Machine Learning Modelling for HVAC Systems
Ekkerman, Genrikh (2021)
2021
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:amk-2021060213499
https://urn.fi/URN:NBN:fi:amk-2021060213499
Tiivistelmä
Vuores is a school building with the HVAC system deployed as part of its building automation system. Indoor conditions are maintained by ventilation and heating systems, which consume a major part of the total energy consumption of the building.
For solving this problem linear regression models were developed for the key parameters of the HVAC system. Models included battery network energy, for future optimization of energy consumption, air temperature and CO2 concentration models for future forecasting Indoor air quality parameters.
For developing these models, statistical and machine learning methods were used. Features for each model were selected using a repeated KFold cross-validation method. Fitted models were successfully evaluated on the unseen data, from the same source.
For solving this problem linear regression models were developed for the key parameters of the HVAC system. Models included battery network energy, for future optimization of energy consumption, air temperature and CO2 concentration models for future forecasting Indoor air quality parameters.
For developing these models, statistical and machine learning methods were used. Features for each model were selected using a repeated KFold cross-validation method. Fitted models were successfully evaluated on the unseen data, from the same source.