Paper Conference
Proceedings of SimBuild Conference 2004: 1st conference of IBPSA-USA
 
 |
Usage of Building Energy Simulation for HVAC Fault Detection
Seung Uk Lee, David E. ClaridgeADM Associates, Inc., Sacramento CA, USA Energy Systems Laboratory, Texas A&M University, College Station TX, USAAbstract: Building energy simulation can be powerful when it is combined with an implanted optimization function. The potential of this combination is tested and analyzed, using a simplified simulation program combined with a global optimization tool. The first test is automatic calibration of the building energy simulation model using global optimization. Automatic calibration can be helpful in quickly generating an accurate energy simulation model. The second test is fault detection of an HVAC system using the automatically calibrated simulation model and an optimization function. The experiments for the automatic calibration and fault detection are performed using simulation data. Small amounts of artificial noise are added to the simulation data to approximate model mis-specification. The method succesfully detected 75-85% of the faults tested when the noise levels present generate scatter in simulated heating and cooling data that approximates that observed in measured data. Hence it seems likely that the approach will often identify the fault but will need to be supplemented by site visits to complete the diagnosis. Pages: 1 - 8 Paper:simbuild2004_SB04T4B2