Application of a novel EWMA-ϕ chart on quality control in asphalt mixtures production. (14th March 2022)

• Record Type: Journal Article
• Authors: Liu, Zhen; Wang, Siyao; Gu, Xingyu; Li, Zhigang; Dong, Qiao; Cui, Bingyan
• Is Part Of: Construction & building materials. Volume 323(2022)
• Publication Date: 2022-03-14
• Subjects: Asphalt mixture -- Quality control -- Variability -- Statistical process control -- Exponentially weighted moving average; Building materials -- Periodicals; 624.18
• DOI: 10.1016/j.conbuildmat.2021.126264
• Languages: English
• ISSNs: 0950-0618

Abstract

Highlights: The model of multiple-source variation in asphalt mixtures production is proposed and the traditional Shewhart control chart is not applicable. The detection efficiency of EWMA control chart is higher than CUSUM control chart, however, the sensitivity of the former is not enough especially when the offset is smaller. The modified EWMA- ϕ by an exhaustive search optimization procedure is recommended for good performance in monitoring the asphalt mixtures production. Abstract: Traditional Shewhart control charts have insufficient detection capabilities for the volatility of means in asphalt mixture production. This paper proposes an exponentially weighted moving average (EWMA) control chart to achieve quality control (QC) in asphalt mixture production based on the statistical process control (SPC) theory. We established a mathematical model of multiple-source variation and considered the effect of three sources of variation including internal variations in materials, variations between materials and variations between material batches in asphalt mixtures production. We analysed the differences between the Shewhart, Cumulative Sum (CUSUM) and EWMA control charts in practical engineering. The results show that the detection efficiency of the EWMA control chart is higher when the offset is larger, while the sensitivity is not sufficient, particularly when the offset is smaller. Therefore, in the proposed EWMA- ϕ control chart, parameters were constantly optimized by an exhaustive search optimization procedure to obtain the minimum extra quadratic loss (EQL) and performance comparison index (PCI) values to optimize the average production length (APL) that could enhance the detection ability over the entire shift range. Finally, a case of key indicator control for asphalt mixtures was introduced, and results showed that the APLs of CUSUM, EWMA and EWMA- ϕ control charts were 13, 9.14, and 7.88 days (taking temperature indicator as an example), respectively, which demonstrates that the EWMA- ϕ has timeliness superior to that of the EWMA control chart and the EWMA- ϕ control chart is more sensitive to the offset (particularly when small) for quality detection data. In addition, it showed significant superiority under any indicator monitoring, which illustrates that the proposed control chart is potentially generalizable to the QC of asphalt mixtures production.