Thesis work in the area of machine learning on drilling data:

Forecasting Drill Rig Operations and Event Sequences Using Machine Learning

Supervisors: Epiroc and Johannes Andreas Stork (Örebro University)

Goal:

Develop and compare machine learning models that can forecast the duration and sequence of drilling operations based on historical rig event data. The objective is to predict how long upcoming actions (e.g., drilling cycles) will take and anticipate key events such as refueling, drill bit changes, or water refills.

Background:

Drilling operations generate detailed time-series data describing the sequence and timing of various actions and events. Understanding these temporal patterns enables more accurate planning, improved resource allocation, and reduced downtime. Machine learning methods for sequence prediction and time-series forecasting can uncover hidden relationships between events and their durations.

Thesis Tasks:

The student will:

• Analyze time-series data describing drilling events and cycle durations.
• Explore feature extraction techniques for event-based time series (e.g., average cycle time, frequency of events, intervals between events).
• Apply and compare different ML models for forecasting and classification, such as Regression models, Random Forests, Gradient Boosting, or lightweight Recurrent or Temporal models.
• Evaluate model accuracy in predicting future operation durations and event occurrence.
• Provide recommendations for practical integration into drilling operation planning systems.

Expected Outcome:

A comparison and evaluation of machine learning approaches for predicting drill rig operational behavior, identifying which techniques best model event sequences and timing.

Student Profile:

• Interest in data science, time-series analysis, and predictive modeling.
• Basic experience with Python (pandas, scikit-learn, possibly PyTorch or TensorFlow)
• Motivation to apply machine learning to real-world industrial data.

For more information, please contact Maria Krantz (Epiroc, maria.krantz@epiroc.com) and Johannes Andreas Stork (Örebro University, johannesandreas.stork@oru.se) Associate Professor