Abstract Conventional geological modelling methods are not capable to provide precise and comprehensive model of the subsurface structures, when dealing with insufficient data. Knowledge based methods employing rule bases techniques are found vast applications in geoscience studies. These methods are applicable for petroleum reservoir geological modelling and characterizations, specifically for geologically complex structures. In this study, we present a knowledge based seismic acoustic impedance inversion method which employs rule based method for porosity estimation. The back propagation algorithm and the fuzzy neural network are also used in the methodology for parameter optimization and definition of nonlinear relationship between seismic attributes and porosity of the reservoir rock. The methodology initiates by seismic acoustic impedance inversion, followed by conventional porosity estimation. Subsequently, a knowledgebase was designed by investigation on more than 24 published case studies. This knowledgebase was used for definition of rules and optimization number of rules and improve efficiency of the inference engine. The porosity model obtained by conventional method in previous step would be used for primary evaluation of the rules. The extracted rules and optimized number rules then would be used for rule-based porosity estimation. The methodology was applied on a petroleum field containing two heterogeneous reservoir formations. Result of application of the proposed approach was evaluated with core analysis, thin sections and drilling data. Consistency of result obtained by the proposed method with geological data has shown its capability to resolve problem of insufficient data in reservoir geological modelling.

中文翻译：

---

基于知识的地震反演与非均质储层沉积学调查的集成

摘要 传统地质建模方法在处理数据不足的情况下，无法提供精确、全面的地下构造模型。采用规则库技术的基于知识的方法在地球科学研究中得到了广泛的应用。这些方法适用于油藏地质建模和表征，特别是地质复杂的结构。在这项研究中，我们提出了一种基于知识的地震声阻抗反演方法，该方法采用基于规则的方法进行孔隙度估计。反向传播算法和模糊神经网络也被用于参数优化和地震属性与储层孔隙度非线性关系定义的方法中。该方法从地震声阻抗反演开始，然后是常规孔隙度估计。随后，通过对超过 24 个已发表案例研究的调查设计了一个知识库。该知识库用于定义规则和优化规则数量，提高推理引擎的效率。将上一步通过常规方法获得的孔隙率模型用于规则的初步评估。提取的规则和优化的数字规则然后将用于基于规则的孔隙度估计。该方法应用于包含两个非均质油藏地层的油田。使用岩心分析、薄片和钻井数据评估了所提出方法的应用结果。