Master Thesis Opportunity: Drilling state classification in rotary drilling

During drilling operations, it is important to adapt to different conditions that influence the outcome like the trajectory of the hole or the time it takes to drill it to the end. Different drilling states have been manually observed in data available from the control system of the drill rig. To automatically monitor this and establish connections to the control of the drill would be an important first step to develop adaptive drill control.

This thesis project offers students the opportunity to work with a real-world dataset to investigate this problem.

Project Objectives

• The project has the following objectives:
• Exploratory data analysis: Analyze the data to identify what different drill states might exist.
• Drill state identification: Develop a method for automatically classification drill states from data.
• Understanding drill states: Investigate possible correlations with drill control or other drilling parameters, e.g. depth.

Research Question

• What classes of drilling state can be distinguished based on the rig control data?
• What accuracy of a classifier can be achieved?
• If there are any, what are the dependence or correlation to the control or any other drilling parameter?

Who Should Apply?

• This project is ideal for students with interest and background in:
• Data science, signal processing, or machine learning
• Mining engineering or mechanical engineering
• Programming and data analysis (e.g., Python, MATLAB)

The thesis can be conducted by one student and is set up as a collaborative project with Epiroc Rock Drills AB in Norra Bro. The thesis is hosted at Epiroc Rock Drills AB in Norra Bro. The project is available to start at any time.