A fresh look at integrated reservoir modelling software

Abstract

Gerard de Jager and Raymond J.W. Pols of JOA Oil & Gas, based in The Netherlands, describe their company’s work in developing a software application suite and reservoir modelling methodology to define, create, and easily maintain up-to-date static and dynamic reservoir simulation models. The Windows .NET based tool is aimed at resolving some of the key bottlenecks in today’s multi-disciplinary reservoir modelling workflow. The main driver for new solutions is the shifting E&P market where production of complex fields and heavy hydrocarbons has become increasingly important. This gives rise to the need for up to date reservoir models with a fast turn around time. The vision of building better and higher quality reservoir models can be realized by no longer focusing on ‘best in class’ for individual parts of the static and dynamic workflow, and instead generating better insight on the total integrated reservoir model, referred to as ‘three models in one’ - the geophysical model, the geological model, and the reservoir simulation model. This is what we have tried to achieve with the JOA Jewel Suite. Modelling complex geology A number of the current industry modelling packages try to solve the 3D gridding of faulted reservoir models by aligning the pillars to the fault in two different ways. The disadvantage of the method shown in Figure 2a is that the lateral sizes of the cells (grid blocks) vary too strongly so that it can only deal with relatively simple fault topologies. In most cases this produces almost vertical branch lines. As a result more complicated fault topologies like y-faults, thrust faults, and x-faults cannot be honoured. The second method shown in Figure 2b can handle both vertical branch lines as well as horizontal branch lines. For horizontal branch lines it creates collapsed cells along the fault surface on which the pillars truncate. However, it creates distorted cells when the fault on which the pillars truncate ends somewhere in the middle of the model. Figure 3a shows an example where Fault A ends and the distorted cells occur because there is no surface to collapse them on. The conclusion of this problem analysis is that the pillar grids based on pillar truncations can handle simple shape fault topologies only. They cannot deal with the more complex topologies of faults like those shown in Figure 3b. The area between the faults indicated with the X cannot be gridded, because the pillars have to reach the top or base of the reservoir. Furthermore, for other complicated geological events, such as salt bags, unconformities, and erosions, pull up of the K-Layers can occur, as shown in Figure 3c, because the pillars are not aligned with the erosion surface. This will also result in distorted cell shapes which will influence the behaviour of the geological and reservoir simulation model in the rest of the workflow.