一种混合优化算法在试井解释中的应用  

Application of a Mixed Optimization Method in Well Test Interpretation

作　　者：林加恩[1] 李霞[2] 贾海军[3] 

机构地区：[1]西安石油大学石油工程学院,陕西西安710065 [2]冀东油田分公司开发技术公司 [3]西北工业大学动力与能源学院,陕西西安710072

出　　处：《油气井测试》2010年第5期1-4,共4页Well Testing

摘　　要：为实现试井资料的自动拟合分析,保证解释结果的唯一性,提出一种改进的非线性回归计算方法,与二进制自适应遗传算法相结合组成新的混合优化算法。应用该混合算法分别对均质无限大和均质变井储模型进行试井分析,将计算结果与单纯遗传算法计算结果进行比较。结果表明,混合优化算法计算精度高、效率高,结果准确、稳定。In order to match and analyze well test data automatically and guarantee the uniqueness of interpretation results,a modified non-linear regression method is proposed;combining this method with binary self-adaptation genetic calculation forms a new mixed optimization algorithm.Well test interpretation for both normal wells and changing wellbore storage wells in infinite homogeneous reservoir with this new mixed method shows that this method has same advantages such as having higher accuracy,higher efficiency and obtaining more correct and stable results etc.than genetic calculation method.

关 键 词：非线性回归分析 遗传算法 迭代算法 数值反演 试井分析 

分 类 号：TE353[石油与天然气工程—油气田开发工程]