Sim time08:00:00
Speed60×
Shift A · Day 1

Calibration anchors

Public HJT process windows the twin's physics is tuned against

Cell-side performance

MetricIndustry windowSourceNotes
Cell efficiency η23.5 – 24.7 %Kaneka, LONGi HPBC, Meyer Burger, MaxeonM10 production HJT, late 2024
Implied V_oc (iV_oc)740 – 750 mVKaneka public PV Tech disclosuresPost-PECVD passivation
J₀5 – 10 fA/cm²ITRPV 2024Post i-layer optimisation
Sheet resistance (TCO front)30 – 80 Ω/□ISFH / Helmholtz HZBITO/IWO 80–100 nm
Transmission (400-1100 nm avg)> 85 %TCO supplier datasheetsCritical for J_sc
Cell P_max (M10)7.2 – 7.4 WLONGi HPBC releaseNominal 7.3 W per cell used as twin baseline

Module-side performance

MetricIndustry windowSourceNotes
Module power (72-cell M10)580 – 620 WTier-1 module datasheets
Cure α (lamination)> 0.85DuPont EVA application notesCrosslink density target
Bubble defect rate< 0.5 %IEC 61215Visual + EL screening
Microcrack rate (M10, MBB)< 1 %VDMA PV roadmap

Process windows

MetricIndustry windowSourceNotes
PECVD i-layer thickness5 – 8 nmKaneka 2014 patent literatureBelow 4 nm → epi growth
PECVD deposition T180 – 220 °CMeyer Burger SmartLine specsAbove 220 °C → H effusion
TCO O₂/(Ar+O₂)1.5 – 2.5 %ULVAC sputter manuals
Solder iron T220 – 260 °CCell-to-ribbon process windows
Lamination plate T148 – 155 °CBürkle / Meier press specs
Vacuum hold240 – 480 sDuPont EVA processing guide

RIL Defect Catalog Adopted (18 defects)

Optimon recognizes, attributes, and reasons about RIL's actual defect taxonomy. The 18 defects below drive the EL Gallery, the Quality Inspection page, the Bayesian Root-Cause hypothesis space, the fault injectors, and Ved's responses. Source for all entries: RIL FI Defect Specification (Choice column accepted).
CodeNameTypeSurfaceTier rule
101Crack cell (middle)Cell DefectELNot allowed in A/B; allowed in C if crack piece doesn't overlap
102Crack cell (outside wire)Cell DefectELNot allowed in A/B; allowed in C if crack piece doesn't overlap
112Crushed cellCell DefectELNot allowed in any module → Scrap
129Cell chipCell DefectELA: ≤1×2 mm, max 1/cell, max 3 cells/module
106Broken fingerCell DefectELA: ≤5 broken fingers, total gap ≤30 mm
108Broad fingerCell DefectELA: width ≤0.5 mm
115Incomplete depositionCell DefectELRefer to contamination scheme
117Cell flakeCell DefectELRefer to contamination scheme
151String spacingCell DefectELA: 0.5–4.5 mm
156Uneven stringCell DefectELA: displacement ≤2.5 mm
198Bubble (active)Bubble DefectELA: single ≤15 mm², total ≤25 mm²
199Bubble (inactive)Bubble DefectELRefer to scheme
143Glass bubbleGlass DefectVISUALA: front/back W<0.5 mm, L≤1 mm, qty 3
635Glass brokenGlass DefectVISUALNot allowed in any module → Scrap
141Glass scratchGlass DefectVISUALA: L<50 mm, W<1 mm, qty 1
402Frame corner gapFrame DefectVISUALReference scheme
414Sealant discontinuousFrame Sealant DefectVISUALReference scheme
906Power output lowPower Output DefectFLASHBelow recipe spec band

Honesty footer

Process Twin (HJT) is a physics-driven simulator built to demonstrate Optimon's industrial intelligence approach for HJT solar manufacturing. Models are calibrated to the public HJT process windows above (Kaneka, LONGi, Meyer Burger, Maxeon, ITRPV, IEC 61215, VDMA) and tuned for operational reasoning — they are not certified to any specific OEM tool. All telemetry is generated by the on-board physics simulator. Live PLC/SCADA/MES ingest is wired but not enabled in this build. State is in-memory; persistence via Lovable Cloud is wired and can be enabled. EL imaging is procedurally generated with a defect signature library; production deployment integrates with EyeZ, Optimon's machine vision module.