Name
Technical use case: Machine Learning approach to predicting oil sand recovery
Date & Time
Thursday, April 27, 2023, 9:50 AM - 10:20 AM
Speakers
Description
- Building a model to predict elevated extraction recovery deviations for the Kearl oil sands mine
- Leveraging process surveillance data to isolate our data set to causes of interest only
- Visualizing Shap explanations over time to shed light on what is driving the model predictions leading up to and during events
- Evaluating the model has uncovered other potential uses of the approach, not only to predict but also to highlight process control adjustments or post-event root cause analysis