基于GM-ANN组合模型的大坝变形监测数据分析方法

Dam Deformation Monitoring Data Analysis Method Based on GM-ANN Model

大坝变形监测数据由于受水位、温度、时效等多因素影响,呈现一定的趋势性和波动性。引入时间序列分析模型和方法,合理分析大坝变形监测数据序列特征,精确预测大坝变形状况,是大坝安全监测的重要内容。重点针对"贫信息"、"小样本"情况下的大坝变形监测数据序列,将灰色模型(GM)与人工神经网络(ANN)模型相结合,研究了分析和预测大坝变形监测数据的方法,并进行工程实例验证。首先利用灰色模型拟合和预测实测样本数据,然后将灰色模型残差作为人工神经网络模型的学习样本进行训练和预测,最后将上述结果进行综合得到大坝变形的预测值。应用文中所述方法对某实际大坝变形监测数据的建模检验表明,本文给出的GM-ANN模型预测精度高于GM(1,1)模型,具有较好的实用性。

Affected by factors such as water level ,temperature and aging ,dam safety monitoring data have a certain trend and volatil‐ity .Analyzing reasonably the features of the measuring data series and accurately predicting the dam deformation by the time series analysis model are two important parts of dam safety monitoring .This paper focuses on the“poor information” and“small sample”monitoring data and proposes combining grey model （GM ） with artificial neural networks （ANN） to analyze deformation data .This combined model has been also validated by engineering examples .Firstly ,data are fitted and forecast by the grey model .And then , according to the grey residuals ,artificial neural network is used to predict residuals .The final result is the combination of the two models .Based on the measured data ,the forecasting research shows that GM‐ANN model has a higher precision than GM （1 ,1） model .And it has good practicability .