Purpose
During bitumen processing, insufficient flow of slurry leads to deposition of coarse sand particles in hydrotransport lines, termed 'sanding' which can lead to blockages. This project aims to develop a method for detection of sand bed formation in hydrotransport lines using real-time predictive analytics on sensor data.
Expected Completion Date
January 2027
Project Overview
Currently, there is a lack of reliable early warning systems for sanding events and partial or total plugging events can cost millions of dollars per incident.
This project analyzes the feasibility of SANDtracTM, a sonar-based velocity profiler that provides velocity at multiple depths of pipeline cross-section, for detection of sanding events. The key element of our study is the interpretation of its one-year deployment data. Methods have been developed and tested for the removal of noise, harmonics, outliers, and other unwanted signals from sensor data, followed by comprehensive understanding, modeling, and analysis of sanding events.
Future/Ongoing Work
A machine learning model will be developed and trained using one year deployment and historical data to detect sanding events. Developed methods will significantly reduce the use of energy and water during the bitumen extraction process.