Building system diagnostics through a network of smart local sensors

The study describes the design and implementation of automated fault detection and diagnostics (AFDD) for office building systems. This work is part of larger research focused on distributed digital collaboration framework used in Facility Management. The implementation fully exploits distributed computing to make it possible remote and smart monitoring of the building, anomaly detection, and eventually fault diagnostics. A huge number of data are gathered locally by smart sensors that are integrated and interconnected by the Internet of Things (IoT). Local intelligence allows both anomaly detection and fault diagnostics: identification of the cause of the fault and choice of the corrective action. IoT allows transferring on the cloud processing information, on the higher level of abstraction, in order to execute diagnostics of the entire building system. Distributed diagnostics requires the collection and harmonization of a huge number of feature data and the extraction of significant sequences by Analytics. By feeding the network with relevant data about the anomalies extracted by local intelligent agents and by sharing the information at every level, the resulting AFDD system becomes a distributed computing application.