SPE 89426 Adomian Solution of Forchheimer Model to Describe Porous Media Flow

Hadi Belhaj, SPE/Dalhousie Universiy, Jafar Biazar/ Guilan University, Stephen Butt, and Rafiq Islam, SPE/Dalhousie University

Copyright 2004, Society of Petroleum Engineers Inc. This paper was prepared for presentation at the 2004 SPE/DOE Fourteenth Symposium on Improved Oil Recovery held in Tulsa, Oklahoma, U.S.A., 1721 April 2004. This paper was selected for presentation by an SPE Program Committee following review of information contained in a proposal submitted by the author(s). Contents of the paper, as presented, have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material, as presented, does not necessarily reflect any position of the Society of Petroleum Engineers, its officers, or members. Papers presented at SPE meetings are subject to publication review by Editorial Committees of the Society of Petroleum Engineers. Electronic reproduction, distribution, or storage of any part of this paper for commercial purposes without the written consent of the Society of Petroleum Engineers is prohibited. Permission to reproduce in print is restricted to a proposal of not more than 300 words; illustrations may not be copied. The proposal must contain conspicuous acknowledgment of where and by whom the paper was presented. Write Librarian, SPE, P.O. Box 833836, Richardson, TX 75083-3836, U.S.A., fax 01-972-952-9435.

Darcys viscous term to correct that deficiency. Brinkman3 disagreed with Forchheimers modification and instead he added another viscous term to Darcys model. Others observed the inadequacy of Darcys model and they suggested that turbulence flow is taking place in porous media at high velocity flow causing deviation from the Darcian linear behavior. They determined that Forchheimer flow resistance and dispersion, which they believe it takes place at high Reynolds number, can be interpreted by using the concept of laminar flow theory4-9. Vafai and Tien4considered the Forchheimer flow resistance as an inertial effect of flow offered through the solid matrix from drag. In addition to these speculations there are many experimental studies10-16 in which the deviation from Darcys linear trend is observed at about Re =10 and not only the effect of Forchheimer flow resistance, but also the effect of the dispersion, gradually become predominant as the Reynolds number become higher. At high porosity porous media, Lee and Howell17 introduced another model that considers the same eddy viscosity for porous media as one which is commonly used for pure fluid. Travkin et al. 18 developed a model for high porosity media along with a statistical and numerical methodology. On the other hand, the argument that punched holes in the essence of adapting turbulence flow phenomena in porous media is the problematic determination of Reynolds number caused by lack of suitable definition to the length parameter. In addition, recent studies19 revealed that even if we consider the length parameter as the media length, diameter or the square-root of the permeability, Reynolds number proven to fluctuate rather than taking a definite value at the point of deviation from Darcys trend. Also, the deviation from the linear to the non-linear behavior in porous media is gradual unlike the sharp change from laminar to turbulence flow in the case of fluid flow in conduits. It has been believed that the main cause of deviation from the Darcian behavior is related to other factors different than those contributed to turbulence behavior takes place in flow between two plates or in a pipe. Judging from above, it seems reasonable to assume that non-Darcy flow is occurring at high spatial velocity in porous media as a consequent to a number of factors rather than attributing this behavior to turbulence phenomena. In this study, we shall consider a new formula that consists of Darcy viscous term, the other viscous term suggested by Brinkman and the inertial term introduced by

Abstract Currently, fluid flow in porous media is mostly calculated by utilizing the well known diffusivity equation based on Darcys law. This diffusivity equation is the core fluid flow equation in all modern reservoir simulators used to predict flow behaviors. Inaccurate predictions of reservoir simulators have been reported nevertheless, history matching has been achieved. This dilemma led to questioning the adequacy of the basic governing equation of fluid flow behavior in porous media. This paper suggesting a new governing equation that includes Darcys viscous term, Forchheimers inertial term and Brinkmans viscous term all in one model called the Modified Brinkman Model MBM. MBM proven to accurately describe fluid flow in porous media in both Darcian and non-Darcian domains and can be used in both oil and gas reservoirs for both matrix and fracture systems. A genuine mathematical solution Adomian decomposition technique has been successfully employed to solve the partial differential model with great deal of accuracy and ease. The proposed MBM is expected to have wide applications in oil, gas and underground water reservoirs. Introduction To date, the diffusivity equation derived from the famous Darcy law1 is in use to describe fluid flow in porous media. Darcys basic assumption at the time he introduced his equation is that the compressibility of both flowing fluid and the media are constant with low spatial velocity flow. Despite the wide spread and the modifications introduced to Darcys empirical equation to suite different flow schemes, this equation has been debated for over one and a half century. In 1901, Forchheimer2 depicted that flow in porous media is not always traveling at low velocity; he argued that Darcys model cannot describe such cases and added another inertial term to

www.petroman.ir

SPE 89426

Forchheimer and we shall consider Darcian behavior at low flow velocity and non-Darcian flow behavior at high velocity. The diffusivity equation model20 adopted to represent the type of behavior believed taking place in porous media is a complex partial differential equation that needed to be solved in order to predict flow behavior. This diffusivity equation which can be used to predict reservoir pressure at any point within the porous media and at any time can be expressed as:

p(x, y, t ) = pi + ax + Cxt
x t t

p dx + 0 y
x

p p 2p 2p + + + ( D ( ) dt E ( )dt)dx x y x 2 y2 0 0 0
Where:

(5)

p i = p (0, x, t ),

p 1 p 1 = v xt c x ck
2

p ( x, y,0) =0 y

(6)

2 v p v x 2 c
2

Adomian decomposition method considers the solution as the sum of the following series: (1)

p(r , z, t ) =

p
n=0

The mathematical Adomian Decomposition Method ADM 21,22 has been used in this study to solve the partial differential Equation (1) mentioned above. Adomian Solution and Results Equation (1) is called the Modified Brinkman Model MBM; detailed derivation of this equation is available in reference (23) 23. To solve the MBM equation by Adomian Decomposition Method (ADM), lets, for simplicity, consider:

Where p 0 , is defined to be as the terms on the right hand side of the canonical form (5), which are derived from initial conditions and those terms that do not depend on the unknown function p .

P0 = pi + ax + Cxt.
And the other terms of the series are defined as:

(7)

2v 1 v C= ( + ) c k

1 D= c

v E=

p n +1 = An ( p0 , ", p n )
n = 0,1,2, "
(2)

(8)

So equation (1) can be rewritten as:

2 p p 2 p =C+D +E 2 xt x x
Writing equation (2) in canonical form we get:

An ( p0 , p1 ," , p n ) so called Adomian polynomials, for an operator say G ( p ) , are defined as [*2].
Where An ( p 0 ,..., p n ) =
dn [G ( u i i )] d i=0 =0

Lp = C + D

p p +E 2 x x
2

(3)

2 1 Where L = , with the inverse L = tx

u 0

t x

(.)dt .

Using an alternate algorithm for computing Adomian polynomials [ ]we would have:
pn p p + C ( n + n )dt + y x y 0 ]dx t 2 2 pn pn 0 E ( 2 + 2 )dt x y 0
t

Applying the inverse operator to equation (3) we derive canonical for of the equation in the two following steps:
t p p p(x, y,0) p p = a + Ct + + D ( + )dt + x y y 0 x y

An ( p0 , ", p n ) =

[ x

(9)

E (
0

2p 2p + )dt x2 y2

(4)

To derive the solution we have to evaluate the constant parameters involved. The permeability K and the porosity are related empirically by the following correlation. Lets consider K = 100 md, so calculated accordingly using the correlation:

can be

www.petroman.ir

SPE 89426

ln(k) = 0.193 2.6649 13.80777202

= 5.181347150 ln(k) +

p1 = 6038.647345( 24154.58938v + k

The non-Darcy coefficient can be determined using the following correlation:

11500 10 3 k

0.2777777778 1012 v2 )t 2 x k(5.181347150ln(k) +13.80777202)

12077.29469 atx

If we assume water is the flowing fluid, then we can take the density and viscosity as: and = 1cp , the compressibility c can cm 3 5 1 be taken as: c = 8.28 10 psi Then, the correlation becomes:

p2 = 0.2431017450 10 8 ( 24154.58938 v + k )t 3 x

= 1 gm

11500000 k (5.181347150 ln(k ) + 13.80777202)

So the set coefficients C, and D would be evaluated

0.2777777778 1012 v 2 + k (5.181347150 ln(k ) + 13.80777202) 2 0.7293052351 10 8 at 2 x 2

as:
C = 0 . 2777777778 10 v 24154 . 58938 v k k ( 5 . 181347150 ln( k ) + 13 . 80777202 )
12 2

p3 =

E=

v k (5.181347150 ln(k ) + 13.80777202)

0.7340028535 10 11 ( +

24154.58938 v k )t 4 x

0.2777777778 10 12 v 2 k (5.181347150 ln(k ) + 13.80777202 ) 3

Considering these coefficients and using equations (7, 8 & 9) we calculate:

p 0 = p i + ax (

24154.58938v + k

0.2936011414 10 12 at 3 x 3
Finally, the pressure P can be predicted using only four approximation terms:

0.2777777778 1012 v 2 ) xt k (5.181347150 ln(k ) + 13.80777202)

p p 0 + p1 + p 2 + p 3 + p 4
For more accurate prediction of P we can calculate a sixth term or more same manner until reaching required accuracy.

www.petroman.ir

SPE 89426

Discussions The principal idea of addressing the non-linear non-Darcy behavior is to investigate the reasons behind this kind of behavior and to translate their effect to a mathematical formulation that can be used to represent such flow behavior in porous media. The fundamental assumption of constructing the basic flow equation is that the pressure drop in porous media is affected not only by Darcys viscous term contribution, but also, by both Brinkmans viscous term and Forchheimers inertial term. In its simplest form for onedimensional flow, such equation is taking the following form:

also, very interesting to notice that at low velocity, just after the flow starts, MBM predicts even less pressure gradient than Darcys model and then crosses with Darcys trend and begins to rise upward predicting higher pressure gradients. It is safe to say that Darcys model should no longer used to represent fluid flow in porous media; the suggested MBM is better alternative for its realistic prediction based on more comprehensive diffusivity equation. Conclusions This paper not only suggesting a better flow model that describes fluid flow in porous media in both Darcian and nonDarcian domains, but, it presents a new mathematical solution of the partial differential equation used by utilizing the Adomian mathematical technique. It can be clearly seen how much error that Darcys model produces as compared to the suggested MBM specially, at high flow velocity, which is the case in many reservoirs. This indicates that Darcys model should be replaced by MBM as governing flow equation in modern reservoir simulators. Adomian decomposition technique is very precise and easy to use and should be considered as the sole solution to the partial differential equation involved. It is expected that with the implementation of the proposed MBM and the suggested solution technique more accurate results would be produced and the faith in reservoir simulators predictions will increase. Acknowledgment Financial support of this research is funded collectively by the Natural Science and Engineering Research Council of Canada, the Atlantic Innovation Fund and the Killam Trustees. Nomenclature P Pressure x Space Coordinate in flow direction y Space Coordinate perpendicular to flow direction Non-Darcy coefficient t Time Viscosity of flowing fluid Fluid density v Volume-Averaged flow velocity in the x-direction c Isothermal compressibility factor K Permeability of the porous system Porosity References 1. H.P.G. Darcy, Les Fontaines Publiques de villle de Dijon, Exposition et Application des Pricipes a Suivre et des Furmules a Emplyer dans les Questions de Distribution dEau, Victor Dalmont, Paris, (1856). 2. P. Forchheimer, Wasserbewegung durch Boden ZVDI, 45, pp. 1781, (1901). 3. H.C. Brinkman, A calculation of the viscous force exerted by a flowing fluid on a dense swarm of particles Appl. Sci. Res. A, 1, pp. 27-34, (1947).

V P = V X K

2V X
2

(10)

The proposed model represented by Equation (1) in this study has been derived based on the above fundamental Equation (10). Figure (1) confirmed the steady state behavior of the predicted pressure within the porous system at any point along the system. This result agrees very well with previous numerical predictions19, it is expected that at the inlet of the porous system the pressure would be a maximum and as we travel towards the other end the pressure will decline linearly until reaching the outlet at which the pressure would be minimum. Figure (2) shows the performance of pressure gradient within the system with increase of flow velocity. At very low velocity we can assume linear behavior, where Darcys model can function with minimum error. But, as velocity increases non-linear behavior non-Darcy is obviously the predominant performance of flow and the discrepancy from Darcys trend is becoming larger. In Figure (3) the effect of permeability and consequently porosity related with an empirical correlation is examined. Permeability has prime effect on flow behavior; the plot indicates that as the permeability decreases, while other parameters are constant, the velocity will increase within the system. With the increase of velocity we see early break away upward from the Darcian linear trend. The effect of the kind of flowing fluid on the Darcian/non-Darcian behavior has been investigated by testing different viscosities and keeping other parameters constant. Figure (4) clearly implies that the less the viscosity (for example gas flow) the early occurrence of non-Darcy flow and the larger the deviation from Darcys linear behavior and vise versa. That explains the inadequacy of Darcys model to describe flow behavior in gas reservoirs and the adaptation of Forchheimers model as an alternative. Figure (5) compares MBM prediction with the traditional Darcys model prediction. This comparison show no agreement between the two models even at low velocity although, the difference in their predictions is very small at low flow velocity compared to that at high flow velocity. It is

www.petroman.ir

SPE 89426

K.Vafai, C.L. Tien, Boundary and inertia effects on flow and heat transfer in porous media , Int. J. Heat Mass Transfer, 24, pp.195203 (1972). 5. D.J. Gunn, Theory of axial and radial dispersion in packed beds, Trans. Instn. Chem. Engrs., 47,pp.T351-359 (1969). 6. D.L. Koch, J.F. Brady, Dispersion in fixed beds, J. Fluid Mech., 154, pp.399-427 (1985). 7. J.G. Georgiadis, I. Catton, An effective equation governing convective transport in porous media, J. Heat Transfer, 110, pp.635-641 (1988). 8. D.L. Koch, R.G. Cox, H. Brener, and J.F. Brady,The effect of order on dispersion in porous media, J. Fluid Mech., 200, pp.173-188, (1989). 9. C.T. Hsu, P. Cheng, Thermal dispersion in a porous medium, Int. J. Heat Mass Transfer, 33, pp15871597, (1990). 10. R.B. Bird, W.E. Stewart, E.N. Lighfoot, Transport Phenomena , Wiley, New York (1976). 11. S.Yagi, D. Kunii, N. Wakao, Studies on axial effective thermal conductivities in packed beds , A.I.Ch.E.J.,6, pp.543-546 (1960). 12. M.F. Edwards,J.F. Richardson, Gas dispersion in packed beds , Trans. Instn. Chem. Engng Sci. 23, pp.109-123 (1968).

4.

13. D.J Gunn, C. Pyrce, Dispersion in packed bed, Trans. Instn. Chem. Engrs., 47, pp.T341-350 (1969). 14. D.J. Gunn, J.F.C, De Souza, Heat transfer and axial dispersion in packed beds, Chem. Engng. Sci. 29, pp.1363-1371 (1974). 15. D. Vortmeyer, Axial heat dispersion in packed beds, Chem. Engng. Sci., 30, pp.999-1001 (1975). 16. N. Han, J. Bhakta, and R.G. Carbonell, Longitudinal and lateral thermal dispersion in packed beds, A.I.Ch.E.J., 31, pp.227-288 (1985). 17. K. Lee, and J.R Howell, Forced Convective and radiative transfer within a highly porous layer exposed to a turbulent external flow field, Proc. 2nd ASME/JSME Thermal Engineering Joint Conference, 2, pp.377-386 (1987). SI Metric Conversion Factors To Convert from To Ft m md m2 psi kPa kg/m3 lbm/ft3 lbm/ft-sec kg/m-sec

Multiply by 0.3048 9.869E-16 6.894757 16.01845 1.488163

www.petroman.ir

SPE 89426

www.petroman.ir

SPE 89426

www.petroman.ir