地震储层预测中贝叶斯反演方法的研究

Research on Bayesian Inversion Method Applied to Predict Seismic Reservoirs

贝叶斯反演方法属于不确定性反演方法,它是以贝叶斯理论为基础,结合随机反演算法构造出的一种新反演方法。该方法能有效解决传统反演方法在应用中的两大问题,其一是难以联合应用不同类型、不同精度的数据约束反演过程,另一个是无法定量描述反演结果的不确定性。研究首先详细分析了贝叶斯反演的原理,并提出了以地震资料进行贝叶斯反演的实现方法;其次建立了塔里木碳酸盐储层典型模型开展贝叶斯反演,结果表明该方法较传统随机反演方法具有一定的优越性;最后还采用实际资料证明了贝叶斯反演方法的有效性。

Bayesian Inversion Method belonged to one of the uncertainty inversion methods.It was a new inversion method based on Bayesian theory and stochastic inversion methods.In practical application,this method can effectively used to solve two limitations that existed in conventional inversion methods.One was not easy for constrain inversion and simultaneously data inversion with different precision information from borehole; the other was difficult to quantify the uncertainty of the estimated parameters.In this study,the principle of Bayesian Inversion Method was analyzed in detail and it was proposed that Bayesian inversion was implemented by using seismic data,Then a model was established based on the characteristics of reef community in carbonate reservoirs in Tarim Basin.The results show Bayesian inversion method is more available than conventional stochastic inversion methods.At last,it is proved that Bayesian inversion method is effect for practical data inversion.