As can be seen from the original VSP record in Figure 6-2-4, the original VSP record is very different from the ground seismic record. It is difficult to directly use it for interpretation and must undergo certain processing. .

1. Processing of zero-offset VSP data

In addition to the conventional processing of zero-offset VSP data, the processing of zero-offset VSP data has the same content as surface seismic data processing (such as frequency analysis and bandpass In addition to filtering, wavelet shaping, true amplitude recovery, deconvolution, etc.), what is more special is the separation of upstream and downstream waves. In addition, the identification and separation of primary waves and multiple waves is also a very important issue. Only relevant special processing methods are introduced here.

1) First arrival picking

Picking up the arrival time of the first arrival direct wave is the basis of VSP data processing. The accuracy of picking directly affects the accuracy of all subsequent processing results, including the accuracy of calculating speed parameters in interpretation. At present, many methods have been developed to use computers to automatically pick up the arrival time of first arrival waves, including general cross-correlation technology and complex neural network technology, but none of them can guarantee the absolute accuracy of the results. In order to ensure the accuracy of picking up first arrival waves, corrections are generally made through man-machine collaboration.

2) Static time shift

Static time shift is to move each channel in the VSP data by a certain time, so that the upgoing wave or downgoing wave event in the record moves according to time. Aligned separately and displayed similar to ground seismic sections.

For the VSP observation of the zero-offset horizontal interface, it is assumed that the direct wave, upgoing wave, and downgoing wave (secondary reflected wave) received by the geophone in the well arrive at t1, t2, and t3 respectively. The travel time of the received reflected wave is t0. From the time interval relationship formula discussed in Chapter 2, we can know that there is the following relationship between them:

t2+t1＝t3-t1＝t0 (6-2-1)

That is, if The arrival time of the upgoing wave in each channel is added to the first arrival time, which is equivalent to the arrival time of the reflected interface reflection wave received by placing the geophone on the wellhead surface. In other words, the upgoing waves will be aligned according to their round-trip time from the surface to the interface. The process of adding a first arrival time to each channel is called static time shift. At the same time, the time of the first arrival will also double, and the slope of the event will also double. Figure 6-2-5 is a schematic diagram illustrating the alignment display of the upgoing wave after static time shift. Figure 6-2-5a is the ray path diagram of the upgoing wave. Figure 6-2-5b is the VSP record before alignment. Figure 6- 2-5c is the VSP record after the upgoing waves are aligned. Figure 6-2-5d is the result after the coordinates are rotated 90°. This display method facilitates comparison with the ground seismic profile. Figure 6-2-6 is a schematic diagram showing the upgoing waves aligned and displayed after static time-shifting when there are two reflection interfaces. If a first arrival time is subtracted from each channel so that the first arrival waves appear at the same time, then all the downgoing waves will be aligned, thus highlighting the downgoing first arrival waves and multiple reflection waves.

Figure 6-2-5 How the upgoing waves are displayed after being aligned

See the text for instructions

3) Separation of up and down waves

< p>In VSP data, upgoing and downgoing waves are recorded at the same time, and they overlap. Effective separation of upgoing and downgoing waves is an important task in VSP data processing.

Separating upgoing and downgoing waves is mainly based on the difference in apparent speed of the two. The methods include multi-channel velocity filtering, f-k filtering, τ-p domain filtering, median filtering and other methods. In fact, after the aforementioned first arrival picking and static time shift, the upgoing waves (or downgoing waves) have been aligned, while the slope of the downgoing waves (or upgoing waves) is longer. If conventional superposition is performed at this time, the upgoing wave (or downgoing wave) will inevitably be strengthened and the downgoing wave (or upgoing wave) will be suppressed. This is also a method of separating upgoing and downgoing waves.

Figure 6-2-6 The situation after the upgoing waves are aligned when there are two interfaces

See the text for instructions

A. Multi-channel velocity filtering. Multi-channel velocity filtering, which is effective in conventional seismic exploration, can also be effectively used to separate upstream and downstream waves in VSP.

Figure 6-2-7 Separation of upstream and downstream waves in frequency-wavenumber domain

See text for description

B. Frequency-wavenumber (f-k) domain filtering . The work of separating upstream and downstream waves in VSP records can also be performed in the frequency wavenumber (f-k) domain. The basic principle is shown in Figure 6-2-7. Figure 6-2-7(a) is the original VSP data. Strong downward waves are represented by thick lines, and weak upward waves are represented by thin lines. Figure 6-2-7 (b) is the result of performing a two-dimensional Fourier transform on Figure 6-2-7 (a) to transform the data in the space-time domain into the frequency-wavenumber domain. At this time, the downgoing wave is in the positive wavenumber plane, upgoing waves are in the negative wavenumber plane. Figure 6-2-7(c) is the result of filtering Figure 6-2-7(b), that is, multiplying the data on the positive wave number plane by a decimal (such as 0.001), so that the downgoing wave is attenuated by about 60 dB, and the negative wave number plane is attenuated by about 60 dB. Planar upward waves are not affected. Figure 6-2-7(d) is the result of performing two-dimensional inverse Fourier transform on Figure 6-2-7(c) back to the time and space domain. The downgoing wave has been attenuated and the upgoing wave has been enhanced.

4) Corridor Superposition (VSPLOG)

Corridor Superposition is a process to suppress multiple waves and strengthen primary waves.

What is used is the difference that the primary wave intersects with the first arrival wave but the multiple waves do not intersect with the first arrival wave, as shown in Figure 6-2-8. Figure 6-2-8 (a) is the original VSP profile. In the figure, there are both primary upward reflected waves and upward multiple waves. U1, U2, and U3 are upward primary waves, and US1, US2, and US3 are upward multiple waves. Wave. Figure 6-2-8 (b) is the VSP profile after static time shift correction, and the time axis is round-trip time. Because the multiple wave terminates at the depth of the interface where it is generated and does not intersect with the direct wave, it is a short event. However, the primary wave intersects with the direct wave, so there is only a primary wave near the direct wave and no multiple waves. Divide a strip (corridor) along the line (oblique line) connecting the oblique event axis of the first arrival wave to the termination point of the multiple waves. Keeping the primary wave in the strip will cut off the multiple waves; superimpose the events of the primary wave to form A single seismic trace, as shown in Figure 6-2-8 (c), can obtain a record with strong primary wave energy; this work is called corridor stacking.

Figure 6-2-8 Corridor Superposition

See text for description

2. Interpretation and application of zero-offset VSP data

VSP data contains rich information on geological strata and lithology. Combining it with surface seismic, drilling, logging and other data can greatly improve the accuracy of interpretation. To sum up, there are the following applications.

1) Extract accurate velocity parameters

Extracting velocity parameters from VSP data is the same as using seismic logging, both are calculated based on the first arrival time. However, the velocity measurement using seismic logging and sonic logging is limited by some conditions and the accuracy is not enough. For example, the point spacing of seismic logging is too large; although the layering of sonic logging is finer, it is affected by well diameter changes and time accumulation, making the accuracy lower. In particular, the physical nature of sonic waves and seismic waves is different, and the results will be Certain differences. The push detector is used in VSP operation, which improves the sensitivity, small point distance, and accurate position. In this way, the accuracy of the speed measured by the first arrival wave will be greatly improved.

2) Calibrating seismic geological horizons

In the past, there were two methods to determine the geological attributes (including age, stratigraphy and lithology) of reflected waves on ground seismic profiles: one is to The surface seismic profile of the well is converted into time and depth, and then compared with the drilling; another method is to use the synthetic recording method, making a theoretical synthetic record based on the well logging data and comparing it with the seismic profile, and then using the geological data obtained from the drilling to calibrate the seismic layers. . Whether the above two methods can obtain satisfactory results depends to a large extent on the accuracy of the speed used. If there is an error in the speed, the calibration will not be satisfactory. Generally speaking, the average velocity used in time-depth conversion may always have errors, and the physical properties of sonic logging data and seismic waves are different, so conventional calibration errors are larger.

Use VSP data to calibrate seismic geological layers, and use VSP to record two major characteristics of the previous wave event in the two coordinate axes: the primary wave event and the first arrival wave in the depth coordinate direction. The event axes intersect, and the depth at the intersection is the depth of the formation that generates the primary wave; the intersection of the primary wave event axis and the time axis in the time coordinate direction is the two-way running time of the primary wave. This allows a direct connection between drilling (depth) and surface seismic records (time), independent of velocity parameters. As shown in Figure 6-2-9, firstly, the VSP record is directly compared with the drilling (well column) and logging data. The four primary reflections marked A, B, C, and D on the VSP record are compared with the initial reflections. The intersection points of the arrival wave events determine the depths of the formations that produce these reflections as A′, B′, C′, and D′. From the well pillars, we can know the age, stratigraphy and lithology of these reflections; then, at the time recorded by the VSP By comparing them with the ground seismic section on the axis, the geological properties of the reflecting horizons on the seismic section can be determined.

Figure 6-2-9 Use VSP records to identify and calibrate reflectors

3) Identify multiple waves

Use VSP data to identify multiple waves in ground seismic records Multiples are accurate and convenient, and can indicate the source and propagation process of multiples. Any upgoing wave that intersects the event axis of the first arrival wave is a primary wave, and any upgoing wave that does not intersect with the event axis of the first arrival wave is a multiple wave; the depth of the discontinuity point of the multiple wave event axis indicates the source of the multiple wave.

Figure 6-2-10 VSP recording of multiple waves

If VSP data is used to determine the upper and lower interfaces where multiple waves are generated, both the upgoing and downgoing waves must be recorded. exists, as shown in Figure 6-2-10. In the figure, the two upgoing wave events (A and B) intersect with the first arrival wave event near 0.7 s and 1.4 s respectively, indicating that the two events are both primary waves. According to the intersection point, the depth of the interface is 1402.08 m. and 3322.32 m. The events C and D are the interlayer multiples that go back and forth between the above two interfaces.

4) Extract deconvolution factors

Deconvolution can improve the resolution of seismic data, but its effect depends on whether the deconvolution factor is correct. In surface earthquakes, because the surface only receives upgoing waves, the deconvolution factor can only be extracted from the upgoing waves. Both theory and practice show that if the deconvolution factor can be extracted from the downgoing wave, the deconvolution effect can be greatly improved.

Because the propagation of upgoing waves in the formation first runs downward and then upward, it is affected by factors such as the formation in both directions. Downgoing waves are only affected by factors such as strata in one way, and the signal characteristics and intensity are better than upgoing waves. The first arrival waves in VSP are just such downgoing waves, which are easy to identify and extract. They can be used to obtain the optimal deconvolution factor.

Figure 6-2-11 Predicting the depth of the reflective layer below the well

5) Predicting the depth of the reflective layer below the well

Drilling data can only understand the formation in the well situation, the conditions of underground strata deeper than the well depth cannot be predicted. VSP can not only receive seismic waves from above the geophone, but also obtain seismic waves below the receiving point, so it can predict the stratigraphic conditions below the bottom of the well. Because the primary reflection wave from the formation below the well cannot intersect with the direct wave event axis in the VSP record (the intersection point should be at the depth of the formation where the primary wave is generated), as shown in Figure 6-2-11, it can be extended along the direction of the direct wave event axis. By intersecting two rays in the extended direction of the reflected wave, and combining them with the ground seismic profile, the depth of the reflective layer can be better predicted. Note that the A event in the figure must be the primary wave confirmed after multiple wave elimination.

In addition, VSP data can also be used to calculate absorption attenuation coefficients, extract Poisson's ratio parameters, perform formation lithology interpretation and reservoir lateral prediction, etc., which will not be described here.