Paper to the 11th Annual Meeting of the Polymer Processing Society, Seoul, Korea, paper 6-02, 27th-30th March 1995.

S F Bush with C G A Clayton

Abstract

Synthetic fibre production is characterized by the fact that while control is actually exerted in the factories, in many cases we can only know if the quality required by the market has been achieved some weeks later or even longer, after several additional operations such as knitting and dyeing have been performed by customers. In some cases, the operation of polymerization, spin-drawing, draw-texturing and weaving or knitting, and dyeing are all performed by separate manufacturers.

While defects may show up at any of these stages, the most difficult defects to counter are those which show up after dyeing, since the appearance of fabrics depends on quite subtle variations in the yarn polymer morphology and on the bulking geometry. Any of the preceding stages can contribute to observed variations in appearance. The problem of multiple stages is compounded by the requirement at fabric assembly to be able to use bobbins made at different times and from a sequence of different positions at each of the preceding stages.

In an earlier paper (Ref 1) the authors described the principles of a Process Analysis-Monitoring System (PAMS) which uses the principles of variance analysis extended on a huge scale to rank and locate the sources of non-uniformity in the factories and on the machines. The present paper builds on this by developing mechanistic relationships between the primary determinants of appearance (bulking, dye uptake and/or pigment dispersion) and the principal process variables such as plate temperatures, windup speeds, bulking speeds, through the relationships of both classes of variables to the fibre morphology.

Broadly speaking the PAMS variance analysis will indicate where in the various manufacturing stages, the sources of variability are likely to be. Combining the models of fibre behaviour with the PAMS allows this fuzzy diagnosis to be made more precise and in some cases determines the precise variable and its position causing a given type of variability. By this means an intelligent process system has been built up which is capable of steady refinement in the light of plant results.

References

[1] S F Bush and C G A Clayton, Analysis and Control of Variability in the Fibre-making Process, 7th Ann Mtg Poly Proc Soc, Hamilton, Ontario, Canada, 11-14 April (1991)

Prosyma Research Ltd report to CaMac Corporation

S F Bush

Introduction

Man-made fiber production is characterized by the fact that while control is actually exerted at the factory level, we can only know if the quality required at the market level has been achieved some weeks later or even longer after several additional operations on the fibers have been performed by our customers. Whereas in the CaMac PET case the texturing for critical end-uses is done by another company, there is potentially a major problem in determining how much that company may have contributed to any problem showing up at the ultimate customer, be they knitter or weaver.

Naturally the intensity and frequency of any quality problems will depend on how critical the end-users are. Generally the most difficult markets to serve are those requiring the piece dyeing of natural fibre knitted or woven from many diffeeent ends. Linings, underwear and nightwear are typical markets for these processes. CaMac’s melt-pigmented technology may well have an advantage in markets held by dyed natural yarn because dyeing is (a) very sensitive to non-uniformity in the polymer morphology, and (b) is an additional relatively expensive operation to be performed. On the other hand of course natural fiber gives the customer the greatest freedom in use, and the dyeing cost is partially offset by the fact that the natural yarn is not carrying a pigment cost.

Polymer morphology refers to the size and shape of crystals within the yarn as well as the arrangement of those parts of the polymer chains which are not in crystals. A fundamental point is that different process variables affect the components of the morphology in different proportions so that there is no direct correlation between a test at say the spinning stage and the appearance of the yarn post-texturing. Moreover the texturing operation not only gives bulk to the flat yarn, it is also responsible for much, if not all of the crystallinity, which in turn affects the bulk achieved.

Thus the bulk achieved is a function not only of the texturing proces,s but also of the morphology of the flat yarn fed to it. This in turn is affected by the spinning operations, the drying and the chemical composition of the polymer. We thus have a series of complex interactions extending potentially over several factory locations.

For a given market the ideal is to be able to put the product from any spinning position on any texturing position and mix (merge) the product with that from any other combination of positions. The practical control task is thus essentially to be able to detect positions and occurrences (e.g. a batch of over-moist polymer) as rapidly as possible in order (a) to segregate the affected product, so as to limit the extent of contamination of good product*; (b) take corrective action in the plant.

In “Proposals for Development of Technology and Capability at CaMac Corporation” (Ref 1) section 3 outlines the basis of a process analysis and monitoring system (PAMS) which has been developed for the control of multistream plant like CaMac’s fiber operations. This system rests on four elements:

1. A map of material flow paths
2. Identification of timescales of changes along the flow paths
3. Classification sources of variability in yarn properties into:
   1. Inherent molecular processes (e.g. hydrolysis)
   2. Equipment functioning (e.g. spin pack behaviour)

    

4. Distinguishing these variables which we can actually manipulate (e.g. quench air flow rates) from those we wish to control (e.g. yarn appearance).

This report shows how a PAMS can be gradually built up for PET from what is done on the plant now. Two particular ingredients of PAMS described below are:

1. Analysis of Variability – to find whereabouts on the material’s flow paths the main sources of variability are likely to be
2. Systems Model – to locate particular sources of variability (e.g. excess moisture, mechanical wear on traverse arms).

An important application of both the Analysis of Variability (AV) and the Systems Model (SM) could be to texturing where critical outlets are involved. This would naturally require the cooperation of the texturing house.

References

* What is meant here is that poor material knitted or woven with good, will cause the whole fabric to be discarded and give rise to compensation claims.

Ref 1: S F Bush, Proposals for Development of Technology and Capability at CaMac Corporation, 16 December 1993.