FOCUS ON THE EVOLUTION OF A LEAN NO X -TRAP MODEL FOR NO/NO 2 DIFFERENTIATION. Alexis MANIGRASSO, Pierre DARCY, Patrick DA COSTA (France)

Transcription

1 FOCUS ON THE EVOLUTION OF A LEAN NO X -TRAP MODEL FOR NO/NO DIFFERENTIATION Alexis MANIGRASSO, Pierre DARCY, Patrick DA COSTA (France) Alexis MANIGRASSO 15/09/009 MODEGAT 009

2 01P003 THE NO ISSUE 0P009 MODELING OF THE NOX-TRAP 03P013 GLOBAL KINETIC STUDY 04P018 MODEL CALIBRATION 05P03 RESULTS 06P07 CONCLUSIONS AND PROSPECTS Alexis MANIGRASSO 15/09/009 MODEGAT 009

3 01 THE NO ISSUE Alexis MANIGRASSO 15/09/009 MODEGAT 009 3

4 TOXICOLOGY AND ENVIRONMENT IMPACTS OF NO X Nitrogen monoxide NO Nitrogen dioxide NO Toxicology: NO, not very soluble in water, causes irritations of respiratory system NO, very soluble in water, reacts with water of the lungs to produce nitric acid which causes lesions and pulmonary oedema. NO is 5 times more toxic than NO Environment impacts: Ozone formation in tropospheric layers which is a powerful greenhouse gas Acid rains formation Ground eutrophication Alexis MANIGRASSO 15/09/009 MODEGAT 009 4

5 ZONES AND POPULATIONS TOUCHED BY NO Sensible zones: Major cities European capitals cities Due to transport pollution Industrial zones European industrial centers Thermal power station At risk population: Children and infants Asthmatic People with respiratory system deficiency Satellite analysis of NO over Europe. [1] [1]: Photo: European Space Agency/IUP Heidelberg Alexis MANIGRASSO 15/09/009 MODEGAT 009 5

6 NO X AND NO LIMITATIONS IN EURO 6 STANDARDS [] Current limitations for EURO 6 standards: NO X limitations: 80mg/km for Diesel type engines Combined mass of HC and NO X : 170mg/km Hypothesis for future limitations: NO limitation in percent of total NO X Off-cycle []: Commission Regulation (EC) No 69/008 of 18 July 008 implementing and amending Regulation (EC) No 715/007 of the European Parliament and of the Council on type-approval of motor vehicles with respect to emissions from light passenger and commercial vehicles (Euro 5 and Euro 6) and on access to vehicle repair and maintenance information Text with EEA relevance ; Official Journal L 199, 8/07/008 P Alexis MANIGRASSO 15/09/009 MODEGAT 009 6

7 NO FORMATION IN COMBUSTION CHAMBER NO formation Mechanisms [3] Thermal formation of NO N + O NO = NO formation NO HO NO + OH + Influence of the temperature on the speed rate of reaction (1) NO + O NO + O Reaction () favored for Diesel engines Reaction (3) inhibited by cold point on the combustion chamber NO principally produce at low engine speed and load [3]: Automotive and pollution, P.Degobert ; Society of Automotive Engineers, 31/1/1995, Book NO rate produce by a Diesel engine for different engine speed and motor load Alexis MANIGRASSO 15/09/009 MODEGAT 009 7

8 EXEMPLE OF NO EMISSIONS ( ENGINE OUT ) ON NEDC CYCLE Engine bench test NO / NOx rate (%) NO / NOx rate engine out for NEDC cycle NO/Nox rate SPEED Time ( sec ) Speed (m/s) Alexis MANIGRASSO 15/09/009 MODEGAT 009 8

9 0 MODELING OF THE NO X -TRAP Alexis MANIGRASSO 15/09/009 MODEGAT 009 9

10 PRESENTATION OF THE LEAN NO X -TRAP MODEL Modeling hypothesis Homogeneous flooding of the monolith ( monolith behavior is 1D ) Multi 0D modeling Different steps of calculation for a canal: PSR i i+1 Gas considered like perfect gas Gas properties = air properties ( viscosity, heat capacity, molar mass, ) Conduction and radiation neglected No pressure loss through the monolith ( pressure is the same all along a canal ) Quasi-steady state approximation ( QSSA ) for gaseous concentrations mass balance For the PSR number i, reactions rates calculation Pollutants mass balances and thermal balances resolution dc dt dc <<, * dt dt dt Calculations results = input of the PSR number i+1 Alexis MANIGRASSO 15/09/009 MODEGAT

11 Reaction rate: Mass balance on pollutants: Mass balance on adsorbed species: Thermal balance ( gas ): QSSA: d ( Ci ( z, t)) = dt 0 Thermal balance ( solid ): Alexis MANIGRASSO 15/09/009 MODEGAT

12 LNT MODEL ON MATLAB / SIMULINK Environment Var: (Speed, Ext Temp, ) Evolution of the gas inside catalyst Gas in composition: (Flow, Conc, Temp, ) Gas out composition: (Flow, Conc, Temp, ) Local Var in: (Tmat, init state, ) Local Var out : ( T mat, init state, ) Pressure out Pressure in Variables in defined in Matlab workspace S-Function: C language Reactions rates and balances written in the S-Function Variables out to Matlab workspace Alexis MANIGRASSO 15/09/009 MODEGAT 009 1

13 03 GLOBAL KINETIC STUDY Alexis MANIGRASSO 15/09/009 MODEGAT

14 BASIC SCHEME The study is based on the scheme given by L.Cao & al. [4]: Reversible NO oxidation on Pt: 1 NO + O = NO NO oxidation and adsorption on Ba: NO adsorption on Ba near a Pt site: NO adsorption on Ba far from a Pt site: 3 NO + O + BaCO3 = Ba( NO3) + CO 3NO + NO + CO + BaCO3 = Ba( NO3) 3NO + NO + CO + BaCO3 = Ba( NO3) [4] L.Cao & al.: Kinetic Modeling of NOx Storage/Reduction on Pt/BaO/Al O 3 Monolith Catalysts, Ind. Eng. Chem. Res. 008, 47, Alexis MANIGRASSO 15/09/009 MODEGAT

15 TWO DIFFERENT KINETIC SCHEMES TESTED NO + 1 O First scheme NO 3 NO + O + BaCO3 = Ba( NO3 ) + CO NO + 1 O Second scheme NO 3 NO + O + BaCO3 = Ba( NO3 ) + CO 3NO + NO + CO + BaCO3 = Ba( NO3) 3NO + NO + CO + BaCO3 = Ba( NO3) 3NO + NO + CO + BaCO3 = Ba( NO3) Ba ( NO NO + BaO + O 3) = Ba ( NO NO + BaO + O 3) = Alexis MANIGRASSO 15/09/009 MODEGAT

16 REACTION RATES COMMON TO BOTH SCHEMES Reversible NO oxidation: r r oxno = k oxno NO adsorption: r exp( Ea RT Thermal NO desorption: r redno adsno desno oxno )[ NO ][ O 0,5 ] ( Ea oxno + Ea eq ) = k / K ) exp( )[ NO oxno eq RT = k ( sol adsno exp( Ea RT adsno )[ NO][ BaCO EadesNO = k exp( )[ Ba( NO3) ] desno RT 1,5 3 ][ O ] ] Alexis MANIGRASSO 15/09/009 MODEGAT

17 Scheme 1: r r ads 1NO ads NO Scheme : r adsno DIFFERENCES ON THE ADSORPTION RATE Ea ads 1NO = k 1 exp( )[ NO ][ BaCO 3 ] ads NO RT Ea ads NO = k exp( )[ NO ][ BaCO 3 ] ads NO RT k = 1+ k adsno inibno EaadsNO exp( RT Ea exp( RT )[ NO inibno With [L] number of all the adsorption sites ][ BaCO [ Ba( NO ) [ L] 3 ) 3 ] ] Alexis MANIGRASSO 15/09/009 MODEGAT

18 04 MODEL CALIBRATION Alexis MANIGRASSO 15/09/009 MODEGAT

19 TESTS DESCRIPTION To calibrate the parameters of the model: Sample catalyst on Synthetic Gas Test Bench ( ~0,L ) Catalyst used: Commercial Pt/Rh/Pd catalyst on Ba/Al O 3 support Gas in ( Diesel like synthetic gas ): NO NO C 3 H 6 CO CO H O O N Gas analyzer: after the catalyst with NO X measurement first and NO in a nd test Isothermal adsorption tests with variation of total gas flow, temperature, NO/NO X ratio, concentrations of HC, CO and NO X To validate the calibration: Real condition test with Engine Test Bench Same impregnation and ageing than the sample catalyst (,L ) New European Driving Cycle ( NEDC ) Gas analyzers: before and after the catalyst with NO and NO X measurement Alexis MANIGRASSO 15/09/009 MODEGAT

20 CALIBRATION METHOD different schemes S-Function Simulink model 10 parameters: 5 k and 5 Ea Construction of a Matlab function: function y=optim(x) y: difference between model and test : Comparison between instantaneous values Comparison between cumulative values x: calibration parameters Use of an optimization tool on the optim function: lsqnonlin Alexis MANIGRASSO 15/09/009 MODEGAT 009 0

21 EXEMPLE OF PARAMETER VARIATION Parameters variation for a largest area of optimization: All the Ea are fixed before the optimization The optimization method make variations on the 5 k in the same time The best solution is kept for the previous Ea Ea values are modified The 10 parameters cannot be negative Alexis MANIGRASSO 15/09/009 MODEGAT 009 1

22 05 RESULTS Alexis MANIGRASSO 15/09/009 MODEGAT 009

23 SCHEME 1 : CALIBRATION ON CUMULATED VALUES Alexis MANIGRASSO 15/09/009 MODEGAT 009 3

24 SCHEME 1 : VALIDATION ON CUMULATED VALUES Alexis MANIGRASSO 15/09/009 MODEGAT 009 4

25 RESULTS SUMMARY SCHEME 1 SCHEME SAMPLE Calib Instant Calib Cum Calib instant Calib Cum Error Inst Cum Inst Cum Inst Cum Inst Cum TOTAL 9,9%,9% 10,0% 1,7% 9,9%,8% 10,7% 1,6% NO X 9,0%,7% 9,1% 1,0% 9,0%,5% 8,0% 0,9% NO 10,6%,9% 10,6% 1,0% 10,6%,9% 11,1% 1,1% NO 10,% 3,% 10,%,9% 10,1%,9% 13,0%,7% TOTAL 7,4% 8,4% 7,3% 6,7% 7,5% 8,4% 9,0% 38,6% REAL NO X 6,5% 8,9% 6,4% 6,8% 6,4% 8,8% 6,4% 5,5% NO 7,1% 9,4% 7,1% 7,1% 7,1% 9,3% 7,% 10,1% NO 8,7% 6,8% 8,3% 6,1% 8,8% 6,9% 13,5% 100% Alexis MANIGRASSO 15/09/009 MODEGAT 009 5

26 06 CONCLUSIONS AND PROSPECTS Alexis MANIGRASSO 15/09/009 MODEGAT 009 6

27 Conclusions: Different schemes of NO/NO adsorption on Lean NO X -Trap have been tested An automatic calibration method has been developed and tested A predictive model has been developed for NO/NO differentiation A calibration has been made with sample catalyst, and the same calibration seems to be predictive on real condition tests Prospects: To validate the calibration method with other real condition tests To make this type of study for other post-treatment models ( DOC, SCR, DPF, ) To improve the calibration method in order to reduce the calibration time To improve the optimization target to better fit curves at low / high temperatures To do specific sample tests on commercial catalyst in order to understand and improve the tested schemes Alexis MANIGRASSO 15/09/009 MODEGAT 009 7

28 END THANK YOU FOR YOUR ATTENTION Alexis MANIGRASSO 15/09/009 MODEGAT 009 8