The production of regular clean fuels is faced with a problem of declining profit under more strict and costly environmental regulations. To satisfy the desire for higher profit and the firm requirements of environmental protection, it is imperative to improve the efficiency of energy systems within refineries. Over the past decade numerous attempts were made to enhance the energy system, addressing the steam power system and hydrogen system in particular. However, the fuel gas system, which serves as the dominant energy source of refineries, has drawn little attention in the research community. Industrial practices indicate that the energy efficiency of the fuel gas systems can be improved remarkably by optimizing the operation schedules. This paper presents a multi-period optimizing model for the scheduling of fuel gas system within refineries. Modeling of the pipeline system is considered important, which was usually ignored in the former studies. Flow reversal and flow transition in the pipe segments are taken into consideration. Pipelines with branching structure and loop structure can be easily modeled and solved with rational computation effort. Complementarity formulations are utilized in modeling of discrete decisions instead of the commonly used binary variables. Application of this method is illustrated with a case study. (C) 2014 Elsevier Ltd. All rights reserved.