UVDIS REQUIRES VALIDATION FOR SIZING UV REACTORS

The UVDis algorithm couples a hydraulic model with an irradiation model to estimate UV reactor dose delivery. The main shortcomings of the hydraulic model are described. The irradiation model itself may be accurate enough, however its main shortcoming is that the primary parameter to the model, lamp output power, can have significant errors associated with it. The actual physics of UV disinfection are highly non-linear, due to the non-linear nature of fluid flow, and the linearized model of UVDis is shown to have significant short comings. In this study, results from a bioassay of a UV reactor are compared to UVDis. A UVDis Factor, defined as the reduction equivalent dose (RED) of a UV reactor based on bioassay field data divided by the UVDis dose, is introduced. It is shown that the UVDis Factor values are substantially below 1.0, indicating non-conservative predictions of dose by UVDis. It is also shown that the UVDis Factor is a function of both UV transmittance (UVT) and flow rate, indicating that UVDis is not accurately modeling either of these operating variables: variations in UVT or variations in flow rate.