Available online at www.sciencedirect.com ScienceDirect Energy Procedia 101 (2016) 886 – 892 71st Conference of the Italian Thermal Machines Engineering Association, ATI2016, 14-16 September 2016, Turin, Italy Natural Gas Stable Combustion under Ultra-Lean Operating Conditions in Internal Combustion Engines L Bartoluccia,*, S.Cordinera, V.Mulonea, V.Roccoa a University of Tor Vergata, via del Politecnico 1, Rome, 00133, Italy Abstract Ultra-Lean operating conditions have a great potential toward the increase of engine thermal efficiency and control of pollutant emissions in Internal Combustion Engines; however, cyclic variability and stability are still critical issues to be solved The use of natural gas may partially solve these issues due to a more efficient mixing process, widening the flammability limits An analysis of natural gas ultra-lean combustion process is presented in this work using local charge stratification to stabilize the process In particular, the predictive capabilities of the numerical model are here validated up to ultra-lean (λ=2.0) mixtures Moreover, an analysis of the ignition timing has been performed, to find the optimal trade-off between flame quenching and turbulent kinetic energy decay as produced by the jet in the vicinity of the spark location The optimal ignition timing has been calculated in the range of 5ms © Published by Elsevier Ltd This ©2016 2016The TheAuthors Authors Published by Elsevier Ltd is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/) Peer-review under responsibility of the Scientific Committee of ATI 2016 Peer-review under responsibility of the Scientific Committee of ATI 2016 Keywords: Partially Stratified Charge Spark Ignition; Turbulent Combustion; Inhomogeneous Combustion; Injection and Combustion Modelling Introduction The automotive industry is currently making strong efforts to move the engine combustion control strategies towards leaner and leaner mixtures to fulfill the emission regulations Several works have been done to study the lean combustion, (in particular Natural Gas (NG) lean mixtures), both in Constant Volume Combustion Chambers (CVCC) [1, 2, 3] and engines [4, 5, 6] _ * Corresponding author Tel.: +39 0672597125 E-mail address: lorenzo.bartolucci@uniroma2.it 1876-6102 © 2016 The Authors Published by Elsevier Ltd This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/) Peer-review under responsibility of the Scientific Committee of ATI 2016 doi:10.1016/j.egypro.2016.11.112 L Bartolucci et al / Energy Procedia 101 (2016) 886 – 892 This is mainly due to the flame temperature reduction with related control of NOx emissions and heat losses giving an increase of thermal efficiency [7, 8] Moreover, combustion control with air-fuel ratio allows for reducing pumping losses at partial load However, several issues related to combustion stability and the cyclic variability arise [9] Prof Evans group at the University of British Columbia (UBC) developed a special combustion strategy, The Partially Stratified Charge (PSC) Combustion [10, 11], to try and solve such issues The concept aims at stabilizing the combustion process through the injection of a small amount of fuel in the vicinity of the spark plug Experimental and numerical studies have already proofed that once the kernel grows in a mixture rich locally, the subsequent flame propagation evolves much more smoothly giving an extension of the flammability limit [12, 13, 14] Previous studies based on the Large Eddy Simulation (LES) modeling approach have shown the key role of turbulence, enhancing the mixing process of the post-injected fuel with the surrounding lean mixture and sustaining the flame propagation after the kernel expansion Other authors have shown this strong correlation between turbulence and flame propagation in other combustion concepts using a DNS approach [15] In this paper a deeper analysis of the turbulence-flame propagation interaction is shown, with several simulations at different ignition timings –aiming at highlighting the role of turbulent kinetic energy level around the spark plug location on the duration of the combustion process Nomenclature λ AS ASOI CVCC PSC RoHR air-fuel equivalent ratio After Spark After Start Of Injection Constant Volume Combustion Chamber Partially Stratified Charge Rate of Heat Release CVCC Configuration and Experimental Operating Conditions An optically accessible Constant Volume Combustion Chamber (CVCC) designed and realized at the University of British Columbia [11], was used for the validation of the numerical model Details of the experimental setup of the CVCC can be found in [16] The injection pressure was fixed at 10 bar (abs.), while the chamber pressure was maintained at bar (abs.) to represent engine like pressure conditions at the beginning of the PSC injection The injection duration was set to 9.9 ms, thus fixing the mass percentage of PSC injected charge as a function of the global λ The spark timing was imposed at the End Of the Injection (EOI) Homogeneous cases have also been simulated for same global λ to have a baseline reference 3D Numerical Model The OpenFOAM [18] based code used in this paper is a CFD solver for compressible, chemically reacting, transient, three-dimensional flows A One Equation Eddy LES approach has been used to model the turbulent flow and mixing, while a PArtially Stirred Reaction (PASR) combustion model has been selected to represent the partially premixed combustion process, taking then into account the interaction between turbulence and chemistry, which is mandatory for the correct description of lean combustion regimes [12][19] A Spark-Energy Deposition Model has been implemented to deal with mixture ignition Detailed descriptions of the turbulence closure and the combustion models can be found in previous papers [12, 17] 3.1 Ignition Source Based on a work proposed by Maly et al [20], the energy source has been modeled as a Capacitor DIscharge (CDI) of mJ during 100 Ps plus a constant arc value of 30 mJ during 230 Ps This energy release profile, although not completely realistic (as it is not able to take into account the breakdown phase), allowed to obtain a reasonable evolution of the kernel during its growth, as will be shown in the results 887 888 L Bartolucci et al / Energy Procedia 101 (2016) 886 – 892 In particular, the model, as well as the flame propagation in the earliest stage of the combustion process have correctly captured the flammability limit of the mixture; thus proving the reliability of the assumptions of the ignition model Fig Ignition Profile Analysis of results A comparison between numerical and experimental results is shown to assess the capabilities of the solver to capture flame propagation by varying operating conditions, in terms of relative equivalent ratios and combustion strategy Results have been presented in terms of density gradient along the light direction, which is proportional to the change in the refraction index and hence representative of the parameters acquired experimentally with the Schlieren visualization experimental technique In Figure 3, the comparison between the experimental Schlieren (first column) density gradient, the numerical density gradient contour (second column) and a superimposition of the numerical isothermal line at 900K with the experimental Schlieren (third column) is given for three global air fuel ratios Results show how the solver is able to predict accurately the flame front location for both the homogenous and the PSC combustion cases for all the λ tested The isothermal line at 900K, which can be assumed to represent the flame front location, matches very well the experimental data It is also worth noting that the numerical model provides a reliable description of the flame front shape The flame propagation, shown in Figure 3, is well described from the very beginning of the process and, in particular, the evolution of the energy source allows for capturing according with results obtained experimentally - the quenching of the flame due to the heat losses at the electrode for homogeneous cases and λ of 1.8 and 2.0 (Figure 4) The value of 1.8 represents in fact the boundary between successful and unsuccessful cases in agreement also with the theory [21] The flame front location is predicted accurately also in the PSC cases for all the air fuel ratios considered (Figure 5) Figure 5, similarly to Figure 3, shows how the solver is able to predict the propagation of the flame front during the combustion process for the whole set of air fuel ratios (between 1.2 and 2.0) The flame plum is elongated mainly due to the interaction of the flame with the PSC jet, and the flame front is highly corrugated by the jet structures The PSC combustion cases shows a much faster flame propagation if compared with the homogeneous cases This is mainly due to the stratification of the charge around the spark plug and to the higher turbulent kinetic energy content provided by the interaction of the jet with the surrounding mixture However, at EOI, the latter contribution is highly reduced (Figure 6(b)), and a specific analysis has been carried out to evaluate its effects on flame propagation The spark timing has been changed, with respect to the Start Of the Injection (SOI) different timings have been studied, ranging from 2ms after SOI to 9ms after SOI to try and find an optimum between quenching due to the high velocity of the jet close to the SOI, and the decay of turbulent kinetic energy around the spark plug (Figure 6(a,b)) 889 L Bartolucci et al / Energy Procedia 101 (2016) 886 – 892 λ = 1.2 λ = 1.4 λ = 1.6 Fig Comparison between the experimental Schlieren (first column) and density gradient along the light direction (second column) for the homogeneous cases Over-position of the isothermal line at 900K with the experimental Schlieren (third column) for the homogeneous cases λ = 2.0 λ = 1.8 Figure Misfiring due to the heat losses for the homogenous cases at λ = 1.8 and 2.0 The analysis allows for isolating the effects due to turbulence evolution since, as can be seen in Figure 6(c), the local air fuel ratio close to the spark plug is very stable by changing spark timing λ = 1.2 λ = 1.8 λ = 1.6 λ = 2.0 Fig Comparison between the experimental Schlieren (first column) and density gradient along the light direction (second column) for the homogeneous cases Superimposition of the isothermal line at 900K with the experimental Schlieren (third column) for the PSC cases 890 L Bartolucci et al / Energy Procedia 101 (2016) 886 – 892 Fig Instantaneous z-velocity component (a), turbulent kinetic energy (b) and methane mass fraction (c) close to the spark plug As expected, results show how, for early spark timings, the high kinetic energy of the jet quenched the kernel before reaching stability (Figure 7(a)) However, once the kinetic energy is decayed to an optimal level, the best compromise is obtained in the overall set (Figure 7(b-c)) This highlights the enhancing role of turbulence in the PSC combustion process Greater spark timings led to a slower flame propagation and thus poor combustion performances (Figure 7(b-c) – 8(a,b,c)) Figure 8(c) shows how, once all the PSC charge is burned, the combustion process slowed down (as evident by the change of slope in the RoHR curve for 5ms ASOI spark timing), still propagating smoothly thanks to the stable formation and growth of the kernel in the richer region Figure 7(b) shows also the different occurring phenomena due to the change of spark timing In particular, it can be seen that 5ms after SOI ignition timing case is characterized by several ignition spots close to the spark plug induced by the jet On the other hand, later ignition timings are characterized by a more localized and continuous kernel This difference has contributed to the occurrence of faster propagation of the 5ms ASOI ignition timing case if compare to all the other ones Fig Temperature 2D maps of misfiring PSC combustion case for early ignition timings (a), Temperature 2D maps of successful PSC combustion cases for later timings at 1ms after the Spark (b) and 5ms after the Spark (c) L Bartolucci et al / Energy Procedia 101 (2016) 886 – 892 Fig Mean Temperature (a), relative pressure (b) and RoHR (c) profiles for all the ignition timing tested Conclusions A 3D CFD analysis of methane combustion process in a CVCC has been proposed in the current work Numerical results have been compared with experimental data gathered at the University of British Columbia to assess the reliability of the model toward the prediction of the flame front propagation Results obtained allowed for stating the following main conclusions: x The solver showed very good capabilities in predicting the location of the flame front for all the operating conditions simulated, highlighting the benefits due to the adoption of the PSC combustion strategies in extending the flammability limit of the methane-air mixtures and providing faster and more stable combustion processes compared with the homogeneous cases with the same λ value x Those benefits are mainly related to two different aspects: - The stratification of the charge around the spark location; - The enhancement of the mixing process at the molecular scales due to the high turbulent kinetic energy level provided by the interaction between the PSC jet and the surrounding mixture x A detailed analysis of the spark timing: switching from 2ms after the SOI to 9ms after the SOI, the distribution of λ around the spark plug was almost constant over time during the mixing process, while a consistent decay of turbulent kinetic energy level was observed affecting the speed of the combustion process x Results showed a trade-off between kernel quenching, turbulent heat losses and decay of turbulent kinetic energy The optimal timing has been found in the order of 4-5ms after the SOI Greater timings give slower flame propagation speed due mainly to turbulent kinetic energy decay Further work will focus on implementing the model in a real engine to evaluate the effects of the actual PSC jet operating conditions, including piston motion, real chamber geometry and spark-plug, etc 891 892 L Bartolucci et al / Energy Procedia 101 (2016) 886 – 892 References [1] Zhang, A., Scarcelli, R., Lee, S., Wallner, T et al., "Numerical Investigation of Spark Ignition Events in Lean and Dilute Methane/Air Mixtures Using a Detailed Energy Deposition Model," SAE Technical Paper 2016-01-0609, 2016, doi:10.4271/2016-01-0609 [2] Askari O., Metghalchi H., Hannani S K., Hemmati H et al., “Lean Partially Premixed Combustion Investigation of Methane Direct-Injection Under Different Characteristic Parameters”, Journal of Energy Resources Technology, JUNE 2014, Vol 136, doi:10.1115/1.4026204 [3] Zhang, Anqi, et al "The impact of spark discharge pattern on flame initiation in a turbulent lean and dilute mixture in a pressurized combustion vessel." SAE International Journal of Engines 6.1 (2013): 435-446 [4] Sevik, J., Pamminger, M., Wallner, T., Scarcelli, R et al., "Performance, Efficiency and Emissions Assessment of Natural Gas Direct Injection compared to Gasoline and Natural Gas Port-Fuel Injection in an Automotive Engine," SAE Int J Engines 9(2):2016, doi:10.4271/201601-0806 [5] Melaika, M and Dahlander, P., "Experimental Investigation of Methane Direct Injection with Stratified Charge Combustion in Optical SI Single Cylinder Engine," SAE Technical Paper 2016-01-0797, 2016, doi:10.4271/2016-01-0797 [6] Park, Cheolwoong, et al "Stratified lean combustion characteristics of a spray-guided combustion system in a gasoline direct injection engine." Energy 41.1 (2012): 401-407 [7] John B Heywood, Internal Combustion Engine Fundamentals, McGraw-Hill, 1988 [8] CHO, Haeng Muk; HE, Bang-Quan Spark ignition natural gas engines—A review Energy Conversion and Management, 2007, 48.2: 608618 [9] Scarcelli, R., Matthias, N., and Wallner, T., Numerical Investigation of Combustion in a Lean Burn Gasoline Engine, SAE Technical Paper 2013-24-0029, 2013, doi:10.4271/2013-24-0029 [10]Reynolds C.C.o’B., and Evans R.L., Improving emissions and performance characteristics of lean burn natural gas engines through partial stratification, International Journal of Engine Research, 5(1):105-114, 2004 [11]Chan, E.C., Chicka, A.S., and Evans R.L., Spark ignition of partially stratified natural gas mixtures, Submitted to Combustion Science and Technology, 2015 [12] Bartolucci, L; Chan, EC; Cordiner, S; Mulone, V; Rocco, V; ,Natural Gas Fueling: A LES Based Injection and Combustion Modeling for Partially Stratified Engines,Energy Procedia,82,,417-423,2015,Elsevier [13] Bartolucci, L., Cordiner, S., Mulone, V., Rocco, V et al., "Partially Stratified Charge Natural Gas Combustion: A LES Numerical Analysis," SAE Technical Paper 2015-01-0398, 2015, doi:10.4271/2015-01-0398 [14] Bartolucci, L., Cordiner, S., Mulone, V., Rocco, V et al., "Natural Gas Partially Stratified Charge Combustion: Extended Analysis of Experimental Validation and Study of Turbulence Impact on Flame Propagation," SAE Technical Paper 2016-01-0596, 2016, doi:10.4271/201601-0596 [15] Sankaran, Ramanan, et al "Response of flame thickness and propagation speed under intense turbulence in spatially developing lean premixed methane–air jet flames." Combustion and Flame 162.9 (2015): 3294-3306 [16] Chan, E.C “Spark ignition of partially stratified natural gas mixtures,” Doctoral Thesis, The University of British Columbia, 2010 [17] Bartolucci, L., Cordiner, S., Mulone, V., Rocco, V et al., "Partially Stratified Charge Natural Gas Combustion: The Impact of Uncertainties on LES Modeling," SAE Technical Paper 2015-24-2409, 2015, doi:10.4271/2015-24-2409 [18] www.openfoam.com [Online] [19]Aspden, A J., M S Day, and J B Bell "Turbulence–flame interactions in lean premixed hydrogen: transition to the distributed burning regime." Journal of Fluid mechanics 680 (2011): 287-320 [20] Maly, R., Vogel, M Initiation and Propagation of Flame Fronts in Lean CH4-Air Mixtures by the Three Modes of The Ignition Spark Proceedings of Seventeenth International Symposium on Combustion, The Combustion Institute, 1976 pp 821-831 [21] Kennet Kuan-yun Kuo, Principles of Combustion, John Wiley, 1986 Biography Lorenzo Bartolucci was born in Rome, Italy, in 1989 He graduated in Energy Engineering from the University of Rome Tor Vergata He is currently a Ph.D student in the same university at the department of Industrial Engineering His fields of interest are CFD modeling of turbulent gas injection and combustion ... Efficiency and Emissions Assessment of Natural Gas Direct Injection compared to Gasoline and Natural Gas Port-Fuel Injection in an Automotive Engine," SAE Int J Engines 9(2):2016, doi:10.4271/201601-0806... "Stratified lean combustion characteristics of a spray-guided combustion system in a gasoline direct injection engine." Energy 41.1 (2012): 401-407 [7] John B Heywood, Internal Combustion Engine Fundamentals,... due mainly to turbulent kinetic energy decay Further work will focus on implementing the model in a real engine to evaluate the effects of the actual PSC jet operating conditions, including piston