IoT sensor data enables to optimize thermodynamical models

Riga Technical University (RTU) has elaborated and optimization algorithm for model estimation describing a home / building thermodynamics during heating season in Latvia. The model is based on so called 5R1C model proposed by Piotr Michalak (AGH University of Science and Technology, Krakow, Poland), which expresses building as RC electrical circuit. RTU has extended the model to make it more flexible and more realistic allowing to model state transitions, which are important in large buildings.

The extended model has been populated with data from sensors located in one of RTU’s campus buildings and combined with weather station data as well as complemented with real building use data. The gathered data provide a “ground truth” to evaluate the developed model and apply optimization for its finetuning. The first experimental results shows that the used approach provides a straight forward way to acquire precise models allowing to forecast actual energy consumption and make proper control decisions during heating season.

Leave a Reply