Geological reservoir modelling of fluvial channel sands
Fluvial channel sandstones form some of the major hydrocarbon reservoirs throughout the world. They are also the most difficult for reservoir geologists and engineers to deal with, because of the variety of sedimentary environments and complex sand body geometries and relationships which directly contribute to the reservoir quality and distributions. Successful geological reservoir modelling of fluvial channel sandstones involves proper understanding of the geological background of the target area, choosing a suitable method for each particular modelling purpose, and making available appropriate well data and analogue data. It is also the key to providing reservoir engineers with essential information for further oil exploitation and production activities. Two 3D modelling methods for fluvial channel sandstone reservoirs have been explored and contrasted in this Ph.D. project: Vertical Sequence Generation (VSG) and Geometry Based Generation (GBG). Computer software had to be written specially for each as it is not available, commercially or otherwise. Markov stochastic theory is employed in VSG in two aspects of geological reservoir characterisation. The first is to analyse the vertical statistical transition characteristics of different reservoir types or geological units of interest; and the second is to generate pseudo-vertical sequences of unsampled reservoirs by using the obtained Markov transition properties. This is a further development of the use of Markov statistics in geological description, and is a new approach to geological reservoir modelling developed in this Ph.D. project. The GBG method takes channel sand body dimensions and orientation into consideration and generates each channel sand body, three-dimensionally in the area of interest, with its location, thickness, width, length and orientation.