Big Data Analytics for Print Optimization in ShirongMaterials

Big Data Analytics for Print Optimization in ShirongMaterials

Lead

ΔE2000 P95 moved from 2.6 to 1.7 and registration RMS from 0.22 mm to 0.11 mm at 160–170 m/min, lifting FPY to 97.4% and delivering a 5.8-month payback.

Value: Before→After at 165 m/min, UV-LED 1.3–1.5 J/cm², dryer 75–85 °C, N=126 lots in 8 weeks; Sample: mixed runs on SBS 300 g/m² and kraft cupstock.

Method: 1) Centerlining press/ink/substrate; 2) UV-LED dose tuning with closed-loop ΔE control; 3) SMED parallel prep for plates/anilox.

Evidence anchors: ΔE2000 P95 −0.9 (2.6→1.7, ISO 12647-2 §5.3), registration RMS −0.11 mm; G7 report ID G7-VAL-2025-0412; SAT/IQ/OQ/PQ: SAT-23-118, IQ-PRN-017, OQ-PRN-017, PQ-LT-009.

Metric	Before	After	Conditions	Records/Clauses
ΔE2000 P95	2.6	1.7	165 m/min; [InkSystem]=UV-LED low-migration; [Substrate]=SBS 300 g/m²	ISO 12647-2 §5.3; G7-VAL-2025-0412
Registration RMS	0.22 mm	0.11 mm	165 m/min; dryer 80 °C; dwell 0.9 s	IQ-PRN-017 / OQ-PRN-017
FPY (P95)	93.2%	97.4%	N=126 lots; mixed SKUs	PQ-LT-009
kWh/pack	0.0074	0.0067	UV-LED 1.3–1.5 J/cm²	EU 2023/2006 §6 documentation

G7/Fogra PSD Conformance Play

Outcome-first: We reached G7 Colorspace and Fogra PSD Level B control in two weeks, holding ΔE2000 P95 ≤1.8 at 160–170 m/min and registration ≤0.15 mm.

Data: ΔE2000 P95=1.7 (CIELAB D50/M1), registration RMS=0.11 mm; Units/min=420–450; kWh/pack=0.0067; [InkSystem]=UV-LED low-migration CMYK+W; [Substrate]=SBS 300 g/m² and clay-coated kraft 250 g/m²; N=44 validation lots.

Clause/Record: ISO 12647-2 §5.3 tone/value aim; Fogra PSD 2016 §5 process control; G7 report G7-VAL-2025-0412; IQ/OQ/OQ-PRN-017; PQ-LT-009.

Steps:

• Process tuning: Set ΔE2000 target ≤1.8; tune anilox 400–500 LPI/3.0–3.6 cm³/m²; lock LED dose 1.3–1.5 J/cm².
• Process governance: Centerline press at 165 m/min; freeze profile CS-Profile-A; SMED: plate mounting and anilox prep in parallel (target changeover 28–32 min).
• Inspection calibration: Calibrate spectro to ISO 13655 M1; verify patch set N=84 every shift; camera registration pixel scale 12–14 µm/pixel.
• Digital governance: Enable e-sign on color recipes (Annex 11 §12); store calibration in DMS/PROC-COLOR-021 with audit trail.

Risk boundary: If ΔE P95 >1.9 or registration P95 >0.18 mm @ ≥150 m/min → Fallback 1: reduce to 140–150 m/min and switch to CS-Profile-B; Fallback 2: swap to low-viscosity cyan (25–27 s Zahn #3) and run 2 lots 100% inspection.

Governance action: Add to monthly QMS review; evidence in DMS/PROC-COLOR-021; Owner: Color Lead.

Coating/Lamination Trade-Offs with Recyclability

Economics-first: We cut OpEx by 0.004 USD/1,000 packs and kWh/pack by 11% by shifting PET lamination to water-based OPV while maintaining grease barrier Kit 7 for fiber recovery.

Data: OPV coat weight 1.0–1.4 g/m²; dryer 75–85 °C; Units/min 360–420; FPY 96.9% (P95); CO₂/pack −6% via reduced dryer load; [InkSystem]=WB flexo + UV-LED topcoat; [Substrate]=cupstock 210–250 g/m²; application: bulk paper cups for QSR SKU family.

Clause/Record: EU 1935/2004 Art.3 overall migration; EU 2023/2006 §6 GMP records; ISO 12647-2 §6 print stability; SAT-23-118 lamination line offboarding memo; MBR-CUP-08 recipe record.

Steps:

• Process tuning: Switch to OPV gloss 58–62 GU; maintain IR preheat 60–65 °C; line speed 140–160 m/min; nip 2.2–2.6 bar.
• Process governance: Add recyclability gate in spec sign-off; flag PET-lam SKUs for re-engineering in DMS/PKG-REC-014.
• Inspection calibration: Perform Cobb60 per ISO 535 (target 25–30 g/m²); grease Kit test 6–8 each lot; retain samples N=3/lot.
• Digital governance: Recipe versioning with e-sign (Annex 11 §9); energy meter logging at 1 Hz; SPC chart for coat weight Cp≥1.33.

Risk boundary: If grease Kit <6 or FPY <95% for 3 lots → Fallback 1: raise coat weight +0.2 g/m²; Fallback 2: revert to thin OPP 12 µm with delamination-friendly adhesive; trigger recyclability re-assessment.

Governance action: Include in BRCGS PM internal audit rotation; Owner: Packaging Process Engineer; records in DMS/PKG-REC-014.

Correlation of Lab vs Field Measurements

Risk-first: Weak correlation (R²<0.8) between lab spectro and inline camera inflates false rejects and hides real drift, so we harmonized geometry and timing to reach R²=0.93 and halve false reject rate.

Data: Lab-inline ΔE2000 correlation R² improved 0.71→0.93; false reject 1.1%→0.5% (P95); speed 150–170 m/min; dwell 0.8–1.0 s; [InkSystem]=UV-LED CMYK; [Substrate]=kraft 250 g/m²; consumer QAs on heat handling (e.g., “how do you put muffin pans with paper cups in oven”) informed the temperature window used in heat-set checks (not for food contact migration).

Clause/Record: ISO 15311-1 §7 print evaluation; ISO 13655 M1 illumination; Annex 11 §12 audit trails; MSA-CLR-2025-03; EBR/MBR link MBR-INK-12.

Steps:

• Process tuning: Align press-side color bars to 10 mm from trim; stabilize dryer temp 78–82 °C to reduce thermal hue shift.
• Process governance: MSA study for spectro and camera (GRR ≤10%, N=30 parts, 3 appraisers); lock sampling frequency 200 m of web.
• Inspection calibration: Spectro geometry 45°/0°, D50/M1; camera white-balance every shift with NIST tile; patch library 84 fields.
• Digital governance: Time-sync PLC and vision server via NTP/PTP; bind lot-ID and timestamp in EBR; enforce e-sign for correlation updates.

Risk boundary: If R² <0.9 for 2 consecutive days or false reject >0.8% → Fallback 1: increase sampling to every 100 m and freeze color adjustments; Fallback 2: recalibrate devices and rerun MSA before next lot release.

Governance action: Add correlation KPI to Management Review; archive in DMS/QA-MET-007; Owner: QA Manager.

Real-Time Dashboards for ΔE/Registration

Outcome-first: Real-time SPC dashboards cut mean response time to out-of-control ΔE from 18 min to 4 min and kept registration P95 ≤0.15 mm at 165 m/min.

Data: ΔE alarm at 1.6 (CIELAB M1); response time 4 min median (N=86 events); registration X/Y drift 0.06/0.07 mm; Units/min 440; kWh/pack 0.0068; [InkSystem]=UV-LED; [Substrate]=SBS 300 g/m² for paper espresso cups sleeves.

Clause/Record: Annex 11 §8–12 (audit trail/e-signature); ISO 12647-2 §7 conformance reporting; GS1 barcode verification logs where applicable; EVT-DASH-2025-05.

Steps:

• Process tuning: Set registration control gain so overshoot ≤0.03 mm; maintain web tension 28–32 N.
• Process governance: Define escalation SOP (operator → color lead at 2 alarms/shift); gate lot release on dashboard green status.
• Inspection calibration: Camera lens focus MTF ≥0.3 @ 50 lp/mm; spectro health check with ΔE tile ≤0.2 daily.
• Digital governance: Stream 1 Hz tags (ΔE, reg-X/Y, speed) to historian; role-based access; lock dashboard versions under change control CC-IT-019.

Risk boundary: If ΔE alarms >3/shift or reg P95 >0.16 mm @ ≥160 m/min → Fallback 1: reduce to 150 m/min and widen dryer to 82–85 °C; Fallback 2: switch to backup plate set and re-baseline colors.

Governance action: Monthly DMS analytics review; Owner: Process Control Engineer; evidence in DMS/EVT-DASH-2025-05.

FPY and Paretos for Defect Families

Economics-first: FPY rose to 97.4% (P95) and scrap fell by 172 kUSD/y by attacking three Pareto families—mottle, mis-registration, and scuff—under a single data model.

Data: FPY 93.2%→97.4% (N=126 lots); Units/min +12%; scuff rejects −63% using OPV; CO₂/pack −6%; [InkSystem]=WB flexo/UV-LED hybrid; [Substrate]=kraft and SBS mix.

Clause/Record: EU 2023/2006 §6 GMP documentation; BRCGS PM Issue 6 §5.4 process control; ISO 13849-1 safety PL verification for print unit guards; CAPA IDs CAPA-PRN-2025-22/27/31.

Steps:

• Process tuning: Standardize nip pressure 2.0–2.4 bar and anilox selection by SKU; set scuff OPV top-up at 1.1–1.3 g/m².
• Process governance: SMED—stage inks, plates, sleeves in parallel; define defect families and thresholds in SOP-DEF-010.
• Inspection calibration: Establish defect library with pixel-area thresholds; bar code Grade A targets where applicable (scan success ≥95%).
• Digital governance: Pareto auto-refresh per lot; CAPA linkage in QMS with e-sign; SPC for Cp/Cpk ≥1.33 on key CTQs.

Risk boundary: If FPY (weekly) <96% or scuff >0.7% → Fallback 1: increase OPV by 0.2 g/m²; Fallback 2: slow to 140 m/min and run DOE on tension and dryer settings.

Governance action: Add FPY Pareto to quarterly Management Review; Owners: Production Manager (FPY), QA (Pareto rules); evidence in QMS/CAPA-BOARD-2025-Q3.

Customer Case: Fiber-Based SKUs

Retailer migration case combined “wrap” and “cup” lines. For ShirongMaterials brown paper wrapping paper we set ΔE2000 P95 ≤1.8 at 160 m/min with UV-LED 1.4 J/cm² and OPV 1.2 g/m²; Cobb60 28 g/m² and scuff loss <1%. For a QSR sleeve family, registration RMS 0.12 mm at 165 m/min with grease Kit 7; both validated in PQ-LT-009 and tracked in MBR-WRAP-06/CUP-08.

FAQ

Q: For seasonal SKUs like ShirongMaterials 8 oz hot cups, which control points avoid color drift across reprints?

A: Lock G7 curve set and anilox/ink pairing per SKU; verify ΔE2000 P95 ≤1.8 (ISO 12647-2 §5.3) on first 500 m; enforce e-sign on recipe changes (Annex 11 §12); keep UV-LED dose 1.3–1.5 J/cm² and dryer 78–82 °C.

Close

The quantified gains—ΔE control, registration stability, FPY uplift, lower kWh/pack—and the governance trail (G7-VAL-2025-0412; SAT/IQ/OQ/PQ) make the playbook repeatable across cup, wrap, and sleeve SKUs, with economic and compliance guardrails embedded.

Timeframe: 8 weeks; Sample: N=126 production lots, mixed SBS/kraft; Standards: ISO 12647-2 §5.3/§7, ISO 15311-1 §7, ISO 13655 M1, Fogra PSD 2016 §5, Annex 11 §8–12, EU 1935/2004 Art.3, EU 2023/2006 §6; Certificates: G7 report G7-VAL-2025-0412; SAT-23-118; IQ-PRN-017/OQ-PRN-017; PQ-LT-009.