A fractionation system is an essential unit in the hydrocracking process. Its optimal operation is challenging because of the complexity in the structure of the distillation tower and composition of the stream. In addition, the series‐parallel structure between the distillation towers of different techniques aggravates the coupling and complexity of the hydrocracking fractionation system (HFS). This, in turn, increases the time complexity of the optimization problem. In this paper, a rigorous mechanism model of an actual HFS is first applied to describe the operating conditions of the HFS. Then, an improved state transition algorithm (STA) with a staged evaluation strategy is proposed to solve the above problem. To overcome problems caused by the series‐parallel structure of HFS, the model is divided into multiple stages for evaluation by mechanism analysis. Furthermore, several typical convergence estimation criteria are introduced to reduce unnecessary model calculations. To solve time‐consuming problems associated with HFS optimization, the adaptive change operator is used to improve the search function of the original algorithm and two performance criteria are presented to reduce the optimization time. The proposed algorithm is successfully applied to the operational parameter optimization problem of HFS with a multi‐fractionator series‐parallel structure. The experimental results indicated that the staged evaluation strategy improved the fast convergence probability of the HFS mechanism model and reduced unnecessary calculations, whereas the improved algorithm increased accuracy and significantly decreased optimization time.

中文翻译：

---

具有多重分馏塔串联-并联结构的加氢裂化分馏系统的多阶段智能运行优化

分馏系统是加氢裂化过程中必不可少的单元。由于蒸馏塔的结构和物流组成的复杂性，其最佳操作具有挑战性。此外，不同技术的蒸馏塔之间的串并联结构加剧了加氢裂化分馏系统（HFS）的耦合和复杂性。反过来，这增加了优化问题的时间复杂度。在本文中，首先使用实际HFS的严格机制模型来描述HFS的运行条件。然后，提出了一种具有阶段性评估策略的改进状态转移算法（STA），以解决上述问题。为了克服HFS的串并联结构引起的问题，该模型分为多个阶段，以通过机理分析进行评估。此外，引入了几种典型的收敛估计标准以减少不必要的模型计算。为了解决与HFS优化相关的耗时问题，使用了自适应变更算子来改进原始算法的搜索功能，并提出了两种性能标准以减少优化时间。该算法成功地应用于具有多分馏器串并联结构的HFS的运行参数优化问题。实验结果表明，分阶段评估策略提高了HFS机制模型的快速收敛概率，减少了不必要的计算，