The situation is as follows: there is a city with different districts, each provided with electric power from a different power plant:

  • Thermal power plant
  • nuclear power plant
  • Hydroelectric power plant
  • Wind power plant
  • Solar power plant

All the people have disappeared at the same time, in the middle of the day, so now nobody maintains the power plants anymore.

  1. Will power plants continue to produce and deliver electricity to the houses?
  2. How long it will be possible to use the electrical energy to the abandoned houses of this city?
  3. How soon will each type of power station break down?
  4. Does disaster happen (like the explosion of a nuclear reactor)?
  5. Which stations will last longer?

5 Answers 5


I'm assuming there would still be some electrical load because appliances and machines would be left running, but that load would drop over time. It would drop sharply at first as heavy load machines stopped running, then the next major drops would probably be caused by grid failures over weeks or months.

Will power plants continue to produce and deliver electricity in the houses?

How long the plants keep running depends heavily on how much management they need, of course. It's possible to automate a lot of stuff even without looking at things like AI management.

A coal plant would run out of fuel fast (as mentioned in another answer) but a nuclear power plant typically refuels every 18-24 months and hydro/wind/solar stuff doesn't have the fuel problem exactly.

However, I think the thing that would kill the grid first is automatic safeguards that trip when the power generation and load get uneven. You can't just generate all the power you want regardless of load, you have to manage generation to match the load. That requires careful, ongoing control in the US power grid. If people stopped doing that, the power grid might start going down in hours. I don't know how automated that is at this point in time, but given what I've read from people who work in the power industry, it seems to require a lot of human effort to keep things running smoothly even day to day.

Regardless of how the grid is managed, it's likely that any active generation power plant like nuclear or coal would simply shut down automatically if people stopped showing up.

How soon will each power station break down?

Coal / Nuclear

Coal plants would run out of fuel or shut down quickly without workers. Nuclear plants should automatically shut down too, if the engineers who designed it were smart. Neither plant would break down per se, but I'd expect both coal and nuclear to stop generating power inside 24 hours or so.


Wind turbines need significant maintenance and management too, as it happens. Electrical components fail, gearboxes need fluid changes and sometimes need to be replaced, and you can't just let a wind turbine do its own thing no matter what the wind is doing. If the wind is too high, it can destroy the tower. Who controls which way the turbine is facing? Is that automated? It could be, and it's an important question.

If the turbines just stopped getting maintenance and didn't get killed by high winds, and were able to stay pointed into the wind, from some maintenance numbers I found online I'd expect the wind farm to be down to <30% capacity in one year and completely dead by ~18 months.


Hydro plants are pretty low maintenance, but again, require a lot of management. This is a simpler problem than the wind farm problem, since it's just deciding where to route water through the dam. But something needs to make those decisions. If it's computer controlled, and the computer also manages the reservoir, the plant could conceivably stay functional for years just producing enough electricity to keep itself powered on. Eventually corrosion would take its toll and turbines would start to fail, but at minimal load that could take a decade or more.

Hydro dams would probably get shut down by clogged intakes before the hardware failed, assuming computer management. So, that could mean months to years depending on what kind of debris problems they have.

If there is no automatic management of the system, though, it's a big fat "it depends". In what state was the hydro plant when everyone vanished? Were the turbines running full bore to match the peak power load for the day? Was it flood season, and they were letting a lot of water through to keep the reservoir level manageable? Neither situation would be good for the plant. If the plant was in the wrong state at the time of the great vanishing, it could be damaged in short order, or drain the reservoir in the dry season. So, days to months and the plant would be down.


Solar panels are pretty durable. I believe a typical solar panel is expected to last ~25 years. They do degrade with time, and naturally the output drops if they are dusty. Snow would stop them working at all, of course.

The weak point in a solar system is the battery bank. Lead acid batteries require maintenance. Connections corrode and electrolyte evaporates off. Left unattended, getting hit with a charge/discharge cycle every day, the battery bank in a house might only last 3 months, and certainly not a lot longer. Some of the batteries would be killed from the lack of maintenance, while others would only require new distilled water to work fairly well again.

There are other battery technologies that work fine with solar and require less maintenance, but they're substantially more expensive than typical lead acid batteries. So the system could be designed to be more robust, but probably wouldn't have been.

The grid

Again, there's still the grid maintenance to worry about too. Transformers fail, switches fail, power lines fall, etc. Once that stuff fails, some of it is effectively irreplaceable. Even fixing a power line is a nontrivial task. I think the power grid would start failing inside 6 months, which would start cutting off parts of the city.

How long it will be possible to use the electrical energy in the abandoned houses of this city?

I take it you want each district to be powered by its own independent grid tied to one of the types of generators?

Assuming automatic control of the power grid so it didn't just shut down, ballpark estimates:

  • Nuclear - <24 hours
  • Coal - <24 hours
  • Wind - Grid will probably fail first from lack of maintenance
  • Hydro - Until the grid failed
  • Solar - Until the grid failed, but it depends, with a bit of elbow grease years are possible

Solar is the best bet for long term power, because even if the houses don't have their own panels you can simply relocate and rewire panels from the solar farm to feed the houses you want to feed. That assumes the person doing it has a clue and some basic tools, including a multimeter, of course.

Does disaster happen (like the explosion of a nuclear reactor)?

Probably not. If the grid is entirely dependent on human control with no automation or safeguards, grid failures would probably start fires.

Which stations will last longer?

See above.

Other comments

Quite a lot of this comes down to "Well, it depends how they designed the systems." In most cases, automation isn't done simply because it's cheaper to pay a load of people $20-30 an hour to watch dials and push buttons (or aim windmills) than it is to design, build, and retrofit an automatic system. That sort of automation typically becomes standard over decades, as new systems are built to replace old, worn out ones. The exception is when human costs rise to the point where it's cheaper to just retrofit the automatic system. That's why fast food restaurants are starting to replace workers with robots.

So if what you want in your world is for power to be available even after the people all vanish, it's certainly possible with current technology if, for example, the society placed a high value on the perceived safety of removing human error from the equation.

There are also other ways to get power. There would probably be emergency generators around, in places like hospitals. Some people would own them for camping or as backups (I have a <30lb 1000W Yamaha generator that's been amazingly useful). You can get power from welder/generator machines. RVs have generators, and so do some refrigerated trailers. Gasoline is usable for a year or more, and diesel can last quite a while longer if properly treated. It's also possible to modify internal combustion engines to run on alcohol. So, there are lots of ways to get power in a city even if the grid is down, if you're not worried about people stopping you looting the place.

  • 2
    $\begingroup$ But how long until the grid fails? Because that seems to be the only thing missing in your answer. Even a rough estimate might be helpful. $\endgroup$ Commented May 15, 2016 at 5:12
  • 4
    $\begingroup$ "Who controls which way the turbine is facing? Is that automated?" Yes, yes it is. Either passively, by using a "tail fin", or by using electronic systems. Humans are too expensive. At least here in Poland, where work is cheap but wind changes often. $\endgroup$
    – Mołot
    Commented May 17, 2016 at 6:52
  • 2
    $\begingroup$ I've seen believable reports of nuclear plants certified to operate unattended for three days. $\endgroup$
    – Joshua
    Commented Aug 19, 2016 at 21:17
  • 2
    $\begingroup$ Nuclear reactors need electricity, even when "switched off", because the cooling circuits need to be active for a lonng time. Without cooling, you get Fukushima. Hydro plants suffer a lot from erosion: particles in the water eating away at the turbine blades. This effect strongly depends on water quality. Apart from that: good answer! $\endgroup$
    – Burki
    Commented Aug 22, 2016 at 10:41
  • 2
    $\begingroup$ @forest actually most will not get hot enough to melt down, we design plants to fail in to safe conditions. without a nuclear moderator the rods sustain a reaction. $\endgroup$
    – John
    Commented Sep 11, 2020 at 21:01

If, by 'thermal power plant', you are referring to a Coal based plant, you can eliminate it as a power source within a day.

A large coal train called a "unit train" may be two kilometers (over a mile) long, containing 130-140 cars with 100 short tons of coal in each one, for a total load of over 15,000 tons. A large plant under full load requires at least one coal delivery this size every day. Plants may get as many as three to five trains a day, especially in "peak season"

According to this entry, it takes steady deliveries of coal, so you will get a day worth of power from this type of power station. No deliveries, no more power.

edit: A note about Loads. A comment above mentioned loads, and it is relevant to this scenario. Different Power plants are designed to respond differently dependent upon the load on the system. Some power plants would automatically shut down if the load on the grid was under a certain point.

From the response on Solar we already see that system shuts down when whatever 'main' source on the grid shuts down. Since Solar and Wind power are both unsteady supplies, I would guess both of these would operate in 'load follower' mode and be shut down nearly immediately when everyone vanished. The Load following article also mentions Hydro working in either mode, so you might be down to hydro and nuclear.(there have been other questions and answers concerning nuclear plants...)

  • $\begingroup$ There may be a week or two of coal on stock in big piles but they will usually require human intervention to direct to the boilers. I would think that half a day from an unattended coal station would be lucky. Perhaps a week or two with knowledgeable people feeding just one boiler with the coal stockpile. $\endgroup$
    – KalleMP
    Commented Apr 24, 2018 at 22:23

Well, the solar power plant on my roof should last for 15 years before the inverters are scheduled for end-of-life. The panels are supposed to resist hail but if one were to be damaged, it's not designed to shunt itself out; someone would have to, at the simplest, unplug the bad one and patch the connectors on either side of it with a length of wire. It's possible to build a system where each panel is isolated rather than in series, but for normal use it's not worth it.

Being a "grid tie" system, it will cut off also if the grid feed is down. However, there is a switch (on each inverter) to take it off line and power a 1500W outlet instead. So if central power was down, the house would not have power but the "Sandy switches" would let me charge my devices and even run the 'fridge or whatnot on an extension cord.

A system that has local storage of power would have more to go wrong. How long would batteries last? Even so, being designed to work without grid connection, it might power the house by flipping a few switches and jumpering some things, even if the batteries are bad or the charger stuff is shot.

I suppose one like mine could power the house (with whatever amount of power it has) if cut off from the grid, if it's opened up and patched around somehow. It might need a 60Hz reference signal to sync with, and fake the indicators.

But the point is, you could walk up to it after a decade and use it. It has no moving parts, the modern inverters don't contain transformers, and the photovoltaic material itself has never been known to wear out. If you don’t specifically need the home standard AC, you could skip the inverters and use low-voltage DC directly from the panels.

If you want something that will work after being abandoned, that's the stuff. In some areas it's very popular.

  • $\begingroup$ Your system wont even power your own house without the grid? $\endgroup$ Commented May 13, 2016 at 15:25
  • 1
    $\begingroup$ Right. If the grid is down, it shuts off. The "Sandy switch" can be thrown to provide 1500W (each) to a dedicated outlet on the inverter. But, when the grid is out it's usually evening/night and during a violent thunderstorm, so no sunlight anyway. $\endgroup$
    – JDługosz
    Commented May 13, 2016 at 15:49
  • 2
    $\begingroup$ @user2448131 If you’re powering your house during a grid outage you’re also powering the bit of the grid near you, and electrocuting the engineer who is working on fixing it. $\endgroup$
    – Mike Scott
    Commented Jun 7, 2018 at 4:53

You can probably go for a few years with a hydroelectric power plant, barring any downed wires or bad transformers.

I've been inside of one of those plants and there is not much to it. The water is piped to the turbine, comes on the other side and is released.

The turbines are generally designed to run for years with little to no maintenance.

  • 1
    $\begingroup$ Heh.... tell that to the bearings when they run out of lube. :D $\endgroup$
    – MichaelK
    Commented Aug 22, 2016 at 6:48
  • $\begingroup$ And be aware that particles in the water erode the turbine blades. $\endgroup$
    – Burki
    Commented Aug 22, 2016 at 10:44

I am not sure how systems are designed these days with nuclear but unless there was an automatic shut down set up without human interaction being required, the nuclear chain reaction could cause a meltdown to occur which if it was located too close to other power generation systems could possibly wipe out them in the process quite quickly....not to mention the effects of the radiation itself could effect the transmission of electric through them through ionization and make the grid useless as a whole anyway...

One could presume that people would be able to locate the switching stations that route the power from the plants and make the adjustments to divert power to key areas from any systems that were still usable presuming these were not totally devastated though one may presume as a need to conserve remaining power systems that it would be wisest to send this power to remote villages as opposed to large cities as refrigerators and such require larger loads and would be instantly active when the grid reconnected in their highest state...so unless they went to each home and flipped breakers rebooting a system after such a problem may be impossible for a system to handle in major cities....millions of fridges needing +20 amps to make that initial turn to get the motors going again would likely fry the systems trying to deliver the power alone even if the existing sources could theoretically provide the power needed to accomplish it there would also be a massive voltage spike backfed into the grid when this was happened as all these motors surged to life....just think when you have old incandescent bulbs and your fridge kicked on how the power would dim go bright and then return to normal now think about the extra voltage that involves multiplied by the number of houses and the fact this will then surge backwards through the transforming substations which would likely explode and if not would still likely overload any electronics hooked up to control it...

Still, we do not know why all these people are disappearing ie from a meteor impact from a nuclear attack alien invasion or "rapture" and what effect it could have globally to have disrupted the equipment and a sudden jarring could also disrupt the safe guards involved in any of these machinery....not to mention the main lines that control stations are in control of running the power to the substations are much more difficult for a typical citizen to gain access to if they were damaged and run on much higher voltages making them more dangerous to handle.

  • 4
    $\begingroup$ I am sorry but no. Gods I am sick to the teeth(!) with this stupid "Nuclear Is Black Magic That Destroys EVERYTHING"-trope. Where did you come up with that nonsense?! $\endgroup$
    – MichaelK
    Commented Aug 22, 2016 at 6:51
  • $\begingroup$ How about a fail-safe pebble bed reactor? I detailed it in this answer. $\endgroup$
    – JDługosz
    Commented Aug 22, 2016 at 7:28
  • $\begingroup$ I have seen vids about the way present nuclear reactors work and they require manual pumping of water to stop what is called a runaway chain reaction...a man has designed what is called a lifter system which would make the system able to maintain itself and safely shut itself down but I am all too aware how difficult for an inventor to get worldsaving technology implemented even when present science is capable of delivering it as a result of my efforts these last 10 years to build a diamond nanobot array to emulate every electronic or machine ... $\endgroup$
    – Firobug
    Commented Aug 22, 2016 at 17:40
  • $\begingroup$ @jdlugosz your pebble bed reactor is not not the present day technology and likely is not even feasible for the reasons stated above... this asker is speaking of present day science implimented today across our world.... $\endgroup$
    – Firobug
    Commented Aug 22, 2016 at 18:10
  • 2
    $\begingroup$ No. Any vaguely modern (ie post 1970) western reactor is so stuffed full of triple redundant safety systems it would auto scram within seconds of whatever event caused the human population to disappear. Even if, worst case, get a sudden total loss of coolant event in a light water reactor going flat out, the water acts as both coolant and moderator so the reaction stops. Some melting of the fuel assemblies would occur but this would be easily contained by the pressure vessel, or at least the containment building. Anything less than that is a non event. $\endgroup$ Commented Aug 23, 2016 at 22:29

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .