Automatic Drill Bit Dull Grading Report Generation to Address the Revised IADC Dull Grading Schema

The current IADC bit dull grading process has served the drilling industry well for over 30 years, and it has helped the industry achieve significant improvements in drill bit selection and design. With rapid development of computational techniques such as computer vision and machine learning, the general consensus is that the previous grading process can now be updated to capture additional information about the bits that was previously discarded to enhance the grading process and the information is delivers while keeping it simple. An IADC committee is currently working on finalizing this new schema. In comparison to the previous bit dull grading process, more details about the drill bit such as cutter damage type have been added to the new schema. Additionally, capturing pre-run and post-run bit images is recommended to enable detailed bit forensic analysis into root causes of observed damage. For the bit dull grading process at rig site, a preliminary grading report needs to be filled out and rig site bit photos also need to be captured. The objective of the work reported here was to develop an algorithm to automatically analyze 2D bit images and quickly generate are port that fulfills the new IADC dull grading requirements. Several deep learning models and classic machine learning algorithms were developed to automate the bit dull grading process. Given a set of high-quality bit images, the proposed set of algorithms will process and analyze the images automatically and generate a dull grading report. Different sets of bit images were tested with the proposed algorithms, and the results show that they are able to achieve good performance under standard scenarios. A discussion on hurdles that remain to be tackled is also included.