Stochastic Prospect Charge Evaluation – Old Challenges and New Solutions

Christian Zwach and Kristian Angard
Hydro Oil and Energy, Global Exploration, Drammensveien 264, N-0240 Oslo, Norway

Sensitivity analysis of basin models is a substantial part of prospect charge evaluation in petroleum exploration. During this work process the most critical geological factors and properties have to be identified by systematic variation of input data and boundary conditions. For a full 3D regional basin model this is at present still challenging due to the simulation run times and data storage capacities required.

Numerous efforts to reduce simulation times and to study sensitivity more systematic have been published. However, simulation times for full 3D basin models of significant size are still relatively long, even when running on large computer clusters with parallelised computer codes. Full 3D Monte Carlo simulations are therefore currently not in use in industry.

Internally, we have therefore established an alternative work process, which is tuned to typical exploration situations. Here evaluation of petroleum systems has often to be performed within a few days to weeks. Our strategy therefore was to simulate a few relevant geologic cases in full 3D and use these data subsequently during post-processing. The geologic cases can be for example a minimum, most likely and maximum heat flow history for a given area to evaluate maturity of an identified source rock interval.

During the post-processing of the geologic cases we use highly optimised Monte Carlo simulations to calculate maturity and expelled petroleum volumes from the source rock intervals of interest. The results from the few full 3D basin models are used as statistic end members (e.g. min/most likely/max) and realizations are sampled in between.

This work process has some obvious advantages, mainly simulation times and identification of the most sensitive factors. We show how the most sensitive factors for source rock maturity and expulsion are easily identified with our approach. In addition, it gives a first probabilistic quantification of e.g. expelled volumes in a defined drainage area. We use these quantifications to focus further detailed work, e.g. on secondary migration evaluation.

We discuss several generic problems of probabilistic basin modelling in our paper, mainly the challenge of defining the appropriate probabilistic distributions of input parameters.

We present several exploration cases where our work process has been applied and show how critical factor identification led to change in further exploration work program.

 

AAPG Search and Discover Article #90066©2007 AAPG Hedberg Conference, The Hague, The Netherlands