An Analysis of Bit Bearing Failures With Field and Laboratory Data

Abstract

Abstract This paper discusses how to detect the onset of bit hewing failure before the failure results in a lost cone. It does not attempt to understand the physics of wear. A history of failures was collected for both journal and roller bearing bits from wells where both surface and down-hole mechanical parameters were recorded. The data were then correlated with formation logs to differentiate lithology changes from bit effects. A separate experimental program was conducted on full size bits to observe how a failing bearing affects a cones speed, torque and temperature. A journal bearing bit was worn from new to failure under atmospheric conditions and data were acquired at regular intervals of wear. Comparisons are made between journal and roller bearing bits. This work showed that journal bearings do not necessarily fail in a catastrophic manner as generally believed, but may degrade over several hours. The results suggest a likely technique for separating the one set of bearing failures from lithology effects. A field example is thoroughly examined. Introduction The loss of a cone down-hole is one of the two possible hazards associated with a roller cone bit. A lost cone is either fished or pushed into the bore-wall. Fishing costs range anywhere from a round-trip to a side-track. Pushing the lost cone aside may appear cheaper, but can cause damage to a new bit and create tripping problems. The second hazard, not discussed here, occurs when an undergauged bit drills a conical hole, which can cause the next bit to wedge in the well. Tooth wear is not considered a hazard. If the rate of penetration (R) can be maintained by increasing the weight on bit (W) or the rotary speed (N), and the cost of the bit is small compared to the total drilling costs, the state of wear of the bit is irrelevant. On the other hand, the drilling parameters are often limited by directional considerations. It is then generally considered more important to keep on drilling in the right direction at a reduced rate than tripping to change the bit. The real concerns of the driller as regards a roller cone bit should then be the good transmission of weight to the bit, the possibility of loosing a cone, and bit gauge wear. The relative importance of weight transmission, loss of a cone or undergauged hole varies locally. In a well documented case', the bit runs were limited to 14 hours after a succession of bearing failures and lost cutters. However some of the failed bits lasted up to 35 hours. Several techniques to detect potential bearing failures have already been proposed, but they need new down-hole sensors (accelerometers) and/or fast acquisition rates. The technique described here uses existing sensors and data transmission rate. Since the introduction of a Measurement While Drilling (MWD) tool to measure weight on bit and bit torque in 1981, we have had the opportunity to gather a database of bearing failures. In parallel we have been experimenting on small- and full-scale drilling machines. The agreement between the results that we obtained in the field and the lab led to the development of a technique to identify bearing failures. BEARING FAILURES DATABASE The database is presented in Tables 1 and 2. The data were collected from the Gulf Coast U.S.A, Canada, North Sea and high Sea. Table I shows the raw data and Table 2 the results of simple computations. P. 87^