HEAT TRANSFER IN EXHAUST SYSTEM OF A COLD START ENGINE AT LOW ENVIRONMENTAL TEMPERATURE

THERMAL SCIENCE: Vol. 14 (2010) Suppl. pp. S219 S232 219 HEAT TRANSFER IN EXHAUST SYSTEM OF A COLD START ENGINE AT LOW ENVIRONMENTAL TEMPERATURE by Snežana D. PETKOVIĆ a Radivoje B. PEŠIĆ b b and Jovanka K. LUKIĆ a Faculty of Mechanical Enineerin Banja Luka Republic of Srpska b Faculty of Mechanical Enineerin Kraujevac Serbia Oriinal scientific paper UDC: 621.43.068.001.573 DOI: Durin the enine cold start there is a sinificantly increased emission of harmful enine exhaust ases particularly at very low environmental temperatures. Therefore reducin of emission durin that period is of reat importance for the reduction of entire enine emission. This study was conducted to test the activatin speed of the catalyst at low environmental temperatures. The research was conducted by use of mathematical model and developed computer proramme for calculation of non-stationary heat transfer in enine exhaust system. Durin the research some of constructional parameters of exhaust system were adopted and optimized at environmental temperature of 22 C. The combination of desin parameters ivin best results at low environmental temperatures was observed. The results showed that the temperature in the environment did not have any sinificant influence on pre-catalyst liht-off time. Key words: Exhaust emission low environmental temperatures catalyst start mathematical model Introduction For more than three decades the catalysts have been used for sinificant emission reduction. Toether with the electronic control of device and system operation in vehicles the catalysts nowadays reduce the vehicle emission up to 95%. However the problem is that the catalyst does not function satisfactorily until it reaches the activatin temperature which is 250 C 350 C. Enine operation period until the catalyst and enine reach the startin temperature is called cold start phase. Accordin to references 60% 80% of the total CO and HC emission is emitted durin this period. At lower environmental temperatures below 0 C this value is 95% 98% [1 2 and 3]. Mainly due to poor fuel evaporation in the first few cycles richer composition (mixture) is required for havin safe startin and ood drivin achievement. This leads to Correspondin author e-mail: pesicr@k.ac.rs

220 THERMAL SCIENCE: Vol. 14 (2010) Suppl. pp. S219 S232 increase of fuel consumption and emission of incomplete CO and HC combustion products. CO emission is increased due to formin of liquid fuel phase on cooler walls of the intake manifold and cylinder the so-called Wall wettin effect. Dependin on the startin temperature and intake system construction the fuel consumption is increased 5 6 times in comparison with the normal workin conditions. Reducin of emission in this period is of particular importance for reducin of the entire vehicle emission. In order to reduce emission durin the cold start phase various constructive solutions are applied in both enine and exhaust system. Two basic approaches are usually pointed out: 1. Influencin the emission source and 2. Makin impact on faster catalyst action and increase of entire efficiency on additional exhaust ases processin. In references this approach is very often called Catalyst Fast Liht-off Techniques (FLTs) [4]. Further distribution of FLTs technique may be done to: 1. Passive system which means construction alterations on exhaust system and 2. Active system which means devices ensurin additional heatin required for temperature increase of exhaust as durin enine cold phase. Passive systems include simple constructional alteration on exhaust pipes and exhaust manifold which may be used for reducin the heat losses in the atmosphere. Thus it increases exhaust as temperature before catalyst and accelerates reachin of its startin temperature. However in heavy loaded enine operation attention should be paid on nonoverheatin the catalyst and occurrence of its faster thermal aein. When desinin an exhaust system the followin parameters should be optimized: 1. Exhaust pipes number of pipes insulated or not havin one or double walls or double and air-ap selection of exhaust pipe materials selection of pipe eometric parameters (lenth of individual parts wall thickness number of pipe curves and their radii selection of insulation pipe and its thickness) and 2. Catalyst selection of catalyst location number of catalyst bricks relation between the lenth and radius cell density material density and bearin coat thickness (Wash Coat) and active layer. The listed FLTs systems provide ood results in emission reducin however it is necessary to determined which system or their combination is the best for the appropriate vehicle enine exhaust system catalyst. Reardin the hue number of constructional parameters of the exhaust system it is almost impossible to perform optimization without use of mathematical models and computers (CAE Computer Aided Enineerin) which sinificantly decrease desin period and make the product cheaper [5]. Therefore a mathematical model was desined to simulate heat and mass transfer processes in the enine pipes and the catalyst [6 7]. Since there are various processes that take place in the catalyst and enine exhaust pipes two models have been developed. One model is developed to simulate heat transfer processes in enine exhaust pipes aimed at determinin of such constructional solution of exhaust pipes (eometry pipes number insulation etc.) to achieve the least temperature loss of the exhaust as before the catalyst and thus accelerate its start.

THERMAL SCIENCE: Vol. 14 (2010) Suppl. pp. S219 S232 221 The other model is developed to simulate heat and mass transfer processes in the catalyst. The exit parameters from the exhaust pipes model are the catalyst input parameters. If there are some additional systems for the catalyst accelerated start on the exhaust system such as: electrically heated catalyst absorber or secondary air input system then it is possible to develop a special mathematical model for them too. Since the catalyst is not active it should be mentioned that durin the enine cold start chemical reactions speed is very low so they are nelected as well as the heat oriinatin from exothermal reactions. It is taken that the catalyst functions as a converter. Heat transfer by radiation is sinificant only at hih temperatures in monolith if its temperature exceeds 1000 K and in exhaust pipes and from the catalyst coatin if their temperature exceeds 700 K [7]. The main oals of the paper are to summarize our own research and to obtain preliminary results which may be used in the desin and makin of the experimental device for verification of the mathematical model. Mathematical model of heat propaation in the enine exhaust system Fiure 1 shows the main parts of the enine exhaust system and models that are used for some parts. Fiure 1. The appearance of the enine exhaust system (1D one-dimensional 2D two-dimensional 3D three-dimensional models) An unsteady turbulent flow of compressible fluids is present in the enine exhaust pipes. When the exhaust as flow throuh the enine exhaust pipe heat transfer takes place throuh forced convection. In addition to convection this model takes into account the axial conduction alon the pipe natural or forced convection and radiation from the external pipe to the environment. In the modelin process of heat transfer to the wall of exhaust pipes should be more to know: whether the pipe with one wall is insulated pipe is the double wall

222 THERMAL SCIENCE: Vol. 14 (2010) Suppl. pp. S219 S232 pipe and an air-ap between them as well as constructional characteristics (diameter and wall thickness of pipes the number of curves of pipe and their curvature radius) and pipes and insulation material (density coefficient of thermal conductivity and specific heat). Durin the process of heat transfer in exhaust pipes at the same time there are the chanes in as and wall temperature alon the pipe lenth and with the time (non-stationary chane). To define a mathematical model of the unsteady chane of as temperature and wall temperature two approaches are used: One approach is to take a uniform heat flux from as to wall (or wall on the as) at a very rapid chane and The second approach is to take a uniform wall temperature (T m = const. Isothermal wall) at the slow chane. The second approach can be used for this model. Process modelin of heat transfer in the exhaust pipe is thorouhly studied within the paper [6]. Here we keep the detail on the model of catalysts. Description of the model catalyst Monolith channel with not cylindrical shape is simulated like a cylindrical shape channel with equivalent input cross-section and lenth. Fiure 2 shows a control volume of the monolith channel. When desinin a model of heat transfer in the catalyst durin the cold start enine the followin were taken into account: 1. Convectional heat transfer from as to the catalyst channel; 2. Heat that oriinates from exothermal reactions of particular components of exhaust ases is nelected since the catalyst is not active. The catalyst is considered to function as a converter; 3. Heat transfer by radiation is sinificant only when temperatures are hiher from the monolith to the catalyst coatin if its temperature is hiher than 1000 K and in exhaust pipes and from the catalyst coatin if their temperature is hiher than 700 K; 4. The loss of heat by convection (either natural or forced) and due to radiation from the catalyst coatin on the environment (substantial at hiher temperatures); Fiure 2. Control volume of monolith 5. Uniform distribution of as temperature and channel velocity at the catalyst input; 6. Laminar as flow in the monolith channel; 7. Conduction and radiation are inconsiderable in as phase in comparison to convection;

THERMAL SCIENCE: Vol. 14 (2010) Suppl. pp. S219 S232 223 8. Gas properties are chanin dependin on as temperature and pressure whereas the catalyst material properties do not chane sinificantly due to its temperature chane; 9. Temperature and concentration of as is identical for all catalyst channels and 10. Within a short time step which is taken in calculations in one small part of monolith catalyst channel (control volume) quasi-stationary assumptions can be applied i. e. stationary chane of temperature and as concentration can be considered whereas nonstationary flow can be nelected therefore derivations by time can be nelected. Within the adopted time calculation step it can be stated that the temperature of the catalyst body does not chane which is realistic due to typical temperature inertia of the catalyst material. Based on the aforementioned assumptions calculation procedure for each time step is done in the followin way: 1. Calculatin temperature of the as phase alon the lenth of the appropriate channel and 2. Calculatin temperature of the catalyst body. Gas phase In the Cartesian coordinate system for one-dimensional problems of enery conservation equation is iven by eq. (1): T t T u x Q V c uk p S 1 V T Tk. uk c p (1) T If 0 is taken then follows: t T u x S T Tk c p (2) where: is heat transfer coefficient from as to the catalyst channel [W m 2 K 1 ] T is as temperature in the catalyst channel [K] T k is temperature of the catalyst surface [K] is mass density of the as [k/m 3 ] c p is specific heat of the as at constant pressure [J k 1 K 1 ] S is eometric surface of the catalyst as per volume unit (= S 1 /V uk ) [m 2 /m 3 ] t is time (=n t n=0123...n) [s] t is time step [s] u is velocity of fluid flow [m/s] x is axial distance from the pipe inlet (=i x i=0123...m) current coordinates [m] x is rid space distance [m] is catalyst porosity [ ] and

224 THERMAL SCIENCE: Vol. 14 (2010) Suppl. pp. S219 S232. Q is heat flux from as to the catalyst material ( dq / dt ) [J/s]. The Catalyst Material Since exothermal reactions in the catalyst are nelected durin the cold start mere equations of enery conservation are sufficient for calculation (it is not needed to consider mass transfer equations) eq. (3): 2 T Tk Q k k 1 k 2 t x. (3) Vuk k 1 c Heat flux (Q ) includes: heat transfer rate by convection from as to the catalyst (Q ) heat transfer rate by electric heatin of the catalyst heat transfer rate by convection onto the environment (Q cvo). As it was mentioned before the heat that is a result of exothermal reactions and the heat transfer rate by radiation to the catalyst coatin and from the coatin to the environment is nelected: Q S ( T 1 Q Q T ) k Q cvo cvo S 1o ( T k T ) where: k is mass density of the catalyst [k/m 3 ] c k is specific heat of the catalyst [J k 1 K 1 ] k is thermal conductivity of the catalyst [W m 1 K 1 ] cvo is heat transfer coefficient from the catalyst onto the environment [W m 2 K 1 ] and S 1o is outer surface of the catalyst [m 2 ]. Differential equations solutions Differential equations of heat transfer are solved usin the finite difference method. The chane of temperature and concentration of as in the catalyst durin a period of time are taken as a series of quasi-stationary states. Initial conditions were: t=0 0 x L= M x T k (x0)=t kp. (5) Boundary conditions were: x=0 0 t N t T (0t)=T ul (6) Since the conduction of heat throuh monolith is considerably hiher than convection on the front and rear monolith surface the heat flux is nelected on these two surfaces i. e.: o (4)

THERMAL SCIENCE: Vol. 14 (2010) Suppl. pp. S219 S232 225 Tk x T k n n x 0 Tk0 Tk1 x 0 (7) L 0 Tk i Tk i1 x n n i M (8) With the known initial condition eq. (5) the catalyst temperature in the first iteration step is known then T is calculated from the eq. for enery conservation (2) toether with the boundary condition (6) usin local analytical solution (similar to the pipes): T n n n n NTU i Tk i ( T i1 Tk i ) e (9) F NTU m c where: F is surface of the catalyst sement throuh which heat is exchaned [m 2 ]. From eq. (3) under boundary and initial conditions iven by eqs. (5) (7) and (8) temperature of the catalyst surface T k is calculated in a point of time and alon the channel lenth by application of the finite difference method. Approximation of time derivative by finite differences is done by applyin backward difference whereas the approximation of second derivatives by space is done by usin central difference [7]. Discretized form of eq. (3) combinin the eq. (4) is as follows: p (10) T Q Q ) c (1 ) V n t i n t 1 n n1 1 n1 n1 n1 cvo k i Tk i ak i ( T k i 1 2Tk i T 2 k i1 x k k uk. (11) Stability criteria are taken to be the same as in the pipe models [6 7]. For hyperbolic equations: t C u 1. (12) x For parabolic equations: d a k t x 2 where: C is Courant number [ ] t is time step [s] x is rid space distance [m] d is diffusion number [ ] and a k is thermal diffusivity of the catalyst ( k /c k ) [m 2 s 1 ]. 2 1 (13)

226 THERMAL SCIENCE: Vol. 14 (2010) Suppl. pp. S219 S232 Based on the defined mathematical model for enine exhaust pipes and the catalyst TERMO computer proramme was developed for calculation of non-stationary heat transfer in the enine exhaust system [7]. The proramme was desined in BASIC computer lanuae. The proramme calculates temperatures of ases and pipes wall i. e. catalyst in the exhaust system that may contain 10 different modules. Gas temperature at the exhaust system input (limit condition) is supposed to be iven constant or second-deree polynomials. Optionally the proramme may take (process) the measured as temperature values at enine exit or the values obtained by calculation usin the appropriate proramme for enine operatin cycle calculation. Since the calculation for the catalyst and the pipes differs the proramme contains the option to decide interactively for each calculated part whether it is about the pipe or the catalyst. Based on made choice the sub-proramme is automatically started for the appropriate calculation (the pipe or the catalyst). The proramme also contains the possibility for takin into consideration radiation durin the calculation process or its nelience. Durin the calculatin process the monitor may show a raph with chanes of both as temperatures and wall pipe temperatures and immediately follow the calculation correctness. Upon completion of calculation the proramme creates database outputs for calculation results of the selected part of the system. For some other part the proramme should be reactivated. The obtained output databases are as temperatures in the selected part (Tas.dat) material temperatures (Tmat.dat) in the selected part mean temperature of the material part (Tmid.dat) and the input data and heat balance database (pod.dat). If it is about the double wall and insulated pipe databases reardin temperature chane of the external pipe (Tmout.dat) and insulation (Tins.dat) are also obtained. Fiure 3 shows TERMO proramme block diaram. pipes.dat enine.dat TERMO Tas.dat Tmat.dat Tmid.dat pod.dat Tins.dat Tmout.dat... Fiure 3. TERMO proramme block diaram The developed model and the proramme have been further used for testin of reat number of pipes and catalyst parameters on the catalyst start as well as the testin on low environmental temperatures. Low environmental temperatures testin was so conducted that the catalyst and exhaust pipes parameters were optimized at the environmental temperature of 22 C and then testin of this system was performed at low environmental temperatures.

THERMAL SCIENCE: Vol. 14 (2010) Suppl. pp. S219 S232 227 Table 1. Characteristics of the exhaust system elements No 1 2 3 4 Exhaust system element Exhaust pipe with 2 mm air-ap Metal precatalyst = 0.9 and 400 cps Exhaust pipe with 2 mm air-ap lower than the pre-catalyst Basic catalyst with ceramic monolith = 0.9 and 400 cps Catalyst insulation Catalyst external panel L [mm] D [mm] [mm] Mater. m [k/m 3 ] m [W m 1 K 1 ] c m [J k 1 K 1 ] 200 39 1.5 Č.4571 7900 15 477 75 70 1.2 Č.4571 7900 15 477 500 39 1.5 Č.4571 7900 15 477 140 100 1800 1.5 1020 3 lass wool 200 0.03 660 1.2 Č.4571 7900 15 477 5 Exhaust pipe 200 39 1.5 Č.4571 7900 15 477 Fiure 4 shows a scheme of a tested exhaust system with all the elements and the tab. 1 shows optimized construction characteristics of the exhaust system. 1 2 3 4 5 Enine Fiure 4. Exhaust enine system scheme (see leend in the tab. 1.) Influence of environmental temperature In order to test the influence of low environmental temperatures on thermal behavior of the exhaust system testin was conducted on environmental temperatures: T o2 = 7 C and T o3 = 20 C. For each testin as temperature at the exhaust system input was adopted and it was T =700 K volumetric efficiency v = 0.4 and enine speed was 1500 rpm. Fiure 5 shows as temperature chane at the pipe exit i. e. at the pre-catalyst

228 THERMAL SCIENCE: Vol. 14 (2010) Suppl. pp. S219 S232 input. The fiure shows that with the environmental temperature lowerin as temperature at the pipe exit falls but not sinificantly. Fiure 6 shows the chane of external pipe temperature. Althouh the differences in external pipe temperatures are more obvious they still have neither sinificant influence on the heat emitted in the environment nor on total heat balance which had effect on a sliht chane of as temperature. 680 670 ] 660 [K 650 T re 640 tu 630 p e ra m 620 te 610 a s G 600 590 n = 1500 [rpm] v = 0.4 T = 700 K To1 To2 To3 0 25 50 75 100 125 150 Time [s] 380 ] 360 [K v 340 m 320 T re 300 tu e ra 280 p m 260 te e 240 ip P 220 200 To1 To2 To3 0 25 50 75 100 125 150 Time [s] (environmental temperature: T o1 = +22 C T o2 = 7 C T o3 = 20 C) Fiure 5. Chane in as temperature at the pipe no 1 exit (at the pre-catalyst input) Fiure 6. Chane in external pipe temperature in pipes 1 before the pre-catalyst 650 600 ] [ 550 K T k 500 re tu e ra p 450 m te st 400 ly a ta C350 300 250 liht of temperature n= 1500 rpm v = 0.4 T = 700 K To1/p To1/s To1/k To2/p To2/s To2/k To3/p To3/s To3/k 0 10 20 30 40 50 60 70 80 90 100110120130140150 Time [s] 700 650 600 ] 550 [K 500 T re tu 450 p e ra 400 m te 350 a s G 300 250 0 25 50 75 100 125 150 Time [s] To1/p To1/k To2/p To2/k To3/p To3/k (p beinnin s middle k end of the catalyst) (environmental temperature: T o1 = +22 C T o2 = 7 C T o3 = 20 C) Fiure 7. Chane in pre-catalyst temperature Fiure 8. Chane in pre-catalyst as temperature Fiure 7 shows the chane in pre-catalyst temperature and fiure 8 shows the chane in pre-catalyst as temperature. Fiures show that there is no bi difference in precatalyst as temperatures. The pre-catalyst has slower start at lower environmental temperatures (about 3 s 5 s slower). There is no sinificant difference in pre-catalyst temperatures at temperatures below 0 C.

THERMAL SCIENCE: Vol. 14 (2010) Suppl. pp. S219 S232 229 Chane in environmental temperature has more influence on the basic catalyst start. Fiure 9 shows chane in catalyst temperature. It is evident that at environmental temperature of +22 C the catalyst has faster start in its first part for 25 s in comparison with the environmental temperature of 7 C. Fiure 10 shows chane in catalyst as temperature at specific locations. Catalyst temperature follows as temperature. 550 ] 500 [K k T 450 re tu e ra 400 p m te st 350 ly a ta C 300 250 liht of temperature 0 25 50 75 100 125 150 Time [s] To1/p To1/s To1/k To2/p To2/s To2/k To3/p To3/s To3/k 550 500 ] 450 [K T 400 re tu e ra 350 p m te a s 300 G 250 0 25 50 75 100 125 150 Time [s] To1/p To1/ s To1/ k To2/p To2/s To2/k To3/p To3/s To3/k (p beinnin s middle k end of the catalyst) (environmental temperature: T o1 = +22 C T o2 = 7 C T o3 = 20 C) Fiure 9. Chane in basic catalyst temperature Fiure 10. Chane in basic catalyst as temperature Conclusions Accordin to the conducted testin the followin may be concluded: Environmental temperature has no sinificant influence on pre-catalyst start speed. The outer surface throuh which heat is emitted into atmosphere is small so reardless of the chane in the external pipe temperature there is no sinificant chane in the quantity of heat emitted into atmosphere. It may be considered that the heat flux between exhaust as and the catalyst is almost the same at low and normal environmental temperatures. Pre-catalyst insulation has also no sinificance on its start speed. Pre-catalyst insulation is only sinificant at hiher loadin in order to protect the neihborin parts of vehicle equipment from overheatin and air-ap insulation should be selected. A sinificantly reater influence of environmental temperature on catalyst activatin speed has been noticed in basic catalyst because the outer surface throuh which heat is emitted into atmosphere is larer hence the heat capacity of the parts before the catalyst is also reater. Nomenclature a k Thermal diffusivity of the catalyst (= / k /c k ) [m 2 s 1 ]

230 THERMAL SCIENCE: Vol. 14 (2010) Suppl. pp. S219 S232 C Courant number ( t u 1) [ ] x c specific heat [J k 1 K 1 ] c p specific heat of the as at constant pressure [J k 1 K 1 ] D diameter of pipe [m] d diffusion number [ ] L lenth of pipe [m] M number of time sements [ ] N number of lenth sements [ ] Q heat flux [J/s] S eometric surface of the catalyst as per catalyst volume unit (= S 1 /V uk ) [m 2 /m 3 ] S 1 inner surface of the catalyst [m 2 ] S 1o outer surface of the catalyst [m 2 ] t time [s] T temperature [K] V uk volume of catalyst [m 2 ] u velocity of fluid flow [m/s] Greek letters heat transfer coefficient [W m 2 K 1 ] F surface of the catalyst sement throuh which heat is exchaned [m 2 ] t time step [s] x rid space distance [m] thermal conductivity [W m 1 K 1 ] v volumetric efficiency [ ] density [k/m 3 ] catalyst porosity [ ] Subscript from as to the catalyst channel cvo from the catalyst onto the environment as k catalyst k end m material o1 environmental 1 o2 environmental 2 o3 environmental 3 p beinnin p as at constant pressure s middle References [1] Laurikko J. Exhaust Emissions Performance of In-Use TWC Cars at Low Ambient Temperatures Proceedins on CD 25 th World Automotive Conress FISITA 1998 Paris France 1998 pp. F98P090 [2] Petrović S. Kuzmanović M. Popović V. The Control of passeners cars exhausts emissions: Mobility &Vehicle Mechanics Vol 20 (1994) pp. 8 15 [3] Marsh P. et al. Application Guideline to Define a Catalyst Layout for Maximum Catalytic Efficiency SAE Paper 2001 01 0929 (2001) [4] Petković S. Mrđa J. Catalyst Fast Liht-Off Techniques for SI Enines (on Serbian lanuae) Proceedins International Scientific Symposium IPS 2001 Podorica Becici SR Yuoslavia 2001 pp. 151 160 [5] Konstantinidis A. P. Kolltsakis C. G. Stamatelos M. A. The Role of Computer Aided Enineerin in the Desin Optimization of Exhaust After-Treatment Systems Journal of automobile enineerin Proceedins Part D 212 (1998) 3 pp. 169 186 [6] Petković S. et al. Heat Transfer Modellin in the Exhaust Pipes of Enine Proceedins 12 th International scientific symposium Motor Vehicle and Enines Kraujevac Serbia 2002 pp. 201 204 [7] Petković S. Exhaust System of Enine Optimization at Low Temperature Workin Reime (on Serbian lanuae) Ph. D. Thesis Faculty of mechanical enineerin Banja Luka BiH 2002 [8] Petković S. Mrđa J. Milašinović A. Influence of Exhaust System Construction Parameters on Thermal Response of Catalyst (on Serbian lanuae) Proceedins on CD International scientific symposium IPS 2003 Podorica Bečići SR Yuoslavia 2003 Paper submitted: May 05 2010 Paper revised: May 31 2010 Paper accepted: July 31 2010