NIR_RealTime_Blend_Endpoint

URS & FS — User Requirements and Functional Specification

This document describes the controlled interface, user requirements and functional behaviour of NIR_RealTime_Blend_Endpoint. The utility is intended for automated verification of laboratory, pharmacopoeial, analytical or manufacturing QC parameters using input.csv and producing a structured output.json result.

Domain limits and critical parameters

• • Traditional approach: fixed blending time + off-line sampling. Slow and risky.
• • PAT approach: NIR probe (e.g., InfraLUM FT-12) continuously acquires spectra from the blender volume.
• • Algorithm calculates metric of difference between sequential spectra or blocks of spectra.
• • When blend becomes homogeneous, spectrum stops changing, and metric drops below threshold.
• • This allows stopping the blender immediately after reaching homogeneity, saving time and energy.
• • Threshold must be set based on validation studies (correlation with reference sampling method).
• • "N consecutive passes" requirement is necessary to exclude false positives due to random spectral fluctuations or mixer blade passing near the probe.
• • Metric can be based on PCA (Principal Component Analysis), Block Variance, or Correlation Coefficient.
• • Probe must be installed in the zone of active powder flow.
• • Real-time NIR data analysis.
• • Robust endpoint determination logic (Consecutive Pass).
• • Integration with blender control systems (PLC/SCADA).
• • Homogeneity Metric: <= Threshold (e.g., 0.1)
• • Consecutive Passes: >= Required Count (e.g., 3)
• • Homogeneity Metric (e.g., Block Variance): ≤ Threshold Value

URS — User Requirements Specification

input.csv contract

BatchNumber,TimeMin,MetricValue,Threshold,ConsecPassCount,ReqConsecPass
BLEND-PAT-2026-001,15.0,0.05,0.1,3,3
BLEND-PAT-2026-002,10.0,0.15,0.1,0,3
BLEND-PAT-2026-003,12.0,0.08,0.1,2,3

FS — Functional Specification

Example output.json

{
  "utilityId": "nir-realtime-blend-endpoint",
  "utilityName": "NIR_RealTime_Blend_Endpoint",
  "overallStatus": "PASS|WARNING|FAIL",
  "sourceFile": "input.csv",
  "checks": [
    {
      "parameter": "BatchNumber",
      "value": "BLEND-PAT-2026-001",
      "status": "PASS|WARNING|FAIL",
      "message": "Rule-based check result"
    },
    {
      "parameter": "TimeMin",
      "value": "15.0",
      "status": "PASS|WARNING|FAIL",
      "message": "Rule-based check result"
    },
    {
      "parameter": "MetricValue",
      "value": "0.05",
      "status": "PASS|WARNING|FAIL",
      "message": "Rule-based check result"
    },
    {
      "parameter": "Threshold",
      "value": "0.1",
      "status": "PASS|WARNING|FAIL",
      "message": "Rule-based check result"
    },
    {
      "parameter": "ConsecPassCount",
      "value": "3",
      "status": "PASS|WARNING|FAIL",
      "message": "Rule-based check result"
    },
    {
      "parameter": "ReqConsecPass",
      "value": "3",
      "status": "PASS|WARNING|FAIL",
      "message": "Rule-based check result"
    }
  ],
  "criticalFindings": [],
  "warnings": [],
  "generatedFor": "QA/QC review and LIMS integration"
}

Traceability matrix

OQ/PQ test scenarios

QA/QC and change control

• Do not rename columns without updating the validator, documentation and test set.
• Retain input.csv, output.json, executable version, documentation and checksum.
• Before production use, perform IQ/OQ/PQ or equivalent CSV/CSA verification.
• Critical limits must be verified against the approved specification, local SOPs and registration dossier.