集输管网规模庞大,枝状管网结构复杂,以往多采用分级优化算法进行优化。全局优化算法多集中于拓扑结构简单的星状管网、小规模枝状管网、站场或参数优化。为解决大规模枝状管网的全局优化问题,建立了枝状集输管网全局优化模型,提出一种生成管网初步连接图的方法,对传统单亲遗传算法的编码方式、遗传算子及遗传操作手段进行了改进,并对我国某油田的集输管网进行了全局优化。通过实例计算证明,改进后的单亲遗传算法的优化速度及效果都显著提高,真正实现了对大规模枝状集输管网的全局优化,达到了费用最省的目的。

Hierarchical optimization algorithm was normally adopted to the optimization of branched gathering pipeline network for its large scale and complex layout, while global optimization algorithm was used for astroid pipe network, small-sized branched pipe network, station or parameter optimization. In this paper, a model of global optimization was developed for large-scale branched pipeline; a method of initial pipeline network was developed; the encoded mode and genetic operators and operating methods were improved; and the improved parthenogenetic algorithm was used in the optimization of a domestic gathering pipeline network. The result showed that the optimization speed and effect of improved parthenogenetic algorithm were significantly improved, and the improved parthenogenetic algorithm was adapted well to the optimization of large scale branched gathering pipeline network.