Of course
That is why there are no people there

Were not gonna make it are we?

In the United States, there are several different power reactor designs.  Since I"m most familiar with the one I'm licensed to operate, I can try and explain a few things if anyone has specific questions.

To the original question, "How much human interaction is needed to keep them running and providing power?"  Quite a bit actually.  At my facility, the integrated control system maintains reactor power at 100%.  We also utilize an "unrodded core" meaning that nearly all of the control rods are completely withdrawn.  The control rods are what is used to shutdown the reactor.  The other control mechanism is boron.  The water circulating through the reactor core of a commercial power reactor in the US is not pure water (like it is on naval vessels). It is really a boric acid solution.  Boron is an element that absorbs neutrons and can be used to inhibit the nuclear chain reaction.  At the beginning of a fuel cycle (typically 2 years), there is a high boron concentration in the reactor.  The concentration varies by core design, but let's say we start out with a 1500 ppmB in the reactor coolant.  As the core ages and the fuel depletes, the control rods will move slightly to control small changes in reactivity and maintain reactor power, but since we run "unrodded", we dilute the RCS boron concentration.  We usually dilute about twice a day and this is a completely manual operation.  At the end of core life, there is virtually no boron (<10ppmB) in the reactor coolant.  This is when the reactor is shutdown for refueling.  If no dilution took place, the remaining few control rods still partially inserted into the reactor would withdraw completely, and the reactor would shut down (more complicated but I've typed a lot already).

There are several transformers that we use at my facility.  The main transformer sends power to the grid.  The aux transformer feeds the plant.  Both of these are fed from the main electrical generator inside the plant.  There is bus work that leads to them both.  We also receive power from the grid when we're not operating and require such power to start up, run pumps, etc.  The power demands at the plant are not insignificant.  Each reactor coolant pump requires about 5MW of power to operate, and each unit has 4 of them.  And there are a lot of other pumps, control systems, etc that need power.  

Hope this helps answer some questions.

Wasn't there a case when a US nuke sub had to be used to black start a power plant on an island?

Side note: never knew we have so many utility/power plant guys on here. Neat thread.

Yeah, I sell an balance of plant system and most of them are still using the ones we put on the market 35 years ago because it is such a pain in the ass to have anything upgraded.  You would think they would want to be on the cutting edge and i can tell you they aren't.

The operation is different from a gas fired turbine, because the "combustion box" is a lot more complicated and dangerous.

For the steam turbine on an NPP, the combustion box is essentially the reactor core.  As the turbine extracts more power, it cools the reactor, which in turn increases power; conversely as the turbine extracts less power, it undercools the reactor, which reduces power output - which is good, but it also puts the fuel rods closer to departure from nucleate boiling [DNB] which can rupture fuel rods.

The increase and decrease in reactor power with decreased / increased energy extraction has to do with the density of steam / water.  Water is a neutron moderator in a reactor - that is, water slows the speed of neutrons, allowing them to better interact with the fissile and fissionable materials.  Heat up water, and the density decreases - decrease the density of the moderator and you decrease moderation and slow the nuclear reaction.  Cool down the water, and density increases, increasing the moderation and increasing the nuclear reaction.

Rapidly changing external power loads change the turbine operating conditions, and ultimately jack around with neutron moderation in the core - so we conservatively scram the plant if these fast transients manifest.  Better safe than sorry (see Chernobyl and their attempts to skirt a turbine trip).  If the turbine trips for whatever reason, the steam isolation valves open and the plant scrams - it's better to fight the transient you know than the one you don't - you may not know exactly why the turbine tripped, and the wrong decisions made on the fly can cause the loss of a reactor core - so isolate the turbine and get it out of the loop so unexpected things don't muck with your shutdown transient.  

But, power conditions on the grid are always in a state of flux, right?  Yes, but the above scenario is for rapid changes.  We have plants in the US that were specifically designed for load following with once-through-steam-generators (Babcock & Wilcox), but even these plants will trip on load induced turbine transients that exceed set limits.  Westinghouse designs use an accumulator-style steam generator that isn't as responsive, but these are the same style of steam generators found all throughout France, and the French specialize in load following - less ideal SG design be damned.

I worked at a geothermal plant that the controls valves were fast enough and the turbines small enough that when the main breaker opened they would catch the turbine go into island mode and keep the plant running. All you had to do was synch across the main breaker to the grid and load it up. Was the greatest thing ever never having to worry about weather or the grid dumping ya.

You in the industry?

This is truth.

To clarify...

In the event of a DBA such as an LOSP, the UFSAR requires an automatic SCRAM. RHR will be activated unless reactor pressure raises above the set points for RCIC and HPCI. CS will keep the top of the core covered and LPCI will activate as pressure comes back down.

The diesel generators are thirsty. Each site has a plan for diesel delivery in the event of a longer event. However in a TEOTWAWKI scenario, the locals are screwed. Of course, most (99%) people are screwed regardless in a TEOTWAWKI scenario.

The safety systems are designed with the health and safety of the public first and foremost. The NRC doesn't care if we generate a single MW. Safety first.

It's safer to stop the reaction and begin cooling the core when a loss of offsite power occurs.

This. The huge magnetic fields from the electric load when connected to the grid "resist" the turbines rotation. Upon disconnect from the grid, you can't dump steam fast enough.

The Russians tried to test one of their reactors ability to use the energy from the spinning turbine after trip to carry the load until the diesel generators kicked in. We know that as Chernobyl.

And I know when loss of offsite power/blackout occurs nuclear sites get priority cranking paths/restart power first and above all. It is first priority.

So, for those playing the home game, an overspeed turbine is bad, very bad?
And I also believe nuclear reactors in general don't play well trying to generate a few % of their operating capacity (which is why I assume you go for straight scram and not "reduce to 2% to supply just enough energy to run the plant isolated from the main grid?)

Worst case scenario for a turbine that spins to quickly is a sudden, rapid separation from itself.

I was never ops and it's been a few years since reactor theory...but the reactor is unstable at low power...maybe something to do with the void coefficient? Or I may have pulled that out of my ass. It would be dependent on fuel makeup, core design, control rod design/life, among other things.

My little brother is a nuke on a fast attack sub.
My little sister is a nuke on an aircraft carrier.
They can't say much, but from the sounds of it the damn things run a thin line between running right and WTFBBQ!

If people go away and the infrastructure goes down ..... you're going to have a smoking radioactive mess in short order.

Naval reactors are a completely different beast. The way a navy nuke explained it to me is that commercial nukes have a bias to shutdown for safety...naval nukes have a bias to run WFO to get into the fight/the hell out of dodge.

And that's about all I got out of them. Didn't realize how classified that stuff is.

Civilian nukes run something like 25% enriched uranium and The navy runs something like 93% enriched.  
I heard something about the army being limited to 25% or whatever because they had an incident with a nuclear reactor and the DOE slapped limitations on them.

Nuclear generator?  

Commercial nukes run between 3-5% enriched.

SL-1 is a notable Army accident back in the early days.

The reactor is too big, they need a grid to put all that power out on. It's not safe (or even possible) to run the plant's reactor with the only load being the plant itself, you wouldn't be able to take all the heat generated by the core and it would boil. This is a very crude explanation.

Now, if you had the grid to put power out to then the plant does pull power from itself to supply in-house loads, it's just a transformer that taps off the main generator output. Once you lose the grid though, the load on the generator/turbine disappears and you need to shut the core down to stop producing so much steam and cool it off. Once the generator trips offline, no more power, then you're relying on diesel generators if you have no grid, this is loss of offsite power. No bueno. If you lose your diesels (loss of all AC), well, you can rely on a steam powered turbine to provide cooling if you can, or the core will just naturally recirc a heavily borated water (hopefully). At least until the cavalry arrives with more diesels and pumps.

Plants will have a rated minimum power for steady state operation.  A few years ago when tornados knocked out sizeable portions of the grid in Arkansas, one of the plants was operating at 40%, and that was their minimum licensed steady state power level IIRC.

You could bootstrap everything, and the plant could idle in a low power producing state, but we want to be conservative in design philosophy, so if your goal is to cool the core the and not really produce power, shut down the reactor, and let your aux systems handle the decay heat.  Running at low power for extended periods of time changes the fuel burn distribution in the assemblies and will impact your ability to meet overall power production targets for that fuel cycle when you return to full power.

The Soviets tried to cowboy around with turbine trips, and it didn't work out so well.  In theory, their plan would have worked if other things hadn't also gone awry, but too much is at risk to deliberately enter a transient where you are only one unexpected event away from entering a severe accident scenario.

Commercial plants use <5% U235 enrichment.  Something like 3.7% enrichment is probably the fleet average.

Harbor Freight.

Nothing.Could.Go.Wrong.

Many modern plants have automated systems for shutdown.  Now as to whether they work the way they're suppose to is another question.

(main question)

I read that a nuclear plant could operate on it's own for several days to several weeks without people.  Unlike coal fired plants that require consistent fuel inputs, nuclear plants typically have enough fuel for a year and half to two years to operate.  

There are automated systems all over a nuclear plant. People aren't having to put their hands on, and manually manipulate all the switches and valves.  But the entire plant itself isn't automated.  Meaning people still have to run those various automated systems, and of course deal with breakdowns.  As you can imagine a nuclear plant is a fairly complex set of systems.  

You know what the difference is between theory and reality?

In theory, there isn't a difference.
In reality, there is.

Most plants have the ability to be fully automated. You could set it up to run unmanned but how long it would run by itself is anyone's guess due to an instrument failure or running out of a consumable that has to be replaced by personel.  All plants I have been too have entire departments set up for instrument maintenance and all they do is verify readings from sensors to mechanical gauges.  They routinely have to re calibrate or replace sensors, IO cards, and switches which there are thousands of. So basically if one sensor would go tits up or out of calibration the plant could shut down in an unmanned facility.  

The water treament in most plants is something that couldn't be unmanned.  Most have a few onsite chemical engineers that treat the water as it goes in and again as it comes out of the plant.

There are two versions.  One is design. 2nd is Reality: human error, and occasionally stupid foulups.

Example:  Pilgrim 1, Massachusetts, -13 level  (13 feet below adjoining just beyond parking lot sea level)   I worked there on a clean-up (burn out crew) daily manual dewatering radioactive overflow caused by pump control malfunction.  Plant had argument going as to who was going to PAY for the repair, whether it was a rad waste responsibility, or a plant engineering responsibility.

I left after two weeks, met my limit of dosage.  Never worked at another nuclear plant of any kind, though a Key Man for Siemens-Westinghouse millwright Crew, in conventional turbines.  Last job was Milennium in Mass, start-up.

Next generation nuclear plant designs look interesting: getting past the "NO NUKES" crowd will be difficult.

That's not even close to being accurate.

That is awesome!  Funny thing is most control rooms look just like that (in coal plants) because many of them are 30-60 years old.  Most of the instrumentation has been abandoned in place and they've added walls of LCD monitors around the top of the console if they haven't replaced the console with a new horseshoe shaped desk.  One small coal plant added a handful of gas turbines onsite and then decommissioned the coal plant but left the control room in the original plant. It's awesomely spooky walking around this giant rusting hulk of a building that is dead silent.  Like being in a horror movie.

I think that was more of a human factors/group dynamic foul up more than anything.

As in the procedure(s) called to do X, but the human operators overrode that and did Y instead.

That's probably the door for their stash.

This is a Trico bulb oiler on a centrifugal pump:



If a nuke plant uses such a pump with an oiler like that, then most likely whatever rounds the plant operators are supposed to do daily or at least once per shift include  checking on the level of oil in those bulbs.

If the oil runs out, then the pump would probably just destroy itself unless there was some instrumentation or controls in place that would kick it out for over-temp or vibration.

There probably is some redundancy in place at a nuke plant should one pump shoot craps, then the pump right next to it kicks on automatically.

As mentioned above by a way more experienced nuke plant person, boron rods are used to control the rate at which the fuel rods "burn".

Should something catastrophic happen to the cooling water inside the reactor's "can", I do NOT know if these boron rods drop automatically on their own without human intervention.

Also, I have no idea how long these boron rods are supposed to last.

Not a door.
It's either a sacrificial zinc anode, or a recent hull patch that was cut for access (and that the yard did after they painted the rest of the underwater hull).

What part?

The control rods on commercial nuclear power plants are typically an alloy of silver, indium, and cadmium., and they last a LONG time.  There are scenarios that will cause them to automatically insert into the reactor core and trip the reactor.  The boron is typically dissolved in the reactor coolant system water which is recirculated through the core.

We do have bubblers like the one pictured on some pumps, and there is redundancy.  Depending on the function of the pump though, it may not be "right next to" a pump with a similar design requirement.  They are often separated by space or barriers for train separation.  This prevents for example a fire in one room from taking out all the equipment designed for a specific function.

Friends father growing up was a navy nuke, started as a bubble head. Got sent to a closet or two by a notoriously demanding admiral.

 
After some kidney stone issues he landed on the enterprise tending that big batch of little tea kettles.


I always enjoyed talking nuke engineering with him, he never came out at said he couldn't talk about cretain things but conversations would sometimes get cut off. Nuke navy guys don't talk about navy PWRs.  He was local to TMI when that went sideways, might have been an engineer for Westinghouse by then, he was called in to consult during the event. Very smart man and I loved learning all I could from him.


I thought the major contributing factor to chernobyl besides the intentional trip and coasting turbine experiment, was the inherent and dramatic instability of the graphite moderated design at low power, and the chief engineers hubris.

Time to upgrade the PLC.