摘要
针对地铁工程地表沉降监测预报问题,分析时间序列分析中求和自回归移动平均模型理论;结合某地铁车站地表沉降的一个监测点连续39期的沉降实测资料,研究地铁施工过程中基坑周边地表沉降监测点的变化趋势。应用ADF单位根检验完成时间序列数据平稳性检验,根据自相关偏自相关函数分析来进行模型参数定阶。在经过比较ARIMA(1,1,1)、ARIMA(1,1,2)、ARIMA(1,2,1)、ARIMA(1,2,2)四种模型,基于AIC准则和BIC准则以及相关系数R2确定最终拟合模型为ARIMA(1,1,2),计算得出模型实际值与预测值吻合较好,并经计算得出平均残差值和平均相对误差分别为-0.0185mm、2.72%,证明了模型的有效性,满足短期预测需求,可为地铁车站施工安全监测预报研究提供一定的借鉴。
Aiming at some problems of the surface settlement monitoring and prediction for the subway engineering,and according to the theory of time series analysis and summation auto- regressive moving average model,this article studies the changing trends of monitoring point of surface subsidence of foundation pit,combining the measured data of the surface settlement of a subway station 39 consecutive period of a monitoring the settlement of the subway construction. Applying ADF unit root test to complete time series data stationarity test,the model parameters are determined by the autocorrelation function analysis. Comparing the four models of ARIMA（ 1,1,1）,ARIMA（ 1,1,2）,ARIMA（ 1,2,1）,ARIMA（ 1,2,2） and based on the AIC criterion and BIC rule and correlation coefficient R2,the final fitting model ARIMA（ 1,1,2） is adopted. The calculated results are in good agreement with the predicted values,and the average residual value and the average relative error are- 0. 0185 mm and 2. 72% respectively. This proves the validity of the model and meets the demand of short- term prediction. Therefore,it can provide some reference for the subway station construction safety monitoring and forecast research.
出处
《长春师范大学学报》
2015年第10期75-80,共6页
Journal of Changchun Normal University
基金
宿州学院大学生科研项目(KYLXLKYB15-06
KYLXLKYB15-09)
宿州学院安徽省煤矿勘探工程技术研究中心开放课题(2013YKF04)
宿州学院校级项目(2011yss03)
关键词
地表沉降监测
求和自回归移动平均模型
预测
surface subsidence detection
summation auto-regressive moving average model
prediction
