| The refining industry is very important for national economy and our daily life. The world refining capacity is increasing rapidly especially in a developing country like China. The imbalance between the relatively low managerial and technological levels of Chinese refining industry and its rapidly growing speed has provided the impetus for improvements especially after China's entry into WTO (World Trade Organization) in 2001. In this thesis, planning, scheduling and economical analysis, which are the main activities of refinery management, are studied and applied to two large-scale refinery complexes in China.; Nearly all refineries have their own utility systems that provide energy in the forms of steam or electricity to support their production activities. Recently, besides the production of petroleum products (e.g. gasoline and diesel), modern refineries are being reformed to become "energy plants" that export electricity and steam. The success of this reformation depends on an effective integration of the production and the energy systems. So far, few researches have been done in addressing the integration problem. In this study, after developing efficient planning/scheduling models for the main areas of the production system, the energy system was integrated into the plant-wide model to obtain the optimal strategy that maximizes the overall profit. Two industrial case studies were used to illustrate the importance of the integrated model.; A new method, MVA (Marginal Value Analysis), was studied to analyze the economic information provided by the optimal solution of the plant-wide planning/scheduling model. Useful information for refinery debottlenecking, retrofitting, investment evaluation and product pricing, etc., can be obtained from MVA. Further economic analysis using PP (Parametric Programming) was also performed by a new procedure proposed, which can explore comprehensive economic information beyond MVA effectively.; Due to the volatile raw material prices, fluctuating product demands, and other changing market conditions, many parameters in a planning/scheduling model are uncertain. Up to now, very few refinery-planning studies took into account these uncertainties. A novel approximation-based approach is proposed to estimate the overall revenues of a refinery under uncertainty. The approach can obtain good accuracy with better solution speed compared to the existing methods. |
