Numerical and experimental analysis of quenching process for cam manufacturing

J. Cent. South Univ. Technol. (2010) 17: 529- -536包SpringerDOI: 10.1007/11771-010-0518-0Numerical and experimental analysis of quenching process forcam manufacturingTANG Qian(唐倩)', PEI Lin-qing(裴林清)'，XIAO Han song(肖寒松)21. State Key Laboratory of Mechanical Transmissions, Chongqing University, Chongqing 400044, China;2. Department of Mechanical and Industrial Engineering, University of Toronto, Toronto, Ontario MSS 3G8, CanadaO Central South University Press and Springer Verlag Berlin Heidelberg 2010Abstract: In order to obtain satisfactory mechanical properties for the cam used in high-power ship diesel engines, a new quenchingtechnology was proposed by designing a two-stage quenching process with an alkaline bath as the quenching medium. Tdemonstrate the effectiveness of the proposed new quenching technology, both numerical analysis and experimental study wereperformed. The new quenching technology was analyzed using finite element method. The combined effects of the temperature,stress and microstructure fields were investigated considering nonlinear material properties. Finally, an experimental studyperformed to verify the effctiveness of the proposed new quenching technology. The numerical results show that internal stress isaffected by both thermal stress and transformation stress. In addition, the direction of the internal stress is changed several times dueto thermal interaction and microstructure evolution during the quenching process. The experimental results show that the proposednew quenching technology significantly improves the mechanical properties and microstructures of the cam. The tensile strength, theimpact resistance and the hardness value of the cam by the proposed new quenching technology are improved by 4.3%, 8.9% and3.5% compared with those by the traditional quenching technology. Moreover, the residual stress and cam shape deformation arereduced by 40.0% and 48.9% respectively for the cam manufactured by the new quenching technology.Key words: quenching process; cam manufacturing; finite element method; numerical; simulation experimental studyquenching technology consists of two stages of1 Introductionquenching，with an alkaline bath as the quenchingmedium. It is believed that the proposed new quenchingAs an essential part of a high-power ship dieseltechnology brings major advantages over the traditionalengine, a cam must have a working surface with enoughquenching process， including better mechanicalhardness, wear resistance, and toughness in order to haveproperties, better microstructures, and more accuratea guaranteed lifespan when working under severeshape and dimensions.conditions, including heavy load, high temperature, anThe quenching process is a complicated process infatigue. To meet the requirements for mechanicalwhich interaction occurs among coolingproperties of the cam, a quenching process is widelytemperature variation, phase transformation, andused in the manufacturing of cams. However, traditionalstress- strain state. For various microstructures during thequenching process, the thermophysical properties ofmanufacturers cannot provide satisfactory mechanicalmaterials (including thermal conductivity, heat transferproperties for the cam. This is indicated by spocoefficient, and specific heat at a constant pressure) andcorrosion, cracks, and rapid erosion on the workingmechanical properties (including elastic modulus, plasticsurface of the cam within the designed work cycle.modulus, Poisson ratio, and yield strength), changeMoreover, traditional quenching processes， whichcontinuously with the temperature variation. Since thegenerally consist of only one stage of quenching atprocess involves high nonlinearity of the cam materialspecified temperature， have difficulty in accurately[1]， it is very difficult to accurately analyze theanalyzing the temperature and stress fields during thetemperature and stress fields during the quenchingquenching process of the cam. In recent years, aprocess of the cam. In recent years， a number oftechnology has been proposed in this subject. Differentnumerical investigations have been performed to predictfrom the traditional quenching process, the newthe changes of temperature and stress fields of the camFoundation item: Project(50875268) supported by the National Natural Science Foundation of Chin中国煤化Iedbyv Foundationof Chongqing Science and Technology Commission, China; Project( 108107) supMYHCNMH GY of Educaio ofRecedae 2000 Poxss 01 258 200 % he Naioanl Sine Fund lor DsinguishsCorresponding author: TANG Qian, PhD, Professor; Tel: +86- 23- -65111276; E-mail: tqcqu@cqu.edu.cn530J. Cent, South Univ. Technol. (2010) 17: 529- -536during the quenching process [2- -8]. For example, YAO(PDE) of heat conduction is expressed as follows byet al [3] studied the oil quenching process of GCr15 tube .using Fourier law:using finite element method after determining the heatJTdn0T、(dT、d1d7-Q=0transfer coefficientof oil. The temperature,microstructure and stress variations during quenching(1)were analyzed. HOSSAIN et al [4] predicted the residualThe initial condition and boundary condition areand thermal stresses that occur during water quenchingof solid spherical balls. KANG and IM [6] established agiven as3D thermo-elastic-plastic model for plain-carbon steelT(x,y,z)|=o = T(x,y,z)(2dTof numerical simulation. KAKHKI et al [9] simulated thex”.+k,0n,+k2-n2 =h(T -T)(3)continuous cooling and kinetics of phase transformationand predicted the final distribution of microstructureswhere p is the density of the material; c is the specificand hardness in low alloy steels using measured heatheat capacity of the material; kx, kry, and k; are thermaltransfer coefficients of quenching media by an inverseconductivity coefficients along three directions of x, y,method. ULYSSE and SCHULTZ [10] investigated theandz; t is the time instant; Q is the latent heat of phaseeffect of various surface treatments on the thermo-transformation; n is the normal direction; h is the wholemechanical response of cylindrical aluminum missilesheat transfer coefficient; T, Ta, and To are the surfaceduring warm water quenching. In the review of the pasttemperature, medium temperature， and initialstudies on analyzing the quenching process, most oftemperature of the cam, respectively.studies were solely focused on numerical simulation forone-stage liquid-quenching. Only a few studies [9- 12]2.2 Microstructure fieldconducted experimental investigations to verify theBecause of the high level of presence of nickel (Ni)accuracy of simulation results. Moreover, the geometryand chromium (Cr) in the cam specimen, an appropriateof the specimens analyzed in those previous studies washardenability is expected. Therefore， the specimenquite simple, such as cylinders or thin plates. .contains a large amount of martensite, a small amount ofIn this work, a new quenching technology wasaustenite, and very lttle carbide after the quenchingproposed by using a two-stage quenching process withprocess with the alkaline bath. Only martensitean alkaline bath as the quenching medium. Totransformation is considered in the numerical analysis.demonstrate the effectiveness of the proposed quenchingThe martensite volume fraction can be expressed astechnology，both numerical and experimental studiesfollows by using the dynamic equation of temperature-were performed. The analytical models were establishedvarying martensite transformation [13]:to investigate the coupling effects of the microstructure,f=1-exp[-k(TMs-T)](4)temperature, and stress fields. The nonlinearity of thecam material was considered by combining the measuredwhere TMs is the temperature when martensite startsheat transfer coefficient and the empirical physicaltransforming; Ti is the temperature of specimen; k is theproperties. The finite element method was then appliedrate of martensite transformation, which is selected as .to analyzing the temperature and stress fields of the cam0.011 in this work.during the quenching process. To verify the effectivenessof the new quenching technology, an experimental study2.3 Stress and strainwas performed.The quenching process consists of both elasticdeformation and plastic deformation. Due to the2 Analytical modelcomplexity, it is difficult to establish the constitutiverelation of the material. Generally, the total strain2.1 Temperature fieldincrement is considered to be the sum of the followingcomponents: elastic strain increment, plastic strainproblem with convective boundary conditions andincrement, and strain increment caused by volumeinternal heat source. The heat flux across a boundarychange, which is expressed as:layer on the surface is dependent on the surfacede; =dej +def +de} +de;(5)temperature, temperature of the quenching medium, andwhere superscri中国煤化工nt the elasticthe convective heat transfer coefficient of the quenchingstrain incremenicNMHGchermalsrin.medium. Considering the influence of latent heat fromincrement, and yuasu L ausiutiatiui ou ain increment,phase transformation, the partial differential equationrespectively.J. Cent. South Univ. Technol. (2010) 17: 529- -536531The phase transformation increment is evaluated as3 Finite element analysisde';" =β"dV(6The finite element method is applied to analyzingβ'=β° +(am-ax)T(7)the two-stage quenching process for the proposed newquenching technology. Fig. 1 illustrates the procedure ofThe thermal strain increment is calculated as [15]:the finite element analysis (FEA) for the prediction of thetemperature，microstructure and stress fields throughde) = m-adT(8)considering thermal interaction, microstructure evolutionnd elastic-plastic deformation during the quenchingwhere β° is the volume expansion coefficient foprocess. The physical model of the cam specimen andmartensite transformation at temperature T; β" is volumemesh system in the FEA is shown in Fig.2. The .computerized model of the cam is meshed usingexpansion coefficient for martensite transformation a0 °C; QM and aA are the thermal expansionhexahedral element, and the whole model consists of36 570 elements. It is assumed that all the elements arecoefficients of martensite and austenite, respectively; V isthe volume of martensite transformation; m is theisotropic and have a linear displacement field.parameter determined from experiments; and a is theStartcoefficient of thermal expansion.When the material stress appears on the yield| Input 3D modelsurface, the elasto-plastic behavior must be taken intoMeasured heatBoundaryaccount if the load further increases. The elasto-plastictransfer- - FE analysis-conditions.stress- strain relationship is derived as follows bycoefficientconstraint and loadCalculate temperatureassuming the strain is small [16] .dσj = D器dEuNoIs phasetranstormationand Dijx and Dik arefinished2the elastic and plastic matrices, respectively; dσj is theJYesstress increment; subscript kl indicates elasto-plastic|Calculate stress/strainbehavior.The explicit expressions of the isotropic materialConverge?can be given as: .]YesS=[$，sy s2 Txv Tyz τx]T(10)( EndFig.1 Flowchart for analyzing quenching process of cam9G2 ssTDp=σ(3G+ Ep)(11)D。=1+)(1-21)*E「1-v v0)~C0 0.5(1-2v)0.5(1-2v)[(0.5(1-2v)」(12)where S; is the component of deviatoric stress tensor; T; isFig.2 Finite element model of camthe component of shear stress; E is the elastic modulus;σ; is the yield limit; and Ep is the hardening coefficient;The materi中国煤化工el. The majorD。and D, are the elastic matrix and the plastic matrix ofgeometry ofYHCN MH GS fllos Inexplicit expressions, respectively; v is the Poisson ratio;x-axis direction SIIUw1 111 rig.c, ule lal diameter isand G is the shear modulus.269 mm, the outer diameter is 405 mm, and the thickness532J. Cent, South Univ. Technol. (2010) 17: 529- -536is 107 mm. Without loss of generality, the followingA, B, and C on the surface of the cam (shown in Fig.2)assumptions are made in solving the problem.for the proposed new quenching technology.(1) During the thermal analysis, the initial900temperatures of the inner parts are the same and heat▲一Node Atransfer coefficient of the outer surface of the cam is800中，一NodeB，■- Node Cuniform.700(2) The temperature of the alkalie is kept at the只600initial temperature throughout the quenching process.(3) During the thermal -stress analysis, the initial500stress of the material prior to quenching is set as zero,and the nodal displacement of the end face is also set aszero.200 IThe suface heat transfer coefficient plays alimportant role in accurately calculating the stress and100temperature fields. Therefore， an experiment was102025 30conducted to measure the surface heat transferTime/mincoefficient of the alkaline bath. The experiment wasFig.4 Cooling history of selected nodes on cam surfacedesigned as follows: a 1Cr18Ni9Ti stainless steel probewas uniformly heated to 1 000 °C, and then cooled downDue to the interaction between the radiation of theusing the alkaline bath. The data measured by thecam and the convection heat transfer of the medium, thethermocouple were stored. After the experiment, thtemperatures of all the three nodes vary with a greatsurface heat transfer coefficient of the alkaline bath wasgradient. At the beginning of the cooling process, nodesobtained by using the inverse heat conduction method toA, B, and C are cooled quickly at a similar speed becauseprocess the measured data [17]. The obtained curve forof direct contact with the alkaline liquid. As the quenchingthe surface heat transfer coefficient is shown in Fig.3. Itprocess continues, the heat sources nodes obtained fromcan be observed in the figure that the coefficient variesthe core are different due to the change of microstructurewith the surface temperature of the specimen, and theield and the heterogeneity of the structure. Node Aheat transfer capacity reaches the highest level at aroundobtains the maximum and cools the slowest; node C650 C. The curve of the measured surface heat transferobtains the minimum and cools the fastest.coefficient，together with various thermophysicalIt is clearly shown that the temperature differencesparameters obtained from Ref.[18], is incorporated intobetween the three nodes decrease gradually during thethe analysis procedure of the quenching process toquenching process.improve the accuracy of the numerical results.It can be observed that there is a protrusion for allthe three curves shown in the figure. For example, for thecurve of node C, the protrusion can be seen at around4.0250 C. The reason is that when the temperature is甲3.5-around 250 C, martensite transformation will occur.Therefore, the latent heat generated by microstructures .3.0-causes a temperature increase on the surface and hence2.5the curve shows a protrusion point. This phenomenonagrees with the known practical process.Fig.5 shows the calculated temperature history at anode on the cam surface. For comparison, an experiment1.(was performed to measure the temperature of the0.5-corresponding node on the surface of the cam specimenby using infrared thermography00400 600800 1000 .Temperature/Cquenching process. The measured temperature history isFig.3 Surface heat transfer cofficient of alkaline bath .also included in the figure. Compared with the measuredtemperature, the simulation results are quite consistent4 Simulation results and discussionwith the measur中国煤化工mall margins.In addition, Fitemperaturefield distributiorTYHC NMH G_4.1 Temperature field. ...... of 200 s afterFig.4 ilustrates the temperature variations of nodesthe quenching starts.J. Cent. South Univ. Technol. (2010) 17: 529- -536_533900 |800“一Center node一Calculated600Surface node■Measured700: 400ξ 600-g 500200。400? 300-2001000105202530-6001015202530Time/minFig.5 Comparison of temperature variations of node on camFig.7 Axial stress of center node and surface nodesurfacetensile stress and then reaches up to 700 MPa, as shownin Fig.7. At the end of quenching process, the directionof stress reverses again for the center node due tomartensite transformation. Therefore, the surface nodeAyexhibits tensile residual stress, while the center nodeexhibits compressive residual stress.5 Experimental verificationSince the temperature and stress fields of the camare obtained through the FEA of the cam quenching253.18177.307 301.432 325.558 349.683 373 808397.93422 .059 446.18470 310process, the accuracy of the numerical results must beTemperature/"checked with the experimental results. The experimentFig.6 Temperature distribution contour of quenching at 200 swas designed as follows: two groups of cam specimens(d110 mmX 130 mm) where their total numbers is 64.2 Stress field(n=6)， were manufactured using the traditionalFig.7 shows the axial stress of a center node and aquenching technology and the proposed new quenchingsurface node of the cam. The internal stress is thetechnology listed in Table 1, respectively. For the tensileoutcome of the interaction between the thermal stresstest, the samples (d16 mmX 110 mm) were chosen fromand phase transformation stress. At the beginning of thethe 1/3 R (R is the radius of specimen) to the camcooling, due to the quick surface cooling, the surfacespecimens surface according to the National Standardshrinkage is restricted by the core structure. Therefore, asGB 6397- -86, where its diameter is d10 mm in standardshown in the figure, the surface node exhibits tensiledistance. A universal tensile testing machine (CCS- -92)stress, while the center node exhibits compressive stress,was used, with the tensile speed of 3 mm/min and loadand the thermal stress quickly increases. As thecell of 50 kN. In addition, for the impact test, the depthquenching process continues, the temperature gradientwas 2 mm for U-type gap of samples (10 mmX 10 mmXdecreases, as shown in Fig.4 or Fig.5. As a result, the55 mm) selected according to the National Standardtensile stress at the surface node reduces accordingly. AtGB/T 229- 1994, an impact tester (PSW- 3000) wasthe time instant around 5 min, the surface temperatureused, and a Rockwell hardness meter was used forreduces to point TMs, where the martensite transformationhardness test. Finally, nitric acid alcohol solution of 4%occurs and hence the microstructure volume expands.(mass fraction) was applied to metallographic specimenWhile the center node is still above point TMs, and thevolume does not change dramatically. Therefore, theTable 1 Quenching technologies of different methodstensile stress on the surface node quickly changes intoTechnology Heat treatment process Quenching mediumthe compressive stress. While in the subsequent cooling,Traditionalthe thermal stress propagates to the inner layer and thetechnology中国煤化工Oilvolume of the center continues to shrink. Due to theNewMYHCNMH G .restraint from the surface structure, the stress of the815 C, 120 minAiKaline bathcenter node changes from the compressive stress to the534J. Cent, South Univ. Technol. (2010) 17: 529- -536for 15-30 s, and then a XJZ- -64 optical microscope was85-used for structure inspection and pictures were taken.Traditional technologyFigs.8-12 show the mechanical properties of the30- 一- New technologytested cam specimens. From those figures, it can beobserved that the mechanical properties obtained from25-the proposed new quenching technology are better thar0-those of the traditional one. A quantitative comparison is5|(135, 12.5)1030-- Traditional technology---- New technology0叶1010-(120,7.5)990(120, 994)100 115 130 145160 175 190970Time/min950(135, 953)ig.11 Comparison of mass fraction of retained austenitebetween new technology and traditional technology930-1.4910-Traditional technology1.2-一- New technology100 115 1314516Fig.8 Comparison of tensile strength between new technologyand traditional technology50.8-120. Traditional technology言0.6-(135, 0.49)一- New technologyr 115后3 1102.2 (120, 0.25)(120, 110.5)105100 115 130145 160 175 19(135, 101.5)e 100Fig.12 Comparison of deformation between new technology首95-90also made and listed in Table 2 to clearly present the100 115 130 145 160175 190results. Specifically for the proposed quenchingtechnology，tensile strength， impact resistance andFig.9 Comparison of impact resistance between newhardness value are improved by 4.3%, 8.9% and 3.5%, astechnology and traditional technologyshown in Figs.8, 9, and 10, respectively.6:Fig.1l ilustrates that less austenites are retainedwith the proposed quenching technology. As a result,62both the irregular deformation tendency and deformation6“(120, 61.4)quantity during the cam operation are significantlyreduced. The deformation is reduced by 48.9% compared(135, 59.3)to that by the traditional technology, as shown in Fig.12.59Fig.13 shows the microstructures of the cam center.Compared to the cam manufactured with the tradition皇58-quenching technology, the cam manufactured with the57 - Traditional technologynew technology demonstrates better microstructures.. New technologySpecifically, the structure of the quench-hardened layer56-carbide level中国煤化工satisfactoryFig.10 Comparison of hardness between new technology andmicrostructuresYHC N M H Gw quenchingtraditional technologytechnology results in better microstructures is explainedJ. Cent. South Univ. Technol. (2010) 17: 529- -536_535Table 2 Comparison of mechanical properties and deformation between new technology and traditional technologyTensile strength/Impact value/Mass fraction of retainedDeformation/TechnologyHardness (HRC)MPa(J.cm 3)austenite/%mmTraditional technology953.0101.559.312.50.49New technology994.0110.561.47.50.25Improvement/%4.38.93.48.90proposed new quenching technology is superior to thetraditional quenching technology.6 Conclusions(1) The simulated temperature field of the camusing the measured heat transfer coefficient as aboundary condition shows a good agreement with themeasured temperature.(2) The directions of the internal stress changeseveral times due to thermal interaction and6microstructure evolution over the quenching process.(3) Using the alkaline bath as the quenchingmedium has better cooling characteristics than using thetraditional oil. The cam cools quickly above the criticaltemperature (martensite transformation temperature) andslowly down afterward. The shape transformation isreduced remarkably and reasonable microstructures aregenerated.4) Mechanical properties and microstructures areFig.13 Final center microstructures of cam obtained fromimproved significantly by adopting the two-stagequenching technology.traditional technology (a) and new technology (b)as follows. During the quenching process, a greatReferencesdifference between the internal and external temperatures[1] LI Hui-ping, ZHAO Guo-qun, HE Lian-fang. Finite element methodprobably generates a high level of thermal stress and heatbased simulation of stress strain field in the quenching process []deformation. This requires a quick cooling above theMaterials Science and Engineering A, 2008, 478(1/2): 276- 290.criticaltemperature(martensite transformationTOPARLI M, SAHIN s, OZKAYA E, SASAKI s. Residual thermalstress analysis in cylindrical steel bars using finite element methodtemperature), and then a slow cooling below the criticaland artificial neural networks []. Computers and Structures, 2002,temperature. As a result, more martensites are formed.80(23): 1763- -1770.Compared with the traditional oil quenching, the new3] YAO Xin， GU Jian-feng, HU Ming-juan, ZHANG Wei-min.two-stage quenching process with the alkaline bath isNumerical simulation of the quenching process of GCr15 steel tubeable to achieve such a cooling process. Therefore, it can[]. Transactions of Materials Heat Treatment, 2003, 24(1): 78- -81.(in Chinese)remarkably reduce the quenching temperature difference4] HOSSAIN s, DAYMOND M R, TRUMAN C E, SMITH D J.and generate reasonable microstructures.Prediction and measurement of residual stresses in quenchedIn summary, the experimental study shows thasanlssteel spheres []. Materials Science and Engineering A,compared to the cam manufactured with the traditional2004, 373(1/2): 334-349.technology, the mechanical properties including tensile[5] GOR C H， TEKKAYA A E. Numerical investigation ofstrength, impact resistance, and hardness of the camnon-homogeneous plastic deformation in quenching process [D].Material Science and Engineering A, 2001, 312/319: 164- 169.manufactured with the proposed new quenchingtechnology are improved. Moreover, the shape中国煤化工，lasti plastic fniteelement modn-carbon steel indeformationis significantly decreased, and thecouple withYHC NMH Gional Jounal ofmicrostructures are more uniform and regular. IMechanical Sctences, 2001, 49(4): 423- -459.conclusion, the experimental results demonstrate that theSEN S, AKSAKAL B, OZEL A. Transient and residual thermal stress536J. Cent. South Univ. Technol. (2010) 17: 529- -536in quenched cylindrical bodies [J]. International Journal ofinite element method []. Computational Material Science, 2007,Mechanical Sciences, 200, 42(10): 2013- 2029.40(2): 282- -291.[8] CORET M, CALLOCH S, COMBESCURE A. Experimental study14] SONG Guang-sheng, LIU Xiang-hua, WANG Guo-dong， XIof the phase transformation plasticityt of 16MND5 low carbon steelXiang-qiu, LI Guo-chen. Numerical simulation of microstructure andunder multiaxial loading []. Intermational Journal of Plasticity, 2002,stress in carburizing and quenching process of 22CrMo steel []18(12): 1707-1727.Journal of Iron and Steel Research, 2006, 18(10): 36- 40. (in[9] KAKHKI M E, KERMANPUR A, GOLOZAR M A. NumericalChinese)simulation of continuous cooling of a low alloy steel to predict15] WANG De-guang, WU Yu-cheng, JIAO Ming-hua, YU Jian-wei,microstructure and hardness []. Modeling and Simulation inXIE Ting. Finite element simulation of infuence of diferentMaterials Science and Engineering. 2009, 17(4): 1-21.compacting processes on powder metallurgic products properties [].[10] ULYSSE P, SCHULTZ R w. The efect of coatings on theChinese Journal of Mechanical Engineering, 2008, 44(1): 205 -211.thermo-mechanical response of cylindrical specimens during(in Chinese)quenching [J]. Jourmal of Materials Processing Technology, 20016] WANG Xu-cheng. Finite element method [M]. Beijing: Tsinghua204(1/3): 39- 47.University Press, 2003: 1-776. (in Chinese)[1] SONGDL, GU JF, PANJ s, HU M J. Numerical;quenching of large sized blocks of 718 steel used for platic dies [J].WANG Ming-hua, XU Jun. The measurement and calculation of heatMaterials Science and Technology, 2004, 20(12): 1567-1572.transfer cofficient under cooling conditions [J]. Transactions of[12] DOLAN G P, FLYNN R J, TANNER D A, ROBINSON J S. QuenchMaterials and Heat Treatment, 2005, 26(4): 115-119. (in Chinese)factor analysis of aluminium aloys using the Jominy end quench18] TAN Zhen, GUO Guang- wen. Thermophysical properties oftechnique []. Materials Science and Technology, 2005, 21(6)engineering alloys [M]. Bejing: Metallurgical Industry Press, 1994:1-228. (in Chinese)[13] LI Hui-ping, ZHAO Guo-qun, HUANG Chuan-zhen, NIU Shan-ting.(Edited by CHEN Wei-ping)Technological parameters evaluation of gas quenching based on the中国煤化工MHCNMH G