Modelling, simulation and optimization perspectives of an industrial steam network - Case study at a major oil refinery on the West Coast of Sweden
Steam is a heat carrier for industrial plants that is relatively cheap to produce and distribute, compared to other heat transfer media aimed for the same temperature range. It is the most used medium in thermal power plants, but can also be used in a utility network in, for example, a refinery, where it is used to handle the heat distribution. Utility steam can be used to heat process streams up to about 250°C, while also having the ability to generate work. Management and optimization of steam networks can lead to heat, electricity and economic savings. However, many times, there are difficulties associated with data collection and measurements. As a matter of fact, some secondary process variables like the outlet temperature of a steam turbine, even though useful to check the isentropic efficiency of the machine, could not be measured. This type of lack in information brings difficulties in the plant modelling and analysis. The present master thesis, which has been carried out at the department of Energy and Environment at Chalmers University of Technology, concerns the modelling and optimization of the steam system at the biggest refinery of Sweden. By using engineering and thermodynamic knowledge, a model of the steam network is developed, first in spreadsheet databases, and then aided by a proper modelling software. Assumptions and thermodynamic relations are used to model the different components that the steam system is composed by and to calculate missing values. Several steady state simulations are made with the model. For each of them a comparison with the real behaviour of the system is made by connection of the software output to a validation spreadsheet through a plug-in. Moreover, once the reliability is tested, examples of how the model can be used for analysis of energy retrofit measures are presented. Finally, an economic optimization of the system operation for a given set of inputs is carried out and its results are analysed.