Textile machinery: yarn and warp tension, dye uptake by trial

Yarn tension and yarn breaks in spinning

What happens. On spinning machines the tension of the yarn changes with speed, bobbin diameter, material and the day’s climate. Where it is not measured at every position, breaks show up first as stops and rejects, and the operators patch the setting by position.

What it costs. Every break stops a position or a whole machine. The costs are downtime, wasted yarn and staff walking between positions to restart them, and they limit the speed at which the machine can be sold.

Why the usual fix fails. A tension sensor at every position is expensive and itself adds a point of failure. Lowering the speed reduces breaks but gives away the output the customer paid for.

How a fixed-price remote diagnosis checks it. From machine logs (speeds, drive currents, break events, bobbin data) I look at whether the tension can be estimated from signals the machine already has. The fixed-price remote diagnosis tells you whether such an estimate can guide the drives, and what a pilot on one machine type would need.

Background: Computing unmeasured states for control

Warp tension on the loom

What happens. On a loom the tension of the warp threads changes as the beam empties, with the shed movement and with the yarn. Setting it is a matter of the weaver’s experience for each article, and a new fabric means a new round of trial and error.

What it costs. Uneven tension gives streaks, thread breaks and fabric defects that show up only in inspection. The result is second-grade goods, stops and long set-up times for new articles.

Why the usual fix fails. A fixed tension setting per article works for the one run it was found on. Adding more sensors and hand-set offsets makes the loom harder to hand over to the next shift or the next customer.

How a fixed-price remote diagnosis checks it. From loom data (beam speed, drive torques, tension readings where they exist, break and stop events) I check whether a model of the warp can hold the tension steady through the whole beam. The fixed-price remote diagnosis answers whether model-based control would help on your looms, and what to measure on a test machine.

Background: Control that adapts to drift and wear

Fabric moisture and temperature in the stenter

What happens. In the stenter (the frame that dries and sets the fabric) nobody sees the moisture and temperature of the cloth itself. The chamber air is measured, so operators set it for the worst fabric and slow the line when a result looks doubtful.

What it costs. Gas for hot air that the fabric did not need, and uneven moisture or shrinkage that has to be corrected later. Both are paid for by every metre that passes through.

Why the usual fix fails. A moisture sensor at the exit measures one point after the damage is done. Running the line slower or hotter by hand works for one article and has to be found again for the next.

How a fixed-price remote diagnosis checks it. From line logs (chamber temperatures, exhaust humidity, speed, fabric weight and the exit values that exist) I estimate what the cloth experiences along the chambers. The fixed-price remote diagnosis tells you whether a computed fabric state can steer the temperatures and speed, and whether this is worth building into the machine.

Background: Predictive control of process plants

Dye uptake is found by trial

What happens. How much dye a batch of fabric takes up depends on the fibre, the liquor, the temperature curve and the lot. In practice, the colourist runs a trial and corrects, until the shade matches the sample.

What it costs. Rework and re-dyeing, extra dye and water, and long lead times for each new colour. A deviation that shows up only after the batch is finished cannot be corrected without another pass.

Why the usual fix fails. Dosing more slowly or in more steps raises the chance of a hit but extends the cycle. A recipe book works as long as the fibre and the liquor stay exactly as in the earlier batch.

How a fixed-price remote diagnosis checks it. From dyeing logs (temperature curves, dosing, liquor levels and colour measurements after the batch) I look at whether the uptake can be estimated during the cycle. The fixed-price remote diagnosis shows whether dosing could be steered on that estimate, and what a machine builder would need to measure in its test hall.

Background: Computing unmeasured states for control