• 목록

[Abstract]

Understanding the impact of equipment and process parameter changes on final outcomes is critical for optimizing efficiency and quality in manufacturing environments. However, the complexity of multi-step processes, along with factors such as equipment drift over time and variations in input materials, makes quantitative analysis challenging. This study proposes a method that integrates causal inference techniques with Double Machine Learning (DML) to estimate the average treatment effect (ATE) of equipment parameter changes in multi-step manufacturing processes. The proposed method is applied to the small steel bar (SSB) hot rolling process in steel manufacturing to quantitatively assess the impact of rolling equipment parameters on process output geometry. The results demonstrate that this approach enables robust and causal analysis of parameter effects, providing practical insights for equipment control and process optimization under varying equipment conditions.