Development of an Indirect Resistance Brazing Technology for Sandwich Metal Panels

Transcription

1 Development of an Indirect Resistance Brazing Technology for Sandwich Metal Panels Jerry Gould, EWI, Columbus, OH Doug Cox, CellTechMetals, San Diego, CA CellTech Metals is developing a new generation of sandwich sheet and panel structures for predominantly high-volume applications. These sandwich structures maintain the stiffness of solid steel sheet at roughly half the product density. The sandwich structure includes thingauge face sheets as well a one or two dimpled core sheets. The product is brazed together to create the metal sandwich. Key to mass production markets is a low-cost method of accomplishing the braze necessary to assemble the product. A roll-brazing technology for assembly of this product has been demonstrated. This technology utilizes a process of running the components of the sandwich through hot rolls, and through heat conduction accomplished the necessary brazing for assembly. Preliminary work examined, using heattransfer analyses, the achievable processing speeds associated with such technology. Through this analysis, it was estimated that high processing speeds (multiple meters/minute) were possible for thin-gauge (<2-mm) sandwich structures. Work to develop this technology has been progressive, first examining a resistance bar-brazing process, and subsequent adaptation to a sub-scale indirect resistance rollbrazing process. For roll brazing, a prototype assembly was manufactured. Indirect resistance roll brazing was demonstrated on this unit using nominally 150-μm thick galvanized sheets. The core in this study was a stamped with uni-directional corrugations, facilitating a final panel thickness of roughly 2-mm. Work conducted to date suggests that continuous indirect resistance roll brazing of honeycomb panels is feasible, and can be done economically in high volumes. Considerable additional work needs to be done to extend this technology to a production level. Figure 1 Schematic Representation of a Roll Brazing System for Low-Cost Sandwich Panel Construction CellTech Metals has been developing a revolutionary new sandwich panel product designed for utility in high-volume manufacturing markets. These markets include automotive, renewable energy, industrial packaging, and transportation sectors. The metallic sandwich itself is fabricated from stack-ups of thin sheet materials using pre-plated braze alloys. Preliminary work at CellTech Metals and Edison Welding Institute (EWI) (1) has identified a unique technology for highspeed assembly of these individual sheets into the full sandwich structure. This technology has been termed resistance roll brazing (RRB). The basic configuration for RRB is shown in Figure 1. RRB functions through local resistance heating of the 1

2 roll faces and subsequent conduction of this heat into the workpieces. For this to occur, a current is introduced by sliding contacts into the rolls at the position of contact with the workpieces, with conduction and subsequent heating in a line along the contact length. NO CURRENT IS CONDUCTED INTO THE WORKPIECES. Preliminary calculations(1) for this technology have allowed some scaling of both the processing dynamics as well as power requirements. The basic focus of this technology is for brazing with alloys which flow at below 900 C. These preliminary calculations have shown that for rolls heated to these temperatures, full thermal equilibrium of the workpiece (and thus flow of the braze alloy) can occur at processing speeds up to 10 m/min. A typical thermal profile at the center of the sandwich stack-up (the last area to heat) is shown in Figure 2. Figure 3 Allowable Line Speeds for Full Heat Penetration as a Function of Strip and Core Thickness. Analysis is based on a roll temperature of 600oC and a 10-mm contact length. if a flood cooling arrangement is employed, the braze can be resolidified within a few millimeters following exit from the rolls. These analyses were also 2 used to demonstrate the suitability of a single roll brazing system for a range of sandwich configurations. As shown in Figure 3, variations in line speed can facilitate brazing of a wide range of both material and total sandwich thicknesses. Figure 2 Calculated Heating and Cooling Profiles for the Dimple-to-Dimple Contact Surface in a Four-Thickness Stack- Up Sandwich at a Processing Speed of 3 m/min This is for a 1.5-mm-thick sandwich construction brazed at 600 C and a speed of 3 m/min. The results here suggest that the panel can be thermally equilibrated in roughly 100 to 200 ms. Further, In this program a series of experimental studies have been completed to advance the rollbrazing approach. This work includes a preliminary demonstration of indirect resistance brazing in a spot welding configuration, incorporating zinc (Zn)-coated steels. This demonstration was then followed by development of a dedicated resistance bar braze capability on equipment at EWI. This resistance bar brazing process was evaluated again using Zncoated steels. Finally, results from the resistance bar brazing trials were used to define requirements for a prototype IRRB system. 2

3 EXPERIMENTAL SET-UPS FOR RESISTANCE BAR AND ROLL BRAZING Initial resistance bar brazing was done using a Newcor Bay City 10-T welding frame combined with a Miyachi 4-kA inverter power supply. To conduct the resistance brazing trials, dedicated tooling was assembled. This tooling is shown in Figure 4. Tooling allows positioning of two horizontal heating metal bars. Current is passed down the length of these bars generating the necessary heat for brazing. Figure 4 shows the tooling including the bars themselves, as well as the conductors necessary to supply the required heating current. Process development included monitoring for both secondary current and temperature measurements. Welding current was largely monitored using a Miyachi 326-mm current meter. In addition, current was monitored using a toroid, active integration system, and oscilloscope. This set-up allowed the details of the current waveform itself to be assessed. Temperature measurements were taken using a Chromel- Alumel (Type K) thermocouple, calibration module, and oscilloscope. It is understood that the actual current waveforms affect the output of such thermocouple configurations, so data immediately after completion of the heat cycle was used as indicative of heating element temperatures. Figure 4 The Dedicated Resistance Bar Brazing Set-up. Note the AlN coatings on the heating elements Figure 5 Pilot Resistance Roll Brazing System. Shown is the frame, rolls, part fixture, and drive. Figure 6 Details of the Rolls used on the Pilot System. Note the segmentation of the elements making up the roll. The heating elements themselves were electrically isolated from the workpieces through application of AlN surface layers. This was done using an active brazing technique with a TiCuNi filler. Active brazing was done in a vacuum tube furnace. This tube furnace allowed control of the specific temperature cycles used, and could maintain vacuums at a level of 10-7 torr. Roll brazing was done on a dedicated system built at King Machine and Tool Company in Massilion, OH. An over view of the system is provided in Figure 5. This system includes direct refractory metal based segmented rolls nominally 150-mm wide by 150-mm in diameter. Details of the rolls are shown in Figure 6. 3

4 Current for these trials was supplied by a 5-kVA AC transformer. An overtly small transformer was used to provide demonstration of the process. Current was supplied to the rolls by spring loaded contacts. Access of the contacts to the rolls, as well as roll segmentation is shown in Figure 7. (1) (2) Figure 7 Roll Edges Viewed Through the Contactor Access Location in the Frame. Note the segmentation of the roll elements. All brazing trials were done using a nominal mild steel sheet with a 20-μm thick hot-dip galvanized coating. Strip thickness was nominally 200-μm. This strip was also stamped to produce 2-mm wide features for the workpiece core. These features included dimples or corrugations for the bar and roll brazing trials, respectively. These features resulted in core heights of nominally 1-mm. These equations were then used to estimate acceptable bar geometries (widths, heights, lengths) that could achieve brazing temperatures for Zn (600 C) within a timeframe consistent with the RRB estimates from earlier work in this program (1). That work had suggested heating times on the order of 100 to 200 ms for processing speeds of roughly 3 m/min. Results from these equations for a 150-mm-long Mo-titanium-zirconium-molybdenum (TZM) bar are shown in Figure 8. RESISTANCE BAR BRAZING RESULTS Initial efforts to develop resistance bar brazing focused on metallic heating elements that came into direct contact with the workpieces. Prior to designing the process, some calculations were done to assess relationships between bar geometry, bar properties, power demand, heating time, and final temperature. A simple lumped heating analysis yielded the following governing equations. For current demand: Figure 8 Power Analysis Indicating the Current and Voltage Requirements for Resistance Bar Brazing as a Function of Heating Time. 4

5 Results suggest that temperatures in the 600 C range can be achieved with 150-mm-long bars in the 200-ms range for power supply capabilities in the range of 14 V. This, in fact, matches the largest medium-frequency direct current (MFDC) power supply (Miyachi 4-kA system) at EWI. Bar brazing was then done with heating elements nominally 10-mm long x 5-mm thick x 50-mm long (effective heating length). Initial trials were done to assess the uniformity of the temperature along the length of the element. Typical results for a trial conduced at 125-ms heating time and 30- ka current are shown in Figures 9 and 10. Figure 9 shows the resulting current pulse and Figure 9 the temperature excursions at the center and end of the bar. in Figure 10. Of note, the high currents used for resistance heating interfere with the generated thermocouple voltages, so data is only available once the heat cycle is completed. As a result, the data shown here has been modified to eliminate the noise associated with the heating current pulse. Results here show that the temperature between the center and end of the heating bar is within 10%. Figure 10 Temperature Profiles from the Center and End Positions of the Heating Element Following Application of the Current Profile shown in Figure 9. Waveforms have been truncated to eliminate induced noise from the current waveform itself. Figure 9 Current Waveform for Resistance Bar Heating. Peak current in this waveform is 30 ka. Note the 30-ms rise time to peak current, and the drop-off in current level for the final 1/3 of the waveform. Figure 8 shows the peak current for this waveform is achieved in roughly 30 ms. Of note, there is a current drop-off over the last third of the waveform. This probably related to changes in the electrical resistivity of the bar as it heats. Corresponding temperature measurements taken at the center and end of the bar are presented Brazing trials were then done to define a practice capable of making prototype coupons with a Zn filler. Initially, the trials were shifted from displacement (compression) control to load control. This was done to assure intimate contact between the heating elements and workpiece. Trials were initiated at short heating times at maximum voltage output for the equipment (14 V). No temperature measurements were taken in these trials, rather peel testing was used to assess performance. Trials progressed to longer weld times, with increasing amounts of 5

6 bonding. Best practice was observed at 140- ms heating time. It was noted that there was distortion of the face sheets at the edges of the heating element. This was leading to secondary process related cracking of the formed brazes. This was corrected through better cooling, allowing formation of small (50-50-mm) demonstrator samples. An example of a demonstrator sample is provided in Figure 11. The sample was made from progressive resistance bar brazes, each overlapping segment. Imprints from the bar are clearly evident on the sample. finish similar to that seen for the indirect spot brazed process described previously. It is also important to note the ridges of Zn on the specimen. This is caused by displaced coating material to the periphery of the heating element. This is strictly an artifact of the resistance bar brazing process. How any Zn (or other braze coating) will be displaced with a roll-brazing process, or how the surface finish would be effected has yet to be determined. Some further details of the exterior sandwich surface after resistance bar brazing are provided in Figure 12. This photograph indicates both the matt finish, as well as markings associated with the dimpled core sheets. These markings are believed to be areas of coating thinning associated with contact forces between the various layers of the sandwich. Figure 11 Sample mm Resistance Bar- Brazed Coupon Produced using the Practices Defined in this Study. As suggested above, the progress nature of sample preparation resulted in some secondary cracking (the in-process braze cracking the previous one) issues that were addressed with supplemental cooling. Resulting samples are flat, show excellent braze consistency across the width, and provide a matt surface Figure 12 a Details of the Surface Morphology at the External Face of a Resistance Brazed Sandwich Panel. Note both the matt finish as well as the reflection of the underlying dimple pattern. 6

7 Some metallographic cross sections of representative joints taken from the resistance bar braze samples are provided in Figures 13 and 14. Figure 13 provides a micrograph showing the morphology of the core sheet/core sheet joint. Figure 13 Details of Zn Braze Material ReFlow at a Dimple-to-Dimple Contact in a Four-High Stack-Up Sandwich Structure following Resistance Bar Brazing Figure 14 Details of Zn Braze Material ReFlow at a Face Sheet-to-Core Sheet Contact in a Four-High Stack-Up Sandwich Structure Following Resistance Bar Brazing. The edge of the joint is shown to examine the reflow characteristics of the Zn coating. The micrograph clearly indicates the Zn has melted across the workpiece area, and pooled in various locations. Most notably, there is pooling of the Zn at the edge of the contact surface, creating a fillet in that location. Of further note, the re-solidified Zn has also cracked, compromising the integrity of the joint in this location. A similar micrograph at the face sheet/core sheet location is provided in Figure 14. This figure again shows pooling of the Zn and formation of a fillet at the edge of the weld. There is some additional evidence of cracking in this micrograph. The specific morphology seen here may be the result of either some contamination affecting flow of the Zn, or may be a polishing artifact associated with the crack itself. RESISTANCE ROLL BRAZING TRIALS Limited work was also done with the pilot resistance roll brazing system described above. As described, this system incorporated rolls of nominally 150-mm in diameter and 150- mm wide. A three high stack-up was used for these trials. This included flat face sheets and a unidirectional corrugated core, with a period of nominally 2-mm and 1.5-mm thick. Samples were prepared as individual pieces roughly 50-mm wide and 300-mm long. The corrugations of the core were oriented parallel to the axis of the rolls. Substrates were identical to those used in the bar brazing studies. As also suggested above, these preliminary trials were done with an undersized (5-kVA) AC transformer. This limited the speed at which the system could be operated. The system itself, with the individual sheets positioned to initiate brazing, is shown in Figure 15. Figure 15 Resistance Roll Brazing of 50-mm wide Hot-Dip Galvanized Steel Sheets into a One Dimensional Corrugated Sandwich Panel. 7

8 A characteristic resistance roll brazed sandwich panel made by reflow of the zinc surface coatings is shown in Figure 16. The resulting product showed effective reflow over the length of the sandwich products using indirect resistance heating. Considerable processing and hardware improvements will be required, however, to create a production capable process. Figure 16 Full width (50-mm) Three-high Sandwich Panel Resistance Roll Brazed over Several Centimeters. product. Initial examinations of these panels also indicated consistent bonding along its length. Heating also appeared to be both uniform across the width of the product as well as from one face sheet side to the other. Some details of the as roll brazed surface is shown in Figure 17. This photograph provides additional detail of apparent heating uniformity as contact is made from peak to peak on the interior corrugated core. These results demonstrate feasibility and a first level capability for creating roll brazed Figure 17 Details of a Galvanized Face Sheet on a Resistance Roll Brazed Sandwich Panel. CONCLUSIONS In this work, the basic feasibility of conducting indirect resistance-heated braze joints was investigated, and developed as a precursor to a full IRRB process. Work was conducted in a number of steps. First, the technology was demonstrated using repeated application of indirectly resistance heated bars or elements. This process was termed indirect resistance bar brazing. This approach was used for joining finite width panels. As part of the resistance bar brazing process, the need for electrically isolated heating elements was identified. Material selection for the composite (ceramic-metal-ceramic) heating element was done, and assembly processes defined. This then became an enabler to demonstration of the resistance bar brazing process. This process was used to produce preliminary technology demonstrator samples. All results were then used to develop requirements for an IRRB variant of the technology. A pilot IRRB system was assembled complete with specially designed segmented refractory metal based rolls. This system provided demonstration for continuous brazing of a candidate three high corrugated sandwich material. These studies demonstrated the feasibility and potential for continuously indirect roll brazed sandwich products. However, work here is preliminary, and will require additional development for scaling to commercially viable products. References (1) Gould, J. E. and Junde, X Development of roll-brazing assembly technology for CellTech Metals designed steel panels Task 1: Basic Process Analysis. EWI Project No CSP. Edison Welding Institute, Columbus, OH. 8