Вы находитесь на странице: 1из 53

Overview Defining Survey Objectives This presentation provides an introduction to the seismic acquisition method.

When faced with the need for more seismic data it is easy to rush ahead with a survey without considering all the factors. It is most important to carefully review in detail the objectives of the survey. For example, in a new exploration play in a gas-prone province having thin reservoirs, it is clear that the exploration target must be large to be of any economic interest. In this case, one might choose a seismic survey with widely spaced lines of sources and receivers. Alternatively, the objective might be to determine the fluid content of the reservoir; in which case, it might be necessary to use long offset data for amplitude versus angle of incidence analysis, or even multi-component acquisition. There are thousands of possible combinations of acquisition parameters and methods; the optimal selection of which must be driven by the objectives of the survey, balanced against economic constraints. The objective most common to all surveys is to obtain clean reflections free of interference from noise, other wave types and multiple events. Review of Historical Information After carefully defining the survey objectives, the next step is to review previously existing information. This may include geologic information and logs from wells, previous seismic data, reports from geologic surface surveys, gravity information and operations reports. The subsurface information will assist in setting the resolution criteria for the survey an important economic factor. Previous seismic data will assist in identifying the noise problems to be encountered. It can be reprocessed and analyzed to provide estimates of attenuation, potential bandwidth, sampling requirements etc. Surface geology will provide information on near-surface dips, the likelihood of static correction problems and also the type of shot hole drilling equipment that may be needed. Gravity data may give dip and depth estimates. Operations reports will provide information on conditions and problems previously encountered; hopefully, they can be avoided in the new operations. Operational and environmental factors have a huge impact on the techniques used and the costs involved. Acquisition requires many different types of equipment, to be deployed depending on the environment. Clearly, with such a range of environments and equipment, no one individual or training module is sufficient to design and plan a seismic survey. However, this presentation will enable the geophysicist to understand the impact that historical information may have on the project and should promote the effective communication and coordination with the acquisition team. Accounting for Noise In this presentation, we will also discuss noise as it affects the seismic survey. Ambient Noise The vibrations from a seismic source are always in competition with naturally occurring vibrations. These vibrations are called ambient noise. The noise may be highly local to the geophone, for example, a leaf brushing the case of the geophone, or a grain of windblown sand striking it. Such noise is called incoherent or uncorrelated. If the ambient noise is ground-borne or airborne (for example, from a distant tractor), the noise will arrive at two closely spaced geophones at the same time and is said to be correlated. An intermediate situation occurs in the jungle or forest; the noise is generated in the tops of the trees but communicated to the geophones through the trunk and root system. The correlation distance may be determined by deploying geophones at a number of different

distances and recording the ambient noise. The data may then be analyzed to find the cross-correlation coefficient. Coherent Noise As discussed earlier, ambient noise can sometimes be coherent, depending on the distance between receivers. However, most coherent noise is generated by the seismic source. It includes the surface wave, known as ground roll on land ( Figure 1:An illustrative record.

Figure 1

Frame a' shows a wiggle-trace, from Antsey, 1981; reprinted by permission of Gebruder Borntraeger. Frame b' shows an interpretation of the wiggle-trace.) and the trapped normal modes at sea ( Figure 2:Water wave), multiple events, out of plane events, and air waves. In land seismic, ground roll is commonly the most significant problem. In marine seismic it is more likely to be multiple reflections.

Figure 2

With the introduction of recording systems capable of acquiring more channels of data, more powerful marine sources and the 3-D method, coherent noise has become a more significant problem than ambient noise in most surveys. Scattered Noise Scattered noise is a term generally given to surface waves or trapped waves, which have been reflected back to the receivers by discontinuities in the near-surface, Figure 3 (Plan view of source-generated scattered noise reflected back to the line by a surface irregularity).

Figure 3

On land, the responsible wave is likely to be ground roll or head waves; in the marine realm, it is likely to be a wave trapped between the surface and the first strong acoustic boundary. A simple example would be a reflection from an oil platform during a marine survey. This would appear as a hyperbolic event on the cross-section. However, there may be many discontinuities in the near surface, and they may be randomly located. This leads to the appearance of a large number of randomly placed diffractions on the seismic record. The superposition of many coherent diffraction events causes the scattered noise to masquerade as random noise by its incoherent appearance on a seismic record. For this reason, only recently has scattered noise been recognized as a significant problem. Survey Methods The following section provides an overview of seismic acquisition methods. 2-D Seismic Method Most seismic work (and particularly reconnaissance work) is conducted along lines. Such work is often called 2-D (two-dimensional, since the final output approximates to a twodimensional section through the earth). However, it is important to remember that only the final result is an approximation to a two-dimensional section through the earth. The actual acquisition process is three-dimensional. Each seismic source propagates energy in a three dimensional earth and each seismic receiver receives energy in a three dimensional manner. The signal and the noise propagate in three dimensions; it is only the analysis, processing and display which are two-dimensional. The objective is to work along a line to produce a vertical section which best represents the parameters of the earth that are of interest to the interpreter. Conceptually, one might achieve this in the same manner as a ship uses an echo sounder (

Figure 4 : Basic single channel recording of reflection times along a 2-D line); coincident sources and receivers would record data at regular intervals along a line, usually a straight line.

Figure 4

However, the echo sounder only produces a vertical cross-section of the sea bottom by making the assumption that the sea floor is horizontal in a direction perpendicular to the traverse of the vessel. Additionally, the echo sounder is only working in one simple layer - the water. It is relatively easy to convert the travel time of the echo sounder into the water depth in such a situation. In the seismic case, the earth has variable velocity. Also, interfaces generate many different seismic waves. These waves may cause interference with the objectives or may be useful, for example, in the case of converted energy in multi-component techniques. This makes it necessary to modify the simple single shot single receiver method to achieve the objective. Multiple versions of the reflection information are always recorded, trying to keep constant the desired component of the geophone signal (usually the primary compression reflections), and to introduce some kind of variability into the undesired components (everything else). Then, the multiple versions are added together, in hope that the desired reflections will reinforce coherently, while the undesired components will add less coherently (or even cancel). This method also provides the ability to estimate velocity within the earth, which allows the travel time to be converted into a more useful parameter to the Explorationist depth. 3-D Seismic Method The previous section stated that the 2-D method is an attempt to produce a vertical cross-section within a three dimensional world. Increasingly, data is collected over a grid of sources and receivers. These two dimensions at the surface, together with the estimations of depth in the subsurface lead to this technique to be termed threedimensional seismic'. This technique frees the geophysicist from the assumption that the subsurface is flat in a direction perpendicular to the seismic line. It has many other advantages including improved resolution and better signal to noise potential. This presentation will introduce most concepts using a two-dimensional approach, since this leads to an easier understanding; however, the concepts can readily be extended to

three dimensions. The presentation will introduce the concepts of the three-dimensional seismic method, but more detailed explanations will be found in other presentation. 2-D or 3-D? A properly designed 3-D survey will always provide a better image than a 2-D survey. Both ambient and coherent noise is better attenuated and out of plane reflections are properly positioned. Factors which sway the decision towards 3-D are structural complexity, extensive faulting, need for improved spatial resolution, and high cost of exploration failure. Frequently a 2-D survey does not solve the exploration problem and will be repeated with closer line spacing. The repeated surveys will then be followed by a 3-D survey. These repeats are expensive and involve lengthy delays. In most circumstances, especially marine, the economics favor an early acquisition of 3-D data (see Aylor, 1995). (See Nestvold & Nelson, 92, for a comparison of interpretations from 2-D and 3D surveys. And see Aylor, 1998, for a discussion of the improvement in exploration success using 3-D. A complete list of authors and papers is given in the Reference section of this presentation.) Multi-Component Seismic Conventional seismic generates both Pressure and Shear waves, but only the vertical component of these waves is measured on land, while only the pressure component is measured at sea. Multi-component seismic (MCS) uses detectors orientated in three directions at each receiver point. MCS might better be called vector seismic, as it attempts to measure the magnitude and direction of particle motion in three dimensions. This allows the simultaneous recording of both Pressure and Shear waves. The main advantages of MCS are the improved ability to predict fluid type, lithology and fractures, permeability, and also sometimes to provide a better structural image. Time Lapse (4D) Seismic The production of hydrocarbons from a reservoir may in some cases affect the seismic image. A simple example of this is the upward movement of the oil-water contact as the oil is produced. Recording seismic data before and during production may provide valuable information by showing differences between the images; for example, it may identify unswept pockets of oil. Before recording a time lapse survey, it is important to model the reservoir to determine whether the production will create a measurable effect

on the seismic image, Biondi, Mavko et al, 1998. It is also necessary to examine the cost benefits of such a survey. (See R.C Uden, R N Anderson, 1999.) The effects of production are likely to be very subtle. Ideally, the acquisition parameters should be identical for all surveys in the time lapse experiment, but changes in equipment and techniques make this unlikely. For this reason, it is important to consider the potential need for a later time lapse survey during the design of the very first survey over a field. Explosive Sources Dynamite An explosive charge is a popular seismic source because it is a compact and light source of acoustic energy. Traditionally, the explosive is high-velocity gelatin dynamite. Increasingly, however, various branded explosives, which are safer to handle, and which have fewer restrictions on transport and storage replace dynamite. Some of these are safe to the degree that the charge must include a booster, of greater sensitivity, before it can be detonated. Most shooters continue to call their explosives "dynamite," whatever they are; some old-timers still call them "powder." Generally, charges are in the range 1-10 kg (2 - 20 lb.). The explosive is typically supplied in cylinders 6 cm (2 in) in diameter, 60 cm (2 ft) in length and 1 kg or 2 lb. in weight. The cylinders are coupled together to make larger charges, and a booster is added at the top if necessary. Then an electric detonator (usually called a cap) is added, and the cap wires are secured to the charge. The assembly of a charge is shown in Figure 1 (Assembling charge).

Figure 1

This charge is usually loaded into a shot hole for both safety and to provide a better source pulse. A shot hole rig drills the shot hole. This rig consists primarily of a rotary drill a scaled-down version of an oil-drilling rig mounted on a truck (

Figure 2 : Elements of a shot-hole rig),

Figure 2

or on a marsh buggy ( Figure 3 : Shot-hole rig in marshland)

Figure 3

or on a tracked vehicle ( Figure 4 : Track-mounted rig) .

Figure 4

For terrain that is generally impenetrable to vehicles, or in environmentally sensitive areas, a small rig ( Figure 5 : Portable shot-hole rig) may be configured for transport by man or by helicopter.

Figure 5

For drilling shallow holes, an airgun may be sufficient. For drilling very soft ground, the rotary capability may not be necessary; the hole can be jetted rather than drilled ( Figure 6 : Jetting operation).

Figure 6

The shot hole is typically drilled with a 12-cm (4 -in) bit a fish-tail bit for soft and unconsolidated materials, or a rock bit for hard materials. The depth of the shot hole is a compromise between considerations of cost, shooting formation (See the presentation entitled THE SEISMIC PULSE ITS GENERATION AND TRANSMISSION), the water table, and ghost reflections; the practical choice of depth is discussed in the presentation on CHOOSING THE FIELD VARIABLES. Typical values range from 3 m (10 ft) to 30 m (100 ft). The drilling fluid may be water, air or a combination of both. If it is water, one or more water tankers are required ( Figure 7 : Shot-hole rig and tanker) (or dozens of workers will need to be employed to carry water cans on their heads).

Figure 7

The drill may encounter lost circulation, which requires the addition of bentonite to form a drilling mud. In serious cases, it may be necessary to case the hole; this process is expensive, and it may be worth salvaging and reusing any undamaged casing. Under favorable conditions, a small shot-hole rig might drill 30 shot holes (to 15 m or 50 ft) in a day. Under unfavorable conditions, or those requiring greater depths, several large rigs might be needed and may still get less production. It is almost always true that shot hole drilling represents a significant part of the cost of seismic fieldwork; in some places it is the dominant part of the cost. Later, we shall discuss ways of tackling this problem, by the use of surface sources. After the hole is drilled, the explosive charge is assembled, then loaded to the bottom of the hole by the shooter. For this he uses a heavy weight, or loading poles of wood or aluminum; these poles are typically 3 m or 10 ft in length, and interlock to form a rod by which the charge may be pushed down the hole ( Figure 8 : The loading operation). Alternatively, where the hole is subject to collapse, the driller may preload it.

Figure 8

The handling and loading of explosives is routinely and safely carried out on hundreds of seismic crews every day. However, this is possible only by the observance of sensible safety practices. The Manual of Safety prepared by the International Association of Geophysical Contractors sets out these practices. A copy should be available on every crew. Our lives are at stake so read it!!! With the charge loaded, we have the situation of Figure 9 (Charge in hole).

Figure 9

If there is a good head of water above the charge, this is sufficient; if not, soil or sand may be loaded down the hole to tamp the charge. When the shot is fired, our feet should feel a good thump, but our ears should hear only a thud; this indicates good tamping the energy has gone into the ground. A thump followed by a whoosh and a big plume of water, is also acceptable. But if the shot blows out with a sharp crack like a gun (or, worse still, if it is so shallow that it craters), the tamping is not sufficient, and the energy has been wasted upwards. Figure 10 (Shot plume)

Figure 10

looks like a good shot; Figure 11 (Blow out from shallow shot) does not.

Figure 11

The blaster ( Figure 12 : Firing the charge) fires the charge, which sends a pulse of electric current through the cap wires to the cap.

Figure 12

Modern practice requires that the moment of application of this pulse can be taken as the moment of detonation of the charge there must be no significant delay. This is achieved by using "seismograph caps" not blasting caps and by using a sufficiently large firing current. In some operations, the blaster is fired manually by the shooter on the instruction of the observer, with the instant of firing the time-break being transmitted to the recording instruments by radio (or by wire). More often, the blaster is made ready and "enabled" by the shooter, but fired remotely from the recording instruments. Also shown in Figure 13 (Charge in hole) is the up-hole geophone, usually placed within 3 m or 10 ft of the shot hole.

Figure 13

As noted previously, the initial output of this geophone is sent by radio to the recording instruments, and is recorded for later use. Obviously, the interval between the timebreak and the uphole arrival represents the travel time up the hole, and this allows us to convert our reflection times to be surface-to-surface times. It also provides a useful measure of the velocity of the near-surface materials provided that the geophone has not been kicked over during the loading operation.

Occasionally, for special or test purposes, the observer calls for a second shot in the hole possibly at a different depth. The shooter loads this in the same way, while the recording operation waits. In general, however, the urge is to shoot only one shot in each hole, and to press on to the next. Before doing so, the shooter is usually required to seal the shot hole, and to restore the surface. The size of the explosive charge must be carefully chosen. First, it must be sufficiently small that it does not create a safety hazard; second, it must be large enough to compete with the ambient noise; and finally, it must be small enough that it has a broad enough amplitude spectrum. The bandwidth of the dynamite source is inversely proportional to, and the amplitude directly proportional to, the cube root of the charge mass (see Sixta, D.P., 1982). This may be likened to the behavior of a loudspeaker; the larger source provides a larger shock wavefront that, just like a larger loudspeaker, generates low frequencies more efficiently. The charge depth must also be carefully chosen; firing the charge in consolidated clay will give a much sharper pulse than in a dry unconsolidated conglomerate. The driller must keep good records of the formations encountered and attempt to load the charge consistently in the same formation. Other Explosives Dynamite can also be packaged into an explosive cord. With explosive cord, the cord (69 mm in diameter, 1/4-3/8 in) is laid in a furrow cut by a vibrating plow ( Figure 14 : Three-furrow primacord plow); the plow continuously cuts the furrow, feeds in the cord, and tamps the soil back on top.

Figure 14

Typically the cord is laid in lengths of 25 meters, fired by a cap at one end; the cord is thus its own source array. Another explosive seismic source is the shotgun, which is typically only used for shallow work as it has low energy. With a shotgun, a projectile is accelerated to strike a baseplate, in contact with the ground, by a shotgun shell ( Figure 15 : Shotgun source), or even fired directly into the ground.

Figure 15

Dynamite is an extremely effective way of transporting very large quantities of energy, but it has its own hazards, and it requires the drilling of expensive shot holes. Clearly, in some situations, such as the middle of town or during a period of civil unrest, it would be most unsuitable. As a result, a number of alternatives have been devised. Gas Gun In a gas gun, a mixture of propane and oxygen is contained in a truck-mounted bell, with a movable diaphragm in contact with the ground ( Figure 16 : Dinoseis truck); since detonation is by a spark plug, multiple units can be fired simultaneously at each position across a source array.

Figure 16

An air gun is a variant on the gas gun. A volume of air at high pressure is discharged abruptly into water contained in a truck-mounted bell ( Figure 17 : Land air gun); the operation of air guns is described more fully later in this presentation.

Figure 17

Non-Explosive Sources Weight Drop In desert areas, much work has been done by the Thumper or weight-drop method. A large weight is carried at the back of a truck ( Figure 1 : Weight-drop truck showing weight just prior to impact); it is raised, and then dropped.

Figure 1

Many drops are made at each shot-point, with the truck moving a few meters between drops; thus sufficient input energy is built up over a source array. The source array is important, because ground roll is much more troublesome with a surface source (see the module entitled THE SEISMIC PULSE ITS GENERATION AND TRANSMISSION). The time-break is obtained from an inertial switch that is actuated as the weight hits the ground, and the vertical stacking is carried out with reference to this time-break. One of the interesting features of the weight-drop system is the ease with which we can calculate the energy employed. If, for example, a mass of 2000 kg is dropped from a height of 3 m, 16 times at each shot-point, the energy employed is about 1 MJ. Probably, a few percent of this appears in the compressional seismic wave. By contrast, a 10 kg charge of explosive releases some 30 MJ. If we say that the seismic results are comparable, the efficiency of the explosive in generating compressional waves despite the benefit of being down a hole, in intimate contact with the earth is probably about one-tenth of one percent. Vibrator An alternative to these short impulsive sources discussed above is the vibrator. Vibroseis, as the technique is known, is by far the most important of the surface seismic sources. This method utilizes the principle of spreading the source energy out over time, so that although the vibrator generates relatively low power compared with dynamite, the total useful energy may be comparable. The source may be small and mounted on a tractor for use in populated or environmentally sensitive areas, or may be large and mounted in huge trucks for use in the desert. Figure 2 (Vibroseis truck) shows a typical vibrator.

Figure 2

Figure 3 (Vibrator system diagram)shows a sketch of a typical vibrator.

Figure 3

Introduction Figure 1 illustrates a classic exploration target-a simple anticlinal structure.

Figure 1

This structure can be defined, along the line of section, by making measurements of the depth of the reservoir below datum ( Figure 2 ).

Figure 2

These values of depth can be posted at the corresponding shot-points on a shot-point location map ( Figure 3 ).

Figure 3

The same operations can be done on another line ( Figure 4 and Figure 5 ).

Figure 4

Figure 5

Finally the map is fully posted ( Figure 6 )

Figure 6

; it can then be contoured ( Figure 7 ).

Figure 7

The maximum hydrocarbon content is then defined by the spill-point contour (OWC of Figure 8 ).

Figure 8

The closure and the maximum area of pay can be read from the map. Seismic interpretation is the meeting-place of geology and geophysics, where the truth about the subsurface environment becomes evident only to the person who accepts a continuum between the two. There are several important principles related to the geological message in the seismic trace. The seismic method measures only physical properties and their changes; geological properties can be obtained only by inference. Reflections evident on a seismic section have a time and a character. The reflection time is an expression of the depth and dip of a geological surface, and of the velocity to it. That surface, in almost all cases, locally approximates the surface of the solid earth at some time in the past; thus it may be a bedding plane (rep resenting a one-time surface of deposition) or a local portion of an unconformity (representing a one-time surface of erosion or non-deposition).

Studies of reflection time allow us to see the folding and faulting of a geological surface, and so to recognize structural traps. Further, they allow us to see angular unconformity between two surfaces, and so to recognize truncation and erosion. Further again, they allow us to see patterns of surfaces, and so to recognize certain depositional mechanisms; from these, in turn, we can make inferences about facies. The reflection character is an amalgam of amplitude, waveform and polarity. It is an expression of the physical properties above and below the reflecting surface, and of interference between reflections from closely-spaced reflecting surfaces. Studies of reflection character allow us to map variations in the nature of the rocks, and so to supplement our inferences about their type and condition. With respect to basic seismic interpretation, our prime concern is with studies of reflection time, and with the correct representation of geological surfaces. This is of paramount importance, of course, to the location of structural traps. Thereafter we may pay more and more attention to reflection character, and to its implications for stratigraphic traps.

INTRODUCTION
Today, in mature petroleum provinces, we seldom or never find classical anticlines with simple four-way closure. All those structures have been found long ago. Nowadays, our standard fare consists of structures bounded in part by dip and in part by faults. We must become expert in recognizing such structures, and in deciding whether or not they have potential as hydrocarbon traps.

Conduits and Seals


Where a crack occurs in the earth, we usually call it a fracture. Where the rock layers are displaced across the fracture, we usually call it a fault. The importance of the fault, as distinct from the fracture, is that the fault has greater potential to form a trap (Figure 1 ).

Figure 1

At the time of cracking, the fault is likely to provide an escape for water (and possibly hydrocarbons) contained in the rocks below; it provides a conduit. For the fault to form a trap, however, the fault must provide a seal. An attractive situation arises where a fault is a conduit in its lower reaches, and a seal in its upper reaches. This may happen, for example, if the conduit formed by the crack allows the escape of mineral-rich waters from below, and the conditions of chemistry and temperature cause the minerals to precipitate out of solution as the water rises in the fault. If a seal forms just above a faulted reservoir we have a trap, while if the conduit remains open below the reservoir we have a path by which the reservoir may be charged from deep source rock. The danger in a fault trap, of course, is that the seal may be broken by later reactivation of the fault, so that the hydrocarbons are lost. Sometimes, therefore, we need faults to act as conduits, allowing the charge of a reservoir from a deep source. In other situations, we need faults to act as seals. And our constant enemy is the breach of a fault seal by reactivation of the fault. In Figure 1,we depend on the fault to act as a conduit, to charge the reservoir from the deep source rock. We depend on the offset cap rock at A to provide an updip barrier. But we also depend on a fault seal at B. If this seal does not exist, we have nothing. if it does, the spill point, C, depends on the thickness and offset of the cap rock.

In Figure 2,with a thicker cap rock, we do not need the fault seal at B, but it must be present higher in the section, perhaps at D.

Figure 2

If it is not present by E, the trap cannot be charged unless both the unconformity surface and its overlying rocks are impermeable-and then only after the development of this impermeability. If the fault has been reactivated as suggested at F, any charge depends on the subsequent sealing of this breach, and can represent only hydrocarbons generated since the resealing. We see, then, the critical decisions we must make in fault interpretation. Is it possible for the fault trap to be sourced? Does the fault plane seal? Did the seal develop before or after the generation and migration of the hydrocarbons? Where the thickness and offset of the cap rock imply a spill point, is the trapped volume sufficient to warrant drilling? Has the fault been reactivated? Is it likely to have resealed? When? Were hydrocarbons still migrating after this time? Having regard to the burial history, should we expect oil or gas? In general, there are uncertainties in our answers, and fault traps always carry an additional element of risk. Some decision-makers, as a matter of principle, will not drill downthrown fault traps, or traps where faults can be seen extending to the surface. Others point to the large number of proven fields having just these conditions. As always, the drilling decision should be based on a reasoned assessment of potential reward and potential risk.

One other seismic decision is very relevant to the assessment of risk: is there evidence of any mobile rock above the trap, in which any fault displacement could be absorbed? Thus, the presence of thick salt above the unconformity in Figure 2would virtually remove any risk from fault reactivation (though in Figure 1 it might not, because of the silt layer). Similarly, a thick overpressured shale is often seen to absorb quite major faulting. Where the salt or shale layer is thin, however, the fault may be transmitted; then we face the judgment of whether it is likely to be sealed by flowage of the mobile material into the fault zone. All in all, therefore, we cannot afford to dismiss fault traps, but working with them will often intensify our problems-and occasionally cause us sorrow.

Interpretation, Past & Present


In the past, a great deal of seismic fault interpretation has been grossly in error. Some of the reasons for this are understandable. With the poorer quality of old data, reflection continuity was often interrupted by seismic artifacts-changes in surface conditions, refraction blind spots, incorrect statics. To a person raised among mountains, every such break was a fault. In the regrettable days when geologists and geophysicists were kept apart, many geophysicists did not even know what the geologists knew about the nature of faulting. What explorationists know about basin dynamics and fault generation has increased enormously in the last decade. There is always a temptation to give the seismic data the weight of a measurement, while regarding the geological input as mere generalization. Today, we resist this, and insist on an interpretation that satisfies both the seismic measurement and the geological generalizations. Perhaps the clearest example of this is that we would no longer accept any fault interpretation, however "obvious," if it did not satisfy conservation of mass before and after faulting. One of the cheapest ways to find oil today is to reinterpret the old fault maps over producing oil fields in the light of present geological knowledge, and to discover pockets of oil never tapped by the drill.

Round and Round


In modern practice, therefore, we are much concerned with establishing the tectonic regime of the area-and how that regime has changed over time-before we attempt final picks on the faults. Here we put ourselves into an interpretive loop: we cannot pick the faults until we know the tectonic history, but we learn the tectonic history, very largely, from the faults. So no fault can be picked in isolation; we search first for faults that show the tectonic message most clearly, we make regional and local inferences about the history, and we go back into all the faults until a coherent pattern emerges.

2-D versus 3-D Seismic Data


Fault interpretation on 2-D seismic data is much less safe than on 3-D data. One reason for this is that 2-D lines are seldom sufficiently close to give definite fault correlations from line to line. Another is that fault interpretation is jeopardized whenever a 2-D line is

not perpendicular to the fault. However, we must still be ready to do the best we can on 2-D data. Fault interpretation on 3-D data has given us new insight into the geometry, distribution and behavior of faulting. Once, where we mapped a few long continuous faults using 2-D data, we now see more numerous short, discontinuous faults separated by some sort of transfer zone. Similarly, mapping around faults has changed. The use of 3-D data has allowed us to see the details of structure along and near the fault planes. We have learned that structural contours at the faults are not necessarily the smooth and flowing artistic lines that so many of us were taught in the past.

General Issues in Interpretation


Figure 1illustrates the fault-picking operation at a simple level.

Figure 1

We join the apparent breaks or steps in the reflections with a geologically plausible fault trace, continuing the pick upwards and downwards until the faulting becomes mere flexure, or disappears. Simple enough. But even in this simple case, we are looking for additional messages. Above the fault we are always vigilant for evidence of reactivation (possibly in the opposite direction), because such reactivation may have broken a seal. However, we also remember the possibility that mere compaction, particularly in shales, can suggest minor reactivation of a fault where none has occurred. Further, the figure illustrates that either type of movement can produce significant local changes in stratigraphy above a fault zone. Figure 2goes to an opposite extreme; it shows us that, even with fair record quality, fault interpretation can sometimes be extremely problematical.

Figure 2

In the face of the changing intervals between major reflections, the requirement for geological plausibility becomes the key; whatever picks we make, we must have an explanation for them. Just joining the reflection breaks in some arbitrary manner will not do. Even when the fault interpretation appears simple, we have to satisfy ourselves that other faulting is compatible with that interpretation. Are the other faults of compatible type and the same age (acceptable), or of different type and different age (acceptable), or of different type and the same age (problematical)? As we have agreed earlier, many of the problems of fault interpretation on 2-D data arise because the lines are not at right angles to the faults. This problem is made worse by the fact that many areas have been subjected to several tectonic episodes, each generating a system of subparallel faults-but at different orientations. Figure 3 is a fairly simple illustration of this, with a system of curved extensional faults crossed by a system of vertical strike-slip faults.

Figure 3

If we lay out a rectangular grid of 2-D lines over such an area, some or all of the lines must be at unfavorable angles. Figure 4illustrates fault picking at a basin edge.

Figure 4

We can see very readily the likely geological development, and the picking is straightforward. However, the earth is also capable of some exceedingly grotesque contortions, and sometimes our courage is taxed ( Figure 5 ,

Figure 5

Figure 6, and Figure 7 ).

Figure 7

Figure 6

In some cases, situations that would otherwise tax our courage are eased when we can recognize the presence of mobile rocks such as salt or overpressured shale. Thus, the major salt-induced fault in Figure 8 gives us few problems.

Figure 8

An essential step in fault interpretation is the recognition of episodes of extension, of neutral subsidence, and of compression. Thus, in Figure 9, we recognize an orderly subsidence at the edge of a rift basin, and the repeated reactivation and adjustment of the main basin-bounding fault.

Figure 9

On the other side of the basin, Figure 10shows the formation of a major left-tilted faulted block as a consequence of extension at rift time, before the filling of the basin; such blocks, of course, may represent attractive petroleum targets.

Figure 10

We would also read extension (followed by stable subsidence) into the tilted fault blocks of Figure 11.

Figure 11.

It is equally important that we recognize the evidence for compression. At a regional scale, this may be fairly clear, and confirmed by the mountainous nature of the surface. Occasionally we may even see such thrust situations at prospect scale (Figure 12 ); the associated rollover can yield a very important structural trap.

Figure 12

A more complex example, involving both compression and extension, is given in Figure 13.

Figure 13.

Many petroleum provinces have been subject to episodes of both extension and compression. When these do not act in the same direction, some sideways movements and adjustments occur, and these create distinctive fault patterns, both in plan and in section. Figure 14 braces us for the difficult decisions that we shall have to make, when we are working in such provinces.

Figure 14

Rock Strength
A substantial amount of literature exists with regard to experimental rock mechanics in the laboratory, the intent of which is to simulate and analyze the process and effects of deformation. Studies are usually performed on cylindrical, core-like samples, which are subjected to compressive or tensile stresses in a chamber whose pressure and temperature are regulated. From these studies, scientists have determined that deformation generally progresses through three main stages (Nadai, 1950). These stages are defined by the behavior of the deformed material, and are most clearly and simply shown by the use of stress-strain diagrams ( Figure 1 ).

Figure 1

In order, the three stages are as follows: 1. Elastic During this initial stage, the stress-strain relationship is linear. If stress is removed, the body reverts to its original dimensions. No deformation (permanent strain) results; strain is said to be completely reversible ( Figure 1 , segment A). 2. Plastic As stress continues to increase, it will eventually reach some limit beyond which the body suffers permanent strain. This is termed the elastic limit, or yield stress. Beyond this limit, material is said to behave plastically: any increase in stress brings a corresponding increase in deformation ( Figure 1 , segment B). 3. Rupture With continued increase of stress, the body will eventually fracture, rupture or fault ( Figure 1, point C). Elastic strain can also end in rupture; a material simply fails without having undergone any measurable deformation. Models of Deformation These three stages through which deformation normally progresses are idealized as the behavior of three hypothetical "bodies" that are subjected to stress, as shown in Figure 2 ,

Figure 2

Figure 3 , and Figure 4(Stress/strain relationships for several ideal materials: (a) Hookean (elastic) body; (b) St.

Figure 3

Venant (plastic) body; (c) Newtonian (viscous) body).

Figure 4

Each style of behavior is denoted by the name of a well-known mathematician and should be apparent from the stress-strain graphs shown. A Hookean body knows only elastic strain before rupture, and is approximated by a simple elastic spring attached to a fixed body. A St. Venant body, in contrast, shows elastic strain up to a yield stress and then deforms indefinitely by shear strain at that same stress. This type of behavior is approximated by a weight that is pulled across some surface by an attached spring; the spring stretches elastically up to the point where friction on the table top is overcome and the weight begins to slide. The last example, called a Newtonian body or, sometimes, Newtonian liquid, has no shear strength at all and therefore exhibits no elastic strain. It deforms by what is called viscous strain. In Figure 4, this type of body is represented as a porous piston pulled through a fluid-filled cylinder. A Newtonian body, then, will deform indefinitely in response to any shear stress, with the total strain being directly proportional to the amount of elapsed time. Strain Hardening These three types of ideal bodies help describe the components in progressive deformation when it is caused by increasing amounts of stress. We might compare them to the graphs shown in Figure 5 , which describe two generalized stress/strain relationships for known materials.

Figure 5

The principal difference between the ideal ( Figure 2, Figure 3 , and Figure 4 ) and reality ( Figure 5 ) lies in the behavior known as strain hardening. Such hardening is often the result of complex readjustment, or even recrystallization, of the material suffering strain. We can envision it on one level as being due to the compaction and realignment of particles during progressive deformation. In sandstones, for example, stress concentrations along grain boundaries result in extensive grain fracturing and dissolution. This causes the progressive filling of pore space with grain fragments and recrystallized quartz. All other conditions (e.g., temperature and pressure) being equal, this will increase rock strength-that is, greater and greater amounts of stress become necessary to impose the same increment of deformation ( Figure 5, upper curve). We expect certain clastic lithologies to respond in this manner. On the other hand, strain hardening can occur up to some ultimate strength, after which the stress necessary to cause a given strain decreases continually ( Figure 5, lower curve). Lithologies such as salt and gypsum-anhydrite may behave this way under certain conditions.

Brittleness and Ductility


On the basis of the differences between elastic and plastic behavior, we are able to characterize the general response of materials as either brittle or ductile.

Brittle materials rupture before any significant plastic deformation occurs. Such behavior in rocks is marked by the development of breakage discontinuities along the planes that represent maximum shear strain. These are not necessarily faults or fractures visible to the unaided eye, but may take place between individual grains or within crystal lattices. In contrast, material is described as ductile if it is able to undergo a large amount of plastic deformation before failing. It should not be assumed that brittleness guarantees faulting, or that ductility inevitably leads to folding. In every situation, whether a rock responds in a brittle or ductile manner depends on several parameters-composition, effective confining pressure, temperature, strain rate and anisotropy. It is, in fact, relatively meaningless to speak of the true "strength" of a rock without reference to these parameters. Figure 6,

Figure 6

Figure 7 ,

Figure 7

and Figure 8 ,

Figure 8

(A series of stress/strain diagrams showing the effects of confining pressure (Figure 6 ), temperature (Figure 7 ), and strain rate (Figure 8 ), on rock deformation on homogeneous (Solenhofen) limestone. Note for ( Figure 8): a strain rate of 10-7per sec is roughly equivalent to 1 year of time; most geologic deformation occurs at strain rates of 10-14per sec) indicate how laboratory analysis has shown deformation to vary as a function of pressure (a), temperature (b), and strain rate (c) on homogeneous limestone. These laboratory tests tell us that near the surface, at low temperatures and pressures, rock will tend to act in a more brittle manner. With growing overburden, which increases both pressure and temperature, ductility generally increases. At the same time, however, the element of time is crucial. If small amounts of stress are applied over sufficiently long periods of time, almost any rock will deform plastically; only those lithologies with very low resistance to shear, such as basalt, may not. But given normal rates of tectonic deformation, there is some transitional depth range over which the response of a particular rock type will grade from dominantly brittle to ductile. Figure 9is a schematic representation of the spectrum from brittle to ductile behavior in limestone, as determined by laboratory tests.

Figure 9

Such testing applies a uniaxial stress (either compression or tension) to a cylindrical sample that is sealed within a chamber whose temperature and confining pressure can be regulated. Limestone and marble have been favored as samples, since these lithologies are more isotropic than clastic rock types. In addition to pressure, temperature and strain rate, the other factor that determines a rock's response to stress is, in one sense, the most obvious-its composition. Figure 10 is a diagram derived by Handin et al.

Figure 10

(1963) from extensive lab experiments on natural rock. Shown are measured ductilities for four common sedimentary lithologies in a water-saturated condition. One interesting point made by this diagram is that, with only one kilometer of burial, marked differences in ductility already exist between limestone and other lithologies. This is partly the result of pore pressure effects that strongly reduce ultimate strength. It is also, however, directly related to the mineral structure of calcite, which allows for significant intracrystalline gliding, even at moderate pressures and temperatures. The same trend of ductility increase is true for sandstones, and is often aided by pore pressure effects. We might also note that dolomite, while slightly more ductile than quartzite at depths below two kilometers, shows the tightest range of permanent strain before rupture. Due to its mineral structure, dolomite does not deform readily by intracrystalline gliding. What is normally referred to as "rock strength," then, is not a fixed property, but a relative response, determined by a specific set of immediate environmental conditions. Since both engineers and geoscientists make use of the terms compressive strength, tensile strength and shear strength, two major points should be made in relation to these qualities. First, almost all materials are weaker under tension than compression ( Table 1). And second, it is very often the shearing strength of materials that determines when and how they will fail. Average crushing Tensile Shearing

strength Sandstone 740 Limestone 960 Quartzite 2020 Granite 1480 Basalt 2500

strength 10-30 30-60 30-90 30-50 -

strength 100-200 150-250 150-250 150-300 50-150

Table 1: Measured strength of various rock types at standard temperature and pressure

Вам также может понравиться