LARES STUDENT

Sara Orlić

  • Precipitation nowcasting based on radar composite imagery and deep neural networks, master's thesis, 2024
  • Precipitation nowcasting based on radar composite imagery and deep neural networks, master's project, 2024
  • Prediction of wildfires using satellite imagery, master's seminar 2, 2023
  • Machine learning in nowcasting, master's seminar 1, 2022
  • Model predictive climate control of a smart residential unit, bachelor's thesis, 2022

BACHELOR'S THESIS

Model predictive climate control of a smart residential unit

The thesis describes the procedure of model-based predictive control for a residential unit. Simulated data from the IDA-ICE program is used to perform system identification. The identification process is carried out in the Matlab environment using a non-derivative Kalman filter, which yields a state space representation for further calculations. MPC takes into account future states to better predict the system behavior and reduce costs. Finally, the sensor data is compared with the data generated by the MPC algorithm.