应用混合微粒群算法优化设计热油管道

APPLICATION OF PSO IN OPTIMAL DESIGN OF HOT OIL PIPELINE

考虑到年输量随油田产量和市场需求的影响而随机变化的实际,采用最佳平方逼近法确定热油管道的设计输量,在此基础上建立了包含年输量模型和参数优化模型的热油管道优化设计两级递阶模型。用微粒群算法和混合离散变量随机搜索法构成的混合微粒群算法解热油管道的参数优化模型,实现了热油管道的整体优化设计。算例表明,该综合算法优化设计热油管道得到的方案比单一采用离散变量随机搜索法、离散变量复合型法和基本的微粒群算法计算得到的方案更能节约年费用,且考虑年输量随机变化的管道设计方案更符合热油管道运行的工程实际。

The annual throughout is often not equal to the designed annual throughout, it changes randomly in actual operation, optimal square approximation is used to identify the pipeline's designed throughout. Based on it, the authors establishe the optimal designed two-level hierarhical model of hot oil pipeline including annual throughout model and optimal parameter model. Combining PSO (Particle Swarm Optimization) and MDRP ( Mixed Discrete Random Programming ) forms the comprehensive PSO-MDRP, which can be applied to actualize the whole optimization design of hot oil pipeline. The calculation examples show that PSO-MDRP can save more annual expense compared with PSO, MDRP and MDCP (Mixed Discrete Complex Programming)for the same design. It also shows using PSO-MDRP improves the design quality.