Using object-oriented databases to model hydrocarbon reservoirs
Hydrocarbon reservoir modelling plays a central role in the exploration and production of reservoirs. This thesis describes significant improvements in the way that data and knowledge about reservoir modelling is used in comparison with other attempts at such modelling. This thesis describes the development of a flexible reservoir modelling environment which allows users to apply their own knowledge in order to influence the modelling process. The modelling environment allows users to explore their own 'What if ...' hypotheses. A Smalltalk/V functional data model has been used as a front end to an object-oriented database (P/FDM). A technique called transparent object migration has been developed which allows objects from P/FDM to be reconstructed in the Smalltalk/V functional data model. The thesis describes how users can configure their own stochastic modelling algorithms. Common stages in stochastic modelling algorithms have been isolated and a number of alternatives developed for each stage. This has been implemented in an object-oriented architecture which allows users to configure their own algorithms from the re-usable parts supplied. A graphical probability distribution function editor has been developed. This provides a graphical representation of a probability distribution function which can be easily modified by the user, without the user having to provide a quantitative description of the changes that are required. By using this technique as part of a feedback loop, the user can develop reservoir models which they have more belief in.