GSM and Tensile Strength Comparison of Woven Fabrics from a Powerloom Cluster

GSM and Tensile Strength Comparison of Woven Fabrics from a Powerloom Cluster is a standards-based testing study that buys fabric the way a buying house does — from working powerloom sheds — and then subjects it to the conditioned laboratory tests an export inspection would demand, connecting loom-shed construction data to measured GSM and strip tensile strength.

India's decentralised powerloom sector produces most of its woven yardage; clusters such as Ichalkaranji in Maharashtra run tens of thousands of looms supplying shirting, suiting and home-textile fabric to domestic and export markets. Buyers, however, complain of inconsistent quality: fabric ordered at a target GSM arrives lighter, and strength varies between sheds running the same construction. For a textile student, this is a measurable, locally sourced research question — not a textbook exercise.

The team samples four woven constructions (100% cotton plain weave, cotton–polyester blend, polyester suiting and cotton poplin) from three sheds each in the cluster, records ends-per-inch and picks-per-inch with a counting glass, and conditions every specimen at the tropical standard atmosphere before testing. GSM is measured per ASTM D3776 with a 100 cm² round cutter on a digital balance, and tensile strength is measured by the strip method to ISO 13934-1 on the department's CRE tester, in both warp and weft directions with five specimens each. One-way ANOVA and correlation analysis in Python then answer the study's two questions: how much GSM and strength vary between sheds for the same nominal construction, and how strongly measured GSM predicts measured strength.

Powerloom clusters supply most of India's woven fabric, but quality control at shed level is informal: fabric is sold against a nominal GSM and construction, yet weight, ends/picks density and strength vary from shed to shed and batch to batch. Buyers respond with rejections and discounts, sheds lose margin without understanding which parameter drifted, and college teaching never demonstrates the full chain from construction data to conditioned standard tests to statistics.

There is a need for a systematic, standards-based comparative study — sampling the same nominal woven constructions from multiple sheds of a real cluster, measuring construction, GSM and strip tensile strength under proper conditioning, and analysing between-shed variation and the GSM–strength relationship statistically — so that the variation buyers complain about is quantified, explained at the loom level, and documented in a protocol future batches and sheds can reuse.

Abstract

This project measures grammage (GSM), fabric construction and strip tensile strength of four woven fabrics sampled from multiple powerloom sheds in a cluster, under standard conditioning, and analyses variation between sheds and the GSM–strength relationship. The deliverables are a reproducible test protocol, a per-shed quality dataset, and an interpretation connecting construction parameters to measured performance.

Introduction

Woven fabric quality begins at the loom: ends and picks per inch, yarn counts and weave decide the fabric's mass and its load-bearing behaviour. GSM (grams per square metre) is the trade's first quality gate — buyers order fabric by it — and tensile strength decides whether the fabric survives processing and service. Both are standardised laboratory measurements (ASTM D3776 for mass, ISO 13934-1 for strip tensile), but their behaviour across a decentralised powerloom cluster, where many small sheds weave the same nominal construction on different looms, is exactly the kind of variation a quality professional must understand.

Existing System and Its Limits

• Trade practice in small clusters checks quality by feel, drape and a cut-strip weigh on a shop balance — unconditioned, unsystematic and undisputed only because nobody records it.
• Export buyers test in accredited labs, but those reports are batch-level and never shared back with the sheds, so loom-side learning is lost.
• College fabric-testing labs teach each test in isolation; students rarely see a dataset — multiple sheds, repeat specimens, statistical comparison — or handle the question buyers actually ask: is variation between sheds real, and does GSM predict strength?
• Limitation: no accessible study links cluster-level sampling with standard-conditioned testing and statistical analysis on a student budget.

Proposed System

• Cluster sampling: four nominal constructions (cotton plain, poly–cotton blend, polyester suiting, cotton poplin), three sheds per construction, two metres per shed, purchased openly from working sheds in the cluster with shed identity recorded.
• Construction analysis: ends/inch and picks/inch with a counting glass, weave identification, yarn count examination, and thickness at defined pressure.
• Standard testing: conditioning of all specimens at 27 ± 2 °C and 65 ± 4 % RH until equilibrium; GSM by ASTM D3776's small-swatch option using a 100 cm² round cutter; strip tensile strength and elongation per ISO 13934-1 on a CRE machine, five warp and five weft specimens per fabric.
• Statistical analysis: one-way ANOVA across sheds within each construction, coefficient-of-variation reporting, and regression of tensile strength on measured GSM, computed in Python with the plots the report needs.

Feasibility

• Technical: the department's textile-testing lab already owns the CRE tester, cutter, counting glass and balance; conditioning uses the lab's conditioning chamber or a sealed desiccator-equilibrium protocol if it does not.
• Economic: fabric samples and consumables total roughly ₹4,000–8,000 including cluster travel, well inside a project grant.
• Operational: sampling and construction analysis complete in the first month; testing and analysis fill the remaining weeks comfortably inside the 14-week two-phase window.

In scope

• Four constructions × three sheds, purchased from working powerloom sheds, with identity and loom details recorded.
• Construction analysis (EPI, PPI, weave, thickness), GSM and strip tensile testing under standard conditioning, with defined repeat-specimen rules.
• Statistical comparison across sheds and GSM–strength regression; practical interpretation for buyers and sheds.

Out of scope

• Colour fastness, abrasion, pilling and drape testing.
• Yarn-level testing (single-end strength, count variance) beyond qualitative examination.
• Loom modification trials or recommendations executed at a shed — the study measures and explains, it does not intervene.
• Knitted and nonwoven fabrics.

The study follows a sample → condition → measure → analyse research design mapped onto the two VTU project phases.

Testing approach: instrument control (balance calibration with a check weight, tester load-cell verification, cutter condition), procedural control (ravelled-strip preparation, jaw-break avoidance rule, gauge length and rate per the standard), and data control (every specimen's value entered on the day of test; any specimen failing at the jaws replaced per the standard's retest rule and logged). The CV of each five-specimen set is reported alongside its mean, so the ANOVA compares like with like.

The study is organised as a four-stage pipeline — sampling, conditioning, measurement, statistics — drawn below; each stage's datasheet feeds the next, and the statistical stage never sees a number that lacks its sampling and conditioning record.

flowchart TD
  CL{"Sampling plan: 4 constructions x 3 sheds"} -->|"2 m per shed, identity recorded"| REG["Sample register"]
  REG --> CON["Construction analysis: EPI, PPI, weave, thickness"]
  REG --> CND["Conditioning: 27 +/- 2 C, 65 +/- 4 percent RH to equilibrium"]
  CND --> GSM["GSM per ASTM D3776: 100 sq cm cutter + balance"]
  CND --> TEN["Strip tensile per ISO 13934-1: 5 warp + 5 weft per fabric"]
  CON --> DB[("Combined dataset")]
  GSM --> DB
  TEN --> DB
  DB --> ANO{"One-way ANOVA across sheds per construction"}
  DB --> REG2{"Regression: strength vs GSM"}
  ANO --> RPT["Variation findings + CV tables"]
  REG2 --> RPT2["Prediction quality per construction"]
  RPT --> FIN["Phase-II report and viva"]
  RPT2 --> FIN

Prose description. The pipeline is deliberately front-loaded: the sampling register is the study's backbone, because every later number is only as good as the shed and loom details recorded at purchase. Conditioning is a separate stage, not a footnote — tensile results on unconditioned fabric are simply not comparable, and the report says so. The two statistical branches answer the study's twin questions independently, so a weak GSM–strength correlation in one construction does not blur the between-shed findings of another.

• Wash-and-care extension: shrinkage and strength retention after repeated laundering cycles, connecting the study to care-labelling decisions.
• Yarn-level bridge: add single-yarn strength and count CV tests so fabric strength differences can be traced to yarn quality at specific sheds.
• Cost–quality mapping: combine shed pricing with measured quality into a value index buyers could actually use for sourcing decisions.
• Faster inspection research: evaluate whether small GSM sample counts can statistically substitute for full tensile testing in routine trade inspection.
• Repeat-season study: re-run the register next weaving season to see whether between-shed variation is systematic or batch-specific.