FINISHING PROCESS OF MOLD IN COMPOSITES MATERIAL WITH AN ABRASIVE DIAMOND TOOL

Transcription

1 Original Article HOME Proceedings of IDMME - Virtual Concept 010 Bordeaux, France, October 0, 010 FINISHING PROCESS OF MOLD IN COMPOSITES MATERIAL WITH AN ABRASIVE DIAMOND TOOL Grégory CHARDON 1, Hélène CHANAL 1, Emmanuel DUC 1 (1) : Clermont Université, IFMA, EA 3867, Laboratoire de Mécanique et Ingénieries, BP 10448, F-63000, CLERMONT-FERRAND Phone: +33 (0) / Fax: +33 (0) { gregory.chardon, helene.chanal, emmanuel.duc }@ifma.fr Abstract: Composites materials made of polymer resin and fibrous reinforcements are increasingly used for aeronautical structures parts. One of the primary processes for production of composites parts is the LCM process (Liquid Composite Molding). This process is based on the resin injection in a mold usually metallic. Today, works are undertaken to produce these molds in Hextool, a carbon fiber-reinforced thermosetting plastic. Molds, realized by draping prepregs, must be finished by free-form machining to ensure the dimensional and surface quality requirements. In particular, an arithmetic roughness of 0.8µm is expected. This quality is not reached just by milling and a manual operation of polishing is required. The aim of this work is to highlight how to replace this manual operation by a machining operation with an abrasive diamond tool on CNC machine. In this paper different machining results are presented in order to show the capacity of an abrasive diamond tool to machine a mold with high surface requirements. Key words: finishing process, abrasive diamond tool, mold, composites material, grinding. 1- Introduction Composites materials made up of polymer resin and fibrous reinforcements are increasingly used for aeronautical structures parts. Increase in production volume of such parts requires an increase in productivity for primary processes, including process LCM (Liquid Composite Molding). In this context, the company Hexcel Composites has developed a new carbon fiber-reinforced thermosetting plastic named Hextool. This material has been developed in order to replace metallic material used for mold of process LCM. After their forming, molds in Hextool require a finishing process by free-form milling to fit the dimensional requirements with an arithmetic roughness of 0.8µm. Thus the work presented in this paper is focus on the definition of the finishing operation of Hextool molds. Indeed, the machining of fiber-reinforced composites is fundamentally different from the machining of metal on many aspects [CC1], [KW1], [BR1]. The cutting of these materials can lead to many problems such as peeling of fiber or pullout [RR1], [WZ1]. Thus, a previous work shows that surface quality requirements for molds can not be reached after a milling operation with carbide or PCD tools [CC1]. To reach surface quality requirements, today, a manual operation is realised. The goal is to automate this operation with a CNC machine tool to improve the productivity. Automation of polishing has many advantages and literature provides various automated experiments for metallic mold. In fact, anthropomorphic robot [WK1], parallel robots [RX1] and 3 or 5-axis NC milling machine could be used [HI1], [PT1]. In the case of automatic polishing, polishing paths should be multidirectional [TH1] and a specific attention on polishing forces is needed which generate difficulties to realise it on a machine tool. Concerning grinding of carbon fibre-reinforced plastics, some works show that is possible to reach a low surface roughness [HZ1], [PL1]. Furthermore, for grinding process forces control is not needed and tool paths could be monotonic i.e. as classical milling tool paths. Thus grinding process is easier to use on CNC machine than polishing process. However, specific grinding wheels need to be designed to be used on CNC machine and the tool geometry should be able to machine industrial mold geometries. This paper aim is to evaluate the feasibility of finishing process of mold with an abrasive diamond tool, using 5-axis machine tools. In the first section, the micro-grain metal bonded diamond tool used and the experimental procedure are presented. Then, with regard to roughness and cutting forces results, the capability of this finishing process of mold is discussed. In particular, the capacity of the tool to tolerate variation on axial depth of cut is studied. - Materials and procedure In this session, the abrasive tool used and the experimental procedure is described..1- Abrasive tool description In order to achieve a grinding operation on a CNC machine tool, a specific tool is developed in collaboration IDMME-P4-1- Copyright of IDMME - Virtual Concept

2 IDMME - Virtual Concept 010 with Asahi Diamond Industrial Europe (Figure 1). This is a micro-grain metal bonded diamond tool, made up of 75% of diamond grit. The grit size is about 5µm. During machining the tool is dressed regularly with a whetstone which is made up of glass and aluminium oxide. Its grit size is about 58µm. becomes equation () with a e the radial depth of cut and R eff the effective cutting radius (Figure 3). Thus, equation (3) ensures to calculate the theoretical surface roughness [G1]. Figure 1 : Abrasive diamond tool The tool geometry is closed to a bull nose end mill. This geometry is more suitable for free form machining than a grinding wheel. Furthermore, the tool diameter is 30mm, this value is high but is required for tool wear reasons. Indeed, for an abrasive tool, the life time is proportional to the cutting diameter [CM1]. The choice of the diameter results from a compromise between tool wear and tool accessibility. In free form machining, the tool orientation is defined with regard to the tool path and surface, so tilt and yaw angles (respectively θ n and θ t ) are used (Figure ). Thus, the surface generated presents scallops which are studied in the next section..- Theoretical scallop height and arithmetic roughness A geometrical study can be conducted in order to calculate the scallop height and the arithmetic roughness perpendicular to the feed rate direction. In fact, a finishing process by sweeping over planes induces a pattern on the machined surface. In the case of a ball end mill, Equation (1) ensures to calculate the theoretical scallop height h c for a concave surface, with R s the local radius of curvature, R the ball-end cutter radius and p the tool-path interval (Figure ) [LK1]. Real cutting edge Reff Surface to machine Effective cutting edge Calculation of effective cutting radius ae hc Rs hc Determination of scallop height Figure 3 : Theoretical scallop height [TD1], [T1] Figure : Concave surface machined by a ball-end cutter [LK1] h c R s (R s+ R) p 1 R s (R s + R R R s )p = (1) In the case of machining a plane surface with a flat end mill oriented with a tilt and yaw angles, the local radius of curvature is infinite, so equation (1) can be simplified and c a e 8R eff h = () e a R a = (3) 18 3R The geometry machined for these cutting tests is plane, but in order to be in the same conditions as free form machining tilt and yaw angles are used. In this case, the effective cutting edge of the tool is an ellipse (Figure 3). eff IDMME-P4 -- Copyright IDMME - Virtual Concept

3 IDMME - Virtual Concept 010 Locally, the effective cutting radius could be compute from the equation (4), with R the tool radius and r the tool corner radius [TD1], [T1]. r(r + rsin(θ t )) R eff = (4) rsin(θ )cos(θ ) + (R + rsin(θ ))sin(θ ) t n Thus the radial depth of cut could be now defined with regard to a specific value of arithmetic roughness expected..3- Experimental procedure The workpiece material used in the present work is a carbon FRP laminates named Hextool, made with a bismaléimide resin and carbon fibres of 7-8µm in diameter randomly disposed [CC1]. In order to carry out these experiments, laminates are previously faced with a PCD end mill. The size of a laminate is 140 mm X 00 mm, and its arithmetic roughness is about µm after milling. The tests are carried out on a 5-axis milling machine (Huron KX15). The tool is used with a classic tool holder and with a cooling by soluble oil. The grinding forces are measured with a three dimensional piezoelectric dynamometer (Kiestler 957B). After the machining, measures of roughness are made using a Mitutoyo Roughness Surftest SV 500. Surface indicator settings are as follow: The length of sampling is 4 mm for measure in the direction of machining and also in transversal direction. The radius of the diamond tracer finger is μm. The cut-off λc is set at 0.8mm. The scanning speed is 0.mm/s. A filter CRPC75 is used. The cutting speed, the feed rate and the axial depth of cut have various values during machining. Their influence on the roughness is analysed in the present study. For each test, an area of 70 mm X 16. mm is machined (Figure 5). A strategy by sweeping over planes is used with a radial depth of cut of.7 mm, a yaw angle of 1 and a tilt angle of 0. In these conditions, the value of the theoretical scallop height and arithmetic roughness are respectively 1 µm and 0.7 µm. The theoretical value of arithmetic roughness aimed is deliberately low in order to evaluate the capacity of this tool. Indeed, with a low value of radial depth of cut, geometric effects of sweeping bring to a defect which can be neglected on the measured roughness with regard to the material behaviour. Furthermore, the machining by sweeping is done in one way in order to limit the movement of rotation axis. Thus, geometric defects due to repositioning are limited. In the next session, different tests are conducted and discussed with regard to roughness and cutting forces results. t n because it is a usual value of finishing allowance. Zone N (rpm) Vc (m/min) Vf (mm/min) f (mm/rev) a (mm) e a (mm) p Tableau 1: Grinding conditions These tests are carried out without incident. The soluble oil used during this operation ensures to realise the grinding operation without loading of the abrasive tool. For each area machined, roughness measurements are conducted in the machining direction and also in the transversal direction. Average values of arithmetic roughness are identical in the two directions. This fact could be explained by the low value of radial depth of cut. Afterwards only results on transversal direction are presented. Furthermore, the dispersion on arithmetic roughness is about 0.1 µm for each area, this phenomenon is due to the heterogeneity of composite material and the random orientation of fiber [HZ1]. This is why the measurements are repeated five times for each area, the average arithmetic roughness measured appears in Figure 4. Figure 4 shows that the cutting speed and the feed rate have just a slight influence on the average arithmetic roughness obtained. However, the tool is dressed with a whetstone between the area two and three. This could be explained the stabilization of roughness for these areas. In fact, this operation cleans the tool and increases its abrasive power. Afterwards, in order to be more productive and close to an industrial application, uses of whetstone are spaced out. Ra (µm) dispersion 1 Dressed with a whetstone Vf (mm/min) 3 Average values of arithmetic roughness Theoritical arithmetic roughness Figure 4 : Roughness measurements in transversal direction 4 3- Results and discussion Fig 6(a) 3.1- Grinding operation on CNC machine A first series of tests is carried out to evaluate the feasibility of a grinding operation on CNC machine and to determinate initial cutting conditions. For that, four areas are machined with the following conditions (Table 1 and Figure 5). Cutting speed and feed rate tested here are classical values in grinding [HZ1], [ZX1]. An axial depth of cut of 0.65 mm is tested Figure 5 : Part used for tests with areas 1 to 4 and geometry 6(a) before grinding IDMME-P4-3- Copyright IDMME - Virtual Concept

4 IDMME - Virtual Concept 010 Experimental results show that arithmetic roughnesses achieved are under the expectation of 0.8 µm for each area. After this experimentation, we consider that a cutting speed of 1080 m/min and a feed rate of 300 mm/min can be used as initial conditions for the other tests. However, in order to show the capacity of this process to machine a mold, tests with variation of axial depth of cut are carried out in the next session. one more time, the variation on axial depth of cut does not seem to influence on the surface quality obtained with the abrasive diamond tool. ap=0.mm 3.- Axial depth of cut variation In the case of an industrial mold production, a finishing allowance of uniform thickness is not easy to reach after a semi-finishing operation [T]. In order to facilitate and to reduce time on semi-finish operation the finishing tool should be able to accept variation on the finishing allowance. Thus, the capacity of the tool to tolerate variation on axial depth of cut is tested here. For that, two geometries are machined with the basic conditions given above (Figure 6). The case 6 (a) permit to test an abrupt variation of axial depth of cut from 0.87 mm to 0.07 mm and then from 0.07 mm to 0.87 mm. Roughness measurements are performed on the whole surface with a specific focus on transition areas. Whatever the direction of measurement and the area studied, the average arithmetic roughness obtained is near to 0.5 µm. The diamond abrasive tool seems to tolerate sudden variation of axial depth of cut. (a) (b) ap=0.87mm ap=0.07mm ap=0.87mm Tool paths ap=0.mm ap=0.5mm Figure 6 : Geometries and tool paths presenting different axial depths of cut In order to confirm this result, another geometry is machined (Figure 6b). In this case, variation of axial depth of cut is linear from 0.5 mm to 0. mm. Furthermore, this geometry is designed with an inclined plane of 30 respect to horizontal. This geometry is obtained by a semi-finishing sweeping over planes with a 16 mm diameter flat end mill. Thus, this geometry presents a pattern, the scallop height is maximal in the inclined plane and equal to 0. mm after semifinishing (Figure 7). Roughness measurements are then carried out after finishing operation in the both direction and in all areas. The average arithmetic roughness obtained in each area is about 0.45 µm. This value is slightly lower than the case 6(a), this difference should be explained by the dressing with a whetstone practiced between the part 6(a) and 6(b). However, ap=0.5mm Pattern with scallop Fillet and inclined plane of 30 Figure 7 : Pattern after semi-finishing on geometry 5(b) During these tests, axial depth of cut varies from 0.07 mm to 0.87 mm. Gradual and abrupt variations are tested and there is no influence on the surface quality produced. The range of value of axial depth of cut tested covers many commonly cases encountered during the finishing of a mold. However, tested values are limited to 0.87 mm in order to not damage the tool. Nevertheless, values tested here are much higher than values commonly used in grinding [HZ1]. Capacity of an abrasive diamond tool to tolerate variation of axial depth of cut is well demonstrated. In the next session, productivity of the abrasive tool is studied Productivity of the abrasive tool Feed rate used in the previous section are low. Tests with feed rates from 300 mm/min to 100 mm/min are carried out here. Eight areas of 16. mm X 70 mm are finished with the abrasive diamond tool like before, with a cutting speed of 1080 m/min, and with axial and radial depths of cut respectively of 0.5 mm and 0.7 mm. The tool is dressed with a whetstone only one time and before the beginning of these tests. Arithmetic roughness measurements are given in Figure 8. Ra (µm) dispersion Vf (mm/min) 6 7 Average values of arithmetic roughness Theoritical arithmetic roughness Figure 8 : Arithmetic roughness obtained according to feed rate Feed rate does not seem to influence on the average arithmetic roughness obtained with these tested conditions. Increase of roughness between the area 1 and could be 8 IDMME-P4-4- Copyright IDMME - Virtual Concept

5 IDMME - Virtual Concept 010 explain by the dressing with a whetstone practiced before the beginning of machining. However, the average arithmetic roughness is constant and about 0.5 µm. The maximal value of feed rate (100 mm/min) used here is lower than value commonly used with PCD end mill (about 1700 mm/min for a 30 mm diameter tool with 4 teeth). However, the surface quality reached with this abrasive tool allows to reduce and even to cancel the manual polishing operation. This is not possible when carbide or PCD tools are used [CC1]. The waste of productivity caused by the use of abrasive diamond tool seems reasonable in view of difficulties and time required for a manual polishing operation. Moreover, radial depth of cut could be increased in order to improve productivity. Indeed, experimental values of roughness are higher than theoretical values. However, complementary tests are necessary to state that Grinding forces During the previous tests, the workpiece is held on a dynamometer to measure forces produced by the abrasive diamond tool. Forces measurements with regard to feed rate are given in Figure 9. Grinding forces measured are lower than values commonly reached in milling but it is coherent with values provide by literature [HZ1]. These low values can avoid problems of mold deformation and tool flexion. Thus, form defects caused by finishing operation are limited. before the beginning and after the machining of eight areas. Figure 10 shows differences between the both profiles. Tool wear mm (a) (b) Figure 10 : Tool profile before (a) and after (b), machining the 8 areas A tool wear near to 10 µm appears after machining eight areas (84 cm²). Tool wear has an influence on the tool profile but has a low impact on the produced roughness for this couple tool-material. Indeed, as Figure 8 shows it, the average arithmetic roughness remains constant. In fact, for abrasive tool, wear has not necessary a negative impact on the roughness produced [ZX1]. However, wear remains a problem. Indeed, the tool profile changes during machining which can lead to form defects on produced mold surface. Further tests should be carried out in order to predict tool wear and so form defects on the machined surface. Nevertheless, the tool profile could be corrected by a dressing operation with a whetstone. Furthermore, development on optimal grit size, bond and machining conditions could improve the wear behaviour. Fy Fx Fz Figure 9 : Grinding forces according to feed rate Vertical grinding force (Fz) is higher than other. This phenomenon could be observed in other study about grinding [HZ1]. Furthermore, when feed rate increases, all the components of grinding forces increase but the Fz components is growing faster. Between the area two and eight, this component has almost tripled (from 10 N to 5 N). Thus, feed rate higher than 100 mm/min are not carried out here in order to not damage the tool. Indeed the strength of the tool and tool wear behaviour are unknown Tool wear Previous sessions give informations about the abrasive diamond tool s performance for a finishing operation but the tool wear behaviour remains unknown. Tests carried out here are not sufficient to predict tool wear but some observations could be realised. During these tests, tool profiles are taken Fz Fx Fy 4- Conclusion Tests and analysis are carried out in this study. They ensure to state that an abrasive diamond tool can realised a finishing operation on a mold in Hextool with a CNC machine. Indeed, a grinding operation conducted on CNC machine allows to reach an average arithmetic roughness about 0.5 µm. Furthermore, the diamond abrasive tool can go through a variation on axial depth of cut from 0.07 mm to 0.87 mm. This variation has no impact on arithmetic roughness produced for this couple tool-material. Moreover, the productivity possibility is partially validated. Indeed, the maximal feed rate tested here (100 mm/min) is lower than value commonly used for PCD tools. However, the surface quality reached with this abrasive tool allows to reduce and even to cancel the manual polishing operation. The waste of productivity caused by the use of abrasive diamond tool seems reasonable comparatively to difficulties and time required for a manual polishing operation. Furthermore, this finishing process leads to low forces. These low values can avoid problems of mold deformation and tool flexion. Thus, form defects caused by the grinding operation are limited to the tool wear. Finally, this study shows that tool wear is not a major problem for surface roughness. In fact, tool wear has an influence on the tool profile but has a low impact on the produced roughness for this couple tool-material. Thus, the tool profile changes and can leads form defects on produced mold. However, the tool profile could be corrected by a dressing operation with a whetstone. The results of this study are full of promise. A finishing IDMME-P4-5- Copyright IDMME - Virtual Concept

6 IDMME - Virtual Concept 010 process of mold in composite material with abrasive diamond tool seems possible. Moreover, this process could significantly reduce or even cancel the manual polishing operation and increase the productivity of manufacturing for composite material mold. 5- Acknowledgement This work was carried out within the LCM-Smart Project (Hexcel, SKF Aerospace France, Issoire Aviation, ESI Group, Isojet, Visuol Technologies) supported by the French Ministry of Industry. This paper was written within the framework of the TIMS Research Group, using grants from the Regional Council of Auvergne, the French Ministry of Research, the CNRS and the Cemagref. This work was carried out within the Manufacturing 1 working group, which comprises 17 French research laboratories. The topics approached are: Modelling of the manufacturing process; Virtual machining; Emergence of new manufacturing methods. The authors wish to thank Asahi Diamond Industrial Europe for this collaboration. 6- References [CC1] Chardon G., Chanal H., Gazel A., Duc E., Prospective study on the surface machining of fiber reinforced composites, Proceedings of 1th CIRP Conference on Modelling of Machining Operations, , 009. [KW1] Konig W., Wulf Ch., Grab P., Machining of fibre reinforced plastics, Ann CIRP, 34 : ,1985. [BR1] Bhatnagar N., Ramakrishnan N., Naik N.K., Komanduri R., On the machining of fibre reinforced plastic (FRP) composite laminate, International Journal of Machine Tools and Manufacture, 35 : , [RR1] Rahman M., Ramakrishna S., Prakash J.R.S., Tan D.C.G., Machining study of carbon fiber reinforced composite, Journal of Materials Processing Technology, 89 : 9-97,1999. [WZ1] Wang X.M., Zhang L.C., An experimental investigation into the orthogonal cutting of unidirectional fibre reinforced plastics, International Journal of Machine Tools and Manufacture, 43 : , 003. [WK1] Wu X., Kita Y., Ikoku K., New polishing technology of free form surface by gc, Journal of Materials Processing Technology, : 81-84, 007. [RX1] Roswell A., Xi F., Liu G., Modelling and analysis of contact stress for automated polishing, International Journal of Machining and Machinability of Materials, 46 (3-4) : , 006. [HI1] Huissoon J., Ismail F., Jafari A., Bedi S., Automated polishing of die steel surfaces, International Journal of Advanced Manufacturing Technology, 37 : , 00. [PT1] Pessoles X., Tournier C., Automatic polishing process of plastic injection molds on a 5-axis milling center, Journal of Materials Processing Technology, 09 : , 009. [TH1] Tsai M.J., Huang J.F., Efficient automatic polishing process with a new compliant abrasive tool, International Journal of Advanced Manufacturing Technology, 30 : , 006. [HZ1] Hu N.S., Zhang L.C., Some observations in grinding unidirectional carbon fibre-reinforced plastics, Journal of Materials Processing Technology, 15 : , 004. [ZX1] Zhou X., Xi F., Modeling and predicting surface roughness of the grinding process, International Journal of Machine Tools and Manufacture, 4 : , 00. [PL1] Park K.Y., Lee D.G., Nakagawa T., Mirror surface grinding characteristics and mechanism of carbon fibrereinforced plastics, Journal of Materials Processing Technology, 5 : , [G1] Grzesik W., A revised model for predicting surface roughness in turning, Wear, 194 : , [CM1] Childs T.H.C., Moss D.J., Wear and cost issues in magnetic fluid grinding, Wear, 49 : , 001. [LK1] Lin R.S., Koren Y., Efficient tool-path planning for machining free-form surfaces, Journal of Engineering for Industry (ASME), 118 : 0-8, [TD1] Tournier C., Duc E., Iso-scallop tool path generation in 5-axis milling, International Journal of Advanced Technology, 5 : , 005. [T1] Tournier C., Contribution à la conception des formes complexes : la surface d usinage en 5 axes, Thèse de doctorat de l ENS Cachan, 001. [T] Toh C.K., Design, evaluation and optimisation of cutter path strategies when high speed machining hardened mould and die materials, Materials and Design, 6 : , 005. IDMME-P4-6- Copyright IDMME - Virtual Concept