基于特征识别与PSO结合的地层对比算法-《计算机技术与发展》

文章信息/Info

Title:: A Stratigraphic Correlation Algorithm Based on Characteristics Identification and PSO

Keywords:: ; stratigraphic correlation; process characteristics; Particle Swarm Optimization ( PSO) ; pattern matching; sequence constraint

摘要:: 针对油田开发区块相邻两井之间的小层对比问题，提出了一种以小层测井曲线直接作为过程特征并与粒子群优化相结合的地层对比算法。通过定义可度量曲线特征相似度的距离指标，直接将相邻两口对比井的原始测井曲线作为小层对比特征变量，以此构建带多级罚函数约束的小层对比目标函数。综合利用粒子群算法( PSO)的经验记忆、信息共享和遍历搜索性质优化目标函数参数，实现两井小层之间的最佳匹配。算法可突出测井曲线的过程特征以及多条曲线的整体组合形态，提高对比井段小层之间对应关系的辨识和匹配能力，对于地层对比问题在机制上具有较好的适应性。对油田开发区块相邻两井进行实际资料处理，小层自动对比结果验证了模型和算法的有效性。

Abstract:: For small layer contrast problem of oilfield development block between two adjacent wells,propose an algorithm with small lay-ers logging curve directly as identification characteristics and combined with particle swarm optimization algorithm in stratigraphic correla-tion. By defining the generalized Hausdorff distance measured curve feature similarity,build small layers compared objective function with multistage penalty function constraint,directly taken the original logging curves of adjacent two contrastive wells as the characteristic vari-ables of detail correlation of reservoir bed. Use the experience memory,information sharing and track traverse searching from PSO for sol-ving global optimization objective function,achieving the optimal matching of small layer between two wells. Algorithm can highlight the well logging curve process characteristics and combination of integral multiple curve form,improve the recognition and matching of corre-sponding relationship between small layers of contrast wells,which has better adaptability for stratigraphic correlation problem in the as-pect of mechanism. Conducting the actual processing for oilfield development block adjacent wells,the results show the effectiveness of the model and algorithm.