Large Marine Engine Technology
Evaluation
Final Report
rnA United States
Environmental Protection
Agency
-------
Large Marine Engine Technology
Evaluation
Final Report
Office of Transportation and Air Quality
U.S. Environmental Protection Agency
Prepared for EPA by
Southwest Research Institute (SwRI)
EPA Contract 68HERC20D0014
Task Order 68HERC24F0014
NOTICE
This technical report does not necessarily represent final EPA decisions
or positions. It is intended to present technical analysis of issues using
data that are currently available. The purpose in the release of such
reports is to facilitate the exchange of technical information and to
inform the public of technical developments.
United States
mpfaEnvironmental Protection
Agency
EPA-420-R-25-007
February 2025
-------
LARGE MARINE ENGINE TECHNOLOGY EVALUATION
EPA CONTRACT 68HERC20D0014
TASK ORDER 68HERC24F0432
OMB CLEARANCE NUMBER 2030-0005
FINAL REPORT
SwRI® Project Number 03.28987
Prepared for:
U.S. Environmental Protection Agency
Prepared by:
Michael G. Ross - Program Director
Christopher A. Sharp - Institute Engineer
Timothy J. Callahan - Staff Engineer
Garrett L. Anderson - Principal Engineer
Kartik G. Adsule - Research Engineer
Southwest Research Institute
6220 Culebra Road
San Antonio, TX 78238
February 28, 2025
Benefiting government, industry and the public through innovative science and technology
-------
LARGE MARINE ENGINE TECHNOLOGY EVALUATION
EPA CONTRACT 68HERC20D0014
TASK ORDER 68HERC24F0432
FINAL REPORT
SwRI® Project Number 03.28987
Prepared for:
U.S. Environmental Protection Agency
February 28, 2025
Prepared by: Approved by:
Michael G. Ross, Program Director Charlie E. Roberts, Jr., PhD, Executive Director
Commercial Powertrain Systems Dept. Commercial Powertrain Systems Dept.
POWERTRAIN ENGINEERING DIVISION
This report must be reproduced in full, unless SwRI* approves a summary or abridgment.
Final Report 03.28987
-------
EPA Task Order 68HERC24F0432
DISCLAIMER
The statements and conclusions in this report are those of the contractor and not necessarily
those of the Environmental Protection Agency. The mention of commercial products, their source,
or their use in connection with material reported herein is not to be construed as actual or implied
endorsement of such products.
Final Report 03.28987
11
-------
EPA Task Order 68HERC24F0432
ACKNOWLEDGEMENTS
This Report was submitted in fulfillment of EPA Task Order 68HERC24F0432 under
contract 68HERC20D0014 by Southwest Research Institute under the sponsorship of the
Environmental Protection Agency.
The cognizant EPA staff members for this task order include:
Charles Fischer - Contracting Officer
Nicolas Witkowski - Contract Level Contracting Officer Representative
Maria Lennox - Task Order Contracting Officer Representative
Chris Laroo - Alternate Task Order Contracting Officer Representative
Jean Marie Revelt
Michael Aldridge
SwRI acknowledges and appreciates the important contributions of marine engine
consultant Dr. Udo Schlemmer-Kelling to the marine engine technology assessment and cost
analysis aspects of this study.
Final Report 03.28987
111
-------
EPA Task Order 68HERC24F0432
TABLE OF CONTENTS
Page
ABSTRACT xiii
EXECUTIVE SUMMARY xiv
1.0 INTRODUCTION AND BACKROUND 1
2.0 ASSESSMENT OF OCEAN-GOING VESSEL LOW-LOAD OPERATION 3
2.1 Datasets 4
2.2 Tools Used for Data Analysis 6
2.3 Region Definition 7
2.4 Description of the Ship Included in the Study 9
2.5 A IS Data 17
2.6 Duty Cycle Estimate & Results 22
2.6.1 Load Estimation 22
2.6.2 Estimated Duty Cycles 27
3.0 CATEGORY 3 MARINE DIESEL ENGINE TECHNOLOGY AS SES SMENT 33
3.1 Tier III Ships in the U.S. EC A 33
3.2 Two-stroke Engine Overview 36
3.3 Current Tier III Thermal Management Technologies 39
3.4 Applicability of Heavy-Duty Engine Thermal Management Technologies 42
3.7.1. Intake Throttling/AFR Reduction 43
3.7.2 Aftertreatment System Heating 43
3.7.3 Exhaust Flow Bypass System 44
3.7.4 Late Combustion Phasing 44
3.7.5 Variable Valve Actuation 45
3.7.6 Cylinder Deactivation 46
3.7.8 Pre-turbine Aftertreatment Location 47
3.7.9 Aftertreatment Insulation 50
3.7.10 ULSD 50
3.5 Other Thermal Management or SCR Technologies 51
3.8.1 Intake Air Heating 51
3.8.2 Direct Use of Ammonia 52
3.6 Summary 53
4.0 MARINE DIESEL ENGINE NOx AFTERTREATMENT AND TECHNOLOGY
PACKAGE ASSESSMENT 56
4.1 Exhaust Temperature Profile for Marine 2-Stroke Engines 56
Final Report 03.28987 iv
-------
EPA Task Order 68HERC24F0432
4.2 Brake Specific NOx Profiles for Marine 2-Stroke Engines 58
4.3 Marine NOx Control Technology Selections for NOx Estimates 58
4.4 Ammonium Bisulfate Deposit Formation Assessment 61
4.5 Selective Catalytic Reduction Performance Curves 65
5.0 ASSESSMENT OF POTENTIAL FOR ADDITIONAL NOx REDUCTIONS IN
EMISSION CONTROL AREAS 70
6.0 SUMMARY AND CONCLUSIONS OF NOx IMPROVEMENT TECHNOLOGY
SCENARIO ANALYSIS 86
7.0 REFERENCES 87
Final Report 03.28987
V
-------
EPA Task Order 68HERC24F0432
LIST OF FIGURES
Figure Page
Figure 1. Load Frequency Distribution for Tier III Vessels in the North American ECA xiv
Figure 2. NOx Mass Distribution by Load Bin for Tier III Vessels in the North American ECA
xv
Figure 3. High Level Diagram of Ship Data Analysis Process 3
Figure 4. Estimated Duty Cycle of Tier III Category 3 Ships Operating Within the North
American and Caribbean Sea EC As (N=255) 3
Figure 5. Map of North American & Caribbean Sea Emission Control Areas as received from
EPA 5
Figure 6. Map of the Boundaries Defined in the C3RIA Dataset as Supplied by EPA 6
Figure 7. Open-Source Tools Used for Analysis & Visualization (Geopandas Developers, 2024)
(OpenStreetMap Foundation, 2024) (Plotly, 2024) (PostgreSQL Global Development, 2024)
(Python Software Foundation, 2025) (QGIS Development Team, 2024) 7
Figure 8. Boundaries Used for Comparing Duty Cycles between ECA Regions 8
Figure 9. Figure Showing Inputs and Intermediate Steps to Derive Regions with Arrows
Indicating Data used for each Step 9
Figure 10. Histogram of IMO NOx Tier from Clarksons World Fleet Register for Ships Included
in the Study (N=12,303) 10
Figure 11. Histogram of Category 3 ShipTypes from Clarksons World Fleet Register
(\ 12.303) 11
Figure 12. Histogram of Main Propulsion Engine Types from Clarksons World Fleet Register for
Category 3 Ships (N=7,264) 11
Figure 13. Histogram of Bore Diameter for Main Propulsion Engines Included in the Study from
Clarksons World Fleet Register (N=12,303) 12
Figure 14. Histogram of Stroke for Main Propulsion Engines Included in the Study from
Clarksons World Fleet Register (N=12,303) 13
Figure 15. Histogram of Individual Cylinder Displacement for Main Propulsion Engines
Included in the Study from Clarksons World Fleet Register (N=12,303) 13
Figure 16. Histogram of Engine Cylinder Count for Main Propulsion Engines Included in the
Study from Clarksons World Fleet Register (N=12,303) 14
Figure 17. Histogram of Engine Displacement for Main Propulsion Engines Included in the
Study from Clarksons World Fleet Register (N=12,303) 15
Figure 18. Histogram of Main Engine Power for Main Propulsion Engines Included in the Study
from Clarksons World Fleet Register (N=12,303) 15
Figure 19. Histogram of Maximum Ship Speed from Clarksons World Fleet Register (N=630) 16
Figure 20. Histogram of Ship Service Speed from Clarksons World Fleet Register (N=8,502). 17
Figure 21. Heat Maps Based on AIS Data with IMO Number Comparing Traffic Patterns for All
Ships Wiyh AIS Data vs Tier III Ships 19
Figure 22. Bar Chart of ship location Counts Inside and Outside of ECA for AIS Data with an
IMO Number Joined to the Clarksons Dataset (N=167,512,041) 20
Final Report 03.28987 vi
-------
EPA Task Order 68HERC24F0432
Figure 23. Bar Chart of Point Counts by Region for AIS Data with an IMO Number Joined to
the Clarksons Dataset (N= 103,552,955) 21
Figure 24. Bar Chart of Unique Vessel Count by Region for AIS Data Based on IMO Number
Joined to the Clarksons Dataset 22
Figure 25. Flow Chart Describing the Process Used to Calculate Percent Load Based upon the
Availability of Maximum Speed or Service Speed in Clarksons Ship Specifications 23
Figure 26. Venn Diagram and Table of Ships with Maximum Speed, Service Speed, Both
Speeds, or No Speed Specified 23
Figure 27. Sankey Diagram of AIS Point Filtering to Identify Useful Points for Duty Cycle
Development 24
Figure 28. Sankey Diagram Depicting Impact of Filtering on IMO Numbers Used for Duty
Cycle Development 25
Figure 29. Histogram Of Time Spent in ECA in Hours (N=7,357) 27
Figure 30. Estimated Duty Cycle of Tier III Category 3 Ships Operating Within the North
American ECA (N=255) 28
Figure 31. Estimated Duty Cycle for Tier II Category 3 Ships Identified in North American ECA
(\ 4.040) 28
Figure 32. Estimated Duty Cycle for Tier I Category 3 Ships Identified in North American ECA
(\ 1.960) 29
Figure 33. Line Chart Comparing Cumulative Distribution of Duty Cycle for Tier I, Tier II,
Tier III, and Certification Test Cycle 29
Figure 34. Estimated Duty Cycle by Ship Type for Tier III Ships Identified within the North
American ECA (N=255) 30
Figure 35. Estimated Duty Cycle by Ship Type for Tier II Ships Identified within the North
American ECA (N=4,040) 30
Figure 36. Estimated Duty Cycle by Ship Type for Tier I Ships Identified within the North
American ECA (N=l,960) 31
Figure 37. Estimated Duty Cycle by Region for All Tier III Ships Identified within the North
American ECA (N=255) 31
Figure 38. Estimated Duty Cycle by Region for All Tier II Ships Identified within the North
American ECA (\ 4.040) 32
Figure 39. Estimated Duty Cycle by Region for All Tier I Ships Identified within the North
American ECA (\ 1.960) 32
Figure 40. Tier III, Category 3, Vessels in the US ECA 34
Figure 41. Comparison of Emissions from Two Dual Fuel Technology Approachs Relative to
Diesel (Werner, 2019) 35
Figure 42. Main Diesel Engine Designer 35
Figure 43. Tier III Options for MAN 60ME Engine 36
Figure 44. Two-Stroke Engine Layout and Operation (MAN Energy Solutions, 2021) 37
Figure 45. Intake and Exhaust Pressure Characteristics of Typical 2-Stoke Engine (Illustration
only) 38
Final Report 03.28987 Vll
-------
EPA Task Order 68HERC24F0432
Figure 46. Intake Air Pressure Improvement with Auxiliary Blower (Illustration Only) 38
Figure 47. Exhaust Temperature Range for Tier II Engines 39
Figure 48. Comparison of Tier II and Tier III Before Turbine Exhaust Temperature for WinGD
8X52 Engine 40
Figure 49. Turbine Bypass Flow for WinGD 8X52 Engine 40
Figure 50. Estimated AFR for Tier II and Tier III Operation 41
Figure 51. SCR Performance and SCR Inlet Temperatures (Fujibayashi, T., et. al., 2013) 41
Figure 52. Effect of Cylinder Deactivation on Exhaust Temperatures of a 16-Cylinder, EMD
Locomotive Two-Stroke Engine at Light Load (Fritz & Riley, 2024) 47
Figure 53. Choice of IMO NOx Tier III Strategy for MAN Two-Stroke Engines in Relation to
Engine Size (Struckmeier, D., et al., 2019) 48
Figure 54. Example of High-Pressure SCR System Layout (Fujibayashi, T., et. al., 2013) 49
Figure 55. Example of WinGD iSCR, Integrated HPSCR System (Kyrtatos, A., et.al., 2016) ... 49
Figure 56. Exhaust Flow Path of WinGD iSCR in Tier II (Left) and Tier III (Right) Modes
(Spahni, M., et. al., 2023) 50
Figure 57. Illustration of Urea Decomposition Unit (DCU) and Burner for LP-SCR Systems
(MAN Energy Solutions, 2021) 53
Figure 58. Exhaust Temperature Data for Marine Two-Stroke Engine 57
Figure 59. Exhaust Temperature versus Load Profile for WinGD X52 Series Engine 57
Figure 60. BSNOx versus Load for a Marine Two Stroke Engine (MAN Energy Solutions, n.d.)
58
Figure 61. scr Reactor Inlet Temperatures for Technology Projections compared to Tier III
Current 60
Figure 62. Engine-Out NOx Curve for Technology Projections 61
Figure 63. ABS Condensation as a Function of Reactant Concentration (Muzio et. al) 63
Figure 64. Influence of Pressure on ABS Condensation Temperature 63
Figure 65. SCR NOx Conversion Curves used for Tier iii Marine Diesel Engine NOx
Projections 67
Figure 66. Ammonia Slip Curve Example versus stoichiometric Ammonia-to-NOx Ratio 68
Figure 67. Tier III ECA Load Distribution - All Vessels 71
Figure 68. Tier III Baseline NOx Distribution for the ECA 72
Figure 69. Comparison of Power Distribution versus NOx Distribution for Tier III Vessels in the
ECA 73
Figure 70. NOx Mass Distribution for ECA Mode ReWeighting Scenario and Current Standard
74
Figure 71. NOx Mass Distribution for Best Case Technology Scenario 75
Figure 72. BSNOx and SCR NOx Conversion by Load Bin for Best Case Technology Scenario
76
Figure 73. Comparison of NOx Mass Distributions for Best Case Technology Scenario and
Tier III Baseline SCENARIO 77
Final Report 03.28987 Vlll
-------
EPA Task Order 68HERC24F0432
Figure 74. NOx Mass Distribution for Maximum Feasible NOx Curve, Intermediate
Temperature Curve, Targeting Conversion Down to 10% Load Scenario 79
Figure 75. BSNOx and SCR NOx Conversion by Load Bin for Maximum Feasible NOx Curve,
Intermediate Temperature Curve, Targeting Conversion Down to 10% Load Scenario 79
Figure 76. Comparison of NOx Mass Distributions for Maximum Feasible NOx Curve,
Intermediate Temperature Curve, Targeting Conversion Down to 10% Load Scenario and
Tier III Baseline 80
Figure 77. NOx Mass Distribution for more Conservative (Low-Temp Only) NOx Curve,
Intermediate Temperature Curve, Targetting Conversion Down to 10% Load Scenario 81
Figure 78. BSNOx and SCR NOx Conversion by Load Bin for Maximum Conservative NOx
Curve, Intermediate Temperature Curve, Targeting Conversion Down to 10% Load Scenario
81
Figure 79. Comparison of NOx Mass Distributions for Conservative (Low-Temp Only) NOx
Curve, Intermediate Temperature Curve, Targeting Conversion Down to 10% Load Scenario
and Tier III Baseline 82
Figure 80. NOx Mass Distribution for Conservative SCR with Best Case Temperature
Distribution, Targeting Conversion Down to 5% Load Conversion Only 84
Figure 81. BSNOx and SCR NOx Conversion by Load Bin for Conservative SCR with Best
Case Temperature Curve Scenario, Targeting Conversion Down to 5% 84
Figure 82. Comparison of NOx Mass Distributions for Conservative SCR with Best Case
Temperature Curve Scenario, Targeting Conversion Down to 5% 85
Final Report 03.28987
IX
-------
EPA Task Order 68HERC24F0432
LIST OF TABLES
Table Page
Table 1. Summary ofNOx Improvement Technology Scenarios Examined xv
Table 2. Table of Software Versions Used for Analysis and Visualization 7
Table 3. Table Describing the Mapping of Ship Classifications in Clarksons to Ship Grouping
Used for Study and Their Distributions within the Clarksons Data 26
Table 4. Table Describing the Mapping of Ship Classifications in Clarksons to Ship Grouping
Used for Study and Their Distributions within the Ship Population Used for Duty Cycle
Development 26
Table 5. Thermal Management Technology Evaluation Criteria 42
Table 6. Thermal Management Technology Evaluation 55
Table 7. ABS Condensaton Temperature Estimates at 5-10% Load 65
Table 8. Actual Operational Distribution in ECA versus E2/E3 TEST Cycle Weighting 71
Table 9. IMO E3 Test Cycle Result for Baseline Tier III Scenario 72
Table 10. NOx Result for Tier III Baseline using Actual Tier III ECA Duty Cycle Weighting,
assuming 82% Peak Efficiency 74
Table 11. Cycle Weighted Emission Results for Current Modes for Best Case Technology
Scenario 77
Table 12. Cycle Results on Best Case Technology Scenario for Updated Test Cycle with 10%
Mode and Updated Duty Cycle Weighting for Tier III 78
Table 13. Cycle Results on Maximum Feasible SCR Curve, Intermediate Temperature Curve,
Targeting Conversion Down to 10% Load for Test Cycle with 10% Mode and Updated
Duty Cycle Weighting for Tier III 80
Table 14. Cycle Results on Conservative (low-temp only) Scenario Targeting Conversion Down
to 10% Load for Test Cycle with 10% Mode and Updated Duty Cycle Weighting for Tier iii
83
Table 15. Cycle Results on Conservative SCR with Best Case Temperatures Scenario for
Updated Test Cycle with 10% Mode and Updated Duty Cycle Weighting for Tier III 85
Table 16. Summary ofNOx Improvement Technology Scenarios Examined 86
Final Report 03.28987
X
-------
EPA Task Order 68HERC24F0432
LIST OF ACRONYMS
ABS
Ammonium Bisulfate
ACD
Auxiliary Control Device
AIS
Automatic Identification System
AFR
Air to Fuel Ratio
ANR
Ammonia to NOx Ratio
ASC
Ammonia Slip Catalyst
AT
Aftertreatment
BMEP
Brake Mean Effective Pressure
BSFC
Brake Specific Fuel Consumption
CAPEX
Capital Expense
CARB
California Air Resources Board
CFR
Code of Federal Regulations
CO
Carbon Monoxide
C02
Carbon Dioxide
COV
Coefficient of Variation
CWF
Carbon Weight Fraction
DCU
Decomposition Unit
DEF
Diesel Exhaust Fluid
deSOx
Desulfation
DOC
Diesel Oxidation Catalyst
DPF
Diesel Particulate Filter
ECA
Emission Control Area
ECM
Electronic Control Module
EEVO
Early Exhaust Valve Opening
EGR
Exhaust Gas Recirculation
EGT
Exhaust Gas Temperature
EHC
Electrically Heated Catalyst
EO
Engine Out
EPA
Environmental Protection Agency
HD
Heavy Duty
IMO
International Marine Organization
IMT
Intake Manifold Temperature
MARPOL
International Convention for the Prevention of Pollution from Ships
MCR
Maximum Continuous Rating
MY
Model Year
NO A A
National Oceanic and Atmospheric Administration
NOx
Oxides of Nitrogen
NTE
Not-to-Exceed
OGV
Ocean-going Vessel
OPEX
Operating Expense
PGM
Platinum Group Metals
PM
Particulate Matter
RAM
Random Access Memory
SCR
Selective Catalytic Reduction (ammonia-based)
SOI
Start of Injection
Final Report 03.28987
xi
-------
EPA Task Order 68HERC24F0432
SwRI
Southwest Research Institute
TDC
Top dead center
THC
Total Hydrocarbons
TM
Thermal Management
ULSD
Ultra Low Sulfur Diesel
USCG
United State Coast Guard
VCR
Variable Compression Ratio
VVA
Variable Valve Actuation
Final Report 03.28987
Xll
-------
EPA Task Order 68HERC24F0432
ABSTRACT
MARPOL Annex IV Regulation 13 requires Tier III NOx reduction technology to be used
within NOx Emission Control Area (ECA). Many Tier III Category 3 (> 30L per cylinder) engines
utilize selective catalytic reduction (SCR) to reduce NOx. The E2 and E3 propulsion test cycles
only test engine emissions down to 25% power and the program allows Administrations to approve
use of auxiliary control devices (ACDs) that permit disengaging the SCR system when the exhaust
gas temperature drops below the SCR minimum operating temperature. This typically occurs
below 25% power, if the engine was not calibrated with thermal management strategies to maintain
the exhaust gas temperature, which is always the case. Because vessels may choose to travel more
slowly within EC As due to economic or operational considerations, NOx reduction benefits may
be curtailed due to frequent operation at less than 25% power.
This study evaluated engine load profiles within the North American ECA for Category 3
ocean-going vessels (OGVs) and assessed the feasibility of extending SCR operation below 25%
power. The findings include: 1) A large percentage of operation in the ECA is less than 25%
power, 2) Existing Category 3 engine technology can maintain adequate exhaust gas temperature
for SCR operation below 25% power, and 3) NOx emissions within ECAs by Tier III Category 3
vessels can be significantly reduced by utilizing SCR at less than 25% power.
Final Report 03.28987
Xlll
-------
EPA Task Order 68HERC24F0432
EXECUTIVE SUMMARY
As part of an ongoing effort to assess the NOx ECAs and how they could be improved, this
study investigated the following Tier Ill-related issues. First, whether the efficiency of
SCR emission control technology on Category 3 vessels can be improved, both underway and
while operating at low load, based on the current engine certification process. Second, an
evaluation of ship operating profiles in the ECAs irrespective of engine category, specifically how
much time is spent at low load operation and, using that information, how the certification test
cycles could be revised to incorporate that operation, and if exhaust gas temperature can be
maintained to extend SCR operation to 10% power or lower.
SwRI engaged in a substantial effort to develop a clear picture of the operating profile of
real Category 3 vessels operating in the North American ECA. This was a key input for examining
the potential for additional NOx reductions. The resulting operating histogram is shown in Figure
1 below. These values indicate that 42% of all operation in the ECA occurs below the 25% load
point, and 75% of all operation in the ECA occurs below the 50% load point.
20%
18%
5" 16%
| 14%
-------
EPA Task Order 68HERC24F0432
Tier III ECA Exhaust Stack NOy Distribution
20%
18%
16%
14%
12%
Z 10%
i 8%
Jj 6%
"cc
E 4%
o
z 2%
0%
CN CN CO CO
O LO O LO O LO
in m CD CD
) Power
| y 1% 1% 0o/o 10%
o m o m o
00 00 cr> o
o in o in o ^
oo oo o) m
O)
FIGURE 2. NOx MASS DISTRIBUTION BY LOAD BIN FOR TIER III VESSELS IN
THE NORTH AMERICAN ECA
SwRI examined a number of technology scenarios based around the application of high
pressure (pre-turbine) vanadium SCR and various levels of thermal management. Both of these
are technologies that are currently used on the main propulsion engines of many Tier III vessels
today. The scenarios examined various combinations of extending NOx control to lower loads
and pushing for additional NOx conversion in the current Tier III control range above 25% load.
These scenarios, and their resulting NOx reductions are shown in Table 1 below.
TABLE 1. SUMMARY OF NOx IMPROVEMENT TECHNOLOGY SCENARIOS
EXAMINED
Scenario
Details
Minimum Load
NOx Reduction
Standard Notes
1
Current Mode Reweights
25%
9%
Reweigh to ECA Duty Cycle
2
Max SCR, BestTemperatures
5%
53%
Add 10% Mode and 5% Mode Cap, Tighten
Standard Limits
3
Max SCR, Intermediate
Temperatures
10%
45%
Add 10% Mode, Tighten Standard Limits
4
Conservative SCR, Intermediate
Temperatures
10%
23%
Add 10% Mode
5
Conservative SCR, Best
Temperatures
5%
30%
Add 10% Mode and 5% Mode Cap
Final Report 03.28987
XV
-------
EPA Task Order 68HERC24F0432
Scenario 1 is included for comparisons, but generally indicates the simple measure of re-
weighting the test cycle modes to more accurately reflect ECA operation will achieve only modest
NOx reductions without additional changes. All of the remaining scenarios involve extending the
range of NOx control for the test cycle down to at least the 10% load point, which would likely
require adding a new test cycle mode at 10% (though it may be possible to achieve the same result
using carefully designed mode caps). As noted earlier, existing literature indicates that this is
likely achievable using technologies already deployed on many Tier III OGVs. It should be noted
the scenarios 2 and 5, which extend NOx control down to the 5% load point will likely require
some additional thermal management technology, such as intake air heating or variable valve
actuation, to achieve. Technology scenarios marked as "Max SCR", include reductions in the NOx
standard that would require improving high load SCR NOx conversion efficiency to 88% from the
current level of roughly 80%.
Based on the scenarios given in Table 1, it appears that NOx reductions on the order of
50% may be available using a reasonable combination of technologies, many of which are already
applied on Tier III vessels today. All of these technology scenarios are considered feasible for
application on Category 3 2-stroke engines, though some, as noted, would require the fitting of
additional technologies beyond what are currently seen today on Tier III vessels.
Final Report 03.28987
xvi
-------
EPA Task Order 68HERC24F0432
1.0 INTRODUCTION AND BACKROUND
The Environmental Protection Agency (EPA) is conducting research to assess the potential
for further reductions in NOx emissions from large Category 3 marine diesel engines (> 30 1/cyl
displacement) used on ocean-going vessels (OGVs).
This study examines OGV operation in the North American NOx Emission Control Area
(ECA), designated through amendment to Annex VI of the International Convention for the
Prevention of Pollution from Ships (MARPOL). Emissions from an engine installed on a ship
with a keel laid date beginning January 1, 2016 may not exceed the Annex VI Regulation 13
Tier III NOx limits while that engine is operating in an ECA. The Tier III NOx emission limits
are as follows, where n = rated engine speed (crankshaft revolutions per minute)1:
• 3.4 g/kWh when n is less than 130 rpm;
• 9 • rirQ T> g/kWh when n is 130 or more but less than 2,000 rpm;
• 2.0 g/kWh when n is 2,000 rpm or more.
Marine diesel propulsion engines are certified to these NOx limits using the standard ISO
marine test cycles: E3 test cycle, for a fixed pitch propeller propulsion engine, or a propeller-law
operated non-propulsion engine; or E2 test cycle for a propulsion engine that does not operate with
a fixed pitch propeller. Examples of applications that use E2 include those used in diesel-electric
installations or operated with a controllable-pitch propeller. The E2 test cycle has emission
measurement mode points at a constant speed of 100% and load points at 25%, 50%, 75% and
100%) of the maximum continuous power rating (MCR) specified on the engine nameplate. The
E3 test cycle has emission measurement mode points at the variable speeds and loads of 100% and
100%), 91%) and 75%, 80% and 50%, and 63% and 25% of the MCR. The NOx emissions at these
mode points are weighted using the test cycle modal weighting factors and then summed for
comparison to the emission limit. The test cycles also include a mode cap that limits the modal
emission values to 1.5 times the Annex VI Regulation 13 Tier III NOx standard. The mode cap
was added to the test cycles to ensure that the emission control system stays functional down to at
least the 25% load mode point.
Over the last 10 years, increases in fuel prices have led ship operators to reduce vessel
speed (slow steaming) to reduce ship fuel consumption and operating costs. In addition, some
coastal areas have Vessel Speed Reduction (VSR) zones to reduce air emissions or address other
environmental or safety concerns. For example, a VSR established in 2001 at the ports of Los
Angeles and Long Beach in California to reduce emissions set a speed limit between 10 and 15
knots, depending on ship type, within 20 to 40 nautical miles from shore. Thus, widespread slow
steaming means that ship operation may not be accurately portrayed by the marine diesel engine
propulsion certification test cycles (E2, E3).
1 See MARPOL Annex VI, Regulation 13.5.1.1.
Final Report 03.28987 1
-------
EPA Task Order 68HERC24F0432
A U.S. government paper submitted to a recent meeting of the IMO Sub-Committee on
Pollution Prevention and Response (PPR) indicated that due to these two conditions, operating at
slow speed and lack of Tier III control below 25% engine load, the MARPOL Annex VI ECA
NOx requirements are not yielding the expected emission reductions and related air quality
benefits in the North American ECA.2
As part of an ongoing effort to assess the Annex VI NOx ECA program and how it could
be improved, this study investigated the following Tier Ill-related issues. First, whether the
efficiency of SCR emission control technology on Category 3 vessels can be improved, both
underway and while operating at low load, based on the current engine certification requirements.
Second, an evaluation of ship operating profiles in the North American ECA irrespective of engine
category to examine how much time is spent at low load operation and, using that information,
how the certification test cycles could be revised to reflect actual operating profiles. Third, if
exhaust gas temperature can be maintained, whether it is possible to extend SCR operation to 10%
load or lower.
2 See PPR 11/INF.4, Assessment of the impacts of the MARPOL Annex VI emission control
regulations in the United States portion of the North American Emission Control Area
2
Final Report 03.28987
-------
EPA Task Order 68HERC24F0432
2.0 ASSESSMENT OF OCEAN-GOING VESSEL LOW-LOAD OPERATION
The purpose of this task was to objectively quantify the duty cycle of category 3, ocean
going vessels operating within the North American ECA. The general approach was to use
publicly and commercially available data sources to identify when ships were operating within the
ECA and estimate the load at which they were operating. A very high-level diagram of this process
is shown in Figure 3.
AIS Data
Operational Data (Lat, Lon, Speed,
IMO Number)
Clarkson's Ship Registry
Ship specifications
(displacement, cylinders,
stroke type, max speed...)
Calculations
Load estimated based
on current speed &
ship specs
Database
House data, join data,
and enable calculations
Visualizations
Sum operating load
points by region to
create histograms
Spatial Join
Identify points within
each region
FIGURE 3. HIGH LEVEL DIAGRAM OF SHIP DATA ANALYSIS PROCESS
The estimated duty cycle for Tier III, Category 3 ships operating within the North
American ECA is shown in Figure 4. It clearly shows that the actual duty cycle is very different
from the duty cycle of the certification test cycle. Significant work was required to generate this
ECA-based duty cycle. An equivalent amount of effort was made to ensure that this derived duty
cycle is representative of ships operating throughout the entire ECA and to verify that the data
used to create it were valid. This section describes the process used to validate and create these
histograms.
25
20
CD
O)
cr,
c
0
o
a>
Q_
15
10
100
Load Factor Distribution
- Cumulative Load Factor
Engine Test Cycle
80
60
40
20
o
o
03
LL
"O
(C
o
J3
3
E
ZJ
O
Load Factor Categories
FIGURE 4. ESTIMATED DUTY CYCLE OF TIER III CATEGORY 3 SHIPS OPERATING
WITHIN THE NORTH AMERICAN AND CARIBBEAN SEA ECAS (N=255)
Final Report 03.28987
3
-------
EPA Task Order 68HERC24F0432
2.1 Datasets
Four datasets were used to perform this analysis.
• Ship operation data collected from the Automatic Identification System (AIS).
• Ship specifications from Clarksons World Fleet Register, augmented by information
from Sea-web™
• ECA boundaries
• C3RIA regions
The first dataset was 2022 AIS data that was acquired by EPA from the United States Coast
Guard (USCG) through a historical AIS data request3. This dataset included ship operational data
including latitude, longitude ship speed, ship dimensions, and ship's assigned International Marine
Organization (IMO) number. Some of the ship dimensions in this dataset were not used because
their accuracy was unknown (United States Coast Guard, 2025). AIS data can be publicly
downloaded from the Marine Cadastre website (NOAA Office For Coastal Management, 2025),
however the dataset used for this study also included satellite data that were provided by the
USCG. This dataset had already been cleaned by the EPA prior to its receipt by SwRI.
The second dataset was from the Clarksons World Fleet Register. This dataset of ship
specifications included items like IMO number, maximum operating speed, service speed, engine
Tier, engine manufacturer, engine displacement, number of cylinders and ship type (Clarksons,
2024). These data were matched with the AIS data based on IMO number to tie a ship's speed
and location to a set of specifications. The combined data were then used to estimate the ship
engine load.
The third dataset included boundary definitions for the North American ECA and the
United States Caribbean Sea ECA, Figure 5. These boundaries are defined by MARPOL Annex
VI (International Maritime Orginization, 2024). Ship operation within these regions requires the
use of lower sulfur ECA compliant fuels, and for ships whose keel was laid beginning 1/1/16,
Tier III engines must be operated with NOx emission control devices engaged. The ECA boundary
definition files for this dataset were supplied by the EPA.
3 https://www.navcen.uscg.gov/contact/ais-historical-request
4
Final Report 03.28987
-------
EPA Task Order 68HERC24F0432
Hcmdw
Manir'albo
FIGURE 5. MAP OF NORTH AMERICAN & CARIBBEAN SEA EMISSION
CONTROL AREAS AS RECEIVED FROM EPA
A fourth set of boundaries was also used for this study to define several regions. This
dataset was supplied by the EPA and was known as the C3RIA dataset. This dataset was used to
define regions within the North American ECA, as shown in Figure 6, to ensure there were no
major differences in ship operation between different regions within the North American ECA.
This set of boundaries did not include areas for the Caribbean Sea ECA or the Canadian portion
of the North American ECA. Some work was performed to address these gaps that is discussed
later in the report.
Final Report 03.28987
5
-------
EPA Task Order 68HERC24F0432
United Srai
HI
FIGURE 6. MAP OF THE BOUNDARIES DEFINED IN THE C3RIA DATASET AS
SUPPLIED BY EPA
2.2 Tools Used for Data Analysis
The analysis required for this project required several different capabilities. The first
requirement was the ability to interact with a very large dataset that exceeded physical random
access memory (RAM). The AIS data included almost 1.5 billion records. The second
requirement was the ability to join the Clarksons data and the AIS data based on the IMO number
field. These first two requirements drove the need for a combined database. The final major
requirement was to be able to understand where each datapoint was geographically located. This
requirement drove the need for several geospatial tools.
The use of open-source tools was a deliberate decision for this project. The use of these
tools will enable the analysis to be reproduced more easily than if proprietary tools had been used.
The major tools used for this project are shown in Figure 7 and include Python, QGIS, Postgresql,
PostGIS, and Geopandas.
Final Report 03.28987
6
-------
EPA Task Order 68HERC24F0432
PostgreSQL PostGIS
python Gf +
Open
|:J> GeoPandas I ^/Street Inn PlOtly
Map
FIGURE 7. OPEN-SOURCE TOOLS USED FOR ANALYSIS & VISUALIZATION
(GEOPANDAS DEVELOPERS, 2024) (OPENSTREETMAP FOUNDATION, 2024)
(PLOTLY, 2024) (POSTGRESQL GLOBAL DEVELOPMENT, 2024) (PYTHON
SOFTWARE FOUNDATION, 2025) (QGIS DEVELOPMENT TEAM, 2024)
Visualization was another major aspect of work. There were three open-source projects
that were leveraged for this project. The first was the OpenStreetMap project. Their maps were
used as the basemap for all maps in this report. The second project was the Plotly plotting library.
The final project was QGIS. This software was used for creating most of the maps in this report.
TABLE 2. TABLE OF SOFTWARE VERSIONS USED FOR ANALYSIS AND
VISUALIZATION
Software Version
Python 3.12
PostgreSQL 16.6
PostGIS 3.4.4
QGIS 3.34
GeoPandas 0.14.2
Plotly 5.24.1
2.3 Region Definition
Several regions were defined for this study to explore whether actual ship duty cycles differ
between regions. It was expected that the ship type composition would differ between regions due
to the differences in industries served by various ports, but it was unknown if the duty cycle was
different.
Final Report 03.28987
7
-------
EPA Task Order 68HERC24F0432
The regions were defined using the North American and U.S. Caribbean Sea ECA
boundaries along with the C3RIA boundaries that were supplied by the EPA. The resulting
boundaries that were derived are shown in Figure 8.
FAIask"a1East«
Canadian Pacific
-North' Pacific
SoutniPacific
Hawaii ,
^uertonRicb@<_Virgin Islands
FIGURE 8. BOUNDARIES USED FOR COMPARING DUTY CYCLES BETWEEN
ECA REGIONS
The regions in Figure 8 were derived by first trimming the C3RIA regions by the North
American ECA. This operation is shown in the lower right comer of Figure 9. Most regions from
the C3RIA dataset were not greatly impacted. However, the Alaska West (AW) region was mostly
eliminated. The decision was made to remove that region entirely because the remaining portions
were mostly over land. The Hawaii region was also reduced in size relative to the C3RIA
definition because the ECA around Hawaii was much smaller. This can be seen by comparing the
right two panes in Figure 9.
Final Report 03.28987
8
-------
EPA Task Order 68HERC24F0432
North American ECA
*
Clipped C3RIA
Regions Subtracted From ECA
AW
AE
< NP
SP
I
C3RIA Regions
GE
NA
GC
' C3RIA Regions
Clipped to ECA
FIGURE 9. FIGURE SHOWING INPUTS AND INTERMEDIATE STEPS TO DERIVE
REGIONS WITH ARROWS INDICATING DATA USED FOR EACH STEP
A few other regions were added including the Canadian Pacific, Canadian Atlantic, and
Puerto Rico & Virgin Islands regions . These regions were created to ensure the entire ECA was
represented by individual regions. The goal of the study was to characterize ship duty cycle within
the complete ECA. These extra regions were derived by subtracting the clipped C3RIA regions
from the ECA. The areas calculated from this operation are shown in the bottom left corner of
Figure 9. The Canadian regions were further trimmed by hand to reflect where ship traffic would
be expected. The Puerto Rico & Virgin Islands region was simply the U.S. Caribbean Sea ECA.
There was no reference to the U.S. Caribbean Sea ECA in the C3RIA dataset.
2.4 Description of the Ship Included in the Study
The Clarksons World Fleet Register was used as the source of the ship specifications for
the 12,303 Category 3 ships that operated in the ECA in 2022 (see Section XX below for how
those ships were identified using AIS data). These specifications were used to estimate engine
load and identify ship categories, for the purpose of the duty cycle analysis. The parameters that
were used from the Clarksons data were IMO number, engine tier level, cylinder bore, cylinder
stroke, maximum speed, and service speed. Each of these is described below.
The breakdown of ships by Annex VI Regulation 13 NOx emissions Tier is shown in
Figure 10. Tier III ships were the smaller population within this group. These ships were of the
most interest for this study because they would be equipped with engines having Tier III emission
control systems. It was speculated that inclusion of aftertreatment systems could cause different
operational characteristics from the other emission tier levels. The Tier I and Tier II ships were
Final Report 03.28987
9
-------
EPA Task Order 68HERC24F0432
separately analyzed to verify that the duty cycles for Tier III ships were similar to Tier I and
Tier II ships.
8000
7000
6000
c
O 0000
o
$ 4000
0)
¦>
3000
2000
1000
0
Main Engine Tier Level
FIGURE 10. HISTOGRAM OF IMO NOx TIER FROM CLARKSONS WORLD FLEET
REGISTER FOR SHIPS INCLUDED IN THE STUDY (N=12,303)
Figure 11 shows the breakdown of ship types for the 12,303 ships studied, obtained from
the Clarksons database. There were 48 different types of ships identified in the dataset, but 11 ship
types made up the bulk of the fleet with over 100 ships per classification. These ship types are
denoted by yellow highlighting in Figure 11. It is important to note the log scale on the y-axis in
Figure 11.
Final Report 03.28987
10
-------
EPA Task Order 68HERC24F0432
FIGURE 11. HISTOGRAM OF CATEGORY 3 SHIPTYPES FROM CLARKSONS
WORLD FLEET REGISTER (N=l 2,303)
A histogram of main propulsion engine types for the 12,303 ships studied is shown in
Figure 12. Two-stroke diesel engines were the most common propulsion engines for Category
3 ships included in the analysis. There were a few diesel-electric ships and a single hybrid ship.
These were used in cruise ships and were not considered for this study because load estimation
would be very difficult.
7000
6000
^ 5000
c
O 4000
-------
EPA Task Order 68HERC24F0432
A histogram of the bore diameter of the 12,303 main propulsion engines included in the
analysis is shown in Figure 13. Most of these engines have bore diameters of 500 mm or 600 mm.
The majority of engines have bore diameters that align with 100 mm increments (i.e. exactly 500,
600, 700, 800, and 900 mm).
4000
3500
3000
§ 2500
o
o
® 2000
1500
1000
500
o"-
I J.
400
500
600 TOO
Bore Diameter [mm]
I
800
I. Jl
900
FIGURE 13. HISTOGRAM OF BORE DIAMETER FOR MAIN PROPULSION
ENGINES INCLUDED IN THE STUDY FROM CLARKSONS WORLD FLEET
REGISTER (N=l 2,303)
A histogram of main propulsion engine stroke for these same engines is shown in Figure
14. Most of these engines have a stroke length between 2,000 mm and 2,500 mm.
Final Report 03.28987
12
-------
EPA Task Order 68HERC24F0432
2500
2000
c
8 1500
CD
1
<13
> 1000
500
1000
1500
J..
2000
Stroke [mm]
2500
3000
3500
FIGURE 14. HISTOGRAM OF STROKE FOR MAIN PROPULSION ENGINES
INCLUDED IN THE STUDY FROM CLARKSONS WORLD FLEET REGISTER
(N=12,303)
This combination of bore and stroke results in the histogram of cylinder displacement
shown in Figure 15. There are two peaks of cylinder displacement for the 12,303 main propulsion
engines included in the data set, centered on 475 liter/cylinder and 675 liter/cylinder.
2500
2000
3 1500
o
m
if)
© 1000
500
-Ll.-L
500
1000 1500
Cylinder Displacement [liters/cyl]
2000
2500
FIGURE 15. HISTOGRAM OF INDIVIDUAL CYLINDER DISPLACEMENT FOR
MAIN PROPULSION ENGINES INCLUDED IN THE STUDY FROM CLARKSONS
WORLD FLEET REGISTER (N=l 2,303)
Final Report 03.28987
13
-------
EPA Task Order 68HERC24F0432
A histogram of cylinder count per engine for the engines included in the study is shown in
Figure 16. The most common cylinder count is 6 cylinders. The highest and lowest cylinder
counts were 16 and 4 cylinders, but these were very uncommon.
8000
7000
6000
c 5000
D
o
o
- 4000
0
a)
U)
:> 3000
2000
1000
4 6 8 10 12 14 16
Cylinder Count [-]
FIGURE 16. HISTOGRAM OF ENGINE CYLINDER COUNT FOR MAIN
PROPULSION ENGINES INCLUDED IN THE STUDY FROM CLARKSONS WORLD
FLEET REGISTER (N=l2,303)
The resulting engine displacements for the engines included in the study are shown in
Figure 17. The histogram shows peaks at 2,000 liters, 4,000 liters, 6,000, and 7,500 liters. The
vast majority of the engines in this population are 8,000 liters or below.
Final Report 03.28987
14
-------
EPA Task Order 68HERC24F0432
2500
2000
c
zs
o
1500
O
o>
U3
U)
0)
1000
>
500
«- 2000-2490 liters
«- 4000-4490 liters
«- 6000-6490 liters
«- 7500-7990 liters
5k
10k 15k
Displacement [liters]
20k
25k
FIGURE 17. HISTOGRAM OF ENGINE DISPLACEMENT FOR MAIN PROPULSION
ENGINES INCLUDED IN TIIE STUDY FROM CLARKSONS WORLD FLEET
REGISTER (N=12,303)
A histogram of main engine power for the 12,303 ships included in the study is shown in
Figure 18. There is apeak centered at 13,500 hp, and most ofthe main engines are rated for 25,000
hp or less.
1200
1000
S 800
o
600
400
200
20k
40k
60k
80k
I -
100k
120k
140k
Power [hp]
FIGURE 18. HISTOGRAM OF MAIN ENGINE POWER FOR MAIN PROPULSION
ENGINES INCLUDED IN THE STUDY FROM CLARKSONS WORLD FLEET
REGISTER (N=l 2,303)
Final Report 03.28987
15
-------
EPA Task Order 68HERC24F0432
A histogram of maximum ship speed for 630 of the ships included in the study is shown in
Figure 19. It is important to note that the vessel counts in this chart are much lower than prior
histograms because only 630 ships in the Clarksons database had data for maximum ship speed.
Most of these ships were rated for a maximum speed between 14 and 17 knots. There was also
another grouping between 20 and 22 knots. The maximum speed was 30 knots. There were four
ships with a maximum speed rating about 29 knots. Three of these were fully cellular container
ships with keel laid dates in 2007 and 2008. The fourth was a cruise ship.
50
40
Maximum Speed [knots]
FIGURE 19. HISTOGRAM OF MAXIMUM SHIP SPEED FROM CLARKSONS
WORLD FLEET REGISTER (N=630)
A histogram of service speeds for 8,502 ships included in the study is shown in Figure 20.
Most ships had a service speed between 13 knots and 16 knots. However, there were smaller peaks
at 20, 22, and 24 knots. Again, the vessel counts in this chart are less than the number of ship
included in the study because only 8,502 ships had a specified service speed specified in Clarksons
database. Several of these ships were also in the group with a specified maximum speed (N=630).
Final Report 03.28987
16
-------
EPA Task Order 68HERC24F0432
1600
1400
1200
% 1000
O
O
"55 800
to
to
> 600
400
200
FIGURE 20. HISTOGRAM OF SHIP SERVICE SPEED FROM CLARKSONS WORLD
FLEET REGISTER (N=8,502).
The composition of the Clarksons World Fleet Register data consists of ships powered with
two-stroke, 6-cylinder engines. The most common ship types that these engine powered were bulk
carriers, fully cellular container ships, tankers, and chemical & oil carriers. It is important to note
that not all fields were populated within this dataset which will impact the data analysis in later
sections.
2.5 AIS Data
The 2022 AIS dataset used for this study consisted of about 1.5 billion ship position records
(data points), both inside and outside the ECA. Because of the large size of this dataset, much of
the characterization and validation of this dataset was done programmatically.
As received from the USCG, the 2022 AIS dataset was aggregated into five-minute
intervals and was split into separate files for data collected by terrestrial AIS (TAIS) and satellite
AIS (SAIS) receivers. Prior to transmitting the AIS data to SwRI, EPA performed the following
cleaning steps. First the SAIS and TAIS files were combined into a single dataset. Entries with
duplicate Maritime Mobile Service Identities (MMSI), IMO, and PERIOD (timestamp) fields were
removed. In cases where there were duplicate records with both SAIS and TAIS records, the TAIS
record was retained and the SIAS record was removed. Next EPA filtered the dataset and retained
entries where the MMSI indicated that the transmitter was one of the following: a ship, an auxiliary
craft associated with a parent ship, or a group of ships. Finally, it should be noted that a database
malfunction at the USCG resulted in sparse data retrieval for the months of May and June 2022.
JL
15 20 25 30
Service Speed [knots]
Final Report 03.28987
17
-------
EPA Task Order 68HERC24F0432
This 2022 AIS dataset included fields for latitude, longitude, GPS reported speed, and IMO
number. There were several other fields included in this dataset, but the four listed fields were the
only ones used for this study. The steps taken to characterize and filter the dataset for the purposes
of this study are described below.
The first major step in filtering the dataset was to remove all AIS data points that did not
have an IMO number. These points were removed because they could not be associated with a set
of ship specifications from the Clarksons dataset. While an attempt was made to remove points
with invalid IMO numbers based on their checksums4, this was not completed because several of
the invalid IMO numbers matched ships in the Clarksons dataset.
The next step was to create a heatmap of the AIS data from all of the 23,245 ships with an
IMO number, regardless of engine Tier or vessel class. Two heatmaps are shown in Figure 21, the
top heatmap is all AIS data from ships with an IMO number, and the bottom heatmap is AIS data
only from 983 Tier III ships as identified in the previous section. These two heatmaps were created
with a 0.1° hexagonal grid. The position data points within each hexagon were counted in order
to create these maps. It should be noted that the color scale is not constant between these two
maps because the total point count is very different. However, roughly similar traffic patterns can
be observed between the two maps suggesting that Tier II and Tier III operate in similar ways.
4 A checksum is a mathematical equation that can be applied to the IMO number to verify that
the contents of the received IMO number are valid.
18
Final Report 03.28987
-------
EPA Task Order 68HERC24F0432
A1S Data Samples With IMO Number
1
r
"X
%
a*.
s
AIS Data Samples for Tier III Ships
FIGURE 21. HEAT 1VIAPS BASED ON AIS DATA WITH IMO NUMBER COMPARING
TRAFFIC PATTERNS FOR ALL SHIPS WIYH AIS DATA VS TIER III SHIPS
There are some subtle differences that can be seen between the two heat maps that may
reflect differences in the Tier III fleet composition, which includes only those with Category 3
propulsion engines. One of these differences can be seen between Newfoundland and Greenland
in Figure 21: the heat map for the complete dataset shows lots of small tear drop shapes in that
area while the Tier III map does not contain these shapes. These features were not present in the
Tier III map because they were generated by vessels that were not Category 3, Tier III vessels.
Final Report 03.28987
19
-------
EPA Task Order 68HERC24F0432
The next step was to consider the context for the part of the dataset that had an IMO number
and were able to be associated with a ship in the Clarksons dataset, regardless of emissions Tier.
The first step was to identify how many of these ship position records fell within the North
American EC A. It was found that about 2/3 or 110 million of these data points were contained in
the ECA. The remaining third fell outside the ECA. The preponderance of points inside the ECA
may occur because ships at anchor or in port may continue broadcasting their position even when
they are not moving or are drifting. The points at anchor or in port were included for the location
data point counts but omitted for estimation of load. These counts are depicted by engine tier in
Figure 22.
100M
BOM
¦4—'
c
O 60M
O
c
o
^ 40M
20M
0
Location
FIGURE 22. BAR CHART OF SHIP LOCATION COUNTS INSIDE AND OUTSIDE OF
ECA FOR AIS DATA WITH AN IMO NUMBER JOINED TO THE CLARKSONS
DATASET (N= 167,512,041)
The next division that was investigated for the AIS data was region. This was done to
understand which regions have the most ship traffic. Understanding this distribution helps to
identify any biases that may exist within the dataset caused by regional differences. Point counts
by region inside the EC As are shown in Figure 23. It was found that the Gulf Coast region
dominated the point counts by region. Alaska East had very few points.
Final Report 03.28987
20
-------
EPA Task Order 68HERC24F0432
40M -Tier I
35M * Tier 11
¦ Tier III
^ 30 M
O 25 M
c 20 M
15M ¦
xf cx cx0 Gx* Gx V "X *X *X *X V*
% ^ x ^ X^X ^
'Xa
N
Region
FIGURE 23. BAR CHART OF POINT COUNTS BY REGION FOR AIS DATA WITH
AN IMO NUMBER JOINED TO THE CLARKSONS DATASET (N=103,552,955)
Another way to look at the data is by IMO number. There were 10,745 unique IMO
numbers for ships that entered the ECA that could be linked to a ship in the Clarksons dataset. The
data presented in Figure 24 shows the number of ship location counts for unique IMO numbers
found in each region within the dataset. It is important to note that the sum of these bars (24,161)
is much larger than the number of ships that were ever present in the EC As (10,745) because many
ships visited several regions. Over half of the ships in the dataset were observed in the Gulf Coast
region. Understanding these counts is important when looking at regional differences. Results for
regions with a small number of unique IMO numbers may be skewed by a small number of ships
that may have atypical operational characteristics relative to the rest of the fleet.
Final Report 03.28987
21
-------
EPA Task Order 68HERC24F0432
6000
¦ Tier I
¦ Tier II
¦ Tier III
Region
FIGURE 24. BAR CHART OF UNIQUE VESSEL COUNT BY REGION FOR AIS DATA
BASED ON IMO NUMBER JOINED TO THE CLARKSONS DATASET
2.6 Duty Cycle Estimate & Results
The next step in the analysis was to estimate the ship propulsion engine load from the AIS
data and summarize it. This involved using ship design speed and current speed for each point.
This data was then grouped by region, ship type, and engine tier level to create a fleet level duty
cycle.
2.6.1 Load Estimation
The load (L) for each ship propulsion engine as a fraction of the maximum continuous
power rating was estimated using the propeller law shown below. This method was chosen
because it only requires the use of current vessel speed (V) and its maximum speed (P max). A value
of 1.0 was chosen for M, the decimal fraction of the maximum continuous engine power rating at
maximum vessel speed, in the equation below to reflect that 100% of MCR corresponds with
maximum speed in normal sea conditions (MAN Energy Solutions, 2023). This approach is
reasonable because...
It was mentioned in the description of the Clarksons dataset above that only a fraction of
ships had maximum or service speed identified. This omission required a special process to
produce the best estimate of engine load possible. If maximum speed was available, this value
was used for the estimation of load with the propeller law. If only the service speed was available,
it was divided by 0.94 to estimate the maximum speed for use with the propeller law. If no speed
was specified, the load was set to a null value so that it would not influence the duty cycle
Final Report 03.28987
22
-------
EPA Task Order 68HERC24F0432
development. This process is illustrated in Figure 25. This approach was only applied to points
with a speed greater than 2 knots. It was assumed that ships traveling slower than 2 knots were
anchored, at berth, or not operating under their own power.
FIGURE 25. FLOW CHART DESCRIBING THE PROCESS USED TO CALCULATE
PERCENT LOAD BASED UPON THE AVAILABILITY OF MAXIMUM SPEED OR
SERVICE SPEED IN CLARKSONS SHIP SPECIFICATIONS
The omission of maximum and service speed led to the omission of a significant amount
of data. The majority of Tier I ships had valid speed data. However, this filtering rendered 66%
of the Tier III ship data and about 34% of Tier II data unusable. An attempt was made to predict
these speeds for each ship, but we were unable to develop a method to confidently verify the ship
design speed predictions. A Venn diagram and table outlining how many ships had maximum or
service speeds specified are shown in Figure 26.
TotalX
Tier 1
Tier II
Tier III
Total
2,894
8,432
977
Valid Max Speed
164
432
34
Valid Service Speed
2,726
5,453
323
Valid Service or Max
Speed
2,733 (94.4%)
5,548 (65.8%)
329 (33.7%)
No Valid Speed
266 (9.2%)
3401 (40.3%)
653 (66.8%)
FIGURE 26. VENN DIAGRAM AND TABLE OF SHIPS WITH MAXIMUM SPEED,
SERVICE SPEED, BOTH SPEEDS, OR NO SPEED SPECIFIED
Final Report 03.28987
23
-------
EPA Task Order 68HERC24F0432
The final validity check on the data for each ship was the engine load. Many ships had
predicted maximum engine loads in excess of 125%. It was assumed that ships that exhibited
loads in excess of 125% anywhere in the dataset had an incorrect specification or estimate of
maximum vessel speed. These ships were omitted to ensure the data used for duty cycle estimation
were as reliable as possible. It would take a speed significantly greater than the rated speed to
cause the estimated load to be in excess of 125%.
Combining all the filters for identifying the points of interest results in a very small
percentage of the total available points being used: 1,960 Tier I ships, 4,040 Tier II ships, and 255
Tier III ships. This filtering process is illustrated in Figure 27. This chart visually describes how
the data points used for this study were defined from the dataset as received. The largest reduction
in useable points was due to the lack of an IMC) number in the AIS dataset. This eliminated about
two-thirds of the data. The next largest filter was the lack of an I MO number match with the
Clarksons data. The lack of a match in most cases was due to Clarksons not having the information
for a specific ship. This could be because data for non-category 3 vessels was not obtained from
Clarksons. Simil arly, the effect of filtering on the reduction in ship counts based on unique IMO
numbers is shown in Figure 28. Taken together, these two figures show that load histograms
generated by this study for Category 3 Tier III vessels in the ECA were based on 2,310,734 AIS
speed data points from 255 ships.
995,551,339
AIS Points w/o
IMO Number
495,402,646
AIS Points w/
IMO Number
327,890,60
[Points With No IMO Number
Match to Clarksons
63,959,086
Points Outside ECA
29,087,031
c Points With No Speed Spec.
19,022,292
0
llOL
Ships With Excess Load
17,763,395_ Tier I
35,369,503a Tier II
2,310,734 Tier III
FIGURE 27. SAN KEY DIAGRAM OF AIS POINT FILTERING TO IDENTIFY
USEFUL POINTS FOR DUTY CYCLE DEVELOPMENT
Final Report 03.28987
24
-------
EPA Task Order 68HERC24F0432
I
10,942
IMO Numbersw/o
Clarksons Match
12,303
IMO Numbers w/
Clarksons Match
1,558
IMO Numbers That Never Entered ECA
3,321
10,745
I
IMO Numbers With No Speed Spec.
1,169
IMO Numbers With Excess Load
7,424
1.960 Tier I
4,040
IMO Numbers IMO Numbers
That Entered ECA with Speed Spec
255
Tier II
Tier III
FIGURE 28. SANKEY DIAGRAM DEPICTING IMPACT OF FILTERING ON IMO
NUMBERS USED FOR DUTY CYCLE DEVELOPMENT
For the purpose of this study, the ship types included in Clarsksons were combined into a
smaller number of categories as shown in Table 3. There were many different types of ships used
to classify the vessels in the Clarksons data. The data presented in Table 3 and Table 4 shows the
mapping used for the study. The counts presented in Table 3 are the counts for each classification
in the Clarksons data. Table 4 presents the counts for the ships used to develop the duty cycles in
the next subsection. The study counts are lower for two reasons. The first is that some of these
ships never entered the ECA. The second is that some of these ships had no speed specification.
Comparison of these two ship counts is important to ensure that one type of ship is not being
eliminated from the dataset and introducing a bias.
Final Report 03.28987
25
-------
EPA Task Order 68HERC24F0432
TABLE 3. TABLE DESCRIBING THE MAPPING OF SHIP CLASSIFICATIONS IN
CLARKSONS TO SHIP GROUPING USED FOR STUDY AND THEIR
DISTRIBUTIONS WITHIN THE CLARKSONS DATA
Study Classification
Clarksons Classification
Tier III
Tier II
Tier I
Tankers
Crude Tankers, Product Tankers,
Chemical Tankers, Spec. Tankers
167
983
785
Gas Carriers
LPG, LNG
85
195
72
Bulk Carriers
Bulkers
19
2,319
681
General Containers
General Cargo
0
67
7
Containers
Containerships
28
753
590
Other
All other ship types
8
532
163
Total
307
4,849
2,298
TABLE 4. TABLE DESCRIBING THE MAPPING OF SHIP CLASSIFICATIONS IN
CLARKSONS TO SHIP GROUPING USED FOR STUDY AND THEIR
DISTRIBUTIONS WITHIN THE SHIP POPULATION USED FOR DUTY CYCLE
DEVELOPMENT
Study Classification
Clarksons Classification
Tier III
Tier II
Tier I
Tankers
Crude Tankers, Product Tankers,
Chemical Tankers, Spec. Tankers
111
692
555
Gas Carriers
LPG, LNG
50
107
44
Bulk Carriers
Bulkers
14
1,763
465
General Containers
General Cargo
0
15
7
Containers
Containerships
27
663
526
Other
All other ship types
5
379
135
Total
207
3,619
1,732
The final aspect of the data is how much time each ship spent within the ECA during the
study period. The calculated time is based upon the number of points that exist within the ECA
boundary for each IMO number. The total time per ship can be for one or several trips in 2022,
depending on how many times the ship entered the ECA. The histogram presented in Figure 29
shows the distribution of how much time ships spent inside the ECA. It should be noted that there
were 451 ships that entered the ECA but were within the ECA for less than an hour. This may
also be the case with ships that operated in the ECA between 1 and 10 hours, or they may have
been in innocent passage (no stop at a U.S. port). These ships are believed to have erroneously
Final Report 03.28987
26
-------
EPA Task Order 68HERC24F0432
entered the EC A and probably made no operational changes while operating within the ECA (i.e.,
they probably operated in cruise mode for the brief amount of time they transgressed the
boundary). While they were not omitted from the duty cycle development, their inclusion should
not have an impact on the duty cycle results because they represent less than 0.01% of the
4.6 million hours of data.
600
500
C
O 400
O
"0)
% 300
:>
200
100
o
FIGURE 29. HISTOGRAM OF TIME SPENT IN ECA IN HOURS (N=7,357)
2.6.2 Estimated Duty Cycles
The duty cycle for ships at each emission tier level were developed by creating engine load
histograms from the dataset generated using the process outlined in the previous section. The
Sankey Diagrams in Figure 27 and Figure 28 depict how the dataset of interest was identified. The
equation in Section 2.6.1 explains how the engine load was estimated.
These histograms were created at the fleet level within the North American ECA for each
emission tier level. They were divided by tier to ensure that differences in engine technology did
not cause differences in operational characteristics. The duty cycle for the Tier III ships is
presented in Figure 30. This figure is identical to Figure 4. It is shown here for ease of comparison.
The results for Tier I and Tier II ships are presented in Figure 32 and Figure 31 respectively.
Total Time In ECA [hours]
Final Report 03.28987
27
-------
EPA Task Order 68HERC24F0432
FIGURE 30. ESTIMATED DUTY CYCLE OF TIER III CATEGORY 3 SHIPS
OPERATING WITHIN THE NORTH AMERICAN ECA (N=255)
jMtIl!lCIU,0{;iOUlOUl0y,OU,OU,OU'OUlOC,^^MW
^'o^rorowwiiiuiuicnros^cocotbi)^
-------
EPA Task Order 68HERC24F0432
¦ Load Factor Distribution
Cumulative Load Factor
Engine Test Cycle
l1»!',fn^o^,o<-r,o^r,ocjiocnocnoc/ioaioaioo-^-Aroro
Load Factor Categories
FIGURE 32. ESTIMATED DUTY CYCLE FOR TIER I CATEGORY 3 SHIPS
IDENTIFIED IN NORTH AMERICAN ECA (N=l,960)
The estimated duty cycles for all three emission tiers appear to be very similar. Figure 33
presents the duty cycles for each tier relative to the certification test cycle. They estimated duty
cycles all have a much lower load than the certification test cycle. The Tier III duty cycle is the
most heavily loaded of the three duty cycles, but it is still dramatically different from the
certification test cycle.
FIGURE 33. LINE CHART COMPARING CUMULATIVE DISTRIBUTION OF DUTY
CYCLE FOR TIER I, TIER II, TIER III, AND CERTIFICATION TEST CYCLE
Final Report 03.28987
29
-------
EPA Task Order 68HERC24F0432
The duty cycles for each emission tier were broken out by major ship classification to
ensure that the aggregate duty cycle is representative of the five major ship types shown in Table
4 or the legend for each duty cycle plot. The estimated duty cycles for each of these ship types are
shown in Figure 34 to Figure 36. These duty cycles are very similar. The largest deviation is that
the container ships have a duty cycle that is consistently lighter than the other ship types. The
Tier III chart in Figure 34 does not have an entry for general container ships because there were
none.
25
20
15
0)
£ 10
$
Q_
—* ro ro oj co
O (Jl O CJl O Oi
m ro co co .k
cn o cn o cn o
4*. cn cn o>
cn o cn o cn
Ol Ui CI) O) s
O cn o cn o
m co co © -i -»
Ul O Oi O Ol o o
N CD CO (D CO -»
an o cr» o o> o
o cn p cn o
cn o cn
100
80
60
40
20
TJ
cc
o
0
>
il
3
a
¦ Tankers - Load Factor
-•-Tankers - Cumulative Load Factor
* Gas Carriers - Load Factor
— Gas Carriers - Cumulative Load Factor
• Bulk Carriers - Load Factor
— Bulk Carriers - Cumulative Load Factor
Containers - Load Factor
— Containers - Cumulative Load Factor
Engine Test Cycle
Load Factor Categories
FIGURE 34. ESTIMATED DUTY CYCLE BY SHIP TYPE FOR TIER III SHIPS
IDENTIFIED WITHIN THE NORTH AMERICAN ECA (N=255)
25
20
15
a>
Q_
10
jrouij_i(oroww
rocn-^9V19V19 ^
. o —^ ro ro co co
o cn o cn
cn cn
.. ^|iU1U10)0>JS0005(D(D—k-i—
cnoyiocnocnooiooiocnoo-f-4
J^i^cn6icncn^j-^joo6ooDCD-^(?V1.V1
ocnocnooiocnocnocno-^.-*-*-*
U1 o m
Load Factor Categories
100
o
80 o
Cl
60
40
20
o
o
ro
LL
TD
CD
O
_l
CD
>
Jo
E
13
o
¦ Tankers - Load Factor
—Tankers - Cumulative Load Factor
¦ Gas earners - Load Factor
— Gas Carriers - Cumulative Load Factor
• Bulk Carriers - Load Factor
— Bulk Carriers - Cumulative Load Factor
Containers - Load Factor
— Containers - Cumulative Load Factor
¦ General Containers - Load Factor
— General Containers - Cumulative Load Factor
Engine Test Cycle
FIGURE 35. ESTIMATED DUTY CYCLE BY SHIP TYPE FOR TIER II SHIPS
IDENTIFIED WITHIN THE NORTH AMERICAN ECA (.N=4.040)
Final Report 03.28987
30
-------
EPA Task Order 68HERC24F0432
ro cn
' o -k w ro w oj cn on O) O)
tn *-n cnocnocnocnocnoai
o cn
Load Factor Categories
100
¦ Tankers - Load Factor
— Tankers - Cumulative Load Factor
^ ¦ Gas Carriers - Load Faetor
80 Gas Carriers - Cumulative Load Factor
¦ Bulk Carriers - Load Factor
— Sulk Carriers - Cumulative Load Factor
O
CU Containers - Load Factor
Containers - Cumulative Load Factor
"U
TO ¦ General Containers - Load Factor
60 LL
u
TO
^ — General Containers - Cumulative Load Factor
oj Engine Test Cycle
'•M
J3
z$
E
20 3
FIGURE 36. ESTIMATED DUTY CYCLE BY SHIP TYPE FOR TIER I SHIPS
IDENTIFIED WITHIN THE NORTH AMERICAN EC A (N=l,960)
The duty cycle was also investigated from a regional perspective. These duty cycles are
presented in Figure 37 to Figure 39. In general, these duty cycles do not deviate from the aggregate
duty cycles, but there are a few exceptions that are discussed in the following paragraphs.
FIGURE 37. ESTIMATED DUTY CYCLE BY REGION FOR ALL TIER III SHIPS
IDENTIFIED WITHIN THE NORTH AMERICAN ECA (N=255)
The Tier III duty cycles show the largest deviations between regions. This may be due to
the small population size. A small population size can allow a single ship to greatly skew the
average if it is being used in an atypical manner. The Hawaii and Alaska East regions had the
lowest number of samples. These counts can be found in Figure 23. The Tier III duty cycle was
impacted most because it had the least amount points. The Tier II and Tier I duty cycles did not
appear to be impact because they had more data. Dividing them by region had less of an impact.
¦ AE - Load Factor
—AE - Cumulative Load Factor
¦ CA- Load Factor
CA - Cumulative Load Factor
¦ CP - Load Facto*
—- CP - Cumulative Load Factor
HI - Load Factor
— Ml - Cumulative Load Factor
¦ NA - Load Factor
NA - Cumulative Load Factor
SA - Load Factor
SA - Cumulative Load Facto*
NP - Load Factor
NP - Cumulative Load Factor
¦ SP - Load Factor
— SP - Cumulative Load Factor
¦ PR - Load Factor
—PR - Cumulative Load Factor
GC - load Factor
GC - Cumulative Load Factor
• * • Engine Tost Cycle
©
Oi
B
C
©
p
©
CL
15
Load Factor Categories
Final Report 03.28987
31
-------
EPA Task Order 68HERC24F0432
¦ AE • Load Factor
— AE - Cumulative Load Factor
¦ CA - Load Factor
CA • Cumulative Load Factor
¦ CP - Load Factor
— CP • Cumulative Load Facto*
HI - Load Factor
- HI - Cumulative Load Factor
¦ NA - Load Factor
— NA • Cumulative Load Fador
SA- Load Factor
SA - Cumulative Load Factor
NP - Load Factor
NP - Cumulative Load Factor
• SP - Load Factor
— SP - Cumulative Load Factor
¦ PR - Load Factor
PR - Cumulative Load Factor
GC - Load Factor
GC - Cumulative Load Factor
Engine Test Cycle
-'roai-i-»MrowwAAuiaifflO)-vjNia)o:(D(D-1-4
^^^p^oc^o^owpwouipaio^ouioo
~-^»isoroci3co-U-k.cncna>o>->l~*NlC»CD^-k-k-*
o o ro ro
en o en o cn
Load Factor Categories
FIGURE 39. ESTIMATED DUTY CYCLE BY REGION FOR ALL TIER I SHIPS
IDENTIFIED WITHIN THE NORTH AMERICAN ECA (N= 1,960)
As shown in the above graphs, the operating duty cycles of ships within this dataset are
very different from the certification test duty cycles. This means that the certification test cycles
are not representative of actual ship operation. The impact of these deviations should be
investigated to ensure that regulations are achieving the goals for which they were developed.
Final Report 03.28987
32
-------
EPA Task Order 68HERC24F0432
3.0 CATEGORY 3 MARINE DIESEL ENGINE TECHNOLOGY ASSESSMENT
3.1 Tier III Ships in the U.S. ECA
Details on ships entering the U.S. ECA in 2022 were obtained from the US Coast Guard
AIS data. Each IMO number in the AIS database was cross referenced to the Clarksons vessel
database to identify the ship. The Clarksons database contained the main engine manufacturer and
some performance details such as the power and IMO emission level registration. Nine hundred
and eighty-six (986) ships were identified as Category 3, Tier III, vessels. Over ninety-seven (97)
percent of these Tier III ships were powered by 2-stroke engines.
There are two main pathways to achieve Tier III NOx levels for 2-stroke marine diesel
engines: exhaust aftertreatment in the form of selective catalytic reduction (SCR) or application of
exhaust gas recirculation (EGR). The SCR system can be placed prior to the turbine (high-pressure,
HP-SCR) or after the turbine (low-pressure, LP-SCR).
The main advantage of HP-SCR is that this location has higher exhaust temperatures
needed for efficient SCR operation. The main disadvantage is the space constraint of having to
place the SCR between the exhaust receiver and the turbine inlet. LP-SCR offers greater flexibility
in the placement of the SCR but has the disadvantage of low exhaust temperatures after the exhaust
has been expanded through the turbine, impacting the SCR efficiency and range of operation.
HP-SCR, where the SCR is placed before the turbine (high pressure side of turbine), seems
prevalent when looking at options on the MAN Computerized Engine Application System (CAES)
website (MAN Energy Solutions, n.d.), but this really depends on the size of the engine, with some
of the largest engines having the LP-SCR as the only option.
Figure 40 illustrates the Tier III ship break down by fuel type. While there are a significant
number of dual fuel ships, over seventy (70) percent of Tier III ships were strictly diesel fuel
powered. It should be noted that dual fuel engines using early cycle, low-pressure, gas injection,
with a small amount of diesel pilot can meet Tier III NOx emissions without SCR but generally
utilize EGR to reduce the tendency to knock. The WinGD X-DF and the MAN ME-GA engines
are examples of engines using low-pressure gas injection. The MAN ME-GA product line was
discontinued in late 2024 (Snyder, 2024) due to upcoming IMO methane emission regulations.
Late cycle, high-pressure, injection of gas with a small diesel pilot will generally need EGR or
SCR to meet Tier III NOx levels. A comparison of the emissions of each approach relative to diesel
only is shown in Figure 41 where the DF/gas LP represents the low-pressure gas injection
approach, and the GD/gas HP represents the "gas diesel" high-pressure gas injection approach. As
illustrated, the DF/gas LP approach significantly reduces engine out NOx emissions and can
achieve Tier III NOx levels without SCR. The GD/gas HP approach would require similar NOx
countermeasures as the diesel (SCR or EGR) to meet the Tier III NOx limits.
Final Report 03.28987
33
-------
EPA Task Order 68HERC24F0432
The Tier III ships operating in the US EC A were predominately powered by MAN engines
as illustrated in Figure 42. MAN reported reaching 1,000 Tier III engine orders in 2021 of which
25-percent were EGR solutions and 75-percent were SCR (Soholt, 2021). The split between EGR
and SCR solutions continues to evolve (in 2022 MAN reported EGR solutions were 36-percent of
the market (Blenkey, 2022), but SCR continues to be the primary path. As an example of the
variety of options to meet Tier III, the MAN 60ME engine was selected. Figure 43 illustrates the
various engine configurations that are Tier III compliant (MAN Energy Solutions, n.d.)Error! B
ookmarknot defined, includes a DI (diesel) configuration as well as dual fuel configurations: GI
(gas injection LNG), LGIM (liquid gas injection methanol), LGIP (liquid gas injection propane),
and GIE (gas injection ethane). The DI, GI, and LGIM versions are offered with either HP-SCR
or various versions of EGR. The LGIP and GIE versions are offered with HP-SCR only.
Tier
III, Category 3 Vessels in US ECA
80%
70%
c
60%
o
50%
40%
o"
o_
30%
"ai
t/i
i/i
20%
Ol
>
10%
_ m — — _
0%
Diesel
DF Ethane DF LNG DF LPG DFMeOH Biofuel
Fuel
FIGURE 40. TIER III, CATEGORY 3, VESSELS IN THE US ECA
Final Report 03.28987
34
-------
EPA Task Order 68HERC24F0432
100
l/l
CU
JD
fO
>
c
o
in
in
E
LU
DF/gas
LP
GD/Gas
HP
Diesel / HFO
FIGURE 41. COMPARISON OF EMISSIONS FROM TWO DUAL FUEL
TECHNOLOGY APPROACHS RELATIVE TO DIESEL (WERNER, 2019)
Tier Hi, Category 3, Ships in the US ECAs
80% -
^ 70% -
vO
=~- 60%
i—H
| 50% -
£ 40% -
o
s 30% ^/
a 20% -
io%
P
0% '
J-ENG MAN B. & W
WinGD
Main Engine Designer
FIGURE 42. MAIN DIESEL ENGINE DESIGNER
Final Report 03.28987
35
-------
EPA Task Order 68HERC24F0432
FIGURE 43. TIER III OPTIONS FOR MAN 60ME ENGINE
3.2 Two-stroke Engine Overview
There are two main challenges for using SCR with 2-stroke engines and fuel that contains
sulfur. First, the SCR NOx conversion efficiency is dependent on temperature, with lower
temperatures having lower conversion efficiency. Second, the sulfur in the fuel results in
combustion products that can form ammonia bisulfate (ABS) in the exhaust by reacting with water
and ammonia (reductant required for SCR operation). The kinetics for ammonia bisulfate
formation are dependent on the concentration of the reactants, pressure, and temperature. More
details on ABS formation kinetics and rates is given in Section 4.4 of this report. High exhaust
temperatures are required to prevent the formation of ABS. Diesel engines in general, and 2-stroke
engines in particular, tend to have low exhaust temperatures at light loads due to very lean
operation (high air-fuel ratios). Additional discussion of ABS formation can be found in
Section 5.0.
Typical 2-stroke engine components and operation are depicted in Figure 44. Key features
include:
1) Intake ports built into the liner fix the intake opening and closing events.
2) Single exhaust valve per cylinder hydraulically driven on modern engines.
3) Exhaust gases, collected in manifold, drive a turbocharger.
4) Turbocharger compressor supplies air through the charge air cooler and water mist
collector to the scavenge air receiver and then the cylinders.
5) Positive differential pressure (Pin - Pex) required to force air into the cylinder.
6) At low load, exhaust energy is not sufficient to drive the turbocharger and create
boost, requiring the addition of an auxiliary blower to supply the scavenge air as
illustrated in Figure 45 and Figure 46.
Final Report 03.28987
36
-------
EPA Task Order 68HERC24F0432
Charge
Air
Cooler
Water
Mist
Catcher
Blower
(not shown)
MAN Project Guide 6G50ME-C9.5
Exhaust
Gas
Manifold
Exhaust Valve
Turbo
Charger
Cylinder Liner
/
Scavenge
Air
Receiver
FIGURE 44. TWO-STROKE ENGINE LAYOUT AND OPERATION (MAN ENERGY
SOLUTIONS, 2021)
Final Report 03.28987
37
-------
EPA Task Order 68HERC24F0432
to
-Q
<1>
a
20
40 60
Power Output (%)
80
100
— 1.00
0.75
0.50
0.25
0.00
-0.25
FIGURE 45. INTAKE AND EXHAUST PRESSURE CHARACTERISTICS OF TYPICAL
2-STOKE ENGINE (ILLUSTRATION ONLY)
2.25
2.00
1.75
1.50
1.25
— — Charge Air Pressure
Charge Pressure w/ Blower
Exhaust Press.
Delta Pressure w. Blower
— — Delta Pressure w/o Blower
Ambient Pressure
'o
CL
3
w
4
Differential Pressure
10 15 20 25
Power Output (%)
30
35
40
FIGURE 46. INTAKE AIR PRESSURE IMPROVEMENT WITH AUXILIARY
BLOWER (ILLUSTRATION ONLY)
Final Report 03.28987
38
-------
EPA Task Order 68HERC24F0432
Typical exhaust temperatures for a range of Tier II WINGD engines are shown in Figure
47 (WinGD, n.d.). The legend refers to the bore diameter, ranging from 52-92 cm. As shown,
turbine inlet temperatures for 50-percent load and above are well above 300°C for avoidance of
ABS formation and for high SCR conversion efficiency. These temperatures are favorable for HP-
SCR. In contrast, the turbine outlet temperatures are much lower and present a challenge for LP-
SCR operation and avoidance of ABS formation. More detail regarding ABS formation rates and
temperatures is given in Section 4.4 of this report.
Power Output (%)
FIGURE 47. EXHAUST TEMPERATURE RANGE FOR TIER II ENGINES
3.3 Current Tier III Thermal Management Technologies
To meet the Tier III NOx emission target with SCR, engine manufacturers have modified
the engine performance to raise exhaust temperatures to enable ITP-SCR operation at loads
between 25- and 50-percent, depending on the engine model. A comparison between Tier II and
Tier III exhaust temperatures is shown in Figure 48 for the WinGD 8X52 engine (WinGD, n.d.).
As shown, Tier III operation has higher exhaust temperatures below 55-percent load. This is
accomplished with a turbocharger bypass valve (wastegate) which reduces flow to the turbine and,
correspondingly, the compressor flow and the scavenging air to the cylinder, thus lowering air-
fuel ratio (AFR) and raising exhaust temperatures. The turbocharger bypass flow is shown in
Figure 49 while the estimated AFR (illustration only) is shown in Figure 50.
Final Report 03.28987
39
-------
EPA Task Order 68HERC24F0432
WinGD 8X52-1.l_14480kW_105rpm
500
u
450
-------
EPA Task Order 68HERC24F0432
WinGD 8X52-1.l_14480kW_105rpm
70
65
60
ai
55
Ll_
<
50
-a
ai
45
ro
E
40
'u
L_J
35
30
25
20
ZZ7—
^ rrn ~r;„„ ¦
—slk, i ier i
1
—•—Tier II
20
40 60 80
Engine Load {%)
100
120
FIGURE 50. ESTIMATED AFR FOR TIER II AND TIER III OPERATION
In another example, Hitachi Zosen Corporation published results for a MAN engine using
HP-SCR down to 8-percent load (Fujibayashi, T., et. al., 2013). At light loads, cylinder by-pass,
which routed air from the scavenging air receiver (compressor outlet) to the turbine inlet, was used
to reduce the air flow to the cylinder (bypassing the cylinder while still maintaining the scavenging
differential pressure). This resulted in lowering the AFR, which raised exhaust temperatures. This
approach was demonstrated to be successful in maintaining approximately 75-percent NOx
conversion down to 8-percent load as shown in Figure 51.
S*m» v«U. m« vwi. 150CT2011. SCR mo0«: Auto
30
20
10
0
0 10 20 30 40 60 60 70 80 90 100
Engine load %
\-m- DeNOx ratio % NOx SCR outlet q/kWh]
200
0 10 20 30 40 50 60 70 80 90 100
Engine toad %
|-B- T1 (exh receiver) [degC] T3 (reactor inlet) |degC][
o
z
&400
9
¦o
g. 350
a>
I zoo
iS 250
FIGURE 51. SCR PERFORMANCE AND SCR INLET TEMPERATURES
(FUJIBAYASHI, T„ ET. AL., 2013)
In addition to controlling AFR via either turbocharger or cylinder bypass, modern two-
stroke engines also employ common rail injections systems and hydraulically actuated exhaust
valves (Kyrtatos, A., et.al., 2016) (Kindt, 2016) providing flexibility over injection timing and
exhaust valve opening and closing. Either feature can be tuned to provide an increase in exhaust
energy (and temperature), albeit, with a small increase in fuel consumption.
Final Report 03.28987 41
-------
EPA Task Order 68HERC24F0432
3.4 Applicability of Heavy-Duty Engine Thermal Management Technologies
Emission regulations for the heavy-duty truck industry have been progressively more
stringent since the first regulation in 1974. As such, HD truck emission abatement technology is
understandably more advanced than in the present-day marine industry. HD thermal management
technologies to optimize and improve SCR light off and performance at light loads are
sophisticated and can included (United States Environmental Protection Agency, 2022):
• Intake throttle
• Heated aftertreatment system
• Exhaust flow bypass systems
• Late combustion phasing
• Variable valve actuation
• Cylinder deactivation
• Pre-turbine aftertreatment location
• Aftertreatment insulation
• Use of 0.0015% ULSD fuel
The applicability of each of these technologies to the two-stroke marine diesel engines was
considered. Engineering assessment was made based on 8 criteria: feasibility, NOx reduction
potential, space consumption (need for space to install the technology), benefit at high loads,
reliability, Cost - CAPEX, cost OPEX, ability to retrofit, and scalability for different size engines.
Table 5 provides a brief description of each evaluation criteria.
TABLE 5. THERM AT, MANAGEMENT TECHNOLOGY EVALUATION CRITERIA
Feasibility
The degree of challenge to implement the technology and applicability.
NOx Reduction
Potential
How far does the technology extend the SCR operating range?
Operating
Range
Is this technology beneficial at higher loads as well as low loads?
Integration
How easy is it to integrate and how much space does it take?
Reliability
How does it impact overall reliability? How reliable is the technology?
CAPEX
Relative to today's Tier III systems, does it increase the CAPEX
expenditure? Additional components?
OPEX
Is there a BSFC penalty, fuel cost increase, or cost of additional fluids?
Retrofit
Can the system be easily retrofitted on Tier II engines?
Scalability
Can this be applied to small and large marine diesel engines?
Final Report 03.28987
42
-------
EPA Task Order 68HERC24F0432
3.7.1. Intake Throttling/AFR Reduction
Intake throttling is used in the HD engine industry to reduce the air flow to the cylinder
and thereby lower AFR and increase exhaust temperatures. Due to the nature of the scavenging
process for two-stroke engines, that requires a higher intake than exhaust pressure, intake
throttling, used by itself, is not applicable to two-stroke engines. One might envision using a
suction blower on the exhaust side of the engine to lower the exhaust pressure (sub-atmospheric)
in combination with throttling to control the air flow and scavenging. Some implementations of
EGR in two-stroke engines use a blower on the exhaust side to drive EGR so using a similar setup
at low load to reduce the exhaust pressure might be feasible. As noted previously, current industry
practice includes some form of AFR reduction at loads below 60-percent in the form of either
cylinder or turbine bypass and this type of setup might be used along with throttling and an exhaust
suction blower to facilitate optimization.
• Feasibility Low for intake throttling, only with exhaust suction
blower, High for other methods of AFR reduction
NOx Reduction Potential:
Space Consumption:
Benefit at High Loads
Reliability:
Cost - CAPEX
C02-OPEX
Retrofit
Scalability
3.7.2 Aftertreatment System Heating
SCR down to -10-15 % MCR
Suction blower installation
Only at low loads, no benefit at high loads
Negative influence due to blower maintenance
Installation cost of blower
Slightly higher, 3 to 5 % BSFC
Possible
Could be scaled to large engines
In the HD engine industry, heat addition is usually accomplished by adding fuel to the
exhaust which then oxidizes over a DOC. Tier III marine diesel engines don't currently use a DOC
in conjunction with the SCR system because the fuel sulfur limit of the ECA is too high (1000
ppm). Implementing that approach would require additional capital expenditure for the DOC, an
additional space requirement, and ULSD fuel. HP-SCR systems are already space constrained so
there may not be room to accommodate a DOC. The space constraints of an HP-SCR system would
also make a burner challenging. The impact of the additional energy on the turbocharger needs to
be considered. Adding energy to the exhaust pre-turbine has the potential of increasing the turbine
and compressor work which would generate more boost increasing air flow and tending to increase
AFR and reduce exhaust temperature. A turbine or compressor bypass valve may be required to
maintain balance and get the desired effect. The energy requirement would represent a BSFC
penalty and additional CO2 emissions.
A burner would be more practical for a LP-SCR system where space is less constrained.
Burners are already in use today in marine applications. One example is the use of a burner to heat
a portion of the exhaust sufficiently to evaporate the dosed urea prior to recombination with the
main exhaust stream (MAN Energy Solutions, 2021).
Final Report 03.28987
43
-------
EPA Task Order 68HERC24F0432
An electric heater prior to the SCR or an electrically heated catalyst (EHC) might be a
possibility, but durability would be a concern when scaled up to large marine engine sizes. In the
HD engine industry, 48V e-heaters are primarily used for cold start to facilitate catalyst light off
and emission reduction, as well as low load thermal management. Electric grid heating in the
exhaust port, much like a grid heater in the intake port for help during cold start on 4-stroke
HD diesel engines, might be adaptable for use in marine applications. The space requirement
would be small, but the energy and heat transfer needed over a short distance would make this
approach challenging.
Feasibility
NOx Reduction Potential:
thermal management
Space Consumption:
space available for LP-SCR
Benefit at High Loads
Reliability:
Cost - CAPEX
C02-OPEX
Retrofit
Scalability
Low for HP-SCR systems
SCR down to -5-10 % MCR combined with other
Minimal space for heater or burner prior to HP-SCR,
Low load only, no benefit at high loads
Negative influence, burner or heater durability
Installation cost
Slightly higher, 3 to 5 % BSFC
Challenging depending on the approach
Could be scaled to large engines
3.7.3 Exhaust Flow Bypass System
Exhaust flow bypass is used for HD engines to retain energy in the exhaust instead of
expanding it through the turbine assisting with catalyst light off and temperature at low loads.
Since this discussion is considering SCR installations that are HP-SCR, installed prior to the
turbine, exhaust flow bypass will be of little help to raise SCR inlet temperatures, improving SCR
efficiency and reducing NOx emissions at low load.
• Feasibility Low for HP-SCR systems
3.7.4 Late Combustion Phasing
Late combustion phasing (retarding the start of fuel injection, SOI) can be helpful in
reducing the cycle efficiency shifting energy from piston work (expansion) to the exhaust. The
increase in exhaust temperature is expected to be small relative to that required so this approach
would be insufficient on its own but may be used in conjunction with other technologies. Another
incremental approach toward making the combustion cycle less efficient would be to reduce the
compression ratio. Technology for variable compression ratio (VCR) currently exists and might
be used in conjunction with late combustion phasing.
• Feasibility High with existing common rail injection systems
Final Report 03.28987 44
-------
EPA Task Order 68HERC24F0432
NOx Reduction Potential:
Space Consumption:
Benefit at High Loads
Reliability:
No, engine internal
Load low, no benefit at high loads
Slightly lower
Similar to current Tier III engines
Slightly higher, 4 to 6 % BSFC
Likely, injector space claim should be similar
Could be scaled to large engines
SCR down to ~ 10 % MCR
Cost - CAPEX
C02-OPEX
Retrofit
Scalability
3.7.5 Variable Valve Actuation
Variable valve actuation (VVA) is used in the HD engine industry for active thermal
management, altering lift and/or timing of intake and/or exhaust valves. For two-stroke engines,
because the intake ports are fixed in the cylinder liner, VVA can only be applied to the exhaust
valve. Modern two-stroke engines, as noted above, have hydraulically actuated exhaust valves
providing timing flexibility for both opening and closing of the valves. Late closing of the exhaust
valve reduces the effective compression ratio and peak cylinder pressure (useful at high loads)
while early opening of the exhaust valve transfers energy from work to the exhaust (useful at low
loads).
Early exhaust valve opening (EEVO) has multiple effects. Since EEVO reduces the
expansion ratio (ratio of the cylinder volume at EEVO to the cylinder volume at TDC), the cycle
efficiency would be decreased which would increase the BSFC, requiring more fuel at a given
power level. If air flow were to remain constant, the additional fuel would reduce AFR and
contribute to higher exhaust temperatures. However, the additional exhaust energy available to the
turbocharger can potentially increase boost and lead to higher airflow, increasing AFR. The
turbocharger response is highly dependent on where the operating conditions fall within the
turbocharger operating range (turbine and compressor maps). At the lightest loads, the
turbocharger is generally not functioning as an additional blower is required. At moderately light
loads, the turbocharger response could be significant and turbine bypass (wastegate) or cylinder
bypass could be required to optimize tradeoffs affecting BSFC, scavenging, AFR, and exhaust
temperatures.
Timing of the exhaust valve closing, in essence, controls the time available for the piston
to force gases from the previous cycle out of the cylinder during the compression event. Later EVC
reduces the compression work and in-cylinder residual exhaust gas. Earlier EVC traps more
residual exhaust gas and increases the compression work. Trapping more residual exhaust gas
leads to higher in-cylinder temperatures and ultimately, higher exhaust temperatures. Optimization
of the exhaust valve events is critical to low load exhaust temperatures.
• Feasibility High with existing hydraulically controlled exhaust
valve
Final Report 03.28987
45
-------
EPA Task Order 68HERC24F0432
• NOx Reduction Potential: SCR down to 10 to 15 % MCR, could extend to 5%
combined with other thermal management
3.7.6 Cylinder Deactivation
The use of cylinder deactivation in modern two-stroke marine diesel engines is possible.
Common rail injection systems enable fuel shutoff on a cylinder basis and the VVA system can
deactivate the exhaust valve. Deactivating one or more cylinders would increase the load of the
firing cylinders and the overall exhaust temperature due to the reduction in AFR. Increasing the
load of some cylinders may also lead to an efficiency improvement, reducing CO2 emissions.
Cylinder deactivation on a 16-cylinder, EMD, locomotive uniflow two-stroke engine at a
low speed, low power, condition (referred to as Notch 1) was demonstrated recently by Fritz and
Riley (Fritz & Riley, 2024). The results are shown in
Figure 52. The figure illustrates multiple strategies that were applied to achieve the desired
exhaust temperatures. The engine was equipped with two Roots blowers. The initial step taken
was to deactivate one of the blowers while operating on 16 cylinders, reducing the AFR from -160
to -90:1. Deactivating 8- and 12-cylinders lead to higher exhaust temperatures. Additionally, while
operating on 4-cylinders, the injection timing was retarded by 8 degrees, from 4°BTDC to
4°ATDC which yielded a slight increase in exhaust temperature. Additional backpressure was
added to reduce scavenging, which led to modest exhaust temperature increase. Finally, the AFR
was further reduced, while operating on 4- and 8-cylinders, by bleeding off air after the compressor
(similar to the cylinder bypass concept) resulting in a reduction in AFR and corresponding increase
in exhaust gas temperature. For slow speed inline marine diesel engines, torsional vibrations
would need to be evaluated to check if cylinder deactivation would be feasible.
Space Consumption:
Benefit at High Loads
Reliability:
No, engine internal
No limit, most benefit at low load
Slightly lower
Low, only software, similar to current Tier III engines
Slightly higher, 2 to 4 % BSFC
Likely, exhaust valve space needed is similar
Could be scaled to large engines
Cost - CAPEX
C02-OPEX
Retrofit
Scalability
Final Report 03.28987
46
-------
EPA Task Order 68HERC24F0432
350
300
°L 250
S_
D
| 200
U
SZL
E
Q)
f- 150
4-
V)
=3
x 100
OJ
50
Notch 1 - Avg Exhaust Port Temperature vs Air Fuel Ratio Control
1
4 Cylinders
•
8 Cylinders^
•V
• \
•
•
16 Cylinders
0 20
• Nl4ATDC
• Nl 8cyl
40
60
80 100
AFR
120
140
160
180
• Nl 4BTDC
• Nl 16cyl
Nl 4ATDCBP
Nl 16cyl 2blwr
Nl 4ATDC BP AFR
FIGURE 52. EFFECT OF CYLINDER DEACTIV ATION ON EXHAUST
TEMPERATURES OF A 16-CYLINDER, EMD LOCOMOTIVE TWO-STROKE
ENGINE AT LIGHT LOAD (FRITZ & RILEY, 2024)
Feasibility
NOx Reduction Potential:
combined with other thermal
Space Consumption:
Benefit at High Loads
Reliability:
Cost - CAPEX
C02- OPEX
Retrofit
Scalability
Possible, but torsional vibrations need to be checked
SCR down to 10 to 15 % MCR, could extend to 5%
management
No, only software
Low load, no benefit at high load
Negative influence possible
Low, only software
Slightly lower, 1 to 2 % BSFC improvement
Likely, injector, exhaust valve systems and controller
Could be scaled to large engines
3.7.8 Pre-turbine Aftertreatment Location
Close coupling of the aftertreatment system in heavy and light-duty applications improves
catalyst light off and thermal management. However, there are no significant instances of pre-
turbine aftertreatment systems in production for on-highway engines. However, in two-stroke C3
Final Report 03.28987
47
-------
EPA Task Order 68HERC24F0432
marine applications, manufacturers are placing the SCR system pre-turbine, when possible, to
achieve the highest exhaust temperatures at the SCR inlet. This approach appears to be most
prevalent for smaller engines. In 2019, MAN reported the trends for LP-SCR, HP-SCR, and EGR
by engine size (Struckmeier, D., et al., 2019), which can be found in Figure 53. A typical HP-SCR
layout locates the SCR separate from the engine but in the engine room as illustrated in Figure 54.
Recommendations from the initial development and demonstration of HP-SCR included keeping
the piping as short as possible and insulating the pipes (Fujibayashi, T., et. al., 2013). Designs have
progressed and WinGD now offers an integrated SCR system (Kyrtatos, A., et.al., 2016) (Spahni,
M., et. al., 2023), called iSCR for about 39-percent (WinGD, n.d.) of diesel engine models where
the SCR is placed on the engine as shown in Figure 55. The exhaust flow path for the WinGD
iSCR system is shown for Tier II and Tier III modes in Figure 56. (Spahni, M., et. al., 2023). The
flow path is controlled by shutoff valves in the two flow paths. In Tier II mode, the exhaust is
routed directly from the exhaust manifold to the manifold outlet. In Tier III mode, the exhaust is
routed from the exhaust manifold (top) to the bottom through passages at the ends of the manifold.
This exhaust then flows through two SCR catalysts back to the center of the unit and then upwards
to the exhaust manifold and manifold outlet. Urea dosing takes place in the exhaust port runner
with air assisted injectors.
Number of Engines by Bore Size
_
LPSCR
M HPSCR
EGR
_
-
1
I
1 1 1
50 and Less 60-70 80 and more
Bore size
FIGURE 53. CHOICE OF IMO NOx TIER III STRATEGY FOR MAN TWO-STROKE
ENGINES IN RELATION TO ENGINE SIZE (STRUCKMEIER, D., ET AL., 2019)
Final Report 03.28987
48
-------
EPA Task Order 68HERC24F0432
Turbo-
charger
Vaporiser
(containing urea injector and mixer)
SCR reactor
(containing catalyst blocks
and soot-blowers)
<=t>
: SCR operation
: Non-SCR operation
FIGURE 54. EXAMPLE OF HIGH-PRESSURE SCR SYSTEM LAYOUT
(FUJIBAYASHI, T., ET. A I.., 2013)
FIGURE 55. EXAMPLE OF WINGD ISCR, INTEGRATED HPSCR SYSTEM
(KYRTATOS, A., ET.AL., 2016)
Final Report 03.28987
49
-------
EPA Task Order 68HERC24F0432
iSCR in Tier II mode-Tier II valve (red) open, iSCR in Tier III mode - Tier II valve (red) closed,
Tier III valves (blue) closed Tier III valves (blue) open
FIGURE 56. EXHAUST FLOW PATH OF WINGD ISCR IN TIER II (LEFT) AND TIER
III (RIGHT) MODES (SP A 11M, M., ET. AL., 2023)
Feasibility
NOx Reduction Potential:
Space Consumption:
Benefit at High Loads
Reliability:
Cost - CAPEX
C02- OPEX
Retrofit
Scalability
High, already implemented on many engines
Needs other technologies to extend SCR operating
range, SCR down to ~ 5-10% with thermal management
Constrained
Beneficial for mid-load operation
No influence
Same as current
Dependent on other technologies implemented
Possible, Exhaust system and piping to turbocharger
SCR volume requirement may limit application
3.7.9 A ftertreatment. Insulation
For stationary engines, insultation of exhaust piping is typically done as a safety measure
to reduce the outer surface temperature. If the piping is not insulated, some sort of shielding would
be used. For HP-SCR systems, additional insultation or shielding would likely be of no benefit.
• Feasibility Low
3.7.10 ULSD
The current sulfur limit for diesel fuel in the ECA is 0.1 percent by weight. The sulfur in
the fuel ends up primarily as SO2 in the exhaust, although a small portion (typically around 5%) is
oxidized to SO3 (and H2SO4 or a few other oxidized SO2 byproducts). SO3 can potentially react
with the SCR reductant, NH3, to form ammonium bisulfate (ABS) at low temperatures. ABS
deposition results in catalyst fouling and corrosion over time, and therefore SCR can only be
utilized above the ABS condensation temperature, which for current systems at loads above 20%
is generally around 300°C. Therefore, for marine SCR systems, avoidance of ABS condensation
is one of the primaiy factors limiting low temperature conversion in the system. Since the
Final Report 03.28987 50
-------
EPA Task Order 68HERC24F0432
formation of ABS is dependent on the concentration of SO3 as well as temperature, reducing the
SO3 concentration can enable the SCR to operate at a lower temperature, in essence, extending the
low load operating zone of the SCR. A more complete analysis of ABS formation is discussed
below in Section 4. Switching to ULSD, with a sulfur limit of 15 ppm, would extend the operating
range of the SCR system, reducing the ABS formation temperature at light loads to as low as
225°C (which is below the effective temperature of the vanadium-SCR catalysts anyway). This
would, in turn, require much less thermal management to operate the SCR at even very light loads
in the range of 5%. Ideally, this could be a direct replacement of the current ECA fuel with ULSD,
however there would be a cost impact to fuel price. It does not seem practical to carry an additional
fuel type for only low load operations in the ECA, so it is likely that this change would have to
apply to all fuel used in the ECA. Ultimately this may be cost-benefit analysis weighing the
additional cost of ULSD compared to 0.1-percent fuel versus the addition and complexity of a
third fuel system.
• Feasibility High, could replace current ECA fuel
• NOx reduction Potential: SCR down to 5% MCR with less thermal
management need than other methods
• Space consumption: May require an additional fuel tank
• Benefit at High Loads No limitation
• Reliability: No influence
• Cost - CAPEX Same as current
• C02 - OPEX Fuel price increase, but some fuel consumption
decrease, net increase in cost expected
• Retrofit Yes
• Scalability Could be scaled to large engines
3.5 Other Thermal Management or SCR Technologies
3.8.1 Intake A ir Heating
Raising the starting point temperature of the process will raise the temperatures throughout
the process. Addition of heat to the intake air would raise the temperature leading to a higher
exhaust gas temperature. Typically, the charge air cooler after the compressor is used to remove
heat added during the compression of the air. However, at light load, there is no compressor work
being done on the fluid, so the temperature rise across the compressor is small. If the charge air
cooler water temperature was increased, it would add heat to the inlet air raising the temperature.
This could be done with modification of the cooling water circuit to add a heat exchanger for using
process steam, if available on board, or installing an electric heater. Approximately, a 1-to-l
temperature increase would be expected. Higher in-cylinder temperatures would lead to some
combustion effects such as a short ignition delay and higher NOx, which could be mitigated by
optimizing the fuel injection timing.
• Feasibility High
Final Report 03.28987
51
-------
EPA Task Order 68HERC24F0432
• NOx Reduction Potential:
SCR down to ~ 5-10 % MCR combined with other
thermal management
• Space Consumption:
• Operational Range
• Reliability:
Low load only, no benefit at high load
No influence
Small
Slightly higher
Likely
Could be scaled to large engines
Small
• Cost - CAPEX
• C02-OPEX
• Retrofit
• Scalability
3.8.2 Direct Use of Ammonia
Urea could be replaced directly by ammonia, or a separate burner or heater could be used
to vaporize the urea and decompose it into ammonia. Ammonia is often used on board for
refrigeration systems so safety protocols may already exist. A burner for urea decomposition is
also in use today on some ships, generally for LP-SCR systems (MAN Energy Solutions, 2021).
This approach would allow SCR introduction of ammonia down to ~150°C. However, low
temperature SCR performance is not currently limited by low temperature DEF injection limits,
especially on HP-SCR systems. Instead, low temperature SCR performance is limited primarily
by the risk of ABS formation, which direct NH3 introduction does nothing to resolve. In addition,
the current vanadium-based SCR technology that is preferred on OGVs does not perform well at
temperatures below 250°C, so the lower temperature introduction of NH3 would not yield much
performance benefit. The estimated potential for using ammonia is summarized below.
• Feasibility Moderate, safety concerns
• NOx Reduction Potential: Limited improvement without other thermal
Cost - CAPEX
C02- OPEX
Retrofit
Scalability
Space Consumption:
Benefit at High Loads
Reliability:
management technologies.
Additional tank system
No benefit at high load
No influence
Installation cost
Probably slightly higher
Likely
Could be scaled to large engines
Final Report 03.28987
52
-------
EPA Task Order 68HERC24F0432
VAR Stack
Figure 41: LPSCR with DCU installation
FIGURE 57. ILLUSTRATION OF UREA DECOMPOSITION UNIT (DCU) AND
BURNER FOR LP-SCR SYSTEMS (MAN ENERGY SOLUTIONS, 2021)
3.8.3 Cylinder and or Turbine Bypass
Cylinder bypass and turbine bypass have been mentioned (see sections 3.3 and 3.7.1)
previously as examples of AFR management technologies currently in use at loads above -20-
percent to raise exhaust temperatures. These technologies should be considered for extension of
exhaust thermal management to operation below 20-percent load, particularly since it would entail
mostly additional calibration work with minimal engine modifications. Also, there may be synergy
with other technologies (i.e. cylinder deactivation or VVA) that may not be currently in-use or
used to the fullest potential.
3.6 Summary
The technologies described above are summarized in the Table 6, below. To visually
illustrate the discussion, engineering judgement was used to assign a numeric score to each
category for each technology. The numeric score was based on a scale of 1 (unfavorable) to
5 (favorable). Conditional formatting (i.e., color coding) was used to help visualize the rankings.
Some comments about the rating process are appropriate.
As discussed above, exhaust bypass and exhaust system insulation are not expected to be
feasible technical solutions for increasing exhaust temperatures. Intake throttling might be possible
with a suction blower on the exhaust side but is ranked as slightly unfavorable. Heated
aftertreatment systems have an unfavorable rating due to questions about durability. The other
technologies have favorable ratings with technologies that are currently existing in some form
having the highest rating.
Final Report 03.28987
53
-------
EPA Task Order 68HERC24F0432
In terms of NOx reduction potential, intake throttling and heated aftertreatment were given
the same score as for feasibility due to challenges in implementation may hamper the NOx
reduction potential. The direct use of ammonia has the least NOx reduction potential as discussed
above. The other technologies received a slightly favorable rating.
None of the technologies considered are beneficial at improving SCR performance at high
loads so all were rated neutral in this regard.
Some of the technologies discussed are already in use so there is no space or integration
issue. These technologies, which included late combustion phasing, VVA, cylinder deactivation,
and cylinder or turbine bypass, received a score of 5. Pre-turbine aftertreatment is already in use
on some engines, but not on others, so while the in-use cases would score a 5, the other cases
would score a 1, resulting in an aggregate of 3. The impact of direct use of ammonia or ULSD on
space and integration really depends on the implementation. These could be drop-in technologies,
replacing systems already in use or add-on technologies. Without knowledge of the
implementation details, these technologies were rated as 3, neutral. Intake throttle, heated
aftertreatment, and air heating intake would require additional hardware and corresponding space
claim. These technologies were rated as slightly unfavorable, as 2.
Heated aftertreatment systems have the biggest question mark regarding reliability.
Whether it is an electrically heated catalyst, a burner prior to the SCR, or an electric grid heater
prior to the SCR, the high temperature required imposes a reliability risk. This technology was
scored as 1. Other technologies not currently in use, cylinder deactivation and air heating intake
were scored as slightly unfavorable for reliability, 2. Late combustion phasing, VVA, cylinder or
turbine bypass, and direct use of ammonia are technologies currently in use in some fashion and
would have little impact on reliability. Pre-turbine aftertreatment, if compared to applications
currently using LP-SCR, would likely improve reliability due to the decreased risk of ABS
formation. So this technology was rated as slightly favorable. The use of ULSD would reduce the
risk of ABS formation and was rated as 4.
No technology is going to reduce CAPEX so, at best, any technology is neutral and rated
a 3. If a given technology is going to include additional expenditure, then the rating will be lower
than 3 but would have to be a major expense to receive a rating of 1.
Operating cost, OPEX, is really a judgment on the BSFC penalty or the cost of additional
fluids (DEF, ammonia, ULSD). Only cylinder deactivation has the potential for improving BSFC,
so it received a favorable rating. Pre-turbine aftertreatment is a "passive" technology that is
currently in-use today so the impact on operating cost would be neutral. Other technologies would
incur some BSFC penalty or require additional or more expensive fluids so received a below
neutral rating. Late combustion phasing and VVA, while effective in increasing exhaust
temperatures, may have a less favorable trade-off with BSFC than other technologies so received
the lowest rating.
Final Report 03.28987
54
-------
EPA Task Order 68HERC24F0432
Except for heated aftertreatment systems, all technologies are viewed as retrofittable to
Tier II engines. This is reflected by a rating of 4 for these technologies. The use of ULSD was
rated 5 since it could be a direct replacement for today's current ECA fuel. Heated aftertreatment
systems, depending on the implementation method as discussed above, may pose more of a
challenge, which is reflected in its rating.
In general, apart from the pre-turbine SCR, all technologies are viewed as scalable to larger
engines. Pre-turbine SCR may be limited for the largest engines due to the volume of the SCR.
TABLE 6. THERMAL MANAGEMENT TECHNOLOGY EVALUATION
Technology
Feasibility
NOx
Reduction
Potential
Benefit at
High Loads
Integration/
Space
Consumption
Reliability
Capital
Cost
Operating
Cost
Suitability
for Retrofit
Scalability
1 nta ke Th rottle/AFR Red u cti on
2
2
3
2
2
2
2
4
4
Heated Aftertreatment System
2
2
3
2
1
2
2
2
4
Exhaust Flow Bypass Systems
1
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
Late Combustion Phasing
5
4
3
5
3
3
1
4
5
Va ri a ble Va Ive Actuati on
5
4
3
5
3
3
1
4
5
Cylinder Deactivation
4
4
3
5
2
3
4
4
5
Pre-Turbine Aftertreatment
5
4
3
3
4
3
3
4
3
Aftertreatment Insulation
1
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
ULSD
5
4
3
3
4
3
2
5
5
Intake Air Heating
4
4
3
2
2
2
2
4
5
Direct Use of Ammonia
4
1
3
3
3
2
2
4
5
Cylinder or Turbine Bypass
5
3
3
5
3
3
3
4
5
Final Report 03.28987
55
-------
EPA Task Order 68HERC24F0432
4.0 MARINE DIESEL ENGINE NOx AFTERTREATMENT AND
TECHNOLOGY PACKAGE ASSESSMENT
The previous report section discussed a number of potential technologies that could be
applied to Category 3 2-stroke marine primary propulsion engines as a means to achieve further
NOx reductions. To support an analysis of the potential for further NOx reductions in the ECA, it
was necessary to develop a projected technology package that could serve as a modeling case.
Ultimately, a set of technology packages based around high pressure (pre-turbine) SCR and
thermal management technologies were developed. These projections included different levels of
aggressiveness in the application of both SCR and thermal management. This section of the report
details the technology choices that were made in the development of these packages, including the
background, assumptions, and supporting data used to develop the final projections.
4.1 Exhaust Temperature Profile for Marine 2-Stroke Engines
SCR efficiency is dependent, in part, on the exhaust temperature at the inlet of the SCR
system. With the objective of extending the SCR system performance below 25-percent load, an
exhaust temperature profile was required in this region. The literature was reviewed to determine
data available for estimating the exhaust temperature profile versus load. Unfortunately, most of
the published data was at 25-pecent load and higher. Fujibayashi, et al. (Fujibayashi, T., et. al.,
2013) published data of SCR system performance and inlet temperatures down to 8-percent load.
WinGD published engine performance data which includes exhaust temperature before the turbine
in their general technical data (GTD) software (WinGD, n.d.). The WinGD data is at 5- and 10-
percent load steps down to 25-percent load. WinGD GTD also provides the nominal exhaust
temperature decrease from 25-percent load for 20-, 15-, and 10-percent loads. MAN publishes
engine performance data on their website in the form of the MAN Computerized Engine
Application System (CEAS) (MAN Energy Solutions, n.d.). While the MAN data does not contain
pre-turbine exhaust temperatures, it does contain the SCR inlet temperature for Tier III
applications at 25-, 50-, 75-, and 100-pecent load. The pre-turbine temperature generally trends
with load with higher loads having higher temperatures. The temperature can vary somewhat
depending on the size of the engine and the engine calibration. A comparison of the available data
from the three sources is shown in Figure 58. The left figure illustrates the pre-turbine exhaust
temperatures in Tier II operation while the graph on the right shows the SCR inlet temperatures.
These graphs illustrate three points. First, general trends for different engines are similar with a
maximum spread of 50°C in the mid-load range. Second, different calibrations can have up to a
50°C spread in exhaust temperature for the engine model and this is most apparent in the mid-load
range. Third, the Tier III temperatures essentially follow the Tier II temperatures until about 50-
percent load (~310°C) at which point thermal management maintains a constant temperature down
to 25-percent load.
Final Report 03.28987
56
-------
EPA Task Order 68HERC24F0432
Of the data shown, the WinGD X52 data with the low load turning (LLT) calibration
seemed to be 011 the low range of temperatures and would be the most conservative choice for
further analysis. The X52-LLT series engine was selected to provide the nominal exhaust
temperature versus load profile. The Tier II and III exhaust temperatures versus load for this engine
are shown in Figure 59. The unknown is the Tier III exhaust temperatures below 25-percent load.
Regardless of the technology used to maintain temperature below 25-percent, the best-case
scenario would be a gradual decline in temperature to 260°C at 5-percent load., which is illustrated
in Figure 59. For analysis purposes, any number of intermediate cases could be selected. An
example intermediate case has also been provided in Figure 59.
Comparison of Pre-Turbine Temperatures - Tier I
i
500
1*450
: 400
i
[ 3S0
¦¦J00
I1®
\ ?oo
' ISO
100
40 GO 80
Fnginp I oad (%)
Comparison of SCR Inlet Temperatures - Tier I
500
i-450
: 400
i
| 3bO
i
[¦JUG
i
: 250
j ?00
' 150
100
40 GO 80
Engine load [%)
FIGURE 58. EXHAUST TEMPERATURE DATA FOR MARINE TWO-STROKE
ENGINE
FIGURE 59. EXHAUST TEMPERATURE VERSUS LO AD PROFILE FOR WINGD X52
SERIES ENGINE
Final Report 03.28987 5 7
Before Turbine
¦WinGDX52 I IT
¦WinGD X5? DBT
t-ujibavashi
WinGDXS? Ill
¦WinGDX52 DBf
I ujibayashi
MAN GGSOMr IIP5CR
-------
EPA Task Order 68HERC24F0432
4.2 Brake Specific NOx Profiles for Marine 2-Stroke Engines
Brake specific NOx data are provided as an output of the MAN CEAS. For Tier III SCR
engines are provided, in Figure 60. The upper and lower ranges from MAN CEAS for the engine
out (SCR in) data represent values that would meet the Tier II standard, in the case of the minimum,
and meet the Tier I standard, in the case of the maximum. Although these data are nominal values
(the same for all engines with HP-SCR), they can be used for calculations looking at the
contribution of low loads to NOx and estimate the benefit if SCR operation were extended to lower
loads. For these calculations, the average of the minimum and maximum values was assumed to
account for the possibility that the engine calibration might be tuned to produce engine out NOx
higher than Tier II to gain back some energy efficiency when using SCR. Another assumption that
was required was the engine out NOx emissions below 25-percent load. For purposes of
calculations, it was assumed to be an extrapolation of the SCR in curve. This is considered a
conservative assumption since the brake-specific NOx may be much higher since operation at
these loads is almost always governed by an approved ACD.
Engine-Out BSNOx Curves for Analysis
35
30
25
~ub
X
o 15
CO
co
10
5
0
0 20 40 60 80 100 120 140
Engine Load {%)
FIGURE 60. BSNOx VERSUS LOAD FOR A MARINE TWO STROKE ENGINE (MAN
ENERGY SOLUTIONS, N.D.)
4.3 Marine NOx Control Technology Selections for NOx Estimates
As discussed earlier, there are a variety of approaches the have been taken in the marine
industry to meeting the current Tier III regulation. For low-speed, two-stroke diesel engines the
primary choice comes down to either EGR or SCR as the primary NOx reduction lever. As shown
Final Report 03.28987
58
-------
EPA Task Order 68HERC24F0432
earlier in Figure 53, it appears that SCR has been the primary NOx reduction lever with 65 to 75%
of Tier III OGV main engines utilizing this technology, although this portion does vary somewhat
by bore size. In addition, it appears that HP-SCR located upstream of the turbine has been the
primary choice for dealing with the relatively low exhaust temperatures associated with low-speed
two-stroke engines. When examining the potential for further emission reductions, it is useful to
make some technology selections as a "best available" approach, although that does not necessarily
mean that the selected strategy is the only approach that will work.
For the examination of available emissions reductions discussed later, the decision was
made to focus on the use of HP-SCR, along with thermal management options necessary to reach
the desired exhaust temperatures. The primary focus of this effort was to examine the possibility
of extending emission controls to loads below the current duty cycle floor of 25% maximum
power, although the potential some reduction in higher load emissions is also examined in the
analysis. For lower load emissions, it was felt that LP-SCR would likely present too much of a
thermal management challenge at loads below 20% without very large energy inputs. , EGR would
also be able to reach at least some of these targets, but would likely face more difficult technical
challenges with condensation risk at very low loads, as well as implementation challenges (and
potentially efficiency issues) with higher EGR rates than are currently used at high loads. Note
that this does not preclude the combination of EGR and SCR as a strategy, but this would likely
be very expensive, complicated, and difficult to package.
In order to reach the necessary temperatures to enable SCR operation, while avoiding ABS
condensation risk, current Tier III engines employ a variety of thermal management techniques at
lower engine loads. The primary approach to thermal management consists of air-fuel ratio
management through the use of either a cylinder bypass valve, a turbine bypass valve, or both.
Other combustion changes affecting injection timing and fueling are certainly also part of these
strategies, but the change in air-fuel ratio is the primary approach.
For the analysis of potentially available NOx reductions, SwRI has chosen to develop a set
of two different scenarios each for SCR efficiency and exhaust temperature, one aggressive and
one more conservative scenario. In general, the more conservative scenarios would represent the
potential of the current technology while extending the NOx controlled envelope down to 10%
maximum power while maintaining current performance levels. The more aggressive scenarios
involve more aggressive thermal management and would extend the range down to 5% of
maximum power, while also pushing SCR conversion to higher levels. The SCR conversion
scenarios are discussed in more detail below.
Regarding the engine and thermal management scenarios, the inputs used are shown below
in Figure 61 for exhaust temperatures and Figure 62 for engine-out NOx. Figure 61 shows
temperature curves versus load for a Best Projected scenario and for an Intermediate Projected
scenario, both in comparison to current Tier III behavior and Tier II temperatures. The current
Final Report 03.28987
59
-------
EPA Task Order 68HERC24F0432
Tier III behavior was used to establish the baseline tailpipe NOx levels which were the basis for
comparison for calculating emission reductions for the other scenarios. It should be noted that
ACD activation does not always occur just below 25%. In practice there is not a sudden switch
off of NOx control just below 25% because some amount of operating margin is needed to ensure
NOx control at 25%. Therefore, the Tier III curve still maintains thermal management and
therefore some NOx conversion down to 20% load. This is important to note because this Tier III
curve is the baseline for comparison of other scenarios, and thus sets the "baseline" cumulative
NOx mass for the projections over the histograms.
The Intermediate Projection represents what is likely to be achievable using the current
technologies installed on marine Tier III compliant engines, with very little modification. In
general, the intent behind the Intermediate Projection was to provide thermal management to
enable successful use of SCR down to the 10% load point. It should be noted that there are several
examples in the literature of temperatures at or greater than these levels on current technology
engines. As mentioned earlier, Hitachi-Zosen showed SCR inlet temperatures at 10% load in the
range of 280°C (Fujibayashi, T., et. al., 2013) in 2013, while WinGD described available engine
thermal management via a "Tier III mode" available down to 10% load at an exhaust temperature
of 310°C (Spahni & et.al., 2023). These data sets indicate the feasibility of reaching these
temperatures on real hardware with the current Tier III technology packages, questioning the need
for approval of ACDs for all operation below 25% load.
-•-Tier 1
yicn
Exhaust Temperature Scenarios versus Load
II -•-Intermediate Project -•-Best Project —•—Current Tier III
Temperature, degC
CJIOCJIOCJIOCJIOC
oooooooooc
r*"i
/
r
1
y
I
r
i i i
i i i
i i i
i i i
1 1 1
1 1 1
1 1
1 1 1
1 1
1 1
)0
i
)
i i i i i i i i i i i i i i i i i i i i
10 20 30 40 50 60 70 80 90 1(
Load, % Max Power
FIGURE 61. SCR REACTOR INLET TEMPERATURES FOR TECHNOLOGY
PROJECTIONS COMPARED TO TIER III CURRENT
Final Report 03.28987
60
-------
EPA Task Order 68HERC24F0432
35
30
25
5 20
•OD
| 15
2 10
5
0 i
(
Engine-Out NOX versus Load
-•-Tier II -*-TierlllEO
i i i
i i i
i i
i i
i i i
i i
i i
i i i
i i i
i i i
)0
i i i i i i i i i i i i i i i i i i i i i
) 10 20 30 40 50 60 70 80 90 1(
Load, % Max Power
FIGURE 62. ENGINE-OUT NOx CURVE FOR TECHNOLOGY PROJECTIONS
The exhaust temperature Best Case Projection is designed to push thermal management
upwards somewhat, allowing for higher conversion at low load points, but also to extend the range
of SCR operation down to 5% of Maximum Power. It is expected that much of this thermal
management can be accomplished using the previously discussed technologies, but in the
particular case of 5% load, some additional technology may be needed to provide the margin
necessary to reach the target temperature, though this heat addition is anticipated to be in the range
of an additional 20°C more than what is available from the current air fuel management and other
engine calibration parameters. From a technology standpoint, options for this would include
combustion phasing, early exhaust valve opening, a relatively small electrical heater or burner, or
potentially the process steam-based intake air heating technology described earlier.
4.4 Ammonium Bisulfate Deposit Formation Assessment
One of the primary factors limiting the use of SCR in lower temperature conditions for
marine SCR is risk of ammonium bisulfate formation (ABS), either in the exhaust upstream of the
SCR reactor, or within the reactor itself. Although small amounts of ABS that do form may be
removed under higher temperature conditions, especially in cases with a low ammonia-to-NOx
ratio (ANR), in general prolonged operation under ABS forming conditions will lead to fouling of
the SCR reactor, deposit growth on upstream piping systems, and corrosion. Given the high mass
rates of engine-out NOx and therefore NH3 injection observed in operations covered by the current
Tier III requirements, as well as elevated exhaust pressures, the approach to controlling this issue
Final Report 03.28987
61
-------
EPA Task Order 68HERC24F0432
is to maintain exhaust temperatures well above the ABS dewpoint, so as to prevent formation in
the first place. Under conditions typical of 2-stroke engines at 25% load and higher, this is
generally considered to be around 300°C. Accounting for some heat loss between the exhaust
receiver and the SCR this generally implies engine-out temperatures of 310°C, considering the use
of exhaust insulation and other design elements of current systems. As discussed earlier, this
generally requires the use of various engine thermal management techniques at loads anywhere
from 30% to 50%, depending on engine and aftertreatment design. The thermal management
requirements become significant around 20 to 25% load and below.
When considering the possibility of achieving NOx control using SCR below 25% load,
the ABS limitation becomes a critical potential limiting factor. Given the limits discussed above,
large amounts of thermal management could be required to reach 310°C at very low loads.
However, because exhaust conditions are quite different in these low load regions, it is important
to examine the behavior of ABS to look more closely at how much thermal management is really
needed. Although ABS formation has been studied extensively for two stroke marine diesel
engines, much of the literature has focused on higher sulfur fuels, well above the current ECA
limit of 0.10%) sulfur. In addition, given the propensity of engine manufacturers to pursue ACDs
to disengage the SCR system at low exhaust temperatures, the operating range below 25% load
has generally not been considered in most of these studies.
The formation of ABS in 2-stroke marine diesel exhaust is a complicated process whose
rate is governed by a combination of exhaust temperature, exhaust pressure, and the partial
pressures of the primary reactants which are SO3 and NH3. Given that the desire is to avoid ABS
formation, the critical point to understand is the condensation temperature of ABS, which in turn
dictates the "safe" temperature, above which SCR may be used. Muzio et al. summarized a variety
of studies of the kinetics of this process as indicated by the chart in Figure 63 (Muzio & et.al.,
2017). The authors examined a variety of studies, as well as conducted their own experimental
work. They concluded that the combined curve described by the earlier work of Menasha
(Menasha & et.al., 2011) and Wei (Wei, 2007) was the best representation of this environment,
and therefore this curve (shown on the chart using the light blue and pink data points) is used for
the current discussion. However, this curve is relevant primarily near atmospheric pressure.
At pressures above atmospheric pressure, ABS formation rates can increase significantly.
This is due to increased partial pressure of the reactants at elevated exhaust pressures. This was
described by Sandelin et.al. (Sandelin & et.al., 2016) at the 2016 CIMAC Congress, and an
example of this relationship is shown in Figure 64. This example was for 165 g/kW-hr fuel
consumption, 8 g/kW-hr NOx, ammonia at lOOOppm, and assumes 5-10% oxidation of SO2 to
SO3. The figure shows that elevated exhaust pressures can easily increase the condensation
temperature by 20-30°C at even moderate exhaust pressures. While this is a crucial consideration
at higher loads, recall that exhaust pressures below 20% load are generally in the range of 1.1 to
Final Report 03.28987
62
-------
EPA Task Order 68HERC24F0432
1.2 bar, even with scavenging blowers active, and therefore will not contribute significantly to the
condensation temperature under low load conditions, especially given the need to limit air fuel
ratio to effect thermal management.
g.
I 100
e
a.
ABS Condensation Temperature
as a Function of ABS Potential
"
i > t r
¦ X
¦ / /
/ r
/ /
/
•
: /
/
J
-/
'¦/
f/
I
400 450 500 550 600
ABS Condensation Temperature (F)
Menasha ABS T vs ABS Potential
¦ Wei ABS T vs ABS Potential
• Radian ABS T vs ABS Potential
Hitachi Zosen ABS T vs ABS Potential
~ Matsuda ABS T vs ABS Potential
FIGURE 63. ABS CONDENSATION AS A FUNCTION OF REACTANT
CONCENTRATION (MUZIO ET. AL)
380
360
340
ZJ 320
o
C
'o
300
280
-------
EPA Task Order 68HERC24F0432
The oxidation of exhaust SO2 to SO3, is an important reaction step in the formation of ABS,
because the reactions are driven by the availability of SO3 and other oxidized SO2 byproducts (at
lower temperatures it is more likely that H2SO4 is the reactant given the reaction with water in the
exhaust). This has been studied extensively in the literature, and SO3 ranges of 3% to 10% of total
SO2 have been reported, though the majority of the reported levels have been 5% and lower. An
assumption of 5% to 10% was used by Sandelin to generate temperature ranges above in Figure
64, and is documented in other literature. For purposes of examining ABS risk in this study, an
SO2 oxidation rate of 7% was utilized to provide a conservative estimate.
As noted earlier, low load operation below 25% of maximum engine power tends to be
characterized by higher air fuel ratios which will reduce the partial pressure of SO3, an important
driver of the ABS formation temperature. In addition, given the lower catalyst temperatures, full
NOx conversion rates in the range of 80 to 90 percent are not expected, and thus it is likely the
ammonia dosing rate will be lowered somewhat below an ammonia-to-NOx ratio of 1.
Finally, we can take into consideration that fuel sulfur levels within the ECA will be less
than 0.10%). According to marine fuel survey data, the median sulfur concentration for fuels under
this specification is 600 ppm (0.06%) (IMO-MEPC, 2024). ABS formation calculations were run
at both 600 ppm and 1000 ppm fuel sulfur. Some calculations were also run at 7 ppm fuel sulfur
to examine the potential impact of the use of ULSD on ABS condensation temperatures. An
ammonia to NOx ratio of 0.8 was used for these conditions, which is also factored into later SCR
conversion curves.
Given these parameters, and the operating parameters typical of low load operation below
25% of maximum power, it is possible to estimate the risk temperatures for lower load operation
in the range of 10% and 5% of maximum power, as compared to 25%. Assuming that thermal
management is accomplished using typical currently-applied methods of air-fuel ratio
management, such as cylinder bypass and/or turbine bypass methods, it is expected that air-fuel
ratio would have to be controlled in the range of 50-60 to achieve SCR temperatures in the range
of 260°C to 300°C. For performance projections, we are expecting NOx values under these light
load conditions in the range of 25-28 g/kW-hr. Considering these inputs and using the ABS
condensation rate kinetics in Muzio (Muzio & et.al., 2017), estimates are given for the "low load"
ABS condensation temperatures in the range of 5-10% of maximum power.
Final Report 03.28987
64
-------
EPA Task Order 68HERC24F0432
TABLE 7. ABS CONDENSATON TEMPERATURE ESTIMATES AT 5-10% LOAD
Fuel Sulfur,
ppmw
S02
Concentration,
ppm
nh3
Concentration,
ppm
ABS
Condensation
Temperature, °C
600 (ECA
median)
8-9
~ 1200
255 -260
1000 (ECA max
limit)
14-15
~ 1200
260 - 265
7 (ULSD
median)
-0.1
~ 1200
-225
These projections indicate that under very light load conditions (i.e., below 20%), and only
under light load conditions, it appears feasible to extend the range of ABS-limited SCR operation
to temperatures lower than the 300°C limit that is generally used at higher loads. Again, it must
be noted that this is not feasible at higher loads (i.e., 30% and higher) due to the combination of
higher pressures, lower air-fuel ratios, and higher ammonia dosing rates.
4.5 Selective Catalytic Reduction Performance Curves
As noted earlier, for the NOx projections, high-pressure SCR is the technology of focus.
For the Tier III marine diesel engine market, the primary technology of choice is vanadium-based
catalysts. Given the size and scale, as well as the need to be resistant to relatively high levels of
sulfur compared to many other applications, vanadium is the generally considered to be the best
overall choice for use in Tier III marine diesel engines. The SCR reactor arrangements often use
technologies similar in many ways to those used for stationary power generation, where SCR was
first applied. More specifically, the catalyst is a V2O5 material on a W03/Ti02 support. These
applications typically use an extruded catalyst rather than a washcoated substrate which is typical
of smaller applications. In addition, given the application a relatively low cell density substrate is
used, often on the order of 64 cpsi. This allows the catalyst assembly to have low backpressure,
while being less susceptible to fouling from soot, ash, and ABS exposure. The reactors are
typically assembled using multiple layers of individual catalyst blocks that are assembled to create
the required reactor cross sectional area for the application. These assemblies are usually designed
so the blocks can be removed and replaced if there is a need to service the reactor. Although design
guidance does vary, the reactors are typically sized for space-velocities (a global measure of
catalyst size relative to flow rate in catalyst volumes per hour) on the order of 10,000 1/hr to
15,000 1/hr. This sizing is changing somewhat as reactor designs continue to improve, but this
will be used for the current projections.
Although NH3 can be used directly as the reductant, it is more typical that urea-water solution
(UWS) similar to DEF and/or AdBlue that is used in land based applications will be used for ease
of storage and safe handling. The concentration of urea in the water in marine applications,
Final Report 03.28987
65
-------
EPA Task Order 68HERC24F0432
however, is typically 40% urea by mass, which is higher than the 32.5% used for land based
applications. The SCR system includes dosing and metering systems for the UWS, and
arrangements for mixing are necessary in the exhaust upstream of the reactor inlet to ensure the
uniformity of the reductant distribution. Although problems with deposit formation can be issues
in other applications, for an HP-SCR marine application dosing is generally not done below the
ABS formation temperature, which is typically much higher than the temperatures where UWS
related deposit formation can be an issue (typically at temperatures below 210°C). Nevertheless,
proper mixing design is important to prevent any such issues.
The active temperature range of these catalyst systems for peak conversion is generally in the
range from 300°C to about 425°C. It should be noted that these catalysts are generally capable of
NOx conversion efficiency in excess of 90% at these temperatures under the right conditions, but
in the case of Tier III certified marine diesel engines, they are generally controlled to a nominal
peak NOx conversion on the order of 80% or slightly higher. Parasitic ammonia oxidation on the
surface of the catalysts begins to compromise performance starting around 425°C to 450°C,
although initially this can be mitigated somewhat by increased dosing at temperatures below
500°C. At temperatures below 300°C, NOx conversion capability typically begins to drop,
although this does not become pronounced until the temperature drops below 250°C. The driving
force, however, that limits low temperature conversion in Tier III marine diesel engine applications
is the need to avoid ABS condensation, as described earlier.
Three conversion curves were generated in support of the NOx mass emission projections
in Section 5. These NOx conversion curves are expressed as NOx reduction efficiency as a
function of catalyst temperature (in this case SCR reactor inlet temperature). The SCR conversion
curves used for these projections are given in Figure 65.
Final Report 03.28987
66
-------
EPA Task Order 68HERC24F0432
c
o
a>
>
c
o
O
X
o
-~-Current Tier III
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
200 0% 250
-~-Conservative Reduction -*-Maximum Feasible Curve
300 350 400 450 500 550
Temperature, degC
FIGURE 65. SCR NOx CONVERSION CURVES USED FOR TIER III MARINE
DIESEL ENGINE NOx PROJECTIONS
The SCR conversion curve labeled as Current Tier III (black line) is intended to reflect a
current system calibrated to meet the Tier III requirements for a NOx ECA assuming the engine-
out NOx levels and temperatures curves shown previously. As noted earlier, peak conversion on
these catalysts is typically controlled to about 80%, which is sufficient to satisfy the Tier III
requirements, although it is common for the actual performance to be slightly better than 80%.
For the case of this baseline Tier III curve, a peak conversion of 82% was used to reach a tailpipe
or stack outlet NOx level of the 2.9 g/kW-hr for the 4-mode E3 cycle, resulting in a reasonable
compliance margin below the standard of 3.4 g/kW-hr. In addition, the individual modes are also
below the individual mode cap of 1.5 times the standard. The SCR does not operate below 25%
load as engine manufacturers routinely request an ACD with a relatively conservative minimum
temperature to avoid ABS formation of290°C at the reactor inlet is used as the cutoff, below which
no UWS will be dosed and thus no conversion will be achieved. At higher loads, the increased
exhaust pressure means that a higher minimum temperature of 310°C is used to prevent ABS
condensation, although this is typically only a concern during transient operation before the
catalyst is fully warmed up. In practice the need for some operation margins means that some
conversion is typically achieved down to 20% load, and sometimes at temperatures as low as
275°C for a short time.
Final Report 03.28987
67
-------
EPA Task Order 68HERC24F0432
The SCR conversion curve labeled Conservative Reduction (blue line) represents a system that
has been calibrated to continue dosing at temperatures below 290°C but only under light load
conditions. A minimum temperature of 260°C is utilized, representing the minimum temperature
that can be allowed to avoid ABS, specifically under very light load conditions. Note that for the
purposes of ABS control at higher loads, the minimum temperature of 310°C is retained, but again,
this is only significant during transients. It is important to note that the target conversion
performance is significantly reduced below the capability of the catalysts at higher loads, in part
to assist with the avoidance of ABS formation at these temperatures, resulting in UWS dosing rates
well below an ammonia-to-NOx ratio of 1. This curve represents an intermediate approach that
retains the high load conversion targets of previous systems, and instead focuses only on extending
the low temperature conversion range, but in a conservative fashion.
The SCR conversion curve labeled Maximum Feasible Curve represents what we believe to be
a more aggressive, yet still achievable performance level given the current technology. It focuses
on both a further improvement step in low temperature performance, and an increase in high load
conversion to better reflect the actual capabilities of the catalyst under reasonable UWS dosing
conditions. For high load, performance can be increased by increasing the UWS dosing rate to
reach an ANR closer to 1. However, this is still limited by the tendency for the generation of
increasing amounts of ammonia slip at higher space velocity conditions. An example of this
relationship, taken from Sandelin (Sandelin & et.al., 2016), is given below in Figure 66 for a
condition of 350°C at 1.3 atmospheres. For the Maximum Conversion curve, the NFb slip limit
was set at lOppm. This effectively limits peak conversion at the upper end to a range of 85 to 90%
depending on the temperature.
1,00
X
z
lOppm stack slip limit
ANR ~ 0.92
Ref.: Pape
No: 111 CIM
\C 2016
0.7 O.S
FIGURE 66. AMMONIA SLIP CURVE EXAMPLE VERSUS
STOICHIOMETRIC AMMONIA-TO-NOX RATIO
Final Report 03.28987
68
-------
EPA Task Order 68HERC24F0432
For the Maximum Conversion curve, the minimum ANR was increased somewhat to bring
NOx conversion closer to the capability of the catalysts at these temperatures, while still
representing a reduced rate to help manage ABS formation. The minimum dosing temperature of
260°C at the reactor inlet is retained for light load operation, but given the higher ANR this is only
feasible at the lowest loads when operating only on ECA fuel with less than 0.10% sulfur content.
Even at 10% load, a higher minimum temperature would likely need to be used, scaling upwards
with drop in load. Careful coordination of thermal management and dosing rates would be needed
to realize reasonable ABS avoidance under these conditions. It is possible that there will be some
ABS formation in catalyst pores at the lightest loads at and around 5% load depending on
conditions. However, overall the very low NOx mass rates would likely result in very low actual
ABS condensation rates that should be manageable as long as there is some operation in the ECA
at higher loads (above 25%).
Other scenarios are possible as well, such as the combination of an SCR conversion curve
that increases high load conversion like the Maximum curve, while retaining the more conservative
low load behavior of the Conservative curve. However, these were not examined as part of this
analysis.
Final Report 03.28987
69
-------
EPA Task Order 68HERC24F0432
5.0 ASSESSMENT OF POTENTIAL FOR ADDITIONAL NOx REDUCTIONS IN
EMISSION CONTROL AREAS
The ultimate objective of this study is to estimate the NOx reduction potential of applying
various technology modifications to Tier III compliant marine diesel engine designs that are
equipped with SCR. Using information contained in the histograms from Section 2.6 showing the
load duty cycle data from vessels in the fleet as derived from AIS position data, the technology
selections discussed in Section 4.3, and the projected SCR performance curves in Section 4.5, it is
possible to estimate potential NOx reductions from Category 3 engines on a mass basis for various
technology scenarios. These projections then can be used to examine different changes to the
standards, and estimate the NOx reduction potential available from those scenarios. These tools
can be used to examine the impact of changes using any engine load histogram as an input, and
the technology assumptions can be adjusted to examine other technology scenarios.
This analysis is focused on the engine load histogram for the fleet of ships with Tier III
main propulsion engines that operate in the EC As, as presented in Section 2. While scenarios were
also generated by individual vessel type, the following discussion focuses on the ships included in
the complete Tier III ECA data set.
The engine load histogram that is used as the input for this analysis is repeated below in
Figure 67. This figure shows the total frequency counts in each of the designated load bins, along
with the cumulative distribution for each bin as a percentage of the total operation data set. It
should be noted that the data indicates that 43% of all operation in the ECA is below the 25% load
point, indicating that a significant portion of operation is below the emission control window of
the current Tier III certification requirements. Even assuming that emission controls are still
operational down to 20% load as discussed earlier, this still results in 35% of all operation not
being considered when assessing compliance with the Tier III NOx limits.
Final Report 03.28987
70
-------
EPA Task Order 68HERC24F0432
Tier III ECA Load Distribution - All Vessels.
50000
45000
40000
35000
(A 30000
§ 25000
O
° 20000
15000
10000
5000
0
LO O LO O .
IV 00 00 0) lO
0)
% Power
FIGURE 67. TIER III ECA LOAD DISTRIBUTION - ALL VESSELS
A comparison of the actual load histogram for the North American ECA to the weight
factors that are currently assigned in the IMO E3 test cycle is shown below in Table 8. Note that
the operational weighting factors are calculated as the sum of operation between a given mode and
the next lower mode, with the 25% mode encompassing operations from 0-25%. From the table,
it is clear that the IMO weighting factors are far from accurately representing actual operation in
the ECA.
TABLE 8. ACTUAL OPERATIONAL DISTRIBUTION IN ECA VERSUS E2/E3 TEST
CYCLE WEIGHTING
IMO Mode
Number
% Power
IMO E2/E3 Weight
Factor
Actual Operational
Weighting in ECA
1
25
0.15
0.43
2
50
0.15
0.32
3
75
0.50
0.21
4
100
0.20
0.05
The basic methodology we used to estimate NOx emissions was to combine the
temperature data and SCR conversion curve along with Tier III engine-out NOx data to generate
a tailpipe brake-specific NOx level for each load bin in the histogram. That brake-specific NOx
level is then multiplied by the duty cycle total counts in each load bin and by the average
percentage of maximum power for that bin, producing a load weighted NOx "mass rate," in units
of mass counts per kilowatt of engine power. Although this could be translated to an actual NOx
mass rate by considering the time associated with each count and the individual vessel power for
Final Report 03.28987
71
-------
EPA Task Order 68HERC24F0432
each data point, this additional transformation was not necessary because the object of this study
was not to estimate a NOx inventory. Rather, the objective of the study was to provide a relative
assessment of NOx reduction potential for certain technologies when applied to 2-stroke
Category 3 engines.
To allow for relative comparisons, a baseline scenario for NOx performance was needed.
Using the ECA duty cycle histogram in Figure 67 and the Tier III NOx baseline performance
curves established for the technology scenarios, a baseline NOx mass scenario representing the
current Tier III fleet was generated. This baseline scenario is shown in Figure 68. This data shows
the relative amount of NOx mass emitted during all operations recorded in the ECA for each load
bin. Table 9 also shows this baseline in terms of the exhaust NOx results over the current E3 cycle.
The composite value of 2.7 g/kW-hr indicates compliance with the 3.4 g/kW-hr IMO Tier III
standard with an expected 20% compliance margin.
Tier III ECA Exhaust Stack NOX Distribution
100%
90%
80% o
z
70% jo
£
60% ~
50% |
a.
40% |
30% *3
E
20% =
10%
0%
% Power
FIGURE 68. TIER III BASELINE NOx DISTRIBUTION FOR THE ECA
TABLE 9. IMO E3 TEST CYCLE RESULT FOR BASELINE TIER III SCENARIO
IMO Mode
Exhaust NOx,
Number
Percent Load
g/kW-hr
E3 Weighting Factor
1
25%
4.2
0.15
2
50%
3.3
0.15
3
75%
2.6
0.50
4
100%
2.3
0.20
Composite
n/a
2.7
n/a
I
i—
®
Q-
V)
O
O
V)
V)
ra
Z
x
O
200000
180000
160000
140000
120000
100000
80000
60000
40000
20000
0
O lO O lO
00 00 O) 0)
o
o
O LO O LO
CN CN 00 00
LO LO CD CD
LO O LO O I
rv oo oo O) LO
o
Final Report 03.28987
72
-------
EPA Task Order 68HERC24F0432
The data indicates that 43% of all NOx mass emitted in the ECA occurs in the load bins
below 20% maximum power. In particular, the load bins between 5%-10%, 10%-15%, and 15%-
20% contribute the majority of that NOx. Although the relative power at these lower bins is low,
the high brake-specific NOx combined with the large amount of operation in those bins results in
a significant contribution to the total NOx mass emissions. The contribution to NOx mass below
5% maximum power is very small and can be ignored. There is also still a considerable amount
of the total NOx mass emitted in the higher bins, indicating potential for further reduction by
applying the technologies discussed above.
Figure 69 shows the contribution of operation in the various load bins to overall total NOx
mass, in comparison to the amount of time spent in each load bin as a percentage of total operation.
These results indicate that the load bins between 5% and 20% have a NOx mass emission
contribution that exceeds their relative amount of operation time, while bins below 5% are not
significant. In addition, bins in the 35% to 65% range still contribute significant NOx mass,
although those values are more proportional to their relative amount of operation time.
Tier III ECA NOX Dist versus Pwr Dist
NOX ¦ Power
20%
18%
"cu
4—»
£
16%
0)
>
14%
re
12%
3
E
3
10%
o
8%
o
*->
£
6%
0)
o
1—
4%
0)
Q_
2%
0%
16%
12%
10%
|9%|
5% 5%
A0/„
B%
¦70/ 7%
7% 6% 6%
4°,
4°,
Lfi Lfi O Lfi
O CM Lfi
- i •
tH CM
Lfi O
OLfiOLfiOLfiOLfiOLfiO
¦ i i i i i i i i i i
LfiOLfiOLfiOLfiOLfiOLfi
O LO O LO O
00 00 Ol O) o
1 1 1 1 —I
Lfi O Lfi O V
rv oo oo o) Lfi
G)
FIGURE 69. COMPARISON OF POWER DISTRIBUTION VERSUS NOx
DISTRIBUTION FOR TIER III VESSELS IN THE ECA
One potential countermeasure to address the disproportionately high NOx emissions at low
load is by re-weighting the current test cycle modes for engines certifying to IMO Tier III NOx
standards, to more accurately reflect the actual duty cycle of engine operation in ECAs. While
this would not address the low load emissions, it is still useful to examine what could be gained
from this relatively simple change. Using the current Tier III NOx baseline performance shown
Final Report 03.28987 73
-------
EPA Task Order 68HERC24F0432
in Figure 68, a revised set of weighting factors would produce the results indicate in Table 10
below. Using revised weighting factors, the composite NOx level rises from 2.9 g/kW-hr to
3.2 g/kW-hr, which would still pass the current standard, but with less compliance margin. To
achieve a sufficient compliance margin, SCR efficiency would have to be increased on some of
the test cycle modes. The current peak efficiency in these modes under the current Tier III scenario
is 82%. Increasing this to 85% efficiency, apart from at 100% load which could be problematic
on some engines, would again achieve 2.7 g/kW-hr NOx, which would meet the NOx standard,
assuming a 20% compliance margin.
TABLE 10. NOx RESULT FOR TIER III BASELINE USING ACTUAL TIER III ECA
DUTY CYCLE WEIGHTING, ASSUMING 82% PEAK EFFICIENCY
IMO Mode
Number
Percent Load
Tailpipe NOx,
g/kW-hr
Revised
Weighting Factor
E3 Weighting
Factor
1
25%
4.2
0.43
0.15
2
50%
3.3
0.32
0.15
3
75%
2.6
0.21
0.50
4
100%
2.3
0.05
0.2
Composite
n/a
3.2
n/a
n/a
The impact of this change on the NOx distribution is show in Figure 70 below. As can be
seen, this adjustment does not address the low load NOx mass emissions, and ultimately results in
only a modest 9% reduction in total NOx mass emitted. It is clear that other changes would be
needed to achieve a more meaningful reduction in total NOx mass.
ECA Reweighting Only Scenario
100%
90% x
O
80% ^
Co
4->
70% ji
60% S
i—
50% =
flj
o
40% aj
Q_
CD
30% >
cn
20% |
10% O
0%
% Power
FIGURE 70. NOx MASS DISTRIBUTION FOR ECA MODE REWEIGHTING
SCENARIO AND CURRENT STANDARD
200000
180000
160000
140000
120000
100000
80000
60000
40000
20000
0
Final Report 03.28987
74
-------
EPA Task Order 68HERC24F0432
To examine if more significant emission improvements are available, we examined the
impact of the various technology scenarios on the overall NOx histogram. It is also possible to
examine different NOx standards that those scenarios could support. Many of these scenarios
involve extending NOx control to the load range below 25%, to address the considerable amount
of NOx mass emitted in the load bins between 5% and 20%. Recall, the assumption that the current
NOx standard at 25% still results in at least some emission control in the 20-25% range due to the
need for operating margin. In addition, some of the scenarios described below also examine
improvements to SCR performance at loads above 25%.
Figure 71 shows the NOx mass distribution for what is considered a "Best Case" scenario.
This scenario combined both the Best Case temperature curve, which results in NOx control down
to 5% load, with the Maximum Feasible SCR performance curve. The results indicate a total
reduction of 53% in NOx mass emitted in the ECA if these technology measures were
implemented. BSNOx levels and NOx conversion by load bin for this scenario are given in Figure
72. This scenario involves leveraging existing thermal management technologies available on
marine diesel engines, pushes high load SCR conversion to what is considered to be the best
performance feasible for the current vanadium SCR catalysts, and requires an additional thermal
management technology to extend emission control down into the 5-10% load bin.
Best Case ECA Exh Stack NOX Distribution
iTotal NOX Mass Counts per Kw
¦cumulative % Tier I
200000
180000
160000
140000
100%
90%
80%
70%
re
4-»
.O
120000
100000
80000
60000
60% .22
o46°/46°/47°/47%
O LO O LO
CM CM CO CO
LO o lq o r
IV oo oo 0) LO
- - - - o>
CD
O
i—
E
2
o
% Power
FIGURE 71. NOx MASS DISTRIBUTION FOR BEST CASE TECHNOLOGY
SCENARIO
Final Report 03.28987
75
-------
EPA Task Order 68HERC24F0432
Best Case Technology Scenario BSNOx versus Load
^^Engine-Out ^^Tailpipe ^^NOX Conversion
28
82%"
28
25
24
O* 15
Z
GO 10
29.2 \
/ 75%^*-
c
o
O
x
O
z
0£
O
CO
% Power
FIGURE 72. BSNOx AND SCR NOx CONVERSION BY LOAD BIN FOR BEST CASE
TECHNOLOGY SCENARIO
A comparison of NOx histograms between the Tier III Baseline scenario and the Best Case
technology scenario is shown in Figure 73. Although NOx reductions are evident in most bins,
and there is a small spillover in the 2.5-5% bin due to operating margin, the majority of NOx mass
reduction, on the order of 60% of the total reduction, occurs in the load bins between 5% and 20%
of maximum power. In the higher load regions, a reduction on the order of 20% is observed,
although this is somewhat lower at the highest loads due to limits associated with ammonia
oxidation. The contribution at high load is still responsible, however, for one-third of the total
NOx mass reduction, and therefore improving NOx reduction at high loads is still a significant
change to consider. Recall that this higher load change involves pushing high load conversion in
the range from 300°C to 425°C from 82% up to 88%.
Table 11 shows the cycle weighted result for the E3 test cycle modes using both the IMO
weighting factors and the ECA duty cycle weighting factors. The table provides potential
composite NOx standards that this best case scenario could support, assuming a 20% NOx
compliance margin to account for production variance. These test cycle weighting factor
adjustments would capture the benefits of the Best Case scenario in the higher load region.
Additional changes would still be needed to realize the large gains achievable at low loads. This
can be done by extending the E3 test cycle to include lower load mode points and mode caps below
25%, which would prevent the use of ACDs in the low load operating range.
Final Report 03.28987
76
-------
EPA Task Order 68HERC24F0432
Comparison of Best Case and Base Tier III NOx Mass Distributions
l Tier I
I Projection NOx Mass Counts
-% Total NOx Mass Reduction
200000
120%
180000
g 160000
% 140000
Q.
£ 120000
c
O 100000
80000
-------
EPA Task Order 68HERC24F0432
the basis of the ECA duty cycle weighting for the 5 modes used in the composite calculation.
Under this scenario, the test cycle and compliance with the standard for operation outside the ECA
could remain at the current Tier II level utilizing the IMO 4-mode E3 test cycle, since the objective
is to capture some of the emissions during Tier III operation that are occurring due to lack of SCR
control at low load.
TABLE 12. CYCLE RESULTS ON BEST CASE TECHNOLOGY SCENARIO FOR
UPDATED TEST CYCLE WITH 10% MODE AND UPDATED DUTY CYCLE
WEIGHTING FOR TIER III
Mode
Tailpipe NOx,
Updated Tier III
Number
Percent Load
g/kW-hr
Weighting Factor
Mode Cap
n/a
5%
6.8
n/a
3.2
1
10%
4.9
0.22
2.3
2
25%
2.7
0.20
1.5
3
50%
2.2
0.32
1.5
4
75%
1.4
0.21
1.5
5
100%
1.6
0.05
1.5
Composite, g/kW-hr
2.2
Potential standard with 20% Margin
2.6
As noted earlier, the extension of the NOx curve to 5% load will likely involve the addition
of thermal management technology that is not currently included on most Tier III compliant marine
diesel engines. Therefore, it is reasonable to examine a less aggressive scenario in which the NOx
curve is extended down to 10% load rather than 5%. This was examined using the Intermediate
Temperature curve and the Maximum Feasible NOx curve. The NOx histogram for that scenario
is shown in Figure 74. Figure 75 shows BSNOx levels and NOx conversion by load bin for this
scenario. This scenario results in a 45% reduction in NOx mass, with the primary changes being
the loss of NOx reduction on the 2.5-5% and 5-10% load bins (note some NOx reduction still
occurs at 5-10% load due to the need for operating margin near 10% load). This scenario involves
applying less thermal management at both 5% and 10% loads, resulting in lower temperatures and
somewhat less NOx performance at the 10% mode point, as well as only a small amount of
conversion below that. NOx reductions for this scenario are shown in Figure 76.
This scenario still includes increasing NOx conversion at temperatures in the range
between 300°C to 425°C from 82% to 88%. In addition, it still requires increased thermal
management utilization at low loads compared to the current Tier III, but to a lesser extent and
only actively down to a little below the 10% load point for operating margin. As noted earlier,
there are several examples in the literature of current hardware on vessels that can achieve the
temperatures required to support conversion down to 10% load without additional hardware.
Table 13 shows potential changes to the standards that would realize these reductions.
These changes would extend the range of emission control down to the 10% load point. This
scenario is similar to the previous scenario shown in Table 12 but without the inclusion of the 5%
load point that provides NOx emission control via a mode cap. Instead, the cycle and caps are
only down to 10%. Functionally, the only cycle change is that performance on the 10% load point
Final Report 03.28987
78
-------
EPA Task Order 68HERC24F0432
is not quite as good, and as a result, the cycle composite NOx result is higher. This assumption
would result in a slightly higher NOx limit, and also requires a mode cap at 3 times the emission
standard for the 10% load point, while leaving the others at 1.5.
Best SCR with Int Temp Case ECA Exh Stack NOX Distribution
200000
100%
180000
160000
Z. 140000
'+-»
ro
E
=
o
) Power
FIGURE 74. NOx MASS DISTRIBUTION FOR MAXIMUM FEASIBLE NOx CURVE,
INTERMEDIATE TEMPERATURE CURVE, TARGETING CONVERSION DOWN TO
10% LOAD SCENARIO
Max SCR with Intermediate Temp Scenario BSN0X versus Load
o 15
m 10
¦ Engine-Out
-Tailpipe
-NOX Conversion
28
28
27
82% .
28.<
32%
72%
L7.5
0%
24 23 22
* 21 21
20 lg
18 17
15 14 14 13
— ¦ ¦ 13 13
2.8 2.7 2.7 2.6 2.5 2.3 2.2 2.1 1.9 1.7 1.5 j 4 1.5 1.6 1.6 1.6
100%
90%
80%
70%
60%
50%
40%
m o
CN CO
mom
o m
m m
m o
m
o m
m o
op ar>
o m
CO CO
1 Power
FIGURE 75. BSNOx AND SCR NOx CONVERSION BY LOAD BIN FOR MAXIMUM
FEASIBLE NOx CURVE, INTERMEDIATE TEMPERATURE CURVE, TARGETING
CONVERSION DOWN TO 10% LOAD SCENARIO
Final Report 03.28987
79
-------
EPA Task Order 68HERC24F0432
Comparison of Best SCR Int Temp Case and Base Tier I
Distributions
I NOxMass
¦ Tier I
200000
180000
a 160000
£ 140000
Q.
J2 120000
c
O 100000
8 80000
en
x 60000
O
z 40000
20000
¦ Projection NOx Mass Counts
b Total NOx Mass Reduction
120%
irilklhlhbi
mmomomom
rH rH CN m
O)
i Power
FIGURE 76. COMPARISON OF NOx MASS DISTRIBUTIONS FOR MAXIMUM
FEASIBLE NOx CURVE, INTERMEDIATE TEMPERATURE CURVE, TARGETING
CONVERSION DOWN TO 10% LOAD SCENARIO AND TIER III BASELINE
TABLE 13. CYCLE RESULTS ON MAXIMUM FEASIBLE SCR CURVE,
INTERMEDIATE TEMPERATURE CURVE, TARGETING CONVERSION DOWN TO
10% LOAD FOR TEST CYCLE WITH 10% MODE AND UPDATED DUTY CYCLE
WEIGHTING FOR TIER III
Mode
Number
Percent Load
Tailpipe NOx,
g/kW-hr
Updated Tier III
Weighting
Factor
Mode Cap
1
10%
7.5
0.22
3
2
25%
2.7
0.20
1.5
3
50%
2.2
0.32
1.5
4
75%
1.4
0.21
1.5
5
100%
1.6
0.05
1.5
Composite, g/kW-hr
2.5
Proposed standard with 20% Margin
3.0
Another possible scenario is to focus only on improvements in the low temperature region,
while leaving the current performance levels intact at higher loads. We examined this case using
the Intermediate Temperature Curve and the Conservative SCR Conversion Curve. This would
limit the low temperature improvements at the 10% load point like the previous scenario, but in
addition would not push the high load conversion. The resulting NOx distribution associated with
Final Report 03.28987
80
-------
EPA Task Order 68HERC24F0432
this more conservative scenario is shown in Figure 77 below, and the BSNOx and NOx conversion
are given in Figure 78. A comparison with the Tier III baseline is shown in Figure 79.
Most Conservative Improvement Case ECA Exh Stack NOX
Distribution
200000
180000
100%
90%
X
o
E
3
o
OLOOLOOLOOLOOLO
cNcsjcoco^j-^j-ununcDcD
O if) © if)
00 00 o> o>
O if) © if) © if)
rl rl M M O O
O if) © if) © if)
•*t •st in m
) Power
o m
O if) © V
CO CO O) lf)
- - - O)
FIGURE 77. NOx MASS DISTRIBUTION FOR MORE CONSERVATIVE (LOW-TEMP
ONLY) NOx CURVE, INTERMEDIATE TEMPERATURE CURVE, TARGETTING
CONVERSION DOWN TO 10% LOAD SCENARIO
Conservative SCR with Intermediate Temp Scenario BSN0X versus Load
^^Engine-Out —^—Tailpipe —^NOX Conversion
100%
82%82%82%82%82%82%82%82%82%82%82%82%82%82%82%82% 90%
80% c
70%
60% >
50%
40%
30%
20%
10%
0%
o
O
cc
o
CO
O Lf) O i
CO CO O) in
G)
% Power
FIGURE 78. BSNOx AND SCR NOx CONVERSION BY LOAD BIN FOR MAXIMUM
CONSERVATIVE NOx CURVE, INTERMEDIATE TEMPERATURE CURVE,
TARGETING CONVERSION DOWN TO 10% LOAD SCENARIO
Final Report 03.28987
81
-------
EPA Task Order 68HERC24F0432
Comparison of Conservative Improvement Case and Base Tier I
NOx Mass Distributions
¦ Tier I
I Projection NOx Mass Counts
•% Total NOx Mass Reduction
200000
180000
g 160000
£ 140000
Q.
£ 120000
c
O 100000
£ 80000
q 60000 12%
z 40000
20000
100M0M0M0M0M)0M0M)0M0M)0M0M)0M0M)0M0M)0%
L 100%
46%
llllllllllli
120%
100%
80%
60%
40%
20%
0%
m
co
m o io o iT)
Tt m m cd cd
omomomomom
tHCNcgcoco^t^tmmcDcD
% Power
m o
00
¦ i
o m
m
o
oo co o) m
a
FIGURE 79. COMPARISON OF NOx MASS DISTRIBUTIONS FOR CONSERVATIVE
(LOW-TEMP ONLY) NOx CURVE, INTERMEDIATE TEMPERATURE CURVE,
TARGETING CONVERSION DOWN TO 10% LOAD SCENARIO AND TIER III
BASELINE
This scenario results in an overall 23% reduction in NOx mass, 88% of which occurs in the
10-20% load range, and a small amount below 10% due to operating margin. This is about half of
the NOx mass reduction observed for the previous scenario that had a similar extension to cover
low load operation but also included higher load reductions.
Table 14 shows a set of results for a 5-mode regulatory scenario with ECA duty cycle
weighting for this scenario. The higher resulting standard does somewhat reduce the multiplier
needed for the 10% mode cap to 2.6, but the NOx limit would be 50 percent higher than the
previous scenario.
Final Report 03.28987
82
-------
EPA Task Order 68HERC24F0432
TABLE 14. CYCLE RESULTS ON CONSERVATIVE (LOW-TEMP ONLY)
SCENARIO TARGETING CONVERSION DOWN TO 10% LOAD FOR TEST CYCLE
WITH 10% MODE AND UPDATED DUTY CYCLE WEIGHTING FOR TIER III
Mode
Number
Percent Load
Tailpipe NOx,
g/kW-hr
Updated Tier III
Weighting
Factor
Mode Cap
1
10%
9.4
0.22
2.6
2
25%
4.1
0.20
1.5
3
50%
3.4
0.32
1.5
4
75%
2.6
0.21
1.5
5
100%
2.3
0.05
1.5
Composite, g/kW-hr
3.7
Proposed Standard with 20% Margin
4.4
Figure 80 shows a scenario which involves maximizing low temperature reductions while
leaving higher temperature performance at 82% conversion where it currently stands. This
scenario involves combining the more conservative SCR reduction curve with the best case
temperature curve, pushing NOx conversion down to 5% of maximum power. BSNOx and NOx
conversion for this scenario are given in Figure 81. As can be seen, pushing the range of NOx
conversion down to 5% from the previous scenario (at 10%) without any change in high load
performance only results in a small added improvement of an additional 7% reduction in NOx
compared to the Tier III baseline scenario, to a total of 30% reduction for this scenario. Figure 82
shows that all of the improvements occur only at low loads as expected, with the additional gains
coming from improvements in the bin from 5-20% maximum power.
Final Report 03.28987
83
-------
EPA Task Order 68HERC24F0432
Cons SCR with Best Temp Case ECA Exh Stack NOX Distribution
200000
100%
180000
90%
160000
CD
Q.
140000
a 120000
c
I 100000
80000
* 60000
Z 40000
20000
0
7%
4%
0%2?
U) ifl ID 0
CO CO O) (D
tH rl CM CM CO CO
O
o
> Power
FIGURE 80. NOx MASS DISTRIBUTION FOR CONSERVATIVE SCR WITH BEST
CASE TEMPERATURE DISTRIBUTION, TARGETING CONVERSION DOWN TO 5%
LOAD CONVERSION ONLY
100%
90%
80% |
70% S2
60% >
50% o
40% o
30% |
20% %
10%
0%
Conservative SCR with Best Temperature Scenario BSNOx versus Load
^^Engine-Out —^Tailpipe ^^NOX Conversion
29
\ 29 28 28 82% 82% 82% 82% 82% 82% 82% 82% 82% 82% 82% 82% 82% 82% 82% 82% 82%
T—1
in
m
o
m
o
m
1
o
CM
m
T—1
1
t-H
CM
i
CM
i
i
T—1
CM
jn
o
t-H
m
t-H
o
CM
o
m
o
m
o
m
o
CO
CO
m
in
CD
in
o
in
o
in
o
in
CM
CO
CO
m
m
% Power
m
O
m
o
m
o
m
CD
00
CO
o
Op
O
in
o
in
o
in
o
CD
CD
CO
00
C)
14 13 13
13
L4 4.3 4.1 4.0 3.8 3.7 3.5 3.3 3.2 3.0 2.8 2.7 2.6 2.5
2.4 2.3 2.3
FIGURE 81. BSNOx AND SCR NOx CONVERSION BY LOAD BIN FOR
CONSERVATIVE SCR WITH BEST CASE TEMPERATURE CURVE SCENARIO,
TARGETING CONVERSION DOWN TO 5%
Final Report 03.28987
84
-------
EPA Task Order 68HERC24F0432
Conservative SCR Best Temp Case and Base Tier III NOx Mass
Distributions
¦ Tier I
I Projection NOx Mass Counts
-%Total NOx Mass Reduction
200000
180000
^ 160000
5 140000
a.
| 120000
I 100000
$ 80000
CB
^ 60000
i 40000
20000
26%
120%
100M0affl0M0affl0M0M0M0M0affl0M0M>0M>0M>0M)0M>0%
100%
100%
56%
llllllllll
80%
60%
40%
20%
0%
O iO O iO
CN CN CO CO
OlOOlOOlOOlO
Tt^tioio
O iO O iO
CN CN CO CO
OLOOLOOLOOLO
^^LOLOCDCDr^r^
o io o i
00 00 O) lO
O)
% Power
FIGURE 82. COMPARISON OF NOx MASS DISTRIBUTIONS FOR CONSERVATIVE
SCR WITH BEST CASE TEMPERATURE CURVE SCENARIO, TARGETING
CONVERSION DOWN TO 5%
Table 15 shows changes to the standard to implement this scenario. A similar set of
structural changes to the Best Case projection is used for this scenario, with a 10% mode point
added to the test cycle, and an additional mode cap added for emissions at the 5% load point.
However, without any high load reductions, calculated composite emissions for the updated cycle
would be somewhat higher than the current Tier III standard, due to the re-weighting of the cycle.
How, this still results in a net 30% reduction in NOx mass emitted.
TABLE 15. CYCLE RESULTS ON CONSERVATIVE SCR WITH BEST CASE
TEMPERATURES SCENARIO FOR UPDATED TEST CYCLE WITH 10% MODE AND
UPDATED DUTY CYCLE WEIGHTING FOR TIER III
Mode
Number
Percent Load
Tailpipe NOx,
g/kW-hr
Updated Tier III
Weighting
Factor
Mode Cap
n/a
5%
8.7
n/a
2.6
1
10%
6.6
0.22
2.0
2
25%
4.1
0.20
1.5
3
50%
3.3
0.32
1.5
4
75%
2.6
0.21
1.5
5
100%
2.3
0.05
1.5
Composite, g/kW-hr
3.3
Potential Standard with 20% Margin
4.0
Final Report 03.28987
85
-------
EPA Task Order 68HERC24F0432
6.0 SUMMARY AND CONCLUSIONS OF NOx IMPROVEMENT TECHNOLOGY
SCENARIO ANALYSIS
A summary of the various technology scenarios is given in Table 16 below. The scenarios
are presented in the order they were examined. The table the total reduction in NOx mass proj ected
for each scenario, along with details of each scenario. These include the minimum load point for
active NOx reduction by SCR. As noted earlier, each scenario projects some NOx reduction in the
next lowest power bin as a result of the need for operating margin, so the scenarios that include
10% minimum load assume some NOx conversion in the 5-10% load bin, and scenarios that
include 5% minimum load assume some NOx conversion in the 2.5-5% load bin. Any scenarios
that include the 10% mode to the duty cycle will require reweighting to account for this new mode,
and these weight factors are based on the weighting for the Tier III ECA power histogram.
TABLE 16. SUMMARY OF NOx IMPROVEMENT TECHNOLOGY SCENARIOS
EXAMINED
Scenario
Details
Minimum Load
NOx Reduction
Standard Notes
1
Current Mode Reweights
25%
9%
Reweigh to ECA Duty Cycle
2
Max SCR, BestTemperatures
5%
53%
Add 10% Mode and 5% Mode Cap, Tighten
Standard Limits
3
Max SCR, Intermediate
Temperatures
10%
45%
Add 10% Mode, Tighten Standard Limits
4
Conservative SCR, Intermediate
Temperatures
10%
23%
Add 10% Mode
5
Conservative SCR, Best
Temperatures
5%
30%
Add 10% Mode and 5% Mode Cap
Scenario 1 is included for comparisons, but generally indicates the simple measure of re-
weighting the test cycle modes to more accurately reflect ECA operation will achieve only modest
NOx reductions without additional changes. All of the remaining scenarios involve extending the
range of NOx control for the test cycle down to at least the 10% load point, which would likely
require adding a new test cycle mode at 10% (it may be possible to achieve the same result using
carefully designed mode caps). As noted earlier, existing literature indicates that this is likely
achievable using technologies already deployed on many Tier III OGVs. It should be noted the
scenarios 2 and 5, which extend NOx control down to the 5% load point will likely require some
additional thermal management technology, such as intake air heating or variable valve actuation,
to achieve. Technology scenarios marked as "Max SCR", include reductions in the NOx standard
that would require improving high load SCR NOx conversion efficiency to 88% from the current
82%.
Final Report 03.28987
86
-------
EPA Task Order 68HERC24F0432
7.0 REFERENCES
Blenkey, N. (2022, May 5). MAN marks another IMO Tier III milestone. Retrieved from Marine
Log: https://www.marinelog.com/news/man-marks-another-imo-tier-iii-milestone/
Clarksons. (2024, 10 5). World Fleet Register. Retrieved from https://www.clarksons.net/wfr/
Freeman, H. A. (1963). Introduction to Statistical Inference. Addison-Wesley Publishing
Company.
Fritz, S., & Riley, M. (2024). Exhaust Temperature Boost for EMD 2-stroke Engines to Make
Tier 4 Feasible. DI2S Conference. Graz: Duret Innovative 2-Stroke. Retrieved from
https://drive.google.com/drive/folders/lZTQPG-bZZUT7HNBMqT3NqlItJd6AGBuj
Fujibayashi, T., Baba, S., & Tanaka, H. (2013). Development of Marine SCR System for Large
Two-stroke Diesel Engines Complying with IMO NOx Tier III. CIMAC Paper 029.
Shanghai.
Fujibayashi, T., et. al. (2013). Development of Marine SCR System for Large Two-stroke Diesel
Engines Complying with IMO NOx Tier III. Paper 029. Shanghai, China: CIMAC.
Geopandas Developers. (2024, 10 5). Introduction to Geopandas. Retrieved from
https://geopandas.org/en/stable/getting_started/introduction.html
IMO-MEPC. (2024). Information related to the global 0.50% sulphur limit (IMO 2020) and
outcome of the sulphur monitoring for 2023. MEPC 82/INF.2. IMO.
International Maritime Orginization. (2024, 10 20). Emission Control Areas (ECAs) designated
under MARPOL Annex VI. Retrieved from
https://www.imo.org/en/OurWork/Environment/Pages/Emission-Control-Areas-(EC As)-
designated-under-regulation-13-of-MARPOL-Annex-VI-(NOx-emission-control).aspx
Kindt, S. S. (2016). MAN B&W Two-stroke Engines Latest Design and Development within
Engine Types, Tier III and Multiple Gas Fuels. Paper 116. Helsinki: CIMAC.
Kyrtatos, A., et.al. (2016). The Development of the Modern Low-Speed Two-Stroke Marine
Diesel Engine. Paper 120. Helsinki: CIMAC.
MAN Energy Solutions. (2021, June 18). MAN Product Guide. Retrieved from MAN Energy
Solutions: https://www.man-es.com/marine/products/planning-tools-and-
downloads/proj ect-guides
Final Report 03.28987
87
-------
EPA Task Order 68HERC24F0432
MAN Energy Solutions. (2023, April). Basic Principles of Ship Propulsion. Retrieved December
11, 2024, from https://www.man-es.com/docs/default-source/marine/tools/basic-
principles-of-ship-propulsion_web_links.pdf
MAN Energy Solutions, (n.d.). CEAS engine calculations. Retrieved from MAN Energy
Solutions: https://www.man-es.com/marine/products/planning-tools-and-downloads/ceas-
engine-calculations
McGill, R., Tukey, J. W., & Larsen, W. A. (1978). Variations of Boxplots. The American
Statistician, 32(1), pp. 12-16.
Menasha, J., & et.al. (2011). Ammonium bisulfate formation temperature in a bench-scale
single-channel air preheater. Fuel Vol. 90, 2445-2453.
Muzio, L., & et.al. (2017). Ammonium bisulfate formation and reduced load SCR operation.
Fuel 206, 180-189.
NOAA Office For Coastal Management. (2025, 1 10). Marine Cadastre Hub. Retrieved from
http s: //hub. marinecadastre. gov/
OpenStreetMap Foundation. (2024, 10 14). OpenStreetMap. Retrieved from
https://www.openstreetmap.org/
Plotly. (2024, 10 5). Plotly. Retrieved from https://plotly.com/python/
PostgreSQL Global Development. (2024, 10 5). PostgreSQL. Retrieved from
https://www.postgresql.org/
Python Software Foundation. (2025, 1 10). Welcome to Python. Retrieved from
https://www.python.org/
QGIS Development Team. (2024, 10 5). Spatial without Compromise ¦ QGIS Web Site.
Retrieved from https://qgis.org/
Sandelin, K., & et.al. (2016). Paper 111 - SCR under pressure - pre-turbocharger NOx abatement
for marine 2-stroke diesel engines. CIMAC Congress. Helsinki.
Snyder, J. (2024, October 31). MANES cuts ME-GA from its two-stroke engine programme.
Retrieved from Riviera Maritime Media: https://www.rivieramm.com/news-content-
hub/news-content-hub/man-es-cuts-me-ga-from-its-two-stroke-engine-programme-82833
S0holt, N. (2021, February 11). Tier III Exhaust-After-Treatment Orders for Low-Speed Engines
Pass 1,000 Mark. Retrieved from MAN Energy Systems: https://www.man-
es.com/company/press-releases/press-details/2021/02/ll/tier-iii-exhaust-after-treatment-
orders-for-low-speed-engines-pass-1 -000-mark
Final Report 03.28987
88
-------
EPA Task Order 68HERC24F0432
Spahni, M., & et.al. (2023). New compact engines from WinGD tailored to the changing needs
of modern vessels. CIMAC Congress Paper 464. Busan.
Spahni, M., et. al. (2023). New Compact Engines from WinGD Tailored to the Changing Needs
of Modern Vessels. CIMAC Congress Paper 464. Busan, Korea: CIMAC.
Struckmeier, D., et al. (2019). MAN Energy Solutions - A Global Update on Emission Reduction
Technology. CIMAC Congress Paper 081. Vancouver: CIMAC.
United States Coast Guard. (2025). Automatic Identification System (AIS) Overview \ Navigation
Center. Retrieved 10 12, 2024, from Navigation Center:
https://www.navcen.uscg.gov/automatic-identification-system-overview
United States Environmental Protection Agency. (2022, December). Control of Air Pollution
from New Motor Vehicles: Heavy-Duty Engine and Vehicle Standards Regulatory Impact
Analysis. Retrieved from https://nepis.epa.gov/Exe/ZyPDF.cgi?Dockey=P1016A9N.pdf
Wasserstein, R., & Lazar, N. (2016, 03). The ASA's statement on p-values: context, process, and
purpose. The American Statistician. Retrieved from
http ://dx. doi .org/10.1080/00031305.2016.1154108
Wei, J. (2007). Formation temperature of ammonium bisulfate at simulated air preheater
conditions. MS Thesis. University of California, Irvine.
Werner, T. F. (2019). Engine Technologies & Aftertreatment Solutions. Retrieved from CIMAC:
https://www.cimac.com/cms/upload/events/circles/circle_2019_SMM/3_Thomas_Werner
_WinGD.pdf
WinGD. (n.d.). General Technical Data. Retrieved from WinGD: https://wingd.com/products-
solutions/engines/general-technical-data
Final Report 03.28987
89
------- |