Erroneous reserve estimates associated with improper application of Arp's relations has led to the development of alternative rate-time models based on empirical considerations. These models can equally provide good fits for only limited duration production data, but they yield significantly different 30-year- estimated ultimate recovery (EUR) estimates.
In our previous work, we extensively demonstrated the reliability of cumulative-production model to analyze the production performance of fractured dominated tight and shale reservoirs.
A number of field and simulation examples are used to generate 30-year-EUR estimates using different rate-time models as well as production-Decline model. As to quantify the uncertainty associated with the aggregating 30-year EUR forecasts, we use the approach of Generalized Likelihood Uncertainty Estimation (GLUE) methodology.
The results show that some of rate-time models overestimate EUR with more than 150% compared with cumulative production-Decline model which yields exact estimates in most of the cases. The results also show that the Generalized Likelihood Uncertainty Estimation provides a robust methodology for quantifying the uncertainty of the EUR estimates.