炼化厂油品装卸过程中产生的油气（VOCs）排放问题，不仅造成资源浪费，还严重威胁周边大气环境与作业人员健康。本文系统分析了油品装卸过程中油气挥发机理，对开放式装卸与密闭式装卸两种工艺的油气损耗特征进行了对比研究，并对吸收法、吸附法、冷凝法及膜分离法四种主流油气回收技术的工艺特点、回收效率和经济性进行了系统评价。研究结果表明：油品装卸过程中油气挥发主要受装卸方式、油品蒸气压、装卸温度及鹤管插入深度等因素影响；密闭装卸改造可使油气逸散量降低85%以上；采用“冷凝+吸附”组合工艺时，油气回收率可达97%以上，排放浓度可控制在25 g/m³以下，显著优于单一工艺；改造后某炼化厂年回收油气约1280 t，投资回收期约1年。研究还指出，目前该领域在多油品共用系统适配性、深冷工艺能耗优化、长周期运行稳定性及智能化监控等方面仍存在不足，有待深入研究。

Abstract : The problem of oil and gas ( VOCs ) emissions generated during oil loading and unloading in refineries not only causes waste of resources, but also seriously threatens the surrounding atmospheric environment and the health of workers. In this paper, the mechanism of oil and gas vocalization in the process of oil loading and unloading is systematically analyzed. The characteristics of oil and gas loss in open loading and unloading process and closed loading and unloading process are compared and studied. The process characteristics, recovery efficiency and economy of four mainstream oil and gas recovery technologies, such as absorption method, adsorption method, condensation method and membrane separation method, are systematically evaluated. The results show that the vocalization of oil and gas in the process of oil loading and unloading is mainly affected by factors such as loading and unloading mode, oil vapor pressure, loading and unloading temperature and crane pipe insertion depth. The closed loading and unloading transformation can reduce the oil and gas emission by more than 85 %. When the 'condensation + adsorption 'combined process is used, the oil and gas recovery rate can reach more than 97 %, and the emission concentration can be controlled below 25 g / m3, which is significantly better than the single process. About 1280 tons of oil and gas are recovered annually in a refinery after the transformation, and the investment recovery period is about 1 year. The research also points out that there are still some deficiencies in the adaptability of multi-oil sharing system, energy consumption optimization of cryogenic process, long-term operation stability and intelligent monitoring in this field, which need to be further studied.