289 CHEMICAL ENGINEERING TRANSACTIONS Volume 21, 2010 Editor J. J. Klemeš, H. L. Lam, P. S. Varbanov Copyright © 2010, AIDIC Servizi S.r.l., ISBN 978-88-95608-05-1 ISSN 1974-9791 DOI: 10.3303/CET102100149 Please cite this article as: Iancu P., Lavric V. and Pleşu V., (2010), Optimisation of water networks using ranking as pre-processing step to get equipartition driving force, Chemical Engineering Transactions, 21, 289-294 DOI: 10.3303/CET1021049 Optimisation of Water Networks Using Ranking as Pre- processing Step to Get Driving Force Equipartition Petrica Iancu*, Vasile Lavric, Valentin Pleşu Centre for Technology Transfer in the Process Industries, University Politehnica of Bucharest, UPB-CTTIP, 1 Polizu Street, 011061 Bucharest, Romania cttip@chim.upb.ro Some ordering criteria are proposed in this paper, as pre-processing step for process design and optimisation of water networks, leading to equipartition of driving forces for mass transfer along of the associated system. The water network is abstracted as an oriented graph, the water using units being ranked based upon maximum of load, fresh water need, inlet or outlet constraints. This way, the local recycling is avoided and the driving force of the mass transfer could be more or less constant for the cascaded units. Optimal solution for a water network using as ranking criterion a properly chosen design variable is presented. 1. Introduction Wastewater minimisation for a water using network is an efficient process integration methodology for process design that considers abatement of technological utilities (supply water flowrate) and diminution of the environmental impact (wastewater flowrate) using an optimisation based framework. Water network is considered as a whole system, the quantity and quality of water is allocated to each water-using unit such as to maximise water-reusing and minimise wastewater discharge. In the end, a complex water network is described by a NLP mathematical model to be solved with an optimisation tool able to tackle problems involving high number of variables. Different simplifications (LP, MILP, MINLP – depending on model complexity) were proposed, considering reuse/recycling and/or regeneration strategies, or including treatment units in the water network. The general approach is to consider water network as superstructure taking into account all possible connections between units, sources and treatment units. For simple water networks, the mathematical models are easily solved using different methods: graphical or mathematical optimisation. But, for complex water networks (like oil refinery and petrochemical water network) with large number of variables, well known solving methods can fail. Sauar et al., 1996, used the driving force equipartition principle for heat, mass, and charge using irreversible thermodynamics combined with optimization procedures as a new tool for process design and optimisation. The principle says that the best trade-off between energy dissipation and transfer area is achieved when the thermodynamic driving forces are uniformly distributed over the transfer area. They claimed that process design should be