So, let us first define “full stick” – at least for use within this post. We define a full stick event as an extreme case of stick-slip, where the bit comes to a complete stop for a slightly “extended” duration. These events are violent and can be extremely damaging to the drill bit, mud motor, etc. Therefore, it is important to track them separately from generally less severe stick slip.
The torque variation index is no doubt a good indicator of the severity of the stick slip. But it is an averaged metric. There is a better way to track these events. Full sticks are characterized by a rapid increase in surface torque reading. See slide 3, for one instance of this event. The sawtooth pattern shown in the slide can be detected in surface data, and so is added as a node to the Bayesian network model. You may notice many nodes similar to the stick slip model we posted last week. Can you find the difference? For one, pressure signal will spike as well. Operationally, full sticks should be tracked, and mitigated/avoided.
As we continue to explain the models used within Sentinel RT, note that these models are available to be licensed either directly from us, or through our partners, who run these models both at the rig site and in the cloud. We partner with operators, service providers, rig contractors, equipment manufactures, and pretty much anyone who would want to optimize drilling operations. So, send us a note if you are interested in learning more, whether it is to use the product through our partners, or for partnering directly with us.
More information about the below slides are available in SPE-204063-MS:
We will move onto the bit balling detection model next week. So stay tuned, and share these posts with your colleagues, so they might benefit from this knowledge share as well.