Visit the Energy Systems Laboratory Homepage.
Black Box Approach for Energy Monitoring of Commercial Buildings
MetadataShow full item record
The potential to save energy by changing operational parameters - especially in existing commercial buildings – is in the magnitude of 5-30%. In order to realize this saving potential in the long term, continuous commissioning of the building is a key issue. Necessary for successful continuous commissioning is real time monitoring of the building performance which allows for Fault Detection and Diagnosis (FDD). This paper presents a method to monitor building operation and detect faulty or unusual behaviour using a black box model approach. The approach is to identify a building’s basic operating characteristics by means of measured data from a building to train a multiple linear regression model based on energy signatures of the building. In addition to supplying measured building data to the regression a clustering process is added which determines the building’s day-types. Once the model is trained it can predict the energy consumption at the building site and unusual or faulty days can be identified by comparing the predictions to real measurements. Models to monitor the daily heating and electricity demand are developed and applied to measured data from two demonstration buildings.
SubjectMultiple Linear Regression
Black Box Models
Komhard, S.; Neumann, C. (2008). Black Box Approach for Energy Monitoring of Commercial Buildings. Energy Systems Laboratory (http://esl.tamu.edu). Available electronically from