Technical Tips

In an effort to share our knowledge with our peers we have compiled a list of important tips.

Tips:

1. Technical Tip #1: Overview of Reservoir Simulation
2. Technical Tip #2: Overview of Reservoir Characterization
3. Technical Tip #3: Inference of Potential Reservoir Compartmentalization from Production and Pressure Data
4. Technical Tip #4: Peppering’ – A quick technique for determining infill locations
5. Technical Tip #5: How Fast does Pressure Move Through Your Reservoir?

Advice:

1. Sage Advice #1: Reservoir Engineering in Your Particular Field
2. Sage Advice #2: Reservoir Engineering and the Economics of Your Field
   
3. Sage Advice #3: Heterogeneities and Permeability Variations
   
4. Sage Advice #4: The Scaling or Averaging Problem
   
5. Sage Advice #5: Aquifers and Natural Water Drives
   
6. Sage Advice #6: Reserve Estimates
   
7. Sage Advice #7: Some Outstanding Problems in the Physics of Oil Production

Technical Tip #1: Overview of Reservoir Simulation

Technical Tip #2: Overview of Reservoir Characterization

Technical Tip #3: Inference of Potential Reservoir Compartmentalization from Production and Pressure Data

Here’s a tip for the novice reservoir engineer: a review of the production and pressure data may reveal potential complex compartmentalization in the field.

The attached figure summarizes the production data at the wells within the area of interest, i.e.:

• the average reservoir pressure in MPa, and
  
• the latest water cut in %, where
  
• a “–“ indicates that no data is available.

The following observations are apparent:

• High reservoir pressure (>40 MPa) has been observed at the two injectors, INJ1 and INJ2. This measured pressure does not dissipate even after long-term shut-in. This phenomenon suggests very high reservoir pressure.
  
• Only one well(PROD14) has experienced high pressure (>40 MPa). All other producers see relatively low pressure (~20 MPa). This and point one above suggest there is very low effective permeability between some of the injectors and producers despite high flow capacity (Kh) seen at the wells.
  
• Except for three wells, all wells see very low water cut (~0%). The three wells with high water cut (~90%) are PROD5, PROD8 and PROD15. The simulation results show that if permeability between the injectors (INJ1 and INJ2) and producers (PROD5 and PROD8) is not reduced, water cut at PROD4, PROD6 and PROD7 should have been higher.

The above observations suggest:

• Barriers between the injectors and the producers with zero water production and low pressure,
  
• Partial barriers between the injectors and the producers with high water cut and low pressure.

Special thanks to Stephen Wong, M.Sc., for the submission of the above technical tip.

Technical Tip #4: ‘Peppering’ – A quick technique for determining infill locations

One of the common objectives of simulation work is defining locations for new infill wells. After achieving an acceptable history match, the next step in simulation is obtaining a variation of forecast runs to find the optimal scenario.

The standard way of selecting new infill locations requires consideration of oil and water saturation plots, pore volume plot that shows fluid-in-place distribution, well productivity, offset well rates etc. A time-saving way is ‘peppering’, i.e. placing maximal number of infill wells in the space between existing wells and then running the case. The better producers of that run should allow the engineer to quickly high-grade their next forecast.

Special thanks to Sonja Stojanovic for the submission of the above technical tip.

Technical Tip #5: How Fast does Pressure Move Through Your Reservoir?

Pressure transient "waves" move at a different (higher) speed than do produced fluids in an oil or gas reservoir. Fluid flow in the reservoir is dictated by the reservoir transmissibility times the applied pressure gradient:

However, the transmission of pressure transients through a reservoir depends on the hydraulic diffusivity of the rock/fluid system:

Many engineers believe pressure transients move much faster in gas than they do in oil reservoirs since the gas viscosity is so much lower than that of the oil. However, this ignores the impact of compressibility, which is much higher in gas reservoirs. In fact, the two parameters tend to cancel one another out, so that pressure transients move at roughly the same speed in gas and oil reservoirs:

As you can see from the Table above, the diffusivity in "typical" oil reservoirs is the same as that of "typical" gas reservoirs. One way to think about pressure transmission is to split the hydraulic diffusivity into two components, one relating to the rock properties and another to the fluids:

From a "fluids" perspective, the (higher) oil viscosity "eats-up" or dissipates the pressure transient to about the same extent as the (higher) gas compressibility. As a result, pressure transients will move through oil and gas reservoirs at similar speeds, depending on the reservoir quality (permeability).

Special thanks to Dennis Beliveau for the submission of the above technical tip.

Sage Advice #1: Reservoir Engineering in Your Particular Field

"Handbook instructions for field operations are not given here. Nor will rule-of-thumb procedures for reservoir exploitation be listed. The subject is not susceptible to dogmatic generalization. If there is any general rule pertaining to it, it is that it is governed by no rule, beyond the basic laws of physics, which is universally valid for all producing reservoirs.

The physics of oil production is unique among all applications of physics. It deals with objects - the oil reservoirs - that are variable continua, subject only to virtually infinitesimal sampling. Their physical histories are irreversible transients that can be observed only at isolated points "on the fly". They are controlled by a trinity of forces: hydrodynamic stress gradient, gravity, and interfacial; superposed on a trinity of phases: oil, gas and water. Except for relatively microscopic laboratory investigations, any major and significant field experimentation creates an irreversible change in the state of the system and destroys the possibility of repeat or comparative experimentation on the same reservoir. And finally, there are no identical specimens in the ensemble of the objects of study. Yet the goal toward which the study of reservoir engineering strives is the prediction and interpretation of the detailed behavior of individual reservoirs! Little wonder is it that the subject is hardly beyond the state of infancy.

Where feasible, actual oilfield observations and data are presented as illustrative parallels of those anticipated from theoretical considerations. However, such parallelisms should not be construed as establishing quantitative rules or generalizations applicable to broad classes or individual examples of fields that are superficially similar to the particular examples cited. For every such generalization that might be formulated, fully as many exceptions will be found as instances of agreement. Such exceptions will result from purely physical factors, which are at least quantitatively different from those tacitly or explicitly assumed in establishing the so-called "generalizations"."

From Physical Principles of Oil Production, by Morris Muskat, 1949

Sage Advice #2: Reservoir Engineering and the Economics of Your Field

"The ultimate goal underlying the development of the science of reservoir engineering is the attainment of a maximum efficiency in the exploitation of hydrocarbon-bearing reservoirs. This implies the maximum recovery of oil and gas at minimum cost. While the consideration of economic factors may seem foreign to a discussion of "physical principles", it must be recognized that these same physical principles would be of very little interest - at least to the oil industry - unless they were applicable to actual reservoirs of commercial significance.

It is indeed the primary function of the reservoir engineer to evaluate the many individual factors characterizing a particular reservoir, and to determine their composite effect in modifying the performance that might be expected of the idealized prototypes illustrating the broad physical principles of oil production. Operating procedures ideally suited to one field may lead to inexcusable inefficiency in another, in spite of their superficial similarity in some respects. Reservoirs are not the product of a mass-production assembly line, designed and constructed to operate according to specifications. They are objects for individual study and analysis. And in the light of such study, they should be developed and exploited to achieve the maximum returns inherent in their individual "personalities"."

From Physical Principles of Oil Production, by Morris Muskat, 1949

Sage Advice #3: Heterogeneities and Permeability Variations

"The virtually infinitesimal sampling of underground strata provided by cores a few inches in diameter taken at lateral separations of several hundred feet can hardly suffice to establish the details of permeability variations, if any, between producing wells. On the other hand, the very fact that the permeabilities of cores at equivalent stratigraphic depths generally do vary from well to well, and sometimes can be correlated in contour form, shows that an assumption of strict uniformity may have as little basis as any assumed type of variation.

These results provide a possible explanation for the large variations in the production capacities that are frequently observed for neighboring and apparently identical wells. For if the formation is not strictly homogeneous in detail, it is not unlikely that neighboring wells may penetrate zones of appreciably different local permeability, and hence show markedly different production capacities, even though the producing stratum as a whole may be considered reasonably homogeneous.

Moreover, owing to variations in the permeability of virtually all formations, no constant or monotonically changing connate water saturation is to be expected in actual reservoir rocks. Except immediately above the water-saturated zone, the values will generally reflect the local capillary structure and permeability of the rock more than its location above the water-oil contact."

From Physical Principles of Oil Production, by Morris Muskat, 1949

Sage Advice #4: The Scaling or Averaging Problem

"On the other hand, "proof" of agreement between theory and field observations in particular instances might be questioned as being accidental, since it is extremely difficult to defend the absolute correctness of any quantitative evaluations of such reservoir data as effective pay thickness, true "average" permeability, the permeability-saturation relationship, the porosity, the connate water saturation, etc.

Finally, it should be noted that even when a reservoir has been subjected to intensive specific study, the data that it is feasible to gather are never as complete as desired. The complete coring and logging of every well in a field would still provide only an infinitesimal sampling - one part in millions - of the reservoir rock. Core analysis data itself is often subject to considerable uncertainty (and bias) in interpretation. Actual producing formations are virtually never strictly uniform, homogeneous and free of stratification. The problems of "averaging" thus created are in no sense completely solved. And even the gross performance data are often subject to serious questions. Reservoir pressures are sometimes in doubt because of insufficient buildup after shutting in the well, and they always involve the problem of averaging. The commonly observed vagaries of individual well fluctuations in pressure attest to the fact that significant field pressures are basically a statistical inference. Water production data are frequently subject to considerable question. And even more serious is the potential uncertainty generally inherent in estimates or record of actual gas production and gas-oil ratios. The uniqueness of such specific data and their applicability to the actual producing intervals between surveys are assumptions that must at best be evaluated as necessary evils.

These basic limitations to the analysis of actual reservoir performance must be recognized. Their discussion may be considered as an apology for the unhappy lack of conclusive field observation evidence with respect to many aspects of reservoir behavior. For this situation is the result of practical difficulties, often beyond the control of the operator and engineer, rather than a consequence of a lack of appreciation of the importance of the problem. Nevertheless, the circumstances underlying these limitations and their implications must be understood clearly if progress is to be made in establishing reservoir engineering analysis on a sound basis."

From Physical Principles of Oil Production, by Morris Muskat, 1949

Sage Advice #5: Aquifers and Natural Water Drives

"Because of the lower compressibility of water, its fractional expansion in volume on pressure release is lower than that for oil. However, when a mobile contiguous water reservoir is present at all, its area (volume) will often be very much greater than that of the oil reservoir, so that in spite of its lower compressibility the total expansion volumes may exceed the whole of the oil reservoir. Thus, whereas the great majority of known oilfields have areas of 10 miles2 or less, water reservoirs extending over 1,000 miles2 are not uncommon.

Complete water drive does not necessarily imply that, if and when volumetric equality between hydrocarbon withdrawals and water-intrusion rates is established, no further decline in reservoir pressure will take place. On the contrary, the pressure may continue to drop throughout the production history even though the rate of water entry is at all times substantially equal to the volumetric fluid withdrawals from the reservoir. For example, the East Texas field pressure declined by 600 psi after an oil production of 2.3 billion barrels, although 98% of that production was replaced by water intrusion. The reason is that to maintain an influx rate equal to the net rate of fluid withdrawal, the pressure at the oil field boundary has had to decline. Thus the water-drive mechanism usually leads to a slow pressure decline with increasing cumulative recovery, after an initial rapid decline required to establish the pressure gradients that induce the water entry."

From Physical Principles of Oil Production, by Morris Muskat, 1949

Sage Advice #6: Reserve Estimates

“It will be evident to the reader that no attempt has been made to provide specific instructions for the exploitation, development, or operation of actual oil-producing reservoirs. Nor has it been the purpose of the discussion presented thus far to provide explicit formulae for predicting quantitatively the recoveries from specific reservoirs or for evaluating them as items for sale or purchase. It has been an aim of this work to provide an exposition of the physical principles underlying the behavior of oil reservoirs so as to permit an understanding of their performance when observed in practice and an anticipation of the broad features of their performance from the consideration of basic data gathered during their development. Although these physical principles have been considered as well established and there is not definite evidence to the contrary, there are still many detailed aspects of reservoir phenomena in need of clarification and much work is yet to be done in correlating idealized theoretical predication and actual field observations before quantitatively significant formulae directly applicable to the complex systems comprising real reservoirs can be constructed. A tabulation of formulae presuming to give quantitatively accurate values of optimum development conditions or recoveries would, at the present time, be both premature and misleading.

From a practical standpoint it is the recovery factor or an equivalent measure of the ultimate oil recovery that is of primary importance. If this factor is not sufficiently high to pay for the cost of the drilling and operation of the producing wells, hypothetical considerations or predictions of the performance of the reservoir, if it were to be exploited, would be only of academic significance. It is true, of course, that the ultimate oil recovery is merely the abandonment value of the cumulative production. As such, it will depend on the nature of the production mechanism and the actual operating history. A prediction of its magnitude before a reservoir is fully developed and its exploitation program established would thus appear to be pure conjecture from a scientific point of view. Yet unless some estimate of the recovery can be and is made early in the development stage, it would be little more than an economic gamble to continue drilling.

The only solution to this dilemma evidently lies in experience and theoretical considerations. From the very first wells, estimates must be made of the reservoir content and the “probable” producing mechanism. A recovery factor is then applied such as experience has shown to obtain in other reservoirs with similar properties and that have produced by the same mechanism. This may be modified by supplementary considerations based on theoretical calculations, where feasible, of the probable reservoir performance and recovery. This is not an ideal solution. It is evidently subject to much uncertainty and fraught with the danger of erroneous assumptions regarding the actual recovery mechanism. Yet it is a necessary procedure if oil-field development is not to degenerate in to blind guesswork.”

From Physical Principles of Oil Production, by Morris Muskat, 1949

Sage Advice #7: Some Outstanding Problems in the Physics of Oil Production

“From the strictly scientific point of view the material presented in this work may well appear to be hardly sufficiently crystallized to be considered as the subject matter of a science. Much of the discussion has been clothed in multitudes of qualifying remarks, and many of the conclusions have been expressed merely as possibilities or at best as probabilities of occurrence. Virtually all the specific analytical and numerical considerations have been restricted to “ideal” systems, with the foreknowledge that these have never been observed to occur in practice.

Perhaps in no other science presuming to deal quantitatively with its subject matter is the latter so ill-defined as in the physics of oil production or reservoir engineering. Not only is each actual reservoir “specimen” in itself of almost infinite complexity, but the ensemble of all those already known and those which one may expect to be discovered contains no strictly “duplicate” samples. Virtually all experiments that may be performed on actual reservoirs not only are irreversible in the thermodynamic sense but are moreover essentially destructive of the specimen, with respect to the basic parameters defining its state prior to the experiment. Individual experimental observations involving extended periods of reservoir performance are not subject to repetition, either to test reproducibility or systematic cause and effect relationships. In its operational sense the “principle of uncertainty”, which is usually considered as limited to the realm of microscopic physics, constitutes the very essence of applied reservoir engineering as a science.

It is because of the inherent sample variability within the general class of oil producing reservoirs that their study does not lead to generalized and universally applicable quantitative solutions or conclusions. However frustrating the situation may appear to be, no two actual reservoirs will perform in exactly the same manner, nor will they react identically to operating control. Nor is this to be expected. As the detailed physical characteristics of the members within the individual major reservoir groups, whether classified according to recovery mechanism, performance, or structure, vary over wider ranges, so will the quantitative aspects of their performance and recoveries. Hence, except in referring to a particular reservoir whose properties are completely specified, general predictions of performance and recovery must of necessity be limited to probabilities of occurrence, and with wide ranges of possible deviations from any general expected trends. Any pretence that actual reservoir behavior will quantitatively follow universal functional histories, independently of their unique and individual characteristics, would simply deny the basic deterministic foundation of macroscopic physics.

The only alternative to hopeless attempts to treat quantitatively each specific known reservoir is the discussion of idealized prototypes. To give such treatments some degree of practical significance the numerical values of the parameters defining the particular system illustrating the analytical developments have been chose to lie within the range known to obtain in practice. Accordingly, in order-of-magnitude, the quantitative aspect of the behavior of the illustrative examples should agree with those observed in actual reservoirs as satisfy the broad physical assumptions underlying the analysis. However, it would still be purely accidental if there were exact agreement between the prototype idealized prediction and the corresponding observations in any specific reservoir.

Perhaps the outstanding unsolved problem in the physics of oil production is that of treating non-uniform flow systems. This is not merely the problem of averaging varying parameters, such as the permeability, connate water saturation, thickness, or ultimate recovery factors. Rather, it is the dynamical statistics of non-uniform reservoirs that would give an expression to the differential behavior of and interaction between localized regions of a common reservoir of different properties. Even the relatively simple case of multiphase transient flow in a stratified horizon has been given no satisfactory treatment in which the crossflow between the zones of different permeability is taken in account, although it is possible to construct the analytical formulation of this problem and solve the equations by laborious numerical procedures. It is because of the tremendous complexity of problems of this type that thus far for all reported studies of heterogeneous fluid systems have been restricted to idealized “uniform reservoir”. While it will probably never be feasible to gather sufficient data completely to define the nature of the heterogeneities in any specific reservoir, the basic reason for such discrepancies as occur between observed and calculated uniform reservoir performance will remain uncertain until some estimate can be made at least of the order of magnitude of the effects of the reservoir non-uniformity.

A rather disturbing situation that has been crystallizing in the last several years, as detailed reservoir data have been accumulated, is the increasing evidence that strict thermodynamic equilibrium is not always observed in reservoirs at the time of exploitation. Bubble point pressures at common datum levels within an interconnected reservoir are not always the same, within experimental errors, even in under-saturated reservoirs. And in reservoirs overlain by gas-caps, under-saturation has apparently been observed below the gas-oil contact. While many of these observations may be associated with changes in the character of the oil, that in itself is a manifestation of the apparent lack of equilibrium. In fact, not infrequently, decreases in API gravity with increasing depth are observed that are far larger than would be expected merely under gravitational equilibrium separation. The heavy and tarry oils often found near water-oil contacts undoubtedly are in neither diffusion nor gravitational equilibrium with the overlying lighter oils in the same formation. Presumably, such agents as may be causing local changes in the character of the oil have been exerting their influence at a more rapid rate than the thermodynamic potential gradients that would tend to establish substantial uniformity. Virtually as a matter of necessity, all these non-equilibrium phenomena have been almost completely ignored in the quantitative study of reservoir behavior. This problem, however, must be faced, and it merits serious study.

The clarification of the whole complex of details of fluid displacement processes still must be considered as an outstanding problem. Not only is the satisfactory treatment of the gravity drainage recovery mechanism beset with formidable analytical difficulties, but the estimation of the residual oil saturations and ultimate recoveries is almost in the state of speculation by analogy. The capillary phenomena in three-phase systems that determine the residual saturations left by gravity drainage, as well as the detailed structure of interphase transition zones, are in some respects uncertain even with respect to their qualitative features. The literature is completely lacking in three-phase permeability-saturation characteristics of consolidated porous media, although literally all gas-drive flow processes involve three phases. The commonly made assumption that the gas-oil permeability ratios are functions only of the total liquid content, and hence can be determined from two-phase measurements, certainly can have no universal validity. Much basic heterogeneous-fluid experimentation is urgently needed to definitely establish the quantitative aspects of the effects involved.

The detailed fluid dynamics and performance of limestones and dolomites are an almost totally unexplored phase of reservoir engineering. However, they deserve particular study, not only because they are of great practical importance as producing formations, but also because their homogeneities are of an essentially local character. While the gross reservoir performance of limestone reservoirs is in many respects substantially similar to that of sand formations, very little is understood of the internal averaging processes that apparently lead to such similarities.

The items discussed above do not indicate an order of importance or priority but are to be considered mainly as illustrative of the types of problems to which further research must be devoted. It is to be anticipated that as these are studied further, still others will be revealed which will demand additional investigation. And it may also be expected that the solution of these problems will lead to even stronger evidence than now available that each specific reservoir must be thoroughly studied individually in order to completely understand or predict its behavior. Although generalizations of physical principles will certainly be developed, their quantitative application in practice will undoubtedly always remain a matter of fitting the broad methods of analysis to the unique properties and characteristics of the particular reservoir of interest.

The physics of oil production is in no sense a “completed science”. The broad physical principles underlying the subject are reasonably well established. It has had sufficient application to justify faith in the soundness of its’ foundations. But its’ scope is far wider than it’s present development encompasses.

The elements comprising the subject matter of this science are so variable among themselves and the individual members – the reservoirs – are so extremely complex that it is very doubtful that it will ever achieve these state of complete description such as characterizes the classical mechanics or electromagnetic theory. According it is virtually impossible to crystallize the developments and discussions of its conclusions. Beyond the physical principles themselves, there is no universal validity to any quantitative assertion presuming to encompass even a restricted class of actual reservoirs. If treated statistically, the physics of oil production would be governed by a statistics of variables – such as the leaves of a tree – rather than identical particles such as gas molecules. Such conclusions as may be derived pertaining to reservoirs and their performance must be carefully qualified to emphasize that any generalization represents only possibilities or probabilities rather than well-defined descriptions of specific reservoirs that actually occur in practice. The probability is very low indeed that any hypothetical reservoir, constructed purely for illustrative purposes, will be duplicated by any reservoir in existence.

Aside from its inherent basic complexity, the physics of oil production is beset with many specific unsolved problems. One of the most serious of these is the treatment of non-uniform reservoirs. The gross statistical averaging into equivalent uniform systems undoubtedly suffices in many respects. Transient phenomena, however, which lead to differential depletion in parts of a common reservoir having different properties, cannot be represented so simply. Certainly the averaging processes will be different from those for the steady state dynamics. In any case, until at least a few simple non-uniform systems are rigorously treated, the quantitative significance of approximation and simplifying procedures cannot be evaluated.

Compared with other major sciences, the research effort thus far devoted to the physics of oil production has been almost infinitesimal. Nevertheless, it has been possible to develop the basic foundations for this complex subject, which already provide at least a semi-quantitative description and correlation of many of the major features of oil reservoir performance. Further effort on an expanded scale is certain to lead to an understanding of the physical phenomena of oil production which will compare favorably with that achieved in other applied physics sciences.

From Physical Principles of Oil Production, by Morris Muskat, 1949