Integrated Optimization Framework for the Design and Scheduling of Batch Cooling Water Networks

Cooling water system (CWS) research traditionally focuses on the interactions between the cooling water network (CWN) and cooling tower performance, with a continuous background process. Very limited research has considered the integration of batch production and utility systems. This work presents a mathematical framework for the synthesis and optimization of a CWS consisting of multiple cooling towers operating with a batch background process. The CWN exhibits a series configuration, which is characterized by cooling water reuse. A batch CWS superstructure is developed which represents all different pairings between the cooling towers and the batch background network. It is worth mentioning that in previous research, the utilization of cooling water reuse opportunities was constrained by temperature only; however, due to the nature of batch processing, the utilization of cooling water reuse opportunities is subject to both temperature and time constraints. The benefits of this research for retrofit design is the increase in cooling tower availability. In the case of grassroots design, the capital expenditure for cooling tower capacity is reduced. The mathematical model is developed as a mixed-integer nonlinear programming (MINLP) problem within the GAMS platform. The MINLP is solved by making use of the BARON solver ensuring global optimality. Two illustrative examples are presented. Results indicate a 0.45% increase in the annual profit, a reduction of 42.1% in the total recirculating cooling water requirement, and a 25.3% increase in the average overall cooling tower effectiveness. Capital costs and plant area savings were introduced through a reduction in the number of cooling towers required from 3 to 2. The makeup water consumed was also reduced by 6.3%; this reduced the operational costs and plant water footprint. One of the drawbacks of the model presented is the computational time requirement. Future work will address this concern by implementing decomposition and investigating the use of alternative batch background scheduling formulations.