Building a representative static model for predicting and monitoring the performance of coal-seam gas (CSG) fields presents several complex and unique challenges. The individual reservoirs possess different coal architectures, often with highly complex seam splitting, amalgamating, and structural deformation. Our objective was to develop an alternative approach that honored log and core data capturing both the lateral heterogeneity and the vertical signature of the Bowen Basin coals in Central Queensland.

In some areas of the Bowen Basin, coals can be thick and laterally continuous; picking the top and base of each seam works well in small models with homogeneous coals. As seam geometries begin to increase in complexity and coals become more heterogeneous in nature with thinner seams in multiple packages, then a net-to-gross (NTG) approach is often more appropriate. Each method has its merits. The former approach describes the reservoir architecture but implies a certain degree of confidence in coal correlation. In a vast field with complex seam splitting and merging with abundant drilling data, it might not be a practical technique. The NTG method disregards coal-seam architecture and reservoir connectivity.

The proposed workflow is a hybrid approach using discrete modeling algorithms and data-analysis tools on continuous NTG properties. The process operates on a relatively coarse chronostratigraphic framework in which coal is captured as contiguous discrete-NTG “facies.” The use of the truncated Gaussian model ensures the ordering of NTG facies and mimics transitions between coals and the surrounding interburdens. With the adoption of facies vertical proportion trends, we are able to replicate a similar coal-seam signature laterally away from the wellbore. The definition of a categorical coal model allows the proper scaling of seams with different coalquality characteristics.

With the successful geocellular-model construction and history match of two historical CSG fields in the Bowen Basin, the discrete- NTG truncated Gaussian simulation (TGS) approach has proved to be a valid alternative CSG-modeling technique.

建立用于预测和监视煤层气（CSG）油田性能的代表性静态模型提出了一些复杂而独特的挑战。各个储层具有不同的煤构造，通常具有高度复杂的煤层分裂，合并和结构变形。我们的目标是开发一种替代方法，以尊重记录和采集昆士兰州中部Bowen盆地煤的横向非均质性和垂直特征的岩心数据。

在博文盆地的某些地区，煤可能很稠，而且横向连续。在使用均质煤的小型模型中，挑选每个煤层的顶部和底部效果很好。随着煤层几何结构的复杂性开始增加，并且煤在本质上变得越来越不均一，且多个包装中的煤层更薄，因此通常采用净高总比（NTG）方法。每种方法都有其优点。前一种方法描述了储层构造，但暗示了对煤相关性的一定程度的信心。在广阔的领域中，复杂的煤层分裂和大量的钻探数据合并，可能不是一种实用的技术。NTG方法忽略了煤层构造和储层连通性。

所提出的工作流程是一种使用离散建模算法和连续NTG属性数据分析工具的混合方法。该过程在相对粗糙的年代地层框架上运行，在该框架中，煤被捕获为连续的离散NTG“相”。截断的高斯模型的使用确保了NTG相的有序性，并模拟了煤与周围突岩之间的过渡。通过采用相垂直比例趋势，我们能够从井眼的侧面横向复制类似的煤层特征。分类煤模型的定义允许对具有不同煤质特征的煤层进行适当缩放。

借助博恩盆地两个历史上的CSG场的成功的地理细胞模型构建和历史匹配，事实证明，离散NTG截断的高斯模拟（TGS）方法是一种有效的替代CSG建模技术。