OEE-Based Manufacturing Loss Diagnostics: A Reproducible Synthetic Case Study

Abstract

This technical paper presents a reproducible Overall Equipment Effectiveness (OEE)-based diagnostic workflow for manufacturing loss analysis. The approach combines correctly aggregated OEE, line-shift and product segmentation, Pareto loss ranking, temporal comparison, and deterministic engineering investigation prompts. The methodology is evaluated using a 30-day synthetic discrete-manufacturing case study comprising three production lines, three shifts, four products, and controlled loss patterns with known ground truth. The experiment is designed to evaluate whether the diagnostic workflow can recover intentionally embedded availability, performance, quality, and temporal loss patterns. The reported results are based exclusively on synthetic data and demonstrate reproducibility and ground-truth recovery within the simulated experiment. They should not be interpreted as industrial or causal validation. The accompanying open-source repository contains the implementation, synthetic data generator, dashboard, documentation, and automated tests required to reproduce the case study.