Performance and Reliability of Water Jacket

Exhaust Gas Waste Heat Recovery Units

Koh Chuan Heng Erik1, MEng(Hons), MSc

SYNOPSIS
Exhaust gas Waste Heat Recovery (WHR) is a proven technique where energy from exhaust gases produced
by power plants that would have been lost, is recovered for useful purposes such as heating and power
generation, thereby reducing operational costs and emission. Research and development efforts to improve
the performance and reliability of waste heat recovery units (WHRUs) are ongoing as they are the most
crucial, but also one of the most vulnerable components in exhaust gas WHR systems. Exposure to harsh
operating conditions such as adverse temperature gradients and the corrosive nature of exhaust gases were
identified as common causes of failure. Although heat recovery performance increases with bigger
temperature difference between the exhaust gas and the working fluid, adverse temperature gradients induces
thermal stresses within and between components resulting in premature failures. This paper presents the
effects of exhaust tube length-to-diameter (L/d) ratio and placement of coolant inlet and outlet on temperature
profile, heat recovery performance, and exhaust side pressure drop in a counter-flow water jacket WHRU. A
water jacket WHRU configuration minimises redesign requirements during the retrofit of existing power
plants with WHR. The configuration also allows modularity, and offers better access for inspection and
maintenance. Non-dimensional parametric studies of the WHRU were conducted using analytical and
Computational Fluid Dynamic (CFD) models. Exhaust gas Reynolds numbers between 20,000 and 400,000,
representative of exhaust gas flow in the exhaust stacks of the MEP803A diesel generator and Rolls Royce
501-K17 gas turbine generator were used in this study. Results indicate heat recovery improved with
increasing L/d. A key finding from the study revealed optimally positioned inlet and outlet of the water jacket
improved heat recovery by up to 19%, and was also very effective at mitigating adverse temperature profiles.
This in turn contributes improves the reliability of the exhaust gas WHRU.

INTRODUCTION
Fossil fuel based power plants such as diesel engines and gas turbines are commonly used on naval ships and shore
installations for propulsion and power generation. Despite technological advances, less than 40% of the energy from
the fuel consumed is actually used for its intended purpose. Nearly 40% of the fuel energy is lost via exhaust gases
[1], [2]. With depleting sources of fossil fuel, increasing concerns of harmful emissions and growing requirements
for energy efficiency, the impetus to research and develop technologies such as waste heat recovery is expected to
increase [3]. Exhaust gas WHR is a proven technology whereby a portion of energy from exhaust gases is recovered
for useful purposes such as heating and power generation, thereby reducing operational costs and emissions [4], [5],
[6]. With the widespread use of diesel engines and gas turbines in installations both ashore and onboard ships,
organisations recognise the potential of WHR and are investing significant efforts to develop and improve WHR
technology [3], [7]. A typical layout of an exhaust gas WHR system is shown in Figure 1.
One area of development is the quest to improve the performance and reliability of exhaust gas waste heat recovery
units (WHRUs). WHRUs are essentially gas-to-gas/liquid heat exchangers where energy from hot exhaust gases is
transferred to a cooler working fluid through heat exchange surfaces such as tubes or plates. WHRUs are the most
crucial, but also one of the most vulnerable components in an exhaust gas WHR system. Although heat recovery
performance improves with larger temperature difference between the exhaust gases and working fluid, adverse
temperature gradients cause differential expansion within and between components. This induces thermal stresses
Erik currently heads the Mechanical section in RSN's Submarine Maintenance and Engineering Centre. He holds a Masters of
Science (Mechanical Engineering) from the Naval Postgraduate School and a Masters of Engineering (Marine Engineering) from
Newcastle University. His research interest includes waste heat recovery, heat and fluid flow as well as renewable energy.

which results in component distortions, stress corrosion cracking, and fatigue making it one of the most common
causes of WHRUs failures [8].

Figure 1: Layout of a typical exhaust gas WHRS used for combined heat and power production [19]
In order to maximise heat recovery, exhaust gas WHRUs are usually positioned within exhaust stacks across the
path of the exhaust gas flow. This arrangement induces back pressure in exhaust system which in turn decreases the
performance of power plants upstream. Additional back pressure is induced when heat recovery enhancement
devices such as fins are added. For a typical gas turbine, every 245 Pascal increase in exhaust gas back pressure
reduces the power output and heat rate by 0.25% and 0.08% respectively [9]. Hield [10] also highlighted that
excessive exhaust back pressure in diesel engines can increases fuel consumption, induces instability, increases
wear, overheating, and thermal failures of engine components, severely affecting the reliability of WHRUs.
During the retrofit of WHRS onto existing power plants, there is a need to adopt a total system approach. In tandem
to achieving the desired heat recovery performance, system reliability, ease of access for inspection and
maintenance, impact of power plant upstream and ability for dry operations should also be considered. This is in
response to lessons learnt from previous unsuccessful attempts to implement WHR onto existing power plants.
Despite being able to fulfill the desired WHR performance, attempts to retrofit WHRS onto United States Navy
(USN)'s DD-963 destroyers, and Canadian Navy's DDH 280 destroyers in the 1970s received limited success due to
low reliability and operational problems. The program was subsequently cancelled [11], [12], [13].
In order to mitigate the issues, a simple but novel counter-flow water jacket WHRU configuration (Figure 2) was
selected and studied using analytic and CFD models to understand the effect of different design options on heat
recovery and pressure drop performance, as well as temperature profiles; a crucial factor concerning reliability in the
WHRU.
BACKGROUND
A counter flow configuration was chosen over a parallel one due to higher heat transfer effectiveness and smaller
temperature differences between the hot and cold fluid domains throughout the WHRU. As shown in Figure 2, the
water jacket surrounds the exhaust stack of the power plant. Water is introduced into the jacket via a pipe situated on
top of the WHRU, flows through the jacket in a direction opposite to the exhaust gas, and exits the jacket through
another pipe situated at the bottom. Heat energy from the exhaust gas is transferred to the water through the exhaust
stack. This configuration allows the dimensions of existing exhaust stacks to remain unchanged.
The water jacket WHRU has the potential to be made modular; allowing it to be scaled accordingly to the exhaust
stacks of different power plants. This configuration also allows easier access for inspection and maintenance. With
no components positioned within the exhaust stack, blockage and back pressure is avoided. This allows existing
exhaust layouts and configurations of power plants undergoing WHR retrofits to remain unchanged. This in turn
minimise redesign and modification of the engine intake and exhaust system and reduces the cost and complexity of
WHR implementation onto existing power plants.

Figure 2: Schematic view of the counter-flow water jacket WHRU configuration

In this paper, the results of different length-to-tube diameter ratio (L* = L/d) and three different water inlet and outlet
placements on heat recovery performance, exhaust side pressure drop, and temperature profiles, based on analytical
and Computational Fluid Dynamic (CFD) models are presented. Jacket to exhaust stack diameter ratio (d* = D/d)
was kept constant at 1.25. Exhaust gas parameters used were based on the MEP803A Diesel Engine, commonly
used by the United States Marine Corp (USMC) as field Diesel Generators (DG); and the Rolls Royce 501K Gas
Turbine, commonly onboard USN ships as Ship Service Gas Turbine Generators (SSGTG). Steady state Reynolds
numbers of the exhaust gas flow (Reexh) from the power plants ranged from 20,000 and 400,000. In order to cover
the range of exhaust parameters and allow results to be applied to non-dimensionally similar configurations,
parameters and models were non-dimensionalised. Non-dimensional analysis using the Buckingham Pi theorem was
also conducted to consolidate and identify the parameter's relations on the impact of heat recovery performance and
exhaust back pressure.
For analytic models, heat recovery performance (q) was obtained using the effectiveness-NTU (-NTU) method.
The Gnielinski correlation, Filonerko Fanning friction factor ( ) were used to determine the required Nusselt
number, the overall heat transfer coefficient and finally the heat recovery effectiveness () required [14], [15].

q C min (Texh , in Twater , in )

(1)

Non-dimensional heat recovery ( ) for both analytical and CFD models were obtained by expressing heat recovery
as a fraction of the maximum recoverable heat energy. This provides a good indication of heat recovery performance
based on the operating environment. The equation for is shown in equation (2). Non-dimensional analysis also
revealed that heat recovery is impacted by WHRU geometry; through WHRU Length to stack diameter ratio (L/d)
and water jacket to stack diameter ratio (D/d), flow regime; through exhaust gas Reynolds number (Reexh) and fluid
properties; through Prandlt number (Prexh).
q
1
L D
q*
(
, , , Prexh )
(2)
m exh C p , exh (Texh , in Twater ,in )
Re exh d d
Equation (3) is used to estimate the back pressure induced by the analytic WHRU models. The surfaces of the
WHRU were assumed to be smooth in this study. Additional back pressure caused by surface roughness must be
considered in actual designs.

2 fLU m

Deff

(3)

The non-dimensional pressure drop ( P * ) is obtained using the equation (4) and is characterised by the WHRU
geometry represented by L/d and D/d, and flow regime represented by exhaust gas Reynolds number (Reexh). P *

is also the reciprocal the Euler number which characterizes the ratio of local pressure drop to dynamic pressure due
to fluid friction in conduits [16]. An Euler number of 1 corresponds to a perfect frictionless flow.

P
*

P
U exh
2

D L
1
, ,
)
d d Re exh

(4)

The temperature (T) generated by CFD models were non-dimensioned as a ratio to the temperature difference
between the inlet exhaust gas (Texh,in) and water (Twater,in) as shown in equation (5).

T*

T Twater ,in
Texh ,in Twater ,in

(5)

Analytical models were able to generate heat recovery and pressure drop results over a wide range of L/d, D/d ratios,
and exhaust gas Reynolds numbers. These results provided an understanding of how WHRU performance (heat
recovery and exhaust side pressure drop) are affected by changes in WHRU geometry and exhaust Reynolds
number. However, the analytical models were unable to provide resolution of the temperature profiles within the
WHRU. The analytical model was also unable to account for the effects of more complex design options such as
different water inlet and outlet placements. These limitations were resolved through the use of CFD modeling.

Figure 3: Isometric view showing locations of temperature measurements

The ANSYS 15 CFX CFD package was used to model a water jacket WHRU with a D/d of 1.25 and nondimensional tube thickness (t* = t/d) of 0.0625 to investigate the effects of L/d and other WHRU features on heat
recovery, pressure drop performance, and temperature profiles within the WHRU. The exhaust Reynolds number
ranged from 20,000 to 400,000 while the Reynolds number of the water is held constant at 8,300. Inlet exhaust
temperature (Texh,in) and inlet water temperature (Twater,in) were kept constant at 773K and 300K respectively. This
corresponds to non-dimensional temperatures of T* = 1 and T*= 0 respectively and represents the maximum and
minimum temperatures in this study. Temperatures were measured along the length of the WHRU tube in the CFD
models at 3, 6, 9, and 12 oclock positions as shown in Figure 3.
The standard K-epsilon (K-) turbulence model was used to evaluate flow within the CFD models. This model is
based on the Reynolds Averaged Navier-Stokes (RANS) and is widely accepted and implemented for turbulence
modeling [17]. The K- model is known to be stable, accurate, and numerically robust, and is considered to be the
4

standard model in the CFD industry. The K- model in CFX also uses the scalable wall-function approach which
allows solutions on arbitrarily fine near-wall grids to be made. This increased robustness and accuracy of the
solutions over standard wall functions. Heat transfer is modeled between the exhaust, water, and the WHRU tube
domain using the ANSYS-CFXs Thermal Energy model, which models the transport of enthalpy through the
fluid domains using conduction and convection [18].

CFD MODEL VALIDATION AND SENSITIVITY ANALYSIS

CFD models were validated by ensuring the overall mass and energy entering and leaving the controlled surfaces
(inlets and outlets) are balanced. In all the cases, energy balances better than 0.019% was achieved. The worst case
had an energy imbalance of less than 0.045%. Mass averaged values of inlet and outlet temperatures, specific heat
capacity, and mass flow rates were used to verify conservation of mass within the controlled volume. A sensitivity
analysis was conducted to find out the effect of different meshing size and solver target residual error value on the
result accuracy of energy balance and the resolution. Runs using three meshing levels (250,000, 500,000, and
1,500,000 nodes) and three levels of CFX solver target residual error values (1E-4, 1E-5, and 1E-6) were conducted
using a model with L/d =10, D/d= 1.25, t/d=0.0625 subjected to the same fluid flow conditions. No significant
improvement in accuracy for output parameters and energy balance were found when the higher residual error
values of 1E-6 were used over 1E-5. Models using a meshing size of 1,500,000 nodes provided slightly better
resolution in temperatures contours and profiles than models with 500,000 nodes. The model with 250,000 nodes
provided the least resolution in terms of contours but took the least amount of time to achieve solution convergence.
The models with increasing mesh sizes took increasing amounts of time run. CFD models with a meshing size of
500,000 nodes and solver residual target error value of 1E-6 offered optimum balance of accuracy, resolution, and
processing time.
In addition, results from CFD models were compared with corresponding results from analytical models. Results
plotted in Figure 4 show heat recovery from CFD and analytical models exhibited similar trends with same orders of
magnitude. Heat recovery improves with higher L/d and lower exhaust Reynolds number. At higher Reexh of
400,000, q* from the CFD models is 6% to 26% lower than corresponding results analytical models. At lower
Reynolds numbers of 20,000, q* from the CFD models are 6% to 24% higher than in analytical models. A
difference up to 26% exists between the results from the CFD and analytical model. The difference increased with
higher L/d ratio and Reynolds numbers. However, it is not unreasonable to expect differences between the CFD and
analytical models, considering the simplifications and assumptions associated with the analytical model. The
difference between the water inlet and outlet placements of the CFD and analytical models would also contributed
toward this difference. In addition, uncertainty of approximately +10% also exists in the analytical correlations used
[14] [15].

Fig 4: Non-dimensional heat recovery (q*) and Non-dimensional pressure drop (P*) of analytical and CFD models
of L/d = 5, 10 and 20 with Reexh from 20,000 to 400,000. D/d=1.25, centerline water inlet / outlet placement Type 1
with Rewater = 8,300.

EFFECT OF WHRU LENGTH TO TUBE DIAMETER (L/d) RATIO

Results from both analytical and CFD models shown in Figure 4 shows that higher L/d increase heat recovery under
similar exhaust gas flow conditions. This was expected as more available area for heat transfer is available with
higher L/d. Increasing L/d from 10 to 20 increases heat recovery by 68% to 81% whereas a reduction of L/d to 5
decreased heat recovery by 41% to 46%. Better heat recovery performance was observed at lower exhaust gas
Reynolds numbers. The maximum possible heat recovery was achieved if the length of the WHRU was long
enough, i.e., L/d=1,000. However, once the maximum heat recovery is achieved (q*=1), further increase of WHRU
length did not improve the heat recovery. On the contrary, additional pressure drop at the exhaust gas side was
incurred. This causes back pressure, negatively impacting the performance and fuel consumption of the power plant
upstream. Results in Figures 4 shows that increasing the L/d from 10 to 20 amplified P* by 96%; while a decrease
of L/d from 10 to 5 reduced P* by 45%. This result was expected as frictional loss is proportional to WHRU
length. It was also observed that P* decreases as the exhaust Reynolds number increases. For the all CFD models,
P* leveled off at Reynolds numbers of 150,000 to 200,000. This is expected for flows that are modeled over
smooth surfaces. CFD models provided valuable insights into the longitudinal and radial temperature distributions
within the counter flow water jacket WHRU. The temperature profiles and contours of the exhaust stack for
WHRUs with L/d ratios of 5, 10, and 20 at exhaust Reynolds numbers of 20,000 and 400,000 are shown in Figure 5
and 6 respectively.

L/d=5

L/d=10

L/d=20

Fig 5: Temperature profile for WHRU with L/d = 5, 10 and 20, D/d = 1.25, t* = 0.0625 at Re exh = 20,000 and
400,000; centerline water inlet / outlet placement Type 1. Re water = 8,300.

L/d = 5

L/d = 10

L/d = 20

Figure 6: Temperature contours of exhaust stack for WHRU with L/d = 5, 10 and 20 at D/d=1.25, t*=0.0625 at Reexh
= 400,000 with centerline water inlet/outlet placement Type 1. Re water = 8,300.

Generally, higher exhaust gas Reynolds numbers increase tube temperatures and amplify temperature differences.
The sharpest temperature gradients were observed at Reexh = 400,000. Temperature profiles on Figure 5 shows that
regions of adverse temperature differences shift along the WHRU with different L/d. A WHRU with L/d = 5, the
largest difference occurs at the lower half of the WHRU where the exhaust inlet and water outlet are situated. When
L/d is increased to 10, this difference is most prominent in the middle. WHRU with L/d = 20 has the most adverse
temperature differences occur at the upper end of the exhaust stack; near the location of the exhaust outlet and the
water inlet. This insight allows potentially problematic areas to be identified for mitigating measures.

Figure 7: Water streamlines depicting uneven water flow distribution inside water jacket of WHRU of L/d =10,
D/d=1.25 with centerline water inlet/outlet placement Type 1 at Reexh = 400,000, Rewater = 8,300.

Another key finding was the uneven flow distribution within the water jacket. This was revealed by examining the
streamlines within the water jacket region (Figure 7). Most of the water entering the water jacket was channeled
towards the 12 oclock profile, as opposed to the 6 oclock profiles (where the water inlet and outlet were located).
Flow distribution along the 3 oclock and 9 oclock profiles were balanced. As such, uneven heat transfer resulted in
adverse temperature profiles between the 6 oclock and 12 oclock profiles. This finding was evident in all WHRU
L/d ratios; as shown in Figure 5 and 6.
This difference was particularly pronounced at Reexh= 400,000. The 6 oclock and 12 oclock temperature profiles
WHRU of for L/d = 5, 10, and 20 at Reexh of 400,000 are plotted in Figure 8 for better comparison. Adverse
temperature profiles causes localized thermal stresses due to differential expansions in the tube. Differential stresses
are produced both longitudinally as well as radially along the tube and severely impact the reliability of WHRUs.
During transient loadings such as startups, shutdowns, or load changes, these stresses could potentially be amplified
due to transient and unsteady temperature differences.

Figure 8: Comparison of temperature profile at 6 oclock and 12 oclock at Re exh = 400,000, Rewater = 8,300.for
WHRU of L/d = 5, 10, and 20
EFFECT OF WATER INLET AND OUTLET PLACEMENT
In order to mitigate the adverse temperature profiles within the WHRU and improve reliability, WHRU models with
three different types of water inlet and outlet placements as shown in Figure 9 were investigated. Type 1 and 2 are
commonly found, unlike Type 3; where the water inlet and outlet are shifted from the centre line in attempt to
improve the flow distribution within the water jacket. Results in Figure 10 shows that there are no significant
difference in heat recovery performance for WHRUs with water placement Type 1 and 2. However, heat recovery
improved by up to 19% when a WHRU equipped with placement Type 3 was used. Results in Figure 10 also
showed that the type of water placement did not have any significant impact on the exhaust side pressure drop. A
WHRU with placement Type 3 produced exhaust side back pressure 1% to 3% lower compared to WHRU with
placement Type 1 and 2.
Placement Type 1
Centerline water inlet and outlet
on 6 o'clock side

Placement Type 2
Centerline water inlet at 6 o'clock
side and outlet at 12 o'clock side.

Placement Type 3
Laterally shifted water inlet and outlet
at 6 o'clock side.

Figure 9: Three types of water inlet and outlet placements for counter flow water jacket WHRU of L/d=10,
D/d=1.25 t*=0.0625.

Figure 10: Comparison of non-dimensional heat recovery and (q*) and dimensional pressure drop (P*) for water
inlet / outlet placement Types 1, 2, and 3 for WHRU of L/d=10, D/d=1.25, t* = 0.0625. Reexh from 20,000 to
400,000, Rewater = 8,300.
Results in Figure 11 show large temperatures differences between the 6 o'clock and 12 o'clock profiles for water
Type 1 and Type 2 placements. Sharp temperature gradients were also present near the exhaust outlet and water inlet.
The placement of the inlet and outlets on opposite sides of the WHRU (Type 2 placement) resulted in temperature
profiles which were more adverse, negatively impacting the WHRUs reliability. This finding is interesting
considering that placement Type 1 and 2 are very common in the heat exchanger industry.
In contrast, placement Type 3 which had water inlet and outlet located off centre produce gradual temperature
profile radially and longitudinally. Inspection of the water streamlines in the water jacket and corresponding
temperature contours in Figure 12 revealed that placement Type 3 improved water flow distribution in the water
jacket significantly. Placement Type 3 can be used to mitigate adverse temperature profiles and the negative effects
of differential expansion and contraction in order to improve the WHRU's reliability.
Placement Type 1

Placement Type 2

Placement Type 3

Figure 11: WHRU temperature profiles of WHRU with for water inlet / outlet placement Type 1, 2 and 3 with
L/d=10, D/d=1.25, t*=0.0625. Reexh = 400,000, Rewater = 8,300.

Placement Type 1

Placement Type 2

Placement Type 3

Figure 12: Water streamlines and exhaust stack temperature contours for WHRU with water placement type 1, 2 and
3 L/d = 10; D/d=1.25, t*=0.0625 at Reexh = 400,000; Rewater = 8,300.

CONCLUSION
The paper presents the effects of exhaust tube length-to-diameter (L/d) ratio and coolant inlet and outlet placements
on temperature profile, heat recovery performance, and exhaust side pressure drop in a counter-flow water Jacket
WHRU. Non-dimensional parametric studies of the WHRU were conducted using analytical and CFD models.
Results indicate that heat recovery increases with higher L/d. However, once the maximum possible heat recovery is
achieved, further increase in L/d only increases the exhaust gas back pressure of the power plant upstream.
Analytical models were able to predict the heat recovery, back pressure performance but were unable to provide
insights on temperature profiles and local fluid behavior within the WHRU. The CFD models generated provided
insights of the flow distribution and corresponding temperature profiles allowing potentially problematic areas to be
identified for mitigating actions.
One key finding from the study is the significance of water inlet and outlet placement toward the control of the
temperature profile within the WHRU. WHRUs with water inlets and outlets located in the centerline of the WHRU
(Type 1 and Type 2) experienced poor and uneven distribution of flow in the water jacket. The finding related to
centerline water inlet and outlet placement was consistently found in WHRUs with different L/d ratios. Poor
distribution of water flow resulted in adverse temperature profiles between the 6 oclock and 12 oclock positions
and sharp temperature gradients at the upper section of the WHRU where the exhaust outlet and water inlet are
located. Adverse temperature profile are known to produce differential expansions and thermal stress, initiating
failure mechanisms such as fatigue and stress corrosion cracking that increase the probability of premature WHRU
failures. The poor distribution of water flow also reduced the heat recovery potential of the affected WHRU. This
finding is crucial as centerline water inlet and outlet placements are very prevalent in cylindrical-shaped WHRUs
and heat exchangers. The problem was subsequently resolved by shifting the water inlet and outlet off center in
placement Type 3. This change improved the flow in the water jacket and eradicated the adverse temperature
profiles. Heat recovery performance of the WHRU also increased as much as 19% with no increase in exhaust side
pressure loss.
The study provides insights into the complex relationships involved in the design of a WHRU. Designers or program
managers of waste heat recovery projects need to consider performance and reliability from a total system point of
view. The increase in heat recovery is often linked to an increased exhaust side pressure drop, which in turn impacts
the power plant upstream in a detrimental manner. In order to further optimise reliability and performance, followon study of the water inlet and outlet of water jacket WHRUs could be conducted looking at the effects of shapes,
vertical or horizontal locations, and sizes of the water inlet and outlet.
ACKNOWLEDGEMENTS
The study was part of U.S. Navy WHR Capability Roadmap undertaken by the Naval Postgraduate School. The
support and guidance from Dr Sanjeev B. Sathe and Dr Knox Millsaps from the Mechanical and Aerospace
Engineering Department of Naval Postgraduate School during the course of study is gratefully acknowledged.

10

REFERENCES

1. Woodyard, D. (2004). Pounders marine diesel engines and gas turbine (8th ed.). Oxford, UK: Elsevier Ltd.
2. Bailey, J, Hardy, Bob, ' The Green Ship Challenge- Delivering more with less', The Naval Engineer, Autumn
2014, pp12-18 (1 Dec 2014)
3. MARKETSANDMARKETS [Internet]. c2009-2015. Dallas: Markets and Markets; [Publishing Date: April 2014
Report Code: CH 2382, cited 2015 Mar 2]. Available from: http://www.marketsandmarkets.com/MarketReports/waste-heat-recovery-system-market202657867.html?gclid=Cj0KEQjw6OOoBRDP9uG4oqzUv7kBEiQA0sRYBFgJE7w3KnAHTujU7ZxaOFPpX3C7dF67T7CIhyWNcoaAuwB8P8HAQ
4. Bailey, M. M. (1985). Comparative evaluation of three alternative power cycles for waste heat recovery from the
exhaust of adiabatic diesel engines. Washington, DC: U.S. Department of Energy.
5. BCS. (2008). Waste heat recovery: Technology and opportunities in U.S. industry. Washington, DC: U.S.
Department of Energy.
6. Carapellucci, R., & Giordano, L. (2012, April). The Recovery of Exhaust Heat from Gas Turbines. In V.
Konstantin (Ed.), Efficiency, performance and robustness of gas turbines (pp. 165190). InTech.
doi:10.5772/37920.
7. Deputy Assistant Secretary of the Navy (DASN) Energy Office. (2010, October). Department of the Navys
Energy Program for Security and Independence. Retrieved August 9, 2014, from http://www.navy.mil/secnav/
8. Dooley, R., Paterson, S., & Pearson, M. (2005). Diagnostic/troubleshooting monitoring to identify damaging
cycle chemistry or thermal transients in heat recovery steam generator pressure parts. Palo Alto, CA: Electrical
Power Research Institute.
9. Boyce, M. P. (2012). Effect of controllable losses to the output and heat rate. In Gas turbine engineering
handbook (4th ed., p. 801). Houston, TX: Gulf Professional Publishing.
10. Hield, P. (2011). The effect of back pressure on the operations of a diesel engine. Victoria, Australia: Maritime
Platforms Division, Defense Science and Technology Organization.
11. Breaux, D., & Davies, K. (1978, April). Design and service of a marine waste heat boiler. Naval Engineers
Journal, 165178.
12. Rains, D. A., Beyer, K. M., Keene, W., Lindgren, J. J., Mogil, E., Page, T., . . . Youngworth, J. (1976, October).
Design appraisal - DD-963. Naval Engineers Journal, 4361.
13. Mastronarde, T. P. (1982, April). Energy conservation utilizing waste heat boilers - The challenges, problems
and solutions. Naval Engineers Journal, 277285.
14. Kakac, S., & Liu, H. (2002). Heat Exchangers Selection, Rating and Thermal design (2nd ed.). Danvers, MA:
CRC Press.
15. Incropera, F. P., Bergman, T. L., Lavine, A. S., & Dewitt, D. P. (2011). Introduction to heat transfer (6th ed.).
Hoboken, NJ: John Wiley & Sons, Inc.
16. Yarin, L. (2012). The Pi-Theorem: Applications to fluid mechanics and heat and mass transfer. New York:
Springer.

11

17. Patankar, S. V. (1980). The turbulence-kinetic-energy equation. New York: Hemisphere Publishing Corporation.
18. ANSYS. (2013, November). ANSYS CFX-Solver theory guide. Canonsburg, PA: ANSYS, Inc.
19. CHP Focus. (2014). Gas turbine performance. Retrieved August 10, 2014, from http://chp.decc.gov.uk/cms/gasturbine-performance/

12