Data Historical of ESR Defect

Smart manufacturing relies on real-time monitoring, adaptability, and quality control. A critical yet often underestimated factor is the effect of thermal variations on the Equivalent Series Resistance (ESR) of electronic components. Previous studies primarily addressed machine-level or system-level failures or adopted static thermal models. However, this study introduces a novel application of system dynamics modeling to component-level ESR behavior, explicitly addressing dynamic thermal conditions. The model simulates baseline, threshold-based, and predictive temperature control strategies. Predictive control, implemented using linear extrapolation combined with SMOOTH3 filtering in Vensim, improved ESR stability by 28% and reduced the defect rate by approximately 20%. Historical defect data (2.18%) from conductive polymer capacitor production validated the model, with ANOVA confirming significant improvements across scenarios (p < 0.05). Sensitivity analysis showed that ESR deviation was most responsive to the ESR sensitivity coefficient and cooling efficiency. The best sensitivity coefficient was 0.36957, with variations of 1.54 reducing NG ESR to 43% and reduce loss cost 151.2KUSD per month. Practical implications include preempting capacitor-related short-circuit defects and aligning predictive cooling with production scheduling.