30

I understand that a two-pole AC generator has to rotate at 3600 rpm (60 Hz USA) in order to maintain a 60 Hz frequency and higher frequencies would require these generators to spin even faster which could be problematic.

I also understand that increasing the number of poles would allow the generator to spin slower, but it would require more complexity and cost. However, are there other reasons that power grids tend to use lower frequencies?

ZekeC
  • 496
  • 4
  • 16
  • 10
    Why do they have to rotate at 3600rpm? They could rotate at any rate depending the number of poles in them, as long as they still output 60Hz. But why 60Hz, that is a whole lot of history how things came to be as they are now. There's even a separate page on Wikipedia about it. – Justme Dec 15 '20 at 22:28
  • 7
    Cuz -- [history](https://en.wikipedia.org/wiki/Utility_frequency#History). Basically, 60Hz was insanely high back in the day, which is why Europe went with 50Hz, and why there's older systems that are even lower frequency. Aircraft AC power, established ca. WW II, is 400Hz because the distances are shorter, and transformers (and smoothing caps in DC power supplies) are lighter. – TimWescott Dec 15 '20 at 22:36
  • i toured a hydro electric dam many years ago and i remember being surprised that the vertical shafts from the turbines turned rather slowly ... my memory may be cloudy, but i remember them turning at only a few of revolutions per second – jsotola Dec 15 '20 at 22:38
  • 6
    there's a whole article about it: https://en.wikipedia.org/wiki/Utility_frequency tlrd: old motors like 60-ish, and the first customers probably were industrial clients who wanted to run motors. – dandavis Dec 15 '20 at 23:34
  • 1
    @jsotola Many modem small scale systems generate in DC and then use a DC/AC converter. This allows the turbines to match rotation speed to water flow rate and removes the need for turbines to be able to sunc to the grid. – Ian Ringrose Dec 16 '20 at 10:21
  • You need to consider the mechanical speed limitations on rotating machinery. The radial (centrifugal/centripetal) stress scales roughly as $r\omega^2$ and even with (high cost) modern materials and manufacturing technology the maximum speed of a large jet engine fan is of the order of 3000 to 3500 RPM. – alephzero Dec 16 '20 at 10:30
  • 1
    Skin effect might be a factor for really thick cables since it makes higher frequencies less efficient. Though I guess you'd use high voltage DC for such large cables anyway so I don't know how much this matters – Navin Dec 16 '20 at 11:21
  • 8
    @TimWescott Europe actually initially went with 40 Hz, but people complained of perceptible light flickering, so they upped to 50 Hz which seemed to solve the problem. – gparyani Dec 16 '20 at 12:00
  • You might check out this article. The origins of 60-Hz as a power frequency https://ieeexplore.ieee.org/document/628099 – David Elm Dec 16 '20 at 17:53
  • 2
    It also makes a nice symmetry: hour = 60 minutes, minute = 60 seconds, sec = 60 turns. – TextGeek Dec 17 '20 at 13:46
  • Most motors and generators run at 1800rpm, actually, as four-pole arrangements are most common. – J... Dec 17 '20 at 16:33
  • 2
    I will note that, in conventional power transformers with steel cores, losses (and heating) increase with frequency. The "sweet spot" is probably in the neighborhood of 50-150 Hz, depending somewhat on the core design. – Hot Licks Dec 17 '20 at 20:51
  • @Justme I just assumed a two-pole generator since that's the least amount I believe you can have for three-phase current/voltage. So couldn't it be said that high frequencies would require more complex/expensive generators due to requiring more poles to achieve lower speeds? – ZekeC Dec 18 '20 at 14:52

7 Answers7

33

Higher frequencies are much more affected by the inductance of the power lines. 400 Hz is fine on an aircraft, but over long distances the power factor would be extremely poor. 60 Hz was an educated guess (as I understand), but it has turned out to be about right.

Peter Mortensen
  • 1,676
  • 3
  • 17
  • 23
Frog
  • 6,686
  • 1
  • 8
  • 12
  • That's a great point! If the frequency was too high there would be significant reactive losses for lines since they are inductive by nature: XL = 2pi*f*L – ZekeC Dec 16 '20 at 00:47
  • 6
    Skin effect also becomes an issue at higher frequencies and power levels. – SomeoneSomewhereSupportsMonica Dec 16 '20 at 12:59
  • 1
    @SomeoneSomewhereSupportsMonica: But only in the MHz (and above) range(?) – Peter Mortensen Dec 16 '20 at 21:53
  • 11
    @PeterMortensen Not really. Even at 50-60 Hz, the skin depth of copper or aluminum is about a centimeter, which already has a very noticeable effect for high-current transmission lines (to the point that it allows using aluminium-clad steel cables because the poor conductivity of the steel doesn't matter, everything flows in the aluminium skin). It just gets worse with increasing frequency: [Skin depth vs. frequency](https://commons.wikimedia.org/wiki/File:Skin_depth_by_Zureks.png) – TooTea Dec 17 '20 at 13:35
31

What @Frog says about losses is true, however, that's not the real reason for utility frequency to be around 50-60Hz. HVDC systems have essentially no reactive losses, yet they did not really become widespread.

The choice of utility frequency is largely historical, and frequencies outside the 25-100Hz range were simply prohibitive around year 1900 from the technology point of view: 25 Hz and lower were too low for most consumer applications and required bulky generators and transformers, and 100Hz and higher frequencies could only be generated with belt-driven generators which were already being replaced by direct-coupled alternators due to higher reliability of the latter.

Dmitry Grigoryev
  • 25,576
  • 5
  • 45
  • 106
  • 23
    Below 40-50Hz also causes flicker issues with lighting. – SomeoneSomewhereSupportsMonica Dec 16 '20 at 13:00
  • 6
    @SomeoneSomewhereSupportsMonica Exactly!. That's one of the reasons why it was "too low for consumer applications". – Dmitry Grigoryev Dec 16 '20 at 13:52
  • 3
    While it is true that using DC instead of AC eliminates reactive losses across lines it makes it more complex to step voltage up or down since DC can't directly interface with a transformer. – ZekeC Dec 16 '20 at 18:32
  • 3
    @ZekeC I'd say converting any significant power from DC to AC was simply next to impossible with 1900 tech, when grid standards were formed. – Dmitry Grigoryev Dec 16 '20 at 18:36
  • 3
    @DmitryGrigoryev Electric railways did it, though. They used rotary converters. Which is not a brute force M-G set; most current passes straight through, but the windings are still doing enough that a rotary converter *actually corrects power factor* on the AC grid in the vicinity of the converter. – Harper - Reinstate Monica Dec 17 '20 at 01:37
  • 3
    @ZekeC there are many HVDC lines linking European countries today, it certainly has caught on for under sea links where losses would otherwise be prohibitive - and rectifiers are certainly possible using 19th century technology. That said, stepping HVDC up or down is still a pain – camelccc Dec 17 '20 at 20:15
15

The frequency of the power grid is a great compromise.

Make the frequency higher and you get smaller (read: cheaper) transformers and (somewhat) smaller generators and motors. But, you get higher hysteresis loses in the transformers' cores and higher radiative loses in long power lines as well.

The above consideration about 100 years ago ended up with the conclusion that frequencies 20-100Hz are OK-ish. US engineers started using 60Hz as they liked the number for being the same for other time divisions (seconds vs minutes vs hours). European engineers (1-2 years late to the party) liked 50 Hz better for being a multiple of 5 and 2 only, just like other unit divisions they used.

Other, independent power grids use other frequencies (like Swiss railways at 16Hz) because they fit their purposes better.

Skin effect is not really a consideration as most power conductors are made out of a number of smaller wires for mechanical reasons anyway.

fraxinus
  • 8,726
  • 11
  • 34
  • 5
    You will have skin effect in a stranded wire too, unless you isolate individual strands from one another. – Dmitry Grigoryev Dec 16 '20 at 13:59
  • 3
    The skin effect is way less. Isolation between individual strands helps with the proximity effect. – fraxinus Dec 16 '20 at 15:06
  • 1
    But the outcome is still the same: only the outer strands are effectively conducting. – Dmitry Grigoryev Dec 17 '20 at 08:17
  • 7
    Maybe. My personal experience in HF is that a stranded wire is worse than Litz wire, but way better than a single circular cross-section. I'll check the theory later. – fraxinus Dec 17 '20 at 08:22
  • 4
    @fraxinus stranded wire contains some dead space, so it has a bigger diameter for the same gauge (cross sectional area). AFAIK that's the *only* advantage it gets in terms of skin effect, and it's not a huge one (should be in the neighborhood of 10%). – hobbs Dec 18 '20 at 06:50
7

I believe the reason for low frequency is as your frequency gets higher the impedance, capacitive and inductive, of the network has an effect on efficiency. The reason for AC is mainly so voltage levels can easily be changed using transformers, so as to carry higher voltages over long distances to minimize copper losses in transmission lines.

If you research under water power transmission systems, you will find, in a majority of cases, they use DC because underwater cabling has a high capacitor type characteristic due to the ocean surrounding it. AC power would not be efficient.

This is a good explanation of power distribution

Submarine Power Transmission

Tom George
  • 188
  • 4
  • 2
    [HVDC](https://en.wikipedia.org/wiki/High-voltage_direct_current) is used not only underwater but also in other circumstances, such as connecting two grids which for whatever reason are not/should not be in sync. – jaskij Dec 16 '20 at 11:43
4

Skin effect gets worse as frequency goes up. Large diameter conductors would not conduct very much in the center. Hollow power wires would be difficult to imagine. Very long lines would be an appreciable portion of a wavelength. This means radiation resistance and radiation losses.

Dan D.
  • 672
  • 7
  • 11
Autistic
  • 14,235
  • 2
  • 27
  • 65
  • 5
    Hollow power wires might be difficult to imagine if you're talking about overhead lines, but note that high-current internal connections in power plants and the like are almost exclusively done using pipes (often with coolant flowing inside). And for overhead lines, cables using poorly conducting yet mechanically advantageous cores are very much a thing: https://en.wikipedia.org/wiki/Aluminium-conductor_steel-reinforced_cable – TooTea Dec 17 '20 at 13:41
  • I've also seen copper-clad steel power lines, although those are primarily used for mechanical strength, not to take advantage of skin effect. – Dmitry Grigoryev Dec 18 '20 at 09:10
  • 3
    @DmitryGrigoryev: The choice of how to physically arrange the steel and copper almost certainly has a lot to do with the skin effect. If current tended to flow more in the center of a conductor, cables would be built with copper in the middle and steel outside, rather than vice versa. – supercat Dec 18 '20 at 17:02
3

Anything over 100Hz was also getting into the audible range. Noise from electricity was far too scary for most folk.

CiaranD
  • 31
  • 1
  • According to http://www.cochlea.org/en/hear/human-auditory-range as low as 20Hz is usually detectable by human ears – AwesomeCronk Dec 18 '20 at 23:09
-1

The frequency output has physical constraints on the generator as well to what Tom has mentioned. A higher frequency = more poles on the 3phase generator and it has smaller copper coils in the winding pair.You want to achieve an optimal current output for the generator for it size.

binaryOps20
  • 155
  • 1
  • 4
  • 3
    The current is also then divided over more poles, so the thinner copper is not such big an issue. Added insulation distance for each coil will take up space though. – jpa Dec 16 '20 at 09:39