Abstract
The Oil and Gas Journal (OGJ) biennial survey on EOR projects has been a partner to the IOR Symposium since early time. Plots of EOR projects versus time, taken from the survey, have been a mainstay of plenary talks. But there is much more to be learned in the database, and the purpose of this work is to find out what can be learned from a broad analysis.
The paper uses standard statistical techniques to infer binary correlations between data sets. With fewer than 300 US projects, the analysis by no means meets the standards of Big Data, but there are things to be learned, nevertheless, because of the scope of survey: It includes all projects worldwide, surveys since 1986, and covers all EOR processes. However, some processes are more represented in the surveys than others, so these are analyzed in the most detail.
We also perform more sophisticated analysis using the conditional probability theorem to infer the probability of success of steam EOR projects. We look at the OGJ survey from the view of screening as exemplified in the work of Taber et al. (1997), and updated by Dickson et al. (2010). In one sense our work is basically an attempt to see how well these guidelines have been followed in practice.
