Previous Lecture Complete and continue  

  7. Data Quality

Lecture content locked

Enroll in Course to Unlock
If you're already enrolled, you'll need to login.

Transcript

- [Instructor] A note on data quality. We have to be really aware of the hoop stress in the available region and that's why we want to pump enough volume to overcome that. Spend more time on the last falloff cycle and do so with natural leakoff. That's the one that's going to be most representative of how the mineral works under stress. You check repeatability of measurements from one cycle to the next. Plot stresses with depth to identify outliers. This sounds quite simple, but I think that it's very important. The best time to do this is when the tool is in the hole. The moment you see any outliers and pressure on the system start, make sure you ask for the injection follow up tests. Or for an actual station to confirm the result. But once you pull it out of the hole, you might never get a chance to run this tool again. Compare results with sonic models, image log, core analyses, other forms of injection tests. Analyze pressure transient plots to pick closure in real-time. It's a very powerful approach to make this decision when the data is being acquired. Pretty much every vendor nowadays has, or should have, a plot where we look at a lock-log time start, or the G function start and iron tight closure. That's the best that's being done. When in doubt, phone a friend. By that I don't mean phone me, I mean phone up the line company or the company's shift engineer. Solving ignition with Microfracturing is different from underlying realities as we cannot "reprocess" or "normalize" the data if it's not acquired correctly the first time. Lastly, we want to make sure we have all hands on the deck when the data is being acquired. Whether it's Vendor Experts, Asset Teams, Completions Experts, we want to have everyone engaged to make sure we make the decision upon the flow rate, the time to wait on each station, and confirm closure threads from each cycle to the next.