Towards modelling of multiple combustion modes: Dual-fuel Concept & Formulation

Transcription

1 Towards modelling of multiple combustion modes: Dual-fuel Concept & Formulation Verbrennungstagung - ETH Zurich D. Farrace, Y. M. Wright and K. Boulouchos

2 Why dual-fuel combustion? The main reason are the fuel costs! price decoupled from liquid fuel price increasing availability of gaseous fuels [] [] M. Ott et al., 3. Rostocker Grossmotorentagung,

3 Why dual-fuel combustion? The main reason are the fuel costs! price decoupled from liquid fuel price increasing availability of gaseous fuels Reduced emissions that can satisfy the stringent regulations of IMO Tier III reduced NO x favored by lean combustion (within Tier III w/o additional treatments) negligible sulfur oxides emissions practically no soot (PM) emissions regulation can be introduced in future theoretically lower GHG emissions [2] [2] E. J. Sixel, 3. Rostocker Grossmotorentagung,

4 Why dual-fuel combustion? The main reason are the fuel costs! price decoupled from liquid fuel price increasing availability of gaseous fuels Reduced emissions that can satisfy the stringent regulations of IMO Tier III reduced NO x favored by lean combustion (within Tier III w/o additional treatments) negligible sulfur oxides emissions practically no soot (PM) emissions regulation can be introduced in future theoretically lower GHG emissions Very flexible can be operated in diesel mode when LNG is not available

5 Why dual-fuel combustion? Challenges GHG emissions are lower in theory, in reality methane slip is an issue CH 4 impact on atmosphere is about times greater then CO 2 (in 2 years time scale) 3-6 g CH 4 slip per kwh during otto LNG operation [] Effective knock control strategies are required UHC emissions due to cold walls can be significant Deep understanding of physico-chemical processes is needed in order to optimise the combustion process and meet the requirements! High-fidelity experiments can provide an insight into the combustion process CFD can support experiments by investigating not measurable data multiple combustion modes and their interaction are a challenge multiple fuels behaviour is a challenge [] T. Mundt et al., 3. Rostocker Grossmotorentagung,

6 State-of-the-art in dual-fuel modelling Knowledge of dual-fuel operation is at early stages, involved processes are not fully understood yet autoignition spots flame kernels Diesel pilot Autoigniting diesel spray atomization and break-up evaporation autoignition diffusion flame non-premixed model flame propagation Regime transition premixed charge ignition interaction flame kernels growth Turbulent premixed flame flame(s) propagation extinction premixed model

7 State-of-the-art in dual-fuel modelling Level set approach + CTC/WM/CMC [][2][3] Spray ignition: CTC / Well Mixed (WM) / Conditional Moment Closure (CMC) depending on model cool flame chemistry could be considered Regime transition: flame kernel initialized when T>2K and R kernel >l mean Premixed flame: flame tracked by a level set method (G-equation/Weller) one-step chemistry, combustion regimes are distinguished SOI=-5 CA atdc SOI=-5 CA atdc Percent NG= 98 [] Δt after SOI Weller 3-eqn for λ CH4 =.5 CH4.7 ms.9 ms 2. ms 2.3 ms 2.5 ms [3] Iso surface b=.3 OH distribution! [] CTC: S. Singh et al., 25 [2] WM: Kokjohn et al., 2 [3] CMC: S. Schlatter et al.,

8 Concept of Multi-mode Combustion Modelling (MC-CMC) dual-fuel combustion non-premixed combustion mode transition premixed combustion CMC non-premixed CMC double-conditioned CMC premixed temperature Temperature [K] CA +. CA +.8 CA +2.2 CA +2.5 CA +3.2 CA +4. CA +6.6 CA temperature (K) Mixture Fraction [ ] mixture progress

9 Non-premixed CMC: Work overview Generic test rigs Aachen CVCC ignition delays, CMC physical space transport terms [] ETH CVCC ignition delay/location, mech. sensitivity [2], stoch. of autoignition [3] Sandia CVCC ignition delays, lift-off heights [4,5], soot formation [6,7] ETH single stroke machine diesel pilot ignition [8] Marine CVCC ignition delay/location, lift-off heights [9,] Engines ETH Liebherr heavy-duty Diesel engine pressure [], HRR and NO x [2], EGR [3] Sandia heavy-duty Diesel engine HRR and soot [4], NO x [5], EGR [6], post injections [7] ETH MTU heavy-duty Diesel engine ongoing work: NO x reduction by extreme Miller timing Wright et al., CNF 43 (25) [] Wright et al., FTaC 84 (2) [2] Wright et al., Procs LES4ICE (22) [3] Borghesi et al., CTM 5 (2) [4] Bolla et al., SAE Int. J. Engines 6 (23) [5] Bolla et al., CST 85 (23) [6] Bolla et al., CTM 8 (24) [7] Schlatter et al., Procs Dessau Conf (2) [8] Bolla et al., Procs COMODIA (22) [9] Bolla et al., SAE World Congress (24) [] De Paola et al., CST 8 (28) [] Wright et al., SAE Int. J. Engines 2 (29) [2] Wright et al., Procs ASME ICCMSE (2) [3] Bolla et al., FUEL 7 (24) [4] Farrace et al., SAE Int. J. Engines 6 (23) [5] Farrace et al., SAE Int. J. Engines 7 (24) [6] Pandurangi et al., SAE Int. J. Engines 7 (24) [7]

10 Non-premixed CMC: Spray Combustion Chamber (SCC) [] Experiment [] Simulation [2] axial distance (mm) Liquid spray region.25 ms d=.875mm 8bar 8K.75 ms ms experiment simulation.75 ms Flame evolution 2 ms 6 ms d=.875mm 9bar 9K [] K. Herrmann et al., CIMAC World Congress, 27 [2] M. Bolla et al., SAE World Congress,

11 Concept of Multi-mode Combustion Modelling (MC-CMC) dual-fuel combustion non-premixed combustion mode transition premixed combustion CMC non-premixed CMC double-conditioned CMC premixed

12 Premixed combustion in real applications Turbulent premixed combustion - regimes from Siemens (Workshop, Cambridge, 25th June 25) Gas turbines (a) (b) (c) (e) (d) (c) (b) IC engines (d) (a) (e) Simulation: Aspden et al., The Astrophysical Journal 689,

13 Premixed CMC: Model formulation flow field!u,!p, k,!!ε,!c, c! 2,Z! i to CMC ρ ζ Q α t CMC + ρu i ζ Q α x i + ρu i Y α ζ P! ( ζ ) P! ( ζ ) x i initial conditions T, Y α Τ required to CFD!!ω c and c!!ω c =!ω α ζ!ω c ζ Q α ζ + ρn c ζ 2 Q α ζ 2 CH 4 O 2 ζ flow field!!ω c =!ω c ζ P! ( ζ ) ζ c!!ω c = ζ!ω c ζ P! ( ζ ) ζ!c!ω c ζ P! ( ζ ) ζ Y! α = Yα ζ P! ( ζ ) ζ convolution N c ζ =! ε c f ( ζ ) f ( c) p( c) dc iterate until CMC convergence get!ω c ζ and Y α ζ tabulated in CHEMKIN linear interpolation or tabulated in CHEMKIN

14 Premixed CMC: Bunsen flame F2 - configuration Bunsen piloted burner (from Chen et al., Combustion and Flame 7, 996) flame F F2 F3 U (m/s) Re (-) τ chem (ms) δ L (mm) τ turb (ms) l turb (mm) F F2 F

15 Premixed CMC: Bunsen flame F2 - results Cold flow (flow field validation) U/U.5.5 X/D =.5 k/k 5 5 X/D =.5 X/D = X/D = X/D =8.5 X/D =8.5 X/D =6.5 X/D =4.5 X/D = X/D = X/D =6.5.5 X/D =4.5 5 X/D = X/D =2.5 5 X/D =2.5.5 Exp CFD r/d 5 Exp CFD r/d Experiment: Chen et al., Combustion and Flame 7,

16 Premixed CMC: Bunsen flame F2- mechanism analysis Chemical mechanism considerations in view of dual-fuel Pitsch-44 (P44) contains the relevant high temperature CH 4 paths r/d P44 GRI

17 Premixed CMC: Bunsen flame F2- mechanism analysis Chemical mechanism considerations in view of dual-fuel Pitsch-44 (P44) contains the relevant high temperature CH 4 paths r/d P44 GRI

18 Premixed CMC: Bunsen flame F2- results Turbulent flame thickness Experiment: Chen et al., Combustion and Flame 7, 996 [] Kolla et al., Combustion and Flame 57, 2 dc/dr (/m) δ t =!c r max δ t (mm) [2] Herrmann et al., Combustion and Flame 45, 26 Experiment Unstrained flamelet CMC based on CH4 CMC based on O2 Strained flamelet 2 G-equation 5 X/D=2.5 X/D=4.5 X/D=6.5 X/D=8.5 X/D= Progress variable X/D

19 Concept of Multi-mode Combustion Modelling (MC-CMC) dual-fuel combustion non-premixed combustion mode transition premixed combustion CMC non-premixed CMC double-conditioned CMC premixed

20 Multi-mode CMC: model formulation just to give an idea about the model complexity AMC model (Nguyen 2) PDF gradient model (Pope 985) algebraic mean or neglected, high Re flows algebraic model (Kolla 2) first order closure (Bilger 24) Bilger 993 Boundary and initial conditions? Probability Density Function (PDF) models? Cross-PDF? etc

21 Conclusions & Outlook Towards a complex formulation for multiple combustion modelling, two numerical frameworks have been first established: CMC for non-premixed combustion has been carefully validated over a broad range of conditions showing promising results CMC for premixed combustion has been formulated and recently implemented A first validation work has been conducted for canonical problems Other configurations will be simulated (spark ignited flames, real engines geometry) Next step is the coupling of the two models for a double-conditioned CMC formulation (ongoing work) A priori DNS studies will be conducted Double-conditioned CMC will be validated with DNS data

22 Acknowledgments Financial support from the Swiss Federal Office of Energy (BfE) and the Swiss Competence Centre Energy and Mobility (CCEM) is gratefully acknowledged. The authors further thank Prof. Mastorakos and Prof. Swaminathan (Cambridge University) for very helpful discussions

23 Thank you for your kind attention!