Abstract Over the last decade the advances in downhole instrumentation have provided increasingly more detailed information about the behavior of drilling systems in real world rather than laboratory conditions. While these downhole systems, now intimately coupled with surface data logging, have provided us with a greater ability to intervene in the process and avoid hazardous conditions, what has proved to be extremely revealing is the detailed post bit run analysis of the data.
This paper sets out to illustrate some of the features that have been observed, how these features have been interpreted, and how this new knowledge of downhole behavior is now being applied to bit selection, BHA design, and the development of new tools.
The paper will explore aspects of the fundamental interaction between drill bits and formations, and how, by making this interaction function optimally, a wide range of associated issues are also addressed. In the course of exploring these issues some of the traditional methods of bit selection and drilling systems design will be shown to be based on incorrect assumptions, and how correcting this situation leads to general and significant improvements in well construction economics.
Introduction Roughly a decade ago our ability to make measurements downhole matured to a point where we had achieved the bulk of what we still measure today in terms of well placement and formation evaluation. While these important areas of measurement have continued to be refined, their primary use continues to be focussed on the reservoir and near reservoir sections of the wells we drill. In the last few years we have had the luxury of being able to add to our armoury a series of sensors that have allowed us to build significantly on our understanding of the drilling process itself. This has lead to an improved ability to impact the economics of drilling the overburden as well as the reservoir.
Initially these sensors were intended to help avoid hazardous conditions while drilling, and might be viewed as providing the ability to observe and manage the symptoms of sub-optimal behavior. These sensors include vibration, pressure while drilling, caliper, weight-on-bit, and torque.We have also added the ability to integrate improved surface data logging with these downhole measurements to provide a more comprehensive record of the behavior of the complete drilling system.
Many of these sensors measure attributes of the process that also are amenable to modeling. This is significant inasmuch as it brings a focus to managing the drilling process that starts with the planning phase. When we plan we build a series of engineering and economic models of expected behavior. There is nothing particularly earth shattering about this. The outcomes of the modeling exercises are encapsulated in the familiar items like the well trajectory, drilling program, torque and drag estimates, BHA, drill string design, bit selection, hydraulics, time-depth curve, mud program, AFE and so on.
It would be peculiar if, having made these predictions about what we are setting out to do, we did not then compare reality with the expectation. However, many of the comparisons we have traditionally made contain significant "fudge factors" that collected together all the elements we either did not understand or could not measure in detail. A good example of this is the use of the euphemism "Friction Factor" in torque and drag calculations. Any engineer worth his salt would cringe at the implication that this bore any relation to a "coefficient of friction" between the BHA/Drillstring and the formations. The friction factor used in Torque and Drag models is a catchall for the detail that we cannot "see" or do not understand. Therefore degrades the models to a gross generalization of the interaction between the drilling systems and the formations they are trying to excavate.