遗传算法用于蒸汽吞吐井注采参数的优化设计

Application of Genetic Algorithm for Optimal Design of Injection-production Parameters for Steam Soak Wells

建立了一套蒸汽吞吐注采参数快速整体优化设计的方法,并编制了相应的软件。该方法以周期采油量计算为基础,以周期平均日收益最大化为优化目标,以遗传算法为手段,对注汽压力、注汽干度、注汽速率、周期注汽量、焖井时间及井底流压进行整体优化设计。鉴于遗传算法快速搜索、整体寻优的特点,该方法可以快速全局优化蒸汽吞吐的注采参数。用该方法对辽河油田冷41稠油区进行注采参数优化设计,并与现场实际资料对比,效果明显。

A set of methods are established for an overall optimization of injection-production parameters of a steam soak and a corresponding software is developed. In the methods cyclic production rate is used as a foundation, cyclic maximum daily average revenue is used as an optimized object and the genetic algorithm is used as a tool for an optimized design of steam injection pressure, dryness fraction of steam, steam injection rate, cyclic injection volume, soak time, and bottomhole flow pressure. In consideration of the characteristics of rapid search, overall optimization of the genetic algorithm, the methods are used for a rapid global optimization of injection-production parameters of steam soak. They are used for optimizing the injection-production parameters for Viscous Oil Region Len 41 in Liaohe Oilfield. The result shows that the effect is obvious by comparing with field data.