Technology from Europe:
Norway

Tackling reservoir uncertainty improves project risk factors

David Hardy, Roxar AS, Stavanger

Reservoir uncertainty is central to any development project, whether it is spatial facies distribution, the existence and location of sub-seismic faults, or the question as to whether there is oil in a portion of the reservoir.

There are many different sources of uncertainty, including data measurement errors, human interpretation errors, velocity models, fluid contact depths, permeability and numerical inputs for flow simulation. Uncertainty is something that today’s oil and gas operators have to live with, due to the expensive alternatives of drilling more wells and conducting more thorough reservoir samplings.

Quantifying uncertainty, reducing risk. The challenge is how best to quantify uncertainty. If one can quantify the effect of the uncertainties, then one can work out the level of risk – normally financial risk – in decisions and actions.

Information about the biggest uncertainties can also be used to identify where more detailed analysis or data collection is required, to get a more complete understanding of the reservoir. With a better understanding of reservoir uncertainties and risk, robust decisions can be made with the maximum information available.

Current approaches. A typical approach to uncertainty consists of the user producing a single geological model consistent with all available data. Sensitivities around this base case model are then generated – in many cases using simple spreadsheet-based Monte Carlo models. Often, only this single base case model is taken through to flow simulation, where the reservoir engineer attempts to force a history match.

There are, however, several flaws to the approach. First, sensitivities around the base case often do not test a broad enough range of scenarios. Second, where simple spreadsheet Monte Carlo methods are used, no spatial information is produced. Third, realizations are ranked on static criteria, such as volume, rather than on dynamic ones. Finally, often just one realization, the base case, is ever simulated. 

New approaches. Properly utilizing multiple scenarios and realizations for risk analysis is a technique that requires a sophisticated, wide-reaching modelling package across the entire reservoir workflow. This should cover everything from well planning (allowing results of the uncertainty workflow to be used to minimize the risk during the planning process) to fault seal analysis and integrated flow simulation.

Uncertainty management should also extend beyond geological modelling to include a range of other uncertainties and scenarios, such as structure, velocity model, geological environments, net to gross, porosity/ permeability, water saturation, fluid contacts and development plans.

Today’s advanced 3D modelling packages allow for a detailed spatial understanding of reservoir uncertainties. They incorporate the best in reservoir modelling with a comprehensive selection of uncertainty management and decision support tools, Fig. 1.

Fig. 1. Advanced 3D modelling packages can generate a detailed spatial understanding of reservoir uncertainties, incorporating the best attributes of modelling with comprehensive selection of management and support tools.

New or existing work-flows can be turned quickly into uncertainty workflows through easy-to-use software solutions, without resorting to scripting or manual setup. A series of dynamic analysis tools can pre-empt the need to build up complex, time-consuming fine-scale models.

Furthermore, rather than storing the complete 3D model with its hundreds of accompanying realizations and the memory implications that come with it, today’s modelling packages can store table-based reports and scenarios that can be reproduced easily for more in-depth analysis.

Roxar is looking to add an uncertainty module to its IRAP RMS modelling solution, which will quantify the effects of uncertainties on volumes and cumulative production, to ensure that less risky decisions are made. Post-processing features will include tornado charts (designed to analyze the results in a statistically robust manner), and probability cubes. to help identify drilling locations.

Capturing uncertainties in dynamic environments. While static analysis can give a lot of valuable information about the reservoir, only dynamic analysis of the reservoir can fully quantify what the impact of uncertainties will be on reservoir performance. This is where simulation comes into its own.

The last few years have seen a growth in the ability of reservoir simulators to measure uncertainty within a dynamic environment. Simulators, such as Roxar’s Tempest are able to examine numerous geological scenarios that can be history-matched to create simulation models that are fully consistent with their underlying geological interpretations. 

The intelligent downhole network. Uncertainty management can only be improved, if one is able to acquire quality data on subsea operations. A key means of achieving this is through permanent downhole and pressure gauges – essential in monitoring reservoir and well data, Fig. 2.

Fig. 2. Formation of an intelligent network that monitors reservoir and well data increases reservoir knowledge and reduces uncertainty.

Placed between each production zone and utilized not only to monitor temperature, pressure and water cut, but also gas fraction, sand rate and flow velocity, the sensors can continuously monitor the production performance parameters of each individually perforated zone of a multilayer well. By forming an intelligent downhole network, reservoir knowledge is increased and uncertainty is reduced.

Conclusion. While drilling more wells, performing new seismic surveys and running multiple log runs might help to reduce uncertainties, it is in the quantification of these uncertainties that operators can more fully understand risk.

By generating multiple plausible realizations, analyzing the results in a robust manner; incorporating the statistics in decision-making within a dynamic, simulation environment; and utilizing one’s measurement tools; reliable decisions can be made with the maximum amount of information.