Buildings Electrical Energy Forecasting using Machine Learning

Improve forecasting algorithm for building consumption to include EV demand

The goal of decarbonizing our energy, calls for the integration of more renewable energy in our energy system. On-site energy generation is becoming more and more the norm for the Dutch buildings. The stochastic nature of renewable energy sources and the limitations of the grid, forces designers to look into demand side management and optimization of energy consumption. To optimize the consumption of the buildings, predicting their energy demand beforehand is valuable.

General project problem description

The project is a collaboration with Kropman building services company. Kropman has been one of the leading forces in the Dutch construction industry towards energy transition and has created a living lab to develop, test, validate and implement new technologies and processes.

The living lab includes on site energy generation through PVs and energy storage through electric batteries. An intelligent building management system is in place that monitors different aspects of the building. A forecasting algorithm is already available to predict building consumption for the next day, but the EV demand is not included.

The student will work on a dataset from Kropman office Breda and use the existing model as a base case to develop a solution which outperform the existing forecast and includes the EV demand.

Types of student project possible

Fontys: Internship, Graduation projects
TU/e: Bachelor project, Master internship, Master thesis

Types of study
  • Electrical engineering
  • Mechanical engineering
  • Mechatronics
  • Applied mathematics
  • HBO-ICT
  • Computer science and Engineering
  • Data science
  • Other study background fitting with the assignment
Starting date/duration of student projects

Flexible depending on the requirements of students’ institutions.

Contact

Wim Zeiler, w.zeiler@bwk.tue.nl 

Register as a student