改进的非等间距GM(1,1)模型在大坝沉降分析中的应用

Application of Improved Non-equidistance GM(1,1) Model in Dam Settlement Analysis

由于影响大坝监测数据因素多,且数据存在非等间距离散的情况,利用传统非等间距GM(1,1)模型的建模原理,分析了数据拟合与预测精度低的原因,提出了优化背景值的改进非等间距GM(1,1)模型,根据两种模型对施工期堆石坝心墙的沉降观测数据进行拟合和预测。实例表明,改进模型精度较高,更接近观测值,可应用于工程实际中。

Because many factors influence the dam monitoring data and the data are non-equidistant discrete,the reasons of low precision on data fitting and prediction are analyzed with traditional GM（1,1）model.And the improved non-equidistance GM（1,1）model is proposed by optimizing background values.According to two models,the settlement observation data of rock-fill dam core wall in construction period is fitted and forecasted.Examples show that the improved model has high precision and it approaches the measured value,which can be applied to the actual project.