Reconciling History Matching and Assessment of Uncertainty in Production Forecasts: A Study Combining Experimental Design, Proxy Models and Genetic Algorithms

This reference is for an abstract only. A full paper was not submitted for this conference. Description The paper presents a method to tackle complex inverse problems where highly non-linear responses are involved. Geological models are built within an experimental design framework and are characterized by an objective function that estimates the quality of the history-match. The goal is to efficiently find combinations of parameters that minimize the objective function. Genetic algorithms are the main optimization tool in the workflow. In order to reduce the number of actual simulations and to accelerate the overall procedure, non-linear response surfaces, built with kriging interpolants at each iteration of the optimization routine, filter out unnecessary combinations of parameters. The models that reasonably honor the historical data are selected via cluster analysis techniques and provide an estimate of future production. The final distribution of the prediction variables defines the range of uncertainty conditioned to production history. Application The practicality of the workflow is demonstrated on the Malampaya field, a Gas Condensate reservoir in the Philippines. The field is a Tertiary Carbonate build-up situated offshore, below 800–1200m of water. It has been supplying gas from five subsea production wells since 2001 to three Gas-to-Power plants. Available subsurface data include a high resolution 3D seismic survey, five production wells with downhole pressure gauges and six appraisal wells with wireline and borehole image data, pressure and well test data. Results The strategy ensures multiple and significantly different history-matched models that provide estimation of the future performance of the reservoir. The coupling of space-filling sampling strategies, optimization algorithms, non-linear response surfaces and high performance computer clusters proved efficient in addressing the issue. In addition to a robust assessment of ranges in production forecast, representative P10, P50 and P90 individuals were selected from the portfolio of models for further analysis including optimization of development plan. Significance Understanding the impact of subsurface uncertainties on production responses is an integral part of the decision making process. A more accurate quantification of the uncertainty band around production forecasts contributes to better business decisions. Traditional experimental design workflows might be well suited for new field developments. However, when a field has been produced for several years, all models have to be conditioned to available production data in order to obtain meaningful predictions. This paper addresses the limitations of conventional techniques and provides a practical, structured workflow to reconcile the processes of data integration and uncertainty assessment.