Impairment by Suspended Solids Invasion: Testing and Prediction
- Eric van Oort (Shell Research, Rijswijk) | J.F.G. van Velzen (Shell Research, Rijswijk) | Klaas Leerlooijer (Shell Research, Rijswijk)
- Document ID
- Society of Petroleum Engineers
- SPE Production & Facilities
- Publication Date
- August 1993
- Document Type
- Journal Paper
- 178 - 184
- 1993. Society of Petroleum Engineers
- 6.5.2 Water use, produced water discharge and disposal, 4.2.3 Materials and Corrosion, 2 Well Completion, 1.8 Formation Damage, 2.7.1 Completion Fluids, 5.3.3 Particle Transportation, 5.1 Reservoir Characterisation, 1.6.9 Coring, Fishing
- 4 in the last 30 days
- 932 since 2007
- Show more detail
- View rights & permissions
|SPE Member Price:||USD 10.00|
|SPE Non-Member Price:||USD 30.00|
A new model was developed to predict wellbore impairment by internal filer-cake formation during water injection. In the model, formation damage is approached semiempirically; no assumptions were made with respect to the specific mechanism underlying impairment. Model constants, derived from linear core-flow experiments, were used to calculate radial flow damage. Radial core-flow experiments subsequently were carried out to validate the proposed model. Theory and experiment were in good agreement. This paper shows that the model can obtain realistic performance and half-life estimates for water injectors. During the experiments, a pronounced velocity effect on impairment rate was found. The extent of impairment increased by several orders of magnitude when injection dropped below a critical inflow velocity of 2 cm/min. The generally accepted "1/3:1/7 rule" (the particle/pore ratio for internal cake formation) is confirmed for inflow velocities exceeding 10 cm/min. At low injection velocities (<2 cm/min), even smaller particles were found to cause increased damage: a "1/3:1/14 rule" may be more applicable for such low inflow velocities.
Reservoir pressure maintenance by water injection is common in secondary oil recovery. Efficient water injection is benefited by proper assessment of the ultimate performance and lifetime of water injectors. To make such an assessment, all factors relevant to injector performance and lifetime must be scrutinized, especially injection-water quality, chemical and physical interactions between the water and the formation, and injection mode.
Injection water may be taken from various sources (e.g., seawater, fresh water, or produced water) and may contain a variety of different particulate materials (e.g., formation particles, insoluble carbonates or sulfates, iron compounds, oil droplets, and bacteria). Deposition of these solid particles in the formation pores may cause well impairment, which will restrict an injector's performance and lifetime. To avoid well impairment, water may have to be treated before injection.
Some water-quality criteria - such as scaling tendency, corrosiveness, and bacterial content - are well-established. By contrast, criteria regarding the allowable size and quantity of suspended solids still are poorly defined. Although formation-damage prediction has been studied extensively,1-3 existing momdels for prediction of solids invasion still describe impairment inadequately.4,5
The objective of this investigation was to find a method for reliable prediction of formation impairment caused by suspended solids for water. If a semiempirical approach to formation damage is followed, no assumptions have to be made regarding the actual impairment mechanism. In general, this mechanism, which may involve various chemical and physical processes, is poorly recognized and difficult to quantify theoretically. The new model takes into account all relevant processes using simple numerical constants, which are derived from linear laboratory core-flow measurements.
According to Barkman and Davidson1 and Abrams,6 impairment by suspended solids invasion occurs as follows:
Particles larger than one-third the pore diameter bridge pore entrances at the formation face to form an external filter cake.
Particles smaller than one-third but larger than one-seventh the pore diameter invade the formation and are trapped, forming an internal filter cake.
Particles smaller than one-seventh the pore diameter cause no formation impairment because they are carried through the formation.
The shortcomings of this model are the lack of realistic values for particle invasion depth and internal filter-cake permeability data. Invasion depth, which is a function of velocity, pore size, and particle size, can be determined only by experiment. Davidson's7 results of a laboratory study on solids movement in porous media have served as a guide for estimating particle invasion depth.
Filter-cake permeability is determined either by a simplified Carman-Kozeny relationship8 or by Millipore™ experiments. While membrane filtration analysis is an attractive way to describe the formation of (external) cake, no obvious correlation exists with particle deposition in rock.
In his expansion of the Barkman-Davidson momdel, Eylander9 maintained the assumption that invading particles are deposited in the pore system some distance from the wellbore and form a filter cake that grows toward the wellbore. His core-flow results, however, implied that internal filter-cake formation starts at the injection face and progresses away from the wellbore.
The Rege-Fogler2 model is rather advanced in accounting for some of the many factors affecting particle capture: geometric sizes, fluid velocity, and deposition morphology. A higher value of the capture probability at reduced fluid velocities was predicted. We also have observed this important effect in our laboratory experiments. According to Rege and Fogler, model predictions were in good agreement with experimental data.10-12 More data and information, however, are required to validate the Rege-Fogler model fully.
Simulations based on the Todd et al.13 model did not match their experimental results. Core damage decreased with an increase in core depth. Smaller particles, however, damaged the cores throughout their entire length; larger particles damaged mainly the first core section. Kumar and Todd14 presented a more comprehensive approach to formation damage modeling. Their linear model predicted the results of the coreflood experiments reasonably well. The permeability damage profile predicted by the radial model, however, is not yet verified by experimental data. A comparison between the data provided in Refs. 13 and 14 and the data from our experiments evinces a similar trend for both data sets.
In the Sharma-Yortsos3 model, the basic mechanisms for particle capture and release are modeled with population balances. Apart from a single paper12 in which "semiquantitative" agreement between theory and experiment is claimed, no other experimental evidence has been reported to support their model.
The van Velzen Model
A mathematical model that describes the reduction in injectivity from internal filter-cake formation under radial flow conditions was developed. The complexity of the problem and the diversity of reservoir rocks and injection waters favors a semiempirical approach. Particles removed from the fluid by porous media are deposited in the matrix pores, resulting in an increasing fluid flow restriction. Using Darcy's law for radial flow, a material balance between the solids in suspension and in the internal filter cake, and a modification of Iwasaki's15 relationship for deep-bed filtration, we can calculate the injectivity reduction of a water injector and the invasion depth of the solids. The Appendix gives a derivation and a more detailed discussion of the model.
|File Size||1 MB||Number of Pages||7|