QG
1
,QWHUQDWLRQDOVHPLQDURQLPSURYLQJPDFKLQHWRROSHUIRUPDQFH
0RUSKRORJLFDODQDO\VLVRIVXUIDFHVJHQHUDWHG LQ+LJK6SHHG0DFKLQLQJ
/DERUDWRLUHGH7ULERORJLHHWGH'\QDPLTXHGHV6\VWqPHV (FROH1DWLRQDOHG¶,QJpQLHXUVGH6DLQW(WLHQQH 5XH-HDQ3DURW6DLQW(WLHQQH&HGH[
'U0LFKHO'XUVDSW0LFKHO&KDPEH 'U+DVVDQ=DKRXDQL'U7KRPDV0DWKLD
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
• • •
RUWRJRQDOGLUHFWLRQLQDGYDQFHZLWKRXWFRYHULQJRIWKHSDG RUWRJRQDOGLUHFWLRQLQDGYDQFHZLWKFRYHULQJRIWKHSDG SDUDOOHOLQDGYDQFHH[FHSWFRYHULQJRIWKHSDG
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
• • •
GLUHFWLRQSDUDOOHOLQDGYDQFHKDV RUWKRJRQDO\HDUGLUHFWLRQLQDGYDQFHZLWKRXWFRYHULQJ RUWKRJRQDO\HDUGLUHFWLRQLQDGYDQFHZLWKFRYHULQJ
7KH UHVXOWV RIWKLV VWXG\LV FRPSOHWHG E\ \HDU DQDO\VLV RI KLJK VSHHG PLOOLQJ SURFHVV LQ FRQWH[W RI WULERORJ\ 6 NQRZOHGJH DQG DOVR VRPH WHFKQRORJLFDO FRQFOXVLRQV LV SRLQWHG RXW IRU GHVLJQHUV RI PDQXIDFWXUHRIPDFKLQHWRROVDQGFXWWLQJWRROVWRR
QG
2
,QWHUQDWLRQDOVHPLQDURQLPSURYLQJPDFKLQHWRROSHUIRUPDQFH
-
3URFHGXUH
It is proposed to analyze the morphology of two families of surfaces obtained in high speed milling and considered as being very representative of this type of work. The first type of surface was obtained by face milling and the second by profiled cutting. The study first of all consisted in carrying out the measurement of surfaces at a particular place, i.e. in the zone covered by the various overlapping passages of tool, in order to highlight the general shape of this part of the machined surface. We then filtered the measurements in order to separate the second-order defects (undulations) from the third and fourth order defects (roughnesses). The goal of this study was to highlight and to analyze the particular surface effects obtained by high speed milling using tools of small diameter, by successive plunges (profile cutting) or by sweeping (face milling), in order to check if they could qualitatively replace the surfaces generated in traditional milling by large-sized tools.
-
)DFHPLOOLQJZLWKVZHHSLQJ
Obtaining a surface by sweeping causes this aspect of small circles (actually epicycloids) apparently tangent ,shading the surface, figure 1 represents this effect. sweeping range
feed movement
diameter of drill
overlap zone
Figure 1 The surface studied comes from a part used an a normative test, made of aluminium
QG
3
,QWHUQDWLRQDOVHPLQDURQLPSURYLQJPDFKLQHWRROSHUIRUPDQFH
alloy (AU4G) with a cutter with a diameter 20 mm, with two teeth, a rotational frequency of 14200 turns per minute, the feed rate being 0.08 mm per tooth and the depth of cut 0.8 mm, these conditions concerning the concept of surface finishing.
2.1 Zone of analysis: The zone studied is part of the surface machined 4 millimeters by 4 millimetres, entered on a three-dimensional measuring apparatus of type SURFASCAN. It will be known by the intermediary of a file of points whose altitude Z is related to X and y; Z = f(x,y). The size of measurement (8m µon X and y) conditions the number of points (here 250000 points) which will be the image of analyzed surface. The measured surface will be rectified by a plan of least squares in order to be freed from the possible variations of position between the reference frame of measurement and the general direction of surface, see figure 2.
- Figure 2 On this image, by the play of the colors assigned to the altitude of the various points of surface, the overlap zone is very clear (in yellow); this zone is materialized on a microscopic scale by a relief that we will call “a ridge” .Several hypotheses can be put forward as to the cause of the formation of this ridge [ KAT 86 ] ; for our part we will speak of a plastic deformation of the matter supporting the machined surface. This plastic deformation is caused by the axial component of the action of the tool on the part,
QG
4
,QWHUQDWLRQDOVHPLQDURQLPSURYLQJPDFKLQHWRROSHUIRUPDQFH
this component causing a contact pressure certainly much higher than the stress limits of the material (260 N/mm2 for AU4G) being given the weak contact surface between the tool and the machined surface resulting from the thinness of the cutting edge of the cutter. Micro-hardness tests actually revealed a definitely higher hardness of the matter constituting the ridge compared to the remainder of material, which seems to indicate that this matter has very probably been hardened and thus well deformed plastically.
2.2 Filtering " low frequencies ": It is known that the output signal h(x,y) of a measuring equipment is equal to the product of convolution of the input signal f(x,y) by the effect of the filter used g((x,y)[ STO 95 ].
( , ) = I ( [, \ ) ∗ J ( [, \ )
K [ \
In our case the treatment is carried out by a software developed within UMR 55 13 [ ZAH 89 ]. The filter is a Gaussian filter of which the length of the cut in x and y is 0.8 millimetres, which corresponds to the ISO recommendation for measurements of conventional surface quality, this filtering thus represents the surface undulation in x and y, see figure 3.
- Figure 3 -
QG
5
,QWHUQDWLRQDOVHPLQDURQLPSURYLQJPDFKLQHWRROSHUIRUPDQFH
In passing it is important to note that in the case of high speed machining, where the cutter is small, therefore with a low number of cutting edges, the defect of undulation, if it exists, is due only to the vibratory phenomena generated by the cut. Indeed the period of passage "of the lowest tooth" (one on two) is small and will cause irregularities of a step of 0.16 mm approximately which are part of roughness. In this connection let us point out a frequent origin of the appearance of a second order defect generated in traditional milling with a large diameter cutter and a high feed rate ; we will take the example represented in figure 4 of a cutter with 9 teeth and a feed rate of 0.3 mm per tooth, which would cause a theoretical step of irregularity of 2.7 mm which effectively corresponds to what is conventionally defined as being undulation. feed per tooth
lowest tooth undulation defect Figure 4
2.3 Filtering " high frequencies ": One of the advantages of the Gaussian filter advantages of is that one obtains the signal filtered in high frequency by substraction of the low frequencies of the initial signal ; high frequency filtering naturally corresponds to the roughness of surface. Figure 5 represents the three-dimensional roughness of the analyzed surface.
( , ) = I ( [, \ ) − K( [, \ )
N [ \
QG
6
,QWHUQDWLRQDOVHPLQDURQLPSURYLQJPDFKLQHWRROSHUIRUPDQFH
Figure 5 In the following paragraphs we will leave the three-dimensional representation and analyze profiles quantitatively. We will take care " to place the feeler lines" following of the positions and the privileged directions.
2.4 Orthogonal roughness in feed without overlap : The criteria selected are the criteria recommended by the ISO [ ISO 97 ]. As figure 6 indicates it is a question of defining a zone outside the "ridge". One will note the value of two criteria frequently used: - total roughness: Rt = 0.94 microns. - arithmetic roughness: Ra = 0.15 microns.
QG
7
,QWHUQDWLRQDOVHPLQDURQLPSURYLQJPDFKLQHWRROSHUIRUPDQFH
GHWDLORISURILOH
OHQJWK PPP
feeling line
Figure 6
2.5 Orthogonal roughness in feed with overlap : Let us establish the same evaluation for a length of feeling "crossing the ridge " It will be then interesting to compare the variations of the two preceding criteria. OHQJWK PPP
feeling line
QG
8
,QWHUQDWLRQDOVHPLQDURQLPSURYLQJPDFKLQHWRROSHUIRUPDQFH
« riage »
Figure 7 The first remark which is essential is that the height of the " ridge" proves to be lower than 0.5 microns; in addition let us notice the variation of the two criteria examined previously: • Total roughness : this goes from 0.94 to 1.38 microns, the difference between these two value corresponds almost exactly to the height of the ridge.
• Arithmetic roughness: goes from 0.15 to 018 microns, which corresponds to an increase of 20%. -
5RXJKQHVVSDUDOOHOWRIHHGRXWVLGHRYHUODS
It is important to make sure that variation in criteria is slight, for the concept of sweeping should not, in theory, generate privileged directions, ensuring qualities of precision and homogeneous and isotropic surface quality. GHWDLORISURILOH
OHQJWK PPP
feeling line
QG
9
,QWHUQDWLRQDOVHPLQDURQLPSURYLQJPDFKLQHWRROSHUIRUPDQFH
Figure 8 In conclusion of this analysis on the face work with sweeping, it appears that in spite of the visual aspect, with the naked eye, the characterization of surface quality is fairly homogeneous, whatever the direction or the zone of covering. In addition it should be stressed that the thickness of the ridge created by sweeping is, in the topographic sense of the term, very slight.
-
0LOOLQJRISURILOHE\VXFFHVVLYHSOXQJHV
It is generally about machining of veins or pockets. In this latter case this type of work is called "pocketing". In the example which will follow we will analyze a sample of an industrial aluminium alloy part machined in blank pocketing with a cast solid carbide cutter with two cutting lips with a diameter of twelve millimetres. The spindle has a rotational frequency of 15000 turns per minute. Let us point out the principle of succession of the operations, figure 9.
QG
10
,QWHUQDWLRQDOVHPLQDURQLPSURYLQJPDFKLQHWRROSHUIRUPDQFH
successive plunges surfacing of profile studied
Figure 9
3.1 Zone of analysis: It is a question of properly identifying the movements of generation on the analyzed surface. One notes on the image of figure 10, representing the studied surface, rectified and not filtered, resulting from three-dimensional measurement on SURFASCAN, the direction of feed and the direction of the successive plunges. Obiously only one overlap “ passage" is analyzed.
direction of plunges
Sens direction des avances of feed
Figure 10
QG
11
,QWHUQDWLRQDOVHPLQDURQLPSURYLQJPDFKLQHWRROSHUIRUPDQFH
3.2 Filtering " low frequencies ": As for the preceding study we will use a filter of cut 0.8 mm in length in X and in y, recalling that undulations, defects of second order, are underlined by the low frequency filtering, figure 11 represents these undulations.
% $
Figure 11 It should be remembered that the undulations observed in the direction A, the direction of the plunge, are due to the geometrical quality of the profile of the tool and to its elastic behaviour faced with cutting stress; and that in the direction B, the direction of feed, these second order defects are due primarily to the feed rate and the vibratory modes which result from this.
3.3 "High frequency" filtering: In direction A the quantification of the ridge created with each plunge is visible in figure 12 ; in addition the periodicity of the advance is it also clearly visible in direction B.
QG
12
,QWHUQDWLRQDOVHPLQDURQLPSURYLQJPDFKLQHWRROSHUIRUPDQFH
% $
Figure 12 Let us now leave the three-dimensional study to work on measures of profile which we will carry out according to feeling lines that we will place parallel to directions A and B.
3.4 Roughness parallel in advance: The line allowing the evaluation of the profile is located out of the zone of overlap of the plunges. GHWDLORISURILOH
OHQJWK PPP
feeling line
QG
13
,QWHUQDWLRQDOVHPLQDURQLPSURYLQJPDFKLQHWRROSHUIRUPDQFH
Figure 12 We note a Ra of 0.85 microns, but especially a total roughness of 3.6 microns. We will confront these values with those obtained following measurements in different places and directions.
3.5 Orthogonal roughness in feed without overlap : GHWDLORISURILOH
OHQJWK PPP
feeling line
Figure 13
QG
14
,QWHUQDWLRQDOVHPLQDURQLPSURYLQJPDFKLQHWRROSHUIRUPDQFH
According to the proposal of the preceding paragraph one sees that total roughness is very stable, Ra being certainly a little more favorable to slightly variable. In terms of industrial application this is important, in particular for distorted shapes, as the obstruction stresses will allow the most convenient working of parts according to the number of digitized axes.
3.6 - Orthogonal roughness in feed with overlap : GHWDLORISURILOH
OHQJWK PPP
feeling line
Recouvrement d’une prise de passe
Figure 14 The very significant variation in total roughness underlines the presence of the ridge ; its height can be estimated at four microns, value nearly ten times higher than that
QG
15
,QWHUQDWLRQDOVHPLQDURQLPSURYLQJPDFKLQHWRROSHUIRUPDQFH
obtained in surfacing of finish but which can be quite acceptable for blank work. The assumption made previously as to the formation of this ridge, namely a plastic deformation of the matter due to the component of the normal cutting pressure on the surface, here appears even more probable. The greatest height of the ridge being explained by the fact that the cutting pressures being more important (blank cutting) the pressure on surface will be stronger and the volume of matter deformed will be larger. It will be noted that the thickness of the ridge being quite clearly lower than the value of the minimum chip usually noted in milling, it is normal that this does not disappear at the time of the successive plunge-cuts.
±&RQFOXVLRQV
The two brief studies that we have just evoked reveal the morphological characteristics of the surfaces obtained by two types of high speed milling frequently employed, a surfacing of finish by sweeping and a profiled blank by successive plunge-cuts. A characteristic particular to these two types of machining is the presence of a ridge of which we explained the appearance by a plastic deformation of the matter close to the zone of covering, this deformation being all the greater since the efforts (thus sections of chips) are large, in the worse of the two cases which we studied (blank by plunge-cuts) the height of this ridge, 5 µµ is absolutely not redhibitory for the majority of applications. The other interesting characteristic underlined is obtaining surfaces whose roughness is isotropic, i.e. the value of the various parameters of characterization of the roughness of profiles measured on these surfaces is about constant whatever the direction in which measurements are taken. Let us recall that anisotropy is a main characteristic of the surfaces obtained by machining [ ROU 92 ] and in particular of those obtained in traditional milling.
±%LEOLRJUDSK\
[ KAT 86 ] KATO K, HOKKIRIGAWA K, KAYABA T, ENDO Y - " Three dimensional shape effect on abrasive wear " - Trans. Of the ASME flight 108. 1986. [ ZAH 89 ] ZAHOUANI H –“ Quantification of the three-dimensional topography of surfaces " - PHD the University of Besancon. 1989. [ STO 95 ] STOOT KJ, SULLIVAN PJ, MAINSAH E, LUO N, ZAHOUANI H, MATHIA T –" The development of methods for characterisation of roughness in three dimensions " - Published one the behalf of the comission of the European communities.
QG
16
,QWHUQDWLRQDOVHPLQDURQLPSURYLQJPDFKLQHWRROSHUIRUPDQFH
1995. [ ISO 97 ] ISO 4287 –" Geometrical Specification of the Surface quality –products: Method of the profile – Terms, definition and parameters of surface quality " - AFNOR. 1997. [ ROU 92 ] ROUSSEAU J, Physics MATHIA –T " Solid state and morphology of surfaces " - LTDS URA CNRS 855 1992.
