Statistical process control of dye solution stream using spectrophotometer

  • Lee, Won-Jae (Department of Chemical Engineering, The University of Texas at Austin) ;
  • Cho, Gyo-Young (Department of Statistics, Kyungpook National University)
  • Received : 2010.04.27
  • Accepted : 2010.10.10
  • Published : 2010.11.30

Abstract

The need for statistical process control to check the performance of a process is becoming more important in chemical and pharmaceutical industries. This study illustrates the method to determine whether a process is in control and how to produce and interpret control charts. In the experiment, a stream of green dyed water and a stream of pure water were continuously mixed in the process. The concentration of the dye solution was measured before and after the mixer via a spectrophotometer. The in-line mixer provided benefits to the dye and water mixture but not for the stock dye solution. The control charts were analyzed, and the pre-mixer process was in control for both the stock and mixed solutions. The R and X-bar charts showed virtually all of the points within control limits, and there were no patterns in the X-bar charts to suggest nonrandom data. However, the post-mixer process was shown to be out of control. While the R charts showed variability within the control limits, the X-bar charts were out of control and showed a steady increase in values, suggesting that the data was nonrandom. This steady increase in dye concentration was due to discontinuous, non-steady state flow. To improve the experiment in the future, a mixer could be inserted into the stock dye tank. The mixer would ensure that the dye concentration of the stock solution is more uniform prior to entering the pre-mixer ow cell. Overall, this would create a better standard to judge the water and dye mixture data against as well.

Keywords

References

  1. Cho, G. Y. (2010). EWMA control chart for Katz family of distributions. Journal of the Korean Data & Information Science Society, 21,681-688.
  2. Ko, S. and Cho, Y. J. (2008). Deriving an overall evaluation index with multiple CTQs in six sigma management. Journal of the Korean Data & Information Science Society, 19, 1255-1267.
  3. Lim, C. and Cho, G. Y. (2008). A new EWMA control chart for monitoring the covariance matrix of bivariate process. Journal of the Korean Data & Information Science Society, 19, 677-683.
  4. Willson, C. G. (2010). Statistical Process Control, Lab Handout ChE 253M, The University of Texas at Austin.