AC versus DC transmission and distribution 1

[Mbrooke]

Starting a new thread not to hijack another one on a different subject. AC vs DC. Which is better? Which do people believe will be used in 50-100 years time?


My thoughts are this: AC is a waste of copper, aluminum, iron, steal, space, resources just to name a few. AC is complex to control, and more unstable than DC in operating large systems. AC is a more difficult to engineer, especially as systems grow. AC power is holding back the burial of transmission and distribution lines. Its holding back renewables. Its holding back energy storage. Its holding back long distance power delivery and exchange. The conversion of AC to DC wastes energy. AC systems can not tolerate large scale none linear loads. AC systems emit fields which can be hazardous to human health. In short AC is impractical and it has always been a bad idea when all is said and done.

 

[Mbrooke]

I think that it will be DC "someday" but not anytime soon. Technology has a ways to go for limits of technology and the costs involved.

Its closer than you think

It is true that AC is more complex. You have to worry about voltage control, stability and so on. But the claim AC is more difficult to engineer as systems grow runs into a massive obstacle: HV DC circuit breaker technology has a ways to go. Right now they're frightfully expensive and limited. When your system is always point to point, not a network, that is rather limiting.

As LEDs were frightfully expensive and limited in the 70s. Today I can buy 4 LED filament lamps for only 5 bucks. Less than a pack of halogen energy saving equivalents.

Not that I see. Certainly not in the United States. 500kV AC can be buried now. Multiple HVDC island power supplies have been decommissioned and replaced with AC. New distribution in the US is almost always underground. The tradeoffs between overhead and underground is comparable for both. As an example, the HVDC systems in the US are all aboveground. They're that way because it's less cost, more reliable, and longer lived when we're talking about transmission voltages.

500kv AC can be buried, but for more than a few miles you need shunt reactors, often variable reactors and complex schemes to control voltage. Also Re-energizing large amounts of underground cable under light load after an outage is a nightmare. Beyond a certain point an underground cable will carry more reactive power than watts work power limiting the distance AC underground cables can go. Those VARS are a perfect example of the waste of material in AC power. Even with 15kv classes you have significant reactive "losses". Its now known that ferroresonance can even occur on a low loss wye gr-wye gr pad mount with enough URD.

Overhead AC lines are also limited. If you built an AC line from Maine to Los Angeles, the power would just radiate off into space.


I know people will bring up that there is no need to bury distribution and transmission- however I think time will force POCOs to entertain that. Consider the severe weather events the US has been plagued with for the last 10 years. Look at Puerto rico. Texas. The storms in Maryland. Just recently there was a Tornado/micro-burst event up in NY state and CT effecting New Fairfield, Southbury, Newtown, Seymour and a few other small towns. The event was only about 10 minutes per area, but it was so destructive people were without power for 5 days, on average some 2 weeks with entire chunks of the distribution system needing to be rebuilt. I am talking about literally hundreds if not a thousand broken poles, hundreds of transformers and miles of reconductoring- enough crews and material for a whole state after a hurricane. Crews from hundreds of miles out of state. In the past New York, New Jersey, New England, ect had Irene, Sandy, a freak October snow storm, a dozen ice storms, smaller tornado events, a lot of serve thunderstorms each of which are considered to be a 100-150 year event. If overhead lines need to be rebuilt every year, eventually going underground will look attractive.

How so?
Just because solar panels are DC doesn't mean you can hang them on a DC power line. They have a 2:1 voltage range and you need maximum power point optimization that varies with insolation level and temperature.

Just because battery technology is DC doesn't mean you can just float it on the system. LiIon batteries voltage over the charge / discharge cycle varies around 40% you can't let it vary that much.

Ultra caps can be connected directly across the line. No need to rectify DC to charge them, then use an inverter to put it back. No need to transform DC when the caps are rated 25kv, 165kv, ect.

If you have DC distribution with either, you have to have a DC-DC converter. If you go with AC, you can have a inverter do the same job (you design the inverter for a large modulation index). Typical large inverters today are 97%-98% efficient. The inverter's advantage is you can hang a transformer on it and have any voltage you want out. With DC, you either are forced to match the battery and line voltage or have a more complex isolated DC-DC converter.

A 50/60Hz trafo is massive, a khz unit is much, much smaller. Plus an inverter to make 3 perfect sine waves is more complex than just something that spits out DC to feed into a T or D line.

I'm not so sure of that. UHV AC was initially developed by the US, Russia and Japan. Other than some 765kV in the US, it hasn't been used much here. Russia has backed down and Japan never implemented anything over 400kV. Since then, India and China have become the leaders in UHV AC. India has achieved a single grid and now operates the largest grid in the world. They still use less power than the US, but the US has 3 major grids, not one. China is headed to a single grid someday. The Eastern and Western interconnections could be synchronized at 765kV by extending AEP's system West just as well as it could be connected more than at present with DC.

All the more reason to have everything DC. The larger you go, the more attractive DC becomes. This becomes substantial when you add renewables which are only reducing the critical clearing time of the bulk power system. Engineers are seeing places where the breaker failure time is actually the CCT. I'm seeing substations now with series AC breakers...

1st: That conversion has to be done for many sources and loads. Anything powered mechanically is an AC generator with conversion. Anything generating mechanical motion is an AC motor with conversion. (Even a "DC" motor is an AC motor with a mechanical inverter.) So you have to add inverters that aren't needed in an AC system.

Correct- But inverters are needed in solar, fuels cells, variable speed generators ect. Solar will only grow and grow, spinning fossil turbines will only dwindle.

2nd: Every change of voltage requires converting DC to AC and back again after a transformation. Non-isolated converters like buck and boost don't send 100% of the power through AC, some goes through at DC; this is analogous to an auto-transformer where some power is transformed and some is coupled galvanically. The advantage is that the frequency can be optimized for the transformer, but when you get to big power, it's only going to be 100's of Hz at the most.

Elaborate. Why can't I use Khz on a 600MVA galvanic isolation trafo?

With an AC system, you have at most one electronic conversion with losses. With DC, you've got many more. Remember, you can't have one voltage. At a minimum, you need 4 different voltages. Bulk transmission (100s of kV to 1MV), local transmission (50-200kV), distribution (10-30kV) and utilization (mostly <1kV). That's a lot more AC-DC conversions than you have in a AC system.

Who says the DC-AC-DC conversion from to from each voltage level will be lossier? Even if so, the smaller size of the DC trafos, reduced transmission line lose, reduced T lines materials, ect will make even an 8% loss attractive.

And? Most SMPS / DC loads are required to be harmonic free. Computers, servers, LED lamps, electric vehicle chargers, solar inverters, battery storage, etc are all low harmonics. Active harmonic filters are going mainstream for those loads that are not.

Again, why more material and equipment? Its cheaper without the filter.

That's not conclusive, at least from what I've ever heard.

Care to be the guinea pig that finds out? I think not. What is esoteric knowledge today will be common knowledge tomorrow. At one point doctors were advocating the health benefits of cigarettes. Who am I in 1955 from 2018 to question a doctor? Or companies making handsome profit where any evidence to the contrary opens major lawsuits.

How is the dominant system that supplies 99% of the world's power "impractical"?

The majority is not always right. Burning fossil fuels gave life to billions of people, gave rise to nations, and creates plastics just to name a few. But coal, oil, ect is not practical because several hundred years from now it will run out. Assuming the environment will tolerate it by then... In sort its not practical to continue doing so.

 

[HamburgerHelper]

My pros and cons:

Pros:

1. Lower losses but losses are only around 5% from generator to customer.

2. No stability, voltage, or var support required.

3. Greater thermal through put on conductors due to no skin effect.

4. Potentially, more convenient for customers with electronic loads and microgrids.

5. Converter stations can operate as a FACTS device and provide voltage support in intermixed systems.

6. There could be stronger ties between the interconnects.


Cons:

1. Conversion losses swamp out any reduction in real losses in the system

2. DC breakers are expensive and don't have a long track record compared to anything AC.

3. It is always better to build your generation near your load. There are dc transmission projects to move wind power east and west from the midwest but that is captive or being used to meet requirements. I don't believe the power people receive is traveling great distances. Why bother with a system that gives you benefits when moving power over great distances if that isn't happening that much to begin with?

4. Load isn't growing that much in the U.S. or quickly in most places to bother with talking about needing a ton of additional capacity. Maybe, when electric cars take off we will need capacity.

5. If you want reliability and longlife, what is more reliable than a transformer with a metal core wrapped with wire sitting in oil. Thyristors, I believe have around a 30 year service life. There was an issue in IEEE PES magazine I think two years ago that was about if it was worthwhile to sift through hockey puck stacks of thrystors or just replace them at older HVDC stations.

6. The system presently works fantastically. How many 9's of availability do we really need?

7. All the utilities in the U.S. are short staffed and have a pervasive attitude of if it ain't broke don't fix it.

8. You will have a very hard time convincing the utility or the PUC for that matter that it is worthwhile to carry out all this additional work. There are dumb projects that get pushed like talk of hardening the system for EMP or solar storms but those are threats that can grab the imagination of whoever is in charge.

9. Electric motors make up 45% of the global energy used. 28% is just for industrial motors. That is a pretty large block that really wants AC.

10. How is it going to make electricity cheaper? My windows on my house are fogged up due to them losing their insulating gas. I can spend $16,000 to have all them replaced with new insulated windows. I will never recover $16,000 in reduction in heating or cooling expenses. Living with what I have is the cheapest option.

-----------------------------------------------

I could see FACTS devices becoming more prevalent if the cost ever came down.

One thing I would be curious to see would be a system with intermixed transmission frequencies. Maybe, something like everything 345 kV and up operating at 20 HZ and having cycloconverters at every step down into 230kv and below. You could get some of the benefits of operating at lower frequencies while keeping some of the benefits of being able to step up and step down with transformers, lower losses over longer distances, keep AC breakers, and not having to overhaul so much of the existing system at the lower voltages. I would be really curious of any information about the variable frequency transformers (VFT) installed in Germany and New Jersey, USA. How closely do the frequencies have to be on either sides?


------------------------------------------------------------------------------------------
If you can't explain it to a six year old, you don't understand it yourself.

 

[ScottyUK]

Elaborate. Why can't I use Khz on a 600MVA galvanic isolation trafo?

Skin and proximity effects. You can mitigate it somewhat using Litz wire, but the economics and practicalities are massively stacked against using it on a huge transformer. In generating plant Roebel bars are used to manage the problem at only 50Hz / 60Hz.

 

[Mbrooke]

@ScottyUK: True, but at what threshold does the material saved vs inefficiency of high frequency meet?

 

[Mbrooke]

Cons:

1. Conversion losses swamp out any reduction in real losses in the system

Do you mean DC to DC voltage transformation or converting an AC system into DC? DC-DC could be made just as efficient as AC-AC.

2. DC breakers are expensive and don't have a long track record compared to anything AC.

As with anything new. GIS was was like that at one point and is not growing in popularity.

3. It is always better to build your generation near your load. There are dc transmission projects to move wind power east and west from the midwest but that is captive or being used to meet requirements. I don't believe the power people receive is traveling great distances. Why bother with a system that gives you benefits when moving power over great distances if that isn't happening that much to begin with?

Ideally, but energy is not always close to home.

4. Load isn't growing that much in the U.S. or quickly in most places to bother with talking about needing a ton of additional capacity. Maybe, when electric cars take off we will need capacity.

Infrastructure is aging, over time everything will be replaced and rebuilt, eventually. Fossil generation is retiring, while solar and wind grow exponentially.

5. If you want reliability and longlife, what is more reliable than a transformer with a metal core wrapped with wire sitting in oil. Thyristors, I believe have around a 30 year service life. There was an issue in IEEE PES magazine I think two years ago that was about if it was worthwhile to sift through hockey puck stacks of thrystors or just replace them at older HVDC stations.

Now, but as semiconductors improve that will change. Think incandescent vs LED.

6. If you switch over, you have to eat the cost of putting in a new system for marginal benefits. The system presently works fantastically. How many 9's of availability do we really need?

Debatable- it works well with top down generation, not so well the other way around.

7. All the utilities in the U.S. are short staffed and have a pervasive attitude of if it ain't broke don't fix it.

I will agree here. But when it does fail, it will have to be changed. And then there is NERC, too.

8. You will have a very hard time convincing the utility or the PUC for that matter that it is worthwhile to carry out all this additional work. There are dumb projects that get pushed like talk of hardening the system for EMP or solar storms but those are threats that can grab the imagination of whoever is in charge.

Time will carry change.

9. Electric motors make up 45% of the global energy used. 28% is just for industrial motors. That is a pretty large block that really wants AC.

Of course they want AC, but why is more and more of it coming from VFD? A motor being run high frequency, high slip alone is more efficient than 50/60 directly across the line. Further variable speeds such as compressors and fans are coming as more efficient.

10. How is it going to make electricity cheaper? My windows on my house are fogged up due to them losing their insulating gas. I can spend $16,000 to have all them replaced with new insulated windows. I will never recover $16,000 in reduction in heating or cooling expenses. Living with what I have is the cheapest option.

Truthfully, nothing will. One way or another people will pay for the changes happening every day in infrastructure.

 

[MatthewDB]

Overhead AC lines are also limited. If you built an AC line from Maine to Los Angeles, the power would just radiate off into space.

What the heck does that mean?
You better get into contact with China because they're building the equivalent right now.

Ultra caps can be connected directly across the line. No need to rectify DC to charge them, then use an inverter to put it back. No need to transform DC when the caps are rated 25kv, 165kv, ect.

Capacitors are even worse in that regard. A capacitor stores energy in the form of CV2. If you don't move the voltage significantly, the stored energy is limited. A power system doesn't like voltage variation. That's mutually exclusive. Any cap storage has a converter to adjust the voltage.

Who says the DC-AC-DC conversion from to from each voltage level will be lossier? Even if so, the smaller size of the DC trafos, reduced transmission line lose, reduced T lines materials, ect will make even an 8% loss attractive.

Transmission transformers are 98% efficient. DC-DC converters aren't that good, especially at higher voltages.

 

[MatthewDB]

@ScottyUK: True, but at what threshold does the material saved vs inefficiency of high frequency meet?

When you hit the GW scale, under 100Hz.

 

[HamburgerHelper]

"What the heck does that mean?
You better get into contact with China because they're building the equivalent right now."


What China is building with its 1,000 kV system will extend out 3,324 kilometers. LA to Maine is like 3,200 miles. In my opinion it is kind of stupid planning. There is going to be a lot of outages. Every lightning storm that roams across that line is going to cause problems. But maybe that isn't the point. The point might be just to have their generation as far as possible from their population centers due to the pollution. I would be curious as to what they have to all do to use up all the vars created by the shunt capacitance to prevent overvoltages and thermal overloading.



------------------------------------------------------------------------------------------
If you can't explain it to a six year old, you don't understand it yourself.

 

[Mbrooke]

What the heck does that mean?
You better get into contact with China because they're building the equivalent right now.

AC or DC?

An AC line of that length would have so much capacitive charging current that it alone would be impractical in terms of power delivery. Think of it like this: how long is a 60Hz wave?

Capacitors are even worse in that regard. A capacitor stores energy in the form of CV2. If you don't move the voltage significantly, the stored energy is limited. A power system doesn't like voltage variation. That's mutually exclusive. Any cap storage has a converter to adjust the voltage.

Nope. Caps can be charged and discharged an unlimited number of times with little energy loss. Batteries on the other hand rely on chemical interactions which are less efficient and break down over time. Not to mention the fire risks, temperature control and other issues.

Transmission transformers are 98% efficient. DC-DC converters aren't that good, especially at higher voltages.

What if one day, they hit a 98.5%?

 

[waross]

As long as the Chinese line is, it is still probably cheaper than moving either the coal or the river closer to the major population centers.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[HamburgerHelper]

Coal by rail or pipeline (gas) starts beating out coal by wire as the distance and volume increases and it can piggyback on existing networks and not require new build out like with the electrical grid. I doubt this is the cheapest option for moving coal power.

------------------------------------------------------------------------------------------
If you can't explain it to a six year old, you don't understand it yourself.

 

[MatthewDB]

Nope. Caps can be charged and discharged an unlimited number of times with little energy loss. Batteries on the other hand rely on chemical interactions which are less efficient and break down over time. Not to mention the fire risks, temperature control and other issues.

My point sailed over your head. It is not about life of the capacitor, it is about effectivly implementing capacitor storage on the grid.

I'll try to break this down further: This is about conflicting voltage requirements. The voltage on any network shouldn't vary by more than a certain range, maybe 10% is a good number. More than that and you run into all sorts of problems. A capacitor stores energy by varying the voltage. Since energy absorbed / released is proportional to the square of the voltage on the capacitor, good utilization of the storage of the capacitor requires a large voltage variation. If you vary the voltage 100%, you get 100% of the capacitor capability. If you vary the voltage by 50%, you get 25%, 25% gets you 12.5%, 10% gets you 1%. Put those two together and you get only a few percent of the capacitors capability if you're just floating the cap on the line.

How do you address that? Put a DC-DC converter between the capacitor and the line.

 

[MatthewDB]

What if one day, they hit a 98.5%?

I think it will, someday. When that time comes we'll start to see a migration. Right now there isn't a focus on improving DC-DC topologies, they're all focused on AC-DC conversion.

 

[cranky108]

AC systems carry more than just power. They carry information on generation/load balance, and information system tuning, where DC systems don't have or carry that information.

AC systems may be more complex, if you don't work with complex numbers very much, but is generally built for long life. DC systems with complex electronics I doubt will last as long. I give it about 15 to 25 years for active components.

 

[MatthewDB]

AC systems carry more than just power. They carry information on generation/load balance, and information system tuning, where DC systems don't have or carry that information.

That's becoming less important with highly networked power systems now. The norm is to include fiber optics along every power line. Without having to worry about as many parameters with DC vs. AC, the information to transmit decreases.

AC systems may be more complex, if you don't work with complex numbers very much, but is generally built for long life. DC systems with complex electronics I doubt will last as long. I give it about 15 to 25 years for active components.

Your numbers are only valid for carefully designed systems with focus on long life. Ordinary electronics doesn't even make it that long. A typical industrial grade VFD lasts 5-10 years. The two big pareto items are electrolytic capacitors and thermal cycling of semiconductor joints (wirebonds and base-plate soldering).

Long life means going with all pressure contact for semiconductors and only film capacitors. For big central systems like HVDC pole converters that's not so bad. The penalty of regular maintenance isn't too bad either, because they're so massive and the labor costs are small compared to the revenue handled.

Making low cost, long life for distribution is going to be much harder. Nobody is going to service the equipment. The costs are much more sensitive. Want to talk about how long transformers can last? Check out this bank:

It's been in service for 80 years now and continues to carry load.

 

[waross]

Is that an original delta:delta or was it upgraded to wye:delta 60 or so years ago?

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[MatthewDB]

I know the business owner. It is 12kV:440V delta-delta. He's got lights inside for ground fault monitoring and a 440:240/120 single phase transformer for lighting and outlets. This setup isn't long for the world. It is an island of 12kV delta surrounded by 20.8kV multi-grounded neutral service. They're stringing grounds in the area and starting to convert to one circuit at a time over.

 

[MatthewDB]

Hopefully if this comes to pass, one of the things that can be improved is worldwide standardization of voltages, utilization connection (cords & plugs), etc.

I would also wonder what voltages would be used on the utilization side? 380VDC is the standard in data centers, but that does seem awful high for residential use. Particularly with the propensity of DC to hold an arc. Maybe time to use the Edison 3 wire system worldwide? That would give you 190V for lower power appliances.

380V was chosen because it is the most typical voltage in PFC SMPS between the boost PFC stage and the down-converter. This idea could be extended by picking 650V DC, the highest nominal on systems design for 400/480V class, which would allow for having shared design VFDs, power supplies, etc.. This could further extend to 930V for 600/690V class equipment and so on going up in voltage.

 

[waross]

We may see dual voltages in residential installations.
A higher voltage for major appliances such as heating loads and motor loads and a lower voltage for consumer electronics and LED lighting.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[waross]

Thank you for sharing that picture, Matthew.
I hope that you can give us an update after the conversion.
If they leave that bank in service, how will they handle the primary wye point?
Grounded with circulating currents?
Floating with higher voltage switching transients?
Fused? One partial solution was to install a fused cutout in the wye point connection. When local maintenance was done, the fuse would be installed in the wye point cutout to avoid switching transients. When all three phases were energized, the neutral fuse would be removed to avoid circulating currents and the wye point would be left floating.
I saw such a bank some years ago. I went back a few years later to take a picture but it was gone.
I suspect that they will retire the old transformers and go with 277 Volt pole pigs for a 480/277 Volt supply.

Bill
--------------------
"Why not the best?"
Jimmy Carter