Bit bounce is generally regarded as not a problem issue when drilling with PDC bits. It is of course well known that they are an issue for roller cone bits especially when drilling hard formations, due to the possibility of forming a three lobed pattern on the rock being drilled. With PDC bits, bit bounce is primarily noticed in association with stick slip, when going back to bottom, or when an auto-driller misbehaves.
Even though they may be less prevalent than other dysfunctions, they can cause damage to the bit and other downhole tools, and therefore should be monitored and mitigated. The main channel we monitor here is the WOB channel. The WOB variation index (which is quite similar to torque variation index) and the WOB trend are the main metrics that are tracked. It is also expected that depth of cut and MSE will be erratic. So, they are tracked as well. Now, even though stick slip and axial bounce can happen at the same time, it was decided to use this model to track bit bounce when the torque variation index is low. The reasoning behind that being, if the two dysfunctions are happening together, taking corrective action for stick slip would also reduce bit bounce.
Slide # 3 has an example of bit bounce detected using this model. As you can appreciate, there are nuances to the various ML models that are developed, and they must be clearly understood, to be used properly. Next week, we will discuss detecting whirl from surface data.
Click below for slides on this topic:
If you find our posts informative, please share with your colleagues. Comments and critiques of the model are welcome.