Thesis subject
Load Disaggregation of Smart Energy Meter Data (MSc)
Disaggregation of load profiles from smart meters in homes and buildings give more robust insights into the energy performance of those buildings and available capacity on the electricity network, is becoming more congested, for both consumption and generation. Potentially this could lead to smoother integration of decentralized renewable energy resources and make buildings more energy sustainable.
Short description
This thesis will focus on the disaggregation of solar PV power generation profiles from smart electricity meters net measurements (difference between demand and supply) in the built environment using data science and machine learning (ML) techniques. The thesis will validate the effectiveness of the developed data analytics models through real-world case studies (actual smart meter data will be made available by the supervisor). At the end, the thesis will benchmark the developed models with traditional and state of the art techniques.
Objectives
- Investigate current methodologies for load disaggregation of solar PV profiles from net smart energy meter data.
- Construct a dataset comprising energy profiles, and relevant weather datasets in the Netherlands.
- Utilize advanced ML algorithms to develop a comprehensive analytics model for disaggregating solar PV profiles accurately.
Tasks
The work in this MSc thesis entails:
- Study literature on the load disaggregation topic.
- Collect and organize the data required (historical PV generation, smart meter data, weather data, other important external variables).
- Develop a robust ML model for disaggregating solar PV profiles accurately.
- Validate and benchmark the developed model using actual dataset.
Requirements:
- Courses: Programming in Python (INF-22306) and (Big Data (INF-34306), Machine Learning (FTE-35306) or Deep Learning (GRS-34806)).
- Required skills/knowledge: Some experience in data analytics and ML algorithms. Interest in sustainable energy transition and energy informatics.
Contact person(s)
Tarek Alskaif (INF) (tarek.alskaif@wur.nl)