Information Transfer Analysis for Ongoing Commissioning of Buildings

Commissioning ensures buildings deliver the performance and energy savings promised by their design. Ongoing commissioning aims to fully automate the commissioning process, providing periodic or even continuous monitoring, identification and possibly remedial actions ensuring persistence in intended operation and performance . One of the major research areas inongoing commissioning at the moment is that of fault detection and diagnosis (FDD) of heating, ventilating, and air conditioning (HVAC) systems. First generation technology mainly focuses on detection of faults in sensors and actuators with manual diagnosis and remedial actions performed by system experts. In this paper, we propose an information theory method to analyse the information transfer between time series of the sensors and actuators (or the agents) of a HVAC system. The information transfer can help in the diagnosis of faults in the system by automatically identifying the related agents that caused the particular faults in another agent. We present some preliminary results from the proposed method using real data gathered from the a commercial office building in Newcastle, Australia.