Drill Bit Wear Detection
Project Overview
Objective: Improve the accuracy & repeatability of dull grading Polycrystalline Diamond Compact (PDC) bits for drilling oil & gas wells
Role: Self-guided independent research project, developed scope & budget, literature review, research & implementation, wrote & presented final report
Skills: Image processing, machine learning (convolutional neural networks), 3D scanning, CAD, technical writing
Result: Filed non-provisional patent application (currently being reviewed by USPTO)
Project Description
For my Master's thesis at Texas A&M University, I investigated three methods for improving the repeatability, resolution, and accuracy of dull grading Polycrystalline Diamond Compact (PDC) drill bits in oil & gas well construction. Through the dull grading process, engineers can gain valuable insight into what caused the damage, be it excessive axial/lateral vibration, improper operating parameters, incorrect bit design, etc. Based on this feedback, refinements can be made to the bit design and operating parameters to extend bit life, improving performance and reducing the cost to drill the well. So, having an accurate and repeatable process to perform these bit forensics is essential to improving drilling operations.
The current industry standard practice is to visually inspect the drill bit after use and manually classify damage to the bit using an 8-digit alphanumeric code; however, this process is subjective, time consuming, and records limited information. For this project, I tested using a combination of image processing, machine learning, and 3D scanning to generate a more holistic picture of bit wear. Image processing was used to identify the cutting elements and quantify the location and amount of damage to each cutter. Machine learning, through the implementation of a convolutional neural network, was used to classify the location and type of damage to each cutter (i.e., erosion, chipping, delamination, breakage, etc.) Finally, 3D scanning was use to evaluate the location and amount of damage to the entire bit including non-cutting elements, such as damage to the body of the bit that would result in it being under-gauge or plugging of the cooling nozzles.
Please click through the gallery below to learn more about the project. Hovering over an image/video will provide a brief summary of what you're looking at and clicking on it provide a more detailed description.
