Stochastic Evaluation of Disinfection Performance in Large-Scale Open-Channel UV Photoreactors

Numerical models based on integrated applications of computational fluid dynamics and fluence rate field models (CFD-I models) are often used to simulate process performance in ultraviolet (UV) disinfection systems and other photoreactors. However, the predictions of these models are often presented as deterministic endpoints, whereas actual performance in these systems is variable. The central hypothesis of this research was that CFD-I models can be applied via a stochastic approach [Monte Carlo (MC) simulations] to simulate process performance of large-scale UV disinfection reactors including variability, by allowing appropriate variations in input variables, including UV dose-response behavior of target or challenge microbes, as well as the initial microbial count. The results of these stochastic simulations indicated that variability in E. coli dose-response model parameters and initial viable E. coli concentration play important roles in the variability among predictions of disinfection efficacy. MC simulation predictions indicated that the likelihood of violating the regulatory permits for viable E. coli concentration for a large-scale UV disinfection system was negligible for many operating conditions that span the range of common use, which suggests that it may be possible to reduce the amount of UV power applied in treatment, hence reducing input electrical power while retaining compliance with treatment objectives.