---

ENHANCING PROCESS SHIFT DETECTION: A MIX CUSUM-DEWMA CONTROL CHART FOR IMPROVED QUALITY MONITORING

Muh.Nurul Hidayat, Erna Tri Herdiani*, Georgina Maria Tinungki

---

Download PDF

---

Abstract 

Quality control in manufacturing processes plays a crucial role in ensuring that products meet predefined standards. The Shewhart control chart is extensively utilized for overseeing and evaluating process performance; Nevertheless, it struggles to detect small shifts in production processes, which may result in slower detection of anomalies. In order to resolve this limitation, this study puts forward the development of a Mix CUSUM-DEWMA control chart, which integrates the advantages of the Cumulative Sum (CUSUM) and Double Exponentially Weighted Moving Average (DEWMA) charts to increase responsiveness to identify shifts. The proposed method is applied to defect data from Butsudan production at PT Maruki International Indonesia to evaluate its effectiveness compared to the Mix CUSUM-EWMA control chart.  Experimental results demonstrate that the Mix CUSUM-DEWMA chart is more effective in detecting out-of-control signals across most  values, particularly within the range of 0.075 to 0.9. In terms of Average Run Length (ARL), the CUSUM-DEWMA chart consistently yields lower ARL values within the  range of 0.05 to 0.9, indicating faster detection of process shifts. The chart achieves its optimal performance at a  0.1, where it most effectively identifies shifts in the mean of the production defect process. The results validate that the newly developed control chart strengthens or enhances the efficacy of quality control systems. It facilitates the early identification of process shifts and helps lower the occurrence of product defects.

Keywords: Statistical Process Control, Cusum Control Chart, DEWMA Control Chart, Process Shift Detection, Quality Monitoring