GB2129938A - Method of stacking seismic data - Google Patents

Abstract

Faults in a mineral seam are located by an in-seam seismic technique which involves locating shots A and detectors E in the seam and firing the shots sequentially. The detectors pick up the reflections of the shots from faults and traces from a reflector segment including a common reflection point D are stacked. A mode conversion point on a target reflector line is found by an iterative process in which the ratio of the sines of the angles of incidence and reflection at an estimated point on the line is compared with the ratio of the velocities before and after mode conversion. <IMAGE>

Description

SPECIFICATION Method of stacking seismic data This invention relates to a method of stacking seismic data to identify a fault in a geological seam.

The invention is particularly, but not exclusively, applicable to locating faults in coal seams.

The advance longwall face method of mining which is typically used in modern coal mines is highly capital intensive. It takes several months to set up a coal face and capital costs are high. A significant proportion of the longwall faces encounter small but serious geological faults. These, almost invariably, disrupt production. In a substantial number of instances, unexpected faulting results in the premature abandonment of faces. Faults usually affect the integrity of a hydraulically supported roof.

Faults encountered head-on bring consequential flooding and/or fire risks. Thus, unanticipated faulting can seriously affect the economics of mining and it is therefore highly desirable to know the geological structure of a seam prior to mining.

A number of techniques are available for investigating subsurface geology. Direct methods include borehole drilling, both vertically from the surface, and horizontally from the coal face itself.

Indirect or geophysical methods include seismology, resistivity, gravity and borehole geophysical techniques. Of these, seismology, as practiced from the surface of the earth, tells most about the subsurface structure. Large faults may be detected; however, small but vitally important faults can be missed. For this reason seismology is now also applied underground since it is known that a coal seam will guide, or channel seismic waves.

In-seam seismology differs from standard surface seismology in two main respects. Firstly, channel waves are dispersive. The signal from a quasi-impulsive source, e.g., a small explosion, is gradually phase encoded as it propagates. The initial sharp signal spreads out as it travels through the coal. Secondly, the underground environment is harsh. There are very stringent safety regulations in most countries governing the use of equipment underground and movement is restricted.

To obtain a viable production system the underground field technique must therefore be made as simple as possible.

The known method of in-seam seismo!ogy is to locate geophones in short, horizontal boreholes in the coal seam. Small charges are fired consecutively from other boreholes in the seam. The signals received by the geophones are recorded using a standard seismic recording system with say, 12 channels and a sampling rate of 0.5 ms.

The problem of dispersion can be handled using specialised data processing techniques. As shown in British Patent No. 2 032 103, dispersive seismic arrivals can be recompressed to appear as impulse-like events.

Reference has been made above to the standard surface seismology which is used. The information gained by such seismology is subjected to processing by a technique known as Common Depth Point (CDP) stacking. This is a method designed to improve the signal to noise ratio in seismograms and highlights seismic reflectors. CDP stacking is greatly facilitated by using a special field recording technique with predetermined locations for shots and geophones. The improvement in signal to noise ratio arises because the totality of seismic traces is processed to produce a smaller number of improved traces. There is in fact data redundancy because reflection points in the earth may be sampled more than once.

In underground seismology the environment does not permit the luxury of data redundancy. The field technique must be made simple.

Unfortunately, the term 'common depth point' is inaccurate. Only in the case of horizontal reflector is the term accurate. For a dipping reflector, the term is a misnomer. In such cases CDP stacking puts the reflectors in the wrong spatial positions. It is necessary to migrate the data to correct for this effect. Migration procedures, such as wave equation migration, require large quantities of data if they are to be accurate. In channel wave seismology large quantities of data are not available, and reflectors may occur at any angle of 'dip' to the geophone line. For large dips the standard midpoint gathers are inappropriate.

In surface seismology dispersion of seismic waves is usually neglected but for in-seam seismology it is vitally important. In British Patent No. 2 032 102 it is shown how channel wave data may be recompressed to produce impulse-like events but even when this is done the group and phase velocities remain different. Standard CDP methods do not differentiate between these two velocities.

It is an object of the present invention to modify the CDP stacking technique to remove or obviate the above difficulties. One way of doing this is disclosed in Patent Application No. 8032259 from which the present application is divided and which claims a method of stacking seismological data to identify a fault in a geological seam including the steps of arranging shots and detectors in a seam or the surrounding strata, firing the shots sequentially and detecting any reflections thereof as seismic traces with the detectors, and selecting from the seismic traces those having a common reflection point, wherein a reflector comprising a target is divided into a number of equal length segments and each trace is assigned to the segment which contains the point at which reflections would occur, belonging to the same line segment are stacked after applying a move out correction according to the formula: p2=xt2+x22+2x,x2 Cos 26+4L2+4L (x1-x2) Sin O in which P is the distance travelled by a seismic signal emanating at a shot, reflecting at a point on a reflector where specular reflection occurs, and being received at a detector; 6 is the angle between the shot-detector line and the reflector; L is the length of a line extending perpendicularly from the reflector to a point of origin on the shot-detector line; x, is the distance between the shot and the point of origin; x2 is the distance between the detector and the point of origin.

The present invention is an extension of this above noted invention and does not assume a simple reflection at a target but acknowledges that there may be mode conversion.

According to the present invention a method of stacking seismological data to identify a fault in a geological seam includes the steps of arranging shots and detectors in a seam or the surrouding strata, firing the shots sequentially and detecting any reflection thereof as seismic traces with the detectors, and selecting from the seismic traces those having a common reflection point, wherein a reflector comprising a target is divided into a number of equal length segments and each trace is assigned to the segment which contains the point at which reflection would occur, traces belonging to the same line segment being stacked, and in which mode conversion occurs; a mode conversion point being fixed by selecting a point as a target reflector line between perpendiculars to the line giving the source and receiver, comparing the ratio of the sines of the angles of incidence and reflection at a point on the reflector where specular reflection occurs with the ratio of velocities before and after mode conversion to determine which side of the midway point on the reflector the conversion point lies and repeating this comparison for the half of the reflection line selected successively until a small enough segment is reached.

In order to illustrate the method of the invention reference will now be made to the three figures of the accompanying drawings in which Figure 1 shows in diagrammatic form the layout of an underground coal seam and Figures 2 and 3 show actual traces obtained by applying the method.

Referring first to Figure 1 , this shows a reflector as the full indicated line at a distance L from a shot-geophone line AE, and at an angle 0 to this line. The shot is located at point A and the geophone at point E. For this arrangement the point D is the point on the reflector where specular reflection occurs. The distances x, and x2 are the distances to shot and geophone respectively from some point of origin G, on the shot-geophone line. The point F is the foot of the perpendicular line from G to the reflector.The total distance travelled by a wave emanating from A, reflecting at D and being received at E is denoted by p and it is possible to show that p2=x12+x22+2x1x2 Cos 26+4L2+4L (x1-x2) Sin 6 (1) The distance FD is S where S=[x1x2 Sin 26-L(x1-x2) Cos 8[2L+(x,-x,) Sin 6] (2) Normally seismic traces are gathered about the midpoint of AE. In this case x,=x2=x and equations (1) and (2) become p2-4(L2+x2Cos26) (3) S=x2 Sin 26/2L (4) The case x=0 corresponds to zero offset. Standard CDP stacking applies a transformation to each seismic trace in the stack to produce effectively a zero offset trace. These traces are then summed.

Equation 4 gives the difference between the actual reflection point and the zero offset reflection point.

In standard surface seismology x/L is usually small and the angle 6 is small. Therefore S is small. That is, all traces in the CDP stack have reflection points near to the zero offset reflection point. It is therefore valid to gather traces with a common midpoint.

For in-seam seismology X/L is much larger, and 6 may take any value. Consequently S is not necessarily small. Traces with quite different reflection points may be stacked if the midpoint is used as the gathering criterion. This effect is overcome by gathering only those traces with similar reflection points.

A target reflector at a certain angle of interest is defined by the processor. This target is subdivided into a number of equal length segments, and each trace is assigned to the segment which contains the point at which reflection would occur. This of course is a dynamic labelling dependent on the target of interest. Those traces belonging to the same line segment are stacked after applying a move out correction given by equation (1). Using this method only traces with similar reflection points are stacked. This technique of gathering traces by their reflection points on a target rules out the normal procedure of sorting traces into gathers before processing.Because trace gathering is target dependent and does not require a certain arrangement of shots and geophones, traces from any combination of shots and geophones may be stacked. Indeed the shots and geophones could be in different mine roadways.

The stack itself may be carried out in distance-distance space, or distance-slowness space. In Figure 1 the line AE was chosen as a reference for defining the target distance and angle, but of course any convenient line may be chosen as a reference. It is sometimes advantageous to choose the reference line parallel to the target. It is then easy to combine adjacent target segments, either with or without overlapping, to increase the fold of the stack, thus highlighting targets parallel to the reference line. The system used is referred to as dynamic trace gathering (DTG).

DTG is extremely flexible. It permits targets at any angle to be imaged. The field shooting technique is simplified. There is little data redundancy. One unfortunate consequence of the method is that the coverage on any particular target may be non-uniform and may vary in density along the target If targets at a certain angle are of particular interest, then it is possible to design the layout of shots and geophones so that coverage and density are uniform along this target. If a target location is already known the method may be used for velocity analysis.

The present invention modifies the above description by firstly, not necessarily assuming a simple reflection at any target but assuming that mode conversion occurs. When the shot and geophone positions are fixed, and with a certain target assumed, then it is possible to derive a complicated algebraic equation for the mode conversion point on the target, provided that velocities before and after mode conversion are known, or assumed. An easier computational procedure is to find the conversion point iteratively. The mode conversion point must lie between the feet of perpendiculars dropped from the shot and geophone to the target reflector. As a first estimate, a point midway bewteen the feet is chosen.The ratio of the sines of the angles of incidence and reflection is then compared with the ratio of the velocities to determine the side of the estimated point that the mode conversion point, given by Snell's law, lies. The next estimate is taken as the midpoint of the section of the target to which has been narrowed down the mode conversion point. Thus each iteration will halve the error in the estimate of the mode conversion point. DTG with mode conversion can also be used to analyse data from transmission shooting. If mode conversion takes place at any target between the shots and geophones, then the target will be imaged.

Referring now to Figures 2 and 3, these are DTG stacked sections of actual reflection data filtered at a centre frequency of 400 Hz. Figure 2 is a distance-distance section using a velocity of 900 m/s. A prominent reflector at a distance of 130 m from the reference line across the central half of the section can be seen. Figure 3 is the complementary distance-slowness section assuming a reflector at 1 30 m.

Not surprisingly, a reflected mode at 1.1 s/km (900 m/s) is apparent. However, no other modes appear to be reflected. These results have been verified by subsequent mining operations.

It will be appreciated that the method of the invention improves the signal to noise ratio at target reflectors which occur at any angle to the shot-detector array and that and convenient arrangement of shots and detectors can be used without any complex field technique being necessary. Furthermore field techniques can easily be designed to optimise the DTG stack for targets at any specific angle and DTG can be used on either transmission or reflection data, with or without mode conversion.

The invention can also be used for velocity analysis and can operate successfully in cases where the group and phase velocities are different. DTG can combine traces from two-component geophones to image with a resultant signal in any direction.

Claims (4)

Claims

1. A method of stacking seismological data to identify a fault in a geological seam including the steps of arranging shots and detectors in a seam or the surrounding strata, firing the shots sequentially and detecting any reflections thereof as seismic traces with the detectors, and selecting from the seismic traces those having a common reflection point, wherein a reflector comprising a target is divided into a number of equal length segments and each trace is assigned to the segment which contains the point at which reflections would occur, traces belonging to the same line segment being stacked, and wherein mode conversion occurs and in which a mode conversion point is fixed by selecting a point as a target reflector line between perpendiculars to the line giving the source and receiver, comparing the ratio of the sines of the angles of incidence and reflection at a point on the reflector where specular reflection occurs with the ratio of velocities before and after mode conversion to determine which side of the midway point on the reflector the conversion point lies and repeating this comparison for the half of the reflection line selected successively until a small enough segment is reached.

2. A method as claimed in claim 1 in which the stack is carried out in a distance-distance space.

3. A method as claimed in claim 1 in which the stack is carried out in a distance-slowness space.

4. A method as claimed in any preceding claim in which a reference line is chosen as being parallel to a target. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

4. A method as claimed in any preceding claim in which the reference line is chosen as being parallel to a target.

New claims or amendments to claims filed on 21/12/83 Superseded claims 1, 4 New or Amended Claims:-

1. A method of stacking seismological data to identify a fault in a geological seam including the steps of arranging shots and detectors in a seam or the surrounding strata, firing the shots sequentially and detecting any reflections thereof as seismic traces with the detectors, and selecting from the seismic traces those having a common reflection point, wherein a reflector comprising a target is divided into a number of equal length segments and each trace is assigned to the segment which contains the point at which reflections would occur, traces belonging to the same line segment being stacked, and wherein mode conversion occurs and in which a mode conversion point is fixed by selecting a point on a target reflector line between the feet of perpendiculars dropped from the shot and detector to the target reflector line, comparing the ratio of the sines of the angles of incidence and reflection at the point on the reflector line where specular reflection occurs with the ratio of velocities before and after mode conversion to determine which side of the midway point on the reflector line between the said feet the conversion point lies and repeating this comparison for the half of the reflection line selected successively until a small enough segment is reached.