The advantages of applying Compressed Natural Gas (CNG)
as a fuel for internal combustion engines are well known. In addition to a signifcant operating cost savings due to a lower fuel price relative to diesel, there is an opportunity to reduce the engine's emissions. With CNG combustion, some emissions, such as Particulate Matter (PM) and Carbon Dioxide (CO2), are inherently reduced relative to diesel fueled engines due to the nature of the combustion and the molecular makeup of the fuel. However, it is important to consider the impact on all emissions, including Total Hydrocarbons (THC) and Carbon Monoxide (CO), which can increase with the use of CNG. Nitrogen Oxides (NOx) emission is often reported to decrease with the use of CNG, but the ability to realize this beneft is signifcantly impacted by the control strategy and calibration applied. FEV has investigated the emissions and performance impact of operating a heavy-duty diesel engine with CNG in a dual fuel mode. The CNG was introduced via injectors mounted to an inlet pipe located upstream of the intake manifold. The fumigation approach included a mixer to improve the distribution of gas prior to delivery to the cylinder. The initial investigations sought to determine how the performance of a heavy-duty diesel engine would be affected by the introduction of CNG. For this effort there was no change to the base engine calibration, and the ability to maximize substitution of diesel with CNG was investigated. It was observed that the ability to maximize substitution of diesel with CNG across the operating map was limited by extremely high THC levels, combustion instability and limitations in peak cylinder pressure and exhaust gas temperature. With the application of a simplifed engine calibration with a single diesel injection and Exhaust Gas Recirculation (EGR), timing adjustments allowed higher CNG substitution levels in several areas of the operating map. A further increase in gas substitution along with higher fuel conversion effciency, improved combustion stability and even lower emissions could be achieved through Reactivity Controlled Compression Ignition (RCCI) combustion. This approach required a unique injection strategy along with a careful balance of EGR rates and boost pressure. Under this combustion regime it was possible to observe a simultaneous reduction of NOx and PM emissions, approaching engine-out emission levels that could avoid, or signifcantly minimize, aftertreatment of these species. With the desire to quickly apply CNG systems to existing diesel engine architecture in an effort to reap the beneft in fuel cost savings, manufacturers and system developers must be careful to understand the full impact on the engine's performance and emissions. Tests conducted as part of this investigation have revealed that an un-optimized approach to CNG introduction can lead to extreme THC emissions that mostly consist of Methane (CH4). In addition, the maximum gas substitution level is signifcantly limited in most regions of the engine operating map. Thus, the ability to specifcally tune the calibration for operation with CNG is essential to achieving the maximum beneft in fuel cost savings and emission control. Introduction The use of CNG in heavy-duty applications has seen a sharp growth in the last couple of years as reported in studies conducted by the Energy Information Administration (EIA) [1]. EIA studies project CNG consumption in heavy-duty applications to increase at 11.4% per year through 2040. The growth is mainly seen in the off-road market, but other sectors such as on-highway, marine and locomotive are also embracing a switch to CNG at a rapid rate. There are two main technology approaches through which CNG can be applied for heavy-duty engines. A common approach for existing diesel engines is to operate in a dual fuel regime where diesel and CNG are combusted simultaneously. Investigation of Diesel and CNG Combustion in a Dual Fuel Regime and as an Enabler to Achieve RCCI Combustion 2014-01-1308 Published 04/01/2014 Mufaddel Dahodwala, Satyum Joshi, Erik W. Koehler, and Michael Franke FEV Inc. CITATION: Dahodwala, M., Joshi, S., Koehler, E., and Franke, M., "Investigation of Diesel and CNG Combustion in a Dual Fuel Regime and as an Enabler to Achieve RCCI Combustion," SAE Technical Paper 2014-01-1308, 2014, doi:10.4271/2014-01-1308. Copyright 2014 SAE International Downloaded from SAE International by Erik Koehler, Wednesday, March 26, 2014 11:09:34 AM A second approach considers a dedicated CNG combustion system that requires the addition of an ignition source. The main technological challenges that apply to this type of engine conversion are studied in detail by Ribas [2]. As an additional consideration, past studies have shown that due to throttling losses and lower compression ratio, effciency of a Spark Ignited (SI) CNG engine is lower compared to a dual fuel diesel/CNG engine, particularly at part loads [3]. Brake Mean Effective Pressure (BMEP) capability is also reduced with an SI CNG engine, leading to lower torque at low speeds when compared to a dual fuel diesel/CNG engine. An SI CNG engine does however have an advantage on the exhaust gas aftertreatment, as stoichiometric operation allows the application of a conventional and comparatively simple 3-way catalyst. In the case of dual fuel diesel/CNG engines, two types of technologies are currently considered depending on the method used for CNG induction. In the frst type, both the diesel fuel and natural gas are directly injected into the combustion chamber using either two separate injectors or a special injector with a dual-concentric needle design [4]. In the second type, natural gas is either fumigated at a single point into the intake path of the engine before or after the turbocharger compressor and is premixed with air and EGR or injected at multiple points in the intake port of the engine. Intake fumigation at single point is the most widely used method of CNG induction for on-highway applications, due to its simplicity, and is the approach applied for this work. Much research has been conducted to understand the combustion behavior of dual fuel diesel/CNG engines [4, 5, 6, 7, 8, 9, 10]. As reported by most researchers, the major diffculty with dual fuel operation is the challenge of providing high levels of CNG substitution, especially at low and medium loads. Attempts to apply high substitution levels lead to lower engine effciency and higher concentrations of CO and unburned hydrocarbon emissions. At higher loads, the improvement in CNG utilization leads to improvements in both engine performance and exhaust emissions. To date a large portion of the research has been focused on either single cylinder research engines or light-duty production engines. Relatively few studies look at engine performance and corresponding emission impact of dual fuel concepts applied to medium-duty and heavy-duty production engines. One study on a heavy-duty engine was conducted by Guzman et. al. [4], wherein the authors recalibrated the base diesel engine to operate in dual fuel mode to quantify the CO2 and PM emissions during the transient portion of the Heavy-Duty Federal Test Procedure (HD FTP) and World Harmonized Transient Cycle (WHTC). Their work found that, with a properly calibrated dual fuel engine signifcant reductions in PM and CO2 emissions could be achieved in transient operation when compared to a baseline diesel engine. In recent years, the heavy-duty market has seen the strongest growth in the application of dual fuel technologies, with refuse trucks and buses leading the way. In terms of on-highway applications, most dual fuel engines currently in use are End Of Life (EOL) retrofts that do not need to comply with stringent emission standards. To date, few companies have certifed dual fuel engines to US2010 emission standards. One of those that has certifed to US2010 has stated a fuel cost savings of up to 30% and a maximum CNG substitution of 70%. One motivation behind this study was to understand the limitations imposed by a US2010-compliant on-highway diesel engine in terms of maximum CNG substitution. Additional complexity to this approach results from the management of cylinder-to-cylinder variations, which in some cases makes the most promising strategy developed on a single cylinder engine not feasible for a production engine without applying signifcant design changes. One of the main requirements for employing a dual fuel approach is to have the fexibility of operating in diesel-only mode. Design changes to specifcally allow better operation with CNG will not be attractive to the end user if they compromise the performance of the engine in diesel-only mode. The frst part of the study aims at understanding engine behavior with introduction of CNG without changes to the base calibration, as is typically done by retroft companies. To accomplish this, 13 load points were selected, and a maximum CNG substitution sweep was conducted at each of these points. The second part of the study investigated the impact of diesel calibration changes on the allowable maximum CNG substitution and thus the achievable fuel cost savings. The calibration optimization was completed at all 13 load points. The fnal part of this study is focused on exploring advanced combustion concepts with CNG and seeks to maximize substitution and lower emissions. Experimental Setup A production heavy-duty diesel engine was used for this investigation. The 2010 inline 6-cylinder engine has a displacement of 13 liters, is rated at 425 HP and is On-Board Diagnostics (OBD) - II compliant. It is equipped with a high pressure common rail fuel injection system, cooled high pressure EGR, and a compression ratio of 16.5:1. CNG was introduced into the intake using eight CNG injectors located downstream of the charge air cooler. A mixer was installed downstream of the CNG introduction location to support equal distribution of the CNG in the intake manifold. Figure 1 depicts the test cell setup applied for this investigation. A Rapid Controller Prototyping (RCP) system was used for controlling the amount of CNG induction. The production engine was equipped with a Diesel Oxidation Catalyst (DOC), Diesel Particulate Filter (DPF) and Selective Catalytic Reduction (SCR) catalyst, however for this investigation only a DOC was installed and the backpressure was adjusted via a back pressure valve to simulate the absent aftertreatment components. The aftertreatment module in the Engine Control Unit (ECU) was deactivated to prevent it from affecting the base engine performance. Downloaded from SAE International by Erik Koehler, Wednesday, March 26, 2014 11:09:34 AM Figure 1. Test Cell Schematic The engine was coupled to a 560 kW Alternating Current (AC) dynamometer. CNG and diesel fuel fow measurements were accomplished using Micro Motion fow meters while the air fow was measured using an ABB air fow meter. Engine-out and tailpipe gaseous emissions were measured with a dual channel Horiba MEXA 7500 DEGR emission bench. The intake manifold was instrumented for CO2 concentration measurement, which was used for calculation of the EGR rate. PM emission was determined through the Motor Industry Research Association (MIRA) calculation, which utilized data provided by a smoke meter drawing sample upstream of the DOC. The engine was instrumented with Kistler 6061B in-cylinder pressure transducers to allow cylinder pressure measurements on all six cylinders. Test Matrix Based on operating points from the Ramped Modal Cycle (RMC), HD FTP and highway cycles for this engine, 13 specifc test points were selected for this study. The points were selected to provide a good balance between certifcation cycles and real-world operation. The program test points are overlaid upon the cycle operating points in Figure 2. For the baseline substitution study, the CNG substitution was increased in 10% increments until the engine became unstable (Indicated Mean Effective Pressure (IMEP) Coeffcient of Variation (COV) less than 10%) or a limit of peak pressure or maximum exhaust temperature was reached. Figure 2. Test points selected based on the operating points on the RMC, HD FTP and highway cycles For optimization studies, the effect of main injection timing and CNG substitution was studied at each of the selected operating points. Figure 3 shows an example test matrix for one particular engine speed. The main injection timing was swept at different substitution levels across the three load points. At high loads the timing could not be advanced because of peak pressure limitations, and it was found that higher substitution could only be achieved by retarding the main timing. RCCI combustion was evaluated at the low load points. The test matrix for RCCI combustion at low loads was an extension of the test matrix shown in Figure 3. The timings were advanced to 50-80 deg Crank Angle (CA) before Top Dead Center (bTDC) at substitution levels of 60-80%. The CNG substitution was calculated on an energy basis. The details of the calculation are outlined in the Appendix. Figure 3. Test matrix applied for optimization at a fixed speed Result and Discussion Baseline Maximum Substitution A review was conducted to understand the methods employed by retroft manufacturers to substitute CNG for diesel without changes to the base calibration. Two popular approaches are to manipulate the boost signal and to bleed off part of the fuel supplied to the injectors back to the fuel tank to reduce the injected diesel fuel for a particular load. These approaches are not directly possible when considering an engine equipped with OBD. Without OBD calibration changes, such approaches would cause a low boost or injector fault and result in a de-rate of the engine. Thus, to avoid this infuence for the current work, the reduction in diesel quantity was obtained by reducing the pedal demand and maintaining load by introducing CNG into the system. With this approach, the control parameter set points, including the injection events, EGR rates and rail pressures, were not consistent for different substitution levels within the same load point. These variables were controlled by the ECU based on the commanded pedal (fueling) and the engine speed. In an effort to maintain this paper within a reasonable length the results presented without calibration changes will be limited to 1500 rpm and three different load points; 6 bar, 14 bar and 24 bar. These points will be analyzed to understand the effect of CNG substitution on engine performance and emissions. Downloaded from SAE International by Erik Koehler, Wednesday, March 26, 2014 11:09:34 AM Figure 4 shows the Apparent Heat Release Rate (AHRR) and cylinder pressure traces for 6 bar BMEP at three different substitution levels. The emission impact and corresponding control variables for different substitution levels at this load point are provided in Figure 5. From these fgures, deterioration in combustion effciency can be seen with an increase in CNG substitution, possibly due to the reduction in injected diesel fuel leading to lower combustion chamber temperatures. This led to signifcant increase in hydrocarbon emissions, Brake Specifc Fuel Consumption (BSFC) (calculated on a diesel equivalence fuel fow rate) and brake specifc cost, which is defned as the net fuel cost per kilowatt hour of energy (Appendix). An increase in NOx emission was observed up to 40% CNG substitution, due to reduction in EGR, but then dropped as the combustion effciency deteriorated. The EGR reduction can be attributed to the richer lambda values observed with increasing CNG substitution as the fresh air charge is displaced with CNG. Increased PM emissions at higher substitution were possibly due to the increased soluble organic fraction at higher loads and a richer lambda value. Figure 4. AHRR and in-cylinder pressure for 6 bar BMEP load at 1500 rpm at three different substitution levels A reduced peak in the heat release trace and a late burning of CNG, also observed by Maxey et al. [5], was clearly visible with an increase in substitution. Observations regarding the combustion behavior are in agreement with those of Maxey et al.; wherein the reduction in diesel fuel quantity reduces the amount of diesel available to ignite the less reactive CNG [5]. Figures 6 outlines the AHRR and cylinder pressure trace for 14 bar BMEP, while Figure 7 shows the emission impact with increasing CNG substitution at the same load. Similar trends for NOx, PM and THC emissions were observed for 14 bar BMEP as seen at 6 bar BMEP, however, here the combustion effciency remained above 90% leading to a net beneft in brake specifc cost. At this load, a maximum substitution of 98% could be reached where CNG was ignited by the diesel pilot injection and an SI type heat release trace was observed. Figure 5. Emissions and performance parameters at 6 bar BMEP, plotted against CNG substitution A NOx emission increase was closely coupled to a reduced EGR rate and an earlier heat release. In general, the combustion COV, as also observed by Sun et al. [6], was affected by CNG substitution percentage and the EGR quantity. The cylinder pressure rise rate and the peak cylinder pressure increased signifcantly at 24 bar BMEP, shown in Figure 8, due to the simultaneous combustion of diesel and CNG. This limited CNG substitution to 36% at this load point. Combustion effciency remained above 98.8% due to the higher combustion temperatures. The NOx, PM and THC emissions trends presented in Figure 9 for 24 bar BMEP are similar to those observed at 6 bar and 14 bar BMEP. Downloaded from SAE International by Erik Koehler, Wednesday, March 26, 2014 11:09:34 AM Figure 6. AHRR and in-cylinder pressure for 14 bar BMEP load at 1500 rpm Figure 7. Emissions and performance variables at 14 bar BMEP, plotted against CNG substitution Figure 8. AHRR and in-cylinder pressure for 24 bar BMEP load at 1500 rpm Figure 9. Emissions and performance variables at 24 bar BMEP, plotted against CNG substitution Downloaded from SAE International by Erik Koehler, Wednesday, March 26, 2014 11:09:34 AM Figure 10 shows the practical CNG substitution map when access to the calibration is not available and thus no calibration changes are applied. This map is determined by maintaining the engine-out emissions within a level that would allow the engine to meet US2010 emission legislation when conventional diesel aftertreatment is applied. The average NOx emission for the baseline diesel calibration was approximately 2.0 g/bhp-hr, therefore a NOx limit of 2.0 g/bhp-hr, along with a NMHC limit of 0.238 g/bhp-hr, was applied as limiting criteria. From Figure 10 we can see that the average substitution would be limited to less than 10% if the base calibration is carried over and no changes are applied. There are three major reasons for the limited substitution shown in Figure 10. First, at low loads the amount of unburned THC is very high. Second, at mid loads the NOx emissions tend to increase due to a reduction in EGR caused by running richer lambda values. Lastly, the peak pressures increase with increase in CNG substitution at higher loads, leading to the necessity to observe mechanical limits. Most dual fuel retrofts on the market today are applied to EOL vehicles that do not require emission certifcation. These applications can therefore accept increases in emission levels and apply higher CNG substitution levels. Figure 10. Practical CNG substitution map without base engine calibration changes and meeting US2010 emission standards Calibration Optimization As shown in the previous section, only very low CNG substitution levels can be achieved when changes to the base diesel calibration are not considered. Therefore, to increase the CNG substitution levels, a modifcation of the base calibration is required. The base calibration used three injections along with EGR to reach the desired NOx, PM and Non-Methane Hydrocarbons (NMHC) targets when operating on diesel fuel. To understand the effect of each injection event, a series of tests were conducted at 1500 rpm and 6 bar BMEP and multiple substitution levels. Three injection strategies were explored: main injection only, main + pilot injection and main + post injection. In order to eliminate the effect of EGR, the EGR rate was reduced and maintained at a near-zero level throughout modifcation of the base calibration. A summary of the injection strategy employed for each case is provided in Table 1. Figure 11 shows the impact of each of these strategies on the NOx and THC emissions at various CNG substitution levels. The NOx emissions decrease with an increase in CNG substitution levels for all three strategies, but none of the strategies show a signifcant advantage in their ability to control THC emissions. Therefore, to simplify the diesel injection control, a main injection only strategy was applied for the calibration optimization. Table 1. Overview of the three injection strategies employed at 1500 rpm, 6 bar BMEP and different CNG substitution Figure 11. Comparison of NOx and THC emission for multiple CNG substitution levels at 1500 rpm and 6 bar BMEP It was also realized that EGR was required to control the NOx emissions, as shown in Figure 12. A study was then conducted with multiple CNG substitution levels and diesel injection timings for the main injection only case. At each speed and load point the EGR rate and injection pressure were held constant. The boost pressure was controlled by the ECU based on the existing calibration. The diesel injection timing advance was restricted to 20 deg bTDC, above which the possibility of spraying the diesel fuel directly on the cylinder liner was high. The CNG substitution was limited to 80%; above this level the combustion was unstable Downloaded from SAE International by Erik Koehler, Wednesday, March 26, 2014 11:09:34 AM and resulted in very high THC emissions. The optimization study was conducted at all 13 points as outlined in the Test Matrix section. To describe the impact of injection timing and CNG substitution on emission and performance, the following text focuses on the results obtained at 1500 rpm and three different load points. Figure 12. Comparison of NOx emissions for multiple CNG substitution levels at 1500 rpm and 6 bar BMEP Figure 13 shows the effect of diesel injection timing and CNG substitution levels on the NOx, PM and THC emissions at 1500 rpm and 6 bar BMEP. A decrease in NOx emissions and an increase in THC emissions can be observed as CNG substitution levels are increased at constant injection timing. Figure 13. Injection timing study for different CNG substitution levels at 1500 rpm, 6 bar BMEP with 35% EGR Figure 14 compares the AHRR and pressure traces for three different substitution levels at a fxed injection timing of 11 deg bTDC. As shown in Figure 14, the initial peak of the AHRR trace reduces as CNG substitution is increased, while the longer burn durations lead to lower combustion temperatures and lower NOx emissions. However, at a fxed substitution level the NOx emissions increase as the injection timing is advanced, but the THC emissions decrease. Figure 14. AHRR and cylinder pressure trace for 1500 rpm, 6 bar BMEP at 11 deg bTDC with three different substitution levels Comparing the AHRR traces in Figure 15, it can be observed that, as the injection timing is advanced, the AHRR traces move toward Top Dead Center (TDC). This leads to higher in-cylinder temperatures causing higher NOx emission formation but at the same time allowing more time for the CNG to burn at higher temperatures and reducing the THC emissions. Although the best NOx emission results are obtained between 60% and 80% substitution levels and approximately 12 deg bTDC injection timing, at this point, the unburned THC emissions are very high. Thus, a compromise must be identifed between the maximum substitution and the allowable NOx and THC emissions. The PM emissions, on the other hand, are very low across the complete optimization range, with a minimum achieved at 20 deg bTDC between 20% and 60% CNG substitution levels. Unfortunately, the NOX emissions in this range are very high. Figure 16 shows the NOx, PM and THC emissions at 1500 rpm and 13 bar BMEP. The NOx emission results are similar in terms of a decreasing trend with increase in CNG substitution levels at the same injection timing and an increase in NOx emissions with injection timing advance at a constant substitution level. Again, as in the previous case, the regions where the NOx emissions are lowest have very high THC Downloaded from SAE International by Erik Koehler, Wednesday, March 26, 2014 11:09:34 AM emissions and vice-versa. Thus, at mid loads as well, the amount of CNG substitution is dependent upon the trade-off between NOx and THC emissions. Figure 15. AHRR and cylinder pressure trace for 1500 rpm, 6 bar BMEP at 60% CNG substitution with three different injection timings Figure 16. Injection timing study for different CNG substitution levels at 1500 rpm, 14 bar BMEP with 38% EGR Figure 17 shows the impact of injection timing at three different substitution levels for 24 bar BMEP. Contrary to the approach at the lower loads, the injection timing was retarded at high load to allow for higher CNG substitution levels. By retarding the injection timing, the mechanical limits of maximum cylinder pressure were avoided. As the injection timing was retarded with increasing CNG substitution levels, the NOx and PM emissions were decreasing with a familiar increase in THC emissions. Comparing the AHRR traces in Figure 18, it is observed that as the CNG substitution is increased the peak of the AHRR trace moves away from TDC, matching the earlier fnding of a longer burn duration with less reactive CNG leading to reduced NOx formation. At a fxed substitution level, a retarded timing reduced the NOx emission as expected, but the THC emissions increased. The ability to further retard the injection timing is limited beyond a certain crank angle since the longer CNG burn duration leads to high THC emissions. Figure 17. Injection timing study for different CNG substitution levels at 1500 rpm, 24 bar BMEP with 28% EGR Based on the main timing and CNG substitution studies conducted at the 13 speed and load points, a map was generated showing the maximum CNG substitution that is possible when allowing changes to the base calibration. The map is again determined by observing the limits imposed by the US2010 emission legislation. Downloaded from SAE International by Erik Koehler, Wednesday, March 26, 2014 11:09:34 AM As shown in Figure 19 the average substitution achieved is approximately 50%, with higher substitution achieved at the higher load points where the retarded injection timing strategy was employed to control the pressure rise rates. With this substitution level, there was a penalty in the CH4 emissions, but these emissions are not regulated by the US2010 emission legislation, and therefore, no limit was imposed on the CH4 levels while generating the optimized map. At low loads the substitution is mainly limited by the unburned hydrocarbons due to a limited diesel quantity available for igniting the premixed CNG charge. One possible approach to overcoming these limitations is to change from conventional diesel combustion to reactivity controlled combustion. RCCI is a dual fuel strategy in which a high reactivity fuel (diesel) is injected early enough to allow complete mixing with a low reactivity fuel (CNG) to create a diverse reactivity map within the chamber, allowing for controlled and complete auto ignition of the premixed charge. The RCCI combustion strategy not only allows higher gas substitution levels but also has the potential to simultaneously reduce NOx and PM emissions. Figure 18. AHRR and cylinder pressure trace for 1500 rpm, 24 bar BMEP with three different substitution levels Figure 19. Optimized CNG substitution map with calibration changes and meeting US2010 emission standards Investigation of RCCI Combustion Figure 20. AHRR and cylinder pressure traces for three different main injection timings at 60% substitution (6 bar BMEP at 1500rpm) The existence of very high hydrocarbon emissions at low loads impeded higher substitution levels during the optimization effort. As a step toward further optimization, it was considered that the higher fuel reactivity difference between CNG and diesel could potentially enable RCCI combustion, despite a Downloaded from SAE International by Erik Koehler, Wednesday, March 26, 2014 11:09:34 AM relatively high compression ratio. To explore this concept, the main injection timing for diesel was advanced well beyond conventional diesel injection timings at a low load point. Figures 20 and 21 show the AHRR and the corresponding emissions at four different main injection timings and a 60% CNG substitution level. Up to a main injection timing of 32 deg bTDC, the peak pressures and temperatures continue to increase, leading to high levels of NOx formation. THC, CO and smoke were reduced due to an increased peak combustion temperature. Figure 21. Emissions and performance variables at 1500 rpm and 6 bar BMEP, plotted against main injection timing However, with further advancement in timing, beyond 32 deg bTDC, the AHRR trace moved closer to TDC. This resulted in lower peak combustion temperatures and pressures causing a sharp reduction in NOx emissions. The heat release trace is wider with a lower peak and clearly visible two-stage heat release that includes a cool fame region. The heat release profle closely resembled that of RCCI combustion [11], and it was concluded that the spatial fuel reactivity gradient led to the controlled heat release [11]. Although the reactivity gradient governed the heat release rate, EGR was necessary to control the combustion phasing for this high compression ratio engine. At 1500 rpm and 6.0 bar BMEP, the maximum possible EGR level was 37%. This EGR level was held constant for each injection timing investigated at this operating point. The NOx emissions at the most advanced injection timing of 65.2 bTDC was below the US2010 emission legislation, indicating the potential of this concept in eliminating NOx aftertreatment. A summary of the operating condition and observed emissions at the injection timing of 65.2 bTDC has been provided in Table 2. Table 2. Operating conditions and performance for RCCI at 6 bar BMEP and 1500 rpm It could be shown that with further advanced timings of up to 80 deg bTDC the combustion phasing did not exhibit further change. However, an increase in CO emissions was observed with further timing advancement, possibly due to the impingement of liquid fuel on the cylinder walls. Figure 22 shows an optimized CNG substitution map when RCCI combustion was included at the low load points. The average CNG substitution was improved from approximately 50% to above 65% for the complete engine map. Figure 23 shows the cost savings based on the optimized CNG substitution map outlined in Figure 22. A maximum cost savings of up to 40% can be achieved with advanced combustion concepts tested at the lower speed and load points. As outlined in Table 3, operation in dual fuel mode (that makes use of RCCI at low loads) allows, on average, a 43% reduction in NOx emissions, a 68% reduction in PM emissions and a 22% reduction in fuel costs. Downloaded from SAE International by Erik Koehler, Wednesday, March 26, 2014 11:09:34 AM Additional work is ongoing to identify the factors that contribute toward achieving RCCI combustion at different speed and load points. Results from these investigations will be further explored and discussed in a future publication. Figure 22. Optimized CNG substitution map with calibration changes, including RCCI combustion at low loads and meeting US2010 emission standards Figure 23. Cost savings in percentage from baseline diesel operation Table 3. Average cost savings and emissions reduction in percent, obtained from the optimized CNG substitution map that includes RCCI combustion at low loads Conclusion 1. Assuming no access to the base engine calibration, a baseline maximum substitution map with practical emissions was created. It was found that the cost and emission benefts of running on CNG were limited at lower loads due to high hydrocarbon emissions and at higher loads due to peak cylinder pressures and turbine inlet temperatures. 2. In an optimization step, strategies were developed primarily involving diesel injection timing change and CNG substitution level. Advancing the diesel injection timing at lower loads led to a reduction in THC emissions. However, at higher loads, slightly retarded injection timing helped in achieving higher CNG substitution. The optimized substitution map achieved an average of 49% substitution across the test point map with 15% reduction in NOx and 43% reduction in PM emissions. At higher engine speeds, it appears the reduced time for CNG/air mixing results in a varying cylinder-to-cylinder distribution of CNG, and thus deteriorated combustion stability. 3. RCCI combustion could be achieved at low loads, which enabled even higher CNG substitution and lower emissions. A maximum of 50% net indicated thermal effciency was observed at 6 bar BMEP load point along with 75% reduction in both NOx and PM emission. The potential benefts of RCCI combustion were limited due to the un-optimized combustion chamber design and high compression ratio. Future Work This work shows that an RCCI combustion strategy for burning CNG and diesel fuel is most promising in terms of NOx emission control and fuel consumption reduction. However, challenges remain that must be resolved to allow extended use of RCCI combustion at higher loads. Studies are currently ongoing to explore the factors that affect the ability to achieve RCCI combustion at different speed and load points. Along with calibration optimization, changes in engine design, including piston bowl design, compression ratio and nozzle confguration, are being considered. Port fuel injection of CNG in an effort to reduce the cylinder-to-cylinder variation observed with RCCI is also being investigated. Solutions enabling RCCI combustion and maximizing CNG substitution in dual fuel mode must also respect the option for diesel-only operation when CNG is not available. References 1. US Energy Information Administration, Natural Gas, http:// www.eia.gov/naturalgas/, Aug 2013 2. Ribas X., Heavy Duty liquefied Natural Gas engine developments to meet future emissions requirements, methodology and real application, FISITA paper F2010F013, 2010. 3. Ouellette, P., Cummins Westport Spark-Ignited (SI) and High Pressure Direct Injection (HPDI) Natural Gas Engines, presented at NGVTF 2003, USA, January 28-29, 2003. 4. Barroso Guzman, P., Ribas, X., Garca, J.Sr., and Pita, M.Sr., PM and CO2 Reduction in a Dual-fuel Heavy- duty Diesel Engine during the Freeway Part of Transient Worldwide Emission Tests, SAE Technical Paper 2013-01- 2759, 2013. 5. Maxey, C., Kalaskar, V., Kang, D., and Boehman, A., Impact of Supplemental Natural Gas on Engine Efficiency, Performance, and Emissions, SAE Technical Paper 2013- 01-0847, 2013, doi:10.4271/2013-01-0847. 6. Sun, L., liu, Y., Zhou, L., and Zeng, K., Experimental Investigation of Cycle-by-Cycle Variations in a Natural Gas/ Diesel Dual Fuel Engine with EGR, SAE Technical Paper 2013-01-0853, 2013, doi:10.4271/2013-01-0853. 7. Aroonsrisopon, T., Salad, M., Wirojsakunchai, E., Wannatong, K. et al., Injection Strategies for Operational Downloaded from SAE International by Erik Koehler, Wednesday, March 26, 2014 11:09:34 AM Improvement of Diesel Dual Fuel Engines under Low Load Conditions, SAE Technical Paper 2009-01-1855, 2009, doi:10.4271/2009-01-1855. 8. Kowalewicz A., and Woloszyn R., Comparison of performance end emissions of turbocharged CI engine fuelled either with diesel fuel or CNG and diesel fuel, Combustion Engines, PTNSS-2011-SC-117, 2011. 9. Papagiannakis, R., Hountalas, D., and Kotsiopoulos, P., Experimental and Theoretical Analysis of the Combustion and Pollutants Formation Mechanisms in Dual Fuel DI Diesel Engines, SAE Technical Paper 2005-01-1726, 2005, doi:10.4271/2005-01-1726. 10. Papagiannakis, R., Hountalas, D., Rakopoulos, C., and Rakopoulos, D., Combustion and Performance Characteristics of a DI Diesel Engine Operating from Low to High Natural Gas Supplement Ratios at Various Operating Conditions, SAE Technical Paper 2008-01-1392, 2008, doi:10.4271/2008-01-1392. 11. Nieman, D., Dempsey, A., and Reitz, R., Heavy-Duty RCCI Operation Using Natural Gas and Diesel, SAE Int. J. Engines 5(2):270-285, 2012, doi:10.4271/2012-01-0379. Contact Information Koehler, Erik Manager, Performance and Emissions Commercial Engines; FEV, Inc. Koehler@fev.com Dahodwala, Mufaddel Senior Engineer, Performance and Emissions Commercial Engines; FEV, Inc. Dahodwala@fev.com Acknowledgments The authors would like to thank FEV, Inc. management for encouraging this research effort and providing the resources necessary to accomplish its goals. The authors would also like to thank those who took the time to review this paper and provided valuable feedback. Abbreviations AC - Alternating Current AHRR - Apparent Heat Release Rate BMEP - Brake Mean Effective Pressure BSC - Brake Specifc Cost BSFC - Brake Specifc Fuel Consumption bTDC - Before Top Dead Center CH4 - Methane CNG - Compressed Natural Gas CO - Carbon Monoxide CO2 - Carbon Dioxide COV - Coeffcient of Variance DOC - Diesel Oxidation Catalyst DPF - Diesel Particulate Filter ECU - Engine Control Unit EGR - Exhaust Gas Recirculation EIA - Energy Information Administration EOL - End Of Life FTP - Federal Test Procedure GGE - Gasoline Gallon Equivalent HD - Heavy-Duty IMEP - Indicated Mean Effective Pressure MIRA - Motor Industry Research Association NMHC - Non-Methane Hydrocarbon NOx - Nitrogen Oxides OBD - On Board Diagnostics PM - Particulate Matter RCCI - Reactivity Controlled Compression Ignition RCP - Rapid Controller Prototyping RMC - Ramped Modal Cycle SCR - Selective Catalytic Reduction SI - Spark Ignited TDC - Top Dead Center THC - Total Hydrocarbons WHTC - World Harmonized Transient Cycle Downloaded from SAE International by Erik Koehler, Wednesday, March 26, 2014 11:09:34 AM APPENDIX CNG SUBSTITUTION CALCULATION (1) Where, = Mass fow rate of CNG QLHV cng = Lower heating value of CNG = Mass fow rate of diesel BRAKE SPECIFIC COST CALCULATION Brake specifc cost is a parameter that accounts for the amount of cost savings that one can get through CNG substitution. The average national cost of CNG (2.09$/GGE) and Diesel (3.84$/Gallon) in May 2013 were used for this calculation. (2) Where, C cng = Average National Cost of CNG in $/Kg (1 GGE CNG = 2.567 Kg CNG) C diesel = Average National Cost of Diesel in $/Kg (1 US Gallon Diesel = 3.149 Kg Diesel) P = Engine Power (W) CNG SUBSTITUTION OPTIMIZER Optimum CNG substitution at different operating points was calculated using the composite desirability index for multiple response optimization. The target, maximum values and weight factors were selected for NOx, PM, NMHC, CO, BSC and ringing intensity according to regulation limits. Desirability can be calculated for individual responses by the following formula: (3) The composite desirability can then be expressed as: (4) Where, Y = Response (NOx, PM, BSCetc) w = Weighing factor n = Number of Responses Downloaded from SAE International by Erik Koehler, Wednesday, March 26, 2014 11:09:34 AM CD = Composite Desirability Index d = Desirability The Engineering Meetings Board has approved this paper for publication. It has successfully completed SAEs peer review process under the supervision of the session organizer. The process requires a minimum of three (3) reviews by industry experts. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior written permission of SAE International. Positions and opinions advanced in this paper are those of the author(s) and not necessarily those of SAE International. The author is solely responsible for the content of the paper. ISSN 0148-7191 http://papers.sae.org/2014-01-1308 Downloaded from SAE International by Erik Koehler, Wednesday, March 26, 2014 11:09:34 AM