The biodosimetric data from the 6-lamp reactor was properly correlated using standard and multiple linear regression analyses. An emphasis was put on the importance of choosing a regression equation that is statistically representative but also relevant to the UV disinfection principles. As a result, Tier 2 design fluences of 268 to 292 J/m2 were found for 3.0-log inactivation of Cryptosporidium, which is much lower than the equivalent Tier 1 fluence of 420 J/m2, hereby confirming the conservativeness of the assumptions underlying the Tier 1 analysis. The Tier 2 analysis detailed in this paper has been implemented in an Excel sizing spreadsheet that is now being used to bid and size UV systems for the disinfection of drinking water.
Implementing tier 2 analysis on biodosimetric test results
With both goals to develop a sizing program and an operation matrix the biodosimetric results of an in-line cross-flow reactor equipped with 6 medium pressure lamps were analyzed per the US EPA Ultraviolet Disinfection Guidance Manual (UVDGM) Tier 2 approach. Using a Tier 1 analysis was quickly ruled out for this reactor, as the surrogate microorganism used during biodosimetry was B. subtilis instead of the standard MS-2 phage. The expanded uncertainty term (E) was by far the most challenging parameter to determine as part of the Tier 2 analysis. The current draft version of the UVDGM does not include a detailed guidance on how to compute uncertainties from correlations that involve a large number of experimental data with multiple parameters such as Reduced Equivalent Fluence (REF), flow, water transmittance, lamp UV output and sensor reading.