Rock and Fluid data provide critical input parameters for reservoir characterisation, production, and reserve forecasts. These datasets are key building blocks for reservoir simulation models, but there has been limited work in the oil industry on the standardisation of their data structures, formats, or quality control.
We will present experiences and results from an ongoing digitalisation initiative. This initiative centres on establishing a database that includes all relevant analysis done on rock and fluid data. The database currently has thousands of wells covering over a hundred NCS fields that can be easily accessed via standard interfaces. As the data are contextualised on e.g., lithostratigraphic units, end-users can analyze data by groups, formation, geological time, and depositional environment.
A key challenge for rock and fluid data compared to simpler oil industry datasets, is that one discipline's analysis and modelling results represent another discipline's input data. We will present an approach to resolve these challenges and discuss how this relates to ongoing joint standardisation efforts in the industry such as the OSDU Forum.
In this paper we will show how industrial DataOps technology and close collaboration between domain experts, data managers and technologists has unlocked great value based on Rock and Fluid data. We will describe the workflow of finding and verifying data, establishing standard formats, implementing automated data validation and ingestion, and how we manage, visualise, and interpret data, ensure full data integrity, and make data available for end-users through different applications.
We will provide examples on how machine learning can be used for automated trend analysis and identification of relationships across different data types to support model generation.
Finally, we will demonstrate a state of the art and seamless workflow of generating static and dynamic SCAL model input for reservoir simulation with uncertainty band. This includes evaluation and verification of both static and dynamic SCAL data. Relative permeability and capillary pressure curves from each laboratory experiment are quality assured, parameterised and stored in the database, along with all relevant information such as plug properties, well name, fluid properties, experimental conditions, flow parameters and endpoint saturations etc.