Models for Improving Management of Biosolids Odors

We describe recent modeling efforts to identify the factors that lead to high biosolids odor levels associated with advanced wastewater treatment plants (AWTP). These factors can be broken down into two groups: (i) those that are beyond the control of the AWTP such as ambient temperature, (ii) those that are controllable such as the number of centrifuges in operation, the amount of lime used, etc.We summarize our findings relative to different statistical models we developed that predict biosolids odor levels based on either subjective or analytic measurements from the District of Columbia Water and Sewer Author (DCWASA). These models take into account a host of factors to predict biosolids odor levels and are then used to generate a relevant probability distribution for odor levels using Monte Carlo simulation. Such probability distributions will guide AWTP managers relative to where to send the biosolids products taking into account the likelihood of high levels and thus indirectly, possible complaints from those living or working near the reuse sites. We also describe our recent efforts in also optimizing the operations of the AWTP and distribution network to balance both biosolids odors and costs. The resulting multiobjective optimization models are computationally challenging due to their size and non-convexities and we discuss some of the salient features as well as representative results. We extend this discussion by noting how to handle stochasticity directly into such optimization models.