CIMAC Congress Bergen 2010 Paper no. 39 Anders Andreasen & Stefan Mayer Basic Research Process Development R&D / Marine Low Speed MAN Diesel & Turbo 16.6.2010 < 1 >
Presentation outline Background and motivation Review of current models for S oxidation Model description and calculational setup Results Proposed model simplifications Summary & Outlook Objective: Provide a realistic, applicable model capturing the essential physics of SO 2 oxidation in large two-stroke diesel engines for in-house 0D to 3D computational codes. Obtain a better understanding of the main mechanisms involved in corrosional wear of the cylinder liner MAN Diesel & Turbo 16.6.2010 < 2 >
Background & motivation HFO average Sulfur content ~ 2.5 wt. % (4.5 % max.) Source: MEPC 57/4/24 MAN Diesel & Turbo 16.6.2010 < 3 >
Background & motivation Sulfur oxidised during combustion MAN Diesel & Turbo 16.6.2010 < 4 >
Background & motivation Acidic species transported to cylinder liner Schramm et al. SAE paper, 940818 Acidic species cause (un)desirable corrosion Corrosion controlled trough TBN of lube Challenge: Low sulfur fuel & predicting scuffing MAN Diesel & Turbo 16.6.2010 < 5 >
Background & motivation Acidic species transported to cylinder liner Acidic species cause (un)desirable corrosion Corrosion controlled trough TBN of lube Challenge: Low sulfur fuel & predicting scuffing Lube base deposits may result in bore polish and scuffing MAN Diesel & Turbo 16.6.2010 < 6 >
Background & motivation Required knowledge Acidic species formed: How, when and how much? Transport mechanism of acidic species to lube oil film Lifetime and behaviour of acidic species in lube oil (Ostwald ripening?) Lube oxidation behaviour (depleting neutralising agents) SO 3 /H 2 SO 4 SO 2 /H 2 SO 3 CO 2 /H 2 CO 3 Oil film surface Oxidation Oil oxidation Neutralization Base Neutralisation (wasteful) Lube oil film Wetting corrosion After van Helden, CIMAC 1987 Cylinder liner MAN Diesel & Turbo 16.6.2010 < 7 >
Review of current models Frozen equilibrium approach A fixed user-defined conversion Detailed kinetic mechanism Conversion,ε Ref. Mode el Experiment Frozen eq. ~20% Teetz, VDI, No. 626/1984 Fixed conversion 3-5% (user defined) 2 stroke Diesel ~4-4.5 4 stroke Diesel 2-8% Schramm, SAE 940818 van Helden, CIMAC 1987 J. J. Valente, J. F. Pessoa Amorim, CEM, Macau, June 2006 Engel et al., J. Eng. Power, vol. 101 (1979) pp. 598 Boilers 0.2-7% Hunter et al. Contract no. ARB 4-421 MAN Diesel & Turbo 16.6.2010 < 8 >
Model description Sulfur oxidation mechanism from Glarborg et al. H/O subset. 28 elementary reactions S subset. 97 elementary reactions Species thermodynamic parameters from NASA polynomials Cantera (http://code.google.com/p/cantera/) used to handle calculation of thermodynamics and integration of kinetic rate equations species(name = "SO3", atoms = " S:1 O:3 ", thermo = ( NASA( [ 1000.00, 5000.00], [ 7.075737600E+00, 3.176338700E-03, -1.353576000E-06, 2.563091200E-10, -1.793604400E-14, -5.021137600E+04, -1.118751760E+01] ) ), note = "BUR0302 J 9/65 ) # Reaction 92 reaction( "SO3 + O <=> SO2 + O2", [2.80000E+04, 2.57, 29200]) # Reaction 88 falloff_reaction( "SO2 + OH (+ M) <=> HOSO2 (+ M)", kf = [5.70000E+12, -0.27, 0], kf0 = [1.70000E+27, -4.09, 0], falloff = Troe(A=0.1, T3=1e-30, T1=1e+30), efficiencies = " H2O:5 N2:1 SO2:5 ") MAN Diesel & Turbo 16.6.2010 < 9 >
Model description Detailed kinetic mechanism coupled to multi-zone approach Post-processing of measured cylinder pressure Differential fuel amount from calc. intgr. heat release (1 CAD) Air@ λ =1 Parcel Equilibrate (HP) Heat release Fuel CAD EVO 1 2 For each CAD repeat Mix air @ air m& mix Integrate rate eq. Sulfur NO x (Zeldovich) SOI 1 1 2 CAD=1º 1 2 3 n-1 n MAN Diesel & Turbo 16.6.2010 < 10 >
Results: Tuning Tuning the mixing rate parameter (75% load) Matching calculated NO with measured Finding corresponding mixing rate and ε Measured NO x ε = 4.43 Mix rate 3.29 MAN Diesel & Turbo 16.6.2010 < 11 >
Results: Details @ 75 % load (MCR) SO 2 Main sulfur species concentrations NO concentration SO 3 SO 2 Conversion, ε Temperature MAN Diesel & Turbo 16.6.2010 < 12 >
Results: Summary and quasi-validation Range of ε Variation in fuel S content Load (%) Epsilon (%) 100 2.59 75 4.43 50 4.25 25 6.72 Experiments on 4-stroke heavy duty diesel engines Source: Engel et al., J. Eng. Power, vol. 101 (1979) pp. 598 Range ε=1.8-7.7 % (0.2-7% for boilers, Hunter et al. Contract no. ARB 4-421) Decreasingεwith increasing S content Decreasing ε with increasing load (and decreasing exhaust oxygen conc.) Good agreement with experiments! MAN Diesel & Turbo 16.6.2010 < 13 >
Results: applications to two-zone combustion in cyclic simulation CycSim: In-house C++ (Object Oriented) Cyclesimulator Zone number reduced from ~50 to 1-2 Computational effort reduced Model concept upgraded from post-processing to prediction also Slightly different, but general trends are preserved! MAN Diesel & Turbo 16.6.2010 < 14 >
Model simplifications Current detailed model computational demanding Decrease calculation time by model reduction Step 1: SO 2 + OH (+M) = HOSO 2 (+M) Step 2: HOSO 2 + O 2 = SO 3 + HO 2 Step 3: SO 2 + O (+M) = SO 3 (+M) Step 4: SO 2 + OH = SO 3 + H Step 5: SO 3 + H 2 O = H 2 SO 4 97 rate equations reduced to 5! No loss in predictions! Reaction flow analysis MAN Diesel & Turbo 16.6.2010 < 15 >
Summary & Outlook Conclusions Detailed model applied with success Qualitative agreement between calculations and experimental findings Model simplifications possible (reduction in steps, zone no. s) Future work Apply reduced model in CFD code for spatial investigations Couple results with mass transport model for lube oil film Other remaining issues Rate of SO 3 to H 2 SO 4 from atmospheric chemistry Influence of Vanadium in fuel oil (catalytic) Influence of N-chemistry on SO 2 oxidation Measurements of SO 3 /SO 2 in exhaust from large two-stroke engines MAN Diesel & Turbo 16.6.2010 < 16 >
Thank you for your attention. Questions? MAN Diesel & Turbo 16.6.2010 < 17 >