Real-time and Retry Don’t Mix: Q&A with gridComm CEO Mike Holt
The willingness to shake off the old ways of doing things may be the only limit for what IoT sensors on a power line network can gather, analyze, and put to smart energy cost- and carbon-saving use.
Pubs, restaurants, music—many ways exist to enjoy a city after dark. Not among these is driving around all night, night after night, noting which streetlights are out. Mike Holt, CEO of Singapore-based power line communications company gridComm, knows about this not so entertaining after dark activity and how to eliminate it. Not long after gridComm announced the role its Orthogonal Frequency Division Multiple Access (OFDMA) PLC transceiver would play in turning Jakarta’s streetlights into smart streetlights, Holt spoke with EECatalog.
EECatalog: gridComm is involved in smart city initiatives. What do cities embarking on these initiatives want to achieve?
Mike Holt, gridComm: Cities are first focused on two things in smart city initiatives. One is smart metering, knowing where the power is being used and the second is streetlight control (because cities use up to 40 percent of their electricity budget to light the streets). Then, once you go beyond just the basic streetlight control part of it, cities have the opportunity now to utilize the power line as a network that forms the infrastructure for smart city implementation—with IoT sensor connectivity. Using the power line as the infrastructure for citywide connectivity is the most scalable and most cost effective way of doing this, even above using wireless to create this kind of a network.
EECatalog: What are some of the things that are possible as a result of networking streetlights over a power line?
Mike Holt, gridComm: There are hundreds of millions of streetlights in the world, and more than a third of these are being upgraded to connected LED streetlights—for power reasons, for maintenance reasons, for cost reasons (Figure 1). Cities can see a payback period of three, maybe at most four years in implementing these upgraded connected streetlights, and they see that payback period through reduced maintenance costs and reduced electricity costs. Streetlights use about 40 percent of a city’s electricity collection. How many power plants get built, reducing carbon emissions—these are big issues cities have to grapple with.
EECatalog: What are some the specific ways to cut street lighting costs?
Holt, gridComm: In the case of gridComm, we use sophisticated data analytics, IBM’s Bluemix, to measure things like the voltage, the current, the burn hours—these kinds of things from streetlights, but also, we are combining it with other data, things like traffic, weather and ambient light. These are all gathered, and from this data analytics we can then set a dimming profile, by individual streetlight automatically that changes over time and over the day to optimize how much light is being used and how much money is being spent.
And we can determine the optimal maintenance schedule, predictively. Right now the way it works for streetlights is—to cite one example—a city has five employees whose job it is to get in their cars as soon as it gets dark, drive around all night and look for burned out streetlights, and then the city will schedule a replacement or whatever they have to do to repair those streetlights. That is the way it’s done today. What we enable is predictively determining a maintenance schedule based on all these parameters and the IoT data that is measured to be able to say, for example, “These 10 streetlights will have an issue in the next 30 days, schedule them together for maintenance, and oh by the way, maybe they’re failing prematurely and you ought to get a refund [from the manufacturer],” and it could correlate to the specs of the light.
EECatalog: What do you see as the significance of power line connectivity for embedded systems developers?
Holt, gridComm: Say you are a company that has developed a very interesting ozone sensor and you say, “We want to deploy a bunch of these ozone sensors, how do we get them deployed?” When a city has power line connectivity, it’s a much simpler discussion [than it would be should power line connectivity not be in place].
The discussion then becomes, “Hey, you have 100,000 streetlights that are all networked over power lines, the proposal is that we will put an ozone sensor on each one of these streetlights, and all of that data will come to our application that runs in your city command center that collects all the ozone data by type or local region and that allows you to make decisions.”
EECatalog: What hurdles have to be overcome to leverage the advantages of power line connectivity?
Holt, gridComm: Communicating over a power line is very difficult because of noise. Historically, this has been all right because the first applications for this were about meter reading. So if you have a city with a million meters, as long as you read each meter once a month you’re okay. So if it doesn’t work on Monday, try Tuesday, Wednesday, Thursday….eventually you are going to be able to read it if you keep retrying, as you only need to read the meter once a month.
But that is not where the world is now. Where the world is now is it’s all about real-time data, and that real-time data is what you need to do any of these IoT applications. Real-time is what you need to do energy management as opposed to meter reading. Or to do streetlight control where you are dynamically controlling the use of power—this involves real-time communications and real-time doesn’t mean retrying—so you have to have reliability.
EECatalog: How is gridComm addressing the challenge of maintaining reliability?
Holt, gridComm: We adapt to noise, and changes in noise, in a way that allows data communications to always happen over the power line regardless of the noise—so the reliability is much higher than [what’s possible] with other competing solutions.
Other solutions use a single channel. They divide one channel over the power line, and because there is only one channel, if there is noise that interferes with that particular channel, that frequency that is being used, then there will not be any data communication—because there is only one channel—there is no choice.
We divide the power line into 18 independent channels and with these independent channels we implement redundancy across those, so even if three of the channels have noise that prevents data communication, it is okay because we have 15 more.
Now in a very clean environment involving dedicated power line communication as opposed to the public grid, it may be very, very clean and noise free, in which case each of those 18 channels could be independent, not redundant, and that allows 18 times the data rate, so you get higher performance.
But—and this is what our software, our networking capability does—as soon as noise comes up that affects performance, we can change the configuration to adapt and add that redundancy real-time to make sure the data goes through.
EECatalog: Street lighting, such as with the city of Jakarta, is one market where gridComm has been involved, but some of the features you’ve just been discussing would be applicable in other areas, correct?
Holt, gridComm: Yes. Right before I walked into this call one of our new customers approached us. This company does embedded systems development for sensors that go on the head of a drill such as [on machinery] that drills for oil, and its challenge is [finding a way to communicate] the data from the sensors to a command center where people make decisions about how they are drilling. And that is a hard thing to do. The company can’t do that wirelessly because the drill is under the ground, and it wants to use power lines—so this company will be buying our modules for its embedded systems to enable this connectivity, so that its application, which enables smartness around drilling, has the data that it needs based on gridComm connectivity.