MAXIMUM EXPONENTIALLY WEIGHTED MOVING AVERAGE WITH MEASUREMENT ERROR (USING COVARIATE METHOD) USING AUXILIARY INFORMATION FOR CEMENT QUALITY CONTROL

Abstract

The main quality characteristic at XYZ Inc. that should be observed is Compressive Strength. Cement production quality control is carried out on the average and process variability jointly with the Max-EWMA control chart. Measurement error can be found in the Compressive Strength. It can affect the sensitivity of the control chart, so quality control will be carried out by considering the presence of measurement error. Handling measurement errors can be done through three approaches (covariate method, multiple measurements, and linearly increasing variance). This research only focuses on the covariate method. Auxiliary variables also explain variance in the production process, so they are also considered in this research, with Blaine used as an auxiliary variable. Therefore, the control chart that will be formed is the Max-EWMA ME (Covariate) AI. The Max-EWMA and Max-EWMA ME (Covariate) AI control charts show that the XYZ Inc. cement production process based on variability and process averages is simultaneously statistically controlled. The controlled Max-EWMA control chart has an upper control limit of UCL=1.503018, and parameters μy = 252.5823 and σy2= 970.1596. Max-EWMA ME (Covariate) AI has in-control parameters μx = 251.49; σx2= 975.809; A = 198.14; B = 0.2165; σm2 = 917.798; μw = 341.05; σw2 = 163.0266; ρ = 0.2550; ρ∗ = −0.144. The Max-EWMA ME (Covariate) AI control chart is more sensitive than the Max-EWMA control chart. Cement production capabilities based on Compressive Strength have a Cpl and Cpk capability index of 1.54, which means that the cement production process is capable, consistent, and has high accuracy so that the quality has reached the target.