Why do we need to shut down power generators if we are producing more power than the grid requires?

How it will damage the system if they are not shut down?


The electricity grid has very very little capacitance. That means that in any and every second, the energy going into it, has to come out of it too.

If it doesn't come out in controlled, intended ways, it will come out in uncontrolled, unintended ways.

For example, by blowing up equipment.

One of the routes for excess energy is to make synchronous spinning things on the grid, spin faster (i.e. the grid frequency increases) - some motors, for example. Some of these will quickly exceed their intended speed.

Another thing that happens with rising supply, is that grid voltage increases. Some equipment will respond by using more energy, which in some cases can lead to faster burnout. Electrical equipment will tolerate some voltage variation, but after a point, increasing voltage will damage it.

So, the excess electricity will get absorbed by some things getting hotter, and by some things going bang. Both on the demand-side, and on the supply-side.

In order to give equipment manufacturers some certainty, limits are set on how far the grid frequency and voltage can vary away from their nominal set points. The system operator will adjust supply and demand to keep frequency and voltage within their limits. One way to stop frequency and voltage getting too high, is by reducing or stopping the output of some generators. Another way is to bring more demand online. All these rules are contained in the Grid Code: each system has its own.

  • 3
    $\begingroup$ All good except the first sentence. Capacitance has nothing to do with this. Note that any capacitance on the AC line is charged and discharged twice per line cycle anyway. You may have meant to say "energy storage", but capacitance doesn't store energy for more than a fraction of a cycle in a AC system. $\endgroup$ – Olin Lathrop Mar 28 '16 at 11:15
  • $\begingroup$ @OlinLathrop NB the supply side is a combined AC / DC system - e.g. PV & wind generators, and HVDC transmission lines. (by the way, electricity systems typically have vast amounts of energy storage - perhaps you meant electricity storage?) $\endgroup$ – 410 gone Mar 29 '16 at 6:54
  • $\begingroup$ @EnergyNumbers, I agree with Olin that your use of capacitance is incorrect. An uninformed reader would think that they could solve this problem by just adding more capacitance; when in reality AC systems only use capacitors to correct power factor. The DC systems you mentioned do not use capacitors except for PWM purposes. There is no such thing as "electricity storage"; we should talk in terms of real types of energy. The very small but main amount energy stored in the grid is kinetic energy in grid tied AC motors. $\endgroup$ – ericnutsch Mar 30 '16 at 0:09
  • $\begingroup$ ...Electric potential energy stored in capacitors and magnetic potential energy stored in transformers lasts for only part of a cycle; so it is not helpful in answering the question. $\endgroup$ – ericnutsch Mar 30 '16 at 0:09
  • $\begingroup$ @ericnutsch there are no pure AC grids that I know of. Maybe in the olden days, but not any more - they're all combined AC + DC. Are you thinking of some micro grid on a small island somewhere? $\endgroup$ – 410 gone Mar 30 '16 at 17:51

There will also be environmental and economic reasons. Since electricity can't be stored in this context, why make more when it's not needed? You can decrease your emissions for the year, including carbon payments. Emission legislation is so onerous these days (in the UK) that we're actually closing some power stations to save carbon. You can also save fuel. And if you never had excess capacity, you could never service or replace the generators. Since it takes 12 hours to spin up a fossil fuel power station they can't be blipped on /off anyway.

The truth of my answer could be proved at finding out whether solar /wind /hydro power plants shut down at surplus times since they run for (effectively) free.


First, conservation of energy says that power input equals power output. For all intensive purposes, the grid does not store any power. Even if there were battery systems tied to the grid, they would be treated just like power generators and not really be part of the grid because of the DC to AC conversion process.

Utilities in first world countries hold tight tolerances on AC grid power. By design, the voltage and frequency change very little, regardless of the amount of power being transferred.

In the United States a frequency 60Hz is held perfectly to the tolerance of the atomic clock in Colorado. Many low cost alarm clocks rely on this for time keeping. This very constant frequency allows AC motors to have a constant RPM, transformers to have predictable performance, and permits for much easier management of grid power over long distances.

To maintain this frequency, the utilities closely monitor and control the rpm of their turbines with with computer control algorithms. For example, if it cools off mid-day and lots of thermostatically controlled air conditioners turn off; the load on the system decreases. The computer algorithm will then quickly see that it needs to decrease steam input to the turbine to reduce power output in order to maintain 60Hz. It works just like "cruise control" in a car; it presses down on the accelerator as needed to maintain your set speed.

Voltage is also kept very constant; with consumer voltage ranging between 115VAC and 120VAC. In my experience the voltage will be different for different locations, but less than 0.1VAC variance at a location when not in a transient state. Regulation of the voltage is a little more complicated, but also computer controlled.

So from a power production stand point, no turbine ever needs to be turned "off", but turbines on the grid do need to be throttled up and down as the demand load changes. The logistics of this may result in some turbines being shut down because it is often more economic to operate a single turbine than it is to operate two turbines at lower output. For example one running at 60% capacity instead of two running at 30% capacity. Turbine operators and computer algorithms smoothly handle these transitions so an end power user never notices a difference. There is always excess capacity ready to handle changes in load.

In the photo below (from wiki) you can see many types of power production interconnected. Nuclear shields its fuel rods to regulate steam production, natural gas and coal reduce fuel input to regulate steam production, and hydro uses low loss valves to regulate water flow. Solar, with the proper electronics could regulate its output, but generally does not because it coincides with peak air conditioning loads and is always utilized 100%. Modern wind turbines with adjustable pitch blades also throttle their production.

Electricity Grid Schematic from wiki electrical grid article

If a controller went bad on a turbine attached to the main electrical grid, it would begin to lead (produce too much power) or lag(pull power from the grid). Once the amount of power it pulled or pushed was too much, it would trip its breakers because of over current and be safely disconnected from the grid.

It is possible for an isolated single-turbine grid (like a back-country hydro turbine) with a control system failure to operate at too high of frequency or too high of voltage. This will destroy sensitive electrical equipment like DC power supplies and over speed AC motors. The main electrical grid however, has many systems in place to prevent that from happening.

  • 1
    $\begingroup$ as far as I can tell, the three US grids do not keep their frequency held perfectly (and they are not synchronised with each other, either): they routinely vary within a defined range that's a lot broader than the tolerance of an atomic clock. See for example fnetpublic.utk.edu/gradientmap.html $\endgroup$ – 410 gone Mar 30 '16 at 6:22
  • $\begingroup$ Yeah, this looks like a great answer except that I highly doubt the assertion of how precisely the frequency of U.S. utility power is regulated. Could you correct or provide an authoritative citation? $\endgroup$ – feetwet Aug 1 '16 at 17:28
  • $\begingroup$ Correct, the resultant frequency measured at any one location will likely not instantaneously be at atomic clock tolerance. The measuring device wouldnt have that resolution anyways. The controlling system however does use the atomic clock as the time control reference, so within a period of time (1 minute, 1 hour, 1 day) the average frequency will be of this tolerance. This is why it is possible to run a low cost grid tied clock for a year and hold atomic clock tolerance. en.wikipedia.org/wiki/… $\endgroup$ – ericnutsch Aug 1 '16 at 18:17

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.