Drill-Bit Diagnosis With Neural Networks

Abstract

Summary A neural network was constructed to determine the grade (state of wear) of a drill bit while it is drilling. Using a three-layer neural network and back-propagation as the learning algorithm, the system was trained with laboratory data collected using bits of known grades drilling lithologies. The inputs to the neural network were rate of penetration (ROP), weight on bit (WOB), torque (T), revolutions per minute (RPM, and hydraulic horsepower per square inch (HSI). The network was tested on synthetic formations of varying bed thicknesses, which were constructed from the test data. Introduction to Drill-Bit Diagnosis During drilling, it is important to have an estimate of the drill bit's condition or state of wear. Drill bits are graded primarily on the length of their teeth. A new bit is said to have primarily on the length of their teeth. A new bit is said to have a grade of 0, and as the bit wears during drilling, the grade goes up linearly until the teeth have completely worn away at which time they are said to have a grade of 8. If the drill bits are continually pulled before they are dull then the cost of the drilling operation will rise significantly due to the cost of the rig time required to pull the bit and install a new one. If the bit is used too long then it will drill inefficiently, and the cost will rise because of the time required to drill at a low ROP. In cases, the condition of the bit may be ignored until catastrophic failure resulting in the loss of one or more of the rolling cone cutters. When this occurs, economic loss can be very significant due to the time consumed in retrieving the broken pieces from the bottom of the hole. Consequently, a more accurate estimate of the grade of the drill bit translates directly into a more efficient and less costly drilling operation. Analytic methods for using drilling parameters and an estimate of the lithology to estimate the grade of a bit while drilling have been only marginally acceptable because of the lack of a good algorithm and the marginal quality of lithology estimates while drilling. As a result, drillers often still use their experience in the particular field and formation together with their knowledge of the particular field and formation together with their knowledge of the bit they are using to estimate the grade of the bit. The result is that often the bit is pulled too quickly, and the cost of drilling is higher than necessary.