The up-to-date approaches
to optimizing water systems only include
fresh water, regenerated water and wastewater and ignore other types
of water in refinery, i.e., desalted water, deaerated water, circulated
cooling water, steam with different pressure levels and condensate
water. Therefore, the existing mathematical model for water system
optimizaiton is not directly applicable for the optimization of practical
refinery water systems. To overcome the limitation and bridge the
theory and application, we first presented a generalized model of
water-using processes including multiple types of water and a general
superstructure for the optimization of refinery water system. The
superstructure consists of water-using processes including multiple
types of water in the main production units (i.e., crude oil distillation,
fluid catalytic cracking), water pretreatment systems (i.e., fresh
water station, desalted water station, steam power station) and wastewater
treatment system. The flow rate balance equations for those components
of a refinery water system and the correlation for all types of water
are formulated. The replacement ratio of altered type of water is
introduced in the flow rate balance equations for water reuse/recycling
and it avoids the imprecise data extraction of limiting water quality
for the inlets of water-using processes. We presented two mathematical
models with different objective functions (minimum flow rate of water
resource (Scenario 1) and minimum partial annualized cost (Scenario
2)). The proposed models are applied for the optimization of the water
system of a large-scale refinery in China. Results show that the water
system with a minimum flow rate of water source can be obtained in
Scenario 1. In Scenario 2, the profit of water conversation for five
strategies cannot offset the investment cost of added pipelines, and
their actual replacement ratios are zero. It leads to an economic
and simpler water system with slightly higher flow rate of water resources.