Gasoline Blending Optimization Combined Process Analytical Technology With Sequential Linear Programming

As a final procedure of gasoline production, gasoline pipe-blending process will directly affect the quality of petroleum products and economic benefit of refineries. The corresponding optimization and control systems have been developed by foreign companies, which can make full use of all component gasoline to satisfly crucial gasoline product quality and increase the blending profit. The blending optimization systems and online analyzers introduced by domestic refineries are usually complex and expensive; In addition, the properties of domestic gasoline vary greatly, which deteriorate the practical application effect. For the above reasons, the relevant blending models and optimization algorithms are summarized firstly in this thesis, then, a real-time optimization algorithm based on the sequential linear programming is proposed, which is validated effectively by comprehensive simulation experiments. Based on the proposed algorithm, the relevant real-time optimization control software is developed. The detailed contents are as follows:(1) Briefly introduce the framework and application of the existing real-time optimization and control software. Systematically analyze the merits and drawbacks of various existing octane prediction model and elaborate the real-time optimization technique in the optimization and control system. Outline the applications of real-time optimization algorithms and process analytical technology in the gasoline blending process.(2) Propose a real-time optimization method based on the sequential linear programming. The existing blending optimization models are complex and the common optimization algorithms are complicated, so dynamic optimization idea is introduced and a new blending scheme is proposed. The octane number can be approximated as a linear blending model during a short-term blending process. For an actual gasoline blending process, the objective function and constraint conditions are built, and then the optimization problem is converted to a simple linear programming problem. A completed simulation experiment is set up to validate the above dynamic optimization scheme. The experimental results show that the proposed scheme based on the linearized models can automatically adjust the flow ratios of all component gasoline to meet the need of gasoline product quality, which will increase the economic benefit.(3) Develop a set of on-line optimization simulation system for the gasoline pipe-blending process. The system consists of two computers, one simulates a simplified blending process, and another runs the optimization control software. Data communication of the two computers is realized through the IP Protocol. Then, the simulation software of blending process and optimization control software are designed. Finally, the typical simulation experiments are finished to validate the above optimization algorithm.