Abstract
It is important to have practical methods for constructing a good mathematical model for a building's thermal system for energy audits, retrofit analysis and advanced building controls, e.g. model predictive control. Identification approaches based on semi-physical model structures are popular in building science for those purposes. However conventional gray box identification approaches applied to thermal networks would fail when significant unmeasured heat gains present in estimation data. Although this situation is very common and practical, there has been little research to tackle this issue in building science. This paper presents an overall identification approach to alleviate influences of unmeasured disturbances, and hence to obtain improved gray-box building models. The approach was applied to an existing open space building and the performance is demonstrated.
Original language | American English |
---|---|
Pages | 50-55 |
Number of pages | 6 |
DOIs | |
State | Published - 2017 |
Event | 2017 American Control Conference (ACC) - Seattle, Washington Duration: 24 May 2017 → 26 May 2017 |
Conference
Conference | 2017 American Control Conference (ACC) |
---|---|
City | Seattle, Washington |
Period | 24/05/17 → 26/05/17 |
NREL Publication Number
- NREL/CP-5500-69120
Keywords
- buildings
- heating systems
- mathematical model
- predictive models
- temperature measurement
- thermal resistance
- thermostats