Ask Hackaday: Is Our Power Grid Smart Enough To Know When There’s No Power?

Just to intensify the feeling of impending zombie apocalypse of the COVID-19 lockdown in the British countryside where I live, we had a power cut. It’s not an uncommon occurrence here at the end of a long rural power distribution network, and being prepared for a power outage is something I wrote about a few years ago. But this one was a bit larger than normal and took out much more than just our village. I feel very sorry for whichever farmer in another village managed to collide with an 11kV distribution pole.

What pops to mind for today’s article is the topic of outage monitoring. When plunged into darkness we all wonder if the power company knows about it. The most common reaction must be: “of course the power company knows the power is out, they’re the ones making it!”. But this can’t be the case as for decades, public service announcements have urge us to report power cuts right away.

In our very modern age, will the grid become smart enough to know when, and perhaps more importantly where, there are power cuts? Let’s check some background before throwing the question to you in the comments below.

A Power Cut On A Sunny Afternoon

In the aftermath I was discussing it among my neighbours, and a difference of opinion emerged over whether our electricity provider has automatic monitoring of their network or whether we need to call a fault in.

We really have reached the end og the line.
We really have reached the end of the line.

It seems the power company operators had issued the advice that there was no monitoring, but my mind had gone back to a summer afternoon in the mid 1990s. Our then neighbour was a lifelong farmer with decades of conjuring a decent yield from the unforgiving heavy Oxfordshire clay behind him, but that was not his lucky day. The long CB whip antenna on the roof of one of his machines touched the village’s 11kV supply line, and our power went out. I never saw what it did to the unfortunate CB transceiver, but I remember him talking about his surprise as when he reached the farmhouse to call it in he found the electricity people calling him instead. Their monitoring equipment had detected the fault and narrowed it down to his property.


But the power people here in 2020 told my current neighbour we have none, and it’s obvious both can’t be true. It’s most likely that they have it on some circuits but not others, but that guess doesn’t really put the whole question to bed.

It’s easy to be certain that there is no power monitoring equipment on any of the poles in our village, by simply looking at them and seeing no untoward equipment in place. We’re in the midst of a nationwide rollout of smart meters so it’s possible that if any of our neighbours have one it might have called home over the cellular data network and alerted their operators to the loss of power. But since the rollout is optional and there is no compulsion on a consumer to have one it’s by no means certain that there are any in the village. It’s thus safe to assume that any power monitoring equipment is centralised rather than distributed. And with that it’s a reasonable guess that some form of time-domain reflectometry would be in use.

Bouncing Pulses Along Your Power Lines

Time domain reflectometry demonstrated on the bench, with the original pulse on the left of the screen and its reflection on the right.
Time domain reflectometry demonstrated on the bench, with the original pulse on the left of the screen and its reflection on the right. Constant314 / CC0

Time domain reflectometry is an extremely simple process that relies on the property of a traveling waveform to bounce off the end of a transmission line and be reflected back to its originator. It’s a standard lab experiment for electronic engineering students that can easily be replicated with a pulse generator, a coil of cable, and an oscilloscope. Adjust the ‘scope to see the end of the generated pulse, and there a short time later will be its reflection. The time between the end of the sent  pulse and the arrival of the reflection is the time it has taken to travel the length of the cable and back again, so from that the length of the cable can be calculated.

Since the pulse will reflect from any faults, the distance to the fault can also be worked out. Once the breaker has been triggered by the fault the electricity company can measure the distance, and send out a repair team to fix it. That’s exactly how an El Segundo steam plant located a fault in ten miles of buried cable.

So goes the theory, but something tells me that even with state-of-the-art equipment it is unlikely to be that simple. For example, is our 11kV power distribution to the 230V transformer in the village a linear one in which a succession of farms and groups of houses as it gets further from the substation tap in with their own transformers, or is it a branched topology in which the line splits, and splits again? Time domain reflectometry would be useless when there are multiple ends of the line. Is the only way to deal with that by waiting for a human (or a smart meter) to phone home?

I’m told mains electricity finally arrived in this village in the 1950s, so I am fortunate to be of a generation for whom it has always been there. It’s so much part of the scenery that we don’t really notice it and certainly don’t appreciate the effort that must lie behind it, so I’m slightly ashamed as an electronic engineer that I don’t know more about its technology. Power engineers, now is your time to shine, and tell us something about your art!

Satisfy my curiosity, Hackaday, what’s the current State of the Art when it comes to automatic power distribution system monitoring? Share your stories in the comments below.

63 thoughts on “Ask Hackaday: Is Our Power Grid Smart Enough To Know When There’s No Power?

    1. When Poland switched from 220V to 230V to conform with EU, at one substation someone made a typo and by accident a neighborhood was switched to 320V. In most cases light bulbs blew along with fuses, but some devices were damaged too. Electrical company at first refused to pay for damages, but when the case was exposed by a TV station, they paid up…

      1. But the EU standard is 230 +10% to -6% Or from 216.2 volts to 253 volts… So 220 is within it, no reason to change.

        And yes, the EU power standard is made to almost not be a standard, since its tolerance is so wide to cover practically all countries that it were supposed to cover. (Its actually made so that equipment manufacturers would be required to make their products usable over that whole range, so that all countries within the union could use said device.)

        It were literally made so that no country in the union would need to change anything, while still being “harmonized” on paper.

        In practice, its close enough to not matter all too much to be fair. (And most equipment manufacturers already designed their products to either support 220 and 240 (with a selection switch), or just center on 230. (and usually have a switch for running on 110-120))

        And in modern times with switch mode power supplies that can run literally anywhere from 90 volts to 250 volts makes it all even less problematic. (Fun fact is that most switch mode power supplies can also run on DC voltages between 120-350 volts.)

        1. It was a cosmetic change for the most part, but usually mains voltages are within 2-5V of the official value. So they did it for the legal reasons not because it matters. Also this separated Poland from Russia which, IIRC, is stuck at 220V…

          As for SMPS, they usually can go as low as 65 volts…

          1. Almost whole Eurasia now switched to 230/400 except countries which have 240/415 and only China and some country from CIS which still use 220 except Russia.
            About SMPS I highly doubt they are will be work without modification as much efficient as even worse and cheapest buck(-boost) converter on DC transformer along should do a better job.

  1. I work for SEL, and as far as I know there is no active Time domain reflectometry devices on the power grid. We just released a product that will do it passively, aka if there is a fault (insulation breakdown) on the line we can see where it is based on reflections, but that depends on how the power fails. It is also only for transmission lines, which are point to point.

    There are other technologies that can detect where a fault is based on the impedance of the line to the fault.

    Other than that you have to have remote battery backed up devices with intelligent controls.

    Its possible your power company does not have remote battery backed up devices with intelligent controls for detection of any outage, but does have (or did have) impedance measurement devices for fault detection.

      1. I work for a utility in the states doing among other things, distributed fault monitoring and sensing. We have current sensing devices that are attached to a line and use the e-field to both sense everything and power themselves. These devices are rather new and require placement on each lateral to locate faults more precisely. Smart meters are capable of reporting power out conditions, but if you’re just rolling them out, the system to move that data to the outage coordination center may be missing.

  2. I have NO idea what the grid in the UK is like but in the US, grid operators know when some portion goes down and demand changes. The “utility” may want reports to help pinpoint downed lines and poles, but even those they know about. Reporting helps expedite “boots on the ground” and verify what they already know.

  3. My neighborhood has “smart meters”, a satellite box on a nearby street lamp gathers our consumption readings and beams it up to the billing office. I doubt the smart meters or lamp post transmitter have the ability to report that they don’t have power. How often the “street lamp” reports to the electric company I don’t know, once a day? So, relying on the absence of a signal to report the location of an outage would be a long delay.

      1. Virtually all “smart meters” do have a battery or supercap backup and when we were evaluating candidates for deployment, each system did support a dying-gasp report of power down. They also rerouted, since each meter talks to other meters along the path to get to the collector device, rather than each talking directly, in the event that a meter along the path was in an outage, but others were not.

  4. It is not. Power went out two weeks ago. Called FPL and they said it is just my house because they could not contact the smart meter. Look on the app and 10 min later 350 people where out then 1700 in my area. I should get a free week for being the first to report it :)
    Yeah we call FPL Florida Flicker and Flash. A few years ago a tree near here caught fire because FPL had not trimmed it and it had grown around the power lines. This was on an Empty piece of land and the right of way should be their responsability.

    1. LWATCDR – I worked in FPL Distribution Information Management serving Distribution Reliability, not Outage Management or Vegetation Management, but I can tell you that they DO use meter last-gasps, the radio signal sent using a battery or super-cap, to notify them of outages and attempt to determine the location of the problem. The Customer Service Reps (CSRs) CAN query the status of your smart meter, but they would NEVER say there was no outage just because they queried your meter – they have other tools available, not the least of which is the outage system that would have given them the ETR (estimated time of restoration) that they quoted you during your phone call. Perhaps you misunderstood what they were telling you.

      FPL doesn’t do tree trimming; they subcontract that work. I’ve had the same type of “something not trimmed” issue that you did, with a pine tree in my neighbor’s yard sending flaming debris down onto dry grass on their right of way. I called and spoke to a CSR, and they sent out the on-call worker for our area to disconnect the power (no more flames, thank you very much) and someone cut the branches that were close to the line. Given that vegetation issues are the #1 cause of outages for virtually ALL US power companies, “it happens”.

      FPL has been proactive in trying to minimize outages. They have installed intelligent reclosers (basically, a circuit breaker that attempts to “retry” energizing power to your line in the event of a fault) all over their distribution system. If you’re an FPL customer, whatever FPL does can never be “enough”, but they do “try”.

      I never knew how “good” FPL service was by comparison to other power companies until I quit and moved to Spokane, WA…

  5. Where I’m at in the US we get a txt from the utility inside a few minutes anytime the power goes out. They just started doing it not to long ago but we got smart meters a few years ago. The routers on the utility poles showed up sometime after so I’d assume it was roled out in phases with the meters being first. And all of it is based on 6LoWPAN tech so I’d assume very realtime.

  6. “In our very modern age, will the grid become smart enough to know when, and perhaps more importantly where, there are power cuts? Let’s check some background before throwing the question to you in the comments below”

    Considering it may be one of the few ways to sneak broadband past the “block all competition” incumbents the grid may already be smart, just not available to us.

  7. I worked at a company called Maxwell Technologies in the 80’s. (They are now part of Tesla). They had a product for finding underground faults. The problem with a TDR is it’s low voltage. And the fault may require high voltage to conduct. So they had big truck mounted power supplies they would connect to underground cables and burn in the fault. They would then do TDR which gives you an idea of the location. They would then use magnetic field detectors above ground to follow the line to the fault. Then dig.

    1. I used to design overhead power lines in the UK (mechanical side, not electrical though I’m afraid). It’s been 20 years since I moved on but even then the power companies used a similar system on the distributions lines, they would mount magnetic field detectors on poles beside roads. They were about milk bottles sized with a light on top, when the detector couldn’t detect the field it would start to flash. That way as the engineers drove around trying to find the fault they could trace it easier.

    2. When a fault developed in the underground wire in front of my house the local utility used a “thumper” to find it.
      I’m not sure what they meant by “thumper”.
      They showed me the piece of defective cable they removed, it was more than a half inch in diameter, and the fault looked like someone had drilled through the insulation to the conductor with a 3/16″ drill without leaving any debris in/around the hole.
      They said it may have been caused by a lightning strike miles away that found a tiny defect at that point.

        1. They called thumpers because they audibly thump. You find the fault for listening to the thump under the ground.

          Basically its a big cap being discharged into the fault.

  8. The answer is “it’s complicated” as usual.

    Power Grids work on demand but most demands aren’t big enough to “see” individually. A group of Air Conditioning units turning off at the same time can look just like any other city block that just lost power. These are small loads for a grid. There is no smarts in most small transformers like those on a power pole supplying a set of houses. If one of those fails they often disconnect by burning up a link that works just like a fuse. For these, there is often no way to see the outage.

    Keeping in mind that power system monitoring is in aggregate helps. They often see power consumption averaged over time. Phenomenons happen that are amazing. For England as an example, large sports events see large spikes in power consumption during a commercial as many start their electric kettles to make tea. When they stop, it looks like one or more large city blocks just lost power. Determining whether that’s just a lot of tea being done or an outage is a not so simple proposition.

    Many smart meters have a dying gasp that can alert though. So areas where smart meters are installed can absolutely tell pretty quickly. Most can use the power in some relatively small capacitors to sense and alert on this.

    Also when the outage happens at a major point of power system protection they can definitely tell as well. These are usually much larger issues though.

  9. I’m in Georgia near the Atlanta area. We had a power outage a few weeks ago; the power company had somebody later knocking on doors after the power was fixed to manually confirm everyone had their power back. So if they had a way of determining which houses had power, it looks like they weren’t 100% confident of the results.

    1. If it went out with a surge it’s possible it could have popped peoples main breakers fractionally before the line went down. So I guess they were letting people know that the power was indeed back on, so they could reset their breaker if needed.

  10. Although it is complicated, TDRs can deal with taps on a line. In fact there are methods to measure the taps lengths. In my area of the United States, the power meters are interrogated over the power lines themselves and the power company can quickly see the loss of connectivity in the network. They can also narrow down the point of failure by the meters that are interrupted. In fact my power company has a website that actually shows you a map of the outage area and within minutes they often know if they have a downed line or faulted transformer. Seems to be pretty accurate.

  11. I live in Sweden, last time I had a power outage, it were a fuse in my own house for only a few sockets. I guess the power company had no clue about it.

    Though, a few years ago, a fairly decent portion of the city went down on winter due to a transformer getting a bit toasty. (Ie, over current protection tripped in… They were servicing one of the other transformers, so no wonder the others were under a bit more stress than usual.)

    In the end, power were out for about 2 hours, but I guess when power goes out at a major substation that the power company has a good idea of where the fault is…

    Other than that, not a single power outage in my area for the last couple of decades. So I guess the grid operator does a decent job.

  12. We had a faulty ring line to our flat (300 appartments and a nearby community building), someone put a traffic sign post right next to the underground cable, and when there was enough wind, something would happen and our buildings 10Kv->230V fuses would trip (as explained by the electrician when they finally repaired it).
    This happened 10 times over the course of a year, and then 10 times over the course of a week.

    I guess half of these times, I was the first to call and usually the conservation went like
    – “Have you checked your RCD?”
    Me: “I can see the entire 300 apt flat is dark. It’s not my RCD”
    – “Can you please check it for me, just in case”
    Me (second time this question came up just making noises as if I walked to the meter and flicking a light switch to stimulate RCD sound) : “It’s not the RCD. The whole flat is without light. Can you send someone over?”
    – “I haven’t had any other reports. I’ll wait till someone else calls in”
    Me: “It’s 2AM. Most people are asleep. This is what happened the previous times too and then we didn’t have power for 5 hours. Just send someone over”
    – “I don’t see any earlier reports on your address”
    Me listing a few earlier reports, with date, time and incident code which I had written down from their website.
    – “Oh, I see. We’ll send someone over.”

    This was between 2006 and 2009 and the fuses that went were those ceramic fuses you see in British plugs, just bigger, inserted with a special tool on a long-sleeved insulating glove that made a lot of sparks when not installed thoroughly enough.

    At last, they had to dig up the entire ±500m of cable
    between our flat and the community building to find and fix the fault.

    What I still don’t understand is how that traffic sign pole managed to not make a dead short but occasionally wiggle the cables such that fuses blew. I’ve once seen the three phases go out with about 20 seconds between them, first one phase starting to flicker, then go out, then the next flickering and out, and then the third, as each third appartment in this flat is on the same phase apparently.

  13. If the outage involves a blown fuse, then there might be monitoring equipment to tell when the fuse is blown.

    However, if the outage involves broken lines, then there might not be monitoring to determine that.

  14. I have worked in power distribution operation, maintenance and control for thirty years in Australia.
    In a rural distribution network (11kV and 22kV) the normal devices to protect the network against CB antennas, cranes, trees, wildlife and the like are fuses and reclosers.
    A fuse operating will not alert anyone until the control room (network operator) is informed, when smart meters and the IOT dream is fully realised the operators will have full visibility down to the individual customer or section of the feeder, I think this ambition is some time off. Consider an individual operator may be monitoring 100,000 or more customers on his network segment, an alarm for a single customer being off is possibly not his highest priority – but if a group of customers in one area are off then that will interest him.
    The second most common device (in rural Australia) is the recloser, these open for a fault and then have another go to see if the fault has cleared (about 80% of the time) usually after a couple of relcose attempts they will lock out and require manual intervention to restore supply. These have been around a long time the original reclosers were electro hydraulic and not remote monitored, a modern recloser such as the Noja OSM (I do not work for Noja) can be fitted with a communications module to inform the control room of its status plus a heap of other useful information, an can be operated remotely if the operator thinks it is prudent to turn the 11kV back on even though an automated device has failed a few times already.
    The protection settings in a basic recloser are overcurrent and earth fault, the earth fault being determined by the vector sum of the three line currents. A more sophisticated recloser is likely to have additional protection functions available, some of which may trip the recloser or simply send an alarm to the control room.
    In the case of the CB antenna, it is likely the upstream protection operated and detected the momentary vapouristation if the antenna and CB as an earth fault, tripping a protective device that was monitored.
    Call the fault in, there have been cases where customers had no supply for days because they assumed there was remote monitoring or someone else had reported it so they did not need to bother. Bystander syndrome.

  15. I didn’t read all the comments but I’m a substation technician. There are levels to this question. The further you get away from the source the less the indications of loss of power. Most substations will have monitoring with alarms and breaker position. When I started at our utility they just installed automatic meter reading equipment. This equipment is mainly used for billing but the system also can let the central operation center know of oranges to residences. Other than the house meter all the equipment is in the substation. So, yes the utility can know if you are out of power. Ours can also give power usage in 15 minute intervals to customer.

  16. During the war games in ’90s in the former Yugoslavia, we experienced frequent blackouts due to restricted power supply, typically 8 hours a day. But from one day in 1996 on, the part of the block where I lived (including my building) suddenly stopped being switched off. Many years after, we’ve heard about the reason for that blessing.

    Service workers came to do a regular maintenance of the substation in our block, asked the main control room to switch the power off remotely, received the confirmation and one worker almost got killed – luckily, he decided to test the HV line which was supposed to be off and clearly indicated so. After five minutes, the brief investigation showed the reason why.

    Someone disconnected the “request” and “acknowledge” lines from the relay block and shorted them together, and thus disabled remote switching off, but also defeated the monitoring system. It’s for sure that he knew well what he was doing, but he probably forgot to reconnect the circuit after the restrictions ended.

    They never found out who did it, but yes – the system is permanently monitored.

    1. Rule number one when working with power systems: “If YOU haven’t verified that the power is off, assume and proceed as if it’s live”.

      Lots of people object with things like “don’t you trust me?”, but my answer is always simple: “Always assume your co-workers are idiots, you’ll be either pleasantly surprised or live longer”

      (I know of at least one company where it was standard procedure to communicate “yes, I really did disconnect all power” by grabbing the bus-bars of the system with bare hands. If you don’t feel comfortable doing that, you shouldn’t be working on it.)

      1. You’re right, but it’s too easy to be be right if all you have to do is just to set up simple rules. For instance, we all know the traffic rules, but people are still dying on streets, because things are no so b&w.

        (alert – dark humor follows)
        A long ago, I knew the carpenter who used to say about his competitor: “What kind of carpenter is he when he has all the fingers on his hands” :)

  17. Has the US power distribution system ever dealt with the “20 hillbillies with bolt-action rifles” vulnerability we have? Like one could just punch a small-caliber hole in the bottom of the oil-cooled and insulated transformers at substations and knock out the grid fairly quickly and easily, there’s basically no security.

    Kind of hope they don’t close that loophole, personally. There might come a time that we have to shut off silicon valley for a little while if they get a little too grabby… I wouldn’t be comfortable in a world where that wasn’t possible.

  18. These days, those monitors (Fault indicators), are supplemented by cable hung monitors (Smart fault indicators), that use a Rogowski coil to both measure the current on a line (Short, open, overload, and overall load monitoring), and either connect to a utility deployed network (such as Tropos, or other radio implementations), or to a cellular APN. This way, major load changes can be reported, no matter what direction it’s in, and average usage data gathered, so that you can try to figure out which sections of a circuit should be upgraded or re-balanced on the phases.

  19. Few years ago I had an internship at a communal power supplier in a very small town in Germany.
    They had a very well planned power lines, automatic circuit breakers for short circuits and earth faults and additional current meters at every substation.
    Short circuits and earth faults are reported to the office and every fault is being printed on a endless paper printer.
    There happened to be an underground cable fault in the middle of the night. In 25 minutes they managed it to wake up, go to the office, find the fault, plan how to switch around it, go to the substations, do the switching and reset the circuit breakers.
    The circuit was formed like a tree with rings as leaves that are cut open to one side. By closing the “gap” in the ring and opening the previous connection they could power the ring the other way around while isolating the faulty supply cable, so everyone had power again before even repairing anything.
    For finding the cable fault we had a device which would bump shocks of electricity into the cable and a magnetic ground sensor for walking the area where the cable runs.
    It turned out that we didn’t even need the ground sensor as you could hear and feel the periodic shorts at the surface above the fault.

  20. In my area we have balloons attached to each meter. When the power goes out the balloon is released and the balloon spotter calls it in. Think this is a fairly standard approach…

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.