In this paper, hydrogen management in refineries is examined considering a new perspective to optimize refinery costs. In order to minimize the total annual cost in refineries a mixed-integer non-linear programming (MINLP) model has been developed. Retrofitting the hydrogen network and determining the hydrogen surplus capacity of refineries are the main outputs of the proposed model. Then, to utilize this surplus capacity and increase the profitability of refineries, surplus hydrogen is used to produce electricity and ammonia. The results of a case study showed that retrofitting of the hydrogen network can save 12% in the consumption of fresh hydrogen from hydrogen plants. After retrofitting of the existing hydrogen network, three scenarios for efficient use of hydrogen surplus are defined as electricity production by SOFC fuel cells, electricity production by hydrogen turbines, and ammonia production. The comparison of proposed scenarios showed that the ammonia production scenario has better economic results with a payback period of 3.52 years and 16.81 MUSD/yr. total saving. Furthermore, the sensitivity analysis of electricity prices in two electricity production scenarios was performed, and the results showed that by considering the exporting electricity price, electricity production scenarios are profitable due to having payback periods of about five years.

在本文中，考虑了优化炼油厂成本的新视角来研究炼油厂的氢气管理。为了使炼油厂的年度总成本最小化，已经开发了混合整数非线性规划（MINLP）模型。改造氢气网络并确定精炼厂的氢气过剩能力是该模型的主要输出。然后，为了利用这种剩余容量并提高炼油厂的利润，多余的氢气被用来生产电力和氨。案例研究的结果表明，对氢气网络进行改造可以节省制氢厂消耗的新鲜氢气的12％。改造现有的氢气网络后，定义了三种有效利用剩余氢气的方案：SOFC燃料电池发电，氢涡轮机发电，氨气生产。对拟议方案的比较表明，氨生产方案具有更好的经济效果，投资回收期为3.52年和16.81 MUSD /年。总节省。此外，对两种电力生产情景中的电价进行了敏感性分析，结果表明，考虑出口电价，由于投资回收期约为五年，因此电力生产情景可获利。