摘要
研究山区高边坡变形多种单一模型组合预测模型,提出边坡变形预测的方法。以GM(1,1)模型、GM(1,1)幂模型和幂函数回归模型为例进行组合预测,分别建立最优定权组合预测模型和最优非负变权系数组合预测模型,以湖南省龙永高速公路响米坳隧道口高边坡变形为例进行了验证和比较。研究表明:2种组合模型预测精度都高于3种单一模型预测精度,最优非负变权系数组合预测模型预测精度最高,定权组合预测模型次之,单一模型中幂函数回归模型精度最低。组合模型对工程模拟预测有一定的实用价值。
A variety of single model combined forecasting model of high slope deformation in mountainous area is studied. The method of prediction of slope deformation is proposed. Taking GM( 1,1)model,GM( 1,1) power model,Power model as example. The optimal weighting combination forecasting model and optimal non negative variable weighting combination forecasting model are established respectively. In Hunan province,high slope deformation of Longyong highway xiang mi-ao tunnel as example has verified and their accuracies are compared. the results shows that: The prediction accuracy of the 2 combined models is better than that of the 3 single models. The prediction accuracy of optimal non negative variable weighting combination forecasting model is the highest,The that of optimal weighting combination forecasting model is second,and that of the power model of the single model is the lowest. The combination model has a certain practical value for engineering simulation prediction.
出处
《公路工程》
北大核心
2016年第2期162-165,195,共5页
Highway Engineering
关键词
边坡工程
变形预测
GM(1
1)幂模型
变权组合模型
slope engineering
deformation monitoring
GM(1
1) power model
variable weight combination model