Thermodynamic Simulation of Exhaust Gas Recirculation Coolers

Used on Diesels Based on Finite Element Method

Yansong Wang, Xiaolin Tang, and Hui He

Abstract—The thermodynamics of exhaust gas recirculation It has been found that it’s almost impossible to obtain an
(EGR) coolers plays a very important role for reducing oxides of exact solution of the airflow in the EGR coolers, due to the
nitrogen (NOX) emissions of diesel engines. Using the ANSYS fact of coupling of the hydromechanics and thermodynamics.
software, some fluid thermodynamic models of EGR coolers Fortunately, the numerical computation methods and some
were established by using the finite element method (FEM); and commercial software, such as the Cfdesign, ANSYS, and
the temperature fields of gas fluids in the EGR coolers were FLUENT, etc., for fluid and thermal analysis were developed
simulated and verified by comparing the experimental results
in the resent years. An example investigation of three-
from some literatures. Based on the FEM models, furthermore,
cooling efficiencies of the EGR coolers with different diffuser dimensional temperature and pressure distributions inside
structures and different numbers of cooling tubes were analyzed EGR coolers using the numerical software FLUENT can be
and discussed. The comparison results show that, for the same found in [10]. It suggested a probability for EGR cooler
dimension of EGR coolers, their cooling efficiencies can be design by using the numerical computations [11]. In this paper,
increased up to five or thirteen percents by replacing their the finite element method was adopted to calculate both
column diffusers with cone diffusers or by properly selecting the temperature and pressure fields of EGR coolers. Based on the
tube numbers of the EGR coolers. ANSYS software, a procedure of FEM fluid thermodynamic
modeling, simulation, and experimental verification of a
I. INTRODUCTION sample EGR cooler were performed. And furthermore,

T HE EGR technique was introduced in the early 1970s and

now has been widely used in diesel engines to reduce the
NOx in their exhaust emissions that was not being
cooling efficiencies of some typical types of EGR coolers
with different structures and numbers of cooling tubes were
discussed. Some constructive measures for structure
cleaned by the other smog controls. NOx are formed when improvement of the EGR coolers were put forward.
temperatures in the combustion chamber at 1370 centigrade
(oC) or hotter. The EGR system recirculates exhaust gases II. BASIC THEORY OF HYDROKINETICS AND
into the intake stream to displace some of the normal intake THERMODYNAMICS
charge. It can slow and cool the combustion process by As known that the discrete methods in the hydrokinetics,
several hundred degrees, thus reducing NOx formation so-called computational fluid dynamics (CFD) in engineering,
[1]-[3]. An EGR cooler is an air-to-water device that works in can be used to solve problems related to the hydromechanics
conjunction with the EGR system, which is used for cooling and thermodynamics. The main idea of CFD method is to
the exhaust gases that go back into the engine through the describe a continuous physical field by a set of variable
intake manifold. For avoiding an unreasonable temperature of collections, build their mathematical equations, and solve the
the exhaust gases returned to the engine combustion, the approximate values of these variables. As a numerical method
desired outlet gas temperature of the EGR cooler was usually of the basic fluid governing equations, CFD can be used to
controlled within an interval from 100oC to 150oC [4,5]. determine the physical parameter values in a complex fluid
To improve the thermodynamics of EGR coolers, a great field, such as the flow velocity patters, pressure distribution,
deal of research was issued from theory and experiment fields and temperature distribution [9].
in the past few decades. Most of these works focused on According to the principle of mass conservation, for an
experimental studies and attempted to get the thermodynamic element block in an ideal fluid field, the mass increment of
parameters by test data processing [2], [3], [6]. A good review the fluid element equals the mass of the fluid flows into the
of EGR-related theoretical methods was performed in [7]-[9]. element block within any time interval. Thus, the equation of
mass conservation can be expressed as,
Manuscript received May 1, 2008. This work is supported by the Scientific
Research Foundation for the Returned Overseas Chinese Scholars, State Dρ (1)
Education Ministry, and partly supported by Research Funds of the Science + ρ∇ ⋅ V = 0
and Technology and the Educational Departments of Liaoning Province, Dt
China under Grant No.20072198, No.20060393. where, ȡ is the fluid density; t denotes time; V={u, v, w} is
Yansong Wang is with the Research Center of Automotive Engineering,
Liaoning University of Technology, Jinzhou, 121001, Liaoning Province, China velocity vector, uǃv and w are the velocity components alone
(phone: 86-416-4199732; fax: 86-416-4199267; e-mail: [email protected]). with the x-, y-, and z-axes, respectively; ∇ and D/Dt are
Xiaolin Tang is with the Research Center of Automotive Engineering,
differential operators,
Liaoning University of Technology, Jinzhou, 121001, Liaoning Province,
China (e-mail: [email protected]). D ∂ ∂ ∂ ∂
Hui He is with the Automobile and Traffic Engineering School, Liaoning = +u +v +w (2)
University of Technology, Jinzhou, 121001, Liaoning Province, China Dt ∂t ∂x ∂y ∂z
(e-mail: [email protected]).

141
978-1-4244-1787-2/08/$25.00 
c 2008 IEEE
 From the law of momentum conservation, the momentum
varying rate (with respect to time) of the fluid element equals
the external force acting on the element block, i.e., the
Navier-Stokes equation,
DV (3)
ρ = −∇P + μ∇ 2V + F
Dt Fig. 1. Two-dimensional FEM model of the sample 19-tube EGR cooler

where, P is the fluid pressure, F is the bulk force vector of the

fluid.
In light of the principle of energy conservation, the energy
increment of the fluid element equals the sum of the entered
heat flux and the powers generated by both the bulk and Fig. 2. Tube layout of the 19-tube EGR cooler (1/2th of cooler represented)
surface forces of the fluid element block. This procedure can
be shown as the following equation,
IV. SIMULATION AND EXPERIMENTAL VERIFICATION
DT k 2 (4)
ρ = ∇ T + ST A. Numerical Simulation
Dt c p
Since the experimental results in [12] were adopted to verify
where, Cp denotes specific heat of the fluid, T denotes the the simulated results, the test conditions in [12] were certainly
fluid temperature, k is heat transfer coefficient, ST is the heat selected as the boundary conditions for numerical computation
energy transferred from mechanical energy, due to interior in this paper. Therefore, the inlet temperature and pressure of
heat source and viscosity of the fluid element block. ST was the exhaust gas were set and kept constants at 420 oC and 600
set to zero in this paper. kPa, respectively. The above FEM model is limited to the
inside volume from the air side. The water side and material
III. FEM MODELING OF EGR COOLERS interfaces were not been included in the simulation. The wall
surface temperature was set constant at 91 oC. This is a fair
A. Assumptions for FEM Modeling approximation since the coolant temperature change is only a
Based on CFD versions of the above equations, some FEM couple of degrees in all working conditions of diesels, due to
models of EGR coolers were developed in the present work, the high coolant flow, from 0.5 to 1 kg/s.
on the assumptions that cooling water with a constant From the hydromechanics, flowing characteristics of the
temperature is filled in the cooling cavities of the EGR exhaust gas in the EGR coolers depend on Renault number,
coolers; the exhaust gas in the tube bundles is isotropic and which can be calculated by the following expression,
has the characteristics of the ideal air; tube bundles are
symmetrically arranged in the radial direction on the cross Re = ρvLc / μ (5)
section of the EGR coolers, and all tubes in each bundle are
Here, ²and­are the density and absolute viscosity of the
straight and smooth, or straight with screw grooves; ignoring
the effects of temperature gradient along the radical direction exhaust gas (air); ® is the gas flowing speed, Lc is a ratio of
of the tube walls, due to their thin thickness [3]-[5]. section area of the gas flow reached and length of the cooling
tubes. The calculated Renault number of the exhaust gas in
B. FEM Model Establishment this work is more than 2300; therefore, the “turbulent”
Based on the above assumptions, using the ANSYS property of gas flow was defined. And a compressing analysis
software, a two-dimensional FEM model of a sample 19-tube needs to be performed by setting a viscosity manually for
EGR cooler was firstly established for comparing with the avoiding minus temperatures appearing in the simulation
tested results in the [12], see Fig. 1. The tube layout of the results. This manual viscosity was initialized with a certain
19-tube cooler is shown in Fig.2. The element FLUID141 2D values, and gradually approaching to zero during the
of the FLOTRAN CFD module in the ANSYS was selected in simulation procedure [11]-[12].
the FEM model. The structure parameters of the EGR cooler The simulation results can be easily obtained by using the
are: the total length of the cooler is 384mm, with a water after-processor. Fig. 3 shows the Temperature distribution in
jacket diameter of 76mm; the diameter of the inlet and outlet the 19-tube EGR cooler. It can be seen that the peak values
port of the EGR cooler are 45mm; the opening angle of the are mainly distributed in the entrance sections of the tubes,
cone is constant at 28o; the tubes are identical in length and which may be explained by the sudden convergence of the
diameter. They are 200mm long, and their inner and outer gas flow. The temperature changing along with the
diameter 10mm and 12mm, respectively. All the computational center-axis of the 19-tube EGR cooler is shown in Fig. 4.
meshes of the FEM model were created manually in the With the gas flowing, the gas temperature in the EGR cooler
pre-processor of the ANSYS. Considering the significantly increases up to 700 Kelvin (K) (426.85oC) and then gradually
changing of thermal gradient near the boundaries, the meshes decreases to a constant value of 443.8 K (170.6oC) in the
were denserly dealt with in the airboxes, the tube bundle and cooler outlet pipe. The same conclusion can be drawn in Fig.3,
the inlet/outlet port for more detail information. Finally the due to the stable outlet field of temperature at around 450 K in
element size is 2511 for the 19-tube EGR cooler. the cooler.

th
142 2008 Asia Simulation Conference — 7 Intl. Conf. on Sys. Simulation and Scientific Computing
 the gas flow is equally divided into the tubes, as well as the
lowest pressure drop. In order to obtain the best possible
thermal efficiency, the overall design of the EGR cooler can
only result from a compromise between the tube bundle and
airbox design, considering the flow range on the gas side. The
aim of this work is to study several cooler design parameters
in term of thermal efficiency. The first parameter is the shape
of the diffuser fitted at the inlet and outlet of the cooler. The
Fig. 3. Temperature distribution in the 19-tube EGR cooler
second one is the arrangement of the tube bundle.
A. Comparison of Two Types of diffuser
The function of the diffuser is to distribute the hot EGR gas
flow into the tubes, and collect the cooled gas and direct it
into the outlet pipe. Two pairs of diffuser in common use
were evaluated and compared in terms of the flow distribution
and cooling efficiency. The diffusers were fitted on the tube
bundle of the 19-tube EGR cooler presented above. The
geometry of the diffusers is shown in Fig.5. Note that the
shapes of the types I and II are cone and column, respectively.
Based on the model of the type I, the modeling and simulation
can be done by just modifying the geometry shape of the
diffuser. The FEM model and simulation results of the type II
are shown from Fig. 6 to Fig. 8. Figs. 9 and 10 are the
calculated gas flow velocities of the type I and type II. The
simulated results are compared in Table II.
Fig. 4. Temperature changing along center-axis of the 19-tube EGR cooler

B. Experimental Verification
As mentioned before, the model structure and simulation
boundary conditions have been determined according to the
sample EGR cooler tested in [12]. Thus the experimental
results in [12] can be directly used to verify the simulation
results. a comparison were listed in Table I, where the cooling
efficiency of an EGR cooler¨ is defined as, Type I Type II
Fig. 5. Two types of EGR cooler diffuser
η = (t i − t o ) (t i − t w ) (6)

where, ti, to and tw denote the temperatures of the inlet gas, the It can be seen that the outlet gas temperature is from 449 K
outlet gas and the cooling water, respectively. It can be seen to 494 K in Fig.7 and a constant in Fig. 8 at 456 K(186oC).
from Table I that there is a 7.6oC difference between the And it can be concluded from the comparisons in Table II that
tested and simulated temperatures of the outlet gas, which the EGR cooling effect has been getting worse after changing
implies a four-percent error in temperature and a two-percent the diffuse from type I to type II, due to five-percent decrease
error in cooling efficiency in the simulation. These errors of the cooling efficiency. The reason may be partly explained
might be attributed to some uncertain errors either in the by the status of the gas flowing in Figs. 9 and 10. The cone
measurement devices and operations of the tests, or in the diffuse with a reasonable opening angle supports a smooth gas
modeling and computation of the simulation. The comparison flow and is helpful for the gas flow and heat transfer in the
result suggests that the simulation error is less than five EGR cooler. But the column diffuse generates a lot of recycling
percents, within the permission limit in engineering. gas flows in the four dead corners, see Fig. 9.
Therefore, conclusion can be drawn that it is feasible and
validate to solve thermal problems by using the FEM TABLE I
approach in the ANSYS software. COMPARISON OF THE TESTED AND SIMULATED RESULTS OF THE SAMPLE COOLER
Comparison items Simulated results Tested results
V. STRUCTURE INVESTIGATION OF EGR COOLERS Cooling water temperature 91oC 91oC
o
Inlet gas temperature 420 C 420oC
The EGR cooler is a shell-and-tube structure. The cooling o
Outlet gas temperature 170.6 C 163oC
unit consists of a tube bundle surrounded by the water jacket.
Cooling efficiency 76% 78%
A pair of diffusers is fitted on both side of the cooling unit. The
tube bundle arrangement should provide the lowest pressure
drop given the gas flow and the most compact design for a
given number of tubes. The airbox design should insure that

th
2008 Asia Simulation Conference — 7 Intl. Conf. on Sys. Simulation and Scientific Computing 143
 TABLE II B. Discussion of Tube Number in EGR Coolers
SIMULATED RESULT COMPARISONS OF THE TWO TYPES OF EGR COOLERS
The second design parameter studied in the present work is
Comparison items EGR cooler (type I) EGR cooler (type II)
the tube bundle arrangement. It has been proved the tube
Cooling water temperature 91oC 91oC
bundle arranged symmetrically in the radial direction of the
Inlet gas temperature 420oC 420oC
cooler cross section is the best way for a high cooling
Outlet gas temperature 170.6oC 186oC efficiency of the EGR systems [10]. Therefore, only the tube
Cooling efficiency 76% 71% number in an EGR cooler will be discussed in the flowing
text.
Based on the 19-tube EGR cooler, a new EGR cooler with
seventy three tubes was deigned for light truck diesel engines,
referring to the structure of a high cooling-efficiency EGR
cooler used on heavy truck diesels [3]. The arrangement of
the new tube bundle is shown in Fig. 11. The 73-tube EGR
cooler has the same geometrical dimensions and boundary
Fig. 6. FEM model of an EGR cooler with the type II diffuser
conditions as those in the 19-tube EGR cooler (type I)
simulation, expect for the tube inner and outer diameters are
5.4mm and 6mm, respectively. In analogy with the above
numerical modelling and simulation procedure, a FEM model
with 2680 elements of the 73-tube EGR cooler was
established in the ANSYS (see Fig.12), and thereby the
temperature distribution and changing along longitudinal axis,
see Figs.13 and 14. Fig.13 shows the outlet gas temperature is
Fig.7. Temperature clouds of the EGR cooler with the type II diffuser within the interval [348K, 450K] and develops towards to a
constant value of 402K (125oC), see Fig. 14. Table III gives a
thermal efficiency comparison of the 19- and 73-tube EGR
coolers. Comparing the 19- and 73-tube EGR coolers,
conclusions may be drawn that: the gas temperature on the
inlet side of the 73-tube bundle is 65oC higher than that of the
19-tube bundle, which makes the 73-tube bundle more
difficult in manufacture, and may perhaps be attributed to an
increased flow resistance in the 73-tube cooler; the gas
temperature in the outlet pipe of the 73-tube cooler is 125oC
(45.6oC lower than that of the 19-tube cooler), which well
meets the diesel EGR requirement of temperature interval
[100oC,150oC]; along the longitudinal axis, the gas
temperature of the 73-tube cooler dropped more quickly, and
the cooling efficiency increased thirteen percents than the
19-tube cooler.

Fig. 8. Temperature changing along center-axis of the type II EGR cooler

Fig.11. Tube layout of the 73-tube EGR cooler (1/2th of cooler represented)

Fig. 9. Gas flow velocities of the EGR cooler with the type II diffuser

Fig.12. FEM model of the 73-tube EGR cooler

Fig.10. Gas flow velocities of EGR cooler with the type I diffuser

th
144 2008 Asia Simulation Conference — 7 Intl. Conf. on Sys. Simulation and Scientific Computing
 dynamical analysis. In the design stage of an EGR cooler, the
proposed methods might be used to predict its performance,
therefore, may be regard as an accessorial tool for EGR
system design.

ACKNOWLEDGMENT
Fig.13 .Temperature distribution of the 73-tube EGR cooler Yansong Wang thanks the Jinzhou Huaneng Welding Tube
Company in Liaoning province, who has given many
supports to this research work.

REFERENCES
[1] U. Lundqvist, G. Smedler, P. Stalhammar. “A comparison between
different EGR systems for HD diesel engines and their effect on
performance”, Fuel Consumption and Emissions, SAE, 2000 Paper
2000-01-0226.
[2] N. Wu, Z. D. Qiu, and Z. M. Tao, “Design of venturi tube of EGR
system in YC6105ZLQ diesel engines”, Internal Combustion Engines.
no.1, pp. 45-47, Feb. 2006.
[3] Joseph R Griffin, “Cooled EGR rate measurement with a thermal
anemometer for EPA02 heavy duty diesel engine emission control”,
SAE, 2003, Paper 2003-01-0263.
[4] Martin Müller, Peter M Olin, Bart Schreurs, Dynamic EGR Estimation
for Production Engine Control. SAE, 2001, Paper 2001-01-0553.
[5] Matthew J. Hall, Patrick Zuzek, “Fiber optic sensor for time-resolved
measurements of exhaust gas recirculation in engines”, SAE, 2000,
Paper 2000-01-2865.
Fig. 14. Temperature changing along the center-axis of the 73-tube EGR [6] X. G. Fu, X. M. Guo, and Y. X. Liu, et al., “Design and experiment of
control system for EGR cooling system on diesel engine”, Journal of
Agriculture Machine, vol. 37, no.5, pp. 34-36, May. 2006.
TABLE III [7] Quoc Phong Le, “Influences of Alternative Fuels and EGR on
SIMULATED RESULT COMPARISONS OF THE TWO TYPES OF EGR COOLERS Thermodynamic Parameters and Emissions of a Heavy-Duty Engine”,
Comparison items 19-tube EGR cooler 73-tube EGR cooler Ph.D. Dissertation, Dept. Mechanical. Eng., University of Magdeburg,
Germany, 2007.
Cooling water temperature 91oC 91oC
[8] A. K. Agrawal, S. K. Singh, S. Sinha, and M. K. Shukla, “Effect of
o
Inlet gas temperature 420 C 420oC EGR on the exhaust gas temperature and exhaust opacity in
Outlet gas temperature 170.6 C o
125oC compression ignition engines”, Sadhana, vol. 29, no. 3, pp. 275-284,
Jun. 2004.
Cooling efficiency 76% 89%
[9] F. Payri, J. Benajes, J. Arrègle and J. M. Riesco, “Combustion and
Exhaust Emissions in a Heavy-Duty Diesel Engine with Increased
Premixed Combustion Phase by Means of Injection Retarding”, Oil &
VI. CONCLUSIONS Gas Science and Technology, vol. 61, no. 2, pp. 247-258, Apr. 2006.
[10] H.-M. Kim, S.-K. Park, K.-S. Choi, et al., “Investigation on the flow
Using the ANSYS software, the thermal FEM simulations and heat transfer characteristics of diesel engine EGR coolers”,
of the diesel used EGR coolers were presented in this paper. International Journal of Automotive Technology, pp 149-153, vol. 9,
Based on the thermodynamic equations, we first established no.2, Apr. 2008.
a FEM model and performed a numerical simulation of a [11] F. J. Wang, Computation Fluid Dynamics Analysis - CFD Software
Principle and Application. Beijing, CA: Tsinghua University Press.
sample 19-tube EGR cooler. Comparisons between the
2005.
simulated and tested results suggest that the simulation [12] Y. P. Gao, Q. Wang, and X. H. Luo, “Investigation of exhaust gas
procedure is accurate and feasible for thermal analysis of recirculation cooler of DI diesel engine”, Journal of Jiangsu University
EGR coolers. For studying structure designs affect on (Natural Science Edition), vol. 25, no.6, pp. 492-95, Nov. 2004.
thermal efficiency of EGR coolers, furthermore, a 19-tube
cooler with a pairs of column diffusers and a newly designed
73-tube cooler with a pairs of cone diffusers were
respectively modeled, simulated and compared with the
sample one. In view of the cooling efficiency, the
conclusions were drawn that, for the same dimension of EGR
coolers, the cone-diffuser coolers are better than the
column-diffuser ones, and the 73-tube cooler is better than
the original 19-tube one. The newly designed 73-tube cooler
o
with an outlet gas temperature of 125 C is well satisfied the
EGR requirement of diesels. This implies a good approach
for cooling efficiency design by properly selecting tube
number of EGR coolers.
It should be mentioned that the work done in this paper
may be extended to other types of EGR coolers for thermo-

th
2008 Asia Simulation Conference — 7 Intl. Conf. on Sys. Simulation and Scientific Computing 145