NTC Project: F01-NS14 (Formerly F01-S14) 1 NTC ANNUAL REPORT PROJECT: F01-NS14 APPLICATIONS OF MICROMACHINES IN FABRIC FORMATION

Size: px

Start display at page:

Download "NTC Project: F01-NS14 (Formerly F01-S14) 1 NTC ANNUAL REPORT PROJECT: F01-NS14 APPLICATIONS OF MICROMACHINES IN FABRIC FORMATION"

Chad Kennedy
6 years ago
Views:

1 1 NTC ANNUAL REPORT PROJECT: F01-NS14 APPLICATIONS OF MICROMACHINES IN FABRIC FORMATION Project team: Leader: Abdelfattah M. Seyam, NCSU Members: George Hodge, William Oxenham, Eddie Grant, NCSU Graduate Students: Jin Ho Lee, Vivek P. Shankam, NCSU WEBSITE: GOAL The goal of this research is to develop fundamentally new approaches for processing fibers into textile structures using MEMS technology. The specific objectives of our research are to: 1. Explore combining MEMS and small-scale robotic devices to automate the process of repairing broken warp yarns. 2. Identify new applications for MEMS in textiles. 3. Improve traditional textile manufacturing processes. ABSTRACT Microelectromechanical systems or MEMS technology has gone from an interesting academic exercise to an integral part of many applications in several industries including telecommunications, medical, automotive and ink jet industries. However, little or no work has been done in researching applications for MEMS in textiles. We elected to demonstrate the capability of combining MEMS with small-scale robotic devices to automate one of the few tasks that still need the weaving machine operator intervention; repair of broken warp yarns. Devices that can detect warp breaks using MEMS were developed and tested on the loom. Analysis of data acquired show that breaks in warp yarns can be successfully detected to a considerable extent. The study has been extended to demonstrate the application of MEMS in yarn processing and quality control systems. Microsystem solutions to measure and control dynamic yarn tension in two for one twisting mechanism are being developed. INTRODUCTION Recent advances in electronics, sensor, computing, and control technologies have led to simplified machine designs and reduction in the cost of automation. With the wide growth of MEMS (Micro-Electro-Mechanical Systems), sensing and actuating can be achieved by microsize electronics. The potential of using MEMS combined with small robotic devices to automate warp break repair in jacquard weaving is one of the tasks of this research. General accelerometers have been used as a device for monitoring vibration. Because of the small size and the low cost, sensor technology can be integrated in textile monitoring system. Regarding the current research, Gahide [3-7] attempted to find applications for MEMS in textile industry. Her concept was based on a strain-gauge sensor for warp break detection with a contact system which was placed between warp beam and stop motions of jacquard weaving machine. Bringing the sensor into contact with a warp yarn allowed monitoring tension. While

2 2 she did not use a MEMS based sensor, Gahide provided design concepts to use such sensors as a mean of monitoring individual warp yarn tension and warp stop motion. Slusser [12] built a test-bed to simulate shedding motion and mounted MEMS accelerometers on harness cords that were moved by cams. In Slusser s work, the MEMS accelerometers were used as a non-contact to warp yarns to monitor breaks. The experiments were conducted with different speeds and warp yarn tension, and the MEMS acceleration signals were acquired by data collection system. To identify a warp yarn break, the yarn was allowed to run and then cut. The difference between the signals before and after cutting was compared. It was found that the accelerometer output detected an increase in acceleration after yarn break. Slusser s work was limited to acquiring the vertical component of the accelerometer despite the availability of horizontal acceleration that could be used to better detect yarn breaks in extremely short time within one weaving cycle. The current work considered the two acceleration components of MEMS accelerometers that were mounted on harness cords of jacquard weaving machine. EXPERIMENTAL WORK Four MEMS accelerometers were mounted on selected harness cords of an ELTEX rapier weaving machine equipped with Staubli Jacquard head (Figure 1). Figure 1 Interface between the MEMS sensor and DAQ board The MEMS accelerometers were wired to a data acquisition board, which is specially designed for counter-timer. MEMS accelerometers generate pulse-width-modulation (PWM) signal, and signal is transferred into data acquisition board (PCI-6602). A program in LabVIEW was developed to convert the modulation into acceleration values. The In order to identify weaving cycles, a proximity sensor was mounted in front of the reed to detect the beat-up event that is used to identify the start of each weaving cycle. The MEMS accelerometers used are dual-axis acceleration measurement systems on a single monolithic IC. It contains a sensor and signal conditioning circuitry to implement openloop acceleration measurement architecture. For each axis, an output circuit converts the analog signal to a duty cycle (Figure 2) modulated digital signal that can be decoded with a capability of measuring both positive and negative accelerations.

3 Figure 2 Pulse Width Modulation Signal The ADXL202AE used as MEMS accelerometer, operates with supply voltages from 3.0 to 5.25 V.

The simultaneous counting PWM signal was embodied with multi-channel counter/timer board (Figure 4).

Figure 3 ADXL202AE (a) Circuit map (b) After soldering Figure 4 DAQ board for multi-channel usage To facilitate changing the

The creel was located behind the weaving machine to feed the experimental warp yarns.

3 3 Figure 2 Pulse Width Modulation Signal The ADXL202AE used as MEMS accelerometer, operates with supply voltages from 3.0 to 5.25 V. Figure 3 (a) shows the circuit designed for the accelerometer and (b) is the picture of complete configuration of ADXL circuit. The simultaneous counting PWM signal was embodied with multi-channel counter/timer board (Figure 4). It provides 16 lines of 5V digital I/O, so that four MEMS accelerometers can be used. Figure 3 ADXL202AE (a) Circuit map (b) After soldering Figure 4 DAQ board for multi-channel usage To facilitate changing the experimental warp yarn tension and type, an IZUMI creel was used. The creel was located behind the weaving machine to feed the experimental warp yarns. The creel capacity is 20 packages and the tension of the yarns fed by the creel was controlled using a servomotor. Four types of yarns were fed: 16/2 100% Cotton (74.67 Tex), 30/2 50%/50% polyester/cotton (35 Tex), 100% continuous filament polyester (124 Tex), and Spectra (69.67 Tex). Tension levels used were 1 cn/tex, 1.5 cn/tex, 2 cn/tex, and 2.5 cn/tex. Before

4 acquiring data from the accelerometers, the weaving machine was run for 5 minutes to reach to steady state.

Figure 5 shows the acceleration data acquired from the four sensors.

sensors during weaving a jacquard patttern.

4 4 acquiring data from the accelerometers, the weaving machine was run for 5 minutes to reach to steady state. Once the steady state was reached, the computer program was started to collect the acceleration data. And then warp yarns was cut manually during weaving process. Figure 5 shows the acceleration data acquired from the four sensors. Figure 5 (left) shows acceleration values collected while the sensors are stationary (weaving machine was not running) while Figure 5 (right) shows the acceleration data acquired from the four sensors during weaving a jacquard patttern. Figure 5 Front Panel of DAQ program To check whether the sensors are producing the expected signals, ideal plots for different weaves were generated using measured time-displacement data of harness cords that could be converted to time-velocity and time-acceleration plots as shown in Figure 6. Figure 7 shows ideal plot and actual data (with vibration) collected during weaving 1x3 Twill weave. As expected, the actual and predicted data are similar. We have compared the ideal and actual data for a range of weaves and concluded that the system is reasonably functioning. Figure 6 Ideal plot of plain weave (left) and 3x1 twill (right)

5 RESULTS AND DISCUSSION Figure 7 Actual data Time and Frequency Domain Low pass filter and Fast Fourier Transform (FFT) were used, which was provided in LabVIEW, to obtain filtered acceleration

The difference between the two signals provides break detection.

5 5 RESULTS AND DISCUSSION Figure 7 Actual data Time and Frequency Domain Low pass filter and Fast Fourier Transform (FFT) were used, which was provided in LabVIEW, to obtain filtered acceleration signals in X and Y directions from the MEMS accelerometers. The MEMS signals were acquired for yarns before and after intentional warp yarns breaks. The difference between the two signals provides break detection. Additionally, time series analysis was used to deduce equations for time-acceleration, time-velocity, and timedisplacement relationships. Figure 8 shows the time domain and the frequency domain of acceleration data of Plain weave with no yarn break. The black line indicates acceleration value and pink line is the proximity sensor signal. Two major peaks are shown in the frequency domain. The one with 2 Hz corresponds to the harness cord speed or the rate of its number of cycles (number of weave repeats) while the 4 Hz corresponds to the weaving speed (number of beat ups). The weaving speed corresponding to 4 Hz is 240 picks/min. The harness cord number of cycles/min is half the weaving speed since the Plain weave repeats on two picks. In yarn break analysis, MEMS sensor acceleration in X-axis (along the harness cords) did not change significantly while the acceleration in Y-axis (perpendicular to harness cord axis) changed dramatically, Figure 9. Figure 8 Time and frequency analysis of Plain weave of X-axis acceleration

6 Figure 9 Time and frequency analysis of Plain weave in Y-axis acceleration The acceleration in the Y-axis shows an increase after yarn break (intentional break was done in the 20 th second).

6 6 Figure 9 Time and frequency analysis of Plain weave in Y-axis acceleration The acceleration in the Y-axis shows an increase after yarn break (intentional break was done in the 20 th second). However, the amplitude of the acceleration in frequency domain is relatively lower than that of acceleration in X-axis. This is due to the fact that harness cord movement in the horizontal direction is limited by design (the harness cord is threaded through a small hole in the comber board). In other words, the movement of the harness cord in the horizontal direction is due to vibration. The reason why the horizontal acceleration showed significant change when a yarn breaks is due to the increase of the cord vibration. The warp yarn, which supports the cord at the contact point of warp yarn and cord, is suddenly cut leaving the long cord without support. This causes higher vibration as compared to the vibration before cutting the warp yarn. Break Detection Extensive experimental runs were performed to reveal the sensitivity of the system in detecting warp breaks. Different weaves, warp yarn types, and warp yarn tension were used. The weaves used are Plain, 2x2 R.H. Twill, 2x2 Basket, 2x2 Warp Rib, 4x4 RH Twill, 4x4 Basket, and 8H Warp Satin. The warp yarns tested were continuous filament polyester of 124 tex, continuous filament Spectra yarn of tex, cotton yarn of 16/2 (74.7 tex), and 50%/50% polyester/cotton yarn of 30/2 (39.4 tex). Tension levels used were 1 cn/tex, 1.5 cn/tex, 2 cn/tex, and 2.5 cn/tex. Figure 10 and 11 show examples of time domain acceleration in horizontal (in red) and vertical (in blue) directions with warp break using 100% polyester yarn. The data of Figures 10 and 11 confirm what has been discussed earlier. The warp yarn breaks can be better detected by the sudden variation in the horizontal acceleration of the harness cords.

7 Figure 100 4x4 RH Twill weave with warp break at 12.32 sec Figure 11 8-H Warp Satin weave with warp break at 12.

the MEMS sensor. While most of the runs of these three experimental designs were conducted, the signal analyses are being performed.

7 7 Figure 100 4x4 RH Twill weave with warp break at sec Figure 11 8-H Warp Satin weave with warp break at sec We are currently working on three experimental designs (shown in Tables 1-3) to reveal the effect of weave, warp yarn type, and warp yarn tension on the acceleration components collected from the MEMS sensor. While most of the runs of these three experimental designs were conducted, the signal analyses are being performed. This part of the project is expected to be completed by December Table 1 Experimental Design 1: Effect of weave Weave Warp Yarn type Warp Yarn Tension Plain 2x2 Warp Rib 2x2 R.H. Twill 2x2 Basket 4x4 R.H. Twill 4x4 Basket 8-H Warp Satin Polyester filament, 124 tex 1.0 cn/tex Table 2 Experimental Design 2: Effect of warp yarn type and weave Variable Weave Warp Yarn Type Warp Yarn Tension Levels Plain, 2x2 R.H. Twill, 4x4 R.H. Twill, 8-H Warp Satin 16/2 cotton (74.67 tex) 30/2 50%/50% polyester/cotton (39.4 tex) polyester filament, 124 tex Spectra, tex 1.0 cn/tex

8 8 Table 3 Experiment 3: Effect of warp yarn tension and weave Variable Weave Warp Yarn Tension Warp Yarn Levels Plain, 4x4 R.H. Twill 1.0, 1.5, 2.0, 2.5 cn/tex filament polyester, 124 Tex APPLICATION OF MEMS IN YARN PROCESSING AND QUALITY CONTROL Dynamic yarn tension monitoring and control in two for one twisting mechanism Introduction Tension control in textile processes plays and important role in quality and productivity. However, the great majority of yarns are tensioned by archaic means, resulting in lost production and inferior quality. The ideal solution for most processes is the ability to set the yarn tension to a desired level and to be assured that it does not change at any time and at any yarn position. However, the ideal tension is rarely achieved. In most cases, it almost becomes a game of trying to find out how much tension variations a process can tolerate and then adjust the process speed accordingly. With precise tension control, many processes can run at least 30% faster [11] and have the added benefit of quality improvement at the same time. There have been several articles [9, 2, 11] based on theoretical models, which discuss the effect of various factors on yarn tension variation in two for one twisters. But very few [10, 8] have addressed the possibility of measuring and controlling it practically while the yarn is being twisted. Quality control system manufacturers like TEMCO and BTSR [1, 13] have come up with smart tension scanning systems that perform online tension monitoring in various textile machines, but these systems cannot be installed on the current two for one twisters due to space restrictions. The fact that the supply yarn package is housed inside the rotating yarn balloon restrains any wired tension sensor from performing online measurement. As such, there needs to be a wireless sensing element on the machine. We propose the use of MEMS (Microelectromechanical Systems) technology with RF transmission for effectively carrying out online measurement of yarn tension variation in two for one twister mechanism. We also propose a technique to execute online control, using the measured real-time data, to acquire uniform tension in the yarn and hence the formation of a homogeneous delivery package. Design and description of the proposed sensing module Shown in Figure 12 below is the top view of the proposed design for the sensor. There are two ceramic yarn guides on either side, which guide the yarn in and out of the sensing area. The ceramic dowel between the guides is held against the yarn path, where the transduction takes place and the tension in yarn is measured. The ceramic dowel used in the design of the sensor is made from alumina pink, which is 95 to 99.7% pure alumina oxide. It has a very low thermal expansion coefficient, which is very essential in industrial environments because the response of the sensor could vary radically due to slight change in temperature.

9 (b) Yarn direction (a) Ceramic yarn guides (c) RF Transceiver Ceramic dowel Figure 12 (a) Tension sensing positions in ICBT CD350; (b) Yarn path in the sensing area; (c) Design of the proposed

Hence a full Wheatstone bridge is developed and the change in resistance developed in the strain gage is quadrupled. This enables increased sensitivity of the transducer in the sensing element.

9 9 (b) Yarn direction (a) Ceramic yarn guides (c) RF Transceiver Ceramic dowel Figure 12 (a) Tension sensing positions in ICBT CD350; (b) Yarn path in the sensing area; (c) Design of the proposed sensing device The idea here is to install four such strain gages on either sides of a ceramic dowel, which will act as a guiding element in the tension sensing device. Hence a full Wheatstone bridge is developed and the change in resistance developed in the strain gage is quadrupled. This enables increased sensitivity of the transducer in the sensing element. Semiconductor strain gages are devices that exhibit change in resistance when subjected to strain. This property makes them very useful in measuring extremely small amounts of force with accuracy and precision. Creative uses for these gages have ranged from the measurement of internal pressures inside solid rocket engines to delicate medical apparatus used in microsurgery. Gages made from semiconductor materials have advantages over more conventional types of strain gages. These include homogeneity, increased sensitivity, and decreased size. Some semiconductor strain gages range down to 0.027" (0.69 mm) in length. The ones proposed are made from Czochoralski pulled boron doped bulk silicon and bear no P/N junction. The silicon is etched to shape, eliminating the potential for molecular dislocation or cracks, thereby optimizing their performance. The resistance exhibited by a bar type strain gage, (shown in Figure 13) bearing a gage factor of 155+/-10 at 78DegF is 540+/-50 Ohms, and so an input of +5V with a strain of 400µstrain would result in an output of 100mV. Figure 13 Bar semiconductor strain gages

10 10 The sensing device of this module will comprise a ceramic dowel embedded with semiconductor strain gages and the terminals from the Wheatstone Bridge connected to an RF transceiver-cc2420. The CC2420 is a low-cost transceiver designed specifically for low-power, low-voltage RF applications in the 2.4 GHz unlicensed ISM band. Signals from the sensor would be acquired using a remote DAQ system interfaced with a host computer using an RS-232 cable. The key advantages of using microminiature sensors with RF interconnect for this application are their small size, high stability, high accuracy, less response time, and low cost. ACKNOWLEDGEMENT We would like to express our sincere thanks to Staubli Corporation for providing technical support and donating harness tie parts. REFERENCES 1. BTSR Quality control system, 2. Fraser, W. B., Air Drag and Friction in the Two-for-One Twister: Results from the Theory, Journal of Textile Institute, Vol.84, , Gahide, S., Exploration of Micromachines to Textiles: Monitoring Warp Tension and Breaks during the formation of woven fabrics, Ph.D. Thesis, North Carolina Sate University, Gahide, S., Hodge, G., Seyam, A. M., Oxenham, W., Franzon, P., Smart Sensors to Monitor Warp Tension and Breaks on a Weaving Machine, Proceedings of the TI 81st World Conference, Melbourne, Australia, April Gahide, S., Hodge, G., Seyam, A. M., Oxenham, W., Franzon, P., Micromachiens and Textiles: Matching Two Industries, Proceedings of the TI 80th World Conference, Manchester, UK, April Gahide, S., Seyam, A. M., Hodge, G., Oxenham, W., Franzon, P., Application of Micromachines to Textiles: Using Smart Sensors to Monitor Warp Tension and Breaks during Formation of Woven Fabrics, Proceedings of the International Mechanical Engineering Congress & Exposition, Orlando, Florida, November Gahide, S., Seyam, A. M., Hodge, G., Oxenham, W., Franzon, P., Micromachines and Textiles: Matching Two Industries, Textile Asia, Vol. 31, 58-56, August Grunert, J. Quality Management System on Cabling Machines, Melliand Textilberichte, pp. 880, Kothari, V. K.; Leaf, G. A. V.; The Unwinding Of Yarns From Packages Part IV and V: Two for One Twisting, Journal of Textile Institute, Vol.5, , Muller, M. The Modern and Economic Production of Tire Cord, Technische Textilien, Vol.42, , Neiderer, K., Achieving Tension Control in Yarn Processing, Quality Control Instrumentation, Slusser, T., The Application of MEMS Accelerometers for Accurately Finding Warp Yarn Breaks in Textile Machinery, M.S. Thesis, North Carolina State University, TEMCO Online quality monitoring system,

textile machinery, specifically the Jacquard Loom, in order to develop a sensor system

textile machinery, specifically the Jacquard Loom, in order to develop a sensor system Abstract Slusser, Timothy James. The Application of MEMS Accelerometers for Accurately Finding Warp Yarn Breaks in Textile Machinery (Under the direction of Dr. Edward Grant.) The textile industry, particularly