Addressing The Impossible History Match Problem: Using a Systematic Approach to Eliminate Model Bias

摘要

Cognitive biases are unconscious errors in judgement and decisions particularly those that are under conditions of uncertainty.The impact of bias in reservoir simulation can be reduced by validating model assumptions with a multidisciplinary team and sourcing required data that can help to resolve significant uncertainties.This approach was utilized for Reservoir_E simulation study where bias in history match parameter was eliminated.In an initial history match attempt,assisted history match method was used to indicate that a single tank model solution was "impossible" for this reservoir.A two-tank history matched model was therefore proposed against prior geologic and engineering understanding of the reservoir.This bias was identified by a multi-disciplinary team after analyzing all available data.This led to a second history match effort which expanded the range of uncertainty parameters and got a good history match with a single tank,in line with understanding of reservoir architecture and flow mechanism.The single tank assumption was later validated with geochemistry analysis of oil samples from existing producers.This paper demonstrates best practices in conducting reservoir simulation studies by showing the nonuniqueness of simulation models and the need to validate model assumptions using a multidisciplinary team and all available data.