Podcast: Play in new window | Download
Subscribe: Apple Podcasts | RSS
This episode is brought to you by Mouser Electronics. This week we hear about Digital Therapeutics. Go to TheAmpHour.com/digitalhealth to learn more.
Welcome, Paul Zawada of Syntonous LLC!
- Paul was nice enough to respond when we said we were hoping to talk to someone about the grid/smart grid. He is an “operations technology engineer”, which means he straddles the line between the power side and the technology side of things.
- What are grid operators’ appetite for new tech?
- How do blackouts happen?
- Generation, transmission, distribution
- i2r losses are why they operate long distance transmission at 765 kV.
- China’s grid has newer tooling, which allows it to go at higher voltages.
- You can tell the voltage of transmission lines by “counting the bells”. It’s roughly 15kV per bell and they do odd things to the shapes of insulators to prevent arcing.
- Video about the Ohio blackouts in 2003 we talked about last week
- Where does control happen for “the grid”?
- Frequency determines the health of the grid
- Bike analogy
- Tesla battery – Hornsdale Power Reserve
- Moving to a renewable grid has some potential issues
- Grid has “inertia” due to the spun up generators. They’re working on modeling “Synthetic inertia” for a more renewable grid.
- “A disturbance in the electromotive force” (I’m sorry Dave, that title was too long)
- Dave asks why does the voltage go up on his setup?
- RMS voltage vs Vpp
- The Power Triangle
- ELI the ICE man, CIVIL
- 99% of the time, loads are inductive
- The role of reactive power in the grid
- VAR tends to control the voltage
- Real power determines the frequency
- Hawaii example of a local grid with a lot of solar (and that’s kind of a problem)
- Capacitors help to balance the grid when it there are too many inductive loads.
- “Grid Following inverter” (vs “Grid Forming”)
- Dynamic inverter
- 4 quadrant inverter
- “Don’t call it imaginary power”
- Inverters have to be able to deal with frequency changes and take setpoints from the grid
- The 50.2 Hz problem
- Smart Grid
- California Rule 21
- Smart meter and Power Line Carrier Standard (PRIME)
- Communication with meters
- Private LTE
- Pool analogy
- Stability of the grid wasn’t built for two way flow
- Managing a wide range of inverters is difficult
- Standard for comms by SunSpec (Smart Inverter Standards)
- Control systems with SCADA
- While analyzing hte grid, they use a state estimator (which is a mathematical model of the various sources in a grid system) and then do real time contingency analysis
- Spinning reserve
- Rolling blackout is when the grid systematically takes different parts of the grid offline to preserve the overall ability to deliver power.
- Texas example
- Dave asked about available power on the grid
- Steel recycling takes a TON of power.
- Different wires from generation (in Texas)
- Community Energy Storage (“Backyard Batteries” w/ Four-Quadrant Inverters)
- Wind turbine go to DC first (because of variable wind speeds), but are then feed to an inverter
- Proposed Soo Green HVDC Link [Underground]
- Grain Belt Express: Proposed Kansas-Missouri-Illinois HVDC Transmission Line
- Pacific DC Intertie
- Silicon transformer
- Paul’s pet peeve is people asking him about “exploding transformers”. It’s almost always a fuse or a cable shorting to ground.
- Expulsion Fuse Blowing (NSFW Language)
- Another Expulsion Fuse Blowing
- DC vs AC arc
- 250VDC control of relax elements in power plants
- Air Blast Circuit Breaker
- “(The grid is) the most complex machine we ‘ve ever built”
- Consulting and day job AND he is working on a Doctorate of Technology at Purdue. We’re grateful he found time to talk to us as well!
- Follow Paul on Twitter! @engineerz
Erich. Wagner says
Re the 50.2 hz problem and oscillation. Dunno about modern inverters but my 2003 Sunny boy inverters once tripped will not restart until 180 seconds of normal parameters have resumed. I think this means the frequency oscillations are on the timescales of minutes!
Gene Schroeder says
This was a great show, and interesting to me, having had a little exposure to the power industry years ago.
I think the answer to Dave’s question about have many generating plant is more economic than technical. Having lots of generating plans operating at low capacity is very inefficient in terms of capital and operating expenses.
I did a little work (well, not much because I was an intern) over 40 years ago on a “Static VAR Generator” that used power electronics to generate VARs, nearly instantaneously, as needed. It was static in the sense there were no moving parts, because previous solutions involved large rotating machines. I knew of a static var generator being used next to a steel recycling plant where high voltage electrodes were lowered into a furnace full of scrap steel, and all hell would break loose (electrically and mechanically) until a puddle got started.
Great show. Glad Dave could join in as well for the discussion.
Paul Zawada says
I finally had a chance to do a little research on Dave’s question regarding voltage rise on distribution feeders with a high solar penetration. I think the conclusion that we came up with that the voltage goes up due to reduced losses is largely correct. i.e. Because the solar inverters are closer to the load, less current flows from the substation, resulting in less of a voltage drop across the feeder. (Note: VARS in the system is largely responsible for maintaining voltage, but at the end of the day, ohms law still applies and more current across the wire will result in higher voltage drop.) This affect would be more pronounced with more generation at the end of the feeder.
However, a report ( https://www.nrel.gov/docs/fy16osti/63114.pdf ) from the National Renewable Energy Laboratory explains that the voltage phenomena with very high renewable penetration is more complex than this simple explanation. Other factors go into the equation, including where capacitor banks are located, cap control type (voltage sensing vs. VAR sensing), voltage regulation at the substation, etc. There’s a whole host of voltage problems that can make a distribution planning engineer’s job very challenging. This is part of the evolution of what used to be a simple, one-way portion of the grid to a more complex power environment.
Jelle Haandrikman says
Thank you for explaining this further. Over here in the Netherlands, the wire operators have been upgrading substations a lot in the rural areas. Because they produce more power than that they consumes. I would imagine that injecting more current in the secondary side of a smaller transfer will also increase the voltage. Because the current can’t go anywhere and all the losses in the transformer.
Øyvind Mjanger says
As the solar inverters are current sources they just follow the grid voltage.
So normally you get a voltage drop across the distribution line from the closest transformer depending on the load. This drop can be significant when you have old, under dimensioned lines. So houses at the end of the line have often too low voltage compared to the houses closest to the transformer.
But when you start injecting/exporting current from that house, the voltage at the transformer still will be the same, but instead of subtracting the wire voltage drop they will be adding to the voltage as the current goes the opposite way.
Solar inverters are not defining the AC voltage. Just following the AC voltage shape and feeding out a current.
Entrance to the rabbit hole: https://en.wikipedia.org/wiki/Synchronous_condenser
See the Gallery for “Synchronous condenser unit at Templestowe substation, Victoria, AUSTRALIA”
(Deep end starts at “See also”:)
Thanks for the show!
Matthew Suffidy says
I do not know about this topic, but I think David Jones sort of ventured into the idea of why the grid could just not act as a high amperage bus bar. I am guessing it was answered as a connection between the mechanical resistance of rotation vs load. It was mentioned it may cause a measurable change in generation frequency. Maybe a plant could have some inverters to supply the power waveform as expected at any turbine frequencies.
In 2003 I was renting in Quebec across from Ottawa just over the river, because it is cheaper than Ottawa. I just arrived at my workplace in Ottawa, when there was a power failure that involved the North East US. I just went home over the river and it was OK because Quebec has it’s own power system. At that point all of Ottawa showed up on the other side looking for something to do with their evening. It was back to normal in a few days. I was wondering why my parents phone worked on the Ottawa side. A reporter on the news asked the Bell guy and he said ‘our networks remain functional’. The reporter asked ‘could this power be tapped?’, maybe as sort of as a joke, and the Bell guy said something to the effect of ‘that was on the list of restricted topics given to you’. The power minister blamed everyone else and was mentioning ‘the eggheads that got us into this’. Anyway.
We had some tornadoes a few years ago and the lights downtown just kept kind of teasing you as they went out for a second a few times. When it was over, the whole city minus the downtown was out of power, so maybe they designed it to be a bit more resilient.
Chris Lott says
I generally don’t follow the power side of the industry, but found this talk fascinating. The part about simulation reminded me of an old analog computer I worked on back in college — I had a part time job for a professor who specialized in power. You can see a picture of this beast here:
David Pastl says
To expand upon the explanation of why you can’t have infinite spinning generation waiting I think it helps to explain more about the generation side and where energy comes from. Also that in a way, we do have more “generation” “on the grid” than is currently being used, as I’ll explain.
Paul mentioned spinning generation and the concept of spinning reserve. This works by having power plants ready to “spool up” and “come online” as needed to keep the frequency stable. In North America there is a directive that we should have spinning reserve equal to the actual power being used, in response to the massive blankouts we had in the east. I may be off on the exact amount, but it’s a very significant amount of reserve. I think reserve also means it needs to be operating within 20 minutes of needing it.
What counts as spinning reserve? This can battery banks like you have in Australia or it can be natural gas fired turbines (amoung other things). The latter is very common here in Saskatchewan, where our generation mainly comes from Fossil Fuels. A natural gas turbine can come online in minutes and be generating to peak demand.
You can have coal spinning reserve but as Gene mentioned it’s not very efficient. It takes a long time to get a coal plant making steam and it’s going to be very expensive to have it generating “peak steam” and just dumping that into the atmosphere. Remember you can’t have that steam going into the turbine as if you put 100MW of mechanical energy into a turbine and only have a 1MW load your turbine will continue to speed up until:
1. The turbine is losing 99MW to mechanical losses in the bearings, air friction, resistive losses, etc. Likely you’ll have a bad day.
2. More likely, It explodes from centripetal forces and you have a bad day.
When bringing online a large turbine from a steam power plant you also have to make the turbine synchronous to the grid BEFORE connecting it to the grid. If you attempt to connect a turbine that’s out of phase to the grid, it will damage the turbine. It’s kind of like changing gears in your car without a clutch or attempting to match revs. You can do it… but something is going to be damaged. The grid has a large amount of inertia (your car) and the individual turbine (your engine) is small in comparison. So you’ll either break the shaft between the turbine blades and the generator, or the wiring (your gearbox) going to the generator.
I took a tour of our largest power station here and they said they actually have an engineer sign off before they connect a spinning turbine to the grid, because of the liability and risk of damaging something.
Everything above is also part of the explanation of why load leveling is so important. Since it cost so much to take a turbine offline then bring it back online and operator would much rather just leave it running. Here in Saskatchewan our solution was to provide cheap power at night to a local metal recycling company. When demand goes down, they turn on the arc furnace and start melting steel.
Looking back at the generation side with large spinning masses, this is why you end up with frequency increases when load goes down. You’re putting 100MW into the grid but if it’s only demanding 99MW, your spinning generation will continue to increase speed until it absorbs 1MW of energy into the generator/turbine itself. This is because they are synchronous generators and not a DC motor. A DC generator would increase voltage but an AC synchronous generator will not. As Paul mentioned, what why you have controls at the plant limiting the mechanical energy (steam) going into the plant. The same thing with hydro, you don’t put water through the turbine unless you’re planning on generating power from it. Otherwise you just let it out the spill gates. (I think Hydro counts as spinning reserve, but I’m not 100% on that).
For further reading you might also enjoy learning about Voltage vs Current transformers, which I was surprised to learn was a thing when I started in the Power Industry: https://www.difference.wiki/current-transformer-vs-voltage-transformer/
Mike Turner says
‘Spinning reserve’ always involves time to come online. Rarely can this happen instantaneously. Between that and the turn down, the minimum power generated compared to full power output, things get complicated in a hurry when trying to balance things and keep it running for hours.
My experience is only with combined cycle plants (combustion turbines (CT’s) augmented with HRSG’s (heat recovery steam generators)) that feed steam turbines (ST’s)), It takes time to start a CT and get it ready to make power. Temperatures need to come up and be stable because CT’s can be finicky and break. I’ve seen a failure or two. Millions of dollars. Insurance.
Government environmental requirements dictate emissions limits and these sometimes dominate whether or how soon the turbine can be run and the penalties are not trivial.
At a Hawaii power plant, temps and emissions had to be stable and it was sometimes hard to get ready inside of an hour. California had even more restrictions on operating CT’s. CEM’s (continuous emissions monitoring) is always looking over your shoulder and the introduction of predictive CEM’s has limited what can be done to ramp up to a generation ready condition.
Once CT’s are up and running, the ST’s also had startup limitations. Things had to heat up and stabilize because the ST components like bearings and valves were surprisingly fragile. It was not a problem but required time to do right and protect the hardware. Running a steam valve at other than fully closed or open causes its own set of problems, generally having to do with steam erosion of valve seats.
Black start, startup without requiring power from the grid, can offer its own set of problems and time is again one of those. All of the power plants I worked on were capable of black start and had at least one diesel generator and a non-trivial UPS to enable a true black start. Sacrificing a diesel water pump for fire protection might have been allowed but doing the same for a diesel startup generator just to get online quicker was never an option.
The power generations switch yards I worked on had Multilin hardware that had serious power requirements and controls the keep things under control. The Multilin’s monitored voltage, current, phase, impedance, and a myriad of other things and had literally hundreds of configuration parameters. Fun.
It can be a rush standing in the middle of a road outside a power plant switch yard fence and use the laptop in your hands to close and open a high line switch for the first time. Arcs several feet long are the norm. Things can go wrong in a hurry in a very spectacular way. Pretty.
Note, all of my experience in power plant control systems is solely with US plants and one non-US plant.
Amphour is cool. Please keep up the good work as long as it’s fun.