Integrated Optimization Of Resource Networks In Refinery

During the past decade, the worldwide crude oil has become heavier and contains more and more hazardous impurities, while the deteriorating environment urges for stringent clean fuels specifications. Furthermore, the market demands require deep processing of crude oils. Hydrocracking and hydro-treating are widely used in refineries to upgrade heavy oils to more valuable and clean products, leading to an increasing demand for hydrogen and energy. The purchasing cost for hydrogen and energy are now among the three dominating contributors to the system operating cost, after the cost of crude oil. Integrated optimization of resource systems plays a significant role in resources conservation, emission reduction and further enhancement of the plant profit. The integrated optimization of resource systems in refinery includes system design/revamping optimization and system operation optimization. The former mainly focus on new network design and revamping of the existing network, while the latter is about the operation optimization of the existing network.The main scope of this thesis is to carry out network design/revamping optimization and system operation optimization to the refinery resource networks including hydrogen network, fuel gas network, and the heat integrated water network. Systematic designing methods as well as scheduling techniques were developed for each system, aiming to improve the system resource utilization. The main work and contributions are as follows:(1) An improved modeling and optimization approach was developed to integrate H2S removing units into hydrogen network integration, thus the restriction placed by the overloaded H2S content for the ultra-high utilization of system hydrogen resource can be eliminated. In order to optimize the desulfurization process and obtain global optimal result, simplified mass exchange network was incorporated into hydrogen distribution network, and desulfurization ratio was introduces and considered as optimizable variables in the optimizing process. Total annual cost was employed as the optimizing object to investigate the tradeoffs between hydrogen distribution network cost and desulfurization network cost. Pressure constraints and impurity concentrations were considered, and cost equations were established to determine the installation of new equipments in order to synthesis economical network design. Case study illustrated that the proposed method was capable of further enhance the efficient utilization of system hydrogen resource.(2) A systematic mathematical modeling methodology for the optimal synthesis of sustainable refinery hydrogen networks was presented. The proposed mixed integer nonlinear programming model took both the economic and the environmental aspect of the hydrogen network into consideration. Total annual cost was employed to evaluate the economic efficiency of the network, while the environmental performance was assessed by the total CO2 emission of the network. Two types of fresh fuels were investigated. Both single-objective and bi-objective optimization model were built for the optimization problem. The bi-objective optimization model was solved by an adaptive weighted-sum method and the economic-environmental Pareto front was obtained, which helped to determine the most promising options for the reuse, purification and combustion of hydrogen streams. Numerical example shows that the proposed approach is efficient and powerful.(3) An optimization model was proposed for the multi-period scheduling of refinery hydrogen networks based on the practical consideration of multi-period operation in refineries. The performance of pipeline network was considered important for the practicability of the scheduling result. Therefore, detailed pipeline model was incorporated into the scheduling model and the storage capability of the pipeline system was considered. The developed model not only can handle the multi-component and the non-ideal nature of the hydrogen pipeline network, but also allow flow reversals and flow transitions inside the pipeline. In order to obtain a more stable and efficient producing schedule, the exergy efficiency as well as the CO2 emission amount of the hydrogen plant were considered. The discrete decisions were modeled by a new method——complementary constraints, instead of binary variables. A real refinery case study showed that applying the proposed method can generate a stabilized production schedule for the hydrogen plant, which improved the efficiency of the system.(4) A multi-period MPEC optimizing model for the scheduling of fuel gas system in refinery was presented. The pipeline model considering flow reversal and flow transition was extended to the fuel gas system. The multi-component nature of the gas system was also took into consideration. Pipeline systems with branching structure as well as loop structure could be solved simultaneously with the scheduling problem. Complementarity formulations were used to model the discrete elements instead of the commonly used binary variables. The case study illustrated that by applying the developed model practical and favorable scheduling scheme can be obtained with rational computational effort.(5) A one-step methodology addressing the simultaneous synthesis of water allocation and heat exchanger network was developed. The superstructure and the mathematical model were improved, thus more possible network structure were included. Direct and indirect heat exchange were fully explored. Additionally, complementary formulations were also applied in modeling the logic variables instead of binary variables. Both the MINLP model and the MPEC model were developed. The capabilities and efficiency of the proposed method were illustrated with two case studies. It was clearly indicated that, the network designs achieved by the proposed method were economically equal to or better than the literature results. Compared to the previous sequential method, the proposed method can generate better results with less effort. Compared to other simultaneous method published in the literature, our method offered an alternative comparable way to synthesis water allocation and heat exchanger networks simultaneously.