Today’s oil and gas industry is increasingly turning toward complex stratigraphic-diagenetic and structural plays. Prediction and risking of reservoir heterogeneity, seal integrity and source rock sweet spots are becoming more important than ever before. Currently, prediction and risking rely primarily on stochastic geostatistical approaches, which have seen an impressive development over the last few decades. However, exploration of and production from increasingly complex plays has revealed higher levels of uncertainty in geostatistical reservoir models, because of a combination of factors:
- Statistical models do not fully capitalize on the geological information available
- Prediction and risk assessment usually apply a single statistical approach
- Different geostatistical approaches produce varying predictive models
- Surface geological studies (analogues) have proven highly pronounced rock heterogeneity
- Multiple, concurrent processes with various feedback mechanisms control reservoir quality
In order to meet current and future challenges of increasingly complex prospect and play types, the industry needs to develop new, additional approaches to reservoir, seal and source rock prediction. The key requirement for reducing uncertainty and risk in exploration and production is a rigorous understanding and quantification of geological processes and controls.
Fundamental research in geological process-based forward modeling started in the 1960s to 1970s in academia. However, the exploration industry has only recently started to more widely deploy geological process-based forward modeling. The initial focus has been on depositional modeling using diffusion, Navier-Stokes and hybrid geometric approaches, but more recently a diverse range of approaches is being adopted. They include fuzzy logic, cellular automata and various other reduced-complexity modeling approaches that produce output information on petrofacies, depositional environment, and textural porosity. Forward modelling is also being applied to diagenetic processes using reaction-transport modeling (RTM) or reduced complexity proxy rules and to geomechanical processes using finite element or discrete fracture network modeling based on post-burial mechanical stratigraphy and local/regional stress patterns. Geological process-based forward modeling has shown highly promising results for e.g., reservoir quality, seal integrity and sweet spot prediction in complex play and trap settings but many challenges persist, including:
- Calibration of numerical input parameters specific to age, climate and structural settings
- How to use physical experiments and outcrop-reservoir analogue studies for model verification
- Automated input parameter optimization
- Multi-scale process-based models from basin to prospect, play and inter-well scale
- Linking and integrating approaches for depositional, diagenetic and structural modeling
- Integrating textural, diagenetic and fault/fracture-related poroperm models
- Sensitivity analysis and quantitative risk assessment of multiple modeling realizations
- Computational expense vs. complexity of numerical approach vs. temporal-spatial resolution
- Effective implementation in existing industry workflows.
In recent years, interest in geological process-based forward modeling has extended to the geothermal exploration industry and the CO2 storage industry, which face some comparable challenges in predicting subsurface rock parameters and their spatial distribution.
The proposed workshop will include invited experts and interested researchers from both industry and academia. We will concentrate on geological process-based forward modeling rather than on geostatistical modeling, flow simulation, or hydrocarbon systems modeling. Six sessions spread over a period of 2 1/2 days will be dedicated to key challenges in geological process-based forward modeling, finishing with a concluding session to define a practical way forward.