An EWMA‐Driven Adaptive CUSUM Scheme With Variable Sampling for Enhanced Process Mean Monitoring

ABSTRACT This study introduces a Variable Re‐weighted Adaptive CUSUM (VRACUSUM) chart with variable sample size, designed to reliably detect unknown process shifts of diverse magnitudes. Unlike traditional CUSUM charts, the VRACUSUM leverages an EWMA‐driven statistic, coupled with an unbiased shift estimator and an adaptive smoothing constant, to continuously re‐calibrate the slack value and sharpen shift detection. Monte Carlo simulations demonstrate its efficiency to pick up delayed and subtle changes promptly when traditional charts perform poorly. The effectiveness of the suggested control chart is measured based on run length profile. Its greater sensitivity is always exhibited in its results. A practical case study of manufacturing piston rings demonstrating the effectiveness of the proposed approach in a number of different monitoring scenarios, also demonstrates the real‐world applicability of the proposed approach.