MQTT Architecture for Stream Analytics of PMU Data

Synchrophasor data, derived from Phasor Measurement Units, offers great potential to provide insights into the operation of electrical power systems in the presence of high capacities of renewable generation. Accessing synchrophasor data is challenging since established protocols are rigidly hierarchical and `point-to-point', based on serialized methodologies requiring `data concentrators'. This makes real-time and in-stream analysis of synchrophasor data very cumbersome. An alternative is explored in this paper. This novel method is database independent, and allows for synchrophasor data transport, analysis and archival. Using a publish-subscribe MQTT broker, real-time stream analytics are made available in real-time through republishing to new MQTT topics. The advantages of MQTT as a development and deployment platform are presented; in terms of scalability, computational efficiency, latency, modular development, data rate and failover. The dependency chain of an event detector is presented in detail, highlighting the JSON communication protocol used and the implemented topic structure. An intention of this paper is to present the topic name convention and JSON schema employed and invite debate on conventions.