Abstract:

As a timevarying signal,the inflow of the Three Gorges Reservoir is nonlinear and nonstationary,which contains the frequency information that not only can reflect the cycle characteristics of above the annual variation scale in the Yangtze River basin,but also is different in every flood season and dry season.In this paper,the HHT method was used to process and analyze the inflow data in the Three Gorges Reservoir area.By using the EMD method,the signal was decomposed and the analysis results could clearly distinguish the intrinsic mode function components,the frequency of which was different and range from high value to low value.Though Hilbert transfer for the intrinsic mode function components,the timefrequency characteristics showed that the frequency amplitude had a little change in the twenty days to one year cycle range.But the frequency depicted the slight increasing tendency since the impoundment in 2003,and the delimitation of frequency variation was not obvious between the flood and the dry season.Furthermore,annual cycle of periodic change information was stable,and the residual signal represented the inflow of the overall trend,which reflected the trend of decrease in the whole reservoir area since 1998