Hacker News Reader

Comment by trollied | Hacker News Reader

Building a solar-only grid is hard. People don't realise that it's a huge challenge kepping the whole grid synced at 50/60Hz when there's a large amount of solar being exported, never mind the entire grid.

View on HN

Fast response battery systems largely solve the frequency issue. Unfortunately there is a lot of incentive to put solar on the grid by individual players, while the cost to do the frequency response is collectivized so tends to lag behind what is needed.

Bit surprised countries aren't just legislating this away. For every MW you put on the grid that doesn't have rotational inertia you need to contribute X money to a fund that builds stabilization capacity.

My understanding is that this is mostly a technology problem. Current solar does a lot of grid-following. If solar instead did some sort of "grid driving" with an external synchronizer I would expect this to be a fairly solvable problem.

Put a motor-generator pair next the solar or wind farm and the problem of frequency stabilization is largely solved. Adds a big flywheel for more initial mass.

But now you've changed the economics by adding a big, inefficient (0.95*0.95%), and expensive piece of equipment.

flywheel nonsense, just add battery.

> flywheel nonsense, just add battery

Flywheels are batteries. If your model is predicated on a particular chemistry of battery, it’s overfitted.

Which requires inverters, which are a huge problem.

Riddle me this, what instantaneous voltage should an inverter target if the grid's at a fault condition? Stated another way, how do you define the frequency of a non-sinuoid?

Careful, get the wrong answer and you get a short and you help melt millions of dollars of equipment, some with year old lead times.

it is no riddle. you are asking questions from 1942. except you put inverter into it. so you either think about current technology or go spark some wires.

you can island and not island at same time. decentralized grid means DECENTRALIZED grid.

every renewable needs battery next to it.

Indeed, not a problem to solve if you have all day to measure it.

The trick is measuring a frequency in the 60s of Hz of a non-sinusoid within a thousandth of a second.

Otherwise you get a short.

You don't need that, though, the problem is entirely solvable through the inverters. Having some batteries also helps. Combined it's much better grid intertia than spinning motors.

It’s crazy that our grids are still held together by the speed of a physically spinning generator. It’s like driving your electric motors with brush motors or using analog tape instead of digital audio, POTS instead of VoIP. A perfectly reasonable but still obsolete technological anachronism that will be gone within our lifetimes.

Watching the enormous progress we’re making on upgrading our energy infrastructure is definitely one of the most exciting things happening in tech. People seem determined to stick to old ideas.

Inverters are the problem. They're closed source and power engineers cant test their fault behavior well.

(One of ) the problem with inverter the angular velocity of a spinning rotor is trivially easy to define and measure. The operating frequency of the grid that an inverter needs to target is hard to define under the realtime constraints of a failing grid.

Which is what brought Spain down. It wasnt the panels and windmills, it was the inverters.

It would, indeed, make sense to have standardised behaviour for the inverters (they can simulate inertia perfectly well, but a lot do not), and incentives to contribute that value to the grid. It's not a hard to define target if you just want something that looks like inertia.

And the root cause of the blackout in Spain is still mostly the subject of speculation. AFAIK there has is no official report yet into the details of how things went wrong.

The term of art you are probably looking for here is grid forming inverter. The ENTSO-E class list on Wikipedia is probably a good starting point for some of the problems that one might want to handle:

https://en.wikipedia.org/wiki/Inverter-based_resource#Grid-f...

Why would anyone build a solar-only grid?

Nobody would, but the nuclear and carbon lobbies like to use straw men models - combining things like laptop batteries with solar to "prove that you cant live without them" and these models leak into public discourse.

A fully green energy grid would most cost effectively comprise of:

* Solar (generally strongest when wind is low, ~5x cheaper per kwh than nuclear power)

* Wind (generally strongest when sun is low, ~5x cheaper than nuclear power)

* Large scale interconnected grids to offset intermittency.

* Batteries and pumped storage for short term storage (~90% roundtrip efficiency but expensive to store long term)

* Syngas for seasonal storage (~50% roundtrip efficiency but cheap to store long term).

Reply on news.ycombinator.com