Beyond 3 Star Emission Capability for Outboard Engines

7655 7-3-5 Beyond 3 Star Emission Capability for Outboard Engines Greg BELL, Simon BREWSTER and Steven AHERN Orbital Corporation Limited. Copyright 7 Society of Automotive Engineers of Japan, Inc. and Copyright 7 SAE International This paper investigates strategies intended for development towards Super Ultra Low emission levels ( Star) from the existing 3 Star emission status for stroke Direct Injection and stroke outboard engines. The current 5 Mode ICOMIA outboard emission status is considered for Hydrocarbons, Nitrogen Oxides and Carbon Monoxide. From these a gap analysis is developed and possible strategies to close the gaps are considered. The potential of these strategies for stroke air assisted Direct Injection outboards is reviewed together with supporting data where appropriate. Initial data show high potential for reduction of emission levels beyond the 3 Star standard, whilst further work will confirm the levels that may be feasible in a production outboard engine. Keywords: Outboard, Marine, Two Stroke, Direct Injection, Emissions. INTRODUCTION The California Air Resource Board (CARB) 3 Star emission rating scheduled for implementation in 8 represents the current and most stringent criteria for gasoline outboard engine emissions on a global basis. Table - CARB Emission Standards Whilst Stroke and Direct Injected -stroke engines can homologate against these standards, it is necessary to investigate and develop solutions for the next possible level of emission requirements, as defined by the Star CARB standard being implemented for Stern Drive/Inboard marine gasoline engines. Whilst there is no legislation in place today for Star emissions for outboards, it can be anticipated that the emission levels will be reduced to around 6 of 3 star levels or to star levels. The Star, or Super Ultra Low, emission levels, 5 gm/kwh HC+NOx, represents around a further 7 reduction from the 3 Star emission standard and presents a significant challenge for both -and -stroke engines. Outboard applications in particular are not well suited to typical exhaust after-treatments such as catalysts due to both packaging constraints, material temperatures and salt water poisoning of catalysts, given the 35 3 hour life expectancy for the product.. EMISSION STANDARDS The CARB emission standards represent the most challenging in the world for outboard engines, and are shown in Table and Figure. In addition deterioration factors are shown in Table. HC + NOx (gm/kw hr) EPA + CARB Star CARB Star CARB 3 Star CARB Star 9 8 7 6 5 3 5 5 Power (kw) Figure - CARB Emission Standards SETC 7 /9

Table - Deterioration Factors 5 7655 7-3-5 HC CO NOx SDI.3.3.5 S.3.3.3 S+CAT.6.6.6 SD/I.6.35.3 Currently there are no restrictions on CO. It is possible that CO will be controlled due to the risk of excessive levels on the boat and possibly shore. The EPA has limited CO to g/kwh for engines larger than kw as a regulatory alternative. 3. TECHNOLOGY STATUS Both and 3 star emission levels can generally be met by - stroke DI and four stroke engines as shown in Figure and Figure 3 6 7. NOX (gm/kwh) NOX (gm/kwh) -STROKE HC vs NOx Emissions MY5 & 6 for >5 Hp 6 8 6 5 5 5 3 HC (gm/kw hr) Star 3 Star (5 Hp) 3 Star (5 Hp) 6 8 6 Figure - Four Stroke EPA Certification Data -STROKE HC vs NOx Emissions MY5 & 6 for >5 Hp 5 5 5 3 HC (gm/kw hr) Star 3 Star (5 Hp) 3 Star (5 Hp) Figure 3 - Two Stroke EPA Certification Data Figure - EPA Proposed CO Limit and Data This data suggests that the reduction of emissions from Star to 3 Star can be achieved with a combination of calibration and engine design. In addition, in some cases post oxidation is used without the addition of a catalyst or specific thermal reactor to further reduce exhaust emissions. 8 star exhaust emission levels have already been introduced for inboard and stern drive engines. Although the base engines were typically capable of achieving 3 star emissions, these engines to date have required catalyst technology to comply with this standard.. EMISSION REDUCTION STRATEGY.. STROKE DIRECT INJECTED ENGINES Overview Although current direct injected -stroke (SDI) engines can meet 3 star emissions a further reduction of 7 is required to meet the CARB star level or around 5 to meet the EPA proposals. Figure 3 shows some SDI engines already have low NOx emission levels, but these should be reduced further to meet the HC+NOx level. Figure indicates proposed CO levels proposed by the EPA be met with the 3 star calibration and hardware. HC emissions clearly exceeded the star level and both engine management and aftertreatment will be required to meet this target. Potential strategies to further reduce emissions of -stroke direct injected engines are: Post-oxidation. This is most applicable to modes, and 3. Post oxidation can be induced by retarding timings and increasing exhaust gas temperature in current engines. This also has the advantage of reducing NOx as well as HCs. It is likely that post oxidation could be made more effective with modifications to exhaust design to reduce the heat loss to the engine. SETC 7 /9

Retarded Injection The emissions and power at mode are both dependant on injection timing. If the injection timing is retarded the HC emissions will be reduced considerably. Power may be compromised with late injection. Injection Pressure and Spray Geometry Both an increase in injection pressure and spray geometry modifications can greatly reduce emissions and increase the power at the Mode point, generally recovering any power loss due to late injection. EGR This has been shown to be effective in the reduction of NOx on both two and four stroke engines. Although two stroke engines retain exhaust gas in the cylinder as the throttle is closed resulting in low NOx emissions, the NOx can further be reduced by cooling and mixing exhaust gas in the inlet or crankcase. This is most effective when used on modes and 3. Backpressure Increasing the backpressure can reduce NOx whilst lowering or keeping HC+NOx similar 9. This would be applicable on modes to 5. Piston Bowl Shape Using an automotive stroke piston bowl design will reduce HC at low loads (ie modes & 5) by better capturing injector spray and improving combustion robustness. Improved Airflow Control As stratified engines are not very sensitive to AFR, current outboard engines have excess air at low load to simplify the EMS and yet still maintain idle and torque control. A reduction of engine airflow will enable at least a proportional drop in emissions. This will be most applicable to modes and 5. Aftertreatment It is expected that it was the intent of the legislators that catalysts would be used to address the star emissions. Catalysts have been successfully used on Orbital automotive and motorcycle two stroke direct injected engines. As the engine out NOx emission is relatively low, it is practical to use an oxidizing only catalyst. For four stroke inboard and stern drive engines, catalysts appear to be the most practical solution, but there are some concerns with the durability in salt water. It is expected that significant changes will be required to the exhaust to accommodate the catalyst and water injection point. Transfer and Exhaust Porting. The scavenging of a -stroke engine has a marked effect of HC emissions at high load. Transfer porting designed to reduce bypass flow has been shown to have a positive effect. Coolant Temperature. Increasing engine temperature at mode 5, mode and possibly mode 3 will reduce HC emissions... STROKE ENGINES Overview 7655 7-3-5 Figure shows that the majority of four stroke outboard engines can already meet the 3 star emission levels. In comparison to SDI engines the four stroke engines have lower HC but higher CO and NOx emissions. As the typical NOx emissions are greater than 5 g/kwh, a reduction of NOx is the main challenge in meeting star emission levels. A few engines have low NOx, but have correspondingly high HC or CO emissions. The strategies for controlling emissions of a four stroke engine are well known and could include: A rich calibration: Merely calibrating the engine rich will reduce NOx at the expense of higher fuel consumption, CO and HC emissions. These could be further reduced with the use of secondary air and / or aftertreatment. EGR: The use of EGR is used effectively on automotive engines to reduce NOx whilst maintaining fuel consumption (Table 3). It may also allow the use of a three way catalyst at higher manifold pressures and hence lower pumping work. EGR is more effective with direct injected four strokes as a higher flow rate of EGR can be tolerated. Table 3 - Effect of EGR on SD/I 7 EGR HC NOx CO Power BSFC Scenario g/kwh g/kwh g/kwh kw g/kwh 6.8 L.7 3. 6.5 5 36 with egr.7 7..3 5 36 7. L.5 8. 7 9 39 with egr.5.8 8 9 356.3 L.9.7 53 39 with egr. 5.3 9 8 35 Catalysts: Automotive technology has shown that low emissions can be achieved with the application of catalysts. The use of catalysts in a marine environment offers more challenges as contact with salt water will reduce the efficiency of the catalyst. In contrast to this, catalyst technology has been demonstrated to be viable on four stroke SD/I. It is likely further development will be required due to the possible risk of salt water exposure and the higher output and fewer packaging opportunities on an outboard engine. SETC 7 3/9

5. EMISSION MODAL AND GAP ANALYSIS FOR - STROKE DI OUTBOARD Strategies such as listed in the discussion in section. for - stroke outboards can reduce engine emissions at specific, or all modes. The final calibration requires a combination of reductions for all the 5 modes. The contribution of each mode has to be considered as per Table. If the power * weight column is considered emissions at mode are most significant on a brake specific basis, followed by those at WOT. Table - CFR-9 for Marine engines Mode Speed Torque Power Weight P*W 6 9 8 7 58 39 3 6 7 8 5 5 5 5 idle - In an analysis, the EPA separated SDI engines into two classes when determining the contribution of each mode 7. Class appears to refer to single fluid direct injected systems, class to air assisted ones. The bulk of emissions from the class engines were at low load, while the bulk of emissions from the class engines were from the mode point. Modes 3 and on both systems had relatively low emissions. This data may not be a completely accurate reflection of each systems ultimate capability as it appears to be based on production engines and it is possible that with different calibrations the emissions of both systems could be further reduced. For both types of engines the more heavily weighted modes and 3 appeared lower than that of a four stroke. A typical breakdown of the emissions for a type engine is shown in Table 5. Modes to 3 are considered to be able to use homogeneous combustion which has relatively high exhaust temperatures, making aftertreatment easier. Modes and 5 use stratified combustion. From this table, it is evident that the vast majority of the HC and NOx emissions result from modes to 3. To achieve emission control beyond 3 star, improving the emission output from these points becomes the critical step 7655 7-3-5 Table 5 - Breakdown of Modal emissions CONTRIBUTION OF EACH MODE Mode HC CO NOx FUEL 35 5 37 8 7 7 3 3 3 6 3 5 9 3 5 6. POTENTIAL APPLICATION OF THE -STROKE STRATEGIES TO REDUCE BEYOND 3 STAR EMISSIONS ON AIR ASSISTED DI OUTBOARD With reference to the modal gap information as outlined in table, the initial investigations being carried out by the authors have centred on the following strategies: Reduce and control HC with calibration and injector selection. At mode, increasing power will enable retarding of injection timing, a key HC controlling mechanism. At points and 3, both calibration and promoting post oxidation with the existing OEM exhaust system will be investigated for reducing NOx. Reduce NOx initially by calibration (retarding ignition and injection timing controlling in cylinder temperatures and pressures, and rate of rise of cylinder pressure), targeting modes to 3. Reduce HC and NOx at modes and 5 by calibration (improved control of airflow). Additional NOx control (mode and 3) by investigation of effect of adding controlled EGR. EGR has been shown on previous automotive -stroke engine programs to be effective for NOx control without sacrifice to fuel economy. EGR is expected to be effective especially at modes and 3. After treatment of HC, especially at modes to 3, to further reduce the overall HC+NOx. Although post oxidation occurs readily in standard engines, it is expected that this effect will be required to be enhanced by the use of: Insulation on the exhaust passages Shifting the water injection point and allowing oxidation to occur in the outboard leg Thermal reactor Oxidation catalyst The lean mixture of the SDI engine makes it an ideal candidate for a thermal reactor, and in fact post oxidation occurs in a well cooled exhaust. It is expected that high efficiencies can be achieved with a well designed thermal reactor and it has the SETC 7 /9

advantage of not having the same sensitivity to salt water as catalyst solutions. Should a catalyst solution be required as well, a lot of care will be required to protest it from salt water as there is considerable risk during abnormal operation of some engines. The basic approach will be to initially optimize the engine HC and NOX capability with calibration and injector optimsation, then to investigate the effects of added hardware for EGR and aftertreatment. The CO of the air-assisted -Stroke DI outboard is typically less than gm/kwh, and as such, the emphasis has been on the investigation of HC and NOx control. 7. ASSESSMENT TO DATE OF THE STRATEGIES TO REDUCE BEYOND 3 STAR CAPABILITY FOR - STROKE DI OUTBOARD Testing was carried out on an off-the-shelf -stroke air assisted DI outboard engine, with testing carried out as per the procedures as detailed in CFR Part 9. 7.. CALIBRATION AND INJECTION OPTIMISATION Following the general strategies as discussed above, the initial assessments were carries out by calibration and injector optimization for modes to 5, with no changes to the base engine or exhaust systems. Mode The key changes to assist in mode was the increase of the system air pressure from 55 kpa to 65 kpa and in the change of the injector spray shape to increase the penetration rates. The change to the injector spray shape, shown in Figure 5, results in improved mixing within the combustion chamber and enabled a power increase of approx 3 kw or 3.5. 7655 7-3-5 about 9ºbtdc for maximum power, but with some trade off in power, HC emissions can be greatly reduced. This figure also shows the potential of either up-sizing, or down-powering the engine for emission control at the maximum power point. Normalised Power () S Emissions vs Injection Timing at Mode 95 9 85 8 75 7 65 6 55 5 8 Injection Timing 6 BSHC BSNOx BSHC+NOx Power Figure 6 - Emissions vs. Injection Timing Mode Mode 9 8 7 6 5 3 Although the contribution to emissions of Mode is less than mode, it is the most heavily weighted point. Hydrocarbons from mode can be reduced by post-oxidation mainly due to late injection and ignition increasing exhaust temperature. It was possible to induce post oxidation even with the standard engine and water-cooled exhaust passages. A further advantage of retarded timings was the corresponding reduction in NOx. Although both retarding injection and ignition were useful tools, retarding ignition was more effective and the effect is shown in Figure 7. S Emissions vs Ignition Timing at Mode Standard Injector Modified Injector Figure 5 - Change in injector spray geometry The emissions were then reduced by retarding the air injection event (SOA). Figure 6 indicates the emission reduction and power verses SOA. Normally the engine would be calibrated with a SOA of 8 3 7 36 68 8 6 8 6 Ignition dbtdc BSHC BSNOx BSHC+NOx BSFC BSCO Figure 7 - Emissions vs. Ignition Timing Mode SETC 7 5/9

Mode 3 Mode 5 7655 7-3-5 A similar approach was used on mode 3 as per mode. This is illustrated in Figure 8. As with the mode point, smaller gains are made by calibrating injection and air fuel ratio, but retarding ignition to promote post oxidation is the major factor. S Emissions vs Ignition Timing at Mode 3 At mode 5 the engine is stratified and emissions are determined by the injector and spark plug. Typically the concentration of emissions remains relatively constant, but the mass emissions can be reduced by reducing airflow. In this case the emission reduction achieved was less than, possibly not justifying airflow control. This is shown in Figure. BSFC (g/kwh) 338 336 33 33 33 38 36 5 35 3 5 Ignition dbtdc. 3.7 3.5 35 3 5 5 5 5 BSHC BSNOx BSHC+BSNOx BSCO BSFC Figure 8 - Emissions vs. Ignition Timing Mode 3 BSHC+NOx+CO (g/kwh) Emissions g/sec S Emissions vs Air Fuel Ratio at Mode 5.35.3.5. 8.5 6..5 6 8 AFR HC HC+NOx Exh Temp. Temperature (deg C) Mode Figure - Emissions vs. AFR Mode 5 Mode operation is significantly different to modes, and three as the engine operated in the stratified mode. To maintain good combustion the ignition and injection timing are relatively constrained and the major parameter controlling emissions is air fuel ratio. Current 3 star engines are calibrated with excess air mainly to simplify calibration and the throttle design. As the excess air is reduced an almost proportion reduction in HC+NOx can be realized, as shown in Figure 9. S Emissions vs AFR at Mode 5 38 5 36 3 5 3 3 6 8 3 3 3 Air Fuel Ratio BSHC BSNOx BSHC+NOx BSFC Figure 9 - Emissions vs. Air Fuel Ratio Mode BSFC (g/kwh) 7.. AFTERTREATMENT AND EGR 7... AFTERTREATMENT Typically in outboard engines, the exhaust manifold and exhaust are water cooled. To encourage post oxidation, the following steps were taken to isolate the cooling water from the exhaust gas. Lining the exhaust manifold with Stainless steel Removing the water cooling from the exhaust Additional baffles to prevent water splash. Mode At mode post-oxidation occurs with the modified exhaust system generally independent of calibration, with a BSHC of approx. g/kwh and BSNOx of 3. g/kwh. There was approximately a 5 reduction in power. It is expected that the bulk of this drop in power is related to the lack of optimization of the exhaust system. Mode Although post oxidation could be induced on modes to 3, the technique is best illustrated at mode. Calibration of the engine using post oxidation was performed in two steps SETC 7 6/9

. Reducing the AFR until the BSHC was minimized. Figure shows the BSNOx reducing as the AFR is reduced, while BSHC has a minimum, in this case 5.5:. The fuel consumption also increases as the AFR is reduced. 6 5 3 AFR Scan Mode with Post Oxidation.5 5 5.5 6 AFR 6.5 7 7.5 35 3 5 5 BSPEHC BSPENOX BSHC+NOx BSPEFC BSPECO Figure - AFR Scan Mode. Figure shows the effect of ignition timing. Both BSHC and NOx are reduced as the ignition is retarded, but the fuel consumption continues to rise. IGN Scan Mode with Post Oxidation 3.5 35 3 3.5 5.5 5.5 5 5 6 7 8 9 IGN BSPEHC BSPENOX BSHC+NOx BSPEFC BSPECO 7... EGR Figure - IGN Scan Mode The objective of the EGR system was to reduce NOx. For development this was achieved by removing the exhaust gas from the test cell system, before the backpressure valve. The flow was controlled by a butterfly valve and it was mixed into the inlet airflow upstream of the engine throttle. No attempt was made to integrate the EGR system into the engine during this phase. 5 7655 7-3-5 cooling water, the maximum temperature was less than 6 deg C. The starting point for the EGR scans is a point optimized using the methods previously discussed to promote post-oxidation with the insulated exhaust. Mode It was possible to reduce NOx with the addition of EGR, but this also reduced power up to 8. At this stage EGR will not be considered at Mode. Mode At mode the AFR was first reduced until HC began to rise, and then the ignition retarded to induce post oxidation and reduce NOx. Without EGR the BSHC was reduced to.5 g/kwh and the BSHC+NOx 3 g/kwh. Figure 3 shows the effect of the addition of EGR enabled the NOx to be reduced from.9 g/kwh to. g/kwh. This reduced BSHC+NOx to.9 g/kwh. EGR Scan Mode with Post Oxidation BSHC+Nox (g/kwh) 3.5 3 35.5 3 5.5 5.5 5 6 8 EGR BSPEHC BSPENOX BSHC+NOx BSPEFC BSPECO Figure 3 - EGR Scan Mode Mode 3 Figure shows a similar calibration technique used at Mode 3 and BSHC was reduced to.6 g/kwh with a BSHC+NOx of.65 g/kwh. Although it was possible to reduce the BSNOx with the addition of EGR the BSHC+NOx could not be reduced, possibly because the addition of EGR reduced the exhaust temperatures. As the exhaust gas was extracted after it was mixed with the SETC 7 7/9

EGR Scan Mode 3 with Post Oxidation.5 5 35.5 3 5 5.5 5 6 8 EGR BSPEHC BSPENOX BSHC+NOx BSPEFC BSPECO Backpressure Figure - EGR Scan Mode 3 Figure 5 shows the effect of backpressure was tested as an alternative to external EGR. The engine backpressure was increased externally using the test bench hardware and throttle opened to maintain airflow. Increasing the backpressure reduced BSNOx due to the increased internal EGR. The result was not quite as effective as application of external EGR, however it remains a viable option as the packaging is easier. Backpressure Scan Mode 3.5 35 3 3.5 5.5 5.5 5 5 5 5 Back Pressure BSPEHC BSPENOX HC+NOx BSPEFC BSPECO Figure 5 - Backpressure Scan Mode 8. DISCUSSION AND NEXT STEPS Significant reductions in BSHC have been demonstrated to date with the use of insulated exhausts and calibrating the engine to induce post oxidation. This calibration requires the AFR to be rich and the spark retarded which results in low NOx. The effectiveness of EGR to reduce NOx was also shown to be effective, whether external or internal. SETC 7 8/9 7655 7-3-5 The effect of up-sizing, or down-powering the engine, basically controlling the emissions at the mode point can be significant. These techniques combined show the potential to achieve significantly improved HC+NOx emission levels on a prototype outboard engine without using a catalyst, on emission reductions to modes to 3 alone. Further work is required to improve the design of the exhaust system to enable the fuel consumption to be improved by enabling post oxidation at leaner mixtures and more advanced timings. The theory of thermal reactors is documented and initial materials for the construction have been determined for automotive applications. The option of using an oxidizing catalyst needs to be further investigated. It is likely to improve efficiency, but requires the issues brought about by operating in a salt water environment to be addressed. Alternatively, a low temperature catalyst could be used for modes & 5 and switched. If required for further HC control at modes and 5, combustion chamber geometry optimization will be required. To date all testing has been carried out using the OEM supplied combustion chamber. Packaging has yet to be addressed. There may be compromises to the EGR system or advancements in the exhaust system in a production design. The analysis to date has looked at the ICOMIA cycle in terms of what technology/hardware/calibration is optimal for each individual point. A final key step in this investigative program is to define the overall system and assess from both a Star emission capability, and a manufacturing/productionising/cost feasibility viewpoint. 9. CONCLUSION A methodology to investigate Star Emission capability has been carried out on a production S air assist DI outboard engine. The test data to date confirms that the strategy to meet Star by gap analysis on the individual ICOMIA points is promising. The test data show: Increased fuel system pressure and higher penetration injectors improve performance and assist in HC reduction at the mode point. Induced post oxidation both by calibration and by

modifying the exhaust system is effective in controlling HC s especially at modes, and 3. However to maximise the post oxidation efficiency, the calibration needs to be rich, which compromises the fuel consumption. External EGR (or backpressure control of EGR) is effective in reducing NOx. The Next Steps include: Investigate alternative post oxidation strategies. Investigate potential oxidation catalyst options. Combustion system optimization. Define overall system specification for assessment relative to the star emission standard and with regard to commercial feasibility aspects. ACKNOWLEDGMENTS The authors would like to thank all of the dedicated personnel at Orbital Engine Company who have either directly or indirectly contributed to this paper. REFERENCES http://www.epa.gov/otaq/regs/nonroad/marinesiequipld/d7chp.pdf EPA, Control of Emissions from Marine SI and Small SI Engines, Vessels, and Equipment Draft Regulatory Impact Analysis Chapter Regulatory Alternatives http://www.arb.ca.gov/planning/sip/7sip/apr7draft/sipmeas.pdf ARB, Draft Air Quality Plan, Page 9 3 http://ecfr.gpoaccess.gov/cgi/t/text/textidx?c=ecfr&sid=d87b3f7cbcfebba9e8a9e&rgn=div8&v iew=text&node=:...5...5&idno= ecfr Title -9.5 7655 7-3-5 8 Strauss S., Zeng Y.,Montgomery D.T., Optimisation of the ETEC Combustion System for Direct-Injected Two-Stroke Engines Towards 3-Star Emissions, SAE paper 3-3-7 / 337 9 Bell G., Finucci C., Exhaust Emissions Sensitivities with Direct Injection on a 5cc Scooter, SAE Paper 97365 Fujimoto, H., Isogawa, A., Matsumoto, N Catalytic Converter Applications for Two Stroke, Spark-Ignited Marine Engines, SAE Paper 958 Houston R, Archer M, Moore M and Newman R, Development of a Durable Emission Control System for an Automotive Two-Stroke Engine, SAE Paper 9636 Jaimee A, Schneider D A, Rozmanith A I, Sjoberg J W Thermal Reactor Design, Development and Performance, SAE Paper 793 CONTACT Steven Ahern Business Development Manager Orbital Corporation sahern@orbitalcorp.com.au ACRONYMS SDI Two-Stroke Direct Injected CARB California Air Resources Board CO Carbon Monoxide EGR Exhaust Gas Recirculation EGT Exhaust Gas Temperature EPA Environment Protection Agency (US) HC Hydrocarbons ICOMIA International Council of Marine Industry Associations NOx Oxides of Nitrogen SOA Start of Air Injection http://arb.ca.gov/msprog/marine/marinectp/reg.pdf ARB, CALIFORNIA CODE OF REGULATIONS, TITLE 3. Motor Vehicles, Chapter 9. Off-road Vehicles and Engines Pollution Control Devices, Article.7. Spark-Ignition Marine Engines page A5. 5 http://www.epa.gov/otaq/models/nonrdmdl/nonrdmdl5/r5 3.pdf EPA Nonroad Spark-Ignition Engine Emission Deterioration Factors 6 http://www.epa.gov/otaq/certdata.htm#marinesi US EPA, Engine Certification Data 7 http://www.epa.gov/otaq/regs/nonroad/marinesiequipld/d7chp.pdf US EPA Control of Emissions from Marine SI and Small SI Engines, Vessels, and Equipment Draft Regulatory Impact Analysis Chapter Feasibility of Exhaust Emission Control SETC 7 9/9