Final Report Summary - DREAMCODE (Development of Reliable Emission and Atomization Models for Combustor Design)
In project DREAMCODE improved computational methods have been developed to predict combustion emissions accurately and reliably. DREAMCODE is part of the SAGE6 demonstration project, which aims to develop and mature a lean burn combustion system. Lean burn provides significant benefits in terms of NOx emissions. However, the emissions of CO, UHC and soot limit the operation of the combustor at different conditions. Reliable predictions of emission trends will lead to optimized combustor designs in a cost effective way. Existing capabilities, however, were inadequate to produce accurate and reliable predictions in direct support of lean burn system design. To develop computational methods that can be used in the design process of low emission combustors, the following essential elements of a CFD combustion emission tool were improved.
First, detailed comprehensive reaction mechanisms for jet fuel surrogates are necessary to describe the complicated chemical processes of fuel oxidation and emission formation. In DREAMCODE an extensive mechanism for jet fuel surrogates was step by step updated, refined and validated. The new mechanism was found to work well for a wide range of test cases with significant improvements in predictions of NOx and soot precursors.
Second, soot models are indispensable to describe the complex physical and chemical phenomena of soot particle formation. In DREAMCODE the modeling of soot formation was advanced by the development of a new multivariate soot model and improved statistical models. Model uncertainties have been analyzed systematically and carefully validated. Satisfactory predictions of soot volume fractions and particle size distributions were obtained.
Third, reduced chemistry method were developed to include detailed chemistry models in CFD calculations at affordable computational costs. More specifically, the Flamelet-Generated Manifold (FGM) reduction method was coupled with the soot models and extended to improve the prediction of combustor emissions. In addition, the Rate Controlled Constrained Equilibrium (RCCE) method was employed to reduce detailed mechanisms for kerosene. By combining the RCCE with artificial neural networks a speed-up of one to two orders of magnitude was obtained.
Fourth, accurate and efficient spray break-up models were developed to predict the fuel distribution in the combustor. To that end, atomization experiments were analyzed with advanced analysis tools to extract the characteristics of the atomization process. In addition, detailed Smoothed Particle Hydrodynamics (SPH) simulations were performed to study the fuel break up in pre-filming air-blast atomizers. Based on these new insights a phenomenological spray model for predicting the droplet size distribution was developed and successfully implemented into a CFD code.
Finally, the various models were integrated in a CFD code for the modeling of turbulent combustion in gas-turbine combustors. Large-Eddy Simulations (LES) of various test cases were performed to assess the accuracy of the models. With the integration of these improved models, a much more reliable CFD tool has been developed that will guide the further design of lean burn combustion systems.
Project Context and Objectives:
The SAGE6 demonstration project aims to develop and mature a lean burn combustion system. An essential enabler to development of such technology is an accurate and reliable computational tool for prediction of emissions. Lean burn provides significant benefits in terms of NOx emissions.
However, the emissions of CO, UHC and soot limit the operation of the combustor at different conditions. Reliable predictions of emission trends will lead to optimized combustor designs in a cost effective way. Existing capabilities, however, were inadequate to produce accurate and reliable predictions in direct support of lean burn system design.
The DREAMCODE project aimed to develop and improve computational methods that can be used in the design process of low emission combustors. Improved models and methods have been developed to predict emissions accurately and reliably. To that end, the following essential elements of a CFD combustion emission tool were considered:
1. Detailed chemistry models for jet fuel surrogates are necessary to describe the complicated chemical processes of fuel oxidation and emission formation in the gas phase.
2. Soot models are indispensable to describe the complex physical and chemical phenomena of soot particle formation.
3. Chemistry reduction methods are inevitable to reduce the computational cost of the complex chemistry model for application in CFD codes.
4. Spray break-up models are necessary to model the liquid fuel break-up, which has a dramatic effect on emissions.
5. Turbulence-chemistry interaction models have to account for the effects that occur on length scales which cannot be resolved by the computational mesh.
These 5 models have been improved and integrated in a CFD code for the validation on real aero engine gas turbine combustors.
Project Results:
Development of detailed chemical reaction schemes for emissions modelling (WP2)
Task 2.1 – Detailed chemistry model for jet fuel surrogates
The use of surrogate blends provides a route towards the inclusion of realistic fuel effects into design computations. The target for work performed as part of Task 2.1 was to provide an assessment of the performance of a comprehensive surrogate fuel mechanism applied to model actual kerosene data. The surrogate formulation followed the CFD4C (G4RD-CT-1999-00075) and FIRST (FP7-AAT-2010-RTD-1) projects with n-decane and n-propyl benzene selected as the alkane and aromatic fuel constituents. The inclusion of further fuel components is currently difficult to justify due to the increased complexity, leading to reduced applicability, and additional uncertainties in the thermochemistry.
The work performed as part of Task 2.1 has contributed to a robust and reliable surrogate fuel mechanism for aviation applications and has focussed on an extensive validation, performed in a stepwise manner, of the developed mechanism. Despite the practical relevance of aromatic fuel components, chemical kinetic and thermodynamic data have generally not been obtained using comparatively accurate ab initio quantum mechanical methods (QM) with current reaction mechanisms typically based on reaction class based estimates for the higher hydrocarbon chemistry. The application of 10 increasingly accurate QM methods, including state-of-the-art DFT (M06, M06-2X and M08-SO) and contemporary composite methods (G4, G4MP2, CBS-QB3 and CBS-4M), was presented for surrogate fuel components as part of the FIRST project relative to data obtained using the CCSD(T)/jun-cc-pVTZ//M06-2X/6-311++G(3df,3pd) coupled cluster based method. Additional updates included the cyclopentadiene/cyclopentadienyl system that forms a critical part in the oxidation chemistry of aromatic fuel components. The impact of such updates and further refinements were assessed as part of the DREAMCODE project.
The chemistry of aromatics in aviation fuels presents a particular challenge and in deliverable D2.1 the impact of updates to the oxidation and pyrolysis chemistry of n-propyl benzene under fuel lean, stoichiometric and rich conditions was covered. In deliverable D2.4 the formation of oxides of nitrogen was added. Test cases included premixed, as well as diffusion flames, with consideration given to nitric oxide formation as well as the reburn chemistry, including the role of ammonia fragments. It was shown that quantitative predictions were substantially improved when compared to alternative suggestions. In D2.5 experimental data obtained for flames and ignition delay times using practical kerosene fuels (e.g. JP-8) was used. The applied surrogate kerosene fuel mechanism featured a blend modelled as 79% of n-decane and 21% of n-propyl benzene (by volume) based on a fuel analysis. Hence, the complexity of the surrogate is substantially reduced compared to alternatives featuring up to six fuel components. The developed model was found to reproduce experimental data successfully with key species concentrations well predicted.
Thermodynamic effects can play a major role in the PAH growth sequences leading to soot inception. Efforts have been made to augment the thermochemistry for large aromatics using composite G3B3, G3MP2B3, G4MP2 and G4 methods with the choice depending on the maximum accuracy possible given the size of a particular molecule. Properties were determined for over 100 poly-aromatic hydrocarbons and the updates included in the revised surrogate fuel mechanism. For the n-propyl benzene system, thermochemical properties of species were determined using the G4 composite method and resulting data were fitted to the 7 term JANAF polynomials covering a range of 200 - 6000 K. The impact of the updated thermochemistry was also assessed in the context of Task 2.2.
The updated model was further evaluated in terms of assessing the ability to predict molecular growth and PAH formation under conditions where aromatics are not present in the fuel – as would be the case for Fischer-Tropsch based alternative fuels. For this purpose, ethylene diffusion flames were used. The choice of ethylene was made due to it is importance in the breakdown process of the higher alkanes present in aviation fuels. It was shown that concentrations of formed aromatics and major species could be well reproduced.
The comprehensive DREAMCODE mechanism features 402 chemical species and 2240 reactions and was found to work well for a wide range of test cases, including actual kerosene combustion, with NOx results surpassing those obtained using GRI-Mech 3.0 with respect to key aspects due to an extended/different validation data suite. The mechanism is suitable for the computation of fuel rich conditions, as well as diffusion flames, and was translated to Cantera and Chem1D (CHEMKIN) formats for use by project partners.
Task 2.2 – Soot modelling
The work performed in Task 2.2 and reported in deliverable D2.6 addresses the emerging need to consider soot particle size distributions (PSDs) due size dependent toxicity effects. The work can be viewed as a step towards the development of an essential modelling capability that enables the prediction of the number of soot particles and their mass in different size ranges from the nanoscale upwards.
Sectional methods hold the promise of delivering accurate PSD predictions with reasonable computational expense and the developed model conserves both soot mass and particle number densities. To date no experimental PSD data has been procured for either laminar diffusion flames or turbulent flames. Hence, the initial assessment of sub-model sensitivities under turbulent flame conditions was evaluated with the soot particle dynamics treated via the method of moments with interpolative closure (MOMIC) combined with a transported joint probability density function (JPDF) based calculation procedure. A variable number of moments of the soot particle size distribution function (PSDF) and the size distribution of the primary soot particles per aggregate were considered and sensitivities assessed. The method solves the transport equation of the JPDF of the position dependent random vector that includes, in addition to the moments describing the soot properties, all species mass fractions, the mixture fraction and the enthalpy. The coagulation/agglomeration terms in the soot model are condition dependent and determined by the value of the Knudsen number. The approach covers for the free molecular and continuum regimes and accurately accounts for turbulence-chemistry-radiation interactions associated with the comparatively long chemical time scales for soot formation/oxidation.
The work included: (i) The selection of turbulent flame test cases consistent with the International Sooting Flame (ISF) workshop flame series; (ii) The computation of the selected test cases; (iii) The evaluation of the sensitivity to oxidation rates of soot; (iv) The evaluation of the sensitivity to assumptions made for the soot surface chemistry; (v) The translation of the developed models to the sectional method framework for turbulent flame computations; (vi) The evaluation of the sectional model subroutines for cases where PSD data is available.
The applied thermochemistry was updated as part of DREAMCODE, as outlined in Task 2.1 and it was shown that the level of agreement that can be expected for key species is satisfactory and probably close to experimental uncertainties in many cases. It is of note that the ability to predict pyrene (A4) is of particular importance due to its role as the designated PAH used to describe soot inception and that, due to the size of the full mechanism (2240 reactions and 402 chemical species), future simplifications will be necessary.
The level of agreement was found to be very encouraging and the sensitivity of soot predictions to oxidation and acetylene based mass growth rates analyzed with recommendations made for the selection of reaction rates. The developed chemistry was transferred to subroutines used for the calculation of soot PSDs in a premixed well stirred - plug flow reactor (PFR/WSR) system. The agreement obtained with the full model was shown to be excellent with the correct sensitivity to the amounts of aromatics present in the fuel. It was further established that the predominant uncertainty in terms of simplifications is associated with differences in the soot nucleation rate between premixed and diffusion flame systems. Encouragingly, when such differences are accounted for, it has been shown that the full PSD is likely to be predicted accurately even with simplified nucleation models. The development of the latter is considered key for future work along with the procurement of experimental data that enables a comprehensive assessment of the accuracy of such models.
Chemistry reduction (WP3)
Task 3.1 – Flamelet Generated Manifolds
The objective considered, is a development of improved FGM model, which is a flamelet based chemistry tabulation method. The main focus is on accurate prediction of CO, UHC, NOx and soot emissions. In an aircraft engine combustor heat losses and pressure changes may occur for example by fuel evaporation, radiation and the expansion of the burnt gases in the stators, also called as nozzle guide vanes (NGV). In order to capture the influence of these processes on formation of emissions, an FGM extended with enthalpy and pressure variations was developed. Also, a method for the reduction of the FGM memory use was investigated. A 3D FGM utilizing enthalpy and pressure as extra control variables was assessed in one-dimensional laminar flame simulations mimicking NGV conditions, characterized by rapidly decreasing enthalpy and pressure. Simulations using this manifold yielded an acceptable level of agreement with detailed chemistry results, NO and CO were predicted with 5% accuracy. Further, an FGM based on two (or more) reactive control variables was developed. Following the ideas of ILDM, the dimension of the manifold was locally extended by means of a time scale analysis of the reaction kinetics. The FGM was extended by including the dimensions related to the slowest chemical time scales. This new model, applied in the aforementioned one-dimensional NGV simulations, yielded a higher accuracy compared to that of a standard FGM. Especially for CO emissions, significant improvements were observed. For NO the absolute improvement was small, but the extended manifold could capture the qualitative behavior of NO consumption. As last, in the current period, the focus is placed on the modelling of soot with FGM reduced chemistry. The first step towards reduced chemistry soot modelling, is an accurate modelling of soot precursors with FGM. Since soot nucleation is strongly dependent on polycyclic aromatic hydrocarbons (PAH) concentration, an accurate modelling of these is crucial. Same holds for the acetylene, molecular oxygen and hydroxyl radical, as these are utilized in soot surface growth- and oxidation closure. The FGM model was assessed for soot precursors against laminar one dimensional detailed chemistry flame simulations. These simulations utilized a partially premixed configuration, where combustion of a premixed (rich) mixture happens while it is mixing with pure air. In this way, an attempt was made to mimic a rich (sooting) pilot-stream in a lean burn engine. In the range of conditions considered, FGM has been shown to predict the detailed chemistry results for the soot precursors and oxidizers within 5% accuracy. Basing on this, an expectation has been made that correct prediction of soot source terms is feasible in the context of the FGM model. Further, in order to enable Large-Eddy Simulations (LES) using the Flamelet Generated Manifolds (FGM) model as reduced chemistry approach, the FGM model for soot precursors has been connected to the Hybrid Method of Moments (HMOM) soot modelling approach developed in another work package of this project. The FGM chemistry tables have been coupled to the LES and DNS code CIAO. Then several preliminary steps have been performed to validate the combined model. First, an a priori verification has been performed to confirm that the computation of the soot model quantities, during the manifold generation with CHEM1D, is consistent with that implemented in FlameMaster. Further the integration of the FGM model into the CIAO code has been validated for a 1D laminar freely propagating premixed ethylene flame. Comparisons have been made between the results found in the CIAO CFD code using FGM and a detailed chemistry computation of the same flame using FlameMaster, where the same soot model as in CIAO is implemented. Both the soot volume fraction and the number density were in very good agreement, confirming a correct functioning of the coupled model. To continue the work on further validation of the combined FGM soot model, an analysis has been performed on a complex test case involving a turbulent non-premixed jet flame. This selected test case is the Target Flame 1 from the International Sooting Flame (ISF) Workshop, the Adelaide Turbulent Jet Flame. The results of the Large-Eddy Simulations (LES) have been obtained utilizing the FGM model for the gas-phase soot precursors and the bivariate HMOM model for soot formation and growth. Part of research involved the simulation of the same test case flame applying the two-equation soot model of Lindstedt et al., namely the model employed at Rolls-Royce Germany in the past years, in such a way that the performance of the two soot models could be observed in parallel. This part of the research shed a first light on the possibility of coupling the HMOM soot model with the FGM method. It could be concluded that to predict the particles distribution and magnitude with the necessary accuracy a superior description of the flow field cannot be disregarded. In fact, all the soot precursors as well as the temperature are determined as a function of mixture fraction. For the Adelaide test flame studied, some criticism may be expressed regarding the treatment of the preferential diffusion of hydrogen, being one of the main components of fuel mixture.
Task 3.2 – Rate Controlled Constrained Equilibrium
In this task Rate Controlled Constrained Equilibrium (RCCE) is used in order to develop a reduced chemical mechanism for jet fuels surrogates, which would then be utilized for the development of Artificial Neural Networks (ANNS).
RCCE provides a physical and mathematical framework for deriving reduced mechanisms, where the kinetically controlled species (slow or constrained species) are determined from the kinetics of the detailed mechanism, without any assumption. The remaining species (steady-state or fast species) are calculated by minimizing an energy function, such as the Gibbs free energy of the system, subject to the constraint that the kinetically controlled species must maintain their concentrations as derived from chemical kinetics and the additional constraints of conserving energy, mass and elements.
In the context of this task, several detailed chemical mechanisms were tested with several surrogate fuels which were based on n-decane or n-dodecane. The reduced chemical mechanisms were mainly tested in a laminar non-premixed flame and several strain rates were compared in order to account for different conditions of these flamelets from one extinction limit to the other. A deep investigation of the timescales of these detailed mechanisms was conducted in order to identify the major species which are constrained in RCCE. Three different methods were studied at first place and were compared systematically to order the species according their importance. These methods are
Computational Singular Perturbation (CSP), Level of Importance (LOI), simply species with the highest concentrations and a simplification of LOI which was developed (Concentration and Sensitivity Criterion, CSC). The reduced mechanisms which were derived for the different strain rates provided very accurate species and temperature profiles in the above mentioned configuration. Species with high interest such CO, CO2 and C2H2 with the last being very important for soot formation, were predicted with very small error.
Subsequently, a methodology for identifying the kinetically controlled species for RCCE was developed and it was based on Computational Singular Perturbation (CSP) to identify the active modes of the system. These modes were then related to the species using their mass fraction as well, and the slowest species for each chemical mechanism were identified. This methodology was found to be able to derive successfully sets of constraints for several fuels with several mechanisms including methane with GRI1.2 propane with USC-Mech II and kerosene surrogate with a detailed mechanism for kerosene.
For training, testing and simulating the ANNs, we use the Self-Organizing Map (SOM) - Multi-Layered Perceptrons (MLPs). In this implementation, the composition space is subdivided into subareas with similar composition preserving topology and improving the performance. In order to assign a point (or each sample) in composition space to one of these subareas, its Euclidean distance to all the subareas is calculated; then this point is assigned to the subarea with the minimum distance. Several sets of ANNs were derived in order to identify the one with the best performance, regarding the number of subareas of the map, training thresholds and training data.
A set of ANNs which was trained with a reduced mechanism which includes 24 species from a detailed with 369 species and 1785 reactions, was found to predict well both species profiles and temperature. It is also of great importance that the concentration of major species, like carbon dioxide, carbon monoxide, acetylene and fuel components were predicted accurately and that other soot precursors such as benzene and styrene are identified and included in the reduced mechanism. The set of ANNs was tested under several conditions in one-dimensional laminar non-premixed flames for several strain rates and it predicted well the profiles of species and temperature in mixture fraction space. The reported computational savings are around one to two orders of magnitude compared to the reduced mechanism, while these savings are highly increased when compared with the detailed mechanism.
Given the degree of reduction of computational savings and the accuracy of the ANNs to predict profiles of species and temperature over a wide range of initial conditions, its high performance in turbulent combustion CFD codes is expected.
Spray break-up model development (WP4)
The spray generation of prefilming airblast atomizers can be split up into the following mechanisms: formation of bags and ligaments and their breakup (so called primary breakup) and the subsequent disintegration of larger droplets (so called secondary breakup). Also other effects like spray / wall interaction and evaporation occur but these are not in the focus of the present project. The secondary breakup of droplets is widely understood, but the primary atomization is still a topic of active research.
The primary atomization process can be either directly simulated, using numerical methods like embedded Direct Numerical Simulation (eDNS) which is based on a Volume of Fluid (VoF) approach, as well as Robust Conservative Level Set (RCLS) or Smoothed Particle Hydrodynamics (SPH) or modeled using more global primary atomization models. In the context of combustion predictions, these primary atomization models are used in combination with an Euler-Lagrange approach for droplet transport. Thus, spray generation and combustion can be predicted together, which enables the prediction of the exhaust gas emissions.
In this context novel data reduction techniques have been assessed for the extraction of characteristics of the atomization process in the context of prefilming airblast atomizers. Thereby, the Proper Orthogonal Decomposition proved to be a valuable tool for extracting characteristic frequencies and wavelengths on the film flow upstream of the atomizing edge.
2D Smoothed Particle Hydrodynamics (SPH) simulations of the prefilming airblast atomization process have been conducted and were found to be able to capture the physical phenomena of primary breakup. In detail these are the wavy film flow as well as the ligament formation and elongation as well as the flapping of the liquid sheet. 3D phenomena like the generation of liquid bags with a thick rim and a thin interior film and the liquid accumulation at the atomizing edge are missed or weakly captured, because of the 2D setup. Nevertheless, the comparison of the 2D SPH results with the experiment and the spray model showed good agreement.
As the final goal for the project, a phenomenological spray model was developed which is based on easily accessible quantities affecting the atomization process of prefilming airblast atomizers. These are the air velocity, the surface tension of the liquid and the thickness of the atomizing edge of the prefilmer. The basic ideas are to use a parametrized Rosin-Rammler function to describe the droplet size distribution and to take into account phenomena like the accumulation of liquid at the atomizing edge. The model constants have been obtained from experimental results of Gepperth. In these experiments planar atomizer configurations with different atomizing edge thicknesses covering a large range of operating conditions and liquid properties have been analyzed using high speed videos. The phenomenological spray model developed within this project was found to predict droplet size distributions reliably for a wide range of geometries and operating conditions.
The phenomenological spray model was successfully implemented into a CFD code and is ready to be used. Within the CFD code the spray model takes into account the local air velocity at the atomizer. Hence, the spray model is capable to take into account the swirl and other local flow features.
In summary, within work package 4 the following achievements have been obtained: Novel data reduction techniques have been successfully applied for the extraction of characteristics of the atomization process. 2D SPH simulations enabled novel insights in the primary atomization process. A phenomenological spray model for predicting the droplet size distribution was developed and successfully implemented into a CFD code.
Validation based on real aero engine gas turbine combustors (WP5)
Task 5.1 – Integration of spray break-up and emission models
The phenomenological spray breakup model developed within task 4.1 was implemented into a CFD code. Due to the nature of the model, a new formulation of the starting conditions for the droplet injection was introduced. The droplets are injected near each face at the atomizing edge of the injector and not at discrete injection points. Consequently, the spatial distribution of the injected droplets is improved. All input parameters of the spray model are user defined, except the local air velocity. Hence, the model is capable to take into account the swirl and other local flow features.
The spray model was demonstrated to work properly with parallel processing, the steady and the transient solver and the following three spray sub models: evaporation, turbulent dispersion (stochastic) and secondary atomization. The spray predicted by the spray model and the injected droplets in the simulation domain have been compared and have been found to be identical. Consequently, the spray model is demonstrated to function properly and is ready to be used within any CFD code.
In this task also the combined FGM soot model was further validated by performing an analysis of a complex test case involving a turbulent non-premixed jet flame. This selected test case is the Target Flame 1 from the International Sooting Flame (ISF) Workshop, the Adelaide Turbulent Jet Flame. The results of the Large-Eddy Simulations (LES) have been obtained utilizing the FGM model for the gas-phase soot precursors and the bivariate HMOM model for soot formation and growth. Part of research involved the simulation of the same test case flame applying the two-equation soot model of Lindstedt, namely the model employed at Rolls-Royce Deutschland in the past years, in such a way that the performance of the two soot models could be observed in parallel. This part of the research shed a first light on the possibility of coupling the HMOM soot model with the FGM method. It could be concluded that to predict the particles distribution and magnitude with the necessary accuracy a superior description of the flow field cannot be disregarded. In fact, all the soot precursors as well as the temperature are determined as a function of mixture fraction. For the Adelaide test flame studied, some criticism may be expressed regarding the treatment of the preferential diffusion of hydrogen, being one of the main components of fuel mixture.
Furthermore, an updated and further validated set of ANNs for the reduced mechanism of kerosene was delivered and presented in this task. The set of ANNs was tested in the one-dimensional counterflow non-premixed flames for a wide range of strain rates and the accuracy of the ANNs in this combustion regime was remarkable for both of main species profiles, including the fuel components, CO2, CO, H2O, H2, O2 and for temperature. The development, training and application of ANNs has mainly based on our previously developed methodologies and it was, for these purposes, validated in a turbulent non-premixed flame, which is described in detail in D5.3.
Task 5.2 – Validation on real GT combustors
In this task the effects of the turbulence-chemistry interaction was studied by implementing the reduced mechanism (the trained ANNs) into a well-studied CFD code, BOFFIN. The selected test case is the non-premixed CH4 flame L of the experimental Sydney piloted burner adopting the fractional step formulation following convection-diffusion, micro-mixing and reaction steps. Results indicate an average of 13 seconds spent on the convection-diffusion step, while only 1 second spent on reaction (ANNs). Subsequently, a second simulation was performed using the RCCE reduced mechanism technique with 17 kinetic constrains. Results indicate an average of 13 seconds spent on the convection-diffusion step, while 130 seconds used on reaction (RCCE optimization). The outcome indicates that the ANN tabulation has very modest CPU time requirements allowing the simulation of a real flame with a method as comprehensive as LES-PDF on a standard workstation, without the necessity for supercomputing facilities.
The prediction of all the reactive scalars (including profiles of methane, carbon dioxide, carbon monoxide, water and oxygen mean and rms mass fractions) shows good agreement with the experiments, especially for instances closer to the nozzle since due to the pilot the flame there maintains a healthy burnt state. The simulation is also able to capture the local extinction and re-ignition, phenomena that exist to a large extent in this flame; these phenomena are very important for turbulent combustion CFD.
Potential Impact:
The main goal of the SAGE6 Demonstration Project is to develop a lean burn combustion system suitable for civil aerospace up to technology readiness level 6. An essential enabler for the development of such technology is an accurate and reliable CFD method for the prediction of emissions. Present methods are still inadequate to produce accurate and reliable predictions in direct support of lean burn system design. The main objective of project DREAMCODE is to resolve this shortcoming and to develop a CFD method for accurate predictions of combustor emissions. This CFD method will be used by the ITD Members to drive combustor designs in the right direction and effectively lead to optimal solutions.
To ensure that the output of this project can be used directly by the ITD, the newly developed models have been implemented and validated in close collaboration with the ITD Topic leader. Furthermore, the results were presented at conferences and workshops and have been published in scientific journals. The availability of high accuracy CFD methods for the prediction of emissions will allow the aero-engine developers to create an engine with a combustion system that will offer competitive emission levels. By gaining a competitive edge over the competitors, the European aero-engine industry will be able to provide a large proportion of the new green aero-engines needed for the aeronautical growth over the following decades. This creates jobs and helps the economy in Europe.
Although experimental testing is bound to play a key role in the development of combustion systems, a reliable CFD method will reduce the number of expensive time-consuming experimental tests. In this way, the time to market for new green technological solutions will be shortened, which corresponds to the objective of the Clean Sky program. Rapid progress in the introduction of green technology into aviation is necessary to make major steps toward the environmental goals set by the Advisory Council for Aeronautics Research in Europe to be reached in 2020.
List of Websites:
There is no public website dedicated to this project.