MXPA99001730A - Method for time lapse reservoir monitoring - Google Patents

Abstract

A method of comparing multiple seismic survey data sets of a reservoir is provided, wherein a first seismic survey data set is taken at a first time and a second seismic survey data set is taken at a second time, to detect changes in the reservoir between the first time and the second time. Generally, the method comprises design of a crossequalization function for the second data set based on a comparison of unchanging portions of the two data sets. Also provided is a processing method for preparing each survey according to similar processing steps with information taken from each survey.

Description

METHOD FOR THE INSPECTION OF LAPSOS DEPOSITS BACKGROUND OF THE INVENTION This application is a continuation in part of an application of E.U.A., serial No. 08 / 713,948, assigned to the assignee of the present invention and which includes an inventor common to the inventors of the present invention. This invention relates to the control of oil and gas deposits and, more specifically, to the processing of seismic signals from lapses deposits. The characterization and inspection of deposits in the field of oil and gas are important parts of the control of deposits and the production of hydrocarbons. Effective control of deposits is a fundamental goal of energy producing companies, since they always seek to produce the costs of finding, optimizing the locations of the drilling and increasing financial gains. One technique that is tried in that effort is the seismic inspection of lapses (also known as 4D). As fluids are extracted, assembled or injected during production and recovery, property changes occur to effective rock and deposit techniques. The ability to inspect deposit changes over time through the use of seismic methods can result in the best location of production and filling wells, the possibility of locating uncovered areas and more efficient field maintenance, elevating thus the general value of the production property in lease. In a two-dimensional approach, seismic inspection in cross-hole procedures has been examined. However, the repeated results have been compared only qualitatively and the two-dimensional studies of lapses do not contain, until now, the type of information desired in the modern control of the deposits (for example, see Paulsson et al., 1994, The Leading Edge, incorporated herein by reference). 3D techniques have also been tried of lapses that indicate complex processing of modeling and require a lot of processing without using the direct data available to the study itself. See, for example, the patent of E.U.A. 4,969,130, incorporated herein by reference. A problem in lapse processing is that many conditions change over time, not just changes in the state of the deposit. For example, the locations of the source, the receiver in the second study will necessarily be different from the first. In addition, the tide in a marine study may be higher or lower, as may the temperature of the air and water. Similarly, the specific characteristics of the sources and receptors used in the second study will be different. Other differences also occur, through changes in the status of the deposit, such as the differences in how the two studies are processed. Thus, there is a need for a method to treat the two studies whereby the processing of the differences does not detrimentally affect the result of the comparison. For example, in the concentration of seismic data, a source is used to generate seismic waves that are reflected from reflectors in the earth (for example, layer boundaries) and are received at the receivers. In some cases, the source identification signal is a tip, although, in reality, it is not perfect. During its travel through the strata and reflectors, the configuration of the signal is changed and the signal of reflection received in the receivers is no longer, therefore, a tip or even something similar. Deconvolution is the process by which the configuration of the reflection signal is "whitewashed" to recreate the configuration of the tip of the information. In another example, a broad limited band signal is used, which is the zero phase. Deconvolution is used in such a case to eliminate the distortions caused by the earth. In another example more, in the realization of the deconvolution in the frequency domain, all the frequency samples are multiplied to make them at the same level, based on the assumption that the source is a minimum phase signal, immediately ascending to a maximum value and then disappearing. This autocorrelated the stroke multiple times in the time domain, in a series of interval samples, which results in a generally symmetric small wave. The power spectrum of the small wave is then analyzed to determine which multipliers are needed at each frequency to clarify the frequency spectrum. This process is performed on a window base, both along each trace and through the record (as used herein, the term "record" refers alternatively for example, or a common receiver record, a record of CMP, a common record of downloads, an expired record of traces, etc). The autocorrelation is performed in several windows and the results are averaged to give the spectrum, based on the spectrum, the operators that are needed to flatten the spectrum are chosen. The operators are then applied to all the strokes used in the input. Typically, the window is approximately 10 times the length of the operator to be generated, measuring the number of samples. The deconvolution process of a deconvolution operator design is well known in the art and the previous example of the frequency domain is not limited, it is also routinely performed in the time domain. See, for example, Yilmaz, Investigations in Geophysics, Vol. 2, Seismic Data Processing, Societi de Exploration Geofisicists (1987) and references cited therein. In the realization of deconvolution, it is important to design a deconvolution operator dependent on the study's treatments, in order to consider the specific identification signal of the source and other equipment distortions that occur in the data. Therefore, the data have been adjusted in each study through the use of a specific optimal deconvolution operator that is not applied to other studies. The result of this difference in the use of separate deconvolution operators in lapse studies is that the structure appears in the difference records when two records are not subtracted. This result is differentiable. However, to date, no one has proposed a practical solution to the problem.

BRIEF DESCRIPTION OF THE INVENTION It is the object of the present invention to deal with the above problems. It has been discovered, contrary to previous beliefs, that a single deconvolution operator can be used in multiple data sets, not only with no detrimental results, but by improving the processing quality of the lapse comparisons of the seismic studies. According to the above, one aspect of the present invention provides a deconvolution method of multiple sets of seismic data from the same geographical area, the method comprising: designating the data dependent deconvolution operator of at least two sets of data. seismic data, where at least two sets of seismic data were recorded at different times or dates of calendars; apply the deconvolution operator in a deconvolution process to both the two data sets at least; and continue to conduct lapse processing to form a record of differences. According to one embodiment of the invention, the conduction of the previous processing of lapses comprises: providing a first reflection event (e.g., a small wave) in the first seismic survey data set having a second corresponding reflection event on the second set of seismic study data for the first reflection event and the second reflection event representing an unmodified portion of the geological structure in or near the deposit and in the first reflection event is represented by a first set of event parameters and the second reflection event is represented by a second set of event parameters. Next, a minimum acceptance difference function is provided between the first set of event parameters and the second set of event parameters. Then, the cross-match function is determined by applying to the second set of event parameters. According to another aspect of the invention, the cross-equalization function is determined in such a way that, with the application of the cross-match function to the second set of event parameters, a set of cross-matched and cross-matched event parameters is defined. the difference between the first set of event parameters and the set of cross-matched event parameters is below the minimum difference function. Then the cross-equalization function is applied to a third reflection event, the third reflection event being related to the second data set, in which a third cross-matching reflection event is defined, in the third reflection event it has a corresponding fourth reflection event in the first data set and where the third and fourth reflection effects represent a changing portion of the deposit. The comparison of the third reflection event equating cross-wise with the fourth reflection event subtracting the third reflection event by cross-matching the fourth reflection event results in the desired information. According to a more specific example embodiment, said provision of said minimum acceptance difference function comprises: the iterative selection of the modifications of the event parameters to the second set of event parameters, the application of the modifications of the parameters of the event. event to the second set of event parameters, where a modified set of event parameters is defined, the comparison of the modified set of event parameters to a first set of event parameters, in that iterative selection continues until it is achieved a convergence and that of the minimum acceptance difference function comprises the modified set of event and convergence parameters. The example event parameters include a combination of amplitude, phase, bandwidth and time, or any of the above individually. According to another modality and example, the determination of the cross-match function comprises: the interactive selection of the modifications of the event parameters to the second set of event parameters, the application of the modifications of the event parameters to the second set of event parameters, where a modified set of event parameters is defined, the comparison of modified set of event parameters with the first set of event parameters and providing a minimum acceptance difference, wherein said interactive selection towards the Mixed comparison field of said comparison designates a difference between the first set of event parameters and the modified set of event parameters below the minimum acceptance difference. According to yet another example embodiment, said provision of a minimum acceptance difference portion comprises: providing a plot difference with windows between a time window and a first trace of the first seismic survey data set and a time window of a second trace of the second seismic study data set, where the second trace includes reflection events corresponding to reflection events in the first trace and where the time window of the second trace is substantially equal to the time window of the second trace. first stroke, and provide a relation of a difference of strokes with windows with respect to the time window of the first stroke and choose that the minimum difference of acceptance is less than the ratio. The time windows both the non-changing and the changing portions of the deposit have similar spectral characteristics. For example, if the reservoir data have a dominant frequency of 30 Hz, the time window of an unchanging portion of the study having a dominant frequency as close to 30 Hz as possible should be taken. Likewise, the phase changes in the deposit and the unchanging portion should be as close as possible. It is preferred, however, to err on the side of the wider time windows. For example, if the dominant frequency of the deposit is 30 Hz, a window that is 35 Hz to 25 Hz is considered preferable. It is believed to remove bandwidth error of less than about 25% in the frequency bandwidth produces adequate results The best results should be seen with the bandwidth error being below 10%. In another embodiment, said production of a minimum acceptance difference function comprises: providing a plot difference with windows between a time window of the square of the first trace of the first seismic survey data set and a time window of the square of the second trace of the second set of seismic survey data, where the second trace includes reflection events corresponding to the reflection events in the first trace and that of the time window of the second trace is substantially equal to the time window of the first trace , and to provide a difference relationship of traces with windows with respect to the time window of the square of the first trace, choosing that the minimum difference of acceptance is smaller than the relation. In yet another embodiment, said application of the cross-match function or a third reflection event in the second data set comprises the composition between the cross-match function and the third reflection event in the second data set, said first set The data comprises a trace of a seismic receiver. Alternatively, said first data set and said second data set comprises an approximate set of traces of a set of seismic receivers or CMP data ("common midpoint"). In still another embodiment, said first data set and said second data set comprises the download data. In additional alternatives, said first data set and said second data set comprises data of previous stacking or migrated data. In many embodiments, said first data set and said second data set are subjected to equivalent pre-stack processes. For example, in addition to the deconvolution described above, in some embodiments the first data set uses the same process of eliminating interrelation relationships as the second data set, the same processing steps of noise attenuation as the second set of data. data and the same multiple attenuation processing as the second data set. Furthermore, in many embodiments, the same DMO operator is used in the first data set with the second data set and migration is conducted on the first data set with the same velocity field as the migration in the second data set . Finally, according to another aspect of the invention, a method is provided for performing the signal processing of the seismic or lapse, the method comprising: a) performing a set of processing steps in the first study; perform a set of processing steps in the second study, where the set of processing steps is dependent on a set of seismic signal parameters, choose at least one of a set of parameters by an independent selection process of the data of both study; and b) applying at least one of the parameter sets in at least one of the set of processing steps, both the first study and the second study.

DESCRIPTION OF THE DRAWINGS For a more complete understanding of the present invention and for greater advantages thereof, reference is made to the following description of the exemplary embodiments of the invention, considered in conjunction with the accompanying drawings, in which: FIG. It is a graph of a first record of a study taken for the first time, without deconvolution. Figure IB is a graph of a first record of figure A with deconvolution, where the deconvolution operator was designated dependent on the data of the first record. Figure IC is a graph of the data not processed in Figure IA applying deconvolution, where the deconvolution operator is designated, dependent on both the data in Figure IA and another study. Figure 2A is a graph of a second record of a second study taken from the same geographical area as that of Figure IA, but on a different occasion, without deconvolution. Figure 2B is a graph of a second record in Figure 2A, with deconvolution, where the deconvolution operator is designated dependent on the data of the second record. Figure 2C is a graph of the second record in Figure 2A, with deconvolution, where the deconvolution operator is designated dependent on the data dealing with registration of Figure IA with the record of Figure 2A. Figure 3A is a graph of the difference between Figures IA and 2C.

Figure 3B is a graph of the difference between Figures IB and 2B. Figure 3C is a graph of the difference between the figures IC and 2C. It is to be noted, however, that the appended drawings illustrate only typical embodiments of this invention and should therefore not be considered limiting of its scope, as the invention may admit other equally effective modalities.

DESCRIPTION OF THE EXAMPLE MODALITIES OF THE INVENTION As used herein, "cross-matching" is the general term in the industry for compensation filtering, amplitude and static correction corrections necessary for seismic lapse inspection. In essence, the operator or the small wave operators are estimated to configure and compensate the reflection data from one study to another. Typically, operators are designed on static reflectors that exclude deposits where significant changes occur in the fluid (pore) environment. Theoretically, the difference between the two data volumes after cross-matching must be zero anywhere (inside and outside the design window, except where there are changes in the deposit.) All static or unrelated events must be subtracted. the deposit, leaving only changes in the dynamic events (ie, replacement of fluid within the pores) It is determined if a change in the state of the reservoir for deposit lithology and the acoustic properties of the pore filters is seismically detectable ( which are dependent on the temperature and pressure), the type of production and the recovery process, the resolution (spatial and temporal) of the data, as well as the repeatability of the seismic system In certain tanks with gas drives, the injection Gas can reduce the acoustic integration (oil being displaced by gas) enough to induce a bright spot. for example, Fulp and Greaves, GEOPHYSICS, 1987, incorporated herein by reference, with respect to a bright spot generated by a flood of fire. In others, the injection of gas or the injection of water (water mechanism) can have the opposite reaction and generate an opaque spot. If vapor flooding is involved, the descent or lowering of speed is often used to indicate the vapor fronts in order to determine to what degree the recovery has progressed. For example, see Lumley, et al., SEG Expanded Abstracts, 1995, incorporated herein by reference. As mentioned above, the petrophysical and seismic interpretation is required to understand what attribute is significant and when change could be observed.

Processing The rudimentary steps for data processing towards the objective of obtaining preserved data of relative amplitude, especially correct ones, include NMO / BMD and the previous stacking migration (zero effect) for the following examples. Other appropriate pre-stacking procedures, including data comprehension and sampling techniques that can be used to reduce computing overhead and preserve the quality of pre-stacking processing that will occur to them, will occur to those skilled in the art. to those skilled in the art. After such processing, the data conditioning on the relative amplitude provides data of migrated CMPs of previous stacking on which cross-matching and data differentiation measurements are made. Depending on the production deposit, according to alternative embodiments of the invention, CMPs, attributes of AVO and / or migrated stacks of each study are cross-matched and subtracted (difference) to observe the fluid displacements between the pores. It is recommended that each study be processed as closely as possible (which is usually done in the processing of lapses). Prior to the present invention, any processed data that existed in the workstations (with visible pre-conditioning algorithms and procedures) was used and, if attempted, there will be illumination of the resulting variations of the data by means of cross-matching. Better cross-matches and more significant seismic differences will be obtained than processing existing data with the same methodology and software as recent seismic inspection studies. This being the case, it is particularly important to eliminate the identification signal, deconvolute, attenuate noise and multiples with the same parameterization, from study to study, just as it is important to use the same DMO operator and look at the data with the same velocity field, for obtain important measurements of lapses. Therefore, according to an aspect of the invention, a method is provided for performing the signal processing in a seismic study of lapses, where a set of processing steps is performed on the first study, the set of Processing steps on the second study and the set of processing steps is dependent on the set of parameters of seismic signals, the method comprising: (a) choosing at least one of the set of parameters through the selection process dependent on the data from both studies. AND (b) apply at least one of the sets of the parameters in at least one of the set of processing steps as well as the first study and the second study. It has been found that the improved deconvolution processing of the two studies results from the design of a deconvolution operator dependent on the multiple data sets. This deconvolution operator is then applied to both data sets. The result of such a design, although it may not be optimal for any particular set, is nonetheless optimal for purposes of comparison of lapses. According to an exemplary embodiment of this aspect of the invention, a method of deconvolution of multiple sets of seismic data from the same geographic area including the method is provided: designate a data-dependent deconvolution operator of at least two of Seismic data sets, where at least two sets of seismic data are recorded at different times; apply the deconvolution operator in a deconvolution process in both of at least two sets of data; and continue to conduct lapse processing to form a difference record. In some cases, there are at least three sets of seismic data, where each of the three sets at least represents records of known studies in different types and in the design of the deconvolution operator is dependent on the data of each of them. the sets of studies. In some of those cases, an additional step is provided which comprises designating a second deconvolution operator dependent on the data of the first and third data set, with the second deconvolution operator applying to the first data set and the third data set. The first data set may have been taken before or after the second data set with equal effectiveness. According to the most specific example mode, the design of a first deconvolution operator comprises: averaging a power spectrum for a first and a second set. In some circumstances, obtaining averages for each subsequent study will be prohibitively expensive. Therefore, according to a still further embodiment of the present invention, multiple studies are handled by storing the averages obtained from the first study, so that when a subsequent study is made, the autocorrelations of the first one do not have to be recalculated. study. According to such modality, designate the first deconvolution operator comprises: obtaining the average of a power spectrum for a first group of strokes of the first set; inversely transform the average to a representation in the time domain of the average for power spectrum of the first group; store the representation in the domain of a time; obtain the average of a spectrum of powers for a second group of strokes of the second set; inversely transform the average to a representation in the time domain of the average power spectrum for the second group; obtain the average of the representations of the time domain of the averages for the spectrum of powers of the first and the second group; and designate a deconvolution operator for the first and the second study from the average of the representations in the time domain of the averages for the power spectrum of the first and second group. Referring now to Figure IA, a specific example is discussed. Figure A shows a graph of a first non-processor stacking (no BMD, migration, other noise suppression, only spherical divergence and geometry correction) of a study taken in a first time, without deconvolution. Figure IB is a graph of the first record of figure A with deconvolution, where the deconvolution operator was designated before stacking depending on the data of having unprocessed stacking. The specific parameters were: Sample speed: 2 milliseconds. Operator duration: 140 milliseconds. Prediction space: tip formation, 2 milliseconds. Analysis window: Branch dependent: for nearby branches, 300-5000 milliseconds for far branches 3600-5000 milliseconds. White sound added: 0.5% Deconvolution application window: 0-600 milliseconds.

Figure 2A is a graph of a second unprocessed stack of a second study taken from the same geographic area of the same figure IA, for the different time of deconvolution, and Figure 2B is a graph of a second record of the figure 2A, without deconvolution, where the deconvolution operator was designated before the stacking dependent on the data of the second unprocessed stack. Here, the specific parameters are as before. Figure 3A is a graph of the difference of Figures IA and 2A. Figure 3B is a graph of the difference between Figures IB and 2B. Ideally, there would be nothing but noise or changes in the structure that will appear in the difference chart of Figure 3A, if the two studies were otherwise identical and there had been no change in the status of the deposit. However, it can be seen that the unmodified structure appears in the 3B difference chart. Figure IC is a graph that in the first stack of Figure IA, with deconvolution of the previous stack, where the deconvolution operator was assigned dependent on the data of the first record and the second stack. Figure 2C is a graph of the second stack of Figure 2A, with deconvolution and previous stacking, where the deconvolution operator was designated dependent on the data of the first record and the second record. The specific parameters were, both for Figure IC and 2C, of at least that for Figure IA.

Figure 3C is a graph of a difference in three graphs of figures IC and 2C. Note the lack of unmodified structure. In the designation of the deconvolution filter used in some modalities, the first data compute (ie, "bases") is processed where the average of the autocorrelation of strokes works within each group (ie download, receiver, CMP, etc.). ) is calculated in the fx domain, to estimate the power spectrum: i = 0 N P (i, k) = (1 / N) or P (i, k), i = 1, ... Q k = i where: N is the number of traces in the x direction, Q is the niquist frequency index, P (i, k) is the power spectrum estimate, the frequency index i and the number of traces k, and i is an index of lapses. Next, an inverse FFT is performed and the autocorrelation function traces are saved for each group. The same estimation of the power spectra is performed on a set of data taken at different time (ie a set of "inspection" data) to calculate: i = 1 P (i, k) where i = 1 is the First study of inspection of lapses. Next, all the average autocorrelation functions of lapses are averaged to generate authorized autocorrelation functions according to the following formula: A BM 77, M i = 1, ... M where M is the number of lapses. The resulting autocorrelation traces are used to design the deconvolution filters that are applied to all lapse studies. It will occur to those exposed in the technique other processes to designate deconvolution filters (for example, deconvolution of single window, deconvolution to time variant, deconvolution compatible with the surface and deconvolution with the common domain) (discharge, receiver, midpoint, branch), which are dependent data sets, which does not deviate from the spirit of the present invention. Further, as the deconvolution operation described above, there are other processing steps that benefit according to one aspect of the present invention by designing the process parameters dependent on the data of each set, rather than the data of a set. Some examples include: cross-matching (discussed later in more detail), f-k filtering, tau-p and Radon domain filtering, static calculations and multiple deletion in FK domains of tau-p, and Radon. There are always variations in real acquisition, no matter that the operations have also been planned or executed. Y, in a real world, the processing would also be identical, because the acquisition would be reproduced the basic study to the inspection study. But that is not realistic. Thus, acquisition variations occur between the studies with many systems that are useful with the present invention (for example, fixed installation systems, heading selector systems and undulatory luminosity systems). These variations could be caused by echoes in the form of a pen on the screen, cable housing, discharge azimuth, source variation, sea state conditions, etc., which may result in processing variations between the studies.

Cross-matching In accordance with the embodiment of the example of the invention, there are "correction" counter-elements that use the cross-match fusion. They are corrections of the regulation of time, balance of energy by mean square value, bandwidth banning and phase compensation. Each element essentially forms a transfer or impulse response function between the two studies or, alternatively, between the two traces in each study. The cross-linked equation (t £ Q) is calculated as follows: fcXEQ = t * f < scorr 'rmscorrstncorr ^ Pcorr "' (D where t is the entry trace, * with another convolution and S_ corr rmscorr mcorr and Pcorr "are elements of cross-matching (in the impulse response f) corresponding to time, amplitude, magnitude and phase, respectively. alternatives of the invention, the stroke response functions are calculated by stroke, in a manner consistent with the surface or globally from the baseline study (bases) and monitor study (repetition) In accordance with the present example embodiment, a monitor study is transformed (equal to cross-wise) to be seen as a baseline study and the cross-equalization pulse functions are designated by horizon-specific time gates excluding the area of the deposit where the change is anticipated. The effects of each component will now be explained by means of models and data from simple examples from a three-dimensional project of lapses.

Elements of crossed equalization Corrections of time (s orr-) In addition to the turbulent sea to a storm and the common conditions that can cause the formation of wavy luminosities in the form of a feather, changes of tide and temperature are also in play. These changes can not be significant when recording, processing or interpreting seismic data for a single 3D study. However, when several 3D studies are involved for lapse inspections, there are potential variations in the state of the sea between the studies. Seasonal temperature and salinity conditions and shorter periods (related to storms) may also cause differences between studies between time reflectors. The degree of variation is a function of the severity of the change between the temperature and severity profiles and the depth of these variations in the water column. For example, studies done in the East China Sea show that during a large scale of a 10-year period they are notable for seasonal variations and are sufficient to cause a temporary occurrence between the two-millisecond (TWT) studies between reference reflectors in the extremes of the season (which are acquired in January and June, itself), where the water depths are equal to or greater than 100 meters.

As for tidal changes, an extreme example would be to download seismic data in the Bay of Fundy, between the Atlantic provisions of New Brunswick and Nova Scotia. Here, the tidal changes are approximately 15 m which is equivalent to 20 ms of time and directional (assuming a typical water velocity of 1500 m / second). If a study was obtained at low tide and another at high tide (assuming very short study periods), the seismic differentiations of the data sets without time correction would not provide satisfactory results. In general, errors increase with increasing delay and small areas of wider bandwidth obtain greater error differences.

Equilibrium of energy of quadratic mean value (rmscorr-l In a perfect experiment of seismic inspection of lapses, the same seismic personnel would acquire the monitor study with the same exact equipment, in the exact conditions (state of the sea) with which the Original baseline study However, if the acquisition differences occur, even the best controlled conditions, for example, it is important to scale the data to equivalent levels of mean square value, especially if the same (medium quadratic value energy) It is significantly altered between studies.

In one example, of approximately 100 CMPs in a 3D data set of lapses acquired by undulating luminosity vessels approximately 18 months apart, the acquisition geometry and instrumentation were almost identical, except that a study was acquired in the Summer months and the other in the winter. The analysis of the average amplitude of the seismic data within continuous windows of 200 milliseconds shows that there is an amplitude deviation between the baseline study and the monitor study in the whole profile. After the application of a smoothed cross-matching element, with amplitude gradation, within each time gate of 200 milliseconds, it was found that the availabilities between the base and monitor studies are smaller after cross-matching. Undoubtedly, variations such as these need to be corrected before the sections / studies of seismic differences of lapses can be interpreted.

Bandwidth Normalization (m rr-.} _ Equalization of bandwidths is also part of the seismic differentiation.If two seismic volumes have to be subtracted, it would be best to do so with each volume having equivalent spectra. Cross-matching (if strong enough) must correct one spectrum to compensate for another, but if cross-matching is not optimized, discrepancies appear.The largest errors occur for those with large bandwidths.The tests indicate residual reflector ( ie energy remaining in the difference section / study that would ideally be eliminated) ranges from 7% to 40% of the input amplitude for the center frequency differences of 2.5 to 12.5 Hz, respectively (for Ricker small waves with frequencies of center separated incrementally by 2.5 Hz subtracted from a small Ricker wave of 30 Hz from the baseline). or 12.5 Hz is reasonable for most typical 3D studies, but a larger difference in bandwidth occurs in some cases and, in those cases, the residual reflector energy is larger. Therefore, according to modalities that process these studies, cross-matching favors bandwidth differences to reduce these potential errors of seismic differentiation.

Phase Compensation (p nrrs) Finally, errors are associated with phase inequalities. For example, for a small Ricker wave of 30 Hz, rotated phase by 5, 10, 15, 30, 60 and 90 ° constant from phase 0, the residual reflector energy caused by the phase differences can be as large as 20% of the input for phase inequalities as small as 15 °. In addition, the human eye may have difficulty in seeing phase differences of 15 ° or less, so that the cross-equalization operator is significant to correct small phase inequalities as well as large ones. Phase difference errors are not affected by bandwidth differences: it has been discovered that residual reflector errors of 20% for small waves but phase by 15 ° can occur.

Seismic differentiation The ultimate goal in seismic lapse monitors is to see the changes that make up the movement of fluid between pores or the absence of fluid movement between pores between calendar dates. This is commonly obtained by subtracting a monitor study from the baseline study after rectifying the data by cross-matching. Differentiated results with reduced amounts of residual reflector energy will have a higher probability of identifying fluid movement than those with higher reflector energy. To characterize the effectiveness of cross-matching, we show several difference graphs with several component cross-match applications using an experiment of lapses in the North Sea. In figures 4 and 5, respectively, there are migrated in-line stacks of the monitor advance study that traverses an e without deposits, and a prominent series of strong static reflectors is visible between 2000 and 2600 milliseconds. For this example, a design window of 1900 milliseconds to 2700 milliseconds is used. An operator of 800 milliseconds was created and applied from the design window and a bandpass filter of 3 / 8-35 / 55 Hz was applied to both datasets before cross-matching. Since these are static reflectors (that is, they are reflection events that represent an unchanged portion of the geologic structure in or near the deposit), the subtraction of the base study of the monitor should ideally result in the residual reflector energy very small, since both studies are acquired identically. However, the difference section obtained without cross-matching showed that the subtraction did not. The energy of the residual reflector is significant. With these large amounts of excess residual reflector energy in that section it would be difficult to identify any fluid movement (elsewhere) in portions of the reservoir volume. The comparison of a difference section obtained without cross-matching to one in which only the amplitude spectrum was cross-matched, one in which only the phase spectrum was cross-matched and one of which cross-matched both, shows that the Phase corrections are a fundamental component of the total cross-matching operation, which reiterates the results of the development of synthetic models. Previous processing is performed on a massively parallel processing platform (eg IBM SP2, Intel Paragon) and software compatible with such technical support (eg PGS Tensor IncXs CUBE MANAGER ™ software).

Suggested readings Two published backgrounds on seismic lapse monitoring are Greaves and Fulp (Geophysics, 1987), incorporated herein by reference, who monitor a fire flood in western Texas, and Johnstad, Uden and Dunlop (First Break, 1993 ), also incorporated herein by reference, which monitor gas injection from the Oseberg field in the new sector of the North Sea. Another appropriate document involving steam flooding in the Duri field of Indonesia was presented by Lumley and others at the SEG assembly of 1995 in Houston, which may be through the SEG. The foregoing description will be by way of example only and will occur to those skilled in the art to other embodiments without departing from the scope of the invention as defined by the following claims and equivalents thereof.

Claims (45)

NOVELTY OF THE INVENTION CLAIMS

1. - A method for comparing multiple data sets of a seismic study in a deposit, characterized in that a first seismic study data set is taken from a first time and a second set of seismic study data is taken in a second time, to detect changes in the deposit in the first time and second time, including the method: provide a first reflection event in the first set of seismic study data that have a corresponding second reflection event in the second set of seismic study data, in where the first reflection event and the second reflection event represent an unchanged portion of the geological structure in or near the deposit and in the first reflection event it is represented by a first set of event parameters the second reflection event is represented by a second set of event parameters; provide a minimum acceptance difference function between the first set of event parameters and the second set of event parameters; determine the cross-match function to apply to the second set of event parameters, the cross-match function being characterized because, with the application of the cross-match portion to the second set of event parameters, whereby a set of cross-matched event parameters, and the difference of the first set of event parameters and the set of cross-matched event parameters is below the minimum difference function; apply the cross-equalization function to the third reflection event, the third reflection event being related to the second data set, where the third cross-matched reflection event is defined, where the third reflection event has a corresponding fourth event of reflection reflection in the first data set, and in the third and fourth reflection event represent a changing portion of the deposit; compare the third reflection event equating crosswise with the fourth reflection event substituting the third reflection event equated crosswise with the fourth reflection event.

2. A method according to claim 1, further characterized in that said provision of said minimum acceptance difference portion comprises: the iterative selection of event parameter modifications to the second set of event parameters, the application of the modifications of event parameter to the second set of event parameters, where a modified set of event parameters is defined, the comparison of the modified set of event parameters with the first set of event parameters, wherein said iterative selection continues until it is reaches the convergence, and where the minimum acceptance difference function comprises the modified set of event parameters in the convergence.

3. A method according to claim 2, further characterized in that one of said event parameters comprises latitude.

4. A method according to claim 2, further characterized in that one of said event parameters comprises the phase.

5. A method according to claim 2, further characterized in that one of said event parameters comprises the bandwidth.

6. A method according to claim 2, further characterized in that one of said event parameters comprises time.

7. - A method according to claim 2, further characterized in that one of said event parameters comprises the amplitude, phase, bandwidth and time.

8. A method according to claim 1, further characterized in that said determination of a cross-match function comprises: the iterative selection of the event parameter modifications to a second set of event parameters; the application of event parameter modifications to the second set of event parameters, where a modified set of event parameters is defined; the conversion of modified set of event parameters to the first set of event parameters; and the provision of a minimum acceptance difference; wherein said iterative selection continues until a comparison result of said comparison designates a difference between the first set of event parameters and the modified set of event parameters below the minimum acceptance difference.

9. A method according to claim 1, further characterized in that said provision of a minimum acceptance difference function comprises: providing a difference of traces with window between a time window and a first trace of the first set of study data seismic and a time window of a second trace of a second set of seismic survey data; wherein the second stroke includes reflection events corresponding to reflection events in the first stroke, and wherein the time window of the second stroke is substantially the same as the time window of the first stroke; and provide a plot difference relationship with window after the time window of the first stroke; choose that the minimum acceptance difference is less than the ratio.

10. A method according to claim 9, further characterized in that the difference of the bandwidth of the time window of the first trace and the time window of the second trace is less than about 25%.

11. A method according to claim 10, further characterized in that the difference of the width of the time window of the second trace the band difference of the time window of the first trace is less than about 10%.

12. A method according to claim 1, further characterized in that said provision of a function of minimum acceptance difference comprises: providing a difference of traces with window between a time window of the square and a first trace of the first data set of seismic study and a time window of the square of a second trace of a second seismic study data set; wherein the second stroke includes reflection events corresponding to reflection events in the first stroke, and wherein the time window of the second stroke is substantially the same as the time window of the first stroke; and provide a plot difference relationship with window after the time window of the first stroke; choose that the minimum acceptance difference is less than the ratio.

13. A method according to claim 7, further characterized in that the time window has a duration of at least about two reflection events.

14. A method according to claim 7, further characterized in that the time window has a duration of at least about five reflection events.

15. A method according to claim 1, further characterized in that said application of the cross-equalization function to the third reflection event in the second data set comprises the convolution between the cross-equalization function and the third reflection event in the second data set.

16. A method according to claim 1, further characterized in that said first data set comprises a trace of a seismic receiver.

17. A method according to claim 1, further characterized in that said first data set and said second data set comprises a summed set of traces of a set of seismic receivers.

18. A method according to claim 1, further characterized in that said first data set and said second data set comprises a summed set of strokes of a set of hole hole receivers.

19. A method according to claim 1, further characterized in that said first data set and said second data set comprises pre-stacking data.

20. A method according to claim 19, further characterized in that said pre-stacking data comprises CMP data.

21. A method according to claim 19, further characterized in that said pre-stacking data comprises discharge data.

22. - A method according to claim 19, further characterized in that said pre-stacking data comprises migrated data.

23. A method according to claim 1, further characterized in that said first data set and said second set of data are subjected to equivalent processes of pre-stacking.

24.- A method according to claim 1, further characterized in that said first data set is subjected to the same method of designation of indication signal of the second data set.

25. A method according to claim 1, further characterized in that the first data set is subjected to the same deconvolution process as the second data set.

26. A method according to claim 1, further characterized in that the first data set is subjected to the same noise alternation processing steps as the second data set.

27. A method according to claim 1, further characterized in that the first data set is subjected to the same multiple attenuation processing as the second data set.

28. A method according to claim 1, further characterized in that the same DMO operator is used in the first data set with the second data set.

29. - A method according to claim 1, further characterized in that migration is conducted in the first data set with the same velocity field as the migration in the second data set.

30. A method according to claim 1, further characterized in that migration is conducted in the first data set with the same migration operator as the migration in the second data set.

31. A method according to claim 1, further characterized in that the filtration is conducted in the first data set with the same felt as the filtration in the second data set.

32.- A deconvolution method of multiple sets of seismic data from the same geographic area is provided, including the method: designate a deconvolution operator depending on the data of at least two of the seismic data sets, where they were recorded at least two sets of seismic data at different times; apply the deconvolution operator in a deconvolution process to both of the at least two sets of data; and continue to conduct lapse processing to form a difference record.

33.- A method according to claim 32, further characterized in that at least two of the data sets comprise at least three sets of seismic data, because each of the at least three sets represents records of known studies in Different times and the designation of a deconvolution operator is dependent on the data in each of the study sets.

34.- A method according to claim 32, further characterized in that the at least two of the data sets comprise at least three sets of seismic data; because each of the three sets represents records of known studies at different times; and in that the designation comprises designating a first deconvolution operator dependent on the data of a first data set and a second set of data applied to a first deconvolution operator to the first data set and to the second data set; and further comprising: designating a second decomposition operator dependent on the data of the first and a third data set, with the second decomposition operator applying to the first data set and the third data set.

35. A method according to claim 34, further characterized in that the first data set represents records of a study made before a study represented by the second and third data set.

36.- A method according to claim 34, further characterized in that the designation of the first decomposition operator comprises: obtaining the average of a power spectrum for a first group of strokes of the first set; transform, inversely, on average, a representation in the time domain of the average by power spectrum of the first group; store the representation in the time domain; obtain the average of a power spectrum by a second set of strokes of the second set; inversely transform the average to a representation in the time domain of an average power spectrum for the second group; obtain the average of the representations in the time domain of the averages for power spectrum of the first and the second group; and to designate a decomposition operator for the first and the second study of the average of the representations in the time domain of the averages of the power spectrum of the first and the second group.

37. A method according to claim 32, further characterized in that the designation of a decomposition operator comprises: obtaining the average of a power spectrum for a first group of strokes of the first set; obtain the average of a spectrum of powers for a second group of strokes of the second set; obtain the average of the averages of the power spectrum of a first and a second group; and designate a decomposition operator of the first and second study of obtaining averages of the power spectra of the first and second group.

38.- A method to carry out signal processing of seismic study of ties, including the method: perform a set of processing steps of the first study; perform the set of processing steps in the second study; further characterize because the set of processing steps is dependent on a set of parameters of seismic signals; choose at least one in the set of parameters by a selection process dependent on the data and both studies; and applying at least one of the parameter sets in at least one of the set of processing steps in both the first study and the second study.

39.- A method according to claim 38, further characterized in that in at least one of the parameter sets comprises a decomposition operator.

40. A method according to claim 39, further comprising: designating a data-dependent decomposition operator of at least two of the seismic data sets, further characterized in that the at least two sets of seismic data are recorded at different times; apply the depraved decomposition in a decomposition process to both of at least two data sets; and form a difference register dependent on the application.

41.- A method according to claim 40, further characterized in that at least two of the data sets comprise at least three sets of seismic data, because each of the at least three sets represents records of known studies in different times and because designation of a decomposition operator is dependent on the data of each of the study sets.

42. A method according to claim 42, further characterized in that the at least two of the data sets comprise at least three seismic data sets; because each of the at least three sets represents records of known studies at different times; and in that the designation comprises designating a first decomposition operator dependent on the data of a first data set and a second data set, applying them in first decomposition operator to the first data set and to the second data set; and further comprising: designating a second decomposition operator dependent on the data of the first and a third data set, applied in the second decomposition operator to the first data set and the third data set.

43.- A method according to claim 42, further characterized in that the first data set represents records of a study made before a study represented by the second or the third data set.

44. A method according to claim 42, further characterized in that the designation of the first decomposition operator comprises: obtaining the average of a power spectrum for a first group of strokes of the first set; transform, inversely, on average, a representation in the time domain of the average by power spectrum of the first group; store the representation in the time domain; obtain the average of a power spectrum by a second set of strokes of the second set; inversely transform the average to a representation in the time domain of an average power spectrum for the second group; obtain the average of the representations in the time domain of the averages for power spectrum of the first and the second group; and to designate a decomposition operator for the first and the second study of the average of the representations in the time domain of the averages of the power spectrum of the first and the second group.

45. A method according to claim 40, further characterized in that the designation of a decomposition operator comprises: obtaining the average of a power spectrum for a first group of strokes of the first set; obtain the average of a spectrum of powers for a second group of strokes of the second set; obtain the average of the averages of the power spectrum of a first and a second group; and designate a decomposition operator of the first and second study of obtaining averages of the power spectra of the first and second group.