基于谐波小波的径流序列变异检测

Detection of the abrupt changes in runoff based on Harmonic Wavelet

【目的】用谐波小波分析非一致性径流时间序列变异点的数量及变异发生的时间。【方法】针对径流变异点检测方法复杂、难于识别全部变异点等问题,引入一种新的时频分析技术,利用谐波小波分析进行径流时间序列的变异点诊断。【结果】对实测的渭河华县站1961-2004年的径流序列进行谐波小波分析,视频图中的网格图和等高线图显示,1972年、1982年和1994年是径流的变异年。【结论】相对于传统的突变检测方法,谐波小波分析方法在分析非线性时间序列时具有方法简便、可行、有效、分辨率高、对突变点敏感等特点,能够有效地诊断出高度复杂、非线性序列的变异点。

【Objective】 This paper focused on the number and time of abrupt changes in non-consistency runoff time series using the harmonic wavelet analysis.【Method】 The traditional detection methods of runoff abrupt changes are complex and difficult to identify the total variance point.Therefore,a new time-frequency analysis techniques is proposed to diagnose abrupt change points using harmonic wavelet analysis.【Result】 Harmonic wavelet transformation was used to analyze monthly runoff at the Huaxian station in Wei river from 1961 to 2004.The grid graph and contour map showed that there were abrupt changes of runoff in 1972,1982 and 1994.【Conclusion】 Compared to traditional detection methods,harmonic wavelet analysis presents simple,feasible,effective,high-resolution and sensitive features to abrupt changes in the analysis of nonlinear time series,and it can effectively diagnose abrupt changes in highly complex and nonlinear sequence.