| The intensive use of water in refineries and process plants was the object of several studies starting in the early eighties. All these Studies have aimed at the reduction of the water consumption as well as the allocation of wastewater to treatment units. Two major problems have been proposed and solved to various degrees of detail. On one hand, the fresh water allocation wastewater reuse problem has been addressed by various researchers. The first line of work started mostly as graphical and evolved lately towards mathematical programming. On the other hand, the problem of designing wastewater treatment units has been also studied using semi-empirical as well as mathematical programming. Some connections between the two problems have also been studied. This thesis focuses on the optimal solution of the single component water/wastewater allocation-planning problem. First, a robust methodology to obtain an optimal design is presented. The method uses a concentration grid water allocation procedure to obtain preliminary optimal structures. A merging procedure provides the final structures. In addition, the use of different water allocation strategies shows that the problem has several alternative solutions. Secondly, this work shows the existence of necessary conditions of optimum for single component for the aforementioned problem. Furthermore, it is shown that using these necessary conditions in combination with some sufficient conditions, the single component problem can be efficiently solved by hand through a simple and systematic algorithm. Moreover, it is also shown that using the necessary conditions of optimality, the water allocation-planning problem can be reduced to one single LP for the single component case. The solutions are rigorous and global, even when compulsory and forbidden connections are imposed. Finally, the intricacies of the heat integration of these systems as well as different aspects of design and retrofit of these systems are explored. |
