【摘要】 考虑乙烯生产过程中实时操作优化特点,在建立精馏塔严格机理开放式方程优化模型的基础上,提出了基于简约空间序列二次规划(RSQP)算法的精馏塔智能操作优化方法。该法根据精馏塔优化操作的实际性和实时性要求,对简约空间SQP算法进行了一些特殊处理,对收敛条件加入了一些智能化规则,使得优化算法综合考虑优化效益、优化求解时间和质量约束等方面。计算结果表明该法的计算效率高于基于Snopt和一般简约空间SQP算法的精馏塔操作优化方法,并且该方法更符合实时操作优化的要求。更多还原

【Abstract】 Based on the strictly mathematical and open formed mechanistic model of distillation column, an intelligent operation optimization method based on reduced SQP(sequential quadratic programming) algorithm used to optimize the debutanization in ethylene process was presented. In the proposed method, the real time characteristics and rationality of operation optimization for debutanizer were considered, and thinking of utilizing the sparse structure, the information of degrees of freedom and the objective’s structure, the reduced SQP algorithm and hybrid automatic differentiation technology were used to solve this problem. Some intelligent rules for accelerating convergence, balancing the objective and infeasibility and dealing with abnormal situations were merged into the reduced SQP algorithm, so the algorithm can make balance between total benefits, solving time and qualification requirement. Computing results demonstrate that the intelligent operation optimization method proposed in the paper is more efficient than the method based on Snopt algorithm and the method based on general reduced SQP algorithm, and it is more fit for the requirement of real time operation optimization.更多还原