摘要
机场道面沉降,严重影响机场安全运行。准确预测跑道工后沉降,对机场的建设与运行极为重要。以西南某机场跑道沉降变形的观测数据为依据,分别用双曲线模型、对数模型、指数模型以及灰色预测模型,对跑道沉降进行预测和对比分析,解决了小样本数下曲线预测精度较低及灰色模型对非线性预测准确度差等问题,提高了预测的精度;同时通过BP神经网络对组合预测模型的残差进行修正,最大限度地提高模型预测的精度和效果,为地基沉降预测提供借鉴。
The settlement of airport pavements seriously affects the safe operation of airports.Accurately predicting the settlement of runway construction is extremely important for airport construction and operation.Based on the observed data of runway settlement and deformation at a southwest airport,the hyperbolic model,logarithmic model,exponential model,and GM(1,1)model were used to predict and compare the runway settlement.The issue of low prediction accuracy of curve prediction under small sample size and the problem of poor accuracy of the GM(1,1)grey model in nonlinear prediction were addressed to improv the prediction accuracy.Furthermore,by using Back Propagation Neural Network to correct the residuals of the combined prediction model,the accuracy and effectiveness of the model prediction were maximized.Reference is provided for predicting ground settlement.
作者
方学东
顾天宇
舒富民
FANG Xuedong;GU Tianyu;SHU Fumin(School of Civil Aviation Flight University of China,Deyang 618300,Sichuan,China;CACC Southwest Design and Research Institute Co.Ltd.,Chengdu 610202,China)
出处
《科技和产业》
2024年第18期196-202,共7页
Science Technology and Industry
关键词
沉降预测
曲线预测模型
灰色预测模型
组合预测模型
BP神经网络
settlement prediction
curve prediction model
grey prediction model
combined prediction model
BP neural network
