基于混合优化匹配追踪算法的礁滩储层流体识别方法

基于匹配追踪算法(MP)的流体活动性检测是礁滩储层油气识别的一种重要手段。针对常规匹配追踪算法效率低、多解性强的缺点,将粒子群优化算法与BFGS算法相结合的混合优化算法引进到匹配追踪算法中。混合优化匹配追踪算法(HO-MP)既具有粒子群优化算法良好全局优化特性,又具有BFGS方法局部加速收敛速度的优点,通过一次性计算描述Morlet小波的五个参数,可以解决匹配追踪过程中的参数寻优问题,提高搜索最优解的精确度和效率。用混合优化匹配追踪算法对信号进行分解,分解后的理论信号和实际地震资料各个信号分量在时频平面的位置精确,具有良好的时频聚集性,搜索速度显著加快。对哈萨克斯坦楚-萨雷苏盆地Tam研究区流体检测结果表明HO-MP可以有效指示礁滩储层含油气的变化,说明该方法在礁滩储层流体预测中有很好的应用前景。

Fluid identification method for reef-beach reservoir based on hybrid optimization matching pursuit algorithm

Fluid activity detection based on matching pursuit algorithm(MP)is an important means of oil and gas identification for reef-beach reservoirs.Aiming at the disadvantage of low efficiency and strong multiplicity in the traditional matching pursuit algorithm,the hybrid optimization algorithm based on particle swarm optimization(PSO)and BFGS is introduced into the matching pursuit algorithm.The hybrid optimization matching pursuit algorithm(HO-MP)has not only the advantages of good global optimization of particle swarm optimization algorithm,but also the local acceleration convergence speed of the BFGS method.The one-time calculation can describe five parameters of the Morlet wavelet,which solve the problem of parameter optimization in the matching pursuit process,and improve the accuracy and efficiency of searching the optimal solution.Using the proposed method,the theoretical signal and the actual seismic data are decomposed.Each signal component after the decomposition is accurate in the time-frequency plane,has good time-frequency aggregation,and greatly accelerates the search speed.In Tam study area of Chu-Sarysu Basin in Kazakhstan,the results of fluid tests show that HO-MP can effectively indicate the change of hydrocarbon-bearing gas in reef bank,indicating that this method has a good application prospect in the prediction of reef beach reservoir fluid.