基于局部加权映射回归的合采井产能预测算法

An Algorithm for the Prediction of Productivity Contribution in Commingled Oil Wells Based on Locally Weighted Projection Regression

在油田开发中,为了节约成本,优化生产,多采用几个油层混合开采。在混合开采过程中,为了使油井处于合理有效的生产状态,必须了解单个油层对混采油的产量贡献。利用原油色谱烃技术,提出了一种基于局部加权映射回归和粒子群算法的预测方法,并采用原油配比实验进行了验证。实验结果表明,该方法具有一定的可行性,与现有的预测方法比较,提高了预测的准确率。

During the exploitation of oil fields,multi-layer hybrid mining methods are employed for the sake of reducing costing and optimizing productivity.In order to insure the efficiency and effectivity of the commingled oil well exploiting,the productivity contribution of every single layer must be acquainted.Using the technology of crude oil chromatography fingerprint,an algorithm for predicting productivity contribution in commingled oil wells based on locally weighted projection regression and particle swarm optimization is proposed.The validity of the method is proved by laboratory artificial experiments.It is shown by the experimental result that the proposed method is viable.Comparing with existing methods,the proposed method improved the precision of the prediction.