Numerical Analysis of a Six-Stroke Gasoline Compression Ignition (GCI) Engine Combustion with Continuously Variable Valve Duration (CVVD) Control Oudumbar Rajput 1, Youngchul Ra 1, Kyoung-Pyo Ha 2 1 Mechanical Engineering-Engineering Mechanics, Michigan Technological University 2 Automotive, Research & Development Division, Hyundai Motor Company 1. Abstract A six-stroke GCI engine combustion employing CVVD technique is investigated numerically. To understand the effects of additional two strokes of the engine cycle on the thermal and chemical conditions of charge mixtures, a wide range of valve control through CVVD technique was considered under both positive valve overlap (PVO) and negative valve overlap (NVO) conditions. Simulation results show that the additional two strokes of a 6-stroke GCI engine can effectively recover the energy that may be lost in 4-stroke cycle engines. In addition, the control of valve timings could successfully control the thermodynamic and compositional conditions of in-cylinder mixtures that enable to control the combustion phasing. 2. Introduction To increase the combustion efficiency and to reduce the emissions, a six-stroke engine technology can be used over the conventional four-stroke technology for lower loads [1]. The primary difference between four-stroke and six-stroke cycles is the addition of a second compression and expansion strokes (called 2 nd power stroke) following the first combustion event during the fiest compression and expansion strokes (called 1 st power stroke). In a six-stroke engine, the additional two strokes capture the waste heat of four-stroke-cycle engines, improve the fuel efficiency, and reduce emissions by using the excess oxygen present in the cylinder. The Continuously Variable Valve Duration (CVVD) technique is one of the Variable Valve Actuation (VVA) technologies for minimizing the pumping loss, enhancing the volumetric efficiency, decreasing the amount of the residual gas inside the combustion chamber, and controlling the expansion and compression work. A new CVVD system developed by Hyundai Motor Company (HMC) [2] allows the control of all the valve events independently, regardless of the engine type and engine operating conditions. The maximum benefit of the CVVD control system can be achieved when it is applied to a six-stroke cycle engines, by controlling thermodynamic conditions of the charge mixtures and their subsequent combustion to result in additional improvement of fuel consumption. The objective of this work is to computationally investigate the effects of valve timing strategies using HMC s CVVD technique on the physical/chemical conditions of charge mixtures and their combustion behavior in a gasoline engine running a six-stroke cycle. 3. Methodology 3.1. MTU in-house KIVA code For simulating the fuel/air mixture preparation and combustion in the cylinder, MTU-KIVA-Geq-CHEMKIN code, an in-house version of KIVA code [3, 4] is used. The code employs various physical sub-models for multi-component fuel spray and evaporation [5-7], drop breakup [8], collision and coalescence [3], drop deformation [9], and wall impingement. For modeling flame propagation, Discrete Particle Ignition Kernel (DPIK) model [11] and G-equation model [12] were employed. Volumetric reaction chemistry in computational cells is calculated using Chemkin library [4], MPI parallelization [13] and the SpeedChem model [14]. For the turbulence calculation, the RNG k-ε model [15] was used. Physical properties of gasoline are modeled using a 14-component gasoline surrogate fuel [16]. Whereas, the combustion and emissions kinetics of the fuel are modeled using Group Chemistry Representation (GCR) model [17] and skeletal reaction mechanisms for four chemical surrogate (CS) fuel components viz., isooctane, n-heptane, ethanol and butane. This reduced mechanism consists of 97 species and 384 reactions whose performance was well validated against experimental ignition delay times, HCCI engine combustion, spray combustion in a constant volume chamber. 3.2. CVVD technique Combining the benefits of the Continuously Variable Valve Timing (CVVT) and Continuously Variable Valve Lift (CVVL) technologies, HMC developed a CVVD mechanism that can change the valve duration, while maintaining the same maximum valve lift; the operable duration ranges from 108 to 320 degca, which is a significant extension from that of conventional technologies. All the four valve timings can be independently controlled irrespective of the engine type and the operating conditions. The examples of valve profile are shown in Figure 1. In Fig. 1, it is seen that the valve closure timing can be controlled without affecting the valve open timing while the maximum valve lift remains constant. Figure 1 shows the controllability over valve duration without disturbing the maximum valve lift. 1
Figure 1: Nominal Valve Lift Profile: With constant opening point With constant maximum opening point. 4. Engine Specifications and Operating Conditions Figure 2 shows the computational grid used for the simulations. The engine simulated has a pent-roof head and a flat piston bowl with four valves. Engine specifications and simulated operating conditions are shown in Table 1. The compression ratio of the engine is 12.5 and a multi-hole injector with an included spray angle of 120 degrees and nozzle diameter of 134 mm was employed. Injection pressure was 450 bar. Boosted engine operations with an intake pressure of 1.36 bar were simulated at an engine speed of 1300 rev/min. Table 1: Engine specifications and operating conditions Engine specification Displacement [L/cyl] 0.764 Bore x Stroke [mm x mm] 92 x 115 Compression ratio 12.5:1 Spray included angle [degree] 120 Number of Holes 6 Hole Diameter [mm] 0.134 Operating conditions Intake manifold pressure [bar] 1.36 Engine speed [rev/min] 1300 Swirl ratio 0 EGR [%] 0 Injection Pressure [bar] 450 Figure 2: A 3-D computational grid with 4-valves used in CFD simulations. A simulation cycle was from 120 to 1200 degca, i.e., from exhaust valve open (EVO) to EVO. In Table 2, valve timings of all simulation cases are listed. IVO/EVO stands for intake/exhaust valve open. IVC/EVC stands for intake/exhaust valve closure. In the case names, the +/- sign after IV or EV indicates retardation/advance and the numbers indicate changed crank angles, i.e., for example, (IV+8) indicates a case with the intake valve open retarded by eight degca maintaining the constant valve duration. The numbers following NVO indicate the duration of NVO period. The numbers following LIVC indicate retarded crank angle of IVC only with IVO fixed. Table 2: Valve timings of simulated cases. Valve timings are in degca Case name IVO IVC EVO EVC Baseline 352 580 134 360 IV+8 360 588 134 360 IV+15 367 595 134 360 IV+25 377 605 134 360 IV+30 382 610 134 360 IV+35 387 615 134 360 EV-10 352 580 124 350 EV-20 352 580 114 340 EV-25 352 580 109 335 NVO34 372 580 134 340 NVO54 382 580 134 328 NVO74 392 580 134 318 LIVC8 352 588 134 360 LIVC15 352 595 134 360 LIVC25 352 605 134 360 LIVC35 352 615 134 360 5. Results and Discussion 5.1. Six-stroke engine cycle Two simulation cases are run to understand a sixstroke engine cycle in detail, viz., SinglePulse (50% fuel 2
injection at 690 degca and skipping the second injection) and DoublePulse (50%/50% fuel injection at 690 and 1050 degca). Figure 3 shows the nominal valve lift profile of intake and exhaust valves for both the cases. The amount of fuel injected is 12.5mg at 450 bar for each injection event. Figure 3 shows the pressure and heat release rate (HRR) profiles and Fig. 3(c) shows the profiles of unburnt hydrocarbon (UHC) and carbon monoxide (CO) amounts in the cylinder. (c) Figure 3: Nominal valve lift profile with fuel injection timing details Pressure and heat release rate profiles (c) UHC and CO profiles From Figure 3, it is observed that additional heat release is made during the 2 nd power stroke, which indicates that the fuel injected during the 1 st pulse burns incompletely and keeps burning during the 2 nd power stroke. Indeed, the combustion efficiency of the 1 st power stroke at 842 degca (this corresponds to the EVO of the four-stroke cycle engine operation) was calculated to be 59.3 % and it becomes 74.5% at the end of the 2 nd power stroke with additional energy of 79 J released. The additional two strokes of a six-stroke cycle provide supplementary time for UHC and CO to further oxidize. In addition, the compression of gases during the 2 nd power stroke helps enhance the oxidation reactions. However, it is notable that the additional two strokes provide time for both oxidation and mixing of the partially burned gases. Due to decrease of equivalence ratio of local mixtures due to excessive mixing, the temperatures of burned gases are not raised high enough to oxidize CO completely. This results in decrease of UHC, but increase of CO emissions at the end of the 2 nd power stroke, as shown in Fig. 3(c). The 2 nd injection during the 2 nd power stroke helps form richer local mixtures to drive ignition and combustion of the partially burned gases. Due to this, in-cylinder gas pressures and heat release are significantly increased (Fig. 3) and UHC and CO emissions are decreased (Fig. 3(c)), reaching the combustion efficiency of 95.6%. More power is produced during the 2 nd power stroke, which makes it the main power stroke. However, the ignition time in the main power stroke is so early that the improvement of thermal efficiency is limited. The indicated thermal efficiency can be further improved by controlling the combustion phasing. The CVVD technology gives the freedom of controlling valve open duration while maintaining the maximum valve lift in a wide range (up to 320 degca). This enables to control the thermodynamic conditions and composition of in-cylinder mixtures by governing the amount of fresh air and residual exhaust gases, crank angle period for compression stroke etc., and thus ignition/combustion phasing and pressure rise rate can be altered, as well as emissions. Furthermore, combining with a six-stroke cycle engine, the CVVD approach gives wider flexibility and controllability, especially in low load engine operation. This allows an alternative to low load control techniques such as the cylinder deactivation, minimizing the pumping losses and ensuring a balanced (vibration less) engine operation. 5.2. Intake Valve (IV) timing variation with PVO As a way to retard the combustion phasing, simulations with retarded IV timing were performed. IV open duration was maintained constant and EV timings were unaltered. Due to retarding IVC, the effective compression ratio reduces, as seen in Figure 4. This reduces the incylinder gas temperature and the pressure at the start of the 1 st injection (SOI1). 3
Figure 4: Effective compression ratio, Ignition delay (1 st and 2 nd fuel injection event), Overall combustion efficiency, Temperature and UHC amount at BDC (900degCA), Pressure profiles during second power stroke Lower pressure affects the fuel atomization and lower temperature reduces the reactivity of the fuel/air mixtures delaying the ignition times (10% release point of injected fuel energy). When IV timing is retarded more than 15 degca, it is seen that combustion becomes so poor that apparent ignition is not observed (partial burning or misfire) during the 1 st power stroke. This misfire timing is extended to more retardation of IV that the charge mixture misfires in the case of IV+25 or with more retardation. Due to the same expiation period, the gas temperatures at 900 degca (BDC) are directly coupled to the compression temperatures and are reduced with increasing retardation of IVC. Similarly, the gas pressures at BDC are decreased with increasing IV retardation. On the contrary, the UHC at 900 degca increases with increasing IVC retardation. Hence, with increasing IV timing retardation, the gas mixtures are at lower temperatures and richer equivalence ratios at the start of 2 nd injection (SOI2). It is predicted that the ignition delay time in the 2 nd power stroke is slightly increased with IV retardation up to 15 degca, but it is substantially increased with additional 10 degca retardation, as shown in Fig. 4. This can be explained by the fact that IV+25 case misfires during the 1 st power stroke that the temperature and pressure at BDC are become much lower than the first three cases. Since the equivalence ratio of the gas mixtures at SOI2 becomes richer, this behavior indicates that the ignition timing of the main power stroke is rather governed by the thermal conditions than the composition effect. With retardation of ignition time towards TDC, the compression work is reduced and the thermal efficiency is improved from 35.2% (baseline) to 39.3% (IV+25). With further retardation of IV timing, the compression temperature reduction in the 2 nd power stroke reduces mixture reactivity too much to make the mixtures ignite within an optimal crank angle window. Near this point mixture ignition becomes so sensitive to IV timing change that the charge mixtures misfire with even additional 5 degca retardation. 5.3. Exhaust Valve (EV) Timing Variation Exhaust valve (EV) timing variation with constant EV open duration was also simulated. With fixed IV timings, advance of EVC reduces PVO duration, eventually leading to NVO conditions. With earlier EVO, in-cylinder gases expand at higher pressures and temperatures to the exhaust manifold pressure. The process can be regarded as an isentropic expansion process, which typically has a larger polytropic coefficient than the expansion stroke before EVO, and thus the final mixture temperatures tend to be slightly lower at the end of expansion. This temperature difference stays quite the same during the subsequent exhaust process until EVC, which is also advanced. With fixed IVO at 352 degca, earlier EVC tend to increase the trapped residual gases to be mixed with the fresh air during the intake process. Thus advancing EV timing has two competing thermal effects; increase of residual gas amount tends to increase the gas temperature at IVC, while lowered expansion temperatures tend to decrease it. It is predicted that the competing effects cancel off each other so that the gas temperatures do not vary much while PVO is maintained. However, as valve timings fall into NVO conditions, IVC gas temperatures are predicted to increase with advancing EV timing (increasing NVO) because the residual gases are compressed during NVO and the temperatures at IVO become higher. Higher residual gas temperatures and larger mass tend to increase the gas mixture temperatures at IVC, which results in advancing ignition timings in the 1 st power stroke. Figure 5 shows pressure and HRR profiles during the 2 nd power stroke, ignition delays and combustion efficiencies. It is seen that the ignition timings in the 2 nd power stroke are quite early and do not change much 4
with the NVO variation, which is attributed to improved combustion during the 1 st power stroke. the 2 nd power stroke are significantly early and do not vary much with NVO variation, as shown in the figure. It indicates that the higher burned gas temperatures from better combustion during the 1 st power stroke tends to enhance reaction rates of the 2 nd injection fuel, while the increased combustion product amounts at SOI2 have reactionsuppressing effects. It seems that these effects cancel off each other, resulting similar ignition delay times. As expected from early ignition in the 2 nd power stroke, all test cases showed combustion efficiency comparable to that of the baseline case, with the NVO30 case showing the highest. Figure 5: Pressure (solid line) and Heat Release Rate (dashed line) Profile for 2 nd power stroke Ignition delay (1 st and 2 nd fuel injection event) and Overall combustion efficiency (%). 5.4. Negative Valve Overlap (NVO) with IVO and EVC change Further investigation of NVO effects were performed. Here NVO variation was obtained by both advancing EVC and retarding IVO timings (refer to Table 2 for detailed timings). As mentioned above, with earlier EVC, the gases in the closed volume are compressed, raising the pressure and temperature of the residual gases higher at TDC, and more residual mass is trapped. However, with later IVO (note that IVO tested are after TDC), the residual gases expand again and the gas temperatures at IVO are significantly reduced, cancelling the longer compression effect to some extent. It is predicted that the gas temperature and pressure at IVO are increased with longer NVO periods. With the same IVC (i.e., effective compression ratio), the thermal conditions and compositions of the gas mixtures at IVC are the main factors to determine the combustion phasing. Hence, ignition delay in the 1 st power stroke is shortened with increasing NVO periods, as shown in Fig. 6. Advanced 1 st ignition leads to better combustion efficiency of the 1 st injection fuel, but more heat loss due to increased burned gas temperature and longer heat transfer time at high temperatures. Interestingly the ignition timings in Figure 6: Ignition delay period (1 st and 2 nd fuel injection event) and Overall combustion efficiency (%). 5.5. Late Injection Valve Closure (LIVC) This section discusses the effects of retarding IVC only with fixed IVO. Compared to the case IV timing variation discussed above, the effects of intake duration can be investigated. With IVC retardation, it is predicted that the gas temperatures at IVC have minimal increase, as shown in Fig. 7. This indicates that the gas mass in the cylinder during the compression stroke is proportionally reduced with decreasing cylinder volume at IVC, cancelling the likely improvement of the volumetric efficiency. With reduced in-cylinder air mass, the overall equivalence ratio of the mixtures formed by the 1 st injection is increased, which has reactivity-enhancing effects. As discussed in IV timing variation, however, the reduced effective compression ratios tend to reduce the compression temperatures of the gas mixtures. As shown in Fig. 7, the behavior of ignition delays and combustion efficiency is very similar to that of IV timing variation (see Fig. 4). Both the 1 st and 2 nd ignition delays are increased with increasing IVC retardation and combustion efficiency drops with more retardation than a threshold value, resulting in misfire. This implies that intake interval has not significant effects on combustion phasing and it is mainly controlled by IVC timings through thermal effects. 5
Figure 7: Temperature profiles after IVC Ignition delays of 1 st and 2 nd power strokes and overall combustion efficiency. 6. Summary and Conclusions A six stroke-cycle engine combustion was simulated and analysis was done in order to understand the effects of additional two strokes on engine performance. It was found that the six-stroke operation successfully recovers the heat energy improving the combustion performance and reducing UHC emission. As an effective way to drive better combustion of the lean unburnt mixture of the 1 st injection fuel, additional fuel injection during the second compression stroke was employed. Double injections were found to be effective to utilize the additional two stokes for the combustion of overly mixed lean charge mixtures during the main power stroke. Parametric simulations were performed to investigate the effects of valve timing strategies on the physical/chemical conditions of charge mixtures and their combustion behavior. It was found that the IVC timing controls combustion phasing in both power strokes since it affects the thermal conditions of the gas mixture at SOI through effective compression ratio variation. Creating NVO with fixed EVO and IVC timings tends to advance SOI of the first power stroke, but has minimal effect on second SOI. The operation of a six-stroke GCI engine could be successfully simulated and the operability range of the engine could be substantially extended by employing the CVVD technique. Reference 1) Arai, M., Amagai, K., and Ida, Y., "New Concept for Six- Stroke Diesel Engine," SAE Technical Paper 941922, 1994, https://doi.org/10.4271/941922. 2) Kyoung-Pyo Ha, Woo Tae Kim, In Sang Ryu, You Sang Son, Development of Continuously Variable Valve Duration (CVVD) Engine, 25th Aachen Colloquium Automobile and Engine Technology, 2016 3) Amsden, A.A., KIVA-3V, Release 2, Improvements to KIVA-3V, LA-UR-99-915, 1999. 4) Kee, R.J., Rupley, F.M., Miller, J.A., CHEMKIN-II: A FORTRAN Chemical Kinetics Package for the Analysis of Gas Phase Chemical Kinetics, Sandia Report SAND 89-8009, 1989. 5) Ra, Y. and Reitz, R.D., A vaporization model for discrete multi-component fuel sprays, Int. J. Multiphase Flow 35, 101-117, 2009. 6) Kister, H.Z., Distillation Design, McGraw-Hill: New York, 1991. 7) Standard Test Method for Distillation of Petroleum Products at Atmosphere Pressures, ASTM Standard D 86-04b. In Book of Standards, American Society for Testing and Materials: West Conshohocken, PA: Vol 05,01, 2004. 8) Beale, J.C., and Reitz, R.D., Modeling Spray Atomization with the Kelvin-Helmholtz/Rayleigh-Taylor Hybrid Model. Atomization and Sprays, 9, 623-650, 1999. 9) Liu, A.B., Mather, D., and Reitz, R.D. Modeling the Effects of Drop Drag and Breakup on Fuel Sprays, SAE Paper 930072, 1993. 10) O'Rourke, P.J., and Amsden, A.A. A Particle Numerical Model for Wall Film Dynamics in Port-Injected Engines, SAE Paper 961961,1996. 11) Fan L., Reitz R.D., Development of Ignition and Combustion Model for Spark-Ignition Engines, SAE Paper 2000-01-2809, 2000. 12) Federico Perini, Youngchul Ra, Kenji Hiraoka, Kazutoshi Nomura, Akihiro Yuuki, Yuji Oda, Christopher Rutland, Rolf Reitz, "An efficient level-set flame propagation model for hybrid unstructured grids using the G-Equation", SAE Int. J. of Engines 9(3): 1409-1424, 2016. 13) http://www.mcs.anl.gov/research/projects/mpi/ 14) Perini F, Galligani E, Reitz RD, "An analytical Jacobian approach to sparse reaction kinetics for computationally efficient combustion modelling with large reaction mechanisms," Energy and Fuels, 26 (8), 4804-4822, 2012. 15) Han, Z., and Reitz, R.D., Turbulence Modeling of Internal Combustion Engines Using RNG k-e models, Comb. Sci. Tech., 106, 267-295, 1995. 16) Sung-Jun Kim, Youngchul Ra, Yongwook Yu, "CFD Simulation of GDI Engine Cold Start under Extreme Condition with Multicomponent Gasoline Fuels", FISITA World Automotive Congress, F2016-ESYG-010, 2016. 17) Y. Ra and R. D. Reitz, "A combustion model for multicomponent fuels using a physical surrogate group chemistry representation (PSGCR)", Combustion and Flame, 162, 3456 3481, 2015. 6