基于因果机理和邻近影响的拱坝温度场监测数据缺值插补

Interpolation of Missing Values in Temperature Field Monitoring Data of Arch Dams Based on Causal Mechanism and Neighboring Effects

利用实测温度场是提高拱坝变形监控模型性能的重要途径之一,但由于监测系统异常等原因,部分温度测点的监测数据存在缺值。为此,根据缺值段在温度时间序列中所处位置及其相对变化幅值制定判断准则,优选单测点缺值插补方法,基于动态时间规整法量化时间序列之间的相似性,构建拱坝温度场缺值插补的多测点分层标准和同层优先级准则,建立兼顾因果机理和邻近影响的插补预测模型。通过对某高拱坝的实施结果表明,所提出的方法和准则可有效实现拱坝温度场全部测点温度时间序列的系统性插补,将因果机理和邻近影响相结合的插补预测模型对84.0%以上的测点具有提升作用。

The use of measured temperature fields is an important way to improve the performance of deformation monitoring model of arch dams.However,the failure of measurement instruments will result in the loss of monitoring data at some times.In this paper,the position of missing values in the measured temperature time series and the relative variation amplitude are both used to determine the interpolation method for a single temperature time series.Similarities between different temperature time series are quantified by the dynamic time warping method,and hierarchical criteria and priorities are defined for multiple temperature time series.On this basis,prediction models used for the interpolation of missing values are established,in which both the causal mechanism and neighboring effects are considered.The results of a high arch dam show that the proposed method can effectively achieve the systematic interpolation of dam temperature fields.The proposed interpolation prediction model has an improvement effect for 84.0%temperature monitoring points.