The Reliability of Core Data as Input to Seismic Reservoir Monitoring Studies
- O.-M. Nes (SINTEF Petroleum Research) | R.M. Holt (NTNU and SINTEFPetroleum Research) | E. Fjær (SINTEF Petroleum Research)
- Document ID
- Society of Petroleum Engineers
- SPE Reservoir Evaluation & Engineering
- Publication Date
- February 2002
- Document Type
- Journal Paper
- 79 - 86
- 2002. Society of Petroleum Engineers
- 1.2.3 Rock properties, 6.5.2 Water use, produced water discharge and disposal, 1.6 Drilling Operations, 5.2 Reservoir Fluid Dynamics, 5.1 Reservoir Characterisation, 5.5 Reservoir Simulation, 4.1.2 Separation and Treating, 4.1.5 Processing Equipment, 1.6.9 Coring, Fishing, 1.14 Casing and Cementing, 5.3.4 Integration of geomechanics in models, 5.1.8 Seismic Modelling, 5.1.9 Four-Dimensional and Four-Component Seismic, 1.10 Drilling Equipment, 3.3 Well & Reservoir Surveillance and Monitoring, 4.3.4 Scale
- 1 in the last 30 days
- 244 since 2007
- Show more detail
- View rights & permissions
|SPE Member Price:||USD 5.00|
|SPE Non-Member Price:||USD 35.00|
There is a potential for improving the reliability of standard core tests for seismic monitoring studies. A primary concern is the ability to quantify and correct for core-damage effects, which significantly enhance the stress dependency of wave velocities. This problem is most relevant for relatively low-strength rocks cored in high-stress environments. We have used synthetic sandstones formed under stress to perform a systematic study of stress-release- induced core-damage effects. The results show that careful laboratory procedures and modeling efforts may reduce core damage effects. However, no simple procedure is currently available to eliminate the problem. The use of simplified laboratory test procedures, particularly the application of an inappropriate effective stress principle, may lead to erroneous interpretations.
Time-lapse (4D) seismic provides a potentially powerful tool to identify changes in a reservoir induced during production. This is accomplished by running repeated seismic surveys throughout the production period and looking for changes in the seismic response. Such changes can, in principle, be ascribed to several parameters, the most obvious being fluid saturation, pore pressure, and temperature. 1,2 Thus, by monitoring the reservoir at various timesteps during an enhanced oil recovery operation such as a water injection, one may identify nonflooded compartments within the reservoir. This information permits subsequent positioning of new production and injection wells or modification of the existing depletion strategy in a way that significantly improves the total recovery of the reservoir. During the past 5 years or so, the number of commercial 4D seismic surveys has increased from fewer than 5 to approximately 25 per year. The cost of a reservoir monitoring project is in many places comparable to that of drilling a new well, and benefits have, in many cases, proven so large that most companies now consider it a natural part of reservoir management.
There are, however, a number of factors that influence the success for such surveys, be they related to the reservoir itself in terms of depth, stress, temperature, and structural and compositional complexity, or to such intrinsic reservoir properties as the rock and fluid properties at the given reservoir conditions. The success also is affected by the quality of the seismic acquisition parameters during the surveys, such as the degree of repeatability between subsequent surveys,3 as well as the final processing of the seismic data (see Lumley et al.4 for a technical risk summary). Because of this substantial variability, one should always perform a seismic monitoring feasibility study in advance to quantify the extent to which expected production-induced changes may be detectable from a planned seismic monitoring study. Such a study needs integrated input from a number of disciplines; after a proper reservoir model is built, reservoir simulations must be undertaken to produce relevant scenarios to be expected throughout production. Thereafter, these must be translated into corresponding seismic parameters from rock physical principles before, finally, seismic modeling can be undertaken for various acquisition geometries and subsequent processing alternatives can be tested.
Traditionally, seismic monitoring parameters have been deduced from post-stack data through changes in the vertical P-wave reflection coefficient, expressed by the corresponding acoustic impedance ZP = ?·VP, where VP = the acoustic P-wave velocity and p=the density. This, essentially, has allowed for inversion for only one effective reservoir parameter. Knowing that there may be concurrent changes in several parameters has made the interpretation of the seismics difficult. More recently, however, a practical use of amplitude-vs.-offset (AVO) data has been introduced5 that enables the determination of the corresponding shear-wave impedance ZS. This simultaneous determination of P- and S-wave impedances has allowed for distinction between changes in multiple reservoir properties such as saturation and pore pressure, assuming that other parameters remain constant.
A crucial point in the initial feasibility study, as well as in the final interpretation of deduced changes in seismic parameters during monitoring, is the quantitative rock physical interpretation of the seismic parameters in terms of changes in reservoir parameters. A number of factors affect the acoustic velocities in a complicated manner, and no theory exists that can be applied generally. Therefore, laboratory testing on core material at representative test conditions is required as a natural part of a feasibility study to quantify the effects of pore pressure, saturation, and temperature that can be encountered during monitoring. The objective of this paper is to elucidate some fundamental questions related to these key issues. In particular, we focus on the neglected effect of core damage upon the laboratory-measured stress sensitivity of velocities6 and the importance of using proper stress conditions during such experiments. We handle this by performing systematic laboratory measurements on synthetic reservoir sandstones formed under stress, and we try to tune the properties of the synthetics to match specific reservoir sandstones. Even if this procedure is not fully representative of all reservoir sandstones, our experience is that it may at least be applicable for weakly cemented, clean sandstone reservoirs. Furthermore, we also illustrate pitfalls in the common use of the so-called effective stress principle.
As in all core testing, one has to deal with two fundamental questions when running experiments to quantify effects of pore pressure, saturation, and temperature on acoustic velocities:
Are the cores representative of the reservoir rock?
Are the tests performed under the appropriate conditions for prediction of in-situ behavior?
The first question has two different aspects. First, the small core may not be representative of a large heterogeneous reservoir. The most obvious way to deal with this is to test many cores and then perform some kind of statistical analysis on the acquired data. Still, the reservoir may contain fractures and faults at subseismic length scales, which are not present in the core samples but contribute to seismic velocities. The second aspect to consider is core damage: a rock is "born" and "lives all its life" in a stressed earth. When drilled and brought to the surface, it meets the hostile world of atmospheric conditions. The stress release may be sufficient to induce microcracks or broken grain bonds in the rock core, leading to altered rock properties. The damage is permanent and has been shown to have strong effects on rock mechanical and acoustic parameters.7 In the present paper, we discuss in more detail how core damage affects the predicted stress sensitivity of the seismic velocity.
|File Size||1 MB||Number of Pages||8|