(140m) Optwatnet- A Software For The Optimal Design Of Industrial Water Networks

Freshwater supply around the world is under pressure due to a growing population and a significant boost in agriculture and utilization in an industrial wide range of products and applications. Water is probably the most broadly used raw-material in the process industries and it has been intensively used in abundant quantities by the chemical, petrochemical, petroleum refining, food and drink, pulp and paper and many other industrial sectors, for multiple purposes. These usages lead to contaminated wastewater, which may pose an environmental pollution problem. Increasing water costs, restrictions on water use and increased environmental awareness have driven designers towards conceiving more efficient water systems, targeted for minimum water consumption and wastewater generation, through the identification of re-use and recycling opportunities and integration of the water-using (WUN) and water treatment networks (WTN).

Since the seminal paper of Takama et al. (Comput. Chem. Eng. 1980, 4, 251) who addressed the simultaneous optimization problem of water-using and treatment networks in a petroleum refinery with a non-linear program (NLP), many contributions have appeared. While modeling the design problem as a non-convex NLP or mixed integer nonlinear program (MINLP) is straightforward, its solution is not a trivial matter. Global optimization algorithms may require too much computational time for large problems, while local solvers may end up with local optimum solutions. In between, there are efficient methods that generate multiple, structurally different starting points, and solve the NLP several times, with very fast local optimization solvers, which aim at finding global optimal solutions in less time than that required by the global solvers (Castro et al., Resources Conservation & Recycling 2007, 50, 158).

This paper presents the software OptWatNet, developed in Visual Studio 2005, which is basically a user friendly interface for data input, and results visualization superimposed over the optimized water networks. It is linked to a few GAMS models containing the different methods that involve the solution of at least one linear problem (LP), for the initialization, and one NLP for finding the optimal network. At the moment, the WUN and WTN design problems are tackled separately, with 4 or 2 alternative methods, respectively, other than using the global optimization model BARON. These methods, some being presented for the first time, rely on a different number of starting points and typically, the higher the number the better the quality of the solution. Thus, this variety of options allows the user to choose the most appropriate method according to the available time he/she has got to produce a solution.

The embedded models in OptWatNet are general enough to deal with multiple contaminant systems, different freshwater sources, fixed contaminant load and fixed flowrate operation units, while treatment units are characterized by fixed removal ratios and fixed outlet concentrations. Future work will include the study of the integrated problem, with the possibility of getting further savings in freshwater consumption and wastewater generation.