Remote Fault Detection of Building HVAC System Using a Global Optimization Program

摘要

The energy efficiency of HVAC systems inbuildings often degrades with the passage oftime. Significant engineering time is normallyrequired to determine the cause(s) of thedegradation. This paper reports a preliminaryinvestigation of a remote fault detectionmethod using the global optimization programSolver? (Frontline Systems, 2000) coupled toa simplified simulation program, which is acoding of the ASHRAE 'Simplified EnergyAnalysis Procedure' (Knebel, 1983). Thisapproach uses the simulation program inconjunction with synthetic measured data toidentify faults in the building operation. Thisfault detection approach has successfullyidentified all of the faulty parameters withnoise levels of 1%, 3% and 6%. Itsuccessfully detected 117 of 120 faultyparameters in the presence of 10% noise.These results are based on acceptable errorsfor each parameter defined as ?5?F for Tcl,?2?F for Tr, ?0.4cfm/sf for Vs, and ?5% forOA(%). If the acceptable errors for eachparameter are reduced to ?2.5?F for Tcl, ?1?Ffor Tr, ?0.2cfm/sf for Vs, and ?2.5% forOA(%), the number of faults successfullydetected decreases to 118 out of 120 faultyparameters with 1% noise, 112 out of 120 with3% noise, 102 out of 120 with 6% noise, and94 out of 120 with 10% noise that wereintroduced into synthetic ?measured data? thatwas generated with the simulation program.