| In the seismic survey domain, the research for raising the seismic survey resolution is always a hot spot topic. Especially in recent years, along with seismic survey development footsteps unceasing forward, sought for the covert oil gas pool to become the key point which the people studied, the thin layer's recognition and the inversion were also a very important research branch.The vertical resolution also calls the longitudinal resolution or the time resolution, is refers to distinguishes some thickness stratum to go against the bottom boundary ability, may also be distinguishes two to have certain gap infertile land level ability.In the thin layer reservoir forecasting technology's research, often"the false appearance"or"the trap"creates the explanation or the forecasting result mistake because of the seismic data. Therefore, uses the advanced computational method, carries on each kind of model to develop, studies and analyzes these models the earthquake to reflect the characteristic, is the thin reservoir forecasting technology important link. According to the current high resolution seismic survey's request, this article conducts the research to the thin layer earthquake geology mathematical model, has established the thin layer geological model, has studied the cuneate thin layer reflection oscillation amplitude with emphasis along with the frequency shift characteristic, This qualitative or the quota seeks in the thin measure to us various single-layer thickness or thin measure thickness, forecast that the thin reservoir the spatial mop rule has the reality guiding sense. Academician Li Qingzhong had said that the wave impedance section plane is the high resolution exploration final achievement. This indicated the wave impedance inversion in the earthquake technology special position. This article first from enhances the wave impedance inversion to seismic data resolution angle embarking, has realized the L1 norm restraint sparse pulse wave impedance inversion. The sparse pulse inversion is based on the sparse pulse deconvolution foundation above wave impedance inversion method, the premise is the reflection factor sequence which the supposition stratum's wave impedance model corresponds is the sparse distribution,is by plays the leading role mainly (strong) the reflection factor sequence to be composed with the Gauss background weak reflection factor sequence iteration. Acts according to the sparse principle extraction reflection factor from the channel, after wavelet convolution produces the composite trace; Using the composite trace and the primitive channel residual's size revision participation convolution's reflection factor's integer, makes the composite trace again; So the iteration, finally obtains one to be able the optimal approximation primitive channel reflection factor sequence.This article uses the sparse algorithm is the L1 norm restraint sparse solution, introduced one kind of improvement inner point algorithm to use in solving in the inversion process L1 restraint linear programming problem, Raised the counting yield, has tested the algorithm feasibility to the Marmousi model's reflection factor's inversion result, And will use in after the improvement wave impedance inversion method to the actual earthquake data inversion, The wave impedance inversion's result, compares with the seismic cross-section, the geological parameter obtained the refinement, the big structure is as before clear, the main position comes out precisely by the inversion. We reflection factor section plane more explicit demonstration which obtains from the inversion process, the inversion process obviously raised the earthquake resolution, the alignment has been also more reliable.The traditional inversion method still could not the inversion leave with 1/4 wave length definition limit resolution following thin layer, must therefore probably study the new thin layer inversion method. The time-frequency parsing technique has provided one kind in the time frequency range analysis, the explanation earthquake data new way. According to the single channel earthquake data's time-frequency distribution, may realize to the three dimensional seismic data analysis, is advantageous to the seismic data carries on the quota the geologic interpretation and the sequence contact surface automatic pick.Gabor the transformation is one development early time-frequency analysis method, it the concept is direct, simultaneously the algorithm is simple, already became studies the non-steady signal very powerful tool, obtains the widespread application in many domains, like signal instantaneous frequency estimate and so on, and is other time-frequency method foundation. Through the model experiment as well as to the actual data's analysis, we may obtain such conclusion: Regarding the specific signal, chooses the good window box to count the Gabor transformation to be able to obtain the very good time-frequency resolution effect, But applies the Gabor transformation to has the wideband characteristic, signal characteristic rich seismic signal time-frequency analysis sometimes, sometimes the window choice question became has overcome with difficulty, very difficult to cause a window simultaneously to satisfy several kind of different requests; Moreover when the window length's choice question has also highlighted. The window box number's length with the time-frequency distribution map frequency resolution which obtains has the direct relation. Must obtain the good frequency range effect, the request has the long signal hour of observation (window box number long), then regarding the change very quick signal, will lose the time information, cannot reflect the frequency and the time variation relations correctly; Otherwise, if takes the window box number is very short, although may obtain the good time domain effect, but according to Heisenberg measured that does not permit the principle, this will certainly to pay a price in the frequency, will obtain the signal frequency band to stretch, frequency range resolution drop. Therefore we to obtain a better effect, must consider that uses the more suitable analysis method. The match traces the (MP) algorithm is one kind of greedy iterative algorithm, it may decompose the earthquake record into a series of wavelets, matches their time-frequency characteristic, regarding massive redundant dictionary iteration selection wavelet. Usually uses the Gabor function or based on the wavelet transformation time-frequency atom regarding the redundant dictionary's establishment. Regarding the reflection seismic signal, for the quota expression sound wave dissemination in the hole medium's energy decay and the speed frequency dispersion, usually selects the Morlet wavelet to carry on the dictionary structure. Most superior wavelet which obtains regarding the MP algorithm, may further obtain meantime the frequency spectrum using the Wigner distribution, this time frequency spectrum has the good time-frequency resolution, may further use in the thin layer recognition, the reservoir portray, the hydrocarbon compound instructed, as well as thin layer thickness inversion.This article uses in the MP algorithm to the actual seismic signal time-frequency analysis, compares other time-frequency analysis method saying that the match tracing algorithm obtains the signal is sparse, can obtain"cleanly"the time frequency spectrum, uses when distinguishes the thin layer, the thin layer can also obtain the very good separation, obtains the best thin layer useful information in the limit resolution. It has been separated from when the window limit, the retention desired signal maximum limit, has overcome to low-frequency signal forfeit. Therefore the match tracing algorithm may use in the region the thin layer recognition, further may use in the geophysics analysis, like hydrocarbon compound reservoir survey. Has established the depositional cycle body's model, confirmed the match tracing algorithm through the experiment regarding to study the thin interbed interior the partial structure validity. What needs to point out: The time-frequency change characteristic and between the geological cycle body structure will not be one kind simple, the only corresponding relationships, but it regarding the forecast cycle structure was still one kind of useful information, it may take the use oscillation amplitude information research cycle structure the important supply.
|
PDF Full Text Request