(a) HPDL beam profile at varying currents (measured on a beam dump not the HY-80 workpiece) and (b) power versus current of the HPDL array. 

Contexts in source publication

Context 1

... so consideration must be given to cooling rates following joining in order to establish a tempered martensitic microstructure. Although considered and established as a future research goal, no specific post welding heat treatment was applied in this research. To determine temperature distribution during laser heating experiments alone, plates of HY-80 were instrumented with Pt/Re thermocouples at depths varying from 0.635 mm (0.025 inches) to 3.175 mm (0.125 inches) at intervals of 25.4 mm (1 inch) in the direction corresponding to the welding direction (Figure 1). In order to maximize material usage, up to three laser heating passes were accomplished on each plate (Figure 1b). In these cases the preheating runs that passed directly over the line of thermocouples (e.g. the 25A pass in Figure 1b) was used for comparison to theoretical data. Temperature measurements were used for comparison to theoretical models to determine laser operating parameters and verify temperature modelling. The HY-80 workpiece was placed on a motion control stage with an air gap between the workpiece and the stage primarily to allow for thermocouple insertion into the workpiece and to protect the stage from heating damage. This is different than would exist in an actual DLAFSW setup since the workpiece would be mounted to an anvil that would further conduct heat from the workpiece. The moving stage moved the workpiece underneath the laser path with the HPDL stationary at desired speeds which were chosen to correspond to typical FSW traverse speeds. A cartoon summary of the setup is shown in Figure 2 and an actual image of the setup is shown in Figure 3. The experimental setup included the following major components: HPDL array with integrated polarization rotator, HY-80 workpiece, moving stage, CCD camera for video recording, thermocouples embedded in the workpiece, and IR camera to measure surface temperature. The HPDL used was a 5 kW diode array that emits incoherent light at a wavelength of 795 nm. Full details of the specification of the HPDL will be published separately. A series of initial experimental runs were completed to measure the beam profile and establish operating conditions of the HPDL array (Figure 4). Increasing the laser current causes a change in the beam profile (Figure 4a) as well as a linear increase in power (Figure 4b). Increasing the beam current causes a proportionally larger expansion of the beam in the slow axis corresponding to the direction perpendicular to the laser and proposed welding path. The beam width (y-direction from Figure 4a) corresponds very closely to the width of the visibly heat regions in Figure 1b, and as shown in Figure 1b, increasing the beam current causes an increase in the width of the visibly heated area. Full details of HPDL specifications including scattered light, optics used, and other laser specific parameters will be published separately. The focus of this paper is temperature response of the workpiece and the suitability of using a HPDL array in combination with FSW. An example of proposed optimal conditions is shown in Figure 5. In this case the workpiece was preheated by the HPDL initially operating at 25A with a 10 sec warmup time followed by a traverse speed of 1.67 mm/sec. This traverse speed correlates to 100 MMPM and was based on previous FSW research on HY-80 that suggests this is a reasonable traverse speed for reasonable tool rotational speeds [35]. After 3 seconds of motion the laser current was increased to 30A (note the increase in heated area of Figure 5 going from right to left). Near the end of the proposed weld path the laser current was reduced to 25A (note the reduction in heated area of Figure 5 going from left to right) prior to securing laser in the vicinity of thermocouple 1 (TC1). The thermocouples in Figure 5 are positioned progressively further from the surface in the line of the laser path (from TC5 to TC1) and as such the subsequent thermocouples each reach a lower peak temperature with the exception of TC5 since the laser began moving away from TC5 before it has fully heated. Temperature response of the HY-80 workpiece was modeled using FlexPDE. The heat input from the HPDL was modeled using the actual beam profiles (Figure 4a) and beam power from beam current (Figure 4b). Experimental to theoretical results for a 25A run at 1.67 mm/sec (100 MMPM) are shown in Figure 6. In this case the laser is operating continuously at 25A with a 10 second warmup time prior to traversing followed by traversing at 1.67 mm/sec (100 MMPM). Unlike the case shown in Figure 5, only 4 thermocouples were used (due to a failure of one thermocouple) and the thermocouples are in increasing proximity to the surface (i.e. TC2 is closer to the surface than TC4) resulting in higher peak temperatures progressing down the thermocouple line. This arrangement of thermocouple depths is opposite to the depths listed in Figure 5 explaining the differences in peak trends between Figure 4 and Figure 5. Theoretical results from FlexPDE in planar view (Figure 7a) and transverse view (Figure 7b) show the theoretical temperature field due to laser preheating alone. In addition to temperature data from thermocouples, thermal lacquers were used on the surface of the workpiece to establish some experimental bounds on the surface temperature of the workpiece. Temperature sensitive thermal lacquers (Omega Engineering Inc., CT, USA) have been used previously for direct measurement of surface temperatures in other laser assisted processing by the authors and results compared closely with other temperature monitoring devices [36]. An example of the comparison between the theoretical and experimental temperature bounds determined by the thermal lacquers is shown in Figure 8a and 8b. Figure 8a is a magnification of theoretical data from Figure 7 and depicts the temperature field for a 25A pass at 1.67 mm/sec. The experimental surface temperature response of this run is shown in Figure 8b. As the thermal lacquers change color corresponding to their temperature value, the width of the changed area can be compared be compared to the theoretical surface temperature. Comparing these methods, an area of approximately 2 cm in width (i.e. perpendicular to the laser and proposed welding path) is heated to approximately 300 ° C on the surface. For comparison, a typical friction stir weld and the FSW tool used to make that weld are shown in Figure 8c and 8d respectively. Temperature In contrast to control previous of HPDL the preheated applications, zone where in hybrid only FSW the surface systems of is the critical part was and under other hybrid consideration, methods in have DLAFSW shown that the entire preheating depth the of workpiece the workpiece may lead needs to to reduced be brought frictional up forces to the and desired improved temperature tool life. (as Where seen in many Figure 7(b)). of these systems This have can come be short an issue is practical for low-conductivity application of the materials systems such either as steel. through Consequently, complicated in connections order to accurately to the workpiece model the or experiment, expensive and temperature- inefficient energy transfer to the weld area alone. This research has shown that a relatively small and inexpensive HPDL system can preheat a workpiece in a very controllable manner that closely approximate theoretical predictions. In contrast to previous HPDL applications, where only the surface of the part was under consideration, in DLAFSW the entire depth of the workpiece needs to be brought up to the desired temperature (as seen in Figure 7(b)). This can be an issue for low-conductivity materials such as steel. Consequently, in order to accurately model the experiment, temperature- dependent properties for the HY-80 steel must be known. Unfortunately, sufficient temperature- dependent data for HY-80 was available. As a plausible substitute, the properties for HY-130 [37] were used and were scaled to the curves of the room-temperature data for HY-80 [38]. Figures 11(a), (b), and (c) show the results for thermal conductivity, density, and specific heat, respectively. Also shown in Figures 11(a) and (b) are analytic fits used in the FlexPDE model. The specific heat data was imported into FlexPDE as a table, so no fits were used in this case. The good agreement between experiment and model suggests that the temperature-dependent properties derived outside these experiments provide a good description of the experiment. As a result, the model provides a useful tool in determining the optimum parameters needed to produce the desired temperature distribution near the tool. One of the primary advantages of FSW is joining without melting. Although a HPDL can easily melt the HY-80 workpiece, it may also be used at varying combinations of power output and welding speeds to heat the workpiece to a desired temperature regime. This research shows that HY-80 may be heated to below it austenitic transformation region (nominally 700 ° C) while also heating only the weld area. Figure 9 shows the strong similarity of FSW SZs and HPDL heated areas for two different steels. This simple comparison shows the primary advantage of proposed DLAFSW system in that the parameters of the HPDL can be optimized such that the heated area corresponds only to the proposed welding area. Similarly the temperature of the workpiece immediately prior to FSW can be controlled to optimize the combined DLAFSW process. This research also shows that theoretical modeling of the laser heating can accurately estimate the temperature field within the workpiece. Full material characterization of the preheated area remains to be completed and an initial proposed DLAFSW has been designed (Figure 10) that uses the exact same HPDL setup as used in this research (Figure 3c). Once the HPDL is combined with a traditional FSW system, full analysis of the reduction in frictional ...

Context 2

... application, of laser a preheating, HPDL would and be determine used to heat the the material workpiece response (ideally due below to the the combined austenite effect region) of FSW immediately and HPDL prior preheating. to FSW such Based that the on stir the zone current (SZ) research, is preheated it is hypothesized and softened that resulting the addition in reduced of preheat tool wear using and laser overall diode less arrays heat input will from significantly the FSW lower process. frictional Subsequent forces, lessen cooling tool could wear, also and be allow assisted for higher by a laser tool heating traverse process speeds. to minimize the cooling rate to prevent the formation of martensite. Given the flexibility of HPDL systems, preheating and post FSW cooling could potentially be controlled; however, the current study focuses primarily on HPDL preheating. The objectives of this research are to: (1) validate theoretical models of temperature profiles of typical steels using HPDL heating alone, (2) use these theoretical models to establish operating parameters of a HPDL system on a typical hardenable steel, and (3) propose a diode laser assisted FSW (DLAFSW) setup that utilizes laser diode arrays for preheating of the workpiece. These objectives were accomplished by initial laser heating tests on a moving sample workpiece with embedded thermocouples. Subsequent research objectives will determine the effect on FSW tool life with the addition of laser preheating, and determine the material response due to the combined effect of FSW and HPDL preheating. Based on the current research, it is hypothesized that the addition of preheat using laser diode arrays will significantly lower frictional forces, lessen tool wear, and allow for higher tool traverse speeds. Based on previous FSW experience and its similarity to other hardenable steels, the low carbon alloy steel HY-80 was selected for this study. Plates of HY-80 (MIL-S-16216) were 6.4 mm (0.25 inches) thick and of varying widths and lengths. HY-80 (0.12-0.18 wt% carbon) is used in Navy applications as well as some pressure vessels and acquires its strength and toughness through quenching and tempering treatments. Additional data on HY-80, its use in Naval applications, and response due to FSW alone can be found elsewhere [33, 34, 35]. Previous research on FSW of HY-80 has shown that peak temperatures in the SZ are sufficient to create austenite which upon cooling will become tempered or untempered martensite based on the cooling rates. Depending on the intended application the formation of martensite in the weld nugget can be beneficial or detrimental, but in most applications requiring toughness, the formation of a brittle untempered martensite phase is undesirable so consideration must be given to cooling rates following joining in order to establish a tempered martensitic microstructure. Although considered and established as a future research goal, no specific post welding heat treatment was applied in this research. To determine temperature distribution during laser heating experiments alone, plates of HY-80 were instrumented with Pt/Re thermocouples at depths varying from 0.635 mm (0.025 inches) to 3.175 mm (0.125 inches) at intervals of 25.4 mm (1 inch) in the direction corresponding to the welding direction (Figure 1). In order to maximize material usage, up to three laser heating passes were accomplished on each plate (Figure 1b). In these cases the preheating runs that passed directly over the line of thermocouples (e.g. the 25A pass in Figure 1b) was used for comparison to theoretical data. Temperature measurements were used for comparison to theoretical models to determine laser operating parameters and verify temperature modelling. The HY-80 workpiece was placed on a motion control stage with an air gap between the workpiece and the stage primarily to allow for thermocouple insertion into the workpiece and to protect the stage from heating damage. This is different than would exist in an actual DLAFSW setup since the workpiece would be mounted to an anvil that would further conduct heat from the workpiece. The moving stage moved the workpiece underneath the laser path with the HPDL stationary at desired speeds which were chosen to correspond to typical FSW traverse speeds. A cartoon summary of the setup is shown in Figure 2 and an actual image of the setup is shown in Figure 3. The experimental setup included the following major components: HPDL array with integrated polarization rotator, HY-80 workpiece, moving stage, CCD camera for video recording, thermocouples embedded in the workpiece, and IR camera to measure surface temperature. The HPDL used was a 5 kW diode array that emits incoherent light at a wavelength of 795 nm. Full details of the specification of the HPDL will be published separately. A series of initial experimental runs were completed to measure the beam profile and establish operating conditions of the HPDL array (Figure 4). Increasing the laser current causes a change in the beam profile (Figure 4a) as well as a linear increase in power (Figure 4b). Increasing the beam current causes a proportionally larger expansion of the beam in the slow axis corresponding to the direction perpendicular to the laser and proposed welding path. The beam width (y-direction from Figure 4a) corresponds very closely to the width of the visibly heat regions in Figure 1b, and as shown in Figure 1b, increasing the beam current causes an increase in the width of the visibly heated area. Full details of HPDL specifications including scattered light, optics used, and other laser specific parameters will be published separately. The focus of this paper is temperature response of the workpiece and the suitability of using a HPDL array in combination with FSW. An example of proposed optimal conditions is shown in Figure 5. In this case the workpiece was preheated by the HPDL initially operating at 25A with a 10 sec warmup time followed by a traverse speed of 1.67 mm/sec. This traverse speed correlates to 100 MMPM and was based on previous FSW research on HY-80 that suggests this is a reasonable traverse speed for reasonable tool rotational speeds [35]. After 3 seconds of motion the laser current was increased to 30A (note the increase in heated area of Figure 5 going from right to left). Near the end of the proposed weld path the laser current was reduced to 25A (note the reduction in heated area of Figure 5 going from left to right) prior to securing laser in the vicinity of thermocouple 1 (TC1). The thermocouples in Figure 5 are positioned progressively further from the surface in the line of the laser path (from TC5 to TC1) and as such the subsequent thermocouples each reach a lower peak temperature with the exception of TC5 since the laser began moving away from TC5 before it has fully heated. Temperature response of the HY-80 workpiece was modeled using FlexPDE. The heat input from the HPDL was modeled using the actual beam profiles (Figure 4a) and beam power from beam current (Figure 4b). Experimental to theoretical results for a 25A run at 1.67 mm/sec (100 MMPM) are shown in Figure 6. In this case the laser is operating continuously at 25A with a 10 second warmup time prior to traversing followed by traversing at 1.67 mm/sec (100 MMPM). Unlike the case shown in Figure 5, only 4 thermocouples were used (due to a failure of one thermocouple) and the thermocouples are in increasing proximity to the surface (i.e. TC2 is closer to the surface than TC4) resulting in higher peak temperatures progressing down the thermocouple line. This arrangement of thermocouple depths is opposite to the depths listed in Figure 5 explaining the differences in peak trends between Figure 4 and Figure 5. Theoretical results from FlexPDE in planar view (Figure 7a) and transverse view (Figure 7b) show the theoretical temperature field due to laser preheating alone. In addition to temperature data from thermocouples, thermal lacquers were used on the surface of the workpiece to establish some experimental bounds on the surface temperature of the workpiece. Temperature sensitive thermal lacquers (Omega Engineering Inc., CT, USA) have been used previously for direct measurement of surface temperatures in other laser assisted processing by the authors and results compared closely with other temperature monitoring devices [36]. An example of the comparison between the theoretical and experimental temperature bounds determined by the thermal lacquers is shown in Figure 8a and 8b. Figure 8a is a magnification of theoretical data from Figure 7 and depicts the temperature field for a 25A pass at 1.67 mm/sec. The experimental surface temperature response of this run is shown in Figure 8b. As the thermal lacquers change color corresponding to their temperature value, the width of the changed area can be compared be compared to the theoretical surface temperature. Comparing these methods, an area of approximately 2 cm in width (i.e. perpendicular to the laser and proposed welding path) is heated to approximately 300 ° C on the surface. For comparison, a typical friction stir weld and the FSW tool used to make that weld are shown in Figure 8c and 8d respectively. Temperature In contrast to control previous of HPDL the preheated applications, zone where in hybrid only FSW the surface systems of is the critical part was and under other hybrid consideration, methods in have DLAFSW shown that the entire preheating depth the of workpiece the workpiece may lead needs to to reduced be brought frictional up forces to the and desired improved temperature tool life. (as Where seen in many Figure 7(b)). of these systems This have can come be short an issue is practical for low-conductivity application of the materials systems such either as steel. through Consequently, complicated in connections order to accurately to the workpiece model the or experiment, expensive and temperature- ...