The Use of Simplified Statistical Processing Techniques to Increase Sensitivity of Detection to Particulate Breakthrough

Abstract

The monitoring of filtration performance is often accomplished through the monitoring of the filtrate (effluent) stream for particulate materials that would be present in a compromised filtration system. Two technologies that are used to monitor for particulates are laser nephelometers and particle counters. These two particulate detection technologies have proven to be effective on certain types of filtration systems, including membranes. However, particulates that would travel through a compromised filtration system can be diluted to below detection levels. This paper provides a means of analyzing the raw laser turbidity signals using simplified statistical procedures that can help to regain sensitivity to the presence of particles under such conditions. The method was demonstrated to be effective in several different filtration integrity studies. Such statistical processing techniques are easily applied to existing instruments through algorithms that ultimately provide an additional means for detection of filtration breakthrough.

Keywords:

• Laser turbidity
• breakthrough detection

Acknowledgments

A significant effort was made by many individuals to develop and test the applications of the various technologies associated with this study. This included the chemists, engineers, technical writers, and technical illustrators at Hach Company who helped conduct the testing, review data, and present the results. Special thanks also goes out to the Leah Oxenford of Hach Company for her efforts in generating many of the results that were presented. In addition Mark Lichwardt of Burns & McDonnell and Frank Watt of the City of Thornton, Colorado's Columbine Water Treatment Facility for providing the opportunity to test the methodologies presented in this paper on several different membrane pilot plants.