Thesis: Optimizing Electric Pump Energy with ML + EIT

Abstract

My doctoral thesis explores combining electrical impedance tomography (EIT) and machine learning to monitor pump operation and reduce energy consumption. By reconstructing internal states with EIT and training data-driven models to detect anomalies and suggest energy-saving control actions, the project delivers a pipeline from sensing to actionable control.

Methods & contributions

• Data collection and EIT reconstruction for transient flow inside pumping systems
• Novel ML pipelines for real-time anomaly detection and condition-based optimization
• Integration of control-theoretic objectives with learning algorithms for operational efficiency

Resources

• Code & notebooks: https://github.com/skarwech/ — includes preprocessing notebooks and model training scripts (link to repository).
• Data: selected datasets will be released via the files/ directory; contact me for full raw datasets.
• Thesis preview: download the draft PDF (upload your final PDF in files/thesis-preview.pdf to make this active).

Downloads

Download thesis preview (PDF)

If you’d like to collaborate, access the data, or request the full thesis, please contact me at abderrahim.akhrouf@etu.univ-amu.fr.