|
Abstract:
|
Currently, there are two common methods for determining the quality of yarn
exiting the false-twist texturing frame. The first method is a dyed knitted sock test,
and the second is an on-line tension monitoring system. The cost of the dyed knitted
sock test is high because it requires fabric to be knit and dyed, and needs operators to
visually compare shades. The on-line tension monitoring system does a good job of
detecting mechanical faults in the process, but recent studies show that on-line
methods fail to detect yarn defects which are related to thermal history of filaments.
Heat related defects are the main cause of poor quality yarn, resulting in
dyeability and bulk problems. Currently, there is no method of measuring heat history
at-line to quickly detect heat-related defects. The purpose of this thesis was to develop
a practical method for measuring heat history of polyester at-line using NIRA to
quickly detect heat-related defects.
For this thesis research, two different NIRA technologies were calibrated using
yarn produced at different temperatures on a high-speed commercial false-twist
texturing machine. Heat history differences were established by altering primary
heater temperatures in the texturing process in increments of 8°C. Temperatures
ranged from 404°C to 596 DC. Samples not used in the calibration were used to
validate prediction equations. Plant trials were then conducted in which doffs were
analyzed using the two NIRA technologies. It was then determined to what degree
NIRA could detect packages produced with heat-related defects. The research showed that NIR systems are sensitive to changes induced by
thermal treatments and thus, are useful tools for determining heat history of textured
PET yarn. NIR systems were able to identify heat defective packages during inplant
trials, and calibrations were improved through optimal usage of package presentation,
sampling method, and advanced mathematical modeling techniques.
|