英語原文.pdf

Journal of Materials Processing Technology 84 1998 47 55 A laser beam machining LBM database for the cutting of ceramic tile I Black S A J Livingstone K L Chua Department of Mechanical and Chemical Engineering Heriot Watt Uni6ersity Riccarton Edinburgh EH14 4AS UK Received 13 December 1997 Abstract This paper covers the cutting of commercially available ceramic tiles using a CO2laser cutting machine with the object of producing a laser beam machining LBM database that contains the essential parameter information for their successful processing Various laser cutting parameters were investigated that would generate a cut in ceramic tile which required minimal post treatment The effects of various shield gases of multi pass cutting and of underwater cutting were also examined 1998 Elsevier Science S A All rights reserved Keywords CO2 Laser cutting Ceramic materials Advanced manufacturing processes 1 Introduction and background Manual methods of cutting ceramic tiles are very similar to that for glass i e scribing the materials with tungsten carbide tipped cutter followed by the applica tion of a bending moment along the scribed line to initiate controlled fracture However manual tech niques are limited to straight line cutting and relatively large radius cuts Internal and undercut profi les are nearly impossible to produce with scoring alone with the possible exception of internal circles more sophis ticated methods having to be applied to achieve these profi les Traditionally diamond saw hydrodynamic water jet or ultrasonic machining are used to create complex geometries in ceramic tiles but these processes are very time consuming and expensive For example typical diamond saw cutting speeds are in the order of 20 mm min 1 1 while ultrasonic drilling of Al2O3 takes over 30 s per hole 2 The most critical factor arising from use of a CO2 laser to cut ceramic tiles is crack damage which is essentially caused by a high temperature gradient within the ceramic substrate during the cutting process These cracks reduce the strength and are sources for critical crack growth which may result in partial or complete failure of the tile substrate 3 Thus a reduc tion of process induced crack formation is paramount for the realistic commercial use of lasers to cut ceramic tiles 2 Laser cutting parameters Laser machining of any material is a complex process involving many different parameters that which all need to work in consort to produce a quality machining operation 4 parameters such as i laser power input ii focal setting iii assist gas type and pressure iv nozzle confi guration v workpiece thickness and vi optophysical properties Previous research within the authors department 1 5 6 has also demonstrated the criticality of the above parameters in effi cient laser cutting 2 1 Laser power Laser power depends on the type of laser used For the work reported in this paper a Ferranti MF400 CNC laser cutter was employed rated at a power output of 400 W However due to upgrading the maximum beam power achievable was between 520 and Correspondingauthor Fax 441314513129 e mail i black hw ac uk 0924 0136 98 see front matter 1998 Elsevier Science S A All rights reserved PIIS0924 0136 98 00078 8 I Black et al Journal of Materials Processing Technology84 1998 47 5548 530 W in continuous wave CW cutting mode The laser also had the ability to work in pulse mode PM and super pulse mode SPM Fig 1 To determine the equivalent power output during pulsing operation a power verses pulsing chart was used in conjunction with the following basic equation 9 Pr Pl Ps f 1 Pl Pr Although the laser cutter could operate between fre quencies of 50 and 5000 Hz a value of 500 Hz was recommended in previous work 1 5 Since this setting proved to be successful only limited investigation into other frequencies was carried out at 250 Hz 750 and 100 Hz 2 2 Cutting speed The CNC table used with the Ferranti MF400 laser cutter had a maximum feed rate of 10000 mm min 1 Previous work 6 indicated that feed rates above 6000 mm min 1proved to be unstable for any standardised testing The optimum cutting speed varied with the power setting and more importantly with the thickness of the workpiece 2 3 Shield gas type and pressure Compressed air argon nitrogen and oxygen were used as shield gases during cutting with pmax 4 bar Different shield gases were used to examined their effect on cut quality after processing since the shield gas not only cools and cut edges and removes molten material but also generates a chemical reaction with the sub strate material 7 The results of this chemical reaction differ for each type of shield gas used For test purposes p was varied in steps of 0 5 bar from 1 to 2 5 bar then in steps of 0 2 bar from 2 6 bar to the maximum attainable gas pressure 2 4 Nozzle confi guration The nozzle diameter contributes directly to the maxi mum achievable gas pressure and hence to the mass fl ow rate of the gas was important for the economics of cutting especially when using cylinders of argon and nitrogen Only circular profi les for the nozzle exits were available 0 6 mm5Ns520 mm but this uniform nozzle exit geometry allowed cutting in any direction 2 5 Nozzle height and focal positioning The height at which the nozzle was set was governed by the position of the focal point The Ferranti MF400 laser cutter only possessed a long focal length of 110 Fig 1 Cutting modes mm originally a short focal length of 46 mm was available before upgrading and this length could be altered by 95 mm If the nozzle height was incorrectly set the beam would clip the nozzle and reduce the equivalent power output to the workpiece 6 For the bulk of the testing the focal height was set so the focal point was on the job i e on the top surface of the workpiece This condition obviously governed the posi tion of the nozzle above the workpiece 3 Experimental procedure Six types of Si Al2O3 based ceramic tiles were exam ined Table 1 originating from different countries Note that the composition of the tiles varied as did the thickness but all possessed a surface glaze and in the case of the 7 5 8 6 and 9 2 mm Spanish tiles the glaze was double layered 3 1 Set up procedure Since there was a need for standard testing condi tions the following procedure was implemented before the start of testing i the beam power was validated to specifi cation i e 520 530 W developed at full power CW although this dropped to around 50 W after Table 1 Types of ceramic tile used ts mm Tile typeBody colour 3 7BrazilianWhite 4 7WhitePeruvian Light redItalian5 2 SpanishRed5 74 Spanish7 5Red RedSpanish8 6 9 2RedSpanish I Black et al Journal of Materials Processing Technology84 1998 47 5549 about 1 h of testing ii the nozzle and the focal lens were checked to ensure that they were in good condi tion i e clean and undamaged iii the shield gas pressure regulator and shield gas tanks were turned on to prevent damage to the focal lens iv the laser beam was centred within the nozzle using a square test a lower energy input in PM being used to cut a square on a mild steel the sparking density that resulted from cutting being checked to see if it was equally distributed about the cut line and v the focal point was set for its desired positioning i e on the job 3 2 Testing A straight line test SLT was used to evaluate the variablelaserparametersforfullthrough cutting FTC Angular cutting was confi gured to investigate how the material reacted during cutting of tight geome try Circular testing and square testing were devised to determine the effects resulting from cutting various geometries The SLT allowed for the combined testing of two separate parameters on one testpiece upon completion the results being present automatically in a cutting matrix in the form of the resulting cuts P and V are the most important laser parameters as they dictate the amount of energy input per unit length of cut therefore they were paired for the SLT as were p and NSwhich govern the mass fl ow rate of the shield gas For the P V test runs the power was held constant while the cutting speed was increased along the cut Fig 2 a The length of cut at constant cutting speed had to be of suffi cient magnitude to accommodate the acceleration or deceleration of the CNC table between feed changes previous work 6 indicated that 50 mm was adequate Interpreting the results was made easier due to their tabular format with the cutting matrix showing clearly any trends or patterns occurring due to the changes in parameter settings The SLT also al lowed a large number of cuts to be carried out over a short time frame This proved advantageous as the laser tended to drift from its initial settings with time Precautions had to taken to avoid localised heating in the tile from continuous close proximity cutting as a change in tile body temperature would invalidate any resulting data Initially a 20 mm separation between cuts was used and this proved suffi cient In order to study how close the cuts could be made to each other the separation between cuts was reduced by increments of 2 mm from an initial 20 mm spacing During the SLT the other laser parameters had to be held constant 6 For P versus V f was held at 500 Hz with NS 1 2 mm and p 3 bar The beam focal point remained on the job The results from the P V cutting matrix determined the fi xed values for the cutting speed and pulse settings for the succeeding SLT For the NS p Fig 2 Testing confi guration a straight line testing b angular testing c circular testing d square testing cutting matrix the nozzle size remained constant along the x axis refer to Fig 2 a while p was increased in steps of 0 2 bar from 2 bar in the y axis the cut separation remained constant at 20 mm A new matrix was created subsequently for each nozzle size Angular testing Fig 2 b was used to investigate how the cut material reacted to sustained exposure from the laser beam during the machining of tight geometries i e where several cuts are made in close proximity to each other The proximity test mentioned for SLT determines how close parallel lines can be cut to each other whereas angular testing is used to deter mine how the cutting of acute angles effects the cut quality The angles cut from a workpiece were reduced from 45 to 10 and the corresponding surface fi nish quality SFQ was noted I Black et al Journal of Materials Processing Technology84 1998 47 5550 Table 2 Multi pass cutting parameters PlCutting modePsNo of passesLast cut CW60 FTC 9000100SPMFTC100 Table 3 Grading of SFQ Grading 1No cracking in surface glaze solid sharp cut edge Minimal glaze cracking WcB2 mm with slight2 loss of sharpness in cut edge Medium cracking 2 mmBWcB4 mm and slight3 damage to unglazed tile substrate Signifi cant damage to glaze coating Wc 6 mm 4 heavy damage to unglazed substrate causing fl aking in the glazed surface 5Same as 4 but with the formation of cracks in the tile s main body leading to structural failure in a part of the tile usually at the end of a cut or within 8 mm of the tile edge There are two reasons for conducting square and circular testing Fig 2 c and d fi rst to determine the optimum method of laser beam introduction to internal cut profi les and secondly to determine if there was any limitation in the dimension of the size of square or hole cut If not correctly introduced the laser beam would cause an internally cut profi le to fail at the point of introduction due to the brief but excessive thermal gradient induced from cutting i e thermal shock Therefore utilising methods of beam introduc tion such as trepanning onto a profi le enabled com plex geometries to be investigated What also became apparent during testing was the importance of the position of beam extraction from the cut profi le and the position of the beam starting point relative to the geometry i e whether it was at a corner or on a straight edge 3 3 Multi pass and underwater cutting Multi pass cutting was begun with a low power P 100 W laser beam The fi rst pass produced a well defi ned blind kerf in the substrate followed by a second pass to cut deeper and so on The process was repeated until the kerf was about 20 mm deep and then the laser power was switched to 500 W and do the fi nal FTC The objective of multi pass cutting was to reduce ther mal overload by use of less input energy per unit length The parameters used in this test are given in Table 2 Underwater cutting was conducted with the objective of reducing the infl uence of heat around the cut area and also to examined the effect on cut quality through accelerated heat dissipation using water 8 The ce ramic tile was placed under water and the nozzle was also dipped in water the shield gas pressure preventing any water from entering the nozzle jet chamber 4 Cut quality Material properties laser parameters and workpiece geometry have a signifi cant effect on the fi nal result of the laser cutting process Cut quality is essentially char acterisedbysurfaceroughnessanddrossheight whereas crack length dictates the strength reduction in the substrate Fig 3 The overall SFQ at the glaze surface was classifi ed according to the grading scale given in Table 3 Therefore the quality of the cut surface and edge were measured with respected to i surface roughness ii surface fi nish and iii dross adherence Fig 3 Quality criteria for the laser cutting of ceramic tiles I Black et al Journal of Materials Processing Technology84 1998 47 5551 Fig 4 Measurement of Rafor the cut surface 4 1 Surface roughness It was important to measure surface roughness as this allowed the cut quality to be gauged alongside values obtained from previous work 1 and values recorded for other manufacturing processes Due to the large number of cuts being made it was necessary to reduce the number of cuts to be analysed Therefore cuts with SFQB2 were not measured The surface roughness of the cut edge was character ised by the formation of striation lines left by the cutting process Ravalues were measured from the centre line of the cut edge Fig 4 Measurements were taken over a 12 5 mm traverse of the stylus with a cut off value of 2 5 mm i e fi ve readings were taken over the traverse which ensured that the stylus trav elled over a reasonable number of striation lines 4 2 Dross adherence Dross adherence directly effected the Ravalue of the cut and the ability to remove internally cut geometries A micrometer was used to measure the dross height at three intervals along the cut section The dross height remained fairly constant approximately 1 mm with all types of cutting Since this value was deemed to be of no practical importance it was not recorded in the database 5 Results Table 4 contains the current LBM database for cut ting ceramic tiles that was compiled from the results of the work reported in this paper The fi rst part of the table contains the parameters and results for substrates cut in atmosphere while the results for underwater cutting are shown in the second part 5 1 Parameter effects 5 1 1 Cutting speed For the thinner tiles tsB7 mm the P V cutting matrix showed a wide region of FTC with SFQ 1 In the case of the Brazilian tile ts 3 7 mm FTC was obtained with cutting speeds of up to 2200 mm min 1 and down to Pr 0 5 with reduced speeds at f 500 Hz This region diminished with the increase in tile thickness and also with the redness of the body colour generally the thicker tiles are darker in body colour Fig 5 shows how the maximum cutting speed for FTC varies with ts The exponential relationship obtained concurs with previous work 6 for different materials such as steel wood and perspex The cutting matrix also showed that once the cutting speed exceeded values for attainable FTC scribing or blind cutting results 5 1 2 Pulsing Pulsing the laser for all but the thick Spanish tiles was not required as the CW setting produced cuts with a good SFQ grading Successful FTC was obtained at Pr 0 4 in the Brazilian tile but Vmaxwas so low that in a practical sense the settings were not viable On the thick Spanish tiles pulsing of the beam was required as CW caused cracking in the glaze This was probably due to an excess of energy input per unit length of cut causing thermal shock as the thermal expansion rate of the glaze differed suffi ciently from that of the parent tile Since pulsing the laser reduced the energy input by approximately 25 W for every 0 1 drop in Pr at f 500 Hz the surface glaze cracking virtually disappeared at Pr 0 6 and optimum cutting speed although tiny cracks of the order of 0 5 mm wide at the cut edge still remained 5 1 3 Gas pressure This parameter has a great effect on the quality and the rate at which cuts could be made successfully Previous work 2 had shown that high gas pressures were required to achieve FTC on thick substrates ts 7 mm This was borne out by the results obtained from thep Nscuttingmatrix Highqualitycutswere achieved in the thinner tiles tsB6 mm at gas pressures of 2 bar but in the double glazed thicker tiles values of SFQB3 were not achieved unless p 3 bar At low pressures pB2 5 bar the maximum cutting speeds for FTC dropped drastically as the gas failed in its role of dross clearer Vmaxfor Brazilian tile in CW dropped from 2200 mm min 1at p 3 8 bar to 1500 mm min 1 at p 3 bar An increase in surface glaze cracking also became apparent at low gas pressures This led to the conclusion that the shield gas was acting as a coolant and thus helping to minimise the large thermal gradient created by the beam I Black et al Journal of Materials Processing Technology84 1998 47 5552 Table 4 4 LBM database for ceramic tiles Atmospheric cutting Ra mm V mm min Tile typets mm BodyShield gasSFQGlaze typeNs mm Geometricp bar Pl Ps cutcolour 1 2 1 51Brazilian3 7WhiteWhiteStraightCW500 1000C air 325 35 25 3511 2 1 5 3180 20500 1000C air 31 2 1 5125 35160 40500 900C air 25 3511 2 1 5 3Internal180 20400 600C air 1 2 1 51 5Angular160 40300 500C air 325 35 25 3511 2 1 5 3Radi