Approach of design influence on air flow rate through the heat exchangers

Transcription

1 Approah of design influene on air flow rate through the heat exhangers Riardo de Oliveira Silva 1, Thaís Piva de Castro 2, Paulo Marel Alves Bandeira 1, Diego Erberelli 3, Carlos Teófilo Salinas Sedano 4 1 Volkswagen do Brasil Indústria de Veíulos Automotores Ltda. 2 Engbras Software e Projetos Ltda. 3 MAHLE Behr Gereniamento Térmio Brasil Ltda. 4 UNICAMP s: riardo.silva@volkswagen.om.br, thais.astro@engbras.om.br paulo.bandeira2@volkswagen.om.br, diego.erberelli@br.mahle.om, salinas@fem.uniamp.br Abstrat The heat exhangers are responsible by the thermal equilibrium in an internal ombustion engine, but it is important to emphasize, that this property is diretly influened by the air mass flow rate whih passes through heat exhanger, given a speifi oolant flux and heat exhanger geometry. The present work aims to show the influene of vehile omponents, suh as air defletors, fan, radiator s shroud, grilles and gaskets on air mass flow rate passing through the heat exhangers and their onsequenes on ITD (Inlet Differene Temperature) estimation. This influene was quantified using CFD (Computational Fluid Dynamis) and 1-D simulation analysis of vehile under-hood. The simulations were performed in top speed ondition of the vehile, whih haraterizes full throttle of the engine and 50kW of rejeted heat. The analysis allows to state that an inrement of air flux was possible reduing leakage and redireting the air flux. 1. Introdution A model of vehile an show different versions, simply making hanges of radius of the wheels, types of tires, set of grilles and bumpers, hoods and headlights, for example. These hanges an define versions of luxury, sportive, family, off-roader or an eonomy ar. Furthermore it is important to highlight the possibility of different setups of powertrain, so that eah one of these onfigurations has signifiant influene on thermal management of the vehile and its employment in a speifi limate zone (hot and temperate lands). Suh premises of operation require some are of sizing of the ooling pakage. For a better pratie of sizing, it is reommended to use CAE resoures like CFD and 1- D analysis, whenever omponent like radiators, ondensers, fans, shrouds, air flow defletors, thermostati valves, expansion tanks and gaskets are been hosen. The appliability of the simulation (CFD) allows foreasting the quantity of air mass flow rate whih passes through the radiator, one of the variables in the formula of the heat gained by the air whih an be expressed as: Q m.( T T ) Eq. 01 a a. pa ai ao

2 Where: Q = Overall heat transfer of the air (kw); a m = air mass flow rate (kg/s); a = Speifi heat apaity of the air (kj/kg.k); pa T = Temperature of the air inlet (K); ai T ao = Temperature of the air outlet (K); The performane of a heat exhanger an be determined by examining the heat loss and gain that takes plae between its working fluids: The lost by the oolant an be expressed as: Q m.( T T ) Eq. 02. p i o Where: Q = Overall heat transfer of the oolant (kw); m = oolant mass flow rate (kg/s); = Speifi heat apaity (kj/kg.k) p T = Temperature of the oolant inlet (K); i T o = Temperature of the oolant outlet (K); The heat loss and gain are alulated by software whih uses the onept of blok s diagram (1-D simulation). This software uses the following input data: a) oolant pump map; b) fan map; ) the lift urve of the thermostati valve; d) the map of rejeted heat by the engine in different onditions of operation (engine speed, torque and power); e) the geometry of the hoses; f) the urves of bakpressure of the engine jaket; g) the urves of bakpressure of the radiator (air side); h) the urves of bakpressure of the radiator (oolant side); i) ondenser parameters, similar of the radiator; The 1-D simulation makes a ompilation of these data searhing or estimating the point of working of the vehile. The parameter ITD is defined by -NTU method of sizing of the heat exhangers whih an be expressed by: Q. C min. ITD Eq. 03 Q. C min.( T i T ai ) Where: = effetiveness; C min = The smaller apaity rate between air and oolant. C m. C a a p m. pa Where: C = Capaity rate of the oolant side; C = Capaity rate of the air side.

3 The effetiveness is given by: Q Q max Q max C min. T i T ai Where Q max is the maximum quantity of heat transfer possible. The parameter of ITD is usually employed to speify if a vehile and its ooling system is able to be used in a determinate ondition. For top speed ondition for example with 45 C of ambient temperature and 30% of relative humidity, the reommended value of ITD is 75K. More details of -NTU method an be seen in Kanefsky et al. (1999). The redution of engineering osts using simulation is widely proved and represents more one attrative point for its appliation, mainly during the phase of development avoiding osts of experimental tests and prototypes. 2. Underhood simulation for ooling system CFD simulation In priniple it is neessary to know what vehile (model and version) will be simulated and the ondition whih the vehile will be studied. In our example it is been studied a vehile in top-speed (185km/h, with 45 C and 40% of relative humidity). After, it is prepared the geometry of the vehile, in order to have a representative virtual model. In this phase of the proess of simulation, the geometry imported from CAD, (independently of the software of CAD) are inappropriate for CFD simulation. In some ase these geometri problems are inreased when the model are exported using.stl file extension. Normally these geometries show the following type of surfae problems whih an be repaired manually by the own software of CFD: Free edges. Zip edges or fill holes with new triangles. Interseting surfaes. Interset then delete surfaes. Imprint surfaes and edges onto target surfaes/bodies. Surfaes an be split and ombined to reate required boundaries. The proess to lean a surfae ould expend a lot of time, if this proedure is usually performed manually, therefore it is reommended to use some resoures to repair surfae whih is available in the pakage of the software. The software used for the proposed simulation is the Star CCM+ whih has a resoure alled wrapping. The wrapping tool tries to solve every problem found among the ells of the virtual model. In the fig. 1, extrated from Mouffouk (2014), it is possible to notie the differene between surfaes after and before the proess of wrapping. Eah one of the omponents of the vehile reeives a speifi setup of surfae mesh ondition, with a speifi surfae size.

4 This proedure is adopted just for the surfiial mesh of the vehile whih solves mainly problems suh as: Piered faes; Free edges; Non-manifold edges; Non-manifold verties. Figure 1 The differene between the imported CAD surfae after wrapper and before the wrapper [6]. The next step is to prepare virtual wind tunnel where the wrapped surfae of the vehile will be plaed. In other to avoid the influene of the tunnel on the airflow, it is adopted the dimensions of the wind tunnel indiated in the fig. 2.

5 Figure 2 Dimension of the virtual wind tunnel. It is reommended to take are with the distane between the floor of the vehile and the ground of the wind tunnel. Part of the tires should be removed, giving the effet to mold the tire on the ground. This effet is obtained making a subtration between the vehile and the wind tunnel. In the fig. 3, it is possible to observe this effet. Fig. 3 Subtration between the tire and the wind tunnel ground The same ommand subtrat is used to insert the radiators and the ondenser (porous medium) in the wrapped surfae of the vehile. This proedure is adopted in order to avoid deformation of the heat exhangers geometry during the proess of wrapping whih ould generate error during the phase of volumetri mesh generation. In the fig. 4, it is possible to see the wind tunnel with the heat exhangers Figure 4 Heat Exhangers in the tunnel.

6 The fan interfae is other important point to take are. Two surfaes are reated in order to represent the differene of pressure between the upstream and downstream. The distane between these surfaes is 0.1mm and during the proess of wrapping an initial interfae is reated (using ommand repeating). During the proess to generate a volumetri mesh the same setup is kept. The fig. 5 shows the model of the fan; in the end of the proess of volumetri mesh the surfaes upstream and downstream are jointed. Figure 5 Fan model The setup of the wind tunnel should onsider indiating the walls inlet, outlet, the floor veloity and the prismati layer (boundary layer) It is indiating in the fig. 6 the mainly onfiguration of the wind tunnel. Outflow Type: Pressure stagnation = 0 Pa Wall floor Veloity of the vehile (185 kph) Inflow Type:Veloity inlet (185 kph) Figure 5 Wind Tunnel

7 The prismati layer ould follow what it is reommended by Mouffouk (2014): For wall funtions, eah wall-adjaent ell s entroid should be loated within the loglaw layer, (30<y + <300). A y + value lose to the lower bound y + = 30 is most desirable. The volumetri mesh is performed using a tool alled remeshing, with polyhedral onfiguration; the wrapped surfae is reworked before this proess. The present flow field is mathematially desribed by ontinuity and the inompressibility via Reynolds Average Navier Stokes (RANS) equations. To predit omplex turbulent flow was adopted Realizable K- turbulene model and High-Re wall treatment. For this simulation, the spae disretization an be seen in the fig. 6. It was utilized a polyhedral mesh with approximately 17 million of ells. In order to inrease the auray and redue error during the proessing, some volume of the domain is refined reduing the size of the ells. 15m Figure 6 Volumetri mesh It is desirable to represent the intake airflow of the air filter and the exhaust gases in the tail pipe, in both ase it is possible to represent a datum plan as an inflow for intake system and an outflow for the exhaust system. The values of these mass flow rates depend on the engine ondition. In top speed these values represents full throttle of the engine. CFD setup The setup of the heat exhangers obeys the equation of Dary (Eq. 04), for porous inertial resistane ( P i ) and for porous visous resistane ( P v ). Where: dp Differential of pressure between upstream and downstream; dx Differential of thikness of the porous medium; v Veloity of the fluid. Eq. 04

8 Pressure differene (Pa) The omponents of Dary s equation are input in the software and it has been shown satisfatory results. The harateristi urve of the fan that it is given by the relation (dp/dv), where dp is the differene of pressure and dv is differene of veloity. It generates a polynomial funtion of 2 nd degree, that is shown in the fig. 7. air volumetri flow rate (m³/s) Figure 7 Charateristi urve of the fan The surfaes of the fan (upstream and downstream) should be put in rotation. For both ases (heat exhangers and fan) the values are obtained by the suppliers. The summary of the boundary onditions used in the flow field harateristis ould be seen in the table 1. Table 1: Boundary onditions used for CFD simulations of the wind tunnel Boundary type Speifi ondition Valor Veloity inlet Veloity magnitude km/h Turbulent intensity 0.01 Turbulent visosity ratio 10 Pressure outlet Atmospheri pressure Pa Floor (Wall) Relative veloity km/h The air dynami visosity was defined onstant and value of 1.85*10e -5 Pa-s. For the density of referene were adopted 1.18 kg/m 3. 1-D Analysis Basially the 1-D analysis follows the sheme shown in the fig. 8 below. The harateristi s maps of the radiator are obtained by the supplier, similar information that it is used in CFD analysis, but in this ase there is a map of heat exhange of the radiator. The main information obtained via CFD is the air mass flow rate whih is input in the software and it is able to estimate the values of ITD. This analysis is able to obtain ITD data following the sript shown in the introdution of the present work. Although it was used simplified iruit of ooling, it is possible to inrement the model, using all of possible variable that ould influene the ITD estimation, as it ould be seen in the fig.9.

9 Figure 8 Simplified sheme of 1-D analysis Figure 9 Cooling System 3. Air Defletors analysis 3.1 Air mass flow rate estimation The overview of the proedures of the simulation is neessary to explain the methodology of alulation and how these results are obtained.

10 The present ase tends to show the influene of the air flow defletor in the air mass flow rate whih passes through the heat exhangers. Case #1: The simulation was performed using air defletor as it is shown in the fig. 10 (a) and (b). Left side Right side (a) Figure 10 Side air flow defletor in the front-end and radiator set: view (a) perspetive; (b) front. Case #2: The ase 2 as inremented an upper defletor, as it ould be seen in the fig. 11 (a) and (b). (b) Upper side (a) (b) Figure 11 Upper defletor. Case #3: In the 3 rd ase was added a gasket between radiator and front-end (fig. 12). Radiators gaskets (a) Figure 12 Upper air flow defletor in the front-end and radiator set: view (a) perspetive; (b) front. (b)

11 Case #4: For this ase a hood gasket was added fig. 13 (a) and (b). Radiators gaskets (a) Figure 13 Hood gaskets. (b) In the table 2, it is possible to observe the inrement of air mass flow rate whih passes through the radiator after eah ation shown for eah ase. Table 2: Comparison of the Air Mass Flow Rate [kg/s] Case#1 Case#2 Case#3 Case#4 Condenser Main radiator Aux radiator Fan Additionally it is shown the air flux through the fan and the ondenser. Our model has an auxiliary radiator used in an interooler system, but the fous of ITD alulation is in the main radiator. 3.2 ITD estimation Based on the results obtained in the table 2, it is estimated the value of ITD. In the Table 3, it is possible to see the redution of ITD after eah ation shown in eah ase. Table 3: Comparison of the ITD [K] Case#1 Case#2 Case#3 Case#4 ITD

12 ITD [K] 4. Conlusion The values shown in the Table 4 are ompiled in the fig. 14. The values of ITD were redued with the ations; it ould be interesting when the analysis passes to onsider the osts of the omponents like the radiator and a defletor. In ase of the values of ITD is on borderline of aeptable values, it is interesting to redue the leakage of the air flow rate rather than to invest in a heat exhanger with more apaity of heat exhange of effetiveness Case1 Case2 Case3 Case4 Série1 64,61 64,09 62,71 60,59 Figure 14 ITD estimation 5. Referenes [1] Lisa L., Helena M. Aerodynami Drag Redution of a Heavy Truk with Variable Cooling Air Intake Area. Master s Thesis in the Master s programme Automotive Engineering. CHALMERS UNIVERSITY OF TECHNOLOGY - Department of Applied Mehanis-Division of Vehile Safety. Göteborg, Sweden [2] Abiodun Matthew Amao. Mathematial model for Dary Forhheimer flow with appliations to well performane analysis. Master of Siene Thesis, Texas Teh University, August [3] GILLESPIE, T. D. Fundamentals of Vehile Dynamis. Warrendale, PA: Soiety if Automotive Engineers, In, [4] Kanefsky P., Nelson V. A., Ranger M., A Systems Engineering Approah to Engine Cooling Design The 44 th L. Ray Bukendale Leture SAE SP [5] Quim N., Desenvolvimento de Uma Metodologia de Simulação Apliada ao Sistema de Arreimento Veiular USP 2007.

13 [6] Mouffouk, M. A., Aerodinnami Development of the Seguraing F1-R01 Prototype Using CFD- University H.L of Batna Department of Mehanial Engineering Speialty Aeronautial Engineering Batma/Alderia 2013/2014