Integrated Subsurface Studies for the Energy Industry

Latest Posts

Instagram

Follow us

Back to top

Sound Forecasting

Outcome Scenarios from Ensembles of Models

Sound Forecasting™ guides decision-makers to identify the entire range of uncertainty in production performance predictions. It combines high-dimensional data visualisation and data mining techniques to rapidly interpret outcome scenarios from a large ensemble of reservoir models without being overwhelmed by massive data.

Assessing hydrocarbon resources and predicting reservoir performance are critical tasks in all field life stages: acquisition, exploration, appraisal, development, production, and management. However, the various input parameters used to perform reservoir evaluation contain inherent uncertainties, and so do the outputs of model forecasts. Therefore, it is imperative to characterise these uncertainties correctly and evaluate the model response probabilistically in providing more informative and robust forecasting results of future production.

In addition, subsurface uncertainty creates significant economic risks for developing hydrocarbon reservoirs, driving the need for a robust decision-making procedure concerning this uncertainty. Strados helps asset managers better understand the business risk associated with subsurface uncertainty and develop insights concerning geological features’ potential business impact.

Sound Forecasting™ applies a robust methodology for petroleum and gas reservoir development under uncertainty. It guides executives in identifying the entire range of uncertainty in production performance predictions, the most impacting geological input parameters, and the relationship between geological scenarios and production outcomes.

The multidimensional nature of the decision-making process requires handling multiple outcome variables. Our approach uses dimensionality reduction techniques to transform high-dimensional datasets of a large ensemble of reservoir models to interpret flow response scenarios rapidly. For example, we use (hierarchical) clustering approaches to simplify thousands of reservoir simulation results by grouping similar instances together. Other data description techniques applied include principal component analysis, association rule mining, and classification and regression trees.

In addition, there are different graph-based approaches, such as the dimensional stacking visualisation approach used to handle and synopsise data, visualise trends, and identify relationships between sets of geologic parameters and associated outcome scenarios.

Ultimately, Sound Forecasting™ services answer our E&P customer’s questions: What are the possible production outcomes? How are they statistically distributed? What are the main geological variables controlling the results? And which reservoir models best represent the subsurface uncertainties relevant to business decisions?

Want to learn more?

Contact us to find more information about our business solution and services.
Optimized by Optimole