What Can The Internet of Things Learn From Substation Automation?

Connected devices responding to each other and the world around them is like grid automation, writes Bill Ray of ABB.
Published: Thu 21 May 2015

The Internet of Things (IoT) is the buzzword of the moment, but the world it proposes, of connected devices responding autonomously to environmental changes, is already familiar to substation engineers.

A precise definition of the Internet of Things is hard to come by, but the hype around it is easy to spot. Governments around the world are spending serious money striving to attain a cultural lead in the field they are convinced is about to explode, while the energy industry is leading the way in the real work of gathering intelligence from things in and around their networks. Intelligent Electronic Devices (IEDs), Remote Terminal Units (RTUs) and relays are all connected things, spewing fountains of data about the electrical network onto fibre-optic cables for aggregation in supervisory control and data acquisition (SCADA) systems. Many of those devices are using packet data too; not on the “internet” as such, but certainly forming an “intranet of things,” and showing the value of data collection, and learning some important lessons along the way.

Don’t let the mundane swamp the essential

The internet was built on the principle of 'best effort', and lost data is an acceptable price for ubiquitous access. Control networks aren’t the same, so packet protocols used in substations prioritize certain kinds of data.

ABB’s AFS range of Ethernet switches automatically suspend all other communications when a GOOSE message is detected – the GOOSE message may be throwing a breaker or sounding an alarm, and “best effort” isn’t good enough. The internet will need similar techniques if the flooding data isn’t going to obscure what really matters.

Bluetooth socks (I kid you not) feed data on toe-and-heel pressure to a connected smartphone, to reveal the style of the runner. But most of that data is redundant as the runner is only interested in steps that fall outside the ideal. Bluetooth socks aren’t smart enough to weed out that redundant data, but ABB IEDs are.

An IED, fitted in a substation, can be programmed to send sensor readings only when they fall outside predetermined limits. That reduces network traffic to the interesting stuff. One day Bluetooth socks will be equally clever, but, for the moment, every step generates transmitted data, clogging up the network.

Storage space is cheap, but not unlimited. Facebook is fitting racks of Blu-Ray discs to archive our photographs and associated musings, but making sense of big data takes serious software as well as storage.

Facebook and Google are big enough to develop analysis tools in-house, but for utilities, there’s ABB Historian, a new feature of MicroSCADA Pro 9.3 that archives gathered data for later analysis. That enables users to spot trends which might not be obvious when they’re happening, but can matter once they’ve happened.

Intelligent things make for intelligent networks

An electricity grid is, arguably, the largest and most-complex mechanism created by man, so the fact that’s it’s also the most intelligent shouldn’t be too surprising. The quantity of data sloshing around fibre-optic rings is astonishing, but equally impressive is the ability of devices to make decisions without it.

Last year, an unknown assailant cut the communication attached to a 500kV substation and then proceeded to sabotage the station’s transformers and associated gear. Clearly the assailant believed that without a communications link, the central control the substation would be unable to act, but even as the temperature of the transformers began to rise, local systems were talking to each other and decided to shut off the current before more serious damage could be done.

The repair bill for the substation was significant, but without the intelligence of those local devices, it would have been a lot higher.

To realize its full value the Internet of Things is more than just sensors feeding data into massive control centres, and needs autonomous intelligence shared between things capable of communicating with each other as well as a central, controlling, database.

The IoT and the electrical industry are developing rapidly, but there remain many networked devices with limited intelligence and many industrialists trying to work out what to do with the gathered data. The future will see numerous devices flooding the internet with their incessant chatter. We can only hope the end result will be as well run as a modern electrical grid, or at least learns some of the lessons the grid can teach it.