大坝变形监测序列的阶段性趋势识别方法研究

Study on the Phased Trend Recognition Method of Deformation Monitoring Sequence for Dams

对大坝变形监测序列进行阶段性趋势识别,有助于加深对不同时间尺度上大坝变形监测序列的演变规律认识,对于大坝安全运行和管理具有重要意义。采用启发式分割算法(BG算法)对大坝变形监测序列进行突变点识别,能够有效避免突变点对趋势识别的干扰;在此基础上,通过改进ITA方法对分段子序列进行趋势识别,改进后的ITA方法能够保留序列内部相关性,具有较好适用性。工程实例分析表明,所提方法能够有效识别出大坝变形监测序列中的突跳点,将监测序列分为具有稳定趋势的子序列,并能识别出每段子序列的变化趋势。

The phased trend recognition of deformation monitoring sequences for dams can deepen the understanding of the evolution laws of deformation monitoring sequences at different time scales,which is of great significance for the safe operation and management of dams.The heuristic segmentation algorithm(BG algorithm)is adopted to identify the mutation points of deformation monitoring sequences for dams,which can effectively avoid the interference of mutation points in trend recognition.On this basis,the segmented sub-sequences are trend identified using an improved ITA method,which can retain the internal correlation of the sequences and has good applicability.The engineering case analysis shows that the proposed method can effectively identify the mutation points in the deformation monitoring sequences for dams,divide the monitoring sequences into sub-sequences with stable trends,and identify the changing trends of each sub-sequence.