Name: | Description: | Size: | Format: | |
---|---|---|---|---|
282.8 KB | Adobe PDF |
Advisor(s)
Abstract(s)
This paper proposes a new strategy for the optimal design of water-using networks in industrial systems featuring possibly more than a single water source and multiple contaminants. The model formulation is supported on a superstructure that exploits reuse opportunities and gives rise to a non-convex nonlinear which often leads to local optimal solutions. To overcome this, the new approach generates multiple initialization points, one for each possible sequence of operations, where a particular starting point is obtained by the sequential solution of a small set of related linear programs. The best solution of the several non-linear problems that are solved is then assumed to be the global optimal solution. The results obtained for a set of case studies have shown that the best initialization point is often the global optimal solution and that the procedure as a whole is efficient in escaping local optima.
Description
Keywords
Global optimal solution
Citation
Teles, J. P.; Castro, P.; Barbosa-Póvoa, A. P.; Novais, A. Q. LP-based heuristic procedure for the optimal design of water using networks with multi-contaminants. In: CHISA 2006 - 17th International Congress of Chemical and Process Engineering, Praha, Czech Republic, August 27-31 2006, 6p.