(372e) Global Optimization of Water Management Problems In Process Plants

Over the last few decades, optimization of water network systems in process plants has been analyzed using several simplifying assumptions. Among them one can highlighted the assumptions in which the costs (both capital and operating) associated with the regeneration processes are only a function of flowrate; and, others where the efficiency of the these processes (in some instances given as fixed removal efficiency, and others as fixed outlet concentrations) are independent of operating conditions such as flowrates and concentrations. In reality, these assumptions imply that for the same amount of flowrate treated, capital and operation costs are the same regardless of the concentration (and/or contaminant mass load) of the streams. In other words, this means that a treatment of a diluted stream costs the same as a concentrated one. Moreover, it is assumed these regeneration processes can reach the same efficiency (in terms of fixed removal rate or fixed outlet concentration) for both diluted and concentrated streams.

Most of these simplifying assumptions have been made with the objective of reducing the mathematical difficulties of handling non-convex non-linear models. The optimization of water systems including these assumptions is already a difficulty problem to be solved due to the existence of bilinear terms arisen from the contaminants balances.

We propose a MINLP model to solve water management problems using more elaborated mathematical description of the regeneration processes. To solve this model, a new global optimization methodology that we developed, which is able to handle both bilinear terms and the non-linear function arisen from the more elaborated models for the regeneration processes, is used.