You are on page 1of 7

Proceedings of IMECE2006

2006 ASME International Mechanical Engineering Congress and Exposition


November 5-10, 2006, Chicago, Illinois, USA

IMECE2006-15982
CFD ANALYSIS OF HEAT EXCHANGER AND AFTERCOOLER FOR
DIESEL MARINE ENGINES

Andrew S. Campbell, Jeff Ladendorf


nd
Hadady Corporation, 510 W 172 Street
South Holland, IL 60473

Chenn Q. Zhou, PhD.


Purdue University Calumet, Department of Engineering
th
2200 S 169 Street
Hammond, IN 46323

ABSTRACT performance. Over the years Hadady Corporation and


Shell and tube heat exchangers and bar and plate its costumers have recognized the uses of CFD as a
aftercoolers are used in a diesel marine engine to remove useful design and cost effective tool for the analysis of
heat from the engine coolant and to cool the pressurized fluid flow and heat transfer phenomenon.
air going into the engine. In order to improve the overall Heat exchanger and heat transfer theory (Sekulic and
effectiveness of these two components, analyses have Shah, 2003) in general allows for the analysis of these
been performed using heat exchanger theory and the components; however these methods do not provide the
Computational Fluid Dynamics (CFD) software visual insight into the phenomena occurring with in the
FLUENT®. Detailed information of flow property device. This is where CFD becomes a powerful tool.
distributions and heat transfer was obtained in order to Through the iteration process CFD allows for the
provide a fundamental understanding of the component detailed numerical analysis of complex geometries and
operation. Comparisons have been made between CFD yields further insight that can not be achieved through
results, theoretical calculations, and measurement data. theoretical methods alone.
Parametric investigations and feasibility studies have In this study, widely used commercial CFD software
been conducted to find optimum conditions for the (FLUENT®) was used to analyze the heat exchanger and
engine and component performance. the aftercooler. The main objective is to obtain detailed
results to provide information showing the flow
INTRODUCTION properties and heat transfer, and therefore obtain a
Heat exchangers are used on many types of engines to comprehensive understanding of heat exchanger
transfer heat away from fluids entering the engine. operation. From the results, parametric studies can be
Specifically, the two components analyzed in this study performed to identify the primary contributions to the
are a heat exchanger and an aftercooler. In this case they overall performance. For both components three
are being applied in a diesel marine engine application. dimensional steady state geometries where analyzed.
The exchanger removes heat from the engine coolant and Comparisons were made between experimental and
the aftercooler removes heat from the charged air coming theoretical results to provide validation and verification
from the turbocharger. The aftercooler cools the air prior of the simulation.
to entry into the engine. This allows for a better
combustion process. Both components use water as the THEORY AND APPROACH
cooling fluid taken from the water the boat is resting in. The components analyzed in this study are designed
Prior art associated with the design of these components for a diesel marine engine application. The shell and
was primarily based on intuition, experience, and trial tube heat exchanger removes heat from the engine
and error. These approaches can be attributed to the which makes it vital that the component be designed to
complexity of the geometries and physical phenomena accomplish this in an effective manor while not
occurring in the heat transfer process. In recent years compromising engine performance. The aftercooler is a
with the growth of computer technology and new heat exchanger that cools charged air from the
methods of numerical analysis, CFD has become an ever turbocharger prior to entering the engine combustion
increasing tool used to provide information on chamber. Correct and effective operation of this
parametric effects for optimization of component component will allow the engine to maintain a higher

1 Copyright © 2006 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 11/17/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use


level of combustion performance ultimately leading to a dimensional modeling software Pro-Engineer. Figure 1
cooler and more efficient engine. A cooler engine keeps shows the base case geometry for the heat exchanger.
the engine compartment of the boat cooler reducing the The left view shows the shell coolant (50 / 50 mix of
risk of fire and thermal stress. In the shell and tube, ethylene glycol and water) volume and the tube water
counter flow, heat exchanger the engine coolant flows volume. The right view shows the internal view of the
through the shell and water flows through a two pass geometry along with the flow paths of the respective
tube bundle to provide the means for heat transfer in the fluids.
engine cooling system. For the aftercooler, the water
flows through the water channels over a series of fins
and the air flows perpendicular to the water flow through
air channels, again containing fins which dramatically
increase the heat transfer surface area.
To analyze the flow and heat transfer commercial
software, FLUENT®, was used in conjunction with 3-D
modeling software PRO-ENGINEER 2.0 WILDFIRE®
and Fluent preprocessor GAMBIT®. These software
programs are considered to cutting edge and are
becoming increasingly popular in industry to solve fluid
flow and heat transfer problems. For this research, a
basic three step approach was used in order to achieve
the objectives.
The first step is to establish the theoretical Figure 1: Shell and Tube Heat Exchanger Geometry
calculations for both components. This was done using
heat exchanger theory and then implemented into a Figure 2 shows the base case geometry for the
MICROSOFT EXCEL® spreadsheet to allow for quick aftercooler. The left view shows the air channels
theoretical analysis. The calculations will be useful in volume and the water channel volume. The right view
determining the boundary conditions that need to be shows the internal view of the geometry and the
applied in setting up the simulation and also used to associated flow paths of the respective fluids.
provide a basis of validating the CFD simulation results.
The second step is to set up and perform the CFD
simulation. Also, in this step, the comparison between
the CFD and theoretical results is analyzed to provide
validation. The CFD provides detailed results, but to
ensure accuracy of the CFD simulation the results are
compared to actual experimental results from Hadady
Corporation to ensure the results are verified.
The third, and final step, is to perform parametric
studies to optimize the performance of the components.
Optimization can be performed by changing the physical
properties of the fluids or by modifying the geometry of
the component itself; however, only one parameter was
changed at a time to ensure that the parametric study
maintains a good relationship to the baseline case.
Future work in this research will be to model the
entire device including the housing and brackets that are
associated. This will allow for the application of thermal Figure 2: Aftercooler Heat Exchanger Geometry
stress analysis. The goal is to take CFD results and
apply them to Finite Element Analysis (FEA) software The geometries where exported from Pro-Engineer and
for a means of a detailed stress analysis. then imported into Gambit where the meshing process
was completed along with defining the respective inlets,
outlets, various walls, and interior volumes. The mesh
NUMERICAL METHODS itself consisted of small three dimensional unstructured
Geometry and Grid tetrahedral hybrid elements (2.9 million, shell and tube;
The geometries were constructed from design prints 1.5 million, aftercooler). Adaptive grids were
from Hadady Corporation and created in three

2 Copyright © 2006 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 11/17/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use


considered in the construction of the mesh. Adaptive form of heat transfer considered. All external walls
grids allow for the elements to densely or coarsely were considered to be adiabatic because of the unknown
packed in certain user defined areas. Due to the thermal conditions and any heat transfer that would be
complexity of the geometries adaptive grid application occurring would be small and therefore negligible.
was not possible. The channels in the aftercooler and the
tubes in the shell and tube heat exchanger are tightly RESULTS AND DISCUSSION
packed not allowing for a suitable area for the grid to be Validation and Verification
coarse. Validation of the CFD results is very important, even
if converged results are obtained it is still possible for
Boundary Conditions the results to be incorrect. Incorrect results can be
Flow conditions are very important to the results of caused by improper boundary conditions, grid
the simulation and required to properly define the distribution, or an invalid assumption. This is why it is
boundary conditions. Table 1 shows an example of very important to compare the results of the base case
typical flow conditions for the heat exchanger. against theoretical calculations and experimental data.

Table 1: Typical Heat Exchanger Flow Conditions Heat Exchanger


Coolant Inputs Units The base case results for the heat exchanger were
Mass Flow Rate 5 (kg/s) mixed. Table 3 shows the comparison between the
Inlet Temperature 90 (°C) theory results and the CFD results.
Inlet Pressure 90 (kPa)
Seawater Inputs
Mass Flow Rate 6 (kg/s) Table 3: Heat Exchanger Results Theory vs. CFD
Inlet Temperature 30 (°C) Seawater Outputs ∆ % Diff. Units
Inlet Pressure 135 (kPa) Outlet Temperature 4.92 10.91 °C
Drop Pressure 1.37 8.55 kPa
Coolant Outputs
The heat exchanger and the aftercooler are in series Outlet Temperature 5.53 7.52 °C
with one another, so to maintain conservation of mass Drop Pressure 8.83 58.84 kPa
through the system the water flow rate is the same. The Heat Transfer Rate 13.65 5.55 kW
water starts in the aftercooler and then travels to the shell Effectiveness 0.11 5.55
and tube heat exchanger. Table 2 shows typical flow
conditions for the aftercooler. The results show that from a thermal standpoint the
results are within 10%, but the shell side pressure drop
Table 2: Typical Aftercooler Flow Conditions is not. The theoretical pressure drop was much higher
Air Inputs Units then the CFD calculated pressure drop. This is due to
Mass Flow Rate 0.5 (kg/s) the inability to put clearances into CFD. The primary
Inlet Temperature 220 (°C)
contribution to the pressure drop is the leakage between
Inlet Pressure 100 (kPa)
Seawater Inputs
the shell wall and the baffles. In the CFD geometry it
Mass Flow Rate 6 (kg/s) was not possible to model these gaps because of the
Inlet Temperature 30 (°C) inability to apply an adequate grid to the small areas.
Inlet Pressure 140 (kPa) Comparison between experimental results and CFD
results is shown in table 4. Again, it is shown that the
Notice that there is no difference in the water correlation between CFD and test results is within 10%
temperatures going from the aftercooler to the heat from a thermal standpoint. The results from a pressure
exchanger. This is a direct result of the fluids used in the drop do not correlate as well. This can again be
exchangers. The specific heat capacity of air is much attributed to the baffle clearances. The experimental
smaller then water, so the amount of energy lost to the pressure drops were much higher then the theory
water from the air only raises the temperature slightly. calculations. This may be attributed to the way the
From the typical flow conditions the boundary pressures were obtained experimentally.
conditions were defined. Mass flow inlets were defined
for both components as well as the outlets which were
defined as pressure outlets. The internal walls were
defined to allow for the average convection coefficient to
be applied. The average convection coefficient was
obtained from theoretical calculations. Convection is the
primary mode of heat transfer; therefore, it was the only

3 Copyright © 2006 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 11/17/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use


Table 4: Heat Exchanger Results Test vs. CFD Using the same cross-sectional approach as figure 4,
Seawater Outputs ∆ % Diff. Units figure 5 shows the temperature contours through the
Outlet Temperature 0.41 1.01 °C tube bundle and the shell. The figure shows that thermal
Drop Pressure 2.57 21.22 kPa exchange between the water in the tubes and the coolant
Coolant Outputs in the shell.
Outlet Temperature 6.82 9.43 °C
High
Drop Pressure 4.22 40.59 kPa
Heat Transfer Rate 13.65 5.55 kW
Effectiveness 0.10 45.33

Heat Exchanger Flow Characteristics


The heat exchanger was simulated in steady state.
The power of CFD allows for insight into the physical
phenomena occurring in the device. Figure 3 shows the
velocity vectors of the coolant flow through the shell.
High
Shell-Inlet Near Inlet/Outlet
Low

Figure 5: Temperature Contours (۫C)

Heat Exchanger Parametric Study


The basis of the parametric study performed on the
heat exchanger was to develop a relationship between
Figure 3: Velocity Vectors through Shell (m/s) the number of baffles in the shell and the height of these
Low baffles. This was accomplished by first calculating the
The analysis of this figure shows that as the flow effects of raising the baffle height. The next step was
passes around the baffles there is a dead area of low then to add 4 baffles compared to the base case 2 baffles
velocity and this area of low velocity is the area where and compare how the relationship changes as the height
the greatest pressure drop occurs and can be a possible is increased. Figure 6 shows the effects of this study.
area for design improvement.
Figure 4 shows the flow of the coolant around the tube Baffle Study
100.00
buddle. This shows the turbulent nature of the flow MAX
90.00
around the tube buddle. Having turbulent flow is
80.00
important in the design of a heat exchanger because it
70.00
Pressure Drop

increases the heat transfer, but also causes higher


60.00
pressure drop.
50.00

40.00

High 30.00

20.00

10.00
2 Baffles 4 Baffles
0.00
70.00
0
75.00 80.00 Baffle Height
85.00 90.00 95.00 100.00 105.0
MAX

Figure 6: Baffle Parametric Study Comparison

As baffle height increases the pressure drop decreases.


This is due to the overall higher velocity through the
shell of the heat exchanger. When the addition of 4
baffles is made one can see at the highest baffle height
the pressure drop is close to the same as the initial
pressure drop in the base case. The addition of the
Low baffles also increases the heat transfer area, thus
increasing the thermal effectiveness of the exchanger.
Figure 4: Velocity Vectors through Shell (m/s) Figure 7 shows the velocity vectors through the shell
with 4 baffles versus the height of the baffle. The top

4 Copyright © 2006 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 11/17/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use


figure shows the minimum baffle height and the bottom with Fluent’s porosity function. After correspondence
shows the baffle height at the maximum. with Fluent software technicians the conclusion was that
by increasing the convection coefficient by 10% to help
High account for the thermal variations it may account for the
limitations associated with the porosity function. The
pressure drop variations could not be solved due to the
lack of leakage between the fins and the walls. This
again, like the heat exchanger, is due to the inability to
apply grids to leakage areas. Tables 7 and 8 show the
effects that of increasing the convection coefficient.

Table 7: Aftercooler Results


Theory vs. CFD (Modified Convection Coefficient)
Seawater Outputs ∆ % Diff. Units
Outlet Temperature 2.95 9.01 °C
Drop Pressure 0.69 7.47 kPa
Low Air Outputs
Outlet Temperature 4.23 10.15 °C
Figure 7: 4 Effect of Baffle Height Drop Pressure 0.54 44.79 kPa
Heat Transfer Rate 1.57 1.69 kW
The conclusion from this parametric study is that 4 Effectiveness 0.03 3.11
baffles at an increased height will yield approximately
the same pressure drop as 2 baffles at the original height. Table 8: Aftercooler Results
The reason to change to 4 baffles is because of the Test vs. CFD (Modified Convection Coefficient)
increase in the heat transfer area, thus making the device Seawater Outputs ∆ % Diff. Units
Outlet Temperature 2.15 6.72 °C
more thermally effective while maintaining Drop Pressure 16.10 61.92 kPa
approximately the same pressure drop as the base case. Air Outputs
This is also a cost effective design modification. Outlet Temperature 6.86 17.59 °C
Drop Pressure 23.33 97.22 kPa
Heat Transfer Rate 0.87 0.94 kW
Aftercooler Effectiveness 0.03 3.11
Like the base case of the heat exchanger the results of
the aftercooler were mixed. A porosity function was The result of increasing the convection coefficient
used to simulate the fins inside the water and air makes the results better from a thermal standpoint by
channels. This porosity function was the source of error increasing the heat transfer between the two fluids. The
in the simulation. In the tables 5 and 6, the results are increasing the of the convection coefficient was a valid
compared in the same fashion as the heat exchanger. coarse of action because an increase of 10% is not a
drastic change to the simulation, but is a tool to help
Table 5: Aftercooler Results Theory vs. CFD account for the limitations in the porosity function.
Seawater Outputs ∆ % Diff. Units
Outlet Temperature 3.49 10.65 °C
As with the heat exchanger CFD allows for insight
Drop Pressure 0.17 1.89 kPa into the flow in the device, however, due to the porosity
Air Outputs function the velocity vectors are not as conclusive to
Outlet Temperature 17.39 41.76 °C what the flow is actually doing. Thermally a sense of
Drop Pressure 0.54 44.79 kPa
Heat Transfer Rate 8.49 9.16 kW
what is occurring inside can still be portrayed. Figures
Effectiveness 0.08 9.16 8, 9, and 10 show the temperature contours of the
aftercooler walls.
Table 6: Aftercooler Results Test vs. CFD
Seawater Outputs ∆ % Diff. Units
Outlet Temperature 2.69 8.41 °C
Drop Pressure 16.61 63.90 kPa
Air Outputs
Outlet Temperature 20.02 51.33 °C
Drop Pressure 23.33 97.22 kPa
Heat Transfer Rate 7.79 8.47 kW
Effectiveness 0.08 8.47

The results for the aftercooler have a poor correlation


due to the thermal and pressure drop issues associated

5 Copyright © 2006 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 11/17/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use


because it directly determines how the flow enters the
High core and thus can be correlated to thermal effectiveness
and pressure drop. For this parametric study an
Air aftercooler was taken with an existing geometry. The
Inlet lower housing and the core geometries were held
constant and the top housing was modified to show the
effects of changing the air outlet location. Figure 11
shows the geometries used in the study with the arrow
representing the air flow path.

Upper Housing

Low
Water
Inlet
Figure 8: Temperature Contours (۫C)
Core
High

Lower Housing

Figure 11: Housing Comparison Geometries

The only volume considered in this study was the air


volume. The water heat transfer was simulated by
applying a heat flux to the walls of the core. This was
done just to simulate some heat transfer and not to
Water Outlet Air Outlet Air Inlet
Low reflect actual real world characteristics.
Figure 12 shows the velocity vectors through the
Figure 9: Temperature Contours (۫C) housings of the aftercoolers. Areas of primary concern
the areas of low velocity because these areas indicate
High
lower heat transfer and an increase in pressure drop.

High
Air Flow Paths

Water Flow Paths


Low

Figure 10: Temperature Contours (۫C) Low Areas of Lower Velocity

By looking at the temperature contour plots we gain Figure 12: Velocity Vectors though
insight into the thermal flow occurring in the aftercooler. Aftercooler Housings (m/s)
The pictures give a visual depiction of the thermal
characteristics inside the aftercooler. The aftercooler on the left has a smaller area of low
velocity and also maintains an overall higher velocity
Aftercooler Parametric Study over the core walls. The aftercooler on the right has a
Referring to the results of the base case aftercooler, larger area of low velocity. The air coming in has to
the aftercooler core maintains a thermal effectiveness of essential do a 180 degree turn. This 180 degree turn
greater then 90%. This leaves little room for limits the velocity over the far end of the core. Figure
improvement from a thermal standpoint. In any flow 13 shows the effects from a thermal basis.
device the housing around the core is very important

6 Copyright © 2006 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 11/17/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use


High been performed using heat exchanger theory and the
Computational Fluid Dynamics (CFD) software
FLUENT®. The results provided visual and numeric
insight into the complex phenomena occurring inside of
these devices.
The conclusion of the parametric studies show that
for the shell and tube heat exchanger the use of 4 baffles
over the original 2 baffle is more desirable from a
pressure drop and thermal standpoint. This agrees with
what theory and CFD show. For the aftercooler the best
design is to have the inlet and exit of the housing normal
Low Air Outlets
to the entrance and exit of the core. The components
Figure 13: Temperature Contours through modeled in the parametric study shows that the pressure
Aftercooler Housings (°C), drop CFD calculated was still useful in showing where
(Reverse Angles to Show Air Outlets) the maximum pressure drops will occur and how flow
can be adjusted to minimize these pressure drops. These
Figure 14 shows the contours of total pressure through
conclusions are valuable to Hadady Corporation and
the housings.
Purdue University Calumet in the continuing research of
High
heat transfer and heat exchangers. The power of CFD
has enabled both institutions to gain insight into the flow
and optimization of these devices.

References
FLUENT® User’s Guide, version 6.0, (2002) Fluent
Incorporated, New Hampshire.
D. P. Sekulic / R. K. Shah, Fundamentals of Heat
Exchanger Design. Hoboken, NJ: John Wiley & Sons,
Inc. 2003.
Low F. P. Incropera / D. P. DeWitt, Introduction to Heat
Transfer 4th ed. Hoboken, NJ: John Wiley & Sons, Inc.
Figure 14: Pressure Contours through
2002.
Aftercooler Housings (kPa)
R. H. Shih, Parametric Modeling with AutoDesk
Inventor 8. Mission, KS: Oregon Institute 2004.
Comparing figures 12, 13, and 14 it becomes apparent
R. E. Sonntag / C. Borgnakke / G. J. Van Wylen,
that the 180 degree turn the air has to travel in the
Fundamentals of Thermodynamics 6th ed. Hoboken, NJ:
aftercooler on the right directly correlates to a higher
John Wiley & Sons, Inc. 2003.
pressure drop. The results of the parametric study show
“Shell and Tube.”
that the aftercooler on the right has a 12 kPa higher
http://www.wlv.com/products/databook/ch2_5.pdf
pressure drop and 8 °C increased exit temperature.
(January 15, 2004).
The conclusion of this parametric study is that in the
“Shell and Tube Heat Exchanger Design.”
design of an aftercooler it is desirable to have the inlet
http://www.wlv.com/products/databook/ch1_4.pdf
and exit of the housing normal to the entrance and exit of
(April 15, 1998).
the core. This design practice will maintain maximum
“Tube Heat Exchanger Design.”
velocity over the core of the aftercooler, thus increasing
http://www.wlv.com/products/databook/ch1_4.pdf
heat transfer and lowering overall pressure drop. The
(April 15, 1998).
use of CFD was important in this study because there is
“Fin Heat Exchanger Design Considerations.”
no theory that directly correlates to the housing around
http://www.ijee.dit.ie/articles/Vol14-3/ijee1015.pdf
the core of an aftercooler.
(May, 2002).
“Ohio Heat Transfer Design.”
CONCLUSION
http://www.ohioheattransfer.com (2005).
Shell and tube heat exchangers and a bar and plate
aftercoolers are used in a diesel marine engine to remove
heat from the engine coolant and to cool the pressurized
air going into the engine. In order to improve the overall
effectiveness of these two components, analyses have

7 Copyright © 2006 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 11/17/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use

You might also like