Path: news.eternal-september.org!eternal-september.org!.POSTED!not-for-mail
From: Don Y <blockedofcourse@foo.invalid>
Newsgroups: sci.electronics.design
Subject: Re: The Spanish Grid Drop-out - recently released information.
Date: Sat, 10 May 2025 23:18:28 -0700
Organization: A noiseless patient Spider
Lines: 43
Message-ID: <vvpffo$1onh$1@dont-email.me>
References: <vvnvto$3kd3i$1@dont-email.me> <vvo0k4$3kq8j$1@dont-email.me>
 <vvo5gv$3lr47$1@dont-email.me> <rf8v1klb6d9djefqfr2e2g8f9k3lgotka2@4ax.com>
 <qRTTP.120685$vK4b.43405@fx09.ams4>
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Injection-Date: Sun, 11 May 2025 08:18:33 +0200 (CEST)
Injection-Info: dont-email.me; posting-host="d3e1c2e5c2de319ca147c8663ff8692e";
	logging-data="58097"; mail-complaints-to="abuse@eternal-september.org";	posting-account="U2FsdGVkX1/eWP5aMxD4zbobMQG2KyyP"
User-Agent: Mozilla Thunderbird
Cancel-Lock: sha1:QN+I18D6j6xeOJ+d8pm2uHfcwjk=
In-Reply-To: <qRTTP.120685$vK4b.43405@fx09.ams4>
Content-Language: en-US

On 5/10/2025 7:22 PM, Chris Jones wrote:
> This is largely unnecessary - if the control signal that was being sent out by 
> the central controller to micromanage each power source was derived from a 
> function of the frequency, phase, voltage etc., then rather than trying to 
> distribute the result of this calculation to millions of devices with low 
> latency, it is better to distribute just the formula (once every few years or 
> as necessary), and run it on a microcontroller in the inverters several times 
> every mains cycle. They already have more than enough processing power.

I think any reliance on a "central controller" is inherently flawed.
Model the network.  Then, develop a distributed algorithm where
every cogenerator understands its role in generation -- not just that
of dumping power into the network but, also, of constraining the
*overall* network's response.

I.e., instead of thinking that the cogenerator needs to disconnect
in an anomalous situation, teach it to rectify that situation
within the constraints taht the network model imposes.

> I believe that there are some new regulatuions in at least one Austrlian state, 
> driven by the (fossil-fuel-stoked) fear of "too much solar destabilising the 
> grid", which require new home solar inverters to stop exporting power, unless 
> they receive continuous "permission to export" signals from our overlords, the 
> network operators. In other words, rather than exporting power in the case of 
> communications failure, it goes into the state of "export no power" in case of 
> communications failure, because otherwise people might unplug their internet to 
> export more scary solar power if exporting power was allowed when the internet 
> connection fails. This is a fairly new requirement, so not many compliant 
> devices are installed now, but once a few gigawatts of these inverters are 
> running, it will be interesting to see what happens when there is a major 
> internet outage on a hot summer day, and all of those gigawatts suddenly go 
> away. Hopefully they thought of that but I doubt it.
> 
> The rapid control algorithms should be distributed, and the only low-latency 
> communication signals they should rely upon are frequency and voltage.

Exactly.  It is surprising how many algorithms that we think of as
centralized can, in fact, be distributed.  But, that has to be
a factor in the algorithm's design; you can't just convert an
algorithm designed with the idea of centralized control into
one where multiple actors are involved and add an "escape clause"
to handle the corner cases that you failed to anticipate.