Thin Bed Thickness Estimation Based On Spectrum Decomposition

Reservoir characterization is a crucial prerequisite to predict the economic potential of a hydrocarbon reservoir or to examine different production scenarios. Unfortunately, it is impossible to determine the exact reservoir properties at the required scale. The most abundant seismic data have a resolution of around 30 m. Wells resolve the reservoir down to the centimeter scale, but only at some points in the vertical direction.In fact, seismic resolution is the key to extraction of stratigraphic detail from seismic data and this has become more important over the last decade or so. Seismic resolution comprises two aspects - the vertical and horizontal resolution. The vertical resolution refers to the ability to distinguish two close seismic events corresponding to different depth levels, and the horizontal or spatial resolution is concerned with the ability to distinguish and recognize two laterally displaced features as two distinct adjacent events. While both aspects are important for interpreting small features on seismic data, here, we focus our attention to the vertical resolution, recognizing that migration procedures are usually put in place for collapsing the Fresnel zones that enhance spatial resolution.If the average spectrum of a seismic wavelet is centered around 30Hz, which is usually the case, reservoirs having a thickness less than 25 m, may not have top and base reflectors resolved. This may suffice for structural objectives, but stratigraphic targets are usually set to look for reservoirs 10 m or less in thickness. Attempts to achieve such objectives often lead to frequency enhancement procedures to be followed on surface seismic data. Conventional wisdom usually follows the conclusions enunciated by Widess (1973) some three decades ago. Widess proposedλ/8 as the resolution limit,λbeing the predominant wavelength in the data. In the presence of noise and the consequent broadening of the wavelet during its subsurface journey, this resolution is usually taken to be onlyλ/4, and geophysicists have been assuming this resolution limit as a gospel truth till now. So, wavelength is the yardstick for resolution, which in turn depends on velocity and frequency. Since there is nothing we can do for velocity, which shows a general increasing trend with depth, the key factor that determines resolution according to the Widess model is frequency. Thus, for getting greater reflection detail from seismic data, utmost care is exercised, first at the seismic data acquisition stage in terms of field parameters, seismic sources and improved recording adopted, and secondly, during processing where attempts are made to enhance the spectral bandwidth.Conventional trace inverse methods such as constrained sparse spike inversion and mode-based inversion can help us to know about the reservoir character. The method of constrained sparse spike inversion supposes that the reflections coefficients of strata are sparse. The pulse calculation can obtain the wide-band result. The inversion result is faithful to seismic data, simultaneously; the constraint of logging and geology can replenish the component of low and high frequencies. The vertical resolution is heightened comparatively to the conventional seismic data. I use the Interior Point method with L2-L1 norm constrain to inverse the reflection coefficient. The inversion result is faithful to seismic data, simultaneously the constraint of logging and geology can replenish the component of low and high frequencies The vertical resolution is heightened comparatively to the conventional seismic data. But this method can not break the resolution limit.Thin bed estimation based on spectrum decomposition can breakthrough the resolution limit. Many methods are introduced in this paper in detail. These methods totally can be divided into 2 types to say, spectrum analysis on reflection coefficient and spectrum analysis on wavelet. Earlier researches mostly use the reflection coefficient spectrum analysis which can provide a robust and phase independent approach to seismic thickness estimation. It builds on traditional tuning thickness estimation techniques documented by Widess and Kallweit & Wood. Whereas these traditional techniques require zero-phase data and careful picking of temporally adjacent peaks and troughs, thickness estimates derived from spectral decomposition- derived discrete Fourier components are wavelet-phase independent and require only one guide pick within the seismic zone of interest. Continuous spectrum decomposition can make a clearly spectrum result compared with the DFT. But those methods must use a wide-band frequency reflection coefficient, but usually we can not obtain an efficient wide-band frequency volume. We want to use the orient data with wavelet and it will be much more complex. Thin bed can regard as a filter which changes the amplitude, frequency and phase of the wavelet. So we use spectrum decomposition to estimate the instantaneous. Marfurt analysed the relationship between the peak frequency and the thickness of the thin bed. He the use Hilbert transform to compute the peak frequency to estimate the thickness of the thin bed. But Hilbert transform has many pitfall, so we use continuous spectrum decomposition to compute the instantaneous attribute.