square sum reciprocal method variable-weight combination model
CLC: TD325.2