Super seismic

Eric Adams discusses keys to success in seismic-to-simulation workflows.

“By rigorously using all the available seismic well, geological and production data to build the model and maintaining the connection to that data throughout the process, the model will be accurate, robust, and updateable”
-Eric Adams

What defines a successful seismic to simulation workflow?
Eric Adams. Companies invest in seismic-to-simulation analysis for the invaluable insight it can provide into the best production plan for the reservoir. To ensure that the investment is well made, the workflow must meet some very stringent criteria. It must include and honour, all well, seismic and geological data. This is the foundation of all that follows, so it must be comprehensive and reliable. It must also match production data, without resorting to artificial barriers or porosity modifiers. The workflow must accommodate iterations such that the model can evolve to include new information without losing its integrity. It must allow creation, evaluation and ranking of production scenarios to aid in building the production plan. From my experience, the only reliable way to achieve these goals is to use a quantitative, iterative approach.

How is this approach different?
EA. Quantitative integration is founded on three core principles. First, all data – including well, seismic, and production data – are included in a single, comprehensive workflow. Second, because of the uncertainty involved in all reservoir modelling, the workflow must be iterative and updateable, incorporating new information into the model as it is available. Third, the model should include highly detailed properties derived from the seismic. The more information included in the model, the more likely the model will match new data added to it. Any workflow based on these principles must be multi-discipline and include all the software tools to handle petrophysics, rock physics, inversion, geostatistics, uncertainty estimation, upscaling, simulation and production scenario development. Most workflows today fall far short of that standard.

Where are the typical points of failure in reservoir modelling?
EA. There are four common areas where problems arise in reservoir modelling. First is in defining the range of highly detailed, plausible 3D models from seismic. The results must look like geological layers or it won’t behave properly in the simulator. Second, typical workflows do not sufficiently establish the relationship between petrophysical reservoir properties – lithology, porosity, permeability, and water saturation – and elastic properties – p-velocity, s-velocity and density. This relationship is essential in obtaining accurate petrophysical models from seismic data, through geostatistical inversion. A third problem area is accurately transferring these 3D rock property models from seismic to corner point grids, without which problems such as discontinuity in fluid flow can be introduced. The final problem area comes in uncertainty estimation and production scenario ranking. If modifiers or artificial barriers are added to the model to force it to match production history, its usefulness is compromised.

What extra value does a quantitative seismic-to-simulation workflow provide?
EA. By rigorously using all the available seismic, well, geological and production data to build the model and maintaining the connection to that data throughout the process, the model will be accurate, robust, and updateable. It can then be used with high confidence to predict future production volumes, ultimate recovery, and to evaluate the impact of various reservoir management options.

In one North Sea operation, such a workflow yielded a final model that honoured all available geophysical, geological, and engineering data with a 95 percent match to historical data in producing wells. In an Arabian Gulf field, a model was built including and honouring logs from 42 wells, as much as 30 years of production history from 92 wells, and a recent seismic survey. The model that resulted is still in use three years later, and current field production still matches that model prediction.

Eric Adams is Managing Director of Fugro-Jason, a leading provider of reservoir characterisation products and services. He joined the company in 1996 as Region Manager, North and South America. Adams has 30 years’ experience in the industry, including previous positions at Schlumberger and Veritas.