Research On AMD-HHT Method And Its Application In Seismic Reservoir Prediction

Seismic signal is a typical non-stationary signal,which contains lots of underground geological information.And it is the purpose of seismic interpretation that to extract the effective information from seismic signals as much as possible and link it with the actual geological information.In seismic data interpretation,seismic spectral decomposition technology and seismic attribute technology are common means of seismic reservoir predication by analyzing and extracting effective information of seismic signals.However,with the increasing difficulty of oil and gas exploration,the accuracy requirement of seismic reservoir prediction is becoming more and more higher.The traditional time-frequency analysis method has been difficult to satisfy the accuracy requirement.It is necessary to research new time-frequency analysis method,so as to improve the accuracy of oil and gas prediction in complex reservoirs.Firstly,several traditional time-frequency analysis methods based on Fourier transform(short-time Fourier transform,wavelet transform,S transform)and adaptive Hilbert-Huang transform(HHT)are studied.Through the analysis of simulation results,it is found that because of the existence of window function,although the short-time Fourier transform,wavelet transform,and S transform have been greatly improved,but they are all limited by the Heisenberg uncertainty principle,their time-frequency resolution and time-frequency focusing always can't achieve the best.However,the HHT method does not have this problem.It has a high time-frequency resolution and focusing.It is very suitable for the analysis of non-stationary signals and can better reflect the characteristics of seismic signals.However,due to the inherent defects of empirical mode decomposition(EMD),which is a key step in HHT,there are some shortcomings in HHT,such as modal aliasing,endpoint effect,inseparability of closely spaced frequencies and so on.The existence of these problems will have a great impact on the results of seismic spectral decomposition and some seismic attributes extraction,and bring difficulties to seismic reservoir prediction.In order to solve these problems of HHT method,a new signal decomposition method,analytical mode decomposition(AMD),is introduced in this paper.AMD method can solve the problems of mode aliasing and close-spaced frequency which exist in EMD decomposition.It can greatly improve the accuracy of signal decomposition and combine it with HHT method to form a new time-frequency analysis method,AMD-HHT method.Compared with the conventional HHT method,the new method has stronger decomposition ability,higher time-frequency resolution and is more suitable for seismic signal analysis.Then,AMD-HHT method is applied to seismic reservoir prediction.The AMD-HHT method is used to decompose the forward model and the real seismic profiles.By searching for the oil and gas response characteristics of "low frequency energy enhancement and high frequency energy attenuation" in the frequency-division profiles,the oil and gas location in the seismic profiles is successfully detected,and the validity of reservoir prediction based on AMD-HHT method is verified.In order to improve the accuracy and efficiency of reservoir prediction,this paper also extracts several oil and gas sensitive attributes on the basis of spectral decomposition,such as frequency division amplitude difference attributes,low frequency energy enhancement attributes and high frequency energy attenuation attributes,and proves the validity of these oil and gas sensitive attributes for reservoir oil and gas detection through several real cross-well profiles.Then,the spectral decomposition based on AMD-HHT and seismic attribute extraction technology studied in this paper are applied to the actual work field to verify the effectiveness of seismic attributes in predicting oil and gas reservoirs.Then,considering the complexity of carbonate reservoir and the multi-solution of seismic single attribute prediction,this paper adopts three methods:seismic attribute fusion method based on t-SNE algorithm,seismic attribute clustering method based on Gauss mixture clustering algorithm,and seismic attribute classification method based on support vector machine,and combines well data to carry out multi-attribute analysis of the seismic attributes calculated previously.The purpose of analysis is to reduce the multi-solution and improve the accuracy and reliability of reservoir prediction.And the favorable oil-containing distribution of the work field is given by the use of the optimal method.Applying the spectral decomposition and multi-attributes analysis technology based on AMD-HHT method in oil and gas prediction in Tahe filed,it is shown that AMD-HHT method can improve the frequency division accuracy and is conducive to extracting more accurate and effective information from seismic data.On the basis of AMD-HHT method,oil and gas sensitive seismic attributes are extracted and multi-attributes analysis is carried out,which significantly improves the accuracy of reservoir oil and gas prediction and forms a set of complex carbonate reservoir prediction technology flow with high accuracy and reliability,which combines AMD-HHT method and multi-attribute analysis.The technical flow can effectively reduce drilling risk and provide new ideas and methods for improving the accuracy of seismic reservoirs prediction.