การควบคุมน้ำหนักบรรจุของผลิตภัณฑ์เครื่องดื่มสมุนไพรด้วยการควบคุมกระบวนการเชิงสถิติ

Abstract

This research aims to apply Statistical Process Control methodologies to monitor, control and reduce variability within the production process of a herbal drink, packaged in a standard 80g bottle. After the initial assessment of the production process and identification of key processes, the fill weight per bottle was selected as the quality characteristic of interest. In Phase I, the trial control limits were established using x̅– R charts with fill weight data collected every 30 minutes for 30 subgroups, each with a subgroup size of 5. The trial control limits were calculated, the control charts were established, and the data was plotted. After the process was in statistical control and stable, the upper control limit (UCL), central line (CL) and lower control limit (LCL) from the x̅– R charts were implemented as a process monitoring system in Phase II, where data were collected every 30 minutes, with 5 samples per subgroup. The mean (x̅) and range (R) of each subgroup were immediately computed and compared to the control limits after collection. Since the data were found to be within range of the control limits, the system was stable and the process capability before improvement was analysed, using process mean and standard deviation. The potential process capability (Cp) and the actual process capability index (Cpk) were calculated. Subsequently, potential causes for the fill weight variability were explored using cause and effect diagram and prioritized according to importance by brainstorming with production teams involved in the process. It was concluded that the main causes for the variations in fill weight were nozzle wear-and-tear, lack of preventative maintenance and lack of established procedures for machine inspection before use. After the causes have been determined, the nozzles were repaired and parts were replaced, a protocol for preventative maintenance was initiated, focusing on ensuring that the machines were in working order before use, and a form for daily routine check was created so that the personnel can check designated risk points before the start of each batch to prevent potential disruptions which can affect the filling process. Data were collected again after the improvements have been carried out and Phase I and Phase II control charts were computed again. The analysis of the process capabilities after improvement revealed a mean fill weight of 79.96g, which was closer to the target weight of 80g compared to 80.29g before improvements. The standard deviation decreased from 1.60g to 0.64g. Cp value increased from 0.21 to 0.52 and Cpk value increased from 0.15 to 0.50.