| Résumé | Fault impact analysis is an important phase of fault detection and diagnosis (FDD), enabling prioritization of the faults detected and allocation of resources to procure services and materials. However, many existing fault impact analysis workflows rely on building performance simulation (BPS) models with high engineering cost, limiting their use in FDD software. To this end, this paper introduces an inverse greybox model-based fault impact analysis method using building automation system (BAS) trend data from variable air volume (VAV) air handling unit (AHU) systems to quantify the heating and cooling energy impact of control system faults. It trains inverse models of a VAV AHU system by employing a sequential parameter estimation algorithm, formulates virtual meters quantifying the sensible heat transfer rate from AHU heating and cooling coils and the perimeter heating devices, and employs the virtual meters to estimate heating and cooling energy use of an existing VAV AHU system and the same system after fault correction. The method was used with a synthetic BAS dataset generated by using an EnergyPlus model of a 26-zone VAV AHU system in Ottawa, Canada. The results indicate that the virtual meters could estimate the monthly energy use for heating and cooling at a coefficient of variation of the root mean squared error (CV(RMSE)) of 4–6%. The method was demonstrated with measured BAS data from a 72-zone VAV AHU system, revealing an energy savings potential of 43% for heating and 22% for cooling through fault correction. |
|---|
