Diolin High Speed Video Trial

Transcription

1 Diolin High Speed Video Trial To: Randy Vaughan, Danny Morgan, David Frost, David Israel, and Craig Rollins From: Peter Tkacik, 20 November 2004 Test Conditions: The testing was done on the textile winders at Diolin in Scottsboro, AL on Tuesday, the 16 th of November The trial was focused on 94 mm tubes yarn transfers on the #43 and #44 winders. The #43 machine was operating at 3600 m/m and the #42 machine was running at 5500m/m. I ran the high-speed video equipment and analyzed the results. The fiber under consideration was a 1500 denier yarn. Mirroring tests were also performed but are not analyzed in this report. The primary issues have been with the transfers. Note: As the frame speed goes up, the high speed video transfer rate becomes limited in vertical resolution so many of the sequences are rotated sideways for improved resolution in the slit direction.

2 Objective: The purpose of this report is to review the results of the high-speed video trials run at Diolin, Scottsboro, AL. The analysis revealed that the transfer position accuracy of the #43 machine traverse mechanism has some mechanical variability and did not position the fiber well during the one recorded dropped transfer. The analysis also revealed a slight improvement with the tubes manufactured on November 15 vs. August 23 on the #42 machine. The analysis also revealed that the #43 machine transfers were much faster with the Red tubes 15 vs. the Orange tubes. Test #1 Machine #43, Smooth Transfer: Test #1 was run at 2000 fps (frames per second) on the outboard position of machine #43 and did not reveal any abnormalities. Test #2 Machine #43, Dropped Transfer: Test #2 was also on the outboard position of machine #43 and was the only test of the day in which a transfer was dropped. In fact, it was only marginally focused and would have been deleted but for the missed transfer. The camera was rotated sideways for increased resolution in the slit direction. To help visualize in review of the sequence with the Midas software, (using the standard pull-down menus), I used the Image Processing then Image Settings of: Brightness = 27 Contrast = 17 Filtering = smooth When reviewing image sequence, the primary observation is that the yarn is traversing over to the slit and stops short, (see figures 1 &2). I measured the position as it moved over to the slit and plotted it in Chart #1.

3 Details of the fiber traverse location: When comparing test #2 to the later tests (#8, and #9), the yarn is brought up to the tube surface early and then moved laterally across the counter-rotating surface. The yarn then holds in the 72 pixel (vertical) in sliding contact just at the edge of the slit for several rotations and then drops into the slit at frame It is immediately snagged by the slit but the motion jerks the yarn and it wavers back and forth between the slit and traverse positions until frame 1583, (see figures 3 through 6). The wavering of the yarn between the slit position and the guiding position of the traverse repeats three times until the yarn is snagged and then broken. Lateral position in pixels (mm) location at transfer frame number (time) Traverse stops short of the slit Chart 1. Yarn position delay next to slit on test #2. Shadow Figure 1. Frame 1700, The fiber is shown prior to string up, (down is to the right).

4 Figure 2. Frame 1606, The fiber is shown at the edge of string up. Figure 3. Frame 1600, The fiber drops into the slit. wavering Figure 4. Frame 1586, The yarn fibers wavering in and out of the slit. Yarn fibers pulling backwards Figure 5. Frame 1580, The fibers are being pulled backwards. Broken Figure 6. Frame 1275, The fiber bundle broke and let loose of the tube.

5 Test #3 Machine #43, Smooth Transfer: Test #3 was run at 2000 fps on the outboard position and did not reveal any abnormalities. Test #4 Machine #42, 8000 frames per second: Test #4 was run at 8000 fps on machine #42 and mostly due to the limited field of view at this speed, did not reveal anything. Tests #5, #6, & #7, Machine #42; August 23 tubes vs. November 15 tubes: Note: sequences were analyzed at a brightness of 38 and a contrast of 21 Test #5 was run at 2000 fps on machine #42 and was a tube manufactured on August 23. Test #6 was also run at 2000 fps on machine #42 and was a tube manufactured on August 23. Test #7 was also run at 2000 fps on machine #42 and was a tube manufactured on November 15. There was a slight delay in the capture of the fiber going into the slit on the tubes manufactured August 23. I recorded the frame at which the fiber dropped into the slit and then also recorded the frame number at which the fiber snags in the slit. I also recorded the frame at which the fibers broke and the slit reversed their direction and pulled them backwards out of view, (table 1 and figs. 7, 8, 9, & 10). Image frame number drops is is Total Time into slit snagged Gone (msec) Test #5, August msec Test #6, August msec Test #7, November msec Table 1. Time delay for the slit to snag the fiber (in milliseconds).

6 The average time to snag for the 8-23 tubes was 3.5 msec whereas the time for the tube was 2.5 msec. I consider this to be marginal and not very statistically significant. Shadow Figure 7. Frame 1319, The fiber is shown prior to string up, (down is down). Figure 8. Frame 1315, The fiber is shown dropping into the slit. snagging in the slit

7 Figure 9. Frame 1311, The fiber is shown snagging in the slit. Yarn fibers pulling backwards Shadows Figure 10. Frame 1308, The yarn fibers are being yanked up out of view. Broken Tests #8, & #9 Machine #42, Red tubes vs. Orange tubes: Note: sequences were analyzed at a brightness of 44 and a contrast of 30 Test #8 was run at 2000 fps on machine #43 and was a Red tube. Test #9 was also run at 2000 fps on machine #42 and was an Orange tube. Image frame number drops is is Total Time into slit snagged Gone (msec) Test #8, Orange tube msec Test #9, Red tube msec Table 1. Time delay for the slit to snag the fiber (in milliseconds). A noticeable delay was measured in the capture of the fiber going into the slit on the Orange tubes vs. the Red tubes. As with tests 5, 6, &7, I manually recorded the frame at which the fiber dropped into the slit and then also recorded the frame number at which the fiber snags in the slit, (see table 1 and figures 11, 12,

8 & 13). Finally, I recorded the frame at which the fibers broke and the slit reversed their direction and pulled them backwards out of view. The time to snag for the Red tube was 3.0 msec whereas the time for the Orange tube was 18.5 msec. Shadow Figure 11. Frame 1319, The fiber is shown prior to string up, (machine #43, down is to the right). Figure 12. Frame 1315, The fiber is shown dropping into the slit. snagging in the slit Figure 13. Frame 1311, The fiber is shown snagging in the slit. Comments regarding the tests 8, and 9 fiber traverse positioning: In reference to the earlier comments on the variability of the traverse mechanism in locating the yarn, I repeated the measurements on tests 8, and 9 and plotted the yarn position versus the slit just prior to the snagging, (see charts 2, and 3). Particularly with test #8, the yarn is brought up smoothly to the slit and then drops in. In test #9, the yarn is brought up to the tube but wavers off of the surface and then is brought up close to the slit and drop in.

9 location at transfer Lateral position in pixels (mm) frame number (time) Chart 2. Yarn position delay next to slit on test #8. location at transfer Lateral position in pixels (mm) frame number (time) Chart 3. Yarn position delay next to slit on test #9. Summary: We reviewed the winders at Diolin for issues with transfers using high-speed video technology. Included were sequences of the Caraustar tube on machines #42 and #43. I later reviewed the sequences and the analysis revealed that the transfer position accuracy of the #43 machine traverse mechanism has variability and did not position the fiber well during the one recorded dropped transfer. It

10 also revealed that the #42 machine transfers showed a small improvement with the tubes manufactured on November 15 vs. August 23. This was even though the August 23 tubes do not statistically demonstrate a serious problem during transfer. It also revealed that the #43 machine transfers showed a sizeable improvement with the Red tubes 15 vs. the Orange tubes. Suggestion: It would be prudent to review the cause of the #43 machine s traverse alignment variability and repair/adjust it. With regards to the tube, some of the machine wear can be negated by changing the narrowing the slit and possibly, adding a score. We made a recommendation that we make tubes with these changes and run new trials. The result of these trials will supply information to better define the tolerance needs of the winders in their current condition. Note: Exposure time for most all sequences is 250 microseconds, and all images were digitally enhanced in both brightness and contrast for clarity.