摘要
地震优化非线性反演方法是一种集遗传算法和人工神经网络技术的优势于一体的新技术,它采用混合智能优化学习方法,这种优化学习方法是将BP算法作为一个算子嵌入到自适应遗传算法中,以概率Pbp的方式进行搜索运算,从而,快速而精确地找到全局最优解。在嘉陵江组的实际反演研究中,由井点出发构造测井资料与井旁地震道的非线性映射关系,并自适应地更新这种非线性映射关系和自适应外推,以获得嘉陵江组的高分辨率反演剖面,这种高分辨率剖面清晰而详细地反映出嘉陵江组在纵横方向上的变化特征,有效识别这种薄储层,获得了很好的储层预测效果,并为钻井所证实,在嘉陵江组储层中获得高产工业气流。因此,地震优化非线性反演方法具有广阔的应用前景。
The method of seismic optimized non-linear inversion consists of genetic algorithm and artificial nerval network and employes mix intelligence optimum learning method.It method uses BP′s algorithm as an operator in genetic adaptive algorithm and search with probability PBP mode. So the global optimal solution is obtained quickly.Non-linear mapping is constructed from well loggin and seismic trace for Jialing river Formation and high-resolution inversion section is obtained from the inversion method. The high-resolution reservoir parameter shows clearly the varying characters of Jialing river Formation and can identify effectively the thin reservoir in Jialing river Formation.Practical drills comfirm the effect of reservoir prediction and high production gas is obtained in Jialing river Formation.Therefore the method of seismic optimized non-linear inversion will have prospect of wide application.
出处
《矿物岩石》
CAS
CSCD
北大核心
2004年第1期101-104,共4页
Mineralogy and Petrology
关键词
遗传-神经网络
优化非线性反演
混合算法
映射
genetic-nerval network
optimized non-linear inversion
mix-algorithm
mapping
