Decision in High-Variability Manufacturing: Integrating SPC, APC, and Metrology through Risk-Aligned Governance

Abstract

This paper proposes a reliability-oriented framework that treats monitoring and control as an engineered decision pipeline rather than as a collection of independent charts and sensors. The framework integrates sampling design, uncertainty quantification, chart governance, verification logic, corrective action mechanisms, and escalation rules into a unified architecture evaluated by decision-relevant reliability metrics. We define quantitative measures for time-to-detection, expected lots-at-risk prior to intervention, false-hold burden, and time-to-disposition under constrained engineering resources. A representative case-based analysis compares alternative governance designs, including conservative versus risk-aligned thresholds, staged verification strategies, and coupling of SPC alerts with APC adjustments and metrology confirmation. Results indicate that reliability improvements are driven less by increased sensing density than by disciplined governance that constrains nuisance alarms while preserving early detection of sustained drift. Risk-aligned thresholds combined with verification tiers reduce lots-at-risk and maintain manageable hold rates, improving release decision traceability without degrading throughput. The study provides practical guidance for designing maintainable monitoring policies that minimize both escape risk and operational overload, with implications for high-mix, high-precision production environments where uncertainty and drift are unavoidable.