Improving Reservoir Forecasts by Understanding the Relative Impacts of Sparse Data, Reservoir Modeling Workflow and Parameter Selection, and Human Bias

摘要

Nandurdikar and Wallace’s (2011) review of nearly 150 major capital projects showed that the industry is actually delivering only 75% of the production volumes forecast at the time of project sanction overall and those projects with identified subsurface issues delivered only 55% of the forecast volumes. In other words, the industry production forecasts are significantly optimistic. There are a variety of factors that contribute to the optimistic forecasts. The most important are areal subsurface model grid size, well location optimization workflows, sparse data bias, and “pro-project sanction” management bias. Each of these individually may contribute on the order of 10-25% or so of the observed overall forecast optimism. The impact of other factors such as stochastic model parameters (e.g. semivariogram range) and vertical upscaling are significantly less important.