EP1196791A1

EP1196791A1 - Method for processing seismic data

Info

Publication number: EP1196791A1
Application number: EP00954275A
Authority: EP
Inventors: Henning Trappe; Carsten Hellmich; Marc FÖLL
Original assignee: Individual
Current assignee: Individual
Priority date: 1999-07-19
Filing date: 2000-06-15
Publication date: 2002-04-17
Also published as: US6754587B1; DE19933717C1; WO2001006277A1

Abstract

The invention relates to a method for processing a seismic 3-D measurement data set, consisting of a multitude of seismic traces, each having a sequence of amplitude values or data points provided with acoustic impedances. Said method consists of the following: selecting a reference section at a predetermined location and depth which comprises neighbouring trace sections of several seismic traces; determining the similarity between the selected reference section and localised sections of seismic data from the measurement data set; creating a volume of data which corresponds to the measurement data record, using the similarity value which has been determined and allocated to each data point as the attribute. During processing of the seismic data, the inventive method enables the sub-surface image to be classified by an absolute comparison of the measurement data with a reference sample section as the means of interpretation.

Description

DESCRIPTION

Methods for seismic data processing

The invention relates to a method for processing a seismic 3-D measurement data set consisting of a multiplicity of seismic traces, each of which has a series of data points occupied with amplitude values.

Seismic exploration methods are used worldwide to obtain information about the spread of geological structures underground in addition to information from sunk wells. Information from seismic data can often be used to dispense with further costly exploration drilling or to limit the number to a minimum.

Sensors (geophones / hydrophones) are used for seismic exploration of the subsurface, which receive sound waves in a row (2D seismic). These waves are excited by a seismic source, such as explosive charge, vibration excitation or air pulses (airguns), and z. T. reflected back to the surface. There they are registered by the sensors and recorded in the form of a time series. This time series represents the incoming seismic energy in the form of amplitude fluctuations. It is stored digitally and consists of evenly arranged data points (samples), which are characterized by the time and the associated amplitude value. A Such a time series is also referred to as a seismic trace. The series of measurements moves over the area to be examined, so that a 2D seismic profile is recorded with this arrangement.

The subsequent processing (processing) has a noise suppression z. B. to the target by stacking or filtering. The resulting results are vertical profiles in which amplitudes and transit times as well as attributes derived from amplitudes are shown, which serve as the basis for the further geological evaluation. The geological layers can be tracked on a profile by the lateral alignment of the amplitudes.

If the data is recorded not only along a line but in an areal grid, a three-dimensional data volume results. In the case of the SD volume, an arbitrary point in the underground, e.g. B. by Cartesian coordinates

Amplitude value assigned. The vertical direction is measured in time (sound propagation time).

During further data processing, the measurement data are corrected, filtered and, if necessary, converted. The result is a seismic volume in the form of an SD data set, which shows the physical properties of the examined subsurface in a seismic image.

Any cuts, e.g. B. vertical profiles and horizontal maps can be extracted from different depths, which are then interpreted by geophysicists and geologists. Because this interpretation of the won Seismic images essentially include an optical correlation, attempts have been made to automate this subjective evaluation, which is dependent on one or more interpreters.

A method for seismic data processing is known from WO 96/18915, in which a 3D seismic volume is divided into a plurality of vertically stacked and spaced horizontal slices, at least one slice being divided into a plurality of cells. Each cell has at least 3 track sections, the first and second track sections being arranged in a vertical plane in the profile direction (inline) and the third track section with the first track section in a vertical plane being essentially perpendicular to the profile direction (crossline). Then a cross-correlation is carried out between two track sections in each of the two vertical planes, which result in inline and crossline values that are dependent on the slice inclination. The

Combining these values in a cell gives a coherency value for the cell, which is assigned to a data point of the cell. The end result is again a 3D data volume from which any section can be extracted and displayed.

EP 0 832 442 A1 discloses a method and a device for seismic data processing by means of coherence characteristics, in which a seismic volume is divided into horizontal disks in a manner similar to the above-mentioned document and these are in turn divided into cells. In the simplest case, these cells are cube-shaped. From the at least two track sections located in a cell, a correlation matrix formed as the sum of the differences between the inner and outer product of the value tuple from the track sections. The quotient of the largest eigenvalue of the matrix and the sum of all eigenvalues is then calculated as a measure of the coherence. The result is a 3D volume consisting of coherence values.

Furthermore, EP 0 796 442 AI relates to a method and a system for seismic data processing, in which a coherence method based on a semblance analysis is carried out. Similar to the two aforementioned methods, a seismic data volume is divided into at least one horizontal time slice and this into a large number of three-dimensional ones

Analysis cells divided, each cell having two predetermined, mutually perpendicular lateral directions and at least five seismic track sections arranged next to one another therein. A semblance value of the track sections located in the cell is assigned to the corresponding data point in the respective cell. The semblance is a known measure for the correspondence of seismic track sections. The incidence and the direction of incidence of the analyzed reflector are determined by the best coherence by searching different layer inclinations and directions. In addition to the semblance value, the calculated inclination data are then displayed for each cell.

Furthermore, EP 626 594 AI discloses a method for determining the physical properties of the subsurface, in which a comparison of a seismic reference track recorded at a drilling location with one obtained synthetically from log data of a drilling is known Reference track is carried out. Modified synthetic seismograms are then generated, which are compared with the other seismic traces. However, only two track segments, namely a track segment of a seismic track and a track segment of a synthetically generated seismic track, are compared with each other. Lateral environments are therefore not taken into account.

Also known from C. Hellmich, H. Trappe and J. Fertig with the title "Image Processing of Seismic Attributes and Geostatistics in Upper Carboniferous" from DGMK Conference Report (1996) an image processing method that enables a quantitative characterization of seismic representations and thus further interpretations of the Allows lithology. Different image processing filters are applied to amplitude maps and the fluctuations or the continuity of the amplitude values of the surrounding area are quantified. These filters represent 2D multi-track filters with which the local environment around a data point is evaluated. Operators used for this include Entropy and dispersion. Maps for interpretation can be created with all attributes. The variables "entropy" or "dispersion" are measures that quantify fluctuations or continuities of the amplitude in the local environment.

It should be emphasized that only relative comparisons in the local environment of a data point are considered in the aforementioned methods. Thus, for example, laterally continuous and slowly changing ambient conditions are not noticeable in the aforementioned evaluation methods. Also in the execution According to claim 19 of WO 96/18915, only a relative similarity, that is to say related to the individual cell, is initially determined. The similarity values calculated in this way are compared with similarly calculated similarity values from a volume with a known hydrocarbon deposit, and the drilling location in the newly examined volume is determined from the comparison of the coherence values with the coherence values of the volume of the known hydrocarbon deposit. Nevertheless, the comparison only includes coherence values determined relatively in a local environment (cell).

The object of the invention is therefore to provide a method for seismic data processing in which the data is classified over an entire volume of measurement data according to absolute criteria.

This object is achieved with a method according to claim 1.

If the geological conditions of the subsurface to be investigated at a location within the area covered by the seismic data volume are known, for example from borehole information, the similarity of the seismic signals in the entire measurement data volume to the signal at this location is determined by known geology. It is assumed that similar geological conditions produce a similar seismic signal in order to be able to use the similarity determination to transfer the geological conditions known at the drilling location to other areas or to find them there again. Essential to the invention is the comparison of the local section considered in each case with a predetermined reference section, which likewise consists of adjacent track sections of several seismic tracks. This creates an absolute reference to a reference pattern that, in addition to the temporal extent along a seismic trace (time series), also has a lateral extent.

Thus, the consideration of lateral changes in the pattern comparison based on the reference pattern can also provide probability statements for geological conditions in the lateral direction. This allows both lateral small-scale changes and based on the absolute comparison based on a

Reference patterns can also detect changes over long distances with a high degree of probability. Furthermore, it is also possible to detect laterally slowly changing structures based on the absolute comparison with the reference pattern by decreasing or increasing similarity.

The selection of a volume-shaped section thus has the advantage that, in addition to the vertical distribution of the amplitude information as a characteristic variable, the lateral change in the seismic signal for characterizing the background is also taken into account. It is scientifically demonstrated that based on the knowledge of the lateral change in geology, statements can be made regarding the thickness of sand bodies or the sedimentological environment. Motivated by these observations, the similarity of the local seismic data to the global reference is determined for the entire data volume. On A measure of this similarity is z. B. the dispersion of reference data and local data, but also an average-optimized semblance function on the combined reference data and the local data is used.

Overall, the data points of the seismic volume are classified according to absolute criteria.

If the size of the reference section and the local sections comprises 3 to 7 data points in each dimension direction, on the one hand there is a sufficient number of data points for data evaluation and on the other hand small-scale structures in the signal image can be recognized with the analysis. For example, hydrocarbon-bearing layers often have a vertical thickness in the seismic signal that is well below 10 samples. It is important here that a sufficient number of adjacent tracks are included in the sections considered in each case, in order to take the lateral characteristics of the environment into account in comparison. In order to be able to also detect small-scale changes laterally, a maximum of 10 data points should also be included in each lateral direction.

In the simplest case, the reference pattern section and the local sections are, in the simplest case, rectangular sections of the seismic data at the respective 3-D measurement data set

Location and depth. At the same time, further volume-like cutout standards for 3-D data are conceivable. If the local sections and / or the reference section are deformed according to a local preferred inclination and inclination direction, the sections used for the analysis are better adapted to the respective geological conditions.

The local preferred directions are determined, for example, in that before determining the similarity between the reference section and local sections, iterative determination of the similarity according to the inclination and inclination direction of adjacent track sections offset with respect to one another for the reference section and in each case for the local section that inclination and inclination direction which results in the greatest similarity of the track sections of the reference section and of the local section in each case.

Alternatively, the inclination and inclination direction can also be determined by searching for the inclination and inclination direction with the greatest similarity of the track sections belonging to the reference section when selecting the reference section, the determination of the similarity between the reference section and local sections then in each case that relative inclination between the reference section and the local section is determined which corresponds to the greatest similarity between the two sections.

In both alternative methods, in addition to the data volume with the similarity values found, a data volume with the determined inclination values and a further data volume with the determined inclination direction values are formed as a result. The reference section is preferably selected on a well with secured lithological information, so that the geological conditions secured by the well can be transferred with great similarity to corresponding areas of the examined data volume.

A reference section can be synthesized by folding a preselected acoustic impedance, e.g. B. from the relevant log, with a representative wavelet, if the seismic data quality at the well, z. B. due to near interference, in the quality is impaired. By creating a detailed model, e.g. B. using geostatistical methods, an arbitrarily complicated reference pattern can be formed. Seismic modeling techniques, such as B. seismic ray tracing, can create a seismic reference pattern. This opens up the possibility of special

Looking for situations, e.g. B. wedging of layers or fault zones, which can be important for the development of deposits. This makes it possible to create a reference sample catalog for 3-D structures. This catalog can be used to assign local seismic signal characteristics to geological parameters such as petrological properties, deposition conditions, tectonic features etc. This assignment can be carried out for any point in the entire data network. The investigation also allows the optimization of the data used. Due to the fact that the data points are classified according to the determined similarity values, an automatic assignment of certain subsurface areas of the examined data record to a specific geological structure can be achieved.

By comparing several reference sections, for example drill locations, with the local sections and thus calculating several similarity values for each data point, the significance of the geological conditions in the examined data record can be increased. With an appropriate classification, similar structures can be laterally assigned to certain geological conditions exposed in boreholes.

The method of the reference pattern described above is also to be used if, instead of the seismic data set, an acoustic volume, e.g. B. is used by a seismic inversion process.

The invention is described below by way of example with reference to the accompanying figures.

It shows:

1 schematically shows a 3-D data volume with a local section,

2 schematically shows a 3-D data volume with an inclined local data section, 3 shows a horizontal section along a layer boundary from seismic data processed according to the invention and

4 shows a horizontal section along a further layer boundary to the data according to FIG. 3.

1 schematically shows a 3-D data volume 1 which comprises a large number of seismic traces which are not explicitly shown. A cuboid section 2 is shown in the data volume 1, on which three time series in the form of seismic track sections 21, 22, 23 are arranged by way of example. The local data section 2 preferably has three to seven adjacent seismic traces in each lateral direction, for example 5 × 5 traces with a temporal length of likewise 5 data points (sample), which results in a sampling rate of 4 msec. a time slice of 20 msec. equivalent.

In FIG. 2 there is a schematic representation in the 3-D data volume 1 corresponding to FIG. 1

"Deformed" data section 2 ', likewise exemplarily shown with three time series in the form of seismic track sections 21, 22, 23, is shown. The deformation of the local data section 2 'reflects the preferred inclination 31 and direction 32 determined at this location and depth area. Corresponding to the preferred inclination 31 and direction 32, the data section shown in FIG. 1 is formed in a parallelepiped shape.

3 is the result of an inventive

Reference analysis for a geological horizon with constant lithology presented. For this, a Cut the slice out of the 3-D data set along this layer boundary. Starting from the bore a with secured lithological information, a cube-shaped reference pattern with 3 x 3 x 3 data points (sample) was chosen analogous to the section shown in FIG. 1.

Then the similarity of local data sections to this reference pattern section was calculated in the vertical environment of the horizon to be examined. The calculated similarity values are assigned as an attribute to the respective center of the local data section being considered and each local data point of interest, if necessary, is taken into account over the entire 3-D data set.

The map shows the greatest possible correspondence, characterized by the very high similarity values close to 1. The determined similarity values can be assigned in accordance with the gray scale scaling shown on the right in FIG. 3. In this example, a check could be carried out on the basis of a reference hole b, which demonstrated the same lithological features of the horizon, namely an anhydride. An exception is the northern part of the study area, shown in the upper left quarter of the map, which reflects the influences of a salt dome in the hanging area, which has had a negative impact on the seismic data quality. In addition to this disturbed area, linear fault zones can also be seen.

In contrast, FIG. 4 shows a less troubled lithology for the same examination area. This one The selected layer boundary is to be assigned to a sandstone storage horizon that is suitable for hydrocarbons. Similarity features were calculated on the basis of a reference pattern section derived from hole a, the magnitude of the similarity values corresponding to the gray scale shown on the right being significantly smaller than in FIG. 3. While, as expected, high similarity values are found in the vicinity of borehole a, there are differences to the eastern part of the study area shown on the map on the right. Hole b has encountered a dense sandstone in this area of lesser similarity, which is unsuitable as a storage horizon. It should be noted that some of the fault zones that can be seen in FIG. 3 can also be seen in the area of this layer boundary in FIG. 4.

1b

LIST OF REFERENCE NUMBERS

1 3-D measurement data set

2 local data section

2 'local data section for inclined lithology

21 seismic track section

22 seismic track section

23 seismic track section

31 Preferential tendency

32 Preferred inclination direction

a hole b hole

Gray Scale

Claims

Method for processing a seismic 3-D measurement data record, which consists of a large number of tracks, each of which is formed by a series of data points occupied with amplitude values or acoustic impedances, in the case of which a

Reference section, which corresponds to a predetermined location and depth and comprises adjacent track sections of several seismic tracks,

determines the similarity between the reference section and local sections of seismic data from the measurement data set and

- A data volume corresponding to the measurement data record is generated with the determined similarity values assigned to each data point as an attribute.

2. The method according to claim 1, characterized in that the size of the reference section and the local sections comprises 3 to 7 data points per dimension direction.

3. The method according to claim 1 or 2, characterized in that the local sections and / or the reference section are deformed in accordance with a respective local preferred inclination and direction of inclination.

4. The method according to claim 3, characterized in that prior to determining the similarity between the reference section and local sections by iteratively determining the similarity according to the inclination and inclination direction offset adjacent track sections for the reference section and in each case for the local section that inclination and inclination direction is sought, which results in the greatest similarity of the spur sections of the reference section and of the local section in each case.

5. The method according to claim 3, characterized in that in the selection of the reference section there is searched for the inclination and inclination direction with the greatest similarity of the track sections belonging to the reference section, the relative being then determined in each case when determining the similarity between the reference section and local sections Inclination between the reference section and local section is determined, which corresponds to the greatest similarity.

6. The method according to claim 3, 4 or 5, characterized in that in addition to the data volume with the similarity values, a data volume with the determined inclination values and a further data volume with the determined inclination direction values are formed.

7. The method according to any one of the preceding claims, characterized in that the reference section through a hole with secured lithological information.

A method according to claim 7, characterized in that the reference section is generated synthetically by folding a preselected 3-dimensional acoustic impedance distribution from the relevant borehole log with a representative wavelet.

9. The method according to claim 7, characterized in that the reference section is formed synthetically with the aid of seismic 3-D modeling techniques from a geological model, determined by lithological, petrophysical and / or structural parameters.

10. The method according to any one of the preceding claims, characterized in that a plurality of reference cutouts, for example drilling locations, are compared with the local cutouts and thus a plurality of similarity values are calculated for each data point.