Colour control is a resource-control system

The starting mechanism is specific to the substrate and colour chemistry. In reactive cotton dyeing, salt promotes exhaustion and alkali supports fixation; the reviewed experimental literature distinguishes these functions. IFC's textile guidance connects water use with wastewater, sludge and heating demand. These relationships justify investigating colour as a coupled production and resource decision, rather than a stand-alone dye-purchasing decision.[S31][S01]

GFT proposes that a route record begin with fibre composition, fabric construction, preparation-lot identity, customer shade reference and required end use. Treat these as conditions of recipe approval. The record should then name the machine, loading envelope, water source, chemical formulations and authorised process sequence. This is a proposed control structure. It does not imply that a recipe validated on a pure cotton substrate is approved for a blend, or that a similar commercial colour name establishes comparable behaviour.

The useful comparison is between routes that deliver the same required product. Before a trial, the process engineer should write down which conditions will be held constant and which are being changed deliberately. A lower dye charge is only a candidate improvement until the resulting shade, fastness and accepted output are checked. Similarly, the purchasing team should retain delivered formulation concentration and dilution instructions, so an apparently cheaper kilogram does not conceal a change in the active material supplied.

Make the water used to prepare the recipe part of that record. The hard-water study found that hardness and dye concentration influenced its right-first-time result. GFT therefore proposes recording water-quality checks beside dosing and temperature histories, rather than interpreting a shade deviation through dye quantity alone. The study's operating conditions remain study-specific; they are not setpoints for a proposed Cameroon production line.[S37]

A recipe should have distinct target, permitted range and stop conditions. The target describes the intended process. The permitted range is supported by product trials. The stop condition identifies a departure requiring authorisation, such as unavailable source-water information or an unapproved formulation substitution. Leave those ranges unset until validation exists. Filling an empty engineering field with a generic industry value would create an instruction without evidence.

The shift operator needs a compact instruction derived from the approved master recipe. The laboratory needs the full analytical history. Purchasing needs an approved-product reference. Management needs the resource consequence of deviations. GFT recommends connecting these views to the same recipe version, while preserving their different purposes. A revised dosing sequence should not silently inherit approval from the previous sequence merely because the product code remains unchanged.

Use an exception record for any substitution made under supply pressure. State what changed, which batches received it, who approved the trial and which release evidence is required. Keep unaffected batches separately identifiable. If the substitution fails, the record should support containment without treating all production in the reporting period as chemically equivalent. The resource account can then distinguish the cost of a tested improvement from the cost of an uncontrolled change.

This chapter stops at the interface with substrate preparation and machine selection. It requires their lot and route identities because colour interpretation depends on them. Detailed preparation design and low-liquor equipment comparison belong in the following chapters. Here, the practical question is whether the colour decision retains enough information to identify what was consumed, what was accepted and what must be investigated when those outcomes diverge.

Before running a new colour family, GFT recommends a short readiness review involving production, quality, utilities and chemical management. Confirm that the route instruction matches the laboratory approval, required input checks are available and the metering boundary can capture any correction. If a required measurement cannot be taken, record the resulting limitation before the trial starts. That prevents the team from discovering after production that its most important comparison depends on an unrecorded assumption. Keep the review proportional to the change, but make its decision recoverable from the batch record.

Exhaustion describes transfer out of the bath; fixation concerns the dye retained after unfixed material is removed under the analytical method. The reactive-dye study measures both, using bath analysis and a subsequent extraction procedure. GFT recommends preserving the measurement basis with each result. In particular, a fixation percentage expressed relative to exhausted dye must not be presented as the same percentage of dye originally charged.[S31]

For the proposed ledger, define charged dye mass as D, dye exhausted onto fabric as U and analytically retained dye as F. Record exhaustion as U divided by D, fixation relative to uptake as F divided by U, and retained share of charge as F divided by D. Express each as a percentage only after naming its denominator. These are GFT accounting definitions for a trial protocol; the laboratory must reconcile them to its validated analytical method before reporting production results.

The illustrative flow in Exhibit 15 preserves those mass boundaries. Its fabric branch represents retained dye, not a disposal credit. Its liquid branches are streams requiring measurement and management, not proof that treatment removes the dye or renders the discharge acceptable. A real investigation would also need retained liquor, sampling losses and any other material paths before calling the balance closed.

For a production trial, GFT proposes sampling the charge, residual bath and wash streams through a laboratory-approved protocol. Record sample time, dilution, bath volume, instrument calibration and analytical uncertainty. Concentration alone cannot close a mass account if volume changes during dosing, drainage or replenishment. Preserve the measurements needed to calculate concentration multiplied by volume at the relevant boundary, together with the uncertainty introduced by unmeasured carryover.

Mixtures need particular care in the proposed protocol. Ask the laboratory whether its method distinguishes the components sufficiently for the decision being made. If it cannot, report the aggregate result with that limitation. Do not allocate a blended residual among individual dyes by guesswork. A spectral signal should not be relabelled as active dye mass unless the calibration and matrix support the conversion.

GFT also recommends recording results by dye identity and recipe, rather than assigning a universal fixation factor to the dyehouse. A comparison of alternative recipes should retain substrate, shade, water and analytical conditions. Where the evidence cannot separate recipe effects from measurement variation, the appropriate outcome is another controlled trial or an unresolved result. Publishing a precise-looking average would not resolve that uncertainty.

Finally, retain the balance residual as an explicit field. Ask whether it is consistent with stated measurement uncertainty before investigating its cause. A non-closing balance is a reason to inspect volumes, sample labels, dilution calculations and material paths. It is not permission to force the retained-dye estimate to the value required by the business case. This prevents an accounting adjustment from masquerading as chemical performance.

Analytical sampling should also have a disposition rule. GFT proposes labelling specimens with the parent lot and process stage, retaining the chain of custody and specifying how repeat analysis will be handled. If a specimen is contaminated or its identity becomes uncertain, replace it through the agreed sampling procedure and preserve the invalidated result with the reason. Do not average a disputed result into the accepted dataset. The same rule should apply to an inconvenient low result and an apparently favourable high result, so the measurement process does not depend on the desired conclusion.