隧道地表沉降预测的时变参数灰序模型TGM-AR

Time-dependent parameter gray model TGM-AR for prediction of surface settlement of tunnels

在趋势项提取法中,针对逐步回归法比较烦琐及灰色模型预测精度低的缺点,结合灰色预报系统的灰色性和时变性,根据变形体的实际变形规律,将时变参数a(t),b(t)及灰色理论应用于时序模型建模中,提出并建立时变参数灰序模型TGM-AR.时变参数灰色模型用于提取趋势项部分,时序模型用于提取随机部分.将模型应用于隧道地表的沉降分析和预测,结果表明模型预报精度高,且趋势项的物理意义明确.

For extraction of trend items, the stepwise regression method is troublesome, and the gray model is of low precision of prediction. Based on the gray and time-dependent characteristics of the gray prediction system and the deformation rules of the deformation body, the time-dependent parameters a （ t ） and b （ t ） and the gray theory were applied to time series modeling, and a time-dependent parameter gray model TGM-AR was developed. The model TGM- AR with a （ t ） and b （ t ） was used for extraction of the trend item, and the time series model was for extraction of the random parts. The application of the model to analysis and prediction of the ground surface settlement of tunnels shows that the prediction precision is high and the physical significance of the trend item is explicit.