在切削过程中建模与仿真刀具的磨损

在切割过程中建模与仿真刀具的磨损摘要对于研究了不同刀具的磨损类型实验和分析方法仍然是主要方式。

数值方法和模拟的快速进步，联系到越来越强大的计算机的存在可能会使用有限元法研究刀具磨损。

这项工作的主要目的是提出一种新的方法来预测在切割过程中刀具磨损的操作。

特别是，能源的消耗，连接刀具磨损量与摩擦消耗所使用的能量。

另外，在诱导切削残余应力和由于磨损的机理使工具几何形状变化之间的相互作用做调查。

为了进行这项研究中，它被提交到刀具磨损的测量实验中，特别是在失量切割中。

正交切削操作使用商用有限元软件ABAQUS/ Explicit的数值模拟。

©2013的作者。

由Elsevier B.V. 发布。

根据第14届CIRP大会上的国际科学委员会负责选择和同行审查在会议的人的加工操作。

关键词：刀具磨损;数值模拟;切割;切屑形成;1.介绍、刀具磨损在加工操作中对经济有很大的影响同时也影响表面加工完整性。

事实上，刀具磨损影响刀具寿命和最终产物中的残余应力的质量。

对于这些raisons对刀具的磨损很多调查都能在文献[1-2-3]中找到。

刀具在正交切削下的磨损模拟的开发要么是验证磨损的机理。

要么是在这些模拟中，研究人员往往会更好地理解刀具磨损的残余应力对最终产物的影响[4]。

在一些研究[5-6]的在一个子程序实现刀具磨损模型，是相对的像磨损和扩散特定的磨损机理磨损。

因此，在本次调查中，具体机制被认为在很大程度上影响了磨损现象。

事实上，刀具的磨损受几个不同类材料的附着力、侵蚀、腐蚀、磨料和断裂。

在切割过程中，刀具几何形状的改变受刀具磨损的影响。

此更新的刀具几何形状主要是参照，在数值仿真，通过该工具面节点的运动[7]。

这个方法是使用一个特定的子程序的评估切削变量，如温度，正常压力，并且在正交切削模拟中每个节点工具滑动的距离。

在这之后，其他子程序启动征收节点的运动。

现有磨损模型可分为两个类型：第一种是切削参数、刀具寿命型，这样的泰勒公式，第二个是切割过程中的变量通常是基于一个或若干磨损机制[8]。

这个模型无力的，因为，一方面，磨损现象被建模为不连续的现象的时间而不是真实的情况。

在另一方面，它是在实施的的限制磨损机理，即磨损问题降低到1或2的磨损机制。

磨损接触的现象说明了通过形成之间的关系微动系统碎片和摩擦中消耗的能量。

这个耗能是更加可控制在接触区中使用量方面[9]。

这种方法是实验性的，一个摩擦磨损试验机，用于量化接触力的值，然后将能量耗散因摩擦以及与它链接遗失的能量耗散在这个区域[10]。

由于这些原因，本文提出了一种新的的方法，它提供了不仅是一个全球性的建模磨损现象，而且还是两个组合方面，正交的切割的配置中工具的磨损和在最终产物中的残余应力的影响。

为了带领这项研究中，提出的方法有三个不同的部分。

在第一部分中，一个工磨损由测量呈现。

此后，能量办法提出修改后用于在应用程序中正交切削。

一种数值模拟正交切割操作正在开发，使用了商用的有限元软件ABAQUS/ Explicit。

最后一部分，包括刀具磨损演变的数值结果在仿真和结论。

2.实验测试在实验测试中，进行验证有限元模型，并测量渐进刀具磨损，包括转制成42CD4与操作盘为70毫米的最初直径。

所使用的工具是未涂覆的用碳化钨（WC）作为TPKN1603参考PPR。

（图1）。

图一实验配置磨损测量完成使用显微镜和有列于表1，根据图2Vc是切割速度，f为进给量。

图2刀具几何形状测量位置3.数值模型3.1建模工具用于切削工具的材料是钨碳化物。

这种选择是由在实验测试中刀具的类型所影响。

因此，要比较经验和那些由数值模拟的问题的结果，需要使用同一类型的材料。

关于用于材料流动模型的方法是选择使用曲线约束塑性应变的功能。

因此，具有弹性塑料的工具建模材料的性能如图2[11]所示。

在能计算实施的损伤模型研究的软件ABAQUS/ Explicit中进行约翰逊和库克的[12-13]模型建模刀具的损伤。

所用的参数列于下表3[9]。

切削工具被定义为可变形、易损的。

切割速度是相当于施加一个刚性体。

这个想法来自于一个事实，即我们必须避免施加在可变形的速度或位移体上。

因此，要避免一切纯粹的数字问题，我们选择使用一个刚性体支撑可变形的刀具。

这两者之间的连接是功能领带。

该刀具具有10°的切削刃半径，以及一个侧翼11°的角。

几何建模是用在实验测试中对切削的插入。

因为接触问题，它确保了工具切削的部分刃口半径是无损伤（D 区）如图3所示。

此解决方案是重要的的，以保持接触在分离区和所有节点的外之间在D 区的表面上。

刀区域D刚性支撑图3几何形状和刀具网断裂能的Emax （在ABAQUS 中）开始实施是为了在相当低的正交切削中找到一个快速的刀具磨损，持续一个数值模拟几毫秒。

3.2工件建模在非常高的应变、应变率和温度的金属切割过程中，对材料的流动有很大影响。

为了得到更好的切屑形成分析结果，它结合了所有这些方面的材料模型必要的。

在该模拟中，它被选择了约翰逊和库克的行为和本构关系[14-15-16]。

3.3接触问题面临的最重要的问题正交切削数值模式中考虑到刀具的磨损是接触管理的发展。

目前，经典模拟和它生长在的情况下接触是动态的这方面是非常敏感的，即两组节点之间，分别为节点工件和刀具的节点。

需要注意的是在ABAQUS 软件中，如何数值模型是开发的，不能处理之间的接触区域两个节点。

因此，它仅仅允许一个表面之间的接触和一组节点。

商业软件是有限的在2D 方面的联系人管理方面建模。

如图4所示。

工作表面节点刀具节点刀具表面节点图4接触问题该解决方案可以管理所有将生成的接刀具与切屑的内表面的节点。

这种方法是可能通过利用正交切削与分离模型的行定义的字段和分离，三个地方选择芯片的内表面上的能力。

如图5所示。

内表面切屑图5刀具和切屑之间的接触问题在刀具/切屑与刀具/工作接触和摩擦决定了切削功率和刀具磨损。

在这种提出了一种库仑摩擦模型被用于以恒定的0.2的摩擦系数。

4．数值模拟4.1仿真持续几秒钟的发展正交切削最后的数值模拟大多数情况下几毫秒。

这些模拟大大减少模拟时间方面有相当大的计算时间可能需要几天的时间。

所以，关于数值模拟，其中仿真时间是为了一分钟后，预期一个相当高的计算时间可能达到几个星期。

计算时间可达到几个星期。

为了开发一个数值模拟这需要几秒钟，我们必须找到减少计算时间的解决方案。

在下面，我们介绍了发达国家提供创新的解决方案来解决这个问题的模式。

在数值模拟计算时间主要需要计算之间划分的在材料内的变形和需要管理该联系人。

这取决于这个时间是可变的配方中使用。

它是已知的计算时间甚至不太重要的在下面的顺序配方，拉格朗日，欧拉和ALE。

由于接触约束的开始可以得出结论，的拉格朗日选择配方已经完成，我们不能改变它。

这要追溯到二次调查的途径是负责增加计算相对的接触管理。

因此可以推断，接触的时间更重要而不是在接触节点的数目。

下面的图，（图6），显示的情况下相对长的数值模拟50毫米，计算时间在接触节点的增加是越来越重要。

这些节点相关的切屑。

在某些情况下，有过多的扭曲，创造一个停止计算。

扭曲的网格是位于切屑。

这是由于弯曲的切屑本身。

切屑加工表面图6。

由接触引起的节点的计算时间增加现在是集中在所造成的问题是在模拟的发展中逐步产生切屑。

的解决方案是在从事实上切屑产生用于管理大量的计算时间接触的节点。

此解决方案还依赖于这一事实，切屑是一种浪费，即使它允许我们对有丰富的关于行为的信息剪切的操作。

因此，选择了消除该切屑。

因此，它不会引起在仿真中落实到计算时间管理芯片本身和原料之间的联系表面的难题。

下面的图7给出了模型使用不连续的切屑的形成，经过一段时间后切屑弹出从切削区域的时间，在切削面积的弹射节省的时间，可以大幅度减少接触表面，然后由此所得的方程。

当前部分的切屑切屑间隔工作图7减少计算时间的解决方案，e=80μm切削移动图8 从切削面积中消切屑4.2磨损数值模拟开发的模型，提供了第一个相当满意结果。

结果表明，月牙洼磨损逐渐发展为模拟研究进展。

从下面的图9，结果表明选择的解决方案来管理切屑和刀具之间的接触使刀具工作良好。

这给了令人满意的结果，在这里我们看到在该切屑包括刀具磨损和保持接触表面的月牙洼和之间的所有形成芯片的节点。

刀具切屑区域D图9数控刀具磨损仿真据观察，这种方法提供了在后刀面磨损比较差结果，因为忽视在前面的发生侧面刀具的回弹现象。

后者产生加工表面和刀具的前刀面之间的联系。

这接触效应提高摩擦系数是造成该地区高温[ 17 ]。

但在研究了进给速度和切削之间的比率刃口半径（F / R），它是可以忽略的后刀面磨损。

结论在本文中，刀具磨损建模开发从能量耗散接触区作为一个结果。

这种做法导致的实现在工具材料的损伤规律。

方法刀具的节点消除时使用后者达到的最大能量值的定义域。

该工具的第一个结果表达有限元法充满精力的方法似乎是有趣的。

当然，这是必要的使用的刀具磨损断裂能量设置完成研究。

后来，一部分是致力于发展一种新的方法，使我们能够显著降低计算所需的时间完成模拟。

方法高光的事实计算时间复杂性产生的管理切屑和刀具之间的接触。

该方法提出了保持历史加工完成的部分，因此，它是可能得到诱导的残余应力部分。

这最后一个方面将在未来的论文。

文献[1] Kgnay, Tchadj M.. 2009, Contribution àl’identification des mécanismes d’usure en usinage d’un WC-6%Co par une approche tribologique etthermique. ENSMP 2009. Thèse de doctorat. ED n°.432.[2] Nouari M., Molinari A. 2002, Modeling of tool wear by diffusion inmetal cutting, Wear 252-1, p. 135–149.[3] Poulachon G., Moisan A. et Jawahir I. S. 2001, Tool-wear mechanisms inhard turn-ing with polycrystalline cubic boron nitride tools, Wear,250(1-12) p. 576–586.[4] Muñoz-Sánchez A., Canteli J.A.,. Cantero J.L, Miguélez M.H. 2011,Numerical analysis of the tool wear effect in the machining inducedresidual stresses, 19-2 p. 872–886.[5] Yen Y. C., Söhner J., Lilly B., and Altan T., 2004, Estimation of tool wearin orthogonal cutting using the finite element analysis. Journal ofMaterials Processing Technology, 262 p. 82-91.[6] Filice L., Micari F., Settineri L., Umbrello D. 2007, Wear modelling inmild steel orthogonal cutting when using uncoated carbide tools, Wear,262, p. 545–554.[7] Attanasio A., Ceretti E., Fiorentino A., Cappellini C., Giardini C. 2010,Investigation and FEM-based simulation of tool wear in turningoperations with uncoated carbide tools, Simulation Modelling Practiceand Theory, 269, p. 344-350.[8] Xie L.-J., Schmidt, Schmidt C., Biesinger F. 2004, 2D FEM estimate oftool wear in turning operation, Wear, 258-10, p. 1479-1490.[9] Ramalho A., Miranda J.C. 2006, The relationship between wear anddissipated energy in sliding systems, Wear 260, p. 361-376.[10] Huq M.Z., Celis J.-P., 2002 Expressing wear rate in sliding contactsbased on dissipated energy, Wear 252, p. 375–383.[11] Moxnes J. F., Teland J. A., Skriudalen S., Bergsrud S. M., Sundem-Eriksen L. and Fykse H. 2010, Development of material models forsemi-brittle materials like, Norwegian Defence ResearchEstablishment (FFI), ISBN 978-82-464-1830-8.[12] Johnson R. and Cook W.K., 1983, A constitutive model and data formetals subjected to large strains, high strain rates and hightemperatures, 7th International Symposium on Balistics p. 541-547.[13] Johnson W.H. and Cook G.R. 1985, Fracture Characteristics of ThreeMetals Subjected to Various Strains, Strain Rates, Temperature andPressures, 21 p. 31-48.[14] Barge M., Hamdi H., Rech J. and Bergheau J. M., 2005, Numericalmodelling of orthogonal cutting: influence of numerical parameters,Journal of Materials Processing Technology, 164-165, p. 1148-1153.[15] Mabrouki T., Girardin F., Asad M., and Rigal J.-F. 2008, Numerical andexperimental study of dry cutting for an aluminium alloy 48, p. 1187 -1197.[16] Salvatore F., Mabrouki T., Hamdi H., 2012, Numerical andexperimental study of residual stresses induced by machiningprocesses, Int. Journal of Surface Science and Engineering, 6-1/2, p.136-147.[17] Salvatore F., Mabrouki T., Hamdi H., 2011, Numerical simulation andanalytical model-ling of ploughing and elastic phenomena duringmachining processes, Int. Journal of Surface Science and Engineering,6-3, p. 185-200.Modeling and simulation of tool wear during the cutting process AbstractExperimental and analytic methods are still the main ways to investigate different cutting tool wear types. Numerous developmentsof numerical methods and simulations, associated to the existence of more and more powerful computer make possible tool wearstudies using FEM. The main purpose of this work is to present a new approach to predict tool wear progression during cuttingoperation. In particular, an energy approach, linking the tool wear volume with the energy dissipated by friction is used. In addition,the interaction between residual stresses induced by cutting and the variation of tool geometry due to wear's mechanisms is investigated.In order to carry out this study, it is presented the experimental measurements of the wear of the tool, in particular the lost volumeduring the cut. Numerical simulation of orthogonal cutting operation using the commercial FEM code ABAQUS/Explicit is employed.1. IntroductionThe tool wear has a large influence on the economicsof machining operation and the influence in the surfaceintegrity. In fact, tool wear affect the tool life and thequality of the final product in terms of residual stress.For these raisons many investigations on tool wear arefound in the literature [1-2-3]. The simulation of wear ofthe cutting tool in orthogonal cutting is developed toeither validate a mechanism of wear. Either in thesesimulations, researchers tend to better understand theinfluence of tool wear in the impact of the residual stressin the final product [4]. In some investigations [5-6] thetool wear model implemented in a subroutine, is relativeto a specific wear mechanism like abrasion and diffusivewear. So in this investigation, a specific mechanism isconsidered as largely influencing the wear phenomena.In fact, tool wear is affected by several and differentmechanisms like material adhesion, erosion, corrosion,abrasive, and fracture. During cutting, tool geometry ischanging due the tool wear effect. This updating in thetool geometry is largely modeled, in numericalsimulation, by a movement of the tool face node [7].This method is used with a specific subroutine whichevaluates the cutting variables like, temperature, normalpressure and sliding distance from every tool nodes thatexist in the simulation of orthogonal cutting. After that,other subroutine is launched to impose node movement.The existing wear model can be classified into twotypes: the first one is cutting parameter-tool life type,such Taylor’s equation, the second one is cutting processvariable often based on one or several wear mechanisms[8]. This modeling is weak because, on one hand, thewear phenomena are modeled as discontinuousphenomena in time which is not the case in reality. Onthe other hand, it is limited in terms of the implementedwear mechanism, i.e. the problem of wear is reduced toone or two wear mechanism.The phenomenon of wear in contact is illustrated infretting system by a relationship between the formationof debris and the energy dissipated by friction in thecontact. This dissipated energy is the more controllablein terms of quantity used in the contact zone [9]. Thismethod is experimental, a tribometer is used to quantifythe value of efforts in the contact and then the energydissipated by friction and links it with the lost volume inthis zone [10].For these reasons, it is presented in this paper a newapproach, which provides not only a global modeling ofthe wear phenomena but also the combination of twoaspects, the progression of tool wear and his impact onthe residual stress of the final product in the orthogonalcutting configuration.In order to lead this study, the methodology proposedhas three different parts. In the first part, a tool wearmeasurements are presented. Afterwards, the energyapproach is presented and modified for the application inthe orthogonal cutting. A numerical simulation of theorthogonal cutting operation is developed, using thecommercial FEM code ABAQUS/Explicit. The last partincludes the numerical results of the tool wear evolutionduring simulation and the conclusion.2. Experimental testsThe experimental tests, conducted to validate theFEM model and measure the progressive tool wear,consisted of turning operation discs made of 42CD4 withan initial diameter of 70 mm. the used tool is uncoatedwith a tungsten carbide (WC) referenced as TPKN 1603PPR. (Fig. 1).Fig. 1. Experimental configurationWear measurements are done using microscope andthere are presented in table 1, according to figure 2. Vc isthe cutting speed, f the feed rate.Table 1. Experimental wear measure in case of tungsten carbide toolCutting conditions TimeFig. 2. Location of measured geometries of the tool3. Numerical model3.1. Tool modelingThe material used for the cutting tool is tungstencarbide. This choice is influenced by the type of toolused in experimental tests, which are tools made oftungsten carbide. So, to compare the results issued bythe experience and those issued by the numericalsimulation, the choice of using the same type ofmaterials is required. In regard to the law used to modelthe flow of material was chosen to use the curveconstraints functions of plastic strain. Therefore, thematerial used to model the tool has an elastic-plasticbehavior like it is acted in table 2 [11].Table 2.Carbide tungsten material modelThe damage model implemented on the calculationsoftware ABAQUS/Explicit to model the damage of thetool is the model of Johnson and Cook [12-13]. Theparameters used are given in the following table 3 [9].Table 3. Carbide tungsten material modelThe cutting tool is defined as deformable anddamageable. The cutting speed is rather imposed a rigidbody. This idea comes from the fact that we must avoidapplying a velocity or displacement on a deformablebody. So to avoid all purely numerical problems, wechose to use a rigid body that supports the deformabletool. The connection between the two is done with thefunction Tie.The tool has a cutting edge radius of 10 , and a flankangle of 11 °. The geometric modeling is on the cuttinginsert used in experimental tests. Because of contactproblems, it is ensured that a portion of the tool cuttingedge radius is without damage (Zone D) like it is showedin figure 3. This solution is necessary to maintain contactbetween all nodes in the separation zone and the outersurface of the D zone.Fig. 3. Geometry and mesh of the toolFracture energy Emax (in ABAQUS) was implementedin the tool quite low in order to found a fast tool wear innumerical simulations of orthogonal cutting that lasts afew ms.3.2. .Work piece modelingVery high strain, strain rate and temperature in metalcutting process, have strong influence on materiel’s flow. To get better chip formation analysis results, amaterial model that combines all these aspects isnecessary. In this simulation, it is opted for the Johnsonand Cook behaviour and constitutive law [14-15-16].3.3. .Contact problemThe most important problem faced in thedevelopment of a numerical model of orthogonal cuttingthat takes into account the tool wear is contactmanagement. Already, this aspect is very sensitive in theclassical simulation and it grows in the case where thecontact is dynamic, i.e. between two sets of nodes,respectively node of work piece and tool node. Note thatthe ABAQUS software, on how the numerical model isdeveloped, cannot handle the contact regions betweentwo nodes. So, it just allows a contact between a surfaceand a set of nodes. The commercial software is limitedin terms of contact management regarding 2Dmodelling. (Fig. 4).Fig. 4. Contact problemThis solution can manage the contact between allnodes of the tool and the inner surface of the chip thatwill be generated. This method is possible through theuse of a model of orthogonal cutting with a separationline defined fields and three areas of separation, wherethe ability to select the inner surface of the chip (Fig. 5).Fig. 5. Contact problem between the tool and the chipThe contact and friction at tool/chip and tool/workdetermines the cutting power and the tool wear. In this paper a Coulomb’s frictio n model is used with a constantfrictional coefficient of 0.2.4. Numerical simulations4.1. Development of a simulation lasting several secondsNumerical simulations of orthogonal cutting last inmost cases a few milliseconds. These simulations greatlyreduced in terms of simulation time have a fairly largecomputational time and can take several days. So,regarding the numerical simulations where thesimulation time is in order of one minute, it is expectedthat a fairly high computational time which may meetseveral weeks. In order to develop a numericalsimulation which takes several seconds, we must findsolutions to reduce the computation time. In thefollowing, it is presented the developed model whichprovides innovative solutions to address this problem.The computation time in a numerical simulation ismainly divided between the need to calculate thedeformations within the material and the need to managethe contact. This time is variable depending on theformulation used. It is known that the computation timeis even less important in the order of the followingformulations,Lagrangian, Eulerian and ALE.It can be concluded that the choice of the Lagrangianformulation is already done and we cannot change it,because of contact constraints imposed from thebeginning. This goes back to investigate the secondchannel which is responsible for increasing thecomputing relative to management of the contact. So it ispossible to deduce that the contact time is moreimportant than the number of nodes in contact is high.The following figure, (Fig. 6), shows the case ofnumerical simulation of relatively long pieces of about50 mm, the calculation time isincreasingly importantgiven the increase of nodes in contact. These nodes arerelated to chips. In some cases, there were excessivedistortions that create a stop calculation. The distortedmeshes are particularly located in the chip. This is due tothe winding of chip on itself.Fig. 6.Increase in computation time caused by the nodes in contactIt is now focused on the problem created by thegradually chip generated as the simulation evolves. Thesolution is in place from the fact that the chip generatessignificant computational time for the management ofnodes in contact. This solution also relies on the fact thatthe chip is something of a waste, even if it allows us tohave a wealth of information concerning the conduct ofthe cut operation. Consequently it is opted for theelimination of this chip. So that it does not causecomplications in the simulation which would bereflected in the computation time implemented tomanage the contact between the chip itself and the rawsurface. The following figure 7 presents the modeldeveloped with the discontinuous chip. After a period oftime the chip is ejected from the cutting area. Theejection of the cutting area saves time thanks to thesizeable reduction of the contact surfaces and then theresulting equations.Fig. 7. Solution to reduce the calculation time, e=80 mFig. 8. Elimination of the chip from the cutting area4.2. Numerical wear simulationsThe developed model, delivers first fairly satisfactoryresults. It shows that the crater wear develops graduallyas the simulation advances. From the following figure 9,it is showed that the solutions chosen to manage thecontact between the chip and all nodes of the tool workwell. These give satisfying results, where we see that theinside of the chip follows the tool wear and keeps acontact surface between the crater and all nodes of theformed chip.Fig. 9. Numerical tool wear simulationsIt is observed that this approach gives rather poorresults in terms of flank wear because of the neglectedthe spring back phenomenon that happening in front offlank tool. The latter generates a contact between themachined surface and the rake face of the tool. Thiscontact effect as raising the coefficient of friction that iscausing high temperatures in this area [17]. But in thepresented study the ratio between feed rate and cuttingedge radius (f/R) is high and it is possible to neglectflank wear.ConclusionIn the presented paper, tool wear modeling hasdeveloped as a result from the energy dissipation on thecontact zone. This approach leads to implementing adamage law in the tool material. The method ofelimination of nodes of the tool is used when the latterreaches the value of a defined maximum energy fields.The first results of the tool wear delivered by FEMenergetic approach seems like interesting. For sure, it isnecessary to complete the study with the setting of thetool wear using fracture energy.Afterwards, a section is devoted to the developmentof a new method that allows us to significantly reducethe computation time required to complete thesimulation. The approach high-lights the fact that thecomputation time is generated by the complexity ofmanaging the contact between the chip and work piece.This method presented in this paper keeps the historyof machining in the finished part and as a result, it ispossible to derive the residual stresses induced in thepart. This last aspect will be presented in a future paper.。