中文摘要：

对于加工多种原油的炼油厂，原油调合调度可以改善进料原油性质，它对保证后续生产装置的平稳操作、先进控制的实施以及生产计划的完成具有重要意义。基于实际生产背景和需求，文献[4—5]提出了原油调合调度模型以及基于2层优化结构和序优化的求解方案。该方案中，原油掺炼序列和流量的混合优化问题转化为掺炼序列的采样和评价问题，它们分别涉及到算法的广度搜索和深度搜索，如何处理二者之间的平衡关系对于合理利用计算量并进一步提高算法效率具有重要意义。因此，作为衡量搜索深度和搜索广度的基础，本文对掺炼序列评价的分布规律进行了研究。在首先分析比较了仿真优化中参数评价与原油调合调度优化中序列评价的异同之后，本文通过蒙特卡洛仿真法统计分析了原油掺炼序列评价的分布变化规律，并利用威布尔分布对其进行了拟合。文章最后展示了基于威布尔分布描述的2个掺炼序列的评价比较，同时指出威布尔分布的参数拟合问题是接下来研究的重点。

英文摘要：

Crude oil blending and scheduling can help refineries improve the feedstock property, which is important to the smooth operation of oil refining unit, implementation of advanced control and fulfillment of production plan. Under the actual production background, a multi-pipeline crude oil blending and scheduling model is established and a novel solution methodology based on a two-level optimization structure and ordinal optimization is proposed correspondingly Ref.[4-5]. In this solution methodology, a blending sequence and flowrate hybrid optimization problem changes into a blending sequence＇s sampling and evaluation problem, which relates to the breadth and depth of search in algorithm, respectively. How to deal with the balance between them is important to rationally allocate the computation and further improve the optimization efficiency. Therefore, as the basis for measuring search depth and breadth, the distribution of blending sequence evaluation is studied in this paper. In first analyzing the similarities and differences between design evaluation in simulation optimization and sequence evaluation in scheduling of crude oil blending optimization, this paper statistically analyzed the sequence evaluation distribution through Monte-Carlo method and then fitted the statistical results using the Weibull distribution. Finally, the evaluation comparison of two blending sequences based on Weibull distribution description is showed and future direction is pointed out at the end of this paper.