Abstract

Identifying the root cause of damage of a pulled bit as soon as possible will aid preparation for future bit runs. Today, such bit damage analyses are often anecdotal, subjective and error-prone. The objective of this project was to develop a software algorithm to automatically analyze 2D bit images taken at the rig site, and to quickly identify the root cause of bit damage and failure.

A labelled dataset was first created whereby the damage seen in bit photos was associated with the appropriate root cause of failure. Particular attention was given to the radial position of the cutters that were damaged. Using the 2D bit images (which can be obtained at the rig site), a convolutional neural network along with other image processing techniques were used to identify the individual cutters, their position on the bit, the degree of wear on each cutter. A classifier was then built to directly identify root cause of failure from these images.

This work utilized a large dataset of wells which included multiple bit images, surface sensor data, downhole vibration data, and offset well rock strength information. This dataset helped relate the type of dysfunction as seen in the downhole and surface sensor data to the damage seen on the bit. This dataset however only covered some types of dysfunctions and some types of bit damage. It was therefore augmented with bit images for which the type of failure was determined through analysis by a subject- matter expert. A classifier was subsequently developed which properly identified the root causes of failure when the bit photo quality met certain minimum standards. One key observation was that bit images are not always captured appropriately, and this reduces the accuracy of the method.

The automated forensics approach to Polycrystalline Diamond Compact (PDC) bit damage root cause analysis described in this paper can be performed using 2D bit photos that can be easily captured on a phone or camera at the rig site. By identifying the potential root causes of PDC damage through image processing, drilling parameters and bit selection can be optimized to prolong future bit life. The algorithm also enables uniformity in bit analysis across a company's operations, as well as the standardization of the process.

You can access this article if you purchase or spend a download.