Multi-period steam turbine network optimisation. Part II: Development of a multi-period MINLP model of a utility system

Efficient utility systems are important both due to the impact utility systems have on the economic efficiency of the process industry, as well as the negative environmental effects of energy conversion. In this paper a methodology is used to build a multi-period MINLP model for steam turbine network of a utility system. The methodology consists of several steps, including the building of a simulation model of the system, the development of regression model(s) of the system based on simulation and D-optimal design, and finally the development of a MINLP model. The results show that the methodology is capable of analysing a relatively complex steam turbine network in a utility system. The systematic methodology can help to make the MINLP model compact and transparent, and thus avoid the major challenges that are connected with solving large (nonconvex) MINLP models. However, the MINLP models developed using the methodology should be regarded as ‘ad hoc’ models, and have certain limitations as the models are not so flexible compared to models where all the units in the utility system are modelled in detail.