As the role of reservoir-flow simulation increasingly affects existing operations and field-development decisions, it follows that rigor, fitness, and consistency should be imposed on the calibration of reservoir-flow models to dynamic data through history matching. To evaluate the applicability of the diverse techniques available, a study was performed to benchmark common assisted-history-matching (AHM) techniques. To benchmark the techniques consistently, a set of standards was defined against which each was evaluated. Of the techniques evaluated, the design-of-experiments (DOE) -based approach uniquely satisfied all requirements.
Introduction
In order to understand the practical utility of the wealth of history-matching techniques reported, a study was performed to benchmark four AHM techniques that have been applied in the oil and gas industry for asset-management applications—DOE-based methods, genetic algorithm, ensemble Kalman filter and smoother, and streamline-based generalized travel-time inversion.
