Experimental Design and Genetic Algorithms Approach to Quantify Model Uncertainty – A Case Study

H034 Experimental Design and Genetic Algorithms Approach to Quantify Model Uncertainty – A Case Study H. Gross* (Chevron) A. Castellini (Chevron) I. Gullapalli (Chevron) & V. Hoang (Chevron) SUMMARY Recent developments combine genetic algorithms and experimental design to provide fast and efficient model building tools. Beyond single-number forecasts probabilistic predictions enhance the accuracy of simulation and allow quantifying risks and opportunities in field developments. Often history-matching focuses solely on production forecast uncertainty. Model uncertainty upstream from simulators has more information content: mapping regions with high variability on its physical parameters but good match with past production data makes it possible