Objectives and problems to be solved: The principal purposes of the project are: - to quantify objectively the sensitivity of geological uncertainty on production forecasts, as a function of generic aspects of both the sediment logical architecture and faulted structure of classic hydrocarbon reservoirs, - to validate these results using real-case reservoir and production data. Because of the case-specific nature of existing production forecasting sensitivity studies, links between geological and production uncertainty cannot be made at present. These links are a prerequisite for early recognition of the most significant geological parameters influencing production-forecasting uncertainty and are a necessary basis for establishing optimal methods for including geological uncertainty in reservoir modelling studies. Description of the work: The project is a systematic assessment of uncertainty in reserves and production estimates within an objectively defined geological parameterisation encompassing the majority of European classic oil reservoirs. A broad suite of shallow marine sediment logical reservoir types are indexed to continuously varying 3D anisotropy and heterogeneity levels. Structural complexity ranges from unsalted to compartmentalize, and fault types from Tran missive to sealing. Several geostatistical realizations each for the geologically diverse reservoir types covering the pre-defined parameter-space are up-scaled, faulted and simulated with an appropriate production strategy for an approximately 20 year period. The production and reserves uncertainty associated with geological uncertainty and methodological imprecision are quantified as a function of the underlying geology. Production results and recovery factors are combined with the geological and development plan parameters in a relational database, allowing the levels and origins of production uncertainty to be defined within the full parameter-space. Existing and new static and dynamic heterogeneity measures and dimensionless parameters are tested against production results for their ability to discriminate between geological architectures and to predict production characteristics. Sensitivity analyses, performed using reservoir and production data from three North Sea fields, combined with results from the large suite of geological models test our principal technical findings. Expected results and exploitation plans: The expected project results are: - Quantification of the relative and absolute influences of sediment logy, structure and up-scaling on reserves estimation and production forecasting from reservoirs with different sediment logical and structural properties- Definition of geologically relevant dynamic and static heterogeneity measures and dimensionless groups for improved production forecasting in faulted classic reservoirs. Project results will be tested and implemented through the consulting and software roles of the two service companies and through the practical implementation of the methods developed by the two oil company partners. These results and methods are expected to contribute towards an improvement in the planning and execution of geological reservoir modelling programs, and towards a reduction of the economic risk associated with field development.
We have generated over 35,000 full field reservoir production profiles from synthetic shallow marine reservoirs encompassing realistic ranges of small and large scale sedimentological variability, structural variability and different upscaling approaches applied to both grid-blocks and faults, using a variety of different field production plans. Results have permitted an assessment of the sources and levels of uncertainty in production forecasts in a generic suite of model reservoirs. A number of mainly structural heterogeneity measures have been identified which approximate relevant aspects of the reservoir geology as transportable dimensionless parameters with predictive capabilities. The secondary objective of validating these results using the natural field examples has been hampered by data formatting issues and the quality of pre-existing reservoir models, but some encouraging results backing conclusions from the synthetic modelling were obtained. The sensitivity analyses, which have included models generated with the more permeable fault property predictors, indicate that variability in sedimentary variables at both the small and large scale dominate uncertainty in reserves, but that structural aspects of the reservoirs are as important as sedimentological ones in terms of recovery factors. Of the large-scale sedimentological parameters the aggradation angle is particularly significant.
This variable is seldom taken into account when planning production. The relative orientations of the progradation and waterflood directions are also important. Both the barrier strength and the shoreline curvature have proven less influential than anticipated. The choice of lamina-scale properties used in the upscaling is a very important control on reserves, and there is therefore a large potential for reducing the uncertainty in production forecasts by placing greater efforts into acquisition of special core analysis data, and into the generation of the relative permeability and capillary pressure curves used. The actual algorithms used to perform the upscaling do not appear to introduce uncertainty, and nor does the inclusion or omission of anisotropic curves. The flow rate assumed during pseudoisation is moderately important. The analyses indicate that the principal way of reducing production uncertainty relating to structural issues is to place greater efforts into 3D fault mapping, since fault juxtapositions can have a significant effect on production irrespective of fault rock permeabilities. If a fault system is disconnected, fault permeabilities are influential on production uncertainty only over a two order of magnitude range. Small-scale fault segmentation can be particularly significant in compartmentalised systems of low permeability faults. The project results provide objective methods for the assessment of the risk associated with field recovery as a function of known and (quantifiably) unknown geological reservoir characteristics. Application of the methods developed could contribute to the decision of how to maximise reserves in particular fields, or, indeed, whether or not to invest in a particular field. Since field development costs are many tens of million euro, the potential value to an oil company of a correct decision are far in excess of the cost of the project.
We have demonstrated that, using state of the art computing and software, it is possible to build and flow simulate tens of thousands of parametrically distinct full-field simulation models at a cost ca. €50 per model and on a time-scale comparable with the appraisal period of a reservoir. We recommend therefore that sensitivity modelling programs on this scale and aimed at specific reservoir developments become standard practice for both minimising and understanding the uncertainty on production forecasts associated with specific geological uncertainties.
Funding SchemeCSC - Cost-sharing contracts
PE23 5NB Spilsby, Lincolnshire
RG6 1PT Reading
EH14 4AS Edinburgh
SW7 2BP London
3508 TA Utrecht
2501 CR Den Haag ('S-gravenhage)
L69 7GP Liverpool