摘要
针对常规叠前AVO反演存在强烈依赖于初始模型、易陷入局部最优值等问题,对基本遗传算法进行了自适应改进,然后将改进遗传算法与粒子群算法相结合,发展了遗传—粒子群算法混合的GA-PSO协同进化智能优化算法;对比改进遗传算法、粒子群算法及GA-PSO协同进化算法反演的理论模型合成地震记录的纵波速度、横波速度及密度,表明后者具有精确的反演结果及更强的稳定性和抗噪能力;最后利用GA-PSO协同进化算法对实际地震数据进行叠前AVO非线性反演,验证了算法的应用效果和适用性。
As the conventional prestack AVO inversion is dependent on the initial model and easily trapped into a local optimal solution,we propose a nonlinear AVO inversion based on the hybrid intelligent optimization algorithm.First we improve adaptively the conventional genetic algorithm.Then combining the improved genetic algorithm and particle swarm algorithm,we put forward a hybrid GA-PSO co-evolution algorithm.After that we apply the improved genetic algorithm,the particle swarm algorithm,and the GA-PSO co-evolution algorithm to model synthetic data.With the comparison of inverted P-wave velocity,shear wave velocity and density,the GA-PSO co-evolution algorithm shows better inversion result,bet-ter stability,and better anti-noise ability than the other two algorithms.In the end,AVO nonlinear inversion with the proposed algorithm is used to real data,and the results confirm its effectiveness and applicability.
出处
《石油地球物理勘探》
EI
CSCD
北大核心
2017年第4期797-804,共8页
Oil Geophysical Prospecting
基金
国家重大科技专项子专题"莺琼盆地高温高压天然气富集规律与勘探开发关键技术"(2016ZX05024-005)
中国地质科学院物化探研究所基本科研业务费项目(WHS201308)
2016年"中央高校基本科研业务费"新青年教师计划等联合资助
关键词
遗传算法
粒子群算法
混合智能优化算法
非线性AVO反演
genetic algorithm
particle swarm optimization
hybrid intelligent optimization
nonlinear AVO inversion
