Improved optimization methods for refinery hydrogen network and their applications

Heavier market competition and tighter environmental legislation lead to the increasing demand for hydrogen in refinery. Hence it is necessary for refinery to seek effective hydrogen management strategies to satisfy the increasing hydrogen requirements. In this paper, two improved systematic mathematical methods are developed based on two-step approach and simultaneous optimization approach respectively to retrofit the hydrogen network in refinery. To make the proposed approaches more suitable for real systems, the flowrate and purity at the reactor inlet of hydrogen consumers and the hydrogen recovery of purification units are considered as variables, and the minimum pure hydrogen of hydrogen consumers must be satisfied. Then the corresponding optimization problem is mathematically transformed to a mixed-integer nonlinear programming (MINLP) problem. However, solving a MINLP model directly will result in inconsistency in solution quality and time. In this paper, the solving of the complex MINLP formulation is avoided by using a mixed-integer linear programming (MILP) linearization technique, resulting in better quality, stability, and efficiency than solving the MINLP model directly. The proposed approaches in this paper could make the best use of resources and consequently provide significant environmental and economic benefits for refinery. A real case study is presented to illustrate the applicability of the proposed approaches.