Energy 66 (2014) 868e880

Contents lists available at ScienceDirect

Energy
journal homepage: www.elsevier.com/locate/energy

Antenna array design for enhanced oil recovery under oil reservoir
constraints with experimental validation
Muhammed Moshin Abdulrahman, Mahmoud Meribout*
Electrical Engineering Department, Petroleum Institute, Abu Dhabi, United Arab Emirates

a r t i c l e i n f o a b s t r a c t

Article history: In this paper, a detailed mathematical model of EM (electromagnetic) heating for EOR (enhanced oil
Received 21 July 2013 recovery) is presented. The model takes into consideration three different designs of the array of EM
Received in revised form sources under the constraint of the dimensions, dielectric constituents, and geophysics properties of oil
5 December 2013
reservoirs. This includes a simple waveguide antenna, an E-plane spectral horn antenna of different
Accepted 1 January 2014
dimensions and an array of waveguide antennas with and without deflected phases. Simulation studies
Available online 24 February 2014
have been performed at frequencies of 915 MHz, 2450 MHz and 5800 MHz for the three aforementioned
designs under different combinations of the placement, positioning, and orientation of antennas.
Keywords:
Enhanced oil recovery
Furthermore, an experimental verification of the suggested model is addressed by designing and building
Electromagnetic heating a new experimental setup which has the advantage to mimic more realistically real oil reservoirs as
Microwaves antennas compared to microwave ovens. These latest which were mainly used as the test platform by other re-
Electromagnetic propagation searchers feature substantial differences with oil field reservoirs. The corresponding results indicate a
good matching (i.e. 95% matching) with the suggested model, while outperforming other traditional
enhanced oil recovery techniques. This leads to state that EM-based EOR can be a good alternative to the
other existing techniques.
Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction reduction in permeability of formation [3]). This recently raises the

interest for other recovery methods, namely EEOR (Electrically
In oil industry, high viscosity of heavy oil remains a major Enhanced Oil Recovery) techniques which include EM (electro-
concern in the area of enhanced oil recovery. Different types of magnetic) and ultrasonic waves, inductive heating, and DC (Direct
conventional EOR (enhanced oil recovery) techniques are being Current) heating [3]. The main advantage of EEOR techniques such
used for increasing the production of oil. This includes CO2, steam, as EM-based techniques is that they require less amount of energy
and chemical injections and water flooding. Each of these tech- than conventional EOR. This is because the propagation of EM
niques helps increasing the static pressure of the oil reservoir waves can be tuned towards a specific region of interest, while the
pushing the residue oil to move away from the trapped regions propagation of fluids such as steam and CO2 depend primary on
[1,2]. In order to enhance the efficiency of these techniques, lot some physical properties of the reservoir such as pressure and/or
of related research efforts continue to be conducted mostly in heat distributions. The ability of controlling the propagation of the
modeling and software simulation using some commercially energy is desirable in the usual case when a specific region of the
available software packages [1]. Joined and persistent research reservoir needs to be recovered. Taking into consideration the ad-
efforts have allowed these software packages to be mature enough vantages of EM heating of water-depleted oil reservoirs which
by incorporating the majority of the oil reservoir characteristics constitute a substantial amount of world-wide oil reserves, this
(e.g. permeability, porosity, and oil viscosity). In spite of some paper investigates the feasibility of using this method for EOR. In
encouraging results obtained in oil fields, these techniques have the this paper, EM heating refers to microwave heating where heating
disadvantages of not being environmental friendly, in addition of is produced at high frequency.
requiring costly storage tanks and transportation infrastructure
[2]. Furthermore, for long term, these techniques may seriously 2. Literature review on EM heating for EOR & work
damage the formation (e.g. corrosion of sub-surface devices and motivation

* Corresponding author. Electromagnetic waves are transmitted, absorbed, or reflected

E-mail address: [email protected] (M. Meribout). through/by a given material. They usually interact with dipoles and

http://dx.doi.org/10.1016/j.energy.2014.01.002
0360-5442/Ó 2014 Elsevier Ltd. All rights reserved.
 M.M. Abdulrahman, M. Meribout / Energy 66 (2014) 868e880 869

free or bound charges such as electrons or ions. Hence, when EM heating as compared to conventional EOR techniques [9]. For
waves are applied, these polar molecules are set into circulatory instance, it is not uncommon that the water in field reservoirs is
motion and they collide speedily with other molecules with the salty and has a non-negligible concentration of ions which has the
collision frequency equal to the frequency of applied EM waves advantage to increase the absorption power of the reservoir. In
which in result produces much heat. These properties were addition, some previous works [8] have suggested that the pres-
extensively explored in the area of biomedical engineering for ence of wet sand enhances the efficiency of EM heating by around
treating cancerous cells. This led to achieve some interesting results ten (10) times as compared to conventional heating such as steam
which allowed the emergence of several mature and commercially- heating. The same results have revealed that in such scenario, a fast
available medical instruments which are used in several parts of temperature increase to up 300e400
C could be achieved. These
the world. Similarly, some works related to bio oil production used features have led many scientists, mainly petroleum engineers, to
microwave heating as a mean of liquefaction to convert biomass develop models for EM heating for heavy oil recovery [8].
into energy and high value chemicals [4]. Hence, microwave heat- In Ref. [10], a two-dimensional (2D) model for EM heating of
ing has the advantage that it is fast and is based on volumetric heavy oil was suggested and solved using two FEM (Finite Element
heating. Experimental results indicate that using 600 W microwave Method)-based simulators (e.g. COMSOL Multiphysics and STARS
power for a duration of 30 min, a liquefaction yield of up to 93.17% simulators) at 915 MHz frequency. The results of both simulations
was obtained under optimum conditions (e.g. upon the catalyst were in concordance and reiterated the merit of using microwaves
concentration). A second order polynomial model was developed to for EOR. Another mathematical study was performed at 915 MHz
validate the experiments and assess the potential of the technique for microwave heating of heavy oil and the results were compared
at larger scales. A similar work was done in Ref. [5] to improve with steam heating. It proposed a comprehensive 2D mathematical
biodiesel yields from waste cooking oil using microwave based- model but lacked the experimental verification of the model and
heating. The experimental setup was very similar to the one used consideration of different electromagnetic source scenarios such as
for EOR and consisted of a microwave unit (750 W) power within design, type and orientation of microwave antennas [11]. In Ref.
which a sample of cooked oil is introduced. Experimental results [12], another interesting model which simultaneously takes into
indicate a 97.7% maximum yield using microwave heating as consideration the electromagnetic and dynamic heating phenom-
compared to 96.6% maximum yield with conventional heating with enon of microwave-based heating is proposed. Even though this
65
C heat temperature. Furthermore, the reaction was much faster model is dynamic in nature i.e. the dielectric properties of material
in case of microwave. In Ref. [6], application of microwave onto a are being updated with the rise in temperature, it does not incor-
heat circulator was successfully done using a new AMR (active porate the fluid viscosity reduction and consequently the fluid flow
magnetic regenerative heat) and the corresponding experimental which remains a main area of concern for microwave heating of
results validated the developed model. A theoretical study on the heavy oil. In addition, they did not consider the permeability,
use of microwaves for endothermic reactions was suggested in Ref. porosity and electric absorption coefficients of the reservoir oil as a
[7]. This efficiency can be even higher when the reactor is coated function of the frequency. In Ref. [13], authors suggested another
with a metallic material having a depth function of the frequency of mathematical model for microwave heating of heavy oil but did not
the microwave radiations. The model comprised a detailed Helm- incorporate the permeability as a function of the force field in the
holtz equation for microwave propagation within a semi infinite reservoir which is an important parameter for fluid flow calcula-
batch reactor. Simulation results indicate that the use of microwave tions. In addition, they did not investigate the effect of microwaves
can significantly save energy for up to 60%, dependently on the kind on the pressure distribution within the reservoir and the thermal
of reactor being used. Other works with EM waves have also been conductivity of heavy oil with temperature rise. The work done in
conducted for EOR [8]. This was motivated by the fact that EM Ref. [13] constitutes one of the rare contributions which simulta-
waves do not present any risk of chemical reactions with other neously considered modeling and experimental validation. Hence,
materials found inside the reservoir [8]. In addition, in practice, the authors have suggested a detailed two-dimensional model of
several scenarios may enhance the effectiveness of using EM microwave heating of heavy oil with an operating frequency of

Fig. 1. Design of oil wells for E. (a) Top view of the reservoir, (b) front view of the underground of the reservoir.
870 M.M. Abdulrahman, M. Meribout / Energy 66 (2014) 868e880

Fig. 2. Block diagram of the suggested model considered for the microwave-based EOR.

915 MH using an FEM-based simulator (STARS). In Ref. [14], another beginning of the microwave heating process and employ slow
model of 2.45 GHz microwave heating was disclosed and few ex- heating after reaching a threshold temperature. One of the limita-
periments using a microwave oven were done in an attempt to tions of these studies is that the core samples are dipped inside a
validate the model. A good matching between theory and practice glass beaker which was then placed inside a microwave oven which
was revealed and the study suggested using fast heating at the does not correctly represent actual oil reservoirs. In Ref. [15],

Fig. 3. An explanatory view of microwave antennas: (a) 3-D single microwave antenna, (b) 3-D E-plane spectral horn antenna, (c) 2-D view of horn antenna, (d) an array of three in
phase, (e) microwave wave pattern.
 M.M. Abdulrahman, M. Meribout / Energy 66 (2014) 868e880 871

another experimental study not based on microwave oven is pre- Table 1

sented. It consists of an RF (Radio Frequencies) curve transmitter Reservoir and electromagnetic simulation data.

which continuously emits electromagnetic waves towards a core Properties Value

sample. The performance of the system under different positions of Frequency 915 MHz, 2450 MHz
the core sample and number of electromagnetic sources was Input power 50 kW
investigated. The corresponding results indicate that three elec- Heating time 01 year
tromagnetic sources could lead to the best recovery and suggest API gravity of crude oil (at 20
C) 15
Initial temperature 55
C
that core samples with low porosity and permeability produced
Initial pressure 6.89 MPa
more oil as compared to the ones with high porosity and perme- Mean heat capacity of oil at constant 2.13 kJ/(kg K)
ability. However, the main limitation of this work is that the elec- pressure
tromagnetic source was operated by a function generator which is Heat transfer coefficient of rocks 1 mW/(m K)
Reservoir dimensions 10 by 3 (meters)
not realistic for reservoir applications.
Heat source dimensions Cut-off freq. based
In summary, it can be concluded that most of these research No. of MW (Micro Waves) 3 with each one of them has output
works considered the microwave source and the oil reservoir at an antennas in array power of 1667 W
abstract level and ignored some realistic and decisive constraints Porosity of reservoir 23%
regarding their applications. This includes the available space for Propagation constant 0.542 (915 MHz), 8.09 (2.45 GHz)
Penetration depth 2.62 m (915 MHz), 1.33 m (2.45 GHz)
EM (microwave) sources in the field reservoirs (e.g. diameter and
Permeability of reservoir 1000 mD
height of the well within which microwave sources are to be
inserted) and the risk of high increase of temperature of the mi-
crowave antenna which may damage the source. Thus design of experimental setup which mimics the actual oil reservoirs more
antenna is based on the cut-off frequency and is limited by the accurately than microwave ovens. The corresponding results
diameter of injection well as well as a proper cooling of the mi- demonstrate the validity of the newly introduced model.
crowave source is required to maintain its temperature at an
acceptable level. In addition, none of the earlier works addressed 3. Technical background
different design alternatives, such as the type and arrangement of
different type of microwave antennas. Indeed, they mainly With the continuous depletion of oil reservoirs, a non-negligible
considered microwave oven with an operating frequency of quantity of remaining oil becomes either highly viscous or features
2.45 GHz as the testing platform. The significant economical benefit low pressure. This makes the residual oil easily trappable between
of the EOR added to the preliminary positive results obtained with rocks and thus hard to extract using conventional primary and
microwave-based EOR suggests the importance of such design secondary recovery techniques (i.e. recovery using submersible
alternatives while considering realistic constraints and adequate multiphase flow pumps). With the tertiary enhanced oil recovery,
experimental setups. Hence, microwave ovens may not be consid- one or few wells are dug within an oil reservoir within which one or
ered as a feasible setup since microwaves are continuously re- few operate as producing wells and one of few as injection wells.
flected between the internal metallic walls of the oven, while they Fig. 1 shows a simplified layout of an oil reservoir with three in-
tend to be rather absorbed or vanished in real oil reservoirs within jection wells and one producing well in both the top and front
the water or oil phase. If such mode of microwave propagation is views (Fig. 1(a) and (b)). Without loss of generality, in some other
not provided in the experimental setup, the validation of the situations one well can be simultaneously a producing and injec-
theoretical model cannot be sustained and consequently the tion well. Usually the diameters of these wells vary from 50 to
effectiveness of using microwave for EOR can only be partially 92 cm, while their height can reach up to 4 km. This constitutes one
appreciated. of the constraints to consider during the design of the source of
This article presents a comprehensive mathematical model for energy since the well is the only accessible space for any EOR
microwave heating of heavy oil, which takes into consideration technique. The orientation of antenna is another important
different physical properties of the reservoir such as the specific parameter as any deflection in the direction of the microwave an-
heat capacity of the heavy oil at constant pressure and the imagi- tennas may cause a substantial attenuation of microwaves leading
nary part of the complex dielectric constant of the medium. The to a loss of energy.
model also includes the thermal conductivity, density, dynamic The aim here is then to insert into the exiting well a suitable
viscosity and absorption coefficients of the medium for absorbing type of microwave sources to effectively heat the oil reservoir at a
microwaves. As another contribution in the field, the paper sug- specific height (e.g. oil layer in Fig. 1) and thus reduce its viscosity or
gests different type of microwave antennas which include simple increase its pressure at a value allowing its displacement towards
waveguide antenna, microwave horn antenna, and an array of the producing well.
microwave antennas. The operating frequencies for all of the
aforementioned antennas are set to 915 MHz, 2.45 GHz, and 3.1. Mathematical modeling
5.8 GHz and the resulting temperature and pressure distributions
along with the oil recoveries are compared and discussed. Even One of the contributions of this paper is that it suggests a
though, some of these frequencies overlap with the frequency comprehensive model which considers all the physical parameters
spectrum which is allocated for mobile communication, they are of the oil reservoir (e.g. dielectric constant, thermal conductivity,
not expected to cause any kind of interferences since the corre- electric conductivity, density, dynamic viscosity, and permeability).
sponding microwave source operates down hole at more than few This would allow a better assessment of the effects of microwave
hundred meters with a radiation beam focused in the horizontal sources for EOR. Fig. 2 illustrates the flow chart corresponding to
direction. In addition, and similarly to other preliminary research the microwave heating of heavy oil within a reservoir.
activities (i.e. wireless sensors kits) where busy frequency spec- Initially the microwaves being absorbed by the target fluid
trums are used, the actual study uses commercially available high dissipate energy resulting in heat distribution which causes rise in
power microwave sources to focus on the validation of the model temperature along with its diffusion inside the oil. Some electro-
and to avoid building the hardware from scratch. Finally, extensive magnetic and thermal properties of oil are periodically updated
experiments were conducted on a newly designed and safe depending on the temperature distribution. These updated
872 M.M. Abdulrahman, M. Meribout / Energy 66 (2014) 868e880

Fig. 4. Schematic view of microwaves heating for heavy oil [6].

properties are in their turn used to determine the new heat energy equation, the first, second, and third terms represent the heat loss
distribution. In the flow chart, the dielectric constant, ε, for the by convective heat transfer to the surrounding environment (where
heavy oil is composed of two parts: the real part (ε0 ), also called the h represents the coefficient of convective heat transfer in W/m2),
permittivity and the imaginary part (ε00 ), also called the dielectric radiative heat loss from the wall surface (where x represents the
loss factor. Hence, the variable ε0 deals with the storage capacity of emissivity of oil (taken as 0.95 in this case [10])), and the heat
electromagnetic energy inside the material while ε00 effects the accumulation in the reservoir. The other variables are s which
dissipation of electromagnetic energy per unit volume, Pv, in the define the electrical conductivity of the fluid, r the density of oil,
form of heat energy [16]: and Cp the heat capacity of oil. It is worth noting here that the
contribution of the third term is much more significant that the one
Pv ¼ uε0 ε00 Ev2 (1) of the first and second terms. The term on the left hand side of
Equation (2) can also be defined [18] as follows:
where f ¼ (u/2p) is the frequency of operation and Ev the strength
of electric field inside the material. aP0 eaðrrw Þ
V$qEM ¼  (3)
r
3.1.1. Heat generation and distribution model
In microwave-based EOR, the energy can be described in the where Po is the power applied at the well bore, a the electromag-
form of per unit volume of oil, V, as: netic absorption coefficient, r the radial distance, and rw the well
Accumulation of energy ¼ Net energy transfer þ energy input bore radius. Hence, the heat absorbed or gained by the material
from the source. depends upon the absorption coefficient, a which also depends on
Mathematically, for a microwave emission into a reservoir of the electrical properties of the medium [18]:
area A, actual temperature T, and ambient temperature Ta, the rffiffiffiffiffiffiffiffivffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
u rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

s 2
ffi !ffi
above equation can be written as [17]: ε0 m 0 ut
a ¼ u 1þ 0 1 (4)
2 εu
hA xAs
4  vT
V$qEM ¼ ðT  Ta Þ þ T  Ta4 þ rCp (2)
V V vt Also, in Equation (2), the temperature distribution, vT/vt, is given
The left hand side term, V$qEM , defines the total energy pro- by the heat transfer equation [18]:
vided by the electromagnetic source. In the right side of the
vT Pv ðx; y; tÞ
rCp ¼ VðkðTÞ; VTÞ þ (5)
vt 2

where the constant Cp is the specific heat coefficient.

3.1.2. Power distribution model

The rise in temperature directly affects the thermal conductivity
and pressure distribution in the reservoir as well. The relationships
of temperature with thermal conductivity K(T) and pressure P are
derived by using the correlation function and a fourth order poly-
nomial and are given as follows [18]:

KðTÞ ¼ 0:135  2:5  105 T (6)

P ¼ 2E09 T 4 þ 5E06 T 3  0:0129T 2 þ 7:2355T þ 1299:4

(7)
While the thermal conductivity is used to compute the heat
transfer (Equation (5)), the pressure is used in the fluid flow model
Fig. 5. Microwaves heating of heavy oil (full model). (see the next section). To the best of our knowledge, it is the first
 M.M. Abdulrahman, M. Meribout / Energy 66 (2014) 868e880 873

Fig. 6. Temperature and pressure distributions for microwave array of antennas operating at 2.45 GHz frequency (dynamic electromagnetic and thermal properties): (a) tem-
perature distribution, (b) pressure distribution.

time that this model is considered for microwave-based EOR. Where i ¼ 1, ., M and j ¼ 1, ., N. Assuming no dispersion and
Though, some other researchers have used a part of the model for no adsorption i.e. kij ¼ wis ¼ 0 and only one component is present
microwave heating of food and microwave treatment for diseases in each phase i.e. wij is either 0 or 1. The mass balance equation can
such as cancer. be written as [6]:
0 1
 !
v@ X X
Np Np
3.1.3. Dielectric loss factor updating models kkrj v2 Pj v2 Pj v2 Pj vg vg vg
B A
rj Sj ¼ rj þ þ þ rj þ þ
The dielectric loss factor in a medium with conductivity sigma vt mj vx2 vy2 vz2 vx vy vz
(s) depends upon temperature and is given by Ref. [18]: j¼1 j¼1

(9)
ε00 ¼ sðTÞ=2pf ε0 (8)
 where B is the porosity of the reservoir, rj and Sj are the density and
where sðTÞ ¼ ðsT¼T Þ½1 þ a1 ðT  Ta Þ þ a2 ðT  Ta Þ2 þ a3 ðT  Ta Þ3 , saturation of phase j. In addition, mj is the viscosity of phase j, k the
a
values of a1, a2 and a3 are taken from Ref. [19] and are: permeability of the reservoir, Pj the pressure in phase j, while krj is
a1 ¼ 0.02604/C , a2 ¼ 0.01312/Co and a3 ¼ 0.00997/Co.
o
the relative permeability of phase j. This variable can also be given
by Ref. [20]:
3.1.4. Mass balance and fluid flow model


qj mj
A second major consideration for microwave heating of heavy
A  k
oil is the mass balance which has been considered by only few krj ¼  
 (10)
vx þ vx þ vx þ rj F 
vP vP vP
researchers [6]. The mass balance equation is used to determine the
flow of the fluid by considering the component ‘i’ for each phase ‘j’
incorporating the porosity and permeability of reservoir along with here qj represents the flow rate of phase j, A the area, and k the
the viscosity of different phases of fluids and is given by: permeability of the reservoir. On the other hand, the variables mj
Accumulation of ‘i’ in phase ‘j’ ¼ net mass transfer of ‘i’ in phase and rj define the viscosity and density of the phase j, while g is the
‘j’ þ mass input of ‘i’ into phase ‘j’. gravitational acceleration. In this case the body force field is due to

Fig. 7. Temperature and pressure distributions for microwave array of antennas operating at 2.45 GHz frequency (static properties): (a) temperature distribution, (b) pressure
distribution.
874 M.M. Abdulrahman, M. Meribout / Energy 66 (2014) 868e880

Fig. 8. Comparative analysis of temperature and pressure distributions for different type of MW antennas.

the gravitational acceleration and can be found from density and two known temperatures. The remaining variable, T(x), is the
volume of fluid particles as given below: temperature at position, x [8].

pd3 4. Antenna design for enhanced oil recovery

F ¼ rf g (11)
6
4.1. Overview
The fluid flow is calculated by using the Darcy’s law:
As was mentioned earlier, to our knowledge, there is no litera-
k dP
V ¼  (12) ture which has addressed the effects of different microwave oper-
m dL
ating frequencies and type of antennas on the temperature
distribution for EOR. using different type of microwave antennas
where m is the viscosity, k the permeability of the reservoir, while dP
e.g. microwave antennas with waveguides, horns or an array of
and dL are the differential pressure and length. Darcy’s law also
microwave antennas with waveguides may play a vital role in
provides the volumetric flow rate by using the area ‘A’ in the above
efficient heating of the heavy oil. Single antenna is suitable for small
expression leading to the volumetric flow rate q ¼ k  A/m  dP/dL.
reservoirs with less area to be heated while for large reservoirs
Once the velocity of fluid flow is known, the drop in pressure can
array of antennas would be more feasible. Fig. 3 shows the
be calculated by making use of this calculated velocity and well
simplified geometries of simple waveguide antenna and E plane
bore pressure (which is already defined) as follows:
horn antenna which have been considered in this paper. The reason

m for considering horn antennas in this paper, in spite of their large
P2 ¼ q  dL  þ P1 (13) size, is that they can provide an effective heating capability, in
kA
addition to be relatively cheap. Hence, a special well bore for the
where P1 indicates the well bore pressure and P2 depicts the new injection well, with eventually larger dimensions, can also be
pressure of the reservoir after having the volumetric flow rate, q. considered to host a horn antenna. This will ultimately lead to a
In addition to the pressure distribution, the fluid flow also de- raising the cost for the construction of the well bore; however this
pends also on the viscosity reduction. Hence, the temperature and can be compensated in the long term with a more effective
viscosity relationship are related according to the following equa- enhanced oil recovery. In Fig. 3, the dimensions of waveguide an-
tion [15]: tenna are mentioned as ‘a’ and ‘b’ while the E-plane horn antenna
has more parameters which depend upon the waveguide param-
mðxÞ ¼ DeF=TðxÞ (14) eters ‘a’ and ‘b’. These parameters include the length (R), width (A),
height (B) and flare angle ‘q’ of horn antenna. Hence, in addition to
where F and D represent the empirical constants which are deter- the flare angle, the dimensions of horn antenna also affect the ra-
mined by the regression analysis of the oil viscosity measured at diation pattern.

Fig. 9. Comparison between oil recoveries from microwave heating with different antennas and steam EOR.
 M.M. Abdulrahman, M. Meribout / Energy 66 (2014) 868e880 875

Fig. 10. Comparison between temperature and pressure distributions with different type of antennas at 915 MHz operating frequency.

For all above antenna, the power level, standing time, cycling, ‘a’ and ‘b’ are the dimensions of the waveguide (as mentioned in
applied frequency, dimensions and antenna position and orienta- Fig. 3 above) and ‘c’ stands for the speed of light given as 3  108 m/s.
tion etc. are found as the main factors influencing the microwave For the case of 2.45 GHz, the exact cut-off frequency is found to be
heating procedure. The issues of non-uniformity in heating and 2.419 GHz while for 915 MHz it came out to be 914.63 MHz. The
inefficient heating are resolved by applying cyclic microwave propagation constant of microwaves is calculated by using the
heating instead of its continuous application. Although increased equation given below:
applied power level of microwaves reduces the heating time, it may qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
result in a salty shielding causing an even increase of the required 2p
Propagation constant ¼ b ¼ f 2  fc2 (16)
exposure heating time [15]. c

here ‘f’ stands for the frequency of operation of microwaves. Note

4.2. The microwave antenna design for EOR
that the values of dimensions of the microwave antenna for
5.8 GHz, 2.4 GHz and 915 MHz are small enough to be introduced
The design of microwave antennas depends upon many param-
easily inside the well bore.
eters e.g. the cut-off frequency & propagation constant of corre-
sponding propagating waves. This section addresses how these
b. The target penetration depth
parameters have been considered in our design process of the mi-
crowave antenna.
Depending upon the target penetration depth in oil reservoir, a
specific cut-off frequency may be selected and hence its dimensions.
a. Dimensions and cut-off frequency
The penetration depth can be given by the following equation [12].

Microwave antenna with the dimensions of 62 mm by 31 mm pffiffiffiffi

c ε0
(which is a waveguide antenna in this case) are used for operating Penetration depth ¼ Dp ¼ (17)
2pf ε00
frequency of 2.45 GHz. Similarly, the dimensions of an antenna for
an operational frequency of 915 MHz are 164 mm by 82 mm. These here ε0 is the permittivity, f is the frequency in Hz, c ¼ 3  108 m/s
dimensions are chosen based on the cut-off frequency of the and ε00 is the imaginary part of the dielectric constant i.e. dielectric
waveguide. This latest is the frequency below which the attenuation loss factor.
increases sharply and can be calculated by using the formula [6]:
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
c
m2
n2 5. Numerical simulations
fc ¼ þ (15)
2 a b
COMSOL Multiphysics was extensively used to solve several two
here ‘m’ and ‘n’ represent the mode of electromagnetic waves’ dimensional models because of its ability to handle multiphysics
propagation which is TE10 in this case (i.e. m ¼ 1 and n ¼ 0). while environments using FEM-based analysis approach. Taking into

Fig. 11. Comparison of oil recovery with different microwave antennas operating at 915 MHz and steam EOR.
876 M.M. Abdulrahman, M. Meribout / Energy 66 (2014) 868e880

center of the reservoir. A complete model of the reservoir is shown

in Fig. 5.
The Coupled model of electromagnetic heating with the heat
transfer and Darcy’s flow was solved in COMSOL according to
Equations (1), (13) and (14).

5.2. Simulation results with 2450 MHz antennas

Initially, microwave frequency of 2.45 GHz has been chosen as

the operating frequency and the model has been solved considering
simple magnetron source. Other simulations were performed using
an array of antennas in phase with each other and two E-plane
microwave magnetron antennas with different heights of the horns
i.e. 40 and 30 cm. The temperature and pressure distributions ob-
tained from the array of magnetron antennas are provided below in
Fig. 6(a) and (b). The Figures clearly show a significant temperature
rise and increase in pressure in the area close to the well bore. High
temperature rise generates high pressure because of evaporation of
low hydrocarbons therefore temperature and pressure distribu-
tions are similar to each other. Figs. 6 and 7 provide the tempera-
ture and pressure distributions with the use of an array of three
antennas operating at 2.45 GHz frequency. These plots are provided
for only array of antennas because the resulting temperature and
pressure distributions are significantly high as compared to other
type of antennas. Fig. 6 provides the temperature and pressure
distribution for the case of having dynamic electromagnetic and
thermal properties of oil i.e. properties of oil change with the
temperature. On the other hand, Fig. 7 shows the temperature and
Fig. 12. Temperature distribution in crude oil as a result of MW heating, (a) simple
pressure distributions for static thermal and electromagnetic
waveguide antenna, (b) an array of in phase waveguide antennas, (c) horn antenna, (d) properties of oil i.e. these properties of oil do not change with
steam EOR [6]. change in temperature.
Note that all the antennas have been set to have the same phase,
input/output power, amplitude, and operating frequency. This leads
consideration the complexity of the reservoir model, a simplified to the creation of a constructive interference which is the main
underground reservoir has been considered in this paper. This is reason of an efficient increase in temperature and pressure.
because the proposed technique is not mature enough to be A comparison of Figs. 6 and 7 shows that considering the elec-
employed comprehensively by commercially-available software tromagnetic and thermal properties as static may lead to higher
packages (e.g. Eclipse, CMG, and Petrel packages) which do not temperature distribution within the reservoir, which leads to an
comprise microwave heating model. Table 1 summarizes the fea- overestimation of the oil recovery.
tures of oil reservoir considered for the simulations along with the Fig. 8 shows the temperature and pressure distributions at
frequencies, power levels and number of microwave antennas. 2.45 GHz frequency by considering all of the four aforementioned
microwave sources. The largest horn antenna in the diagram has a
5.1. The reservoir model and simulation results height of horn antenna equal to 40 cm while small horn antenna is
of 30 cm in height.
Usually symmetry is assumed for 2D models in COMSOL. Thus, Fig. 8(a) provides the temperature distribution across the
as shown in Fig. 4, only a half part of a two dimensional model of reservoir as a function of the distance from the well bore. This latest
the reservoir has been considered in the simulation. Hence, the is the bottom of the injection well at which the microwave source is
simulator considers the well shown in the reservoir as lying at the placed. It is evident from this figure that the use of antenna array

Fig. 13. Comparison between temperature and pressure distributions with different type of antennas at 5.8 GHz operating frequency.
 M.M. Abdulrahman, M. Meribout / Energy 66 (2014) 868e880 877

It is evident from Fig. 9(a) that the array of three antennas

operating at 2.45 GHz frequency led to the highest oil recovery
throughout reaching the peak recovery after about 300 days of
heating. This is mainly because of the constructive interference of
the signals emitted by all the antennas of array. A similar trend was
obtained with the other types of antennas, but with less oil
production.

5.3. Simulation results with 915 MHz antenna

Following the above simulations, other microwave antennas

were designed to operate at a frequency of 915 MHz depending on
the cut-off frequency corresponding to the TE10 mode. Here, the
cut-off frequency is the one which sets the limit for waveguide
below which the attenuation increases significantly. The simula-
tions performed with 915 MHz frequency are similar to those
performed at 2.45 GHz except difference in frequency and size of
Fig. 14. Comparison of oil recovery with different microwave antennas operating at antennas. Some changes observed in oil recovery and heat distri-
5.8 GHz.
butions are discussed in this section. Fig. 10 shows the comparison
between temperature and pressure distributions as a function of
contributes significantly to the increase of temperatures in the case the distance from well bore. Four types of microwave sources have
of an operating frequency of 2.45 GHz. The corresponding antenna been considered in the simulations providing microwave power of
is also practical to use in oil reservoirs since it does not require 50 kW for a period of one year.
much space and can be easily installed inside the injection well It is evident from Fig. 10 that the array of microwave antennas is
[21]. Microwave heating time is taken as one year for all the sim- also best suited for operating frequency of 915 MHz. In fact it
ulations with an output power of 50 kW. Fig. 8 (b) shows that the provides high temperatures along with in depth penetration of low
highest pressure distribution is obtained with the antenna array. frequency microwaves because of a higher penetration depth of
The reason may come from the fact that at the injection well, the 915 MHz as opposed to 2.45 GHz. Small microwave horn antenna
high temperature which appears nearby the microwave source with a height of 30 cm is also suitable as compared to large mi-
causes light hydrocarbons to evaporate which in turn adds up to- crowave horn antenna of 40 cm height and single waveguide an-
wards the vapor pressure in the reservoir giving rise to the reser- tenna. However, it leads to lower temperature distribution and
voir pressure ultimately. depth penetration as compared to the array of microwave antennas.
Steam flooding is one of the commonly used thermal enhanced Furthermore, large E-plane horn antenna with 0.4 m height and
oil recovery. Fig. 9 shows the performance of this technique, in single magnetron antenna have almost same temperature distri-
terms of temperature increase and pressure distribution, with re- butions within the first two meters of the reservoir. However, after
gard to microwave heating. The production rate here, Q, can be 2.5 m, single magnetron antenna becomes more economical and
calculated using the following equation: efficient as compared to small horn antenna.
A plausible justification of these results is that with horn an-
 tenna, microwave radiations are not focused but spread throughout
Q ¼ K Pi -Pp (18) a larger area. Hence, similarly to the work done with 2.45 GHz
antennas, the use of array of antennas achieves a high pressure
where (Pi  Pp) are the values of pressure at the injection well and distribution, resulting from the emergence of large volume of hot
production well and K a variable inversely proportional to the vis- vapor.
cosity (the viscosity in its turn depends on the temperature) and Taking into consideration the fluid flow equation provided in
proportional to the oil relative permeability and the well porosity. Equation (13), Fig. 11(a) shows the oil recovery with an operating

Fig. 15. Temperature and pressure distributions for microwave array of antennas operating at 5.8 GHz frequency (static properties): (a) temperature distribution, (b) pressure
distribution.
878 M.M. Abdulrahman, M. Meribout / Energy 66 (2014) 868e880

both the temperature and pressure distributions as a function of

distance from the well bore. It can be observed that all the sources
achieve almost same pressure distributions except the array of
microwave antennas, which provides the highest pressure distri-
butions. This is mainly attributed to the constructive interference of
the three microwave signals which led to a fast increase of tem-
perature, especially nearby the well bore.
Similarly, Fig. 14 shows the amount of oil recovered from an oil
reservoir based on the achieved temperature and pressure rises and
viscosity drops. Hence, except for the case of array of microwave
antennas, all other designs were found less efficient in terms of oil
recovery. This is due to the fact that at 5.8 GHz frequency, micro-
waves cannot penetrate well into the depth of oil which limits its
temperature rise.
Fig. 15 shows the temperature and pressure distributions when
the oil reservoir is exposed to 5.8 GHz microwave radiations. Hence,
Fig. 16. Experimental setup.
similarly to the results obtained for 2.45 GHz and 915 MHz radiations,
the temperature and pressure distributions exhibit high values near
frequency of 915 MHz. It can be observed that a substantial re- the microwave source and then decrease gradually with distance.
covery of oil can be achieved, proportionally to the duration of
microwave heating. However, beyond a certain period (310 days in
6. Experimental results and discussion
this case), the oil recovery remains constant. This can be explained
with the fact that the vaporization process starts after 310 days,
In order to validate the model, an experimental setup has been
keeping the temperature constant after this time which stabilizes
built (Fig. 16). It consists of two tanks: an outer tank, within which
the recovery rate at a particular level.
an inner tank holds the magnetron inside an isolated area which
Fig. 11 also shows the efficiency of microwave enhanced oil re-
mimics the well reservoir. The dimensions of the inner tank were
covery, compared to the most popular conventional EOR, i.e. steam
(60  60  60) (L, W, H) cm3 while the ones for the outer tank were
flooding. The results prove that microwave heating with even sin-
(75  75  60) cm3. The inner tank was fixed from the bottom
gle magnetron antenna are more effective than steam heating. A
exactly at the center leaving 75 mm space between the walls of the
major drawback noticed in the steam flooding is the inability of
inner and outer tanks on each side. Water was filled in this area
steam to penetrate into the depth of fluid and behind the obstacles
surrounding the inner tank to absorb microwaves for the safety
e.g. cap rock. For instance, Fig. 12 provides a comparison between
purpose. In addition, another glass tank of the dimensions of
the temperature distributions obtained with different type of mi-
(60  60  10) cm3 was placed on the top of the internal tank and
crowave sources operating at 915 MHz and steam heating. The
filled with water to not to allow any escape of microwaves from the
results bring clear evidence that the penetration capability of steam
top. In addition, since the area surrounding the microwave is not
is far less as compared to microwave heating.
exposed to the air a pipe carrying cool water was provided within
this area to prevent sharp increase of temperature. A water pump is
5.4. Simulation results with 5.8 GHz antenna employed on the ground surface for sending the cold water inside
and extracting the hot water from the well. A similar cooling system
Similar simulations have also been conducted at 5.8 GHz fre- should be considered in real reservoirs as well since the well bores
quency by considering the four types of antennas. Fig. 13 shows are initially hot and lack the circulation of fresh air.

Fig. 17. Experimental setup: (a) real, (b) microwave generating setup.
 M.M. Abdulrahman, M. Meribout / Energy 66 (2014) 868e880 879

Fig. 18. Experimental validation of the model: (a) temperature distribution (simulation for 60 s), (b) comparison of temperature rise with different MW excitation durations.

Besides being safe for the operator, it is clear that the suggested almost 1.5 mW per cm2 has been detected by the meter which is
experimental setup prevent the waves emitted by the microwave well below the standard limit of 5 mW per cm2.
source to be reflected back to the source, which is the case of real- Next, the microwave source was powered on for 30 s and the
life reservoirs. Fig. 17 shows a photograph of the experimental temperature of the oil sample raised from 24.3 to 29.6
C was
setup, along with its different constituents. observed. In the next set of experiments, the microwave source was
During the experiments, the temperature distribution was powered on for 1 min and the temperature raised of oil reached
measured by five thermocouples (Pt 100, WS 130 Series, Deange, 29.8
C. This small increase of temperature may be explained by the
Taiwan) fixed with temperature displays and placed at equally at large quantity of oil exposed to the microwaves. However, when the
defined distance of 20 cm from each other radially away from microwave source was powered on for 1.5 and 2 min duration, the
the source. These temperature sensors are connected to a host temperature rise was 5.6 and 5.8
C respectively.
computer via three meter long extension wires coated with Teflon The same experimental setup was built in COMSOL to perform
(a microwave resistant material). Note that the thermocouples have similar simulation procedures. Hence, the oil medium was heated
the temperature sensitivity ranging from 50
C to 300
C which is up for 30 s, 1 min, 1.5 min and 2 min. The corresponding results as
enough for the requirement of experiment. shown in Fig. 18 indicate a similar trend with the experimental
In the experiments, a microwave magnetron operating at results. This validates to some extent the theoretical model sug-
1080 W input power, 700 W output power and 2.45 GHz frequency gested in this paper. Furthermore, Fig. 18 also demonstrates the
has been used. A Microwave leakage detection meter operating at difference between microwave heating of a sample of oil placed
2.45 GHz frequency and with 9 V DC battery has been used to detect inside the oven and by using the suggested experimental setup. A
any leakage of microwaves. significant difference between the temperature distributions has
Microwave leakage detector used is capable of measuring the been noticed strengthening the claim that microwave oven is not
electromagnetic radiations present in the air from 0 to 9.99 mW/ suitable for such experiments especially related to enhanced oil
cm2, with an accuracy of 1 dB. Safety standard considered for this recovery. This is because of the metallic walls of microwave oven
experiment is 5 mW/cm2 as set by FDA (U.S Food and Drug Asso- which reflect the electromagnetic energy and the target material
ciation) and EPA (U.S Environmental Protection Agency). Initially, absorbs this energy from all the sides. This is the reason for rapid
and prior to switch on the microwave source, a microwave power of rise of temperature of oil placed inside the microwave oven.
880 M.M. Abdulrahman, M. Meribout / Energy 66 (2014) 868e880

7. Conclusion [3] Zhou Z, Gunter W, Kadatz B, Cameron S. Effect of Clay swelling on reservoir
quality. J Can Pet Technol 1996;35.
[4] Liu J, Zhuang Y, Chen L, Guo J, Li D. Optimizing the conditions for the
Heating effects on the crude oil have been analyzed by consid- microwave-assisted direct liquefaction of Ulva prolifera for bio-oil production
ering different frequencies and types of microwave sources using response surface methodology. Energy (Elsevier) 2013;60:69e75.
including array of antennas and horn antennas. The results have [5] Bhattacharya M, Basak T. A theoretical study on the use of microwaves in
reducing energy consumption for an endothermic reaction: role of metal
been compared with steam heating as well as with other studies coated bounding surface. Energy (Elsevier) 2013;55:278e94.
done by other authors [6]. Temperature distributions and the drop [6] Kotani Y, Kansha Y, Tsutsumi A. Conceptual design of an active magnetic
of temperature as a function of distance from the well bore have regenerative heat circulator based on self-heat recuperation technology. En-
ergy (Elsevier) 2013;55:127e33.
been analyzed. In addition, oil recovery has been calculated using [7] Baday A, Abdullah A, Lee Ad K. Ultrasound-assisted transesterification of crude
the flow rate resulting from pressure distribution. Special attention Jatropha oil using cesium doped heteropolyacid catalyst: interactions between
has also been taken in defining the antenna parameters along with process variables. Energy (Elsevier) 2013;60:283e91.
[8] Acar C. Enhancing petroleum recovery from heavy oil fields by microwave
the orientation and placement of antennas which were not heating. Ankara: M.Sc, Petroleum and Natural Gas Engineering, Middle East
addressed earlier. Results of microwaves heating of the crude oil Technical University; 2007.
have been found quite satisfactory proposing them as a good choice [9] Kim ES. Reservoir simulation of in situ electromagnetic heating of heavy oils.
College Station, USA: Texas A and M University; 1987.
for extraction of heavy oil by heating. The most efficient method [10] Carrizales MA. Recovery of stranded heavy oil by electromagnetic heating.
was found to be the one correspond to the array of microwave Ph.D. Austin, Texas: Petroleum and GeoSciences Engineering, The University
antennas for the operating frequencies of 915 MHz as well as 2.45 of Texas; 2010
[11] Carrizales M, Lake L, Johns R. Multiphase fluid flow simulation of heavy oil
and 5.8 GHz. It was also noticed that the use of array of antennas for
recovery by electromagnetic heating. In: SPE improved oil recovery sympo-
915 MHz provides even better results as compared to the array of sium, Oklahoma, USA; 2010.
antennas operating at 2.45 GHz because of the less penetration [12] Kopyt P. A one-dimensional semi-analytic model of the microwave heating
depth of high frequency microwaves. Based on the above discus- effect in verification of numerical hybrid modeling software. In: IEEE inter-
national conference on “computer as a tool”, Warsaw; 2007. pp. 65e72.
sions, it is proposed to use a frequency of 915 MHz only in the case [13] Carrizales M, Lake LW. Two-dimensional COMSOL simulation of heavy-oil
where a thick and deep layer of oil is present. E-plane Horn an- recovery by electromagnetic heating. In: COMSOL conference, Boston;
tennas have also been assessed by simulation to find out that they 2009.
[14] Hascakir B, Akin S. Recovery of Turkish oil shales by electromagnetic heating
can be very effective for EOR. However their large size prevents and determination of the dielectric properties of oil shales by an analytical
them to be used in conventional well bores. Finally, the suggested method. J Energy & Fuels 2009;24:503e9.
model has been validated by performing a microwave heating [15] Kashif M, Yahya N, Zaid H, Shafie A, Jasamai M, Nasir N, et al. Oil recovery by
using electromagnetic waves. J Appl Sci 2011;11:1366e70.
experiment with single antenna and comparing its results with [16] Evans M, Warboys M. Coupled heat and moisture transfer during microwave
those provided by software simulations. vacuum drying. J Microw Power Electromagn Energy 1997;32.
[17] Fang C, Lai P. Microwave heating and separation of water-in-oil emulsions.
J Microw Power Electromagn Energy 1995;30:46e57.
References [18] Lake LW. Enhanced oil recovery. Old Tappan, NJ, USA: Prentice Hall Inc.; 1989.
[19] Chute F, Vermeulen F. Physical modeling of systems involving electromag-
[1] Stevens S, Kuuskraa VA, O’Donnell J. In: Alto Palo, editor. Enhanced oil re- netic heating of materials. IEEE Trans Educ 1990;33:190e7.
covery scoping study, final report. Electric Power Research Institute; 1998. [20] Helba A, Sahimi M, Scriven L, Davis H. Percolation theory of two-phase rela-
[2] Martson P. A regulatory framework for migrating from enhanced oil recovery tive permeability. SPE Reserv Eng J 1992;7:123e32.
to carbon capture and storage: the USA experience. Energy Procedia, (Elsevier) [21] COMSOL Multiphysics. Microwave oven. Available from: http://www.comsol.
2011;4:5941e7. com/showroom/documentation/model/1424/; 2005 [accessed 05.09.12].