节点文献
基于粗糙集理论和支持向量机的套管损坏动态预报方法
Dynamic prediction method of casing damage based on rough set theory and support vector machine
【摘要】 针对油水井套管损坏受地质、工程综合因素影响,具有不确定性、模糊性和时变性的特点,提出基于粗糙集与支持向量机集成的套损动态预报方法。通过粗糙集理论对套损影响因素进行属性约简,确定导致油水井套管损坏的主导因素,进而应用支持向量机算法建立套管损坏动态预报模型,从而克服传统支持向量机无法处理动态数据以及大样本条件下易出现维数灾难的缺点。榆树林油田的现场应用结果表明,该方法具有较高的预报精度,预报吻合率达72.7%,可为油田现场调整开发参数、尽早制定套管防护措施提供指导。
【Abstract】 Aiming at casing damage problem affected by both geological factors and engineering factors,with characteristics of uncertainty,fuzziness,time-varying,a dynamic prediction method of casing damage was proposed based on rough set integrated with support vector machine.The influencing factors for casing damage were reduced by using rough set theory and the main factors causing casing damage were obtained.Then prediction models were established based on support vector machine,and the problem that the traditional support vector machine can not deal with dynamic data and is prone to dimension disasters with large samples,was avoided in this method.The field application in Yushulin Oilfield indicates that the prediction consistent rate is 72.7% in this method,which provides guidance for oilfield adjusting development parameters and developing casing protection measures early.
【Key words】 casing damage; rough set; attribute reduction; support vector machine; dynamic prediction;
- 【文献出处】 中国石油大学学报(自然科学版) ,Journal of China University of Petroleum(Edition of Natural Science) , 编辑部邮箱 ,2010年06期
- 【分类号】TE931.2
- 【被引频次】12
- 【下载频次】185