From machine control to drilling control

Abstract

The complexity of drilling operations is constantly increasing. By changing the perspective of automation from a pure machine control mindset to a drilling control one, it is possible to provide proactive assistance to the drilling crew during drilling operations. This needs to take in use physical models of the drilling process. Such mechanical, hydraulic and heat transfer models are dependent on physical quantities that can be derived from real-time sensors. The correlation between the existing measurements and the physical model parameters needs to be analyzed in order to extract which corrections need to be applied on the measurement to extract the required physical values and whether the total sensory system, at the rig site, is complete and robust.

Toward Drilling Automation: On the Necessity of Using Sensors That Relate to Physical Models

“…Note that those analyses must take into consideration the entire automated-drilling process, starting from the measurements themselves, to their interpretation and numerical treatment by models, and finally to the response that is generated by the automation system. The sensitivity analyses of this process are available from the literature, including uncertainty analyses for the wellbore position, the wellbore stability (Udegbunam et al 2013), pressure estimations (Cayeux and Lande 2013), and the uncertainty associated with the numerical resolution of the equations (Abgrall 2008). A global analysis is still lacking.…”

“…In fact, the PVT property of the mud is never measured in situ, and is only derived from the PVT behavior of the base constituents. The sensitivity of the overall PVT estimation to the exact formulation of the mud is discussed by Cayeux and Lande (2013), in which it appears that in most cases the obtained precision is sufficient, but may not be in particularly sensitive drilling operations. It should therefore be evaluated whether an automatic PVT measurement tool could be developed.…”