Textile Testing & Quality Control

Measuring and assuring textile quality. Learn sampling and conditioning, fibre testing (HVI, AFIS), yarn testing (count, strength, CSP, evenness, hairiness), fabric testing (tensile, tear, bursting, abrasion, pilling, dimensional stability, air permeability), colour measurement (CIELAB, ΔE) and colour fastness (wash, light, rubbing — grey scale ratings), statistical quality control (mean, CV%, confidence intervals, control charts), and standards (ISO, ASTM, BIS, AATCC) — with Python for strength statistics, CSP and colour difference.

Textile materials are inherently variable — natural fibres differ between bales, yarns vary along their length. Testing quantifies properties and variability so products meet specifications and customers' expectations.

Quality control combines standard test methods, statistical analysis and process control to prevent defects rather than just detect them.

An analogy

Testing a yarn is like checking a chain: the average link strength matters, but the weakest links (variability) decide when it breaks — which is why textile QC focuses so much on CV%.

Why and how we test

Tests check raw materials, control processes and verify final products. Correct sampling (representative, adequate size) and conditioning (20 ± 2 °C, 65 ± 4% RH for 24 h) are essential because moisture affects mass, strength and dimensions. Standards (ISO, ASTM, AATCC, BIS) define procedures so results are comparable.

Yarn and fabric tests

Common tests:

Colour and fastness

Spectrophotometers measure reflectance → CIELAB coordinates; ΔE (or CMC/CIEDE2000 which weight differences like human vision) quantifies shade difference. Colour fastness to washing, rubbing (crocking), light, perspiration and sublimation is rated with grey scales: 5 = no change, 1 = severe. Metamerism: two samples match under one light but not another.

Statistical quality control

Control charts (x̄–R, p, c charts) detect process shifts in counts, GSM or defects; process capability (C_pk) compares variation with specification limits. Online monitoring (sliver/yarn evenness sensors, automatic fabric inspection with machine vision) enables real-time control. Six Sigma and root-cause analysis reduce defects and rework.

Formulas

Worked example: CV%

Given: Mean 290.4 cN, SD 10.84

Find: CV.

Answer: 3.73%

Worked example: ΔE

Given: ΔL −0.7, Δa +0.9, Δb −0.6

Find: ΔE*ab.

Answer: 1.29

Worked example: Shrinkage

Given: Mark 500 mm → 482 mm after wash

Find: S%.

Answer: 3.6%

Strength statistics, CSP and colour difference

import statistics as st
# Quality control: yarn strength CV%, confidence interval and CSP
strength = [285, 298, 276, 310, 290, 281, 302, 295, 288, 279]    # single-yarn strength, cN
mean = st.mean(strength)
sd = st.stdev(strength)
cv = sd / mean * 100
print(f"mean {mean:.1f} cN, SD {sd:.2f}, CV {cv:.2f}%")
t = 2.262                                 # t(0.975, 9)
half = t * sd / len(strength) ** 0.5
print(f"95% confidence interval {mean-half:.1f} - {mean+half:.1f} cN")
tex = 19.7
print(f"tenacity {mean/tex:.2f} cN/tex")
# Lea CSP: count * lea strength (lb)
count, lea_lb = 30, 90
print(f"CSP = {count*lea_lb}")
# Colour difference (CIELAB)
L1, a1, b1 = 52.3, 38.1, 12.4
L2, a2, b2 = 51.6, 39.0, 11.8
dE = ((L1-L2)**2 + (a1-a2)**2 + (b1-b2)**2) ** 0.5
print(f"colour difference dE*ab = {dE:.2f} ({'accept' if dE < 1.0 else 'reject'} at tolerance 1.0)")
# mean 290.4 cN, SD 10.84, CV 3.73%
# 95% confidence interval 282.6 - 298.2 cN
# tenacity 14.74 cN/tex
# CSP = 2700
# colour difference dE*ab = 1.29 (reject at tolerance 1.0)

Only 10 tests give a mean with a ±7.8 cN uncertainty — standards specify minimum test numbers for this reason. Real single-yarn tests use hundreds of breaks (automatic testers); CV% is often more important than the mean. The colour batch is ΔE 1.29 from standard — visible to a trained eye and rejected at a 1.0 tolerance (many brands use CMC ΔE instead).

Common mistakes

Applications

Key takeaways

Going further