A typical refinery consists of the production system and the utility system. Traditional refinery-wide optimization is carried out in a sequential process which not only misses opportunities of increasing the overall refinery economic margin

but also leads to the economic losses of extra fuel gas vent and environmental cost of harmful gas emission. A novel integrated approach which couples the production system and the utility system is presented and optimized simultaneously in a refinery-wide mixed-integer nonlinear programming (MINLP) model. The production unit energy consumption model is introduced by considering the relationship of energy consumption

unit throughput

and operation modes. Material balance of intermediate product such as fuel gas and fuel oil and energy generation and consumption balance of different utilities are considered to improve energy utilization. A real industrial example is investigated to demonstrate the performance of the proposed method which leads to sharp decrease in refinery-wide operation cost and a significant improvement in energy saving and emission reducing.