The mostly widely reported production data proved to be a bad indicator of future output.
The study presented on Wednesday (26 January) at the SPE Hydraulic Fracturing Technology Conference and Exhibition in The Woodlands, Texas, considered whether the widely reported, 30-day production total is good way to evaluate a well compared with less frequently offered longer-term numbers.
“Production at 30 days tends to be an unreliable estimate,” said Edward Ifejika, a field operations engineer for Total who did the research while he was a student at Texas A&M University (SPE 184817). The number after 90 days offers a much better estimate of future gas production and the improvement from there is limited.
In comparison, the output of condensate wells over 30 days is a useful indicator, though 90 days offers a better fit with the longer-term data.
The surprises include production that was both better and worse than expected, with a significant number on the upside, he said.
The paper used data from more than 600 gas and condensate producing wells in the Utica formation over a 2-year period. It used a linear regression analysis to test whether there was a correlation between actual production at earlier and later time periods—30, 60, 90, 180, 360, and 720 days.
The result of the calculation is called R2, with a number greater than .85 being a good correlation, from 0.6 to 0.85 a moderate correlation, and below that little correlation.
A short-term time frame is significant because the presentation showed that output over 30 days is the measure normally reported to investors and the media. The impact those reports can have on a company’s stock price puts a premium on the importance of early well performance, which can influence how the wells are completed and operated in ways that may alter later results. The first month of a production is also a time when production may still be influenced by the well cleanup after fracturing.
While the 30-day number is a pretty good fit for the 90-day number (R2 of .91), the fit declines from there dropping to a poor correlation after 270 days (R2 of .53) for gas. In comparison, after 90 days, the correlation with later production was well correlated, and waiting until 180 days does not improve it much.
When the wells are producing condensate, correlations based on 30-day production tend to be much better than with gas wells, suggesting results are likely to vary in other fields.
“It might be applicable to other plays,” Ifejika said, adding that the low correlation for gas wells in the Utica indicates that “this analysis should be applied.”
A questioner from the audience suggested that he consider whether his database properly evaluated wells with special circumstances, such as production limited by tight pipeline capacity, which might have altered the results.
One of those asking about the methodology was Jason Seshadri, a project manager for IHS, a major energy data provider and consultant. When asked if he sees the industry moving to longer-term production reports, he said there will be resistance, because of the appetite “for quick feedback.”
[The technical paper will be available on OnePetro.org at the conclusion of the conference.]
Also From the Conference
Initial Impressions of Wells May Be Misleading
Stephen Rassenfoss, JPT Emerging Technology Senior Editor
26 January 2017