The Research On Reservoir Prediction Based On The HHT

Time-frequency analysis technology is one of commonly used technology in seismic data processing and interpretation.Compared with traditional time-frequency analysis method,the HHT has more obvious advantages in dealing with the nonlinear nonstationary signal.The core part of HHT is the Empirical Mode Decomposition(EMD).The signals can be decomposed in terms of IMFs according to its characteristics by EMD and we don't need to consider other interferences.However,it has two serious problems: end effect and mode mixing.To deal with the problems,we discuss some improved algorithms.We use the mirror continuation method,polynomial continuation method and RBF neural network prediction method to suppress the end effect.The Ensemble Empirical Mode Decomposition(EEMD),Complementary Ensemble Empirical Mode Decomposition(CEEMD)and EMD with average value constraint are the improved algorithms for the mode mixing.The principle of the EEMD and the CEEMD are similar: Adding white noise to the signal.We all know that,the spectrum of the white noise is continuous.The extreme point features of the signal will be changed if we add much white noise to the signal,which is a effective solution to the mode mixing problem.The third method is the EMD with average value constraint.This method constraints the way of the the signal envelope by the mean filter.It can detect the small change of the signal and get the precise envelope.In this paper,on the basis of the improved algorithms of HHT,we applied it to the seismic data interpretation.The extraction of the instantaneous attributes has very important significance.Three instantaneous information can not only reflect the underground structure information such as lithology,can also reflect the quality factor.In this paper,we decompose the seismic signal into a series of Intrinsic Mode Function(IMF)by using the EMD.Then we extract the three instantaneous profile by Hilbert Transform.The HHT can also be used to divide the sedimentary cycles.We use it to process the seismic signal rather than the logging signal,so our method is not restricted by the logging data.We can get the high frequency of instantaneous phase profiles by the EEMD and Hilbert.These profiles can reflect the pinch out of the sand more clearly.The different scales of attributes' s recognition ability for the geological target is different.We extract the seismic attributes in different scales to identify the fault and sand bodies.Finally,all the seismic attributes are optimized and analyzed to get the prediction for favorable reservoirs.In general,modeling computation and practical data processing prove that the effect of improved EMD is better than that of EMD.HHT has achieved satisfactory results in the application of fine interpretation,division of sedimentary cycles,axtraction of multiscale attributes and so on.