为更好地分析监测序列的波动特征进而识别大坝整体性态演变规律,将多重分形消除趋势波动分析( MF-DFA)方法应用到大坝监测序列分析,并以一混凝土坝垂线监测序列为例,通过单个测点、环境量的长程相关性和多重分形特征分析及3个不同坝段测点比较,结果表明:坝体变形序列具有较强的多重分形特征,大波动受温度和库水位的周期性变化影响,小波动则受水位的频繁涨落控制;同时,溢流坝段和挡水坝段又因其环境荷载条件和工作机理不同波动特征略有差异;总体看,变形序列受环境量控制,大坝整体性态正常。实例分析结果表明MF-DFA可以从局部和整体两个层面评价大坝的工作性态及其演变规律。
In order to analyze the fluctuant features of monitoring series for recognize the evolution pattern of dam performance, we apply the Multifractal Detrended Fluctuation Analysis ( MF-DFA) to time series analysis of dam safety monitoring. Taking the vertical monitoring series of a concrete dam as an example, we analyze the long-range correlations and multifractal characteristics of a single monitoring point and environmental variables as well as compare three different dam monitoring points. The results show that there exists stronger multifractal characteristics in dam deformation series. The periodic change of temperature and reservoir water level produce large fluctuation and frequent ebbs and flows of water level control small fluctuation. The fluctuation characteristics of overflow dam section and water retaining dam section are different because of their environmental loading conditions and working mechanism. Overall, the environmental variables control the dam deformation series and the dam performance is normal. The analysis of the examples shows that MF-DFA can evaluate the dam's working state and its evolution pattern from the part and the global.