Modified PSO And The Application For Gasoline Blending Optimization

Oil blending is an typical nor linear optimization problem.As the most important process of producing refined oil, This step refers to a serial of factors including components, products, oil storage tanks, different order demands and rigid quality test, relates to a large number of complicated physics and chemical reactions. It is complicated to analyze, optimize and simulate such a large system, especially mathematical model is the emphasis. Therefore, mathematical model become an important research direction in refinery scheduling. And resolving blending receipt is key and difficult.This paper first integrally analyze study status about oil blending problems in and abroad and the application status of refinery, compare apart the existing blending methods and blending technology, and build a new oil blending proportion model for the problems about imprecision and speed slowness of the existing oil blending well-proportioned model. Proposing improved Particle Swarm Optimization arithmetic and introducing variation gene in the arithmetic evolution, greatly reduce the probability of convergence for the whole arithmetic, keep the swarm diversity, and improve the arithmetic ability of searching the best solution in the entire space. Applying for the oil blending optimization problem, the simulation result indicates that arithmetic can obtain a better oil blending recipe, and guarantee NLP for the blending constrains while contenting the maximal blending profit, and greatly reduce the constraining margin of refined blending oil. Comparing the method the existing methods, the result of simulations verifys that the method in this paper is effective and adaptive widely, it also creates conditions for improving economic benefit of refinery.Finally, this thesis researches the online pipeline oil blending method by applying PID control methods to realize the recipe online optimization and excution. and points out future study direction.