AP03_5.vp 1 Introduction The peak temperatures of burning gases inside the cylin- der of diesel engines are of the order of 2500 K. To prevent overheating, the maximum temperatures of the metal sur- faces enclosing the combustion chamber are limited to much lower values and, therefore, cooling must be provided for the cylinder, cylinder head and piston. The substantial heat fluxes and temperature nonuniformities arising from these conditions lead to thermal stresses, which further escalate the otherwise significant mechanical loading from combustion pressures. The design must take into account all these consid- erations to ensure trouble-free operation of engines, which, especially in the case of parts of complex design, requires an extended analysis based on detailed information of all the processes involved. The cylinder head is one of the most complicated parts of an internal combustion engine. It is directly exposed to high combustion pressures and temperatures. In addition, it needs to house intake and exhaust valve ports, the fuel injector and complex cooling passages. Compliance with all these require- ments leads to many compromises in design. As a result, cylinder heads tend to fail in operation (distortions, fatigue cracking) due to overheating in regions of limited cooling. In this study, we have put emphasis on the problematic regions around the valve seats and narrow bridges between valves. These regions experience especially severe thermal loading, as they receive heat not only from in-cylinder burn- ing gases during the combustion period but also from burned gases flowing through the exhaust valve and along the exhaust-port walls during the exhaust. Although the temper- atures of exhaust gases are significantly lower than peak in-cylinder temperatures, rapid movement of flowing gases there promotes the heat transfer to the walls. Most of the heat accumulated in the valve is rejected through the contact sur- face of the valve seat. Therefore, any deformations of these parts accompanied by improper contact and occurrence of leakage on the conical valve contact face dramatically increase the thermal loading of the valves and, therefore, may lead to their destruction. A detailed FE heat-transfer analysis can provide valuable information on the temperature distribution in the overall as- sembly of the cylinder head, especially in those regions where experimental data is almost impossible to gather. Moreover, this is the first logical stage of cylinder head strength analysis. In the next step, the temperature and mechanical stresses have to be analysed using temperature field and pressure (and other mechanical loads, e.g., belt pre-stress). The result- ing displacement/stress fields may be utilised for evaluation of operational conditions, e.g., contact pressure between valves and valve port uniformity, as well as strength and failure resistance of the assembly. Such information contributes to a detailed understanding of the thermal and mechanical pro- cesses in the cylinder-head assembly under engine operation, which is a prerequisite for further optimisation of engine design. 2 Cylinder-head assembly In the present study, the cylinder head of a big turbo- charged direct-injection diesel engine is analysed. The engine is used in power generation units. The basic parameters of the engine are: bore 275 mm, stroke 330 mm, maximum brake mean effective pressure 1.96 MPa, nominal speed 750 rpm. 34 © Czech Technical University Publishing House http://ctn.cvut.cz/ap/ Acta Polytechnica Vol. 43 No. 5/2003 Heat Transfer Analysis of a Diesel Engine Head M. Diviš, R. Tichánek, M. Španiel This paper documents the research carried out at the Josef Božek Research Center of Engine and Automotive Engineering dealing with extended numerical stress/deformation analyses of engines parts loaded by heat and mechanical forces. It contains a detailed description of a C/28 series diesel engine head FE model and a discussion of heat transfer analysis tunning and results. The head model consisting of several parts allows a description of contact interaction in both thermal and mechanical analysis. Keywords: heat transfer analysis, FEM, internal-combustion engine. Fig. 1: Cylinder head assembly The cylinder head (Fig. 1, link 1) is made of cast iron. It contains two intake valves (Fig. 1, link 6) and two exhaust valves (Fig. 1, link 7) made of forged alloy steel. The valve guides (2, 3) as well as the valve seats (4, 5) are pressed into the head. The exhaust-valve seats are cooled by cooling water flowing through the annular cavities around the seats. The fuel injector is situated inside the axis of the cylinder. The bottom face of the cylinder head, directly exposed to the in-cylinder gases, is cooled by special bores, which, how- ever, represent a further complication in the design of this mechanically highly loaded region of the cylinder head. 3 The FE model Our FE model includes all components mentioned above, see also Table 1. The real design of the cylinder head was slightly modified in details to enable manageable meshing. The model of the cylinder head block was created using PRO/ENGINEER 3D product development software and was imported as a CAD model, unlike the models of other compo- nents (valves, seats, valve guides and fuel-injector), which were developed directly in ABAQUS CAE. Some parts of the valves and fuel-injector were considerably simplified or com- pletely left out, as they were considered to have a negligible influence on the results. More information on the mesh ge- ometry and statistics are provided in Fig. 2 and Table 1. 4 Interactions and boundary conditions Although the thermal loadings of engine parts vary considerably in time due to the cyclical nature of engine operation, the computations were performed assuming steady-state heat fluxes evaluated on the basis of time- -averaged values. Taking into account the speed of the periodical changes and the thermal inertia of the compo- nents of the cylinder head, the temperature variations are damped out within a small distance from the wall surface (~1mm), and this simplification is therefore acceptable. The thermal contact interactions between individual parts of the cylinder head assembly are described by heat flux qAB from the solid face A to B, which is related to the difference of their surface temperatures TA, TB according to � ��q k T TAB B A� � , where k is the contact heat-transfer coefficient. The values of the coefficient used in the present analysis are summarised in Table 2. They follow the values reported in [3]. The boundary conditions of surfaces contacted by flowing gases are described as a steady-flow convective heat-transfer problem, where the heat flux q transferred from a solid surface at temperature T to a fluid at bulk temperature T0 is determined from the relation, � ��q h T T� � 0 , where h denotes the heat-transfer coefficient. This depends on the flow, the properties of the fluid and the geometry of the surfaces. Functional forms of these relationships are usu- ally developed with the aid of dimensional analysis. In the present study, the values of gas-side heat-transfer coefficients and bulk gas temperatures (i.e., for in-cylinder surfaces and intake and exhaust port walls) were obtained from a de- © Czech Technical University Publishing House http://ctn.cvut.cz/ap/ 35 Acta Polytechnica Vol. 43 No. 5/2003 Fig. 2: Mesh geometry Part Nodes Elements DC3D8 DC3D6 DC3D4 DS4 DC1D2 brick prism tetrahedron shell link Cylinder head 67 282 21 636 174 149 162 – – Inlet valve seat (2×) 4 800 3 360 – – – – Exhaust valve seat (2×) 4 640 2 880 – – – – Inlet valve (2×) 5 038 3 448 56 2 205 – – Exhaust valve (2×) 6 112 4 008 56 2 048 – – Inlet valve guide 1 560 1 000 – – – – Exhaust valve guide 1 716 1 100 – – – – Fuel injector 6 644 4 412 – 5 171 32 2 Table 1: Mesh statistics of individual parts (elements signed according to their ABAQUS names) 36 © Czech Technical University Publishing House http://ctn.cvut.cz/ap/ Acta Polytechnica Vol. 43 No. 5/2003 Fig. 3: Interactions and boundary conditions on a cylinder-head block Interaction description Link Contact heat transfer coefficient k [W m�2K�1] Contact of cylinder head vs. cylinder head gasket 1 6000 Contact of cylinder head vs. inlet valve seat 2a,b,c 6000 Contact of cylinder head vs. exhaust valve seat 3a,b,c 6000 Contact of inlet valve vs. seat 4 6000 Contact of inlet valve vs. valve guide 5 600 Contact of exhaust valve vs. seat 6 6000 Contact of exhaust valve vs. valve guide 7 600 Contact of inlet valve guide vs. cylinder head 8 6000 Contact of exhaust valve guide vs. cylinder head 9 6000 Contact of cylinder head vs. fuel injector (O-ring) 10 6000 Contact of cylinder head vs. fuel injector 11 6000 Table 2: Interactions between single parts considered in FE model (acting surfaces linked in Fig. 3 and 4) Fig. 4: Interactions and boundary conditions on other parts of the cylinder-head assembly tailed thermodynamic analysis of the engine operating cycle performed with the use of the 0-D thermodynamic model OBEH, see [1]. This analysis uses the well-known Eichelberg empirical heat-transfer coefficient correlation. The remaining boundary conditions on outside surfaces most exposed to the ambient air temperature are described using estimated values of the heat transfer coefficient; in special cases the heat-transfer is neglected. More detailed information on the used values are provided in Table 3 in con- junction with Fig. 3 and 4. The coolant-side boundary conditions for the water-cool- ed passages are based on values reported in the literature, see [4]. However, if local boiling occurs at the surface, different relationships for h must be used. Heat-transfer coefficients for boiling features even more complicated dependencies, since in addition to all the mentioned influences affecting the values of heat-transfer in convection, in boiling processes additional variables play a role, e.g., those linked to the phase change, microstructure and material of the surface. In the first computed case (A) the possibility of occurrence of local boiling effects was neglected. However, the computed results suggest that the boiling point of water was exceeded in some parts of the cooling passages. In further computations (B, C) the temperature-dependent heat-transfer coefficient was as- sumed, see Fig. 5, the values of which approximately follow, the steep dependence of the heat-transfer coefficient on the temperature of superheating, as reported in [4, 5]. The reasons for this simple approximation were twofold: the com- plexity of the problem, resulting in significant deviations of the coefficient values calculated according to the different equations, and the multiplicity of variables needed, some of which were not at our disposal. In this connection, two differ- ent dependencies � �h h T� were assumed in cases B and C, which made it possible to assess the sensitivity of the computa- tional results to the potential errors in approximation of the heat-transfer coefficient for boiling. 5 Results and discussions Experimental data provided by the engine manufacturer enabled us to make a comparison with the computed results. The temperature measurement arrangement with the posi- tioning of all the measured points is sketched in Fig. 6. The thermocouples were placed in special bores. All the bores were situated at a distance of 18 mm from the bottom margin of the cylinder head. Despite the lack of further detailed in- formation on conditions of the experiment (errors caused © Czech Technical University Publishing House http://ctn.cvut.cz/ap/ 37 Acta Polytechnica Vol. 43 No. 5/2003 Boundary condition description Link Heat transfer coefficient [W m�2K�1] Bulk temperature [K] Insulated surfaces (negligible heat-transfer rate) 20 0 (adiabatic) – Free surfaces (contact with ambient air) 29 5 320 Cooling passages 30 3 000* 350 In-cylinder surfaces 21 450 1120 Intake-port surfaces 22 800 330 Exhaust-port surfaces 23 800 700 *For cases B and C, the heat-transfer coefficient was assumed dependent on surface temperature Table 3: Boundary conditions considered in FE model (acting surfaces linked in Fig. 3 and 4) 340 360 380 400 2000 4000 6000 8000 10000 12000 14000 16000 Case A Case B Case C h e a t tr a n s fe r c o e ff ic ie n t h [W m - 2 K - 1 ] T [K] Fig. 5: Dependency of the heat-transfer coefficient in the cooling passages on surface temperature for cases A, B, and C Measured point Temperature [K] Case A Case B Case C Experimental data 1 568 488 488 425 2 558 485 485 509 3 497 430 429 442 4 517 453 453 – 5 493 434 433 412 6 528 454 453 448 7 508 444 443 415 8 553 501 501 468 9 501 443 443 394 10 553 501 501 430 11 516 448 448 400 12 489 445 443 361 13 518 486 483 414 Table 4: Comparison of computed with measured temperatures points P1 – P13 by the measuring equipment, influence of the location and fixation of the thermocouples in the bores, etc.), the authors found the data provided a useable and useful resource for verification of the presented model. Direct comparison is disabled by the fact that the experi- mental data and the computed values do not correspond exacly to the same load of the engine. Thermal bound- ary conditions for the FE analysis were computed for about 5 % higher BMEP (break mean effective pressure) than that prevailing during the experiment. Accordingly to [2], the experimental temperatures for the same BMEP as those used in FE analysis might cause an increase in observed temperatures of about 15 K. Table 4 provides a comparison of the results of all the computed cases (A, B, C) with those from measurement. The tabulated values indicate significantly closer agreement of results of cases B, C with experimental data, which confirms the importance of includ- ing the local boiling effects. The negligible differences of values computed for cases B and C indicate the insensitivity of the steady-state temperature field to any possible short- coming in the evaluation of the heat-transfer coefficient for boiling, see Fig. 5, which confirms the suitability of the ap- proach chosen for incorporating of the local boiling effects into the FE model. Due to some uncertainty about the precision of the place- ment of quite lengthy bores, the sensitivity of the calculated values to the positioning of the measuring points was tested. Therefore, for case C, all the temperatures were observed at a distance of 20 mm from the bottom margin of the cylinder head. The comparison displayed in Fig. 6 affords a rough esti- mation of the possible errors arising from the inaccurate fit of the points at which the temperatures were observed in the model to the real measuring points. The results of the calculations shown in the cross-section in Fig. 8 provide some interesting information on the temper- ature distribution within the valves and seats. Especially, there is apparently a positive influence of cooled seats of exhaust valves on the thermal loading of those parts, as the temper- ature values within the exhaust valves just slightly exceed those in the intake valves (experiencing significantly lower thermal loading). 38 © Czech Technical University Publishing House http://ctn.cvut.cz/ap/ Acta Polytechnica Vol. 43 No. 5/2003 Fig. 6: Computed temperature vs. path distance. Paths go through measured points P6, P11 defined in Fig. 7. Squares indicate measured temperature. Fig. 7: Measured points P1–P13 distribution over the head. All measurement holes are positioned 18 mm from the bottom of the head. Colored isotherms represent the computational results of case C. (measured values linked directly with points) 6 Conclusion A steady state heat transfer analysis of a C 28 big diesel engine cylinder head assembly was peformed using FEM. The model was verified using measured temperatures. The computed values correspond to the experimental data. The results of computation confirmed the need to incorporate the possible occurrence of local boiling and associated steep changes in the values of the heat-transfer fluxes. Thermal loading of the head assembly will be performed in the near future. References [1] Macek, J., Vávra, J., Tichánek, R., Diviš, M.: Výpočet oběhu motoru 6c28 a stanovení okrajových podmínek pro pevnostní a deformační výpočet dna hlavy válce. (in Czech) ČVUT v Praze, Fakulta strojní, VCJB, 2001. [2] Macek, J., Vítek, O., Vávra, J.: Kogenerační jednotka s plynovým motorem o výkonu větším než 3 MW – II. (in Czech) ČVUT v Praze, Fakulta strojní, 2000. [3] Horák, F., Macek, J.: Use of Predicted Fields in Main Parts of Supercharged Diesel Engine. Proceedings of XIX. Conference of International Centre of Mass and Heat Transfer. New York: Pergamon Press, 1987. [4] Kreith, F., Black, W.: Basic Heat Transfer. New York: Harper and Row, 1980. [5] Baehr, H. D., Stephan, K.: Heat and Mass Transfer. Berlin: Springer-Verlag, 1998. Ing. Marcel Diviš phone: +420 224 351 827 e-mail: divis@student.fsid.cvut.cz Ing. Radek Tichánek phone: +420 224 352 507 e-mail: tichanek@student.fsid.cvut.cz Department of Automotive and Aerospace Engineering Czech Technical University in Prague Josef Božek Research Center Technická 4 166 07 Prague 6, Czech Republic Ing. Miroslav Španiel, CSc. phone: +420 224 352 561 e-mail: spaniel@lin.fsid.cvut.cz Department of Mechanics Czech Technical University in Prague Josef Božek Research Center Technická 4 166 07 Prague 6, Czech Republic © Czech Technical University Publishing House http://ctn.cvut.cz/ap/ 39 Acta Polytechnica Vol. 43 No. 5/2003 Fig. 8: Temperature field on exhaust valve and seat (case C)