- Email: [email protected]
A new understanding of the wear processes during laser assisted milling 17-4 precipitation hardened stainless steel
Wear 328-329 (2015) 518–530
Contents lists available at ScienceDirect
Wear journal homepage: www.elsevier.com/locate/wear
A new understanding of the wear processes during laser assisted milling 17-4 precipitation hardened stainless steel M.J. Bermingham a,b,n, D. Kent b, M.S. Dargusch a,b a b
Defence Materials Technology Centre, School of Mechanical and Mining Engineering, The University of Queensland, Australia Queensland Centre for Advanced Materials Processing and Manufacturing (AMPAM), The University of Queensland, Australia
art ic l e i nf o
a b s t r a c t
Article history: Received 12 November 2014 Received in revised form 26 March 2015 Accepted 28 March 2015 Available online 3 April 2015
Laser assisted machining is known to improve the machinability of several difficult to cut materials. For the first time, this study investigates the tool wear rates and the wear mechanisms associated with milling a precipitation hardened martensitic stainless steel with and without the assistance of laser preheating. Across both traditional low feed milling and emerging high feed milling techniques, laser assistance was found to reduce the tool wear rates by up to 50% and lower the cutting force by up to 33% in comparison to conventional room temperature machining. In all cases it is observed that tool coating breakdown by abrasive and adhesive wear processes is the dominant tool failure mechanism. Laser assisted milling is effective in prolonging tool life by delaying these processes in comparison to conventional machining. & 2015 Elsevier B.V. All rights reserved.
Keywords: Steel Laser processing Cutting tools Wear testing
1. Introduction Thermally assisted machining (TAM) has been employed for over a century to improve the machinability of difficult to cut materials. Believed to be first pioneered by B.C Tilghman in 1889 [1], the principle involves using heat to soften materials thereby making them easier to remove with a cutting tool. The process was slow to gain interest but around the 1950s research efforts accelerated to investigate the merits of the process [2–5]. Since this time a great deal of research has been performed and in the last few decades alone, it has been well established that thermally assisted machining is a viable method for improving the machinability of numerous materials. In ferrous alloys, the process is beneficial in terms of reducing cutting forces and improving tool life compared to conventional machining [6–16]. The process is also reported to improve the surface finish and reduce chatter during machining [13,14,17]. In addition, TAM is also capable of extending the machining speeds to levels not otherwise possible during conventional machining due to limitations of catastrophic tool failures [7], thus, the process has great potential to improve productivity [8,18]. When machining hardened steels it is reported that TAM significantly reduces the rate of abrasive wear, notching wear and chipping [7,13]. n Corresponding author at: Defence Materials Technology Centre, School of Mechanical and Mining Engineering, The University of Queensland, Australia. E-mail address: [email protected] (M.J. Bermingham).
http://dx.doi.org/10.1016/j.wear.2015.03.025 0043-1648/& 2015 Elsevier B.V. All rights reserved.
It is believed that the most dramatic improvement in the machinability of ferrous alloys using thermally assisted machining practises occurs for alloys at the harder end of the spectrum. For example, Chung-Fai [19] performed TAM on AISI 4140 heat treated to different hardness levels and reports that the tool life during TAM increases by 100%, 600% and 900% at 330 HV, 380 HV and 400 HV respectively. Other reports also suggest that TAM is most effective on alloys with higher hardness than softer alloys [18]. It is understood that alloys with high hardness have high yield strengths and are prone to strain hardening, which is reduced by increasing the temperature during thermally assisted machining. However, in machining softer alloys, increasing the temperature has less effect on reducing the shear stress since the material already has a weak tendency to strain harden at room temperature [19]. 17-4PH is a common precipitation hardened martensitic stainless steel used in numerous applications including oil field valve parts, aircraft fittings, fasteners, pump shafts, gears, nuclear reactor components, missile fittings and jet engine parts [20]. Unlike some applications where machining is performed in a softened state followed by post-machining heat treatments, hardened steels including 17-4PH used in aerospace applications are commonly machined in their fully hardened final state to ensure the highest possible quality products (i.e. no risk of dimensional changes or contamination from post-machining heat treatment). Machining steels in the fully hardened state can also lower costs associated with the alternative traditional approach of machining in the annealed state followed by heat treatment, grinding and manual finishing [21]. In a survey of aerospace machining manufacturers, The Cincinnati Milling Company reported that an
M.J. Bermingham et al. / Wear 328-329 (2015) 518–530
average of 40% of all machining operations on these materials occur when they are hardened as high as 58 HRC [18]. The practise of machining aerospace components from ferrous materials in their fully hardened state still occurs today. Consequently, hardened 17-4PH may be a candidate for thermally assisted machining processes. Despite the quantity of research undertaken in developing thermally assisted machining technologies for ferrous alloys, few studies have applied the concept to milling and no literature is available on thermally assisted milling 17-4PH stainless steel. The aim of this research is to compare laser assisted milling (LAM) and conventional room temperature milling of 17-4PH and determine if LAM can reduce the rate of tool wear and extend cutter life. This study also aims to thoroughly characterise the dominant wear mechanisms during the cutting process. In this study two different types of milling cutters are explored. Firstly, milling is performed using a traditional low feed square shoulder carbide cutter at ‘conventional’ cutting speeds and feeds, a practise common in industry. In addition to this, a cutting trial using high feed milling tools is also performed. High feed milling is an emerging roughing process that is believed to reduce cutter vibrations through efficient cutter design. The key difference between traditional low feed milling and high feed milling is that the feed rate is significantly larger (i.e. 5 times larger) and the depth of cut is much smaller (typically o1 mm) in high feed milling compared to traditional low feed milling. We hypothesis that laser assisted high feed milling may be an effective process considering that laser preheating is fundamentally localised to shallow depths, and therefore, traditional low feed milling processes that employ deep cutting depths are less suitable for laser preheating.
2. Experimental
519
milling cutters were tested in conventional (room temperature) and laser assisted milling. All machining was performed on a 5-axis HASS VF3 YTR CNC in climb milling. Two cutting speeds were tested for each tooling condition with the lower speed of each test determined by the tooling manufacturer's recommended parameters and the upper speed increased 50% beyond the recommended speed. Speeds were increased beyond the recommended speed because the ambition of laser assisted milling is to enable higher than usual cutting speeds and thus, improve productivity. The feed rate (chip load) was kept constant during all tests, consequently, as the cutting speed increased the table speed also increased in order to maintain a constant feed rate (which was 0.11 mm/tooth during the traditional low feed milling experiments and 0.5 mm/tooth with the high feed milling experiments). Table 1 shows a summary of the cutting parameters used during testing. It is worth noting that the tools used for this test are generally considered ‘roughing’ tools and are not suited for final ‘finishing’ machining operations due to the tendency to produce poor surface finish. 2.2. Work-piece material Milling was performed on martensitic precipitation hardened 17-4 stainless steel (nominal composition 17 wt% nickel, 4 wt% chromium, and 4 wt% copper) in the H900 heat treated condition (solution treated at 1040 °C plus aging/tempering at 480 °C for 1 h, bulk hardness 450 720 HV0.05). The material was supplied in a 178x109 100 mm3 block and milling was performed in a linear direction parallel to the edge of the work-piece. After machining the entire length of the block (cutting along the 178 mm edge) with the depth and width of cut specified in Table 1, a parallel step over was performed and milling resumed on the adjacent row. A rotating force dynamometer (Kistler 9124B) was used to measure the cutting forces during machining.
2.1. Experimental design 2.3. Laser assisted machining Linear face milling tests were performed to evaluate the tool performance across a number of different cutting scenarios. High feed milling cutters and traditional square shoulder low feed
A 2.2 kW diode laser was used to preheat material directly ahead of the cutter. Details of how this system is integrated with
Table 1 Details of machining experiments. Test details
Justification
Low feed (traditional) milling: Cutter body Seco R217.69-1616.3-09-2A Tool type Seco XOEX090304FR-ME06 MP1500 Speed Vc ¼78, 117 m/min Feed f¼ 0.11 mm/tooth Table speed 171, 256 mm/min Spindle speed 1552, 2328 RPM Depth and width of cut ap ¼ 1 mm, ae ¼6 mm High feed milling: Cutter body Tool type Speed Feed Table speed Spindle speed Depth and width of cut Room temperature machining LAM
R217.21-0816.RE-LP06.2A Seco LPHW060310TR-MD07 MP2500 Vc ¼100, 150 m/min f¼ 0.5 mm/tooth 2000, 3000 mm/min 2000, 3000 RPM ap ¼ 0.5 mm, ae ¼ 9.9 mm
Ø16 mm twin insert square shoulder milling cutter. Single insert used to conserve material CVD-coated Ti(C,N)–Al2O3 carbide grade for medium rough milling, recommended for milling hardened steels Standard recommended cutting condition is Vc ¼ 60 m/min, f ¼0.08 mm/tooth at full width of cut (ae ¼ 16 mm). Seco recommend increasing speed and feed by 1.3 at 37% radial cutter engagement. Single insert cutter used per test to conserve material Width of cut limited by laser spot size for LAM trials. Depth of cut selected to conserve material
Ø16 mm twin insert high feed milling cutter. Two inserts used CVD-coated Ti(C,N)–Al2O3 carbide grade for milling difficult stainless steels Recommended cutting condition is Vc ¼ 100 m/min, f¼ 0.08 mm/tooth at full width of cut (ae ¼16 mm). Seco recommend increasing speed and feed by 1.3 at 37% radial cutter engagement. Two inserts were used for testing
Width of cut is limited by laser beam size for LAM trials and recommendation from Seco to use large cutter engagement during high feed milling. Depth of cut also recommended from Seco and limited by tool geometry No coolant or compressed air used, machining performed at room temperature 2.2 kW Laserline diode laser. Traditional milling cutter: circular beam size approximately 4–5 mm, Power 210 W and 260 W; High feed milling: Line beam (10 1 mm2), Power 1500 W and 2100 W (more details in Section 2.3)
520
M.J. Bermingham et al. / Wear 328-329 (2015) 518–530
the milling machine can be found elsewhere [22]. Since two different cutting widths were tested for the low feed and high feed milling experiments (see Table 1), it was necessary to use two different laser optics lenses capable of producing different beam patterns. For traditional low feed milling, a circular laser focus was used (approximately 5 mm diameter spot). For high feed milling, a line beam was used (approximately 10 1 mm2). In each case the laser optic to tool distance was 200 mm and focused the beam pattern (circular spot or line) onto the work-piece, positioned approximately 2 mm ahead of the cutter. This heats the material directly ahead of the cutter. Fig. 1 shows the two different beam profiles. A thermal calibration test was performed in order to determine the laser power requirements for the machining experiments. In thermally assisted machining trials it is commonly reported that an ‘optimal temperature’ exists whereby heating the work-piece to this temperature maximises the tool life. This temperature is a balance between work-piece softening, tool softening and workpiece tool reactions. The desired heating temperature for the laser assisted experiments was targeted at 300 7 25 °C and it was necessary to maintain this constant for all tests. This temperature was selected because in laser assisted machining stainless steel, Anderson and Shin [16] reported that improved tool life occurred when the heating temperature was between 250 and 340 °C and that heating above 400 °C decreased tool life. Although 17-4PH is a
non-austenitic stainless steel, in this work it is anticipated that heating to approximately 300 °C may also improve the tool life during machining. Wu and Lin [23] have shown that the tensile strength of 17-4PH in the H900 condition decreases by approximately 15% at 300 °C, and consequently, it is expected that this material will respond positively to LAM. The testing procedure for the thermal calibration is identical to that described in other work (full details can be found elsewhere [24]); but in summary involved embedding fast response K-type thermocouples (time constant o0.2 s) into the work-piece at fixed depths from the surface and traversing the surface above the embedded thermocouples with the laser beam at different speeds and laser power levels. The data logging rate was 10 ms to ensure that the relatively fast heating and cooling profile was captured as the laser traverses the work-piece. Fig. 2 shows some heating plots for different travel speeds and laser powers, measured by a thermocouple embedded 1 mm directly below the surface of the work-piece being heated by the laser. It is clear that the laser promotes rapid heating and cooling of a unit of material as the laser beam travels over the surface. In laser assisted milling, a brief cooling period occurs during the time it takes for the laser to move off a unit of heated material and the cutter to approach a unit of heated material. This cooling period is determined by the distance separating the laser beam
Fig. 1. Set-up images for laser assisted milling: (A) photograph showing position of laser optic lens (arrow) and the dynamometer and tooling configuration, (B) photograph of thermal calibration process showing four Ø0.81 mm K-type thermocouples embedded below the surface being heated by the laser beam (note, thermally conductive white paste used to improve measurement accuracy), (C) circular spot beam profile used for traditional low feed milling experiments and (D) line beam used for high feed milling experiments. A red visible ‘pilot’ laser is shown in the photographs as the actual heating laser used for the experiments is invisible (800–980 nm wavelength). (For interpretation of the references to colour in this figure, the reader is referred to the web version of this article.)
M.J. Bermingham et al. / Wear 328-329 (2015) 518–530
521
focus and cutter (in all cases 2 mm) as well as the travel speed of the cutter. In this work, the distance between the laser beam focus and the cutting tool was kept constant at 2 mm for all experiments, which corresponds to cooling periods of 0.7, 0.35, 0.06 and 0.04 s at 171, 341, 2000 and 3000 m/min laser travel speeds, respectively. At the high travel speeds associated with high feed milling (i.e. 2000 and 3000 m/min), this cooling period has insignificant effect on the temperature of the material being cut. However, at the lower travel speeds associated with low feed milling (171 and 341 mm/min), the material cools by approximately 10–20 °C during this brief period. Fig. 3 shows the results of the calibration trial and considers the brief cooling period mentioned above. From this graph, the laser power requirements for the LAM experiments were determined as 210 W, 260 W, 1500 W and 2100 W at 171, 341, 2000 and 3000 m/min, respectively. Fig. 2. Example of laser heating profile measured by a thermocouple embedded 1 mm below the surface of work-piece. Four offset heating plots are shown at different laser travel speeds and power levels. As the travel speed increases, a higher laser power is required to achieve a desired temperature and the rate of heating increases.
2.4. Wear testing and tool life The rate of wear development and tool lives were monitored during the experiments. This involved periodically removing the tools from the cutter body and measuring the wear at various intervals. The intervals were regular and typically occurred after every second or third cut (where each cut was one length of the material block, equal to 178 mm). Flank wear was measured during the tests using an optical microscope, after which the tools were returned to the cutter body and the test resumed. Fig. 4 shows an example of conventional and laser assisted milling in cut.
3. Results and discussion 3.1. Cutting force
Fig. 3. Results of the laser calibration trail for each travel speed. The brief cooling period has been considered in these plots and the data presented is the temperature at the cutter (after 0.7, 0.35, 0.06 and 0.04 s delay at 171, 341, 2000 and 3000 m/min).
Analysis of the measured cutting forces reveals a noticeable difference between conventional and laser assisted milling, shown in Fig. 5. In all cases, laser assisted milling reduced the cutting force by between 16% and 33%, depending on individual cutting conditions. It is also observed that milling speed has a strong correlation with cutting force during both conventional milling and LAM. In both low and high feed milling, it was found that a 50% increase in cutting speed resulted in an average cutting force increase of approximately 30% during conventional milling and LAM. In addition to reducing the cutting forces, laser preheating also eliminates the large cutting force spike that is occasionally
Fig. 4. Images of conventional room temperature milling (A), and laser assisted milling (B), at Vc ¼ 150 m/min (travel speed 3000 mm/min). Note that the chips during conventional milling are dull grey but during LAM they glow bright orange. (For interpretation of the references to colour in this figure, the reader is referred to the web version of this article.)
522
M.J. Bermingham et al. / Wear 328-329 (2015) 518–530
Fig. 5. Cutting forces measured when milling with new tools (cut length is 178 mm). A noticeable reduction in cutting force occurs when the laser is used. Note that the peak in force at the end of cut for LAM at Vc ¼ 78 m/min and Vc ¼ 117 m/min was due to the laser being prematurely shut off, consequently, this material was not fully pre-heated.
M.J. Bermingham et al. / Wear 328-329 (2015) 518–530
523
Fig. 6. Average flank wear for each cutting condition. Note that ‘L.F.’ refers to low feed milling; ‘H.F.’ refers to high feed milling; ‘R.T.’ refers to room temperature (conventional) machining; and ‘LAM’ refers to laser assisted machining (preheated to approx. 300 °C). Error bars show 7 1 standard deviation (based on the second tool wear test performed during high feed milling).
experienced as the tool first engages the cut (e.g. refer to Fig. 5, Vc ¼150 m/min), presumably because the work-piece is softer when heated. The sudden increase in force as the tool enters the cut is a well-known phenomenon in machining that is usually eliminated by slowly ramping up the cutter's feed rate as the cut is engaged. The sudden shock loading on the tool can negatively influence cutter tool life and hence it is desirable to eliminate this loading whenever possible. Consequently, in this case an advantage of LAM is that feed rates do not need to be reduced upon engaging the cut, which inevitably increases productivity. After careful analysis of the cutting force, in all tested conditions it is evident that some degree of chatter/vibration occurs during the milling process. Two types of vibration are evident; a high frequency type of chatter observed in all milling processes (frequencies up to 100 Hz, depending on the testing condition) and a low frequency chatter only noticeable during high feed milling. The low frequency chatter is characterised by cyclic force oscillations of about 100 N that occur at frequencies approaching 0.5–1 Hz, depending on the cutting speed. LAM has no effect on this low frequency chatter, although the amplitude of the oscillations may be slightly less than in conventional room temperature milling. The traditional square shoulder milling cutter (low feed milling) did not experience any noticeable low frequency chatter, although a similar vibration was encountered in other work milling Ti–6Al–4V with square shoulder milling cutters [24]. It is believed that this vibration is caused by the stiffness of the specific CNC mill used in these experiments and is not related to laser assisted milling or milling this particular material in general. The high frequency chatter associated with individual tools engaging and disengaging the cut is observed in both low feed and high feed milling. In most cases, it is easy to see from Fig. 5 that the force between individual cutter engagements oscillates by as much as 200 N (High Feed Milling at room temperature Vc ¼ 150 m/min). All room temperature machining conditions experienced some form of high frequency chatter; however, the severity varied significantly (the most stable condition was high feed milling at Vc ¼100 m/min where the force only oscillated by about 50 N). In some cases, LAM was very effective in reducing this high frequency chatter by lowering the amplitude of force oscillation or by eliminating the oscillations altogether. The most notable case occurred during low feed milling at Vc ¼ 78 m/min where a 100 N oscillating
amplitude experienced during room temperature machining was completely eliminated during LAM. This indicates that LAM is very beneficial in terms of reducing both the magnitude of force subjected to the cutting tools as well as promoting more stable tool loading compared to room temperature machining. 3.2. Tool wear mechanisms The average flank wear at each condition is presented in Fig. 6, plotted against time and volume of material removed. In both low feed and high feed milling, preheating with the laser reduced the wear rate and improved the tool life of the tools in three of the four test conditions. The exception is high feed milling at 100 m/ min which did not show any difference in tool wear rate between LAM and conventional milling. Considering that these two tools had the lowest wear rates of any test, it is expected that longer machining durations are required before any differences in wear rates become apparent. Although it is not possible to distinguish any differences in the wear rates between conventional and LAM during high feed milling at Vc ¼100 m/min, under other machining conditions there are clear differences between the wear rates during conventional and LAM. For example, for a given level of wear, laser assisted milling more than doubled the amount of material removed compared to conventional milling at both Vc ¼ 117 m/min and Vc ¼150 m/min. At Vc ¼78 m/min, LAM increased the amount of material removed by 50% compared to conventional milling. Overall, high feed milling produced lower tool wear rates than traditional low feed milling. Examination of the worn tools suggests that the geometry of the tooling inserts is likely a key factor influencing the wear rates. Figs. 7 and 8 show images of the worn tools after testing (Fig. 7 shows the square shoulder low feed milling tools and Fig. 8 shows the high feed milling tools). Although not particularly clear in the images, the rake and clearance angles for the traditional square shoulder tool is 31° and 9° respectively, whereas, the rake and clearance angles for the high feed tool is 0° and 11° respectively. Consequently, the square shoulder inserts are much sharper tools which influences the flow of the chip being formed. All of the worn square shoulder inserts experience Built Up Edge (BUE) and adhered chip material to the
524
M.J. Bermingham et al. / Wear 328-329 (2015) 518–530
Fig. 7. Backscatter electron images of the worn low feed square shoulder milling inserts (XOEX090304FR-ME06 MP1500), machined at (A) Vc ¼ 78 m/min room temperature machining for 3.2 min, (B) Vc ¼ 78 m/min LAM for 4.5 min, (C) Vc ¼ 117 m/min room temperature machining for 2.1 min and (D) Vc ¼ 117 m/min LAM for 5.1 min. At each cutting speed, LAM enables the tool to be cut for longer and still achieve a similar amount of wear during conventional room temperature machining. Note that an EDS map of the region identified by the red box in image (A) is presented in Fig. 9. (For interpretation of the references to colour in this figure, the reader is referred to the web version of this article.)
M.J. Bermingham et al. / Wear 328-329 (2015) 518–530
525
Fig. 8. Backscatter electron images of the worn high feed milling inserts (LPHW060310TR-MD07 MP2500), machined at (A) Vc ¼100 m/min room temperature machining for 4.9 min, (B) Vc ¼ 100 m/min LAM for 4.9 min, (C) Vc ¼ 150 m/min room temperature machining for 1.3 min and (D) Vc ¼150 m/min LAM for 2.8 min.
526
M.J. Bermingham et al. / Wear 328-329 (2015) 518–530
Fig. 9. EDS maps showing the four major regions of chemical variation. The work-piece material (Fe rich) coats the top surface of the tool, which is Al rich (Al2O3 coating). A narrow band of material Ti rich (TiC) is also observed (this is the base coating of the tool onto which the Al2O3 coating is deposited). In heavily worn areas, the bare WC tool is exposed. The location of this EDS map is shown in Fig. 7A (Cutting conditions: Vc ¼78 m/min conventional milling with square shoulder tool).
M.J. Bermingham et al. / Wear 328-329 (2015) 518–530
Fig. 10. A narrow Al2O3 band is present below the adhered work-piece on the flank of the tools. The band contains grooves typical of abrasion marks (cutting condition: Vc ¼ 117 m/min LAM). (For interpretation of the references to colour in this figure, the reader is referred to the web version of this article.)
tool. In contrast, the blunt high feed cutting inserts show far less BUE, although small deposits of chip are still present. The atomic contrast of backscatter electron imaging reveals four distinct regions of atomic variation that occur on the tools. These variations are observed in all tools and include adhered work-piece (iron rich), the tool's top coating layer (Al2O3), the tool's base coating layer (TiC) and the bare WC–Co tool. These variations are captured in the Energy Dispersive X-ray maps presented in Fig. 9. In the manufacture of these tools, it is believed that a layer of TiC is first deposited onto the straight WC–Co sintered tool, followed by a subsequent coating of Al2O3 onto the TiC layer. A peculiar feature can be observed below the main layer of work-piece adhered to the flank face, which is shown in the images of Figs. 7–9. In all cases, a narrow uniform band of Al2O3, approximately 20 mm wide, follows the profile of the adhered work-piece material. The exterior coating of the tool is Al2O3, as confirmed by EDS and as per the tool manufacturer's specification. It is not surprising that Al2O3 is found in this region as this is expected given that this is the border between the unworn flank face below and the worn flank face above (covered in work-piece material). What is surprising is that this narrow band of Al2O3 is completely uncontaminated of iron and other work-piece constituents, whereas the tool exterior outside of this band is abundantly smeared with small pieces of work-piece (refer to Fig. 9 Fe map). The thin region of TiC coating also appears uncontaminated. A more detailed image of this region shown in Fig. 10 further reveals that this narrow band is covered with abrasion grooves, and is much smoother than the original unworn Al2O3 tool coating. It is not yet clear why a clean band of Al2O3 exists below the adhered work-piece. The fact that abrasion marks are present in this band and the surface is much smoother than the original coating would suggest that this region is rubbing against the freshly machined surface of the work-piece. No further abrasion marks are visible below the narrow Al2O3 band. From surface examination alone it is not possible to determine the bond strength between the small smeared pieces of work-piece adhered to the Al2O3 coating on the flank (purple arrow in Fig. 10); however, it is plausible that this material is only weakly adhered which could explain why it is easily removed when rubbing on the freshly machined surface (i.e. the area of the clean ‘Al2O3 band’). Abrasion and adhesion play a significant role in the wear process of these tools. In multiple regions it is observed that the narrow Al2O3 band merges with a layer of TiC (see Fig. 11C). No
527
clear boundary between these two layers is discernible. This gentle erosion of the top layer (Al2O3) can only be explained by mechanical abrasion or chemical diffusion. In contrast, there are also instances where the removal of the top Al2O3 layer is far more aggressive and occurs by a plucking process (see Fig. 11D). In this example an entire region of the coating has been ripped out to reveal the sub layer of TiC. For this event to occur the work-piece material must have chemically bonded strongly enough with the top Al2O3 coating and overcome the bonds between the Al2O3 and TiC layers. The cracks in the BUE confirm (Fig. 11A) that this material is built up and then removed by the flow of chips. Overall the adhesion mechanism appears to dominate the wear process of these tools. More often than not the removal of the Al2O3 coating by a plucking process also removes the sub TiC coating to reveal the bare WC–Co. Fig. 11E shows a region where the BUE, Al2O3 and TiC layers have been removed to expose the WC particles. Once exposed, these angular particles are subject to further erosion as they protrude proud of the surface (Fig. 11F). According to Matweb Material Property Data, the hardnesses of TiC, Al2O3 and the WC–6%Co tools are 3200, 2085 and 1550 microvickers respectively, indicating that once exposed, erosion of the bare WC–6%Co tool by abrasion and attrition processes is likely to occur at an accelerated rate once the protective coatings are lost. Therefore, increased tool life is associated with the prolonged presence of the tool coatings. Cracking through the coating layers also occurs along the edge of the adhered work-piece, as shown in Fig. 11B and E. All tools (square shoulder low feed inserts and high feed inserts) experienced the same processes, independent of whether laser assisted or not. However, preheating the work-piece with the laser to approximately 300 °C does reduce the rate of abrasive wear significantly, probably a result of lower alloy flow stress and hardness. Wu and Lin [23] report that the tensile strength of 17-4PH in the H900 condition is highly sensitive to temperature and decreases by approximately 15% during lowstrain rate deformation at 300 °C. 3.3. Heat affected zone during LAM A concern associated with laser preheating work-piece materials before machining is the degradation or alteration of microstructure and properties as a consequence of the laser heating. It is highly important to ensure that no heat affected zones exist after LAM otherwise the part may not operate correctly in service. The material used in this study, 17-4PH, is a martensitic steel strengthened by coherent copper precipitates that form during aging at 480 °C. The stability of these precipitates is highly influenced by thermal exposure. At low temperature aging (400 °C), Rack and Kalish [25] reported that the coherent particles transform to incoherent fcc-Cu particles. Long term aging at 400 °C is also known to decompose the martensitic phase as well as influence the formation of other phases, which degrades properties [26]. To check whether LAM alters the microstructure of the 17-4PH a test was conducted in which the laser traversed over the surface of the work-piece. In this test no machining was performed in order to reveal if any heat affect zone occurs in top layer that would otherwise be removed by the cutter if machining did occur. Since temperature calibration confirmed that in all cases the peak temperature exposure is comparable (Section 2.3), heat-affected zone testing was conducted with the line laser beam used for high feed milling trials. Fig. 12 shows the microstructure of the material and hardness at various depths from the surface. While there is some variation in hardness at the surface (an indication that laser has thermally damaged the material), this is contained within the first 50– 200 mm and the hardness at the depth of cut is comparable with the bulk hardness. This confirms that no heat-affected zone is
528
M.J. Bermingham et al. / Wear 328-329 (2015) 518–530
Fig. 11. Wear patterns observed on the worn tools including cracks, adhesion and delamination of coatings. Phases that have been labelled (i.e. Al2O3 / TiC) have been verified by EDS (cutting conditions: (A) Vc ¼ 150 m/min LAM, (B) Vc ¼ 117 m/min LAM (C) Vc ¼ 150 m/min LAM, (D) Vc ¼150 m/min CM, (E) Vc ¼78 m/min CM and (F) Vc ¼ 100 m/min CM).
contained within parts machined with laser assistance using the parameters tested and that any thermally damaged material is confined to shallow sections removed by the cutter.
4. Conclusions This work investigated the tool wear mechanisms during laser assisted milling (work-piece preheated to approximately 300 °C)
and conventional room temperature milling of fully hardened 17-4PH stainless steel. Two different milling tools and strategies were used including a traditional square shoulder low feed milling cutter and a high feed milling cutter. The main findings of the work include the following:
In both low feed and high feed milling, tool life during laser assisted milling improved at various cutting conditions. In some cases the wear rate halved when laser preheating was adopted.
M.J. Bermingham et al. / Wear 328-329 (2015) 518–530
529
Fig. 12. Microstructure below the surface that has been directly heated by laser (1500 W at 2000 mm/min) and corresponding microhardness values at different subsurface depths. At the depth of cut used in the experiments (0.5 mm), no heat-affected zone is present. Note that ‘removed by cutter’ indicates the region that would have been removed if machining was performed.
The dominant tool wear mechanisms were identical in all tools
(both laser assisted and conventional room temperature machining) and included abrasion and adhesion. Abrasion plays an important role in the breakdown of protective tool coatings. Chemical adhesion of the work-piece material to the tool also resulted in the removal of tool coatings and the exposure of tungsten carbide particles. Occasionally the bonding of the work-piece to the tool became sufficiently strong that entire sections of tool were plucked out by built-upedge and flowing chip motion. Laser assisted milling was found to reduce cutting forces by up to 33% and also reduce or eliminate some of the high frequency chatter encountered during room temperature cutting. Laser assisted milling did not create heat affected zones in the machined component.
Acknowledgements The authors would like to acknowledge Mr Daniel Graham and Dr Yanfeng Gao for assistance during machining trials. The authors would also like to acknowledge the support of the Defence Materials Technology Centre (DMTC) and the Queensland Centre for Advanced Materials Processing and Manufacturing (AMPAM). The DMTC was established and is supported under the Australian Government's Defence Future Capability Technology Centres Programme.
References [1] B.C. Tilghman, Cutting Metals by Electricity, United States Patent Office 416873, 1889. [2] S. Tour, L.S. Fletcher, Hot-spot machining, Iron Age 164 (1949) 78–79. [3] E.T. Armstrong, A.S. Cosler, E.F. Katz, Machining of heated metals, Trans. ASME 73 (1951) 35–43. [4] E.J. Krabacher, M.E. Merchant, Basic factors in hot machining of metals, Trans. ASME 73 (1951) 761–776. [5] A.O. Schmidt, Hot milling, Iron Age 163 (1949) 66–70. [6] H. Ding, Y.C. Shin, Laser-assisted machining of hardened steel parts with surface integrity analysis, Int. J. Mach. Tools Manuf. 50 (2010) 106–114. [7] P. Dumitrescu, P. Koshy, J. Stenekes, M.A. Elbestawi, High-power diode laser assisted hard turning of AISI D2 tool steel, Int. J. Mach. Tools Manuf. 46 (2006) 2009–2016. [8] G. Madhavulu, B. Ahmed, Hot machining process for improved metal removal rates in turning operations, J. Mater. Process. Technol. 44 (1994) 199–206. [9] L. Özler, A. İnan, C. Özel, Theoretical and experimental determination of tool life in hot machining of austenitic manganese steel, Int. J. Mach. Tools. Manuf. 41 (2001) 163–172. [10] N. Tosun, L. Özler, A study of tool life in hot machining using artificial neural networks and regression analysis method, J. Mater. Process. Technol. 124 (2002) 99–104. [11] K.P. Maity, P.K. Swain, An experimental investigation of hot-machining to predict tool life, J. Mater. Proc. Technol. 198 (2008) 344–349. [12] M.A. Lajis, A. Amin, A. Karim, H. Radzi, T. Ginta, Hot machining of hardened steels with coated carbide inserts, Am. J. Eng. Appl. Sci. 2 (2009) 421. [13] A.K.M.N. Amin, I. Abraham, N. Khairusshima, M.I. Ahmed, Influence of preheating on performance of circular carbide inserts in end milling of carbon steel, J. Mater. Process. Technol. 185 (2007) 97–105. [14] A.K.M.N. Amin, M. Abdelgadir, The effect of preheating of work material on chatter during end milling of medium carbon steel performed on a vertical machining center (VMC), J. Manuf. Sci. Eng. 125 (2003) 674–680.
530
M.J. Bermingham et al. / Wear 328-329 (2015) 518–530
[15] A.K.M. Nurul Amin, N.V. Talantov, Influence of the instability of chip formation and preheating of work on tool life in machining high temperature resistant steel and titanium alloys, Mech. Eng. Res. Bull. 9 (1986) 52–62. [16] M. Anderson, Y.C. Shin, Laser-assisted machining of an austenitic stainless steel: P550, Proc. Inst. Mech. Eng.: Part B 220 (2006) 2055. [17] A.K.M.N. Amin, S.B. Dolah, M.B. Mahmud, M.A. Lajis, Effects of workpiece preheating on surface roughness, chatter and tool performance during end milling of hardened steel D2, J. Mater. Process. Technol. 201 (2008) 466–470. [18] I.A. Dickter, C.L. Mehl, R.F. Henke, High Temperature Machining Methods, DTIC Document AD0401303, Cincinnati Milling Maching Co., 1963. [19] H. Chung-Fai, Hot Machining of Alloy Steels, The University of Hong Kong, Hong Kong, China, 1976 315 pp.. [20] C.N. Hsiao, C.S. Chiou, J.R. Yang, Aging reactions in a 17-4 PH stainless steel, Mater. Chem. Phys. 74 (2002) 134–142. [21] P. Koshy, R.C. Dewes, D.K. Aspinwall, High speed end milling of hardened AISI D2 tool steel (58 HRC), J. Mater. Process. Technol. 127 (2002) 266–273.
[22] M.J. Bermingham, P. Schaffarzyk, S. Palanisamy, M.S. Dargusch, Laser-assisted milling strategies with different cutting tool paths, Int. J. Adv. Manuf. Technol. 74 (2014) 1487–1494. [23] J.-H. Wu, C.-K. Lin, Tensile and fatigue properties of 17-4 PH stainless steel at high temperatures, Metall. Mater. Trans. A 33 (2002) 1715–1724. [24] M.J. Bermingham, W.M. Sim, S. Gardiner, M.S. Dargusch, Tool life and wear mechanisms in laser assisted milling Ti–6Al–4V, Wear 322–323 (2015) 151–163. [25] H.J. Rack, D. Kalish, The strength, fracture toughness and low cycle fatigue behavior of 17-4 PH stainless steel, Metall. Trans. 5 (1974) 1595–1605. [26] M. Murayama, Y. Katayama, K. Hono, Microstructural evolution in a 17-4 PH stainless steel after aging at 400 °C, Metall. Mater. Trans. A 30 (1999) 345–353.
Contents lists available at ScienceDirect
Wear journal homepage: www.elsevier.com/locate/wear
A new understanding of the wear processes during laser assisted milling 17-4 precipitation hardened stainless steel M.J. Bermingham a,b,n, D. Kent b, M.S. Dargusch a,b a b
Defence Materials Technology Centre, School of Mechanical and Mining Engineering, The University of Queensland, Australia Queensland Centre for Advanced Materials Processing and Manufacturing (AMPAM), The University of Queensland, Australia
art ic l e i nf o
a b s t r a c t
Article history: Received 12 November 2014 Received in revised form 26 March 2015 Accepted 28 March 2015 Available online 3 April 2015
Laser assisted machining is known to improve the machinability of several difficult to cut materials. For the first time, this study investigates the tool wear rates and the wear mechanisms associated with milling a precipitation hardened martensitic stainless steel with and without the assistance of laser preheating. Across both traditional low feed milling and emerging high feed milling techniques, laser assistance was found to reduce the tool wear rates by up to 50% and lower the cutting force by up to 33% in comparison to conventional room temperature machining. In all cases it is observed that tool coating breakdown by abrasive and adhesive wear processes is the dominant tool failure mechanism. Laser assisted milling is effective in prolonging tool life by delaying these processes in comparison to conventional machining. & 2015 Elsevier B.V. All rights reserved.
Keywords: Steel Laser processing Cutting tools Wear testing
1. Introduction Thermally assisted machining (TAM) has been employed for over a century to improve the machinability of difficult to cut materials. Believed to be first pioneered by B.C Tilghman in 1889 [1], the principle involves using heat to soften materials thereby making them easier to remove with a cutting tool. The process was slow to gain interest but around the 1950s research efforts accelerated to investigate the merits of the process [2–5]. Since this time a great deal of research has been performed and in the last few decades alone, it has been well established that thermally assisted machining is a viable method for improving the machinability of numerous materials. In ferrous alloys, the process is beneficial in terms of reducing cutting forces and improving tool life compared to conventional machining [6–16]. The process is also reported to improve the surface finish and reduce chatter during machining [13,14,17]. In addition, TAM is also capable of extending the machining speeds to levels not otherwise possible during conventional machining due to limitations of catastrophic tool failures [7], thus, the process has great potential to improve productivity [8,18]. When machining hardened steels it is reported that TAM significantly reduces the rate of abrasive wear, notching wear and chipping [7,13]. n Corresponding author at: Defence Materials Technology Centre, School of Mechanical and Mining Engineering, The University of Queensland, Australia. E-mail address: [email protected] (M.J. Bermingham).
http://dx.doi.org/10.1016/j.wear.2015.03.025 0043-1648/& 2015 Elsevier B.V. All rights reserved.
It is believed that the most dramatic improvement in the machinability of ferrous alloys using thermally assisted machining practises occurs for alloys at the harder end of the spectrum. For example, Chung-Fai [19] performed TAM on AISI 4140 heat treated to different hardness levels and reports that the tool life during TAM increases by 100%, 600% and 900% at 330 HV, 380 HV and 400 HV respectively. Other reports also suggest that TAM is most effective on alloys with higher hardness than softer alloys [18]. It is understood that alloys with high hardness have high yield strengths and are prone to strain hardening, which is reduced by increasing the temperature during thermally assisted machining. However, in machining softer alloys, increasing the temperature has less effect on reducing the shear stress since the material already has a weak tendency to strain harden at room temperature [19]. 17-4PH is a common precipitation hardened martensitic stainless steel used in numerous applications including oil field valve parts, aircraft fittings, fasteners, pump shafts, gears, nuclear reactor components, missile fittings and jet engine parts [20]. Unlike some applications where machining is performed in a softened state followed by post-machining heat treatments, hardened steels including 17-4PH used in aerospace applications are commonly machined in their fully hardened final state to ensure the highest possible quality products (i.e. no risk of dimensional changes or contamination from post-machining heat treatment). Machining steels in the fully hardened state can also lower costs associated with the alternative traditional approach of machining in the annealed state followed by heat treatment, grinding and manual finishing [21]. In a survey of aerospace machining manufacturers, The Cincinnati Milling Company reported that an
M.J. Bermingham et al. / Wear 328-329 (2015) 518–530
average of 40% of all machining operations on these materials occur when they are hardened as high as 58 HRC [18]. The practise of machining aerospace components from ferrous materials in their fully hardened state still occurs today. Consequently, hardened 17-4PH may be a candidate for thermally assisted machining processes. Despite the quantity of research undertaken in developing thermally assisted machining technologies for ferrous alloys, few studies have applied the concept to milling and no literature is available on thermally assisted milling 17-4PH stainless steel. The aim of this research is to compare laser assisted milling (LAM) and conventional room temperature milling of 17-4PH and determine if LAM can reduce the rate of tool wear and extend cutter life. This study also aims to thoroughly characterise the dominant wear mechanisms during the cutting process. In this study two different types of milling cutters are explored. Firstly, milling is performed using a traditional low feed square shoulder carbide cutter at ‘conventional’ cutting speeds and feeds, a practise common in industry. In addition to this, a cutting trial using high feed milling tools is also performed. High feed milling is an emerging roughing process that is believed to reduce cutter vibrations through efficient cutter design. The key difference between traditional low feed milling and high feed milling is that the feed rate is significantly larger (i.e. 5 times larger) and the depth of cut is much smaller (typically o1 mm) in high feed milling compared to traditional low feed milling. We hypothesis that laser assisted high feed milling may be an effective process considering that laser preheating is fundamentally localised to shallow depths, and therefore, traditional low feed milling processes that employ deep cutting depths are less suitable for laser preheating.
2. Experimental
519
milling cutters were tested in conventional (room temperature) and laser assisted milling. All machining was performed on a 5-axis HASS VF3 YTR CNC in climb milling. Two cutting speeds were tested for each tooling condition with the lower speed of each test determined by the tooling manufacturer's recommended parameters and the upper speed increased 50% beyond the recommended speed. Speeds were increased beyond the recommended speed because the ambition of laser assisted milling is to enable higher than usual cutting speeds and thus, improve productivity. The feed rate (chip load) was kept constant during all tests, consequently, as the cutting speed increased the table speed also increased in order to maintain a constant feed rate (which was 0.11 mm/tooth during the traditional low feed milling experiments and 0.5 mm/tooth with the high feed milling experiments). Table 1 shows a summary of the cutting parameters used during testing. It is worth noting that the tools used for this test are generally considered ‘roughing’ tools and are not suited for final ‘finishing’ machining operations due to the tendency to produce poor surface finish. 2.2. Work-piece material Milling was performed on martensitic precipitation hardened 17-4 stainless steel (nominal composition 17 wt% nickel, 4 wt% chromium, and 4 wt% copper) in the H900 heat treated condition (solution treated at 1040 °C plus aging/tempering at 480 °C for 1 h, bulk hardness 450 720 HV0.05). The material was supplied in a 178x109 100 mm3 block and milling was performed in a linear direction parallel to the edge of the work-piece. After machining the entire length of the block (cutting along the 178 mm edge) with the depth and width of cut specified in Table 1, a parallel step over was performed and milling resumed on the adjacent row. A rotating force dynamometer (Kistler 9124B) was used to measure the cutting forces during machining.
2.1. Experimental design 2.3. Laser assisted machining Linear face milling tests were performed to evaluate the tool performance across a number of different cutting scenarios. High feed milling cutters and traditional square shoulder low feed
A 2.2 kW diode laser was used to preheat material directly ahead of the cutter. Details of how this system is integrated with
Table 1 Details of machining experiments. Test details
Justification
Low feed (traditional) milling: Cutter body Seco R217.69-1616.3-09-2A Tool type Seco XOEX090304FR-ME06 MP1500 Speed Vc ¼78, 117 m/min Feed f¼ 0.11 mm/tooth Table speed 171, 256 mm/min Spindle speed 1552, 2328 RPM Depth and width of cut ap ¼ 1 mm, ae ¼6 mm High feed milling: Cutter body Tool type Speed Feed Table speed Spindle speed Depth and width of cut Room temperature machining LAM
R217.21-0816.RE-LP06.2A Seco LPHW060310TR-MD07 MP2500 Vc ¼100, 150 m/min f¼ 0.5 mm/tooth 2000, 3000 mm/min 2000, 3000 RPM ap ¼ 0.5 mm, ae ¼ 9.9 mm
Ø16 mm twin insert square shoulder milling cutter. Single insert used to conserve material CVD-coated Ti(C,N)–Al2O3 carbide grade for medium rough milling, recommended for milling hardened steels Standard recommended cutting condition is Vc ¼ 60 m/min, f ¼0.08 mm/tooth at full width of cut (ae ¼ 16 mm). Seco recommend increasing speed and feed by 1.3 at 37% radial cutter engagement. Single insert cutter used per test to conserve material Width of cut limited by laser spot size for LAM trials. Depth of cut selected to conserve material
Ø16 mm twin insert high feed milling cutter. Two inserts used CVD-coated Ti(C,N)–Al2O3 carbide grade for milling difficult stainless steels Recommended cutting condition is Vc ¼ 100 m/min, f¼ 0.08 mm/tooth at full width of cut (ae ¼16 mm). Seco recommend increasing speed and feed by 1.3 at 37% radial cutter engagement. Two inserts were used for testing
Width of cut is limited by laser beam size for LAM trials and recommendation from Seco to use large cutter engagement during high feed milling. Depth of cut also recommended from Seco and limited by tool geometry No coolant or compressed air used, machining performed at room temperature 2.2 kW Laserline diode laser. Traditional milling cutter: circular beam size approximately 4–5 mm, Power 210 W and 260 W; High feed milling: Line beam (10 1 mm2), Power 1500 W and 2100 W (more details in Section 2.3)
520
M.J. Bermingham et al. / Wear 328-329 (2015) 518–530
the milling machine can be found elsewhere [22]. Since two different cutting widths were tested for the low feed and high feed milling experiments (see Table 1), it was necessary to use two different laser optics lenses capable of producing different beam patterns. For traditional low feed milling, a circular laser focus was used (approximately 5 mm diameter spot). For high feed milling, a line beam was used (approximately 10 1 mm2). In each case the laser optic to tool distance was 200 mm and focused the beam pattern (circular spot or line) onto the work-piece, positioned approximately 2 mm ahead of the cutter. This heats the material directly ahead of the cutter. Fig. 1 shows the two different beam profiles. A thermal calibration test was performed in order to determine the laser power requirements for the machining experiments. In thermally assisted machining trials it is commonly reported that an ‘optimal temperature’ exists whereby heating the work-piece to this temperature maximises the tool life. This temperature is a balance between work-piece softening, tool softening and workpiece tool reactions. The desired heating temperature for the laser assisted experiments was targeted at 300 7 25 °C and it was necessary to maintain this constant for all tests. This temperature was selected because in laser assisted machining stainless steel, Anderson and Shin [16] reported that improved tool life occurred when the heating temperature was between 250 and 340 °C and that heating above 400 °C decreased tool life. Although 17-4PH is a
non-austenitic stainless steel, in this work it is anticipated that heating to approximately 300 °C may also improve the tool life during machining. Wu and Lin [23] have shown that the tensile strength of 17-4PH in the H900 condition decreases by approximately 15% at 300 °C, and consequently, it is expected that this material will respond positively to LAM. The testing procedure for the thermal calibration is identical to that described in other work (full details can be found elsewhere [24]); but in summary involved embedding fast response K-type thermocouples (time constant o0.2 s) into the work-piece at fixed depths from the surface and traversing the surface above the embedded thermocouples with the laser beam at different speeds and laser power levels. The data logging rate was 10 ms to ensure that the relatively fast heating and cooling profile was captured as the laser traverses the work-piece. Fig. 2 shows some heating plots for different travel speeds and laser powers, measured by a thermocouple embedded 1 mm directly below the surface of the work-piece being heated by the laser. It is clear that the laser promotes rapid heating and cooling of a unit of material as the laser beam travels over the surface. In laser assisted milling, a brief cooling period occurs during the time it takes for the laser to move off a unit of heated material and the cutter to approach a unit of heated material. This cooling period is determined by the distance separating the laser beam
Fig. 1. Set-up images for laser assisted milling: (A) photograph showing position of laser optic lens (arrow) and the dynamometer and tooling configuration, (B) photograph of thermal calibration process showing four Ø0.81 mm K-type thermocouples embedded below the surface being heated by the laser beam (note, thermally conductive white paste used to improve measurement accuracy), (C) circular spot beam profile used for traditional low feed milling experiments and (D) line beam used for high feed milling experiments. A red visible ‘pilot’ laser is shown in the photographs as the actual heating laser used for the experiments is invisible (800–980 nm wavelength). (For interpretation of the references to colour in this figure, the reader is referred to the web version of this article.)
M.J. Bermingham et al. / Wear 328-329 (2015) 518–530
521
focus and cutter (in all cases 2 mm) as well as the travel speed of the cutter. In this work, the distance between the laser beam focus and the cutting tool was kept constant at 2 mm for all experiments, which corresponds to cooling periods of 0.7, 0.35, 0.06 and 0.04 s at 171, 341, 2000 and 3000 m/min laser travel speeds, respectively. At the high travel speeds associated with high feed milling (i.e. 2000 and 3000 m/min), this cooling period has insignificant effect on the temperature of the material being cut. However, at the lower travel speeds associated with low feed milling (171 and 341 mm/min), the material cools by approximately 10–20 °C during this brief period. Fig. 3 shows the results of the calibration trial and considers the brief cooling period mentioned above. From this graph, the laser power requirements for the LAM experiments were determined as 210 W, 260 W, 1500 W and 2100 W at 171, 341, 2000 and 3000 m/min, respectively. Fig. 2. Example of laser heating profile measured by a thermocouple embedded 1 mm below the surface of work-piece. Four offset heating plots are shown at different laser travel speeds and power levels. As the travel speed increases, a higher laser power is required to achieve a desired temperature and the rate of heating increases.
2.4. Wear testing and tool life The rate of wear development and tool lives were monitored during the experiments. This involved periodically removing the tools from the cutter body and measuring the wear at various intervals. The intervals were regular and typically occurred after every second or third cut (where each cut was one length of the material block, equal to 178 mm). Flank wear was measured during the tests using an optical microscope, after which the tools were returned to the cutter body and the test resumed. Fig. 4 shows an example of conventional and laser assisted milling in cut.
3. Results and discussion 3.1. Cutting force
Fig. 3. Results of the laser calibration trail for each travel speed. The brief cooling period has been considered in these plots and the data presented is the temperature at the cutter (after 0.7, 0.35, 0.06 and 0.04 s delay at 171, 341, 2000 and 3000 m/min).
Analysis of the measured cutting forces reveals a noticeable difference between conventional and laser assisted milling, shown in Fig. 5. In all cases, laser assisted milling reduced the cutting force by between 16% and 33%, depending on individual cutting conditions. It is also observed that milling speed has a strong correlation with cutting force during both conventional milling and LAM. In both low and high feed milling, it was found that a 50% increase in cutting speed resulted in an average cutting force increase of approximately 30% during conventional milling and LAM. In addition to reducing the cutting forces, laser preheating also eliminates the large cutting force spike that is occasionally
Fig. 4. Images of conventional room temperature milling (A), and laser assisted milling (B), at Vc ¼ 150 m/min (travel speed 3000 mm/min). Note that the chips during conventional milling are dull grey but during LAM they glow bright orange. (For interpretation of the references to colour in this figure, the reader is referred to the web version of this article.)
522
M.J. Bermingham et al. / Wear 328-329 (2015) 518–530
Fig. 5. Cutting forces measured when milling with new tools (cut length is 178 mm). A noticeable reduction in cutting force occurs when the laser is used. Note that the peak in force at the end of cut for LAM at Vc ¼ 78 m/min and Vc ¼ 117 m/min was due to the laser being prematurely shut off, consequently, this material was not fully pre-heated.
M.J. Bermingham et al. / Wear 328-329 (2015) 518–530
523
Fig. 6. Average flank wear for each cutting condition. Note that ‘L.F.’ refers to low feed milling; ‘H.F.’ refers to high feed milling; ‘R.T.’ refers to room temperature (conventional) machining; and ‘LAM’ refers to laser assisted machining (preheated to approx. 300 °C). Error bars show 7 1 standard deviation (based on the second tool wear test performed during high feed milling).
experienced as the tool first engages the cut (e.g. refer to Fig. 5, Vc ¼150 m/min), presumably because the work-piece is softer when heated. The sudden increase in force as the tool enters the cut is a well-known phenomenon in machining that is usually eliminated by slowly ramping up the cutter's feed rate as the cut is engaged. The sudden shock loading on the tool can negatively influence cutter tool life and hence it is desirable to eliminate this loading whenever possible. Consequently, in this case an advantage of LAM is that feed rates do not need to be reduced upon engaging the cut, which inevitably increases productivity. After careful analysis of the cutting force, in all tested conditions it is evident that some degree of chatter/vibration occurs during the milling process. Two types of vibration are evident; a high frequency type of chatter observed in all milling processes (frequencies up to 100 Hz, depending on the testing condition) and a low frequency chatter only noticeable during high feed milling. The low frequency chatter is characterised by cyclic force oscillations of about 100 N that occur at frequencies approaching 0.5–1 Hz, depending on the cutting speed. LAM has no effect on this low frequency chatter, although the amplitude of the oscillations may be slightly less than in conventional room temperature milling. The traditional square shoulder milling cutter (low feed milling) did not experience any noticeable low frequency chatter, although a similar vibration was encountered in other work milling Ti–6Al–4V with square shoulder milling cutters [24]. It is believed that this vibration is caused by the stiffness of the specific CNC mill used in these experiments and is not related to laser assisted milling or milling this particular material in general. The high frequency chatter associated with individual tools engaging and disengaging the cut is observed in both low feed and high feed milling. In most cases, it is easy to see from Fig. 5 that the force between individual cutter engagements oscillates by as much as 200 N (High Feed Milling at room temperature Vc ¼ 150 m/min). All room temperature machining conditions experienced some form of high frequency chatter; however, the severity varied significantly (the most stable condition was high feed milling at Vc ¼100 m/min where the force only oscillated by about 50 N). In some cases, LAM was very effective in reducing this high frequency chatter by lowering the amplitude of force oscillation or by eliminating the oscillations altogether. The most notable case occurred during low feed milling at Vc ¼ 78 m/min where a 100 N oscillating
amplitude experienced during room temperature machining was completely eliminated during LAM. This indicates that LAM is very beneficial in terms of reducing both the magnitude of force subjected to the cutting tools as well as promoting more stable tool loading compared to room temperature machining. 3.2. Tool wear mechanisms The average flank wear at each condition is presented in Fig. 6, plotted against time and volume of material removed. In both low feed and high feed milling, preheating with the laser reduced the wear rate and improved the tool life of the tools in three of the four test conditions. The exception is high feed milling at 100 m/ min which did not show any difference in tool wear rate between LAM and conventional milling. Considering that these two tools had the lowest wear rates of any test, it is expected that longer machining durations are required before any differences in wear rates become apparent. Although it is not possible to distinguish any differences in the wear rates between conventional and LAM during high feed milling at Vc ¼100 m/min, under other machining conditions there are clear differences between the wear rates during conventional and LAM. For example, for a given level of wear, laser assisted milling more than doubled the amount of material removed compared to conventional milling at both Vc ¼ 117 m/min and Vc ¼150 m/min. At Vc ¼78 m/min, LAM increased the amount of material removed by 50% compared to conventional milling. Overall, high feed milling produced lower tool wear rates than traditional low feed milling. Examination of the worn tools suggests that the geometry of the tooling inserts is likely a key factor influencing the wear rates. Figs. 7 and 8 show images of the worn tools after testing (Fig. 7 shows the square shoulder low feed milling tools and Fig. 8 shows the high feed milling tools). Although not particularly clear in the images, the rake and clearance angles for the traditional square shoulder tool is 31° and 9° respectively, whereas, the rake and clearance angles for the high feed tool is 0° and 11° respectively. Consequently, the square shoulder inserts are much sharper tools which influences the flow of the chip being formed. All of the worn square shoulder inserts experience Built Up Edge (BUE) and adhered chip material to the
524
M.J. Bermingham et al. / Wear 328-329 (2015) 518–530
Fig. 7. Backscatter electron images of the worn low feed square shoulder milling inserts (XOEX090304FR-ME06 MP1500), machined at (A) Vc ¼ 78 m/min room temperature machining for 3.2 min, (B) Vc ¼ 78 m/min LAM for 4.5 min, (C) Vc ¼ 117 m/min room temperature machining for 2.1 min and (D) Vc ¼ 117 m/min LAM for 5.1 min. At each cutting speed, LAM enables the tool to be cut for longer and still achieve a similar amount of wear during conventional room temperature machining. Note that an EDS map of the region identified by the red box in image (A) is presented in Fig. 9. (For interpretation of the references to colour in this figure, the reader is referred to the web version of this article.)
M.J. Bermingham et al. / Wear 328-329 (2015) 518–530
525
Fig. 8. Backscatter electron images of the worn high feed milling inserts (LPHW060310TR-MD07 MP2500), machined at (A) Vc ¼100 m/min room temperature machining for 4.9 min, (B) Vc ¼ 100 m/min LAM for 4.9 min, (C) Vc ¼ 150 m/min room temperature machining for 1.3 min and (D) Vc ¼150 m/min LAM for 2.8 min.
526
M.J. Bermingham et al. / Wear 328-329 (2015) 518–530
Fig. 9. EDS maps showing the four major regions of chemical variation. The work-piece material (Fe rich) coats the top surface of the tool, which is Al rich (Al2O3 coating). A narrow band of material Ti rich (TiC) is also observed (this is the base coating of the tool onto which the Al2O3 coating is deposited). In heavily worn areas, the bare WC tool is exposed. The location of this EDS map is shown in Fig. 7A (Cutting conditions: Vc ¼78 m/min conventional milling with square shoulder tool).
M.J. Bermingham et al. / Wear 328-329 (2015) 518–530
Fig. 10. A narrow Al2O3 band is present below the adhered work-piece on the flank of the tools. The band contains grooves typical of abrasion marks (cutting condition: Vc ¼ 117 m/min LAM). (For interpretation of the references to colour in this figure, the reader is referred to the web version of this article.)
tool. In contrast, the blunt high feed cutting inserts show far less BUE, although small deposits of chip are still present. The atomic contrast of backscatter electron imaging reveals four distinct regions of atomic variation that occur on the tools. These variations are observed in all tools and include adhered work-piece (iron rich), the tool's top coating layer (Al2O3), the tool's base coating layer (TiC) and the bare WC–Co tool. These variations are captured in the Energy Dispersive X-ray maps presented in Fig. 9. In the manufacture of these tools, it is believed that a layer of TiC is first deposited onto the straight WC–Co sintered tool, followed by a subsequent coating of Al2O3 onto the TiC layer. A peculiar feature can be observed below the main layer of work-piece adhered to the flank face, which is shown in the images of Figs. 7–9. In all cases, a narrow uniform band of Al2O3, approximately 20 mm wide, follows the profile of the adhered work-piece material. The exterior coating of the tool is Al2O3, as confirmed by EDS and as per the tool manufacturer's specification. It is not surprising that Al2O3 is found in this region as this is expected given that this is the border between the unworn flank face below and the worn flank face above (covered in work-piece material). What is surprising is that this narrow band of Al2O3 is completely uncontaminated of iron and other work-piece constituents, whereas the tool exterior outside of this band is abundantly smeared with small pieces of work-piece (refer to Fig. 9 Fe map). The thin region of TiC coating also appears uncontaminated. A more detailed image of this region shown in Fig. 10 further reveals that this narrow band is covered with abrasion grooves, and is much smoother than the original unworn Al2O3 tool coating. It is not yet clear why a clean band of Al2O3 exists below the adhered work-piece. The fact that abrasion marks are present in this band and the surface is much smoother than the original coating would suggest that this region is rubbing against the freshly machined surface of the work-piece. No further abrasion marks are visible below the narrow Al2O3 band. From surface examination alone it is not possible to determine the bond strength between the small smeared pieces of work-piece adhered to the Al2O3 coating on the flank (purple arrow in Fig. 10); however, it is plausible that this material is only weakly adhered which could explain why it is easily removed when rubbing on the freshly machined surface (i.e. the area of the clean ‘Al2O3 band’). Abrasion and adhesion play a significant role in the wear process of these tools. In multiple regions it is observed that the narrow Al2O3 band merges with a layer of TiC (see Fig. 11C). No
527
clear boundary between these two layers is discernible. This gentle erosion of the top layer (Al2O3) can only be explained by mechanical abrasion or chemical diffusion. In contrast, there are also instances where the removal of the top Al2O3 layer is far more aggressive and occurs by a plucking process (see Fig. 11D). In this example an entire region of the coating has been ripped out to reveal the sub layer of TiC. For this event to occur the work-piece material must have chemically bonded strongly enough with the top Al2O3 coating and overcome the bonds between the Al2O3 and TiC layers. The cracks in the BUE confirm (Fig. 11A) that this material is built up and then removed by the flow of chips. Overall the adhesion mechanism appears to dominate the wear process of these tools. More often than not the removal of the Al2O3 coating by a plucking process also removes the sub TiC coating to reveal the bare WC–Co. Fig. 11E shows a region where the BUE, Al2O3 and TiC layers have been removed to expose the WC particles. Once exposed, these angular particles are subject to further erosion as they protrude proud of the surface (Fig. 11F). According to Matweb Material Property Data, the hardnesses of TiC, Al2O3 and the WC–6%Co tools are 3200, 2085 and 1550 microvickers respectively, indicating that once exposed, erosion of the bare WC–6%Co tool by abrasion and attrition processes is likely to occur at an accelerated rate once the protective coatings are lost. Therefore, increased tool life is associated with the prolonged presence of the tool coatings. Cracking through the coating layers also occurs along the edge of the adhered work-piece, as shown in Fig. 11B and E. All tools (square shoulder low feed inserts and high feed inserts) experienced the same processes, independent of whether laser assisted or not. However, preheating the work-piece with the laser to approximately 300 °C does reduce the rate of abrasive wear significantly, probably a result of lower alloy flow stress and hardness. Wu and Lin [23] report that the tensile strength of 17-4PH in the H900 condition is highly sensitive to temperature and decreases by approximately 15% during lowstrain rate deformation at 300 °C. 3.3. Heat affected zone during LAM A concern associated with laser preheating work-piece materials before machining is the degradation or alteration of microstructure and properties as a consequence of the laser heating. It is highly important to ensure that no heat affected zones exist after LAM otherwise the part may not operate correctly in service. The material used in this study, 17-4PH, is a martensitic steel strengthened by coherent copper precipitates that form during aging at 480 °C. The stability of these precipitates is highly influenced by thermal exposure. At low temperature aging (400 °C), Rack and Kalish [25] reported that the coherent particles transform to incoherent fcc-Cu particles. Long term aging at 400 °C is also known to decompose the martensitic phase as well as influence the formation of other phases, which degrades properties [26]. To check whether LAM alters the microstructure of the 17-4PH a test was conducted in which the laser traversed over the surface of the work-piece. In this test no machining was performed in order to reveal if any heat affect zone occurs in top layer that would otherwise be removed by the cutter if machining did occur. Since temperature calibration confirmed that in all cases the peak temperature exposure is comparable (Section 2.3), heat-affected zone testing was conducted with the line laser beam used for high feed milling trials. Fig. 12 shows the microstructure of the material and hardness at various depths from the surface. While there is some variation in hardness at the surface (an indication that laser has thermally damaged the material), this is contained within the first 50– 200 mm and the hardness at the depth of cut is comparable with the bulk hardness. This confirms that no heat-affected zone is
528
M.J. Bermingham et al. / Wear 328-329 (2015) 518–530
Fig. 11. Wear patterns observed on the worn tools including cracks, adhesion and delamination of coatings. Phases that have been labelled (i.e. Al2O3 / TiC) have been verified by EDS (cutting conditions: (A) Vc ¼ 150 m/min LAM, (B) Vc ¼ 117 m/min LAM (C) Vc ¼ 150 m/min LAM, (D) Vc ¼150 m/min CM, (E) Vc ¼78 m/min CM and (F) Vc ¼ 100 m/min CM).
contained within parts machined with laser assistance using the parameters tested and that any thermally damaged material is confined to shallow sections removed by the cutter.
4. Conclusions This work investigated the tool wear mechanisms during laser assisted milling (work-piece preheated to approximately 300 °C)
and conventional room temperature milling of fully hardened 17-4PH stainless steel. Two different milling tools and strategies were used including a traditional square shoulder low feed milling cutter and a high feed milling cutter. The main findings of the work include the following:
In both low feed and high feed milling, tool life during laser assisted milling improved at various cutting conditions. In some cases the wear rate halved when laser preheating was adopted.
M.J. Bermingham et al. / Wear 328-329 (2015) 518–530
529
Fig. 12. Microstructure below the surface that has been directly heated by laser (1500 W at 2000 mm/min) and corresponding microhardness values at different subsurface depths. At the depth of cut used in the experiments (0.5 mm), no heat-affected zone is present. Note that ‘removed by cutter’ indicates the region that would have been removed if machining was performed.
The dominant tool wear mechanisms were identical in all tools
(both laser assisted and conventional room temperature machining) and included abrasion and adhesion. Abrasion plays an important role in the breakdown of protective tool coatings. Chemical adhesion of the work-piece material to the tool also resulted in the removal of tool coatings and the exposure of tungsten carbide particles. Occasionally the bonding of the work-piece to the tool became sufficiently strong that entire sections of tool were plucked out by built-upedge and flowing chip motion. Laser assisted milling was found to reduce cutting forces by up to 33% and also reduce or eliminate some of the high frequency chatter encountered during room temperature cutting. Laser assisted milling did not create heat affected zones in the machined component.
Acknowledgements The authors would like to acknowledge Mr Daniel Graham and Dr Yanfeng Gao for assistance during machining trials. The authors would also like to acknowledge the support of the Defence Materials Technology Centre (DMTC) and the Queensland Centre for Advanced Materials Processing and Manufacturing (AMPAM). The DMTC was established and is supported under the Australian Government's Defence Future Capability Technology Centres Programme.
References [1] B.C. Tilghman, Cutting Metals by Electricity, United States Patent Office 416873, 1889. [2] S. Tour, L.S. Fletcher, Hot-spot machining, Iron Age 164 (1949) 78–79. [3] E.T. Armstrong, A.S. Cosler, E.F. Katz, Machining of heated metals, Trans. ASME 73 (1951) 35–43. [4] E.J. Krabacher, M.E. Merchant, Basic factors in hot machining of metals, Trans. ASME 73 (1951) 761–776. [5] A.O. Schmidt, Hot milling, Iron Age 163 (1949) 66–70. [6] H. Ding, Y.C. Shin, Laser-assisted machining of hardened steel parts with surface integrity analysis, Int. J. Mach. Tools Manuf. 50 (2010) 106–114. [7] P. Dumitrescu, P. Koshy, J. Stenekes, M.A. Elbestawi, High-power diode laser assisted hard turning of AISI D2 tool steel, Int. J. Mach. Tools Manuf. 46 (2006) 2009–2016. [8] G. Madhavulu, B. Ahmed, Hot machining process for improved metal removal rates in turning operations, J. Mater. Process. Technol. 44 (1994) 199–206. [9] L. Özler, A. İnan, C. Özel, Theoretical and experimental determination of tool life in hot machining of austenitic manganese steel, Int. J. Mach. Tools. Manuf. 41 (2001) 163–172. [10] N. Tosun, L. Özler, A study of tool life in hot machining using artificial neural networks and regression analysis method, J. Mater. Process. Technol. 124 (2002) 99–104. [11] K.P. Maity, P.K. Swain, An experimental investigation of hot-machining to predict tool life, J. Mater. Proc. Technol. 198 (2008) 344–349. [12] M.A. Lajis, A. Amin, A. Karim, H. Radzi, T. Ginta, Hot machining of hardened steels with coated carbide inserts, Am. J. Eng. Appl. Sci. 2 (2009) 421. [13] A.K.M.N. Amin, I. Abraham, N. Khairusshima, M.I. Ahmed, Influence of preheating on performance of circular carbide inserts in end milling of carbon steel, J. Mater. Process. Technol. 185 (2007) 97–105. [14] A.K.M.N. Amin, M. Abdelgadir, The effect of preheating of work material on chatter during end milling of medium carbon steel performed on a vertical machining center (VMC), J. Manuf. Sci. Eng. 125 (2003) 674–680.
530
M.J. Bermingham et al. / Wear 328-329 (2015) 518–530
[15] A.K.M. Nurul Amin, N.V. Talantov, Influence of the instability of chip formation and preheating of work on tool life in machining high temperature resistant steel and titanium alloys, Mech. Eng. Res. Bull. 9 (1986) 52–62. [16] M. Anderson, Y.C. Shin, Laser-assisted machining of an austenitic stainless steel: P550, Proc. Inst. Mech. Eng.: Part B 220 (2006) 2055. [17] A.K.M.N. Amin, S.B. Dolah, M.B. Mahmud, M.A. Lajis, Effects of workpiece preheating on surface roughness, chatter and tool performance during end milling of hardened steel D2, J. Mater. Process. Technol. 201 (2008) 466–470. [18] I.A. Dickter, C.L. Mehl, R.F. Henke, High Temperature Machining Methods, DTIC Document AD0401303, Cincinnati Milling Maching Co., 1963. [19] H. Chung-Fai, Hot Machining of Alloy Steels, The University of Hong Kong, Hong Kong, China, 1976 315 pp.. [20] C.N. Hsiao, C.S. Chiou, J.R. Yang, Aging reactions in a 17-4 PH stainless steel, Mater. Chem. Phys. 74 (2002) 134–142. [21] P. Koshy, R.C. Dewes, D.K. Aspinwall, High speed end milling of hardened AISI D2 tool steel (58 HRC), J. Mater. Process. Technol. 127 (2002) 266–273.
[22] M.J. Bermingham, P. Schaffarzyk, S. Palanisamy, M.S. Dargusch, Laser-assisted milling strategies with different cutting tool paths, Int. J. Adv. Manuf. Technol. 74 (2014) 1487–1494. [23] J.-H. Wu, C.-K. Lin, Tensile and fatigue properties of 17-4 PH stainless steel at high temperatures, Metall. Mater. Trans. A 33 (2002) 1715–1724. [24] M.J. Bermingham, W.M. Sim, S. Gardiner, M.S. Dargusch, Tool life and wear mechanisms in laser assisted milling Ti–6Al–4V, Wear 322–323 (2015) 151–163. [25] H.J. Rack, D. Kalish, The strength, fracture toughness and low cycle fatigue behavior of 17-4 PH stainless steel, Metall. Trans. 5 (1974) 1595–1605. [26] M. Murayama, Y. Katayama, K. Hono, Microstructural evolution in a 17-4 PH stainless steel after aging at 400 °C, Metall. Mater. Trans. A 30 (1999) 345–353.