论文标题
水文地质模型的多点统计模拟：3-D培训图像开发和调节策略
Multiple-point statistical simulation for hydrogeological models: 3-D training image development and conditioning strategies
论文作者
论文摘要
大多数基于多点统计（MP）在水文地质模型中应用地统计模拟的研究的研究集中在相对细大的模型以及相应级别的结构不确定性的估计上。对输入数据和培训图像的构建（TIS）的关注较少。例如。即使TI应捕获一组空间地质特征，但大多数研究仍然依赖于2D或准3D训练图像。在这里，我们展示了一种以（i）现实3D TI和（ii）为特征的3D MPS建模的新型策略，以及（ii）合并一组地质和地球物理数据集的有效工作流程。该研究涵盖丹麦南部的2810 km^2。 MPS模拟是在地质继承的子集（中年至中期沉积物）的子集上进行的，该子集的特征是相对均匀的结构，并以沙子和粘土为主。模拟域很大，每个地统计实现都包含大约45 x 10^6素，尺寸为100 m x 100 m x 5 m。用于建模的数据包括水井原木，地震数据和先前发表的3D地质模型。我们根据数据质量应用了一系列不同的策略，并开发了一种新颖的方法来有效地创建观察到的空间趋势。 Ti被构造为相对较小的3D体素模型，覆盖90 km^2的面积。我们使用迭代培训图像开发策略，发现即使在TI中进行了轻微的修改也会在模拟中产生重大变化。因此，本研究表明了如何同时包括地质环境和输入信息的类型和质量，以实现MPS建模的最佳结果。我们提出了一个实用的工作流程来构建TI并有效处理不同类型的输入信息以执行大规模的地统计建模
Most studies on the application of geostatistical simulations based on multiple-point statistics (MPS) to hydrogeological modelling focus on relatively fine-scale models and on the estimation of facies-level structural uncertainty. Less attention is paid to the input data and the construction of Training Images (TIs). E.g. even though the TI should capture a set of spatial geological characteristics, the majority of the research still relies on 2D or quasi-3D training images. Here, we demonstrate a novel strategy for 3D MPS modelling characterized by (i) realistic 3D TIs and (ii) an effective workflow for incorporating a diverse group of geological and geophysical data sets. The study covers 2810 km^2 in southern Denmark. MPS simulations are performed on a subset of the geological succession (the lower to middle Miocene sediments) which is characterized by relatively uniform structures and dominated by sand and clay. The simulated domain is large and each of the geostatistical realizations contains approximately 45 x 10^6 voxels with size 100 m x 100 m x 5 m. Data used for the modelling include water well logs, seismic data, and a previously published 3D geological model. We apply a series of different strategies for the simulations based on data quality and develop a novel method to effectively create observed spatial trends. The TI is constructed as a relatively small 3D voxel model covering an area of 90 km^2. We use an iterative training image development strategy and find that even slight modifications in the TI create significant changes in simulations. Thus, this study shows how to include both the geological environment and the type and quality of input information in order to achieve optimal results from MPS modelling. We present a practical workflow to build the TI and effectively handle different types of input information to perform large-scale geostatistical modelling
