Database technology for geological modelling of hydrocarbon reservoirs

地下沉积岩的地质特征控制着它们内部存在的石油、天然气和/或水的数量(作为碳氢化合物储藏层或含水层)，以及这些流体将如何流动。石油地质学家建立了三维数值模型，以确定石油和天然气的可能数量和流量以及最佳井位。这些模式最终决定了不同油气生产战略的成败。同样，水文地质学家开发相应的地质模型来预测产水量或污染物运移，以便为含水层开采和清理提供信息。还需要建立这些地质模型来评估地下碳捕获和储存方案的可行性。在建立这些模型时，地质学家只能获得有限的直接地下数据，用来约束地下地质体的类型和几何形状，从而限制它们的流体流动特征。为了补充稀少的直接数据，类似类型岩石的暴露露头，或沉积有可比较沉积物的现代沉积环境，可用作碳氢化合物储集层或含水层的“类似物”。这些类似物提供了关于确定储集层或含水层非均质性的地质特征的代理信息。在储油层或含水层内，这些地质非均质性对油井连通性、流量、生产行为和/或清理策略起主要控制作用，从而决定储油层可能生产多少石油或天然气，或是否成功地从地下水中清除污染物。需要定量的模拟数据作为约束地下地质模型的输入。从数据库中得出这类数据是地下建模工作流程的一个组成部分，但由于现有数据库中存储的数据量和质量有限，而且目前与现有建模工具的集成很差，因此当前的方法是不够的。利兹大学IP由三个不同的关系数据库组成，其中包含关于岩石体积类型的模拟数据，这些数据构成了水库或含水层地质模型的构建块；每个数据库都与特定的地质环境有关。所有数据都以一种可以产生定量输出的格式存储，这种格式可以输入到用于建立地下非均质性模型的所有常见数值方法中。知识产权的技术在数据标准化、数量、质量和格式上都超过了同类数据库。这些数据库实现了对沉积非均质性的更全面的描述，这一事实使现有建模算法所需的输出得以推导：这使得IP在其类别中是独一无二的。然而，由于缺乏与用于生成和管理地下地质模型的软件平台(如斯伦贝谢的Petrel)的集成，IP的当前价值受到限制。目前，没有一个现有的模拟数据库产生的输出格式易于被Petrel中用来生成地质模型的算法所理解。因此，开发数据库与PETREL优化整合的接口是使知识产权可销售的关键要求。接口的理想体现形式是Petrel的“插件”软件组件。成功开发后，IP将使大量高质量数据的访问和应用变得容易：这项技术将帮助碳氢化合物和水管理行业的地质学家和工程师生成地质上敏感的储层和含水层模型。该项目将由卢卡·科伦贝拉负责，他目前是利兹大学的博士后研究助理，并由沉积学家兼河流研究小组主任奈杰尔·芒特尼和比尔·麦卡弗里(沉积学家)监督。

The geological characteristics of subsurface sedimentary rocks control the amount of oil, gas and/or water present within them (as hydrocarbon reservoirs or aquifers), and how such fluids will flow. Petroleum geologists build three-dimensional numerical models to ascertain the likely amount and flow rates of oil and gas and optimum well locations. These models ultimately determine the success of different strategies of hydrocarbon production. Similarly, hydrogeologists develop corresponding geological models to predict water yield or contaminant transport, in order to inform aquifer exploitation and clean-up. The creation of these geological models is also required to assess the feasibility of programmes of underground carbon capture and storage. When these models are built, geologists have available only limited direct subsurface data with which to constrain the type and geometry of subsurface geological bodies with and thus their fluid-flow characteristics. To complement the sparse direct data, exposed outcrops of similar types of rocks, or modern sedimentary environments where comparable sediments are deposited, can be used as 'analogues' to hydrocarbon reservoirs or aquifers. These analogues provide proxy information regarding geological features that determine reservoir or aquifer heterogeneity. Within a reservoir or aquifer, these geological heterogeneities exert a primary control on well connectivity, flow rates, production behaviour and/or to clean-up strategies, thereby dictating how much oil or gas is likely to be produced from a reservoir, or whether contaminants are successfully removed from groundwater. Quantitative analogue data are required as input for constraining geological models of the subsurface. The derivation of this type of data from databases is an integral part of subsurface modelling workflows, but current approaches are inadequate because of the limited volume and quality of data stored in existing databases, and their current poor integration with existing modelling tools.The Leeds University IP consists of three different relational databases that contain analogue data about types of rock volumes that constitute the building blocks of geological models of reservoirs or aquifers; each database relates to a particular geological setting. All data are stored in a format that allows quantitative output to be produced, in forms that can be fed into all the common numerical methods used to build models of subsurface heterogeneity. The technology of the IP surpasses similar databases in terms of data standardisation, quantity, quality and format. The fact that a fuller characterisation of sedimentary heterogeneity is achieved by these databases enables the derivation of the output required by existing modelling algorithms: this makes the IP unique in its class. However, the current value of the IP is limited by the lack of integration with software platforms employed to generate and manage geological models of the subsurface, such as Schlumberger's Petrel. At present no existing analogue database produces output in a format readily digestible by the algorithms used in Petrel to generate geological models. Thus, the development of an interface for the optimal integration of the databases with Petrel is a key requirement for making the IP marketable. The ideal incarnation of the interface would be in the form of a Petrel 'plug-in' software component.Upon successful development, the IP will enable easy access and application of large volumes of high-quality data: this technology will aid geologists and engineers in the hydrocarbon and water-management industries in the generation of geologically sensible reservoir and aquifer models. The project will be undertaken by Luca Colombera, currently a post-doctoral research associate at Leeds, and supervised by Nigel Mountney, who is a sedimentologist and director of the Fluvial Research Group, and Bill McCaffrey (sedimentologist).