The Research And Application In The Oil Blending Based On Intelligent Control

Oil blending is one of the most important processes of oil storage and movement system in refineries. Great benefits can be achieved from product oil blend optimization. The development trend for oil blending means is the advanced on-line blending technique which can be controlled and optimized with computer, and also it is the trend to apply multi-period blend planning and scheduling optimization technology as part of the whole refinery's information system.A BP network predictive control model is established by this article since it is a nonlinear and delay blending objiect, which is simulated with Matlab. The results for two kinds of situations-match and mismatch to original model are analyzed as well as Compared with tradition's ratio control. It indicates that the merit which the BP network predictive model is applied to the blending model is compatible; the parameters can be studied online and adjusted to track the actual system's slow change, it owns good robustness when the model isn't imprecise. A deep research on the control and optimization for oil blending are made by author, who has finished software's development work in the gasoline pipe-line online blending system, and it is integrated in the Deltav control system via the OPC way. a big reference value could be provided for the refinery enterprise in the process of oil blending.Above this deepen research, a new control method-the control model for oil blending based on the multi-Agent is also proposed. And its stability is analyzed. The production scheduling of oil blending is the center of organization and administrative in the refinery enterprise, also it is a research subject that academic circle and the industrial world take it seriously and widely. Now, the multi-Agent technology has applied in the industry production scheduling in factory. It provided a new way for the dynamic question's complexity and the uncertainty. Along with artificial intelligence's development, the multi-Agent technology will certainly become the powerful tool for the optimization for oil blending and scheduling.At last, the research and development of the theory and application in this area are also presented.