Geological Risks and Resources-Based Portfolio Ranking and Completion Optimization by Areas: An Eagle Ford Case Study

Abstract

AbstractAsset ranking plays a critical role for the business success. Given the risk and value of asset, choosing an asset by the highest expected value is not guaranteed to be the best decision because it does not consider the uncertainty in the expected value and decision maker's changing tendency toward the risk. The first part of this paper presents an application of the modern portfolio theory to the Eagle Ford unconventional resources and shows how the assets can be prioritized by taking into account the subsurface-based expected value and the uncertainty. For the fair comparison of the asset's potential, the influence of completion parameters is excluded in the asset ranking. With the stationary completion designs across the assets, the asset is evaluated and ranked only by the subsurface quality. Having the selected asset area, the second part of this paper focuses on optimizing the completion designs. Well production trend is predicted against the varying completion designs over the selected asset area and displayed onto the axes of the completion parameters of interest. Optimizing over the selected asset makes a smaller number of data available. Some completion parameters may be fixed due to operational issues – e.g. limited range of porosity and reservoir pressure in a selected asset, with the limited lateral length and total proppant amount. A data-analytics approach with a data simulation technique has been developed. Tens of thousands additional data for the input variables (subsurface and completion) is generated by a data simulation technique. These simulation data are inputted to the statistical predictive model built for the well performance so that many data points for completion, subsurface and well performance are available. With the help of simulation data, any well performance trend analysis with key completion variables of interest with some constraints for other variables is possible.