LESSCCV
Large-Eddy Simulation and System simulation
to predict Cyclic Combustion Variability
in gasoline engines
A R&D project on piston engine combustion financed by the Europen Union in the frame of the 7th Framework prohgramme (FP7).
Contact
Dr. Christian Angelberger
IFP Energies nouvelles
christian.angelberger at ifpen.fr
Project context and objectives
The operation of spark-ignition engines (SIE) is characterised by a non-repeatability of instantaneous combustion rate between different cycles with nominally identical engine operating parameters, commonly referred to as cyclic combustion variability (CCV). CCV inevitably appear over the whole engine operation range, due in particular to the unsteady, cyclic and turbulent nature of flow and combustion in SIE. As long as their amplitude is sufficiently low, engine simulation software or model based engine control can predict combustion in any given cycle using deterministic models that neglect cyclic variations.
It is common practice to consider that CCV amplitudes (measured in terms of standard deviation of IMEP, normalised by the mean value) of less than around 5% are acceptable. For higher CCV amplitudes, individual cycles will behave very differently from the statistically most probable cycle predicted by control algorithms or engine simulation tools. As a result, predictions of fuel consumption or emissions over a number of cycles may differ quite substantially from the one based on the mean engine cycle. These differences increase with the CCV amplitude, and may reach extreme levels in the case of misfires or extreme knocking cycles.
In a context of more and more stringent constraints on fuel consumption, CO2 production, and pollutant emissions from road transport, it becomes crucial to be able to predict and control individual engine cycles, and thus to address the occurrence and effects of CCV. Engine technologies as direct injection (DI), controlled auto-ignition (CAI) or downsizing are key elements on the way to reducing the CO2 emissions from future SIE. Yet the occurrence, under certain operating conditions, of excessive CCV when implementing these technologies is one of the factors limiting their theoretical performance or range of operation. Being able to predict CCV in early design phases based on an improved knowledge of their sources and effects could be essential to exploit the full potential of these promising SI technologies under real operation.
The starting point of the LESSCCV project is the fact that the understanding of how the complex combination of different sources leads to the occurrence of CCV for a specific engine design or mode of operation is still limited; and that there is a need to make such knowledge available in system simulation tools, which are increasingly used in the engine development and optimisation process.
LESSCCV proposes an innovative approach based on advanced Computational Fluid Dynamics (CFD) engine simulation tools to gain a better understanding of CCV due to flow phenomena. 3D CFD approaches as the emerging Large-Eddy Simulation (LES), particularly suited for addressing non-cyclic phenomena, allow studying in detail complex phenomena in real configurations, but the required CPU time is far too large to comply with industrial requirements on return times. On the other hand, 1D CFD tools are much less CPU time intensive, and can run as fast as real time, which makes them suitable in particular for developing control strategies. Yet they cannot predict complex phenomena as CCV.
The originality of the research proposed in LESSCCV is to combine advanced simulation studies based on LES to gain a deeper insight into the complex combination of phenomena at the origin of CCV in different engine types, and to capitalize the acquired understanding in the form of phenomenological models that would be able to reproduce sources and effects of CCV in 1D CFD tools, which could then potentially be used to explore control strategies aimed at reducing the occurrence and negative impact of CCV.
LESSCCV started on 1 December 2009, and will last 3 years. It is coordinated by Dr. Christian Angelberger from IFP Energies nouvelles, and brings together 8 European partners, all active members of the EARPA Task Force Modelling & Simulation who was at the origin of the project: IFP Energies nouvelles (IFPEN), AVL List GmbH (AVL), FEV Motorentechchnik GmbH (FEV), LMS-Imagine (LMS), Ricardo UK Limited (Ricardo), Technical University of Prague (JBRC), Politecnico di Milano (PoliMi) and University of Western Macedonia (UOWM).
The overall project budget is 3,2M€, with an EC funding of 2,1M€.
Work performed and main achievements
R&D work achieved in the first 18 project months was globally dedicated to apply LES (or a variant called PANS) to study the occurrence of CCV in three different types of SI engines. The final aim of these computing time intensive activities is to provide a detailed insight into the causes and effects of CCV that will serve as a basis in the second project half to propose phenomenological models.
A first team formed by researchers from IFPEN and PoliMi worked on studying CCV in an indirect injection single cylinder SIE fuelled with a perfect pre-mixture of air and propane. The engine serving as a basis for this work was extensively studied in the French ANR project SGEmac coordinated by IFPEN, the purpose of which was to acquire a unique database for validating the predictivity of LES in terms of CCV. On-going work concerns the development of a 2-way coupling between PoliMi's 1D CFD GASDYN and IFPEN's LES code AVBP, that would allow studying CCV while including the whole engine set-up. In this approach, GASDYN would simulate the flow in the complete intake and exhaust ducts, while AVBP would address flow and combustion inside the combustion chamber and its neighbourhood. While this coupling is under development, collaborative work started with a simpler approach consisting of a chaining of both codes. GASDYN hereby simulates the studied operating points for the whole engine, and furnishes temporal evolutions of flow variables that are used by AVBP to impose boundary conditions on the limits of the 3D domain. Even if this does simplify interactions between the flow in the cylinder and the ducts, its application to the LES of 25 consecutive engine cycles for a motored reference operating point proved sufficient to achieve a very satisfactory reproduction of the flow patterns observed experimentally. Once the chained approach was validated on this first point, work started on a fired operating point with low CCV, which will serve as a basis for studying two further operating points characterised by high CCV levels in the coming period.
A second team bringing together researchers from JBRC and AVL (assisted by RICARDO on coupling aspects) applied the LES/PANS approach (with PANS corresponding to LES in highly resolved regions and RANS in those where this is not possible as close to solid walls), to the study of CCV in a direct injection (DI) engine. The basis for this work was a single cylinder engine for which experimental results were available to AVL. Extensive work was first dedicated to setting up the LES/PANS simulation methodology on a motored operating point of a comparable engine for which up to 45 cycles were simulated. First, comparing LES with single cycle RANS predictions highlighted how the former allow reproducing cycle to cycle variations in the intake- and compression-induced in-cylinder flow field especially in the vicinity of the spark plug. Second, the LES of a direct fuel injection was set-up for the same engine, and results in terms of flow velocity conditions at the spark plug compared with the case without injection, allowing to conclude that the low pressure injectors of this engine only had a very small impact on the flow variability at the spark (and this potentially on CCV). However a clear variability of composition at the spark plug is induced by the DI, which could have an impact on early flame kernel growth and thus CCV. The application of the developed LES/1D methodology then started to be applied to AVL's SCRE engine for which experimental data on CCV are available. The 1D sketch and the 3D mesh were set up, and the first achieved simulations will be continued in the second project part.
The third LES study is being undertaken by FEV using the Star-CD code with it's LES option, and boundary conditions coming from experiments to study CCV in a CAI engine. After the development and validation of a LES sub-model for liquid injection, a main part of the work consisted of the development, implementation and validation of a LES combustion model for CAI based on reduced chemical kinetics and accounting for air, fuel, residual gas and temperature stratification on auto-ignition. The resulting models were then applied to the LES of two different CAI strategies and allowed exploring the role of stochastic perturbations of the in-cylinder charge and mixture formation distribution and the role of preceding cycles, which directly affect the state of the residual gas content.
As local conditions at the spark plug, and the way they affect early flame kernel growth are probably of importance for the onset of CCV, UOWM used a Direct Numerical Simulation (DNS) approach – in which all flow scales are resolved and not modelled – to study in detail these phases. A first series of DNS allowed to characterize the relationship between the flame displacement speed and the stretch rate at early instants, and to find the functional dependence between heat release rate and laminar flame speed. A second study was used to investigate how the flame surface area is affected by mean convection (as can be observed at the spark plug), completed by a study investigating how the flame surface area is affected by turbulence. The parameters of these DNS were chosen to be as close as possible to those observed in SI engines. The obtained results provide valuable information to other partners on the physics of early flame kernel growth, and can serve to validate elements of the spark ignition models they use in their LES.
Finally, on-going work by IFPEN and LMS concerns the exploration by dedicated 1D CFD and LES of the impact unsteady injector outflow conditions may have on a liquid fuel spray under engine conditions, and how this may affect CCV in a DI engine.
Expected final results and their potential impact and use
By proposing to acquire knowledge aimed at deepening the understanding of the sources of CCV related to flow phenomena, and of their effects on the operation of advanced SI engines, LESSCCV aims at contributing to a better understanding of combustion in gasoline engines under realistic operating conditions. The improved 1D engine simulation codes – many of which developed based on commercial software largely used in the automotive industry - able to reproduce causes and effects of CCV on a CPU time inexpensive manner that will be developed and validated in the second project part will be readily available to OEMs in order to contribute lessening by design the negative effects related to the occurrence of CCV in SIE. Furthermore, the 1D-CFD engine simulation tools resulting from LESSCCV will find a usage in developing and testing model based control strategies aimed at reducing as much as possible the negative effects of CCV.
LESSCCV shall also contribute to the further development of innovative 3D-CFD codes based on the LES technique, which has unique capabilities as compared to today's standard, RANS, as increased predictivity and the capacity to address non-cyclic phenomena not accessible up to now. LES could gain an even more important role and perhaps lead to the emergence of a numerical engine test bench on the long term beyond 2016, by exploiting the fast development of massively parallel computing power in Europe.
By contributing to rendering road transport more efficient and less polluting, LESSCCV will also have an indirect positive impact on the achievement of the EU's goals for a sustainable and green transport, as well as on the competitiveness of European industry.