Prøveforelesning og disputas – Nasrin Kianpoor / Trial lecture and defense – Nasrin Kianpoor

Nasrin Kianpoor disputerer for ph.d.-graden i ingeniørvitenskap og vil offentlig forsvare avhandlingen / Nasrin Kianpoor will defend his thesis for the PhD degree in Engineering Science:

“Machine Learning-Based Flexible and Intelligent Energy Distribution Systems for Residential Buildings”.

Avhandlingen er tilgjengelig her (lenke kommer) / The doctoral thesis is available here (link is coming).

Auditoriet er åpent for publikum. Disputasen vil også bli strømmet. Opptak av disputasen vil være tilgjengelig i en måned. / The auditorium is open to the public. The defense will be streamed. A recording of the defense will be available for one month.

Prøveforelesningen starter kl. 10:15 / The trial lecture starts at 10:15. Tittel / title:

“Designing Intelligent HEMS under Uncertainty: Probabilistic Load Modeling and Robust Optimization”.

Disputasen starter kl. 12:15 / The defense starts at 12:15.

Prøveforelesning strømmes her, disputas strømmes her. / The trial lecture will be streamed here, and defense will be streamed here.

Sammendrag av avhandlingen / Summary of the thesis:

Buildings and residential sectors have considerable potential for demand-side flexibility, which can be used for optimization of electric energy usage and electricity cost reduction. In this context, Home Energy Management Systems (HEMS) can provide solutions to manage electricity use efficiently by controlling flexible assets and responding to dynamic electricity price signals. This thesis investigates methodologies for non-intrusive load monitoring (NILM) and forecasting of short term load demand to enhance the performance of HEMS in residential buildings, particularly under the challenging conditions of the Arctic climate in Northern Norway.
The research is based on four key contributions. First, a new dataset was measured and prepared to reflect the climate characteristics of Northern Norway, which is the basis for evaluating the proposed methodologies. Second, a Non-Intrusive Load Monitoring (NILM) approach is developed using a combination of deep learning and signal processing techniques to disaggregate the electricity usage of individual household devices from the aggregated load. The third contribution involves developing machine learning-based models for forecasting short-term load demand to predict the aggregated power consumption of residential buildings. Finally, an optimization algorithm is proposed to enable smart charging of an electric vehicle to minimize overall electricity costs. The findings demonstrate that the proposed NILM methodologies, especially those incorporating signal processing methods, achieve higher accuracy in identifying appliance-level consumption than approaches that do not utilize signal processing techniques. Furthermore, the optimized EV charging strategy significantly reduces electricity costs by flexible demand rescheduling to off-peak periods. The results highlight the benefits of economic efficiency and environmental sustainability achieved through advanced machine learning methods and smart energy management systems in residential settings. By addressing the challenges of energy management in cold climates and validating the proposed methods with real-world data, this thesis offers valuable insights and practical solutions for improving the performance of HEMS, contributing to the broader goal of creating energy-efficient and sustainable residential buildings.

Veiledere / Supervisors:

Hovedveileder / Main supervisor:

As. Professor Bjarte Hoff, Department of Electrical Engineering, UiT The Arctic University of Norway,

Biveileder / Co-supervisor:

As. Professor Trond Østrem, Department of Electrical Engineering, UiT The Arctic University of Norway.

Bedømmelseskomité / Evaluation committee:

• Dr. Ingrid Munné Collado, senior ML engineer at Electricity Maps ApS – Univate, Copenhagen, Denmark, 1st opponent,
• Dr. Harsha Kumara Kalutarage, Senior lecturer at School of Computing, Robert Gordon University, Aberdeen, UK, 2nd opponent,
• Prof. Bernt Arild Bremdal, UiT The Arctic University of Norway, Faculty of Engineering Science and Technology, Department of Computer Science and Computational Engineering, internt medlem og komiteens administrator / leader of the committee.

Prøveforelesning og disputas ledes av prodekan for forskning, Svein-Erik Sveen /

The trial lecture and defense are led by Vice Dean of research, Svein-Erik Sveen.

De som ønsker å opponere ex auditorio kan sende e-post til Svein-Erik Sveen / Opponents ex auditorio should contact Svein-Erik Sveen.