Algorithms for Green Buildings: Learning-Based Techniques for Energy Prediction and Fault Diagnosis
Daniel Holcomb, Wenchao Li, and Sanjit A. Seshia. Algorithms for Green Buildings: Learning-Based Techniques for Energy Prediction and Fault Diagnosis. Technical Report UCB/EECS-2009-138, EECS Department, University of California, Berkeley, 2009.
Download
Abstract
We consider two problems in the design and operation of energy-efficient buildings. The first is the predictionof energy consumption of a building from that of similar buildings in its geographical neighborhood. The secondproblem concerns the localization of faults in building sub-systems with a focus on faults that lead to anomalousenergy consumption. For both problems, we propose algorithmic techniques based on machine learning to addressthem. Simulation results using EnergyPlus show the promise of the proposed methods.
BibTeX
@techreport{holcomb-tr09, Author = {Holcomb, Daniel and Li, Wenchao and Seshia, Sanjit A.}, Title = {Algorithms for Green Buildings: Learning-Based Techniques for Energy Prediction and Fault Diagnosis}, Institution = {EECS Department, University of California, Berkeley}, Year = {2009}, Month = {October}, Number = {UCB/EECS-2009-138}, Abstract = {We consider two problems in the design and operation of energy-efficient buildings. The first is the prediction of energy consumption of a building from that of similar buildings in its geographical neighborhood. The second problem concerns the localization of faults in building sub-systems with a focus on faults that lead to anomalous energy consumption. For both problems, we propose algorithmic techniques based on machine learning to address them. Simulation results using EnergyPlus show the promise of the proposed methods.}, }