Sustainable CFD

Hey there,

It’s Robin from CFD Engine & this is an email that has been bobbling around my drafts (& in my head) for quite a while & it still doesn’t qualify as a fully-thought-out thought. I wanted to share it with you though, in case you can help me to crystallise it.

I’ve been thinking about the environmental impact of CFD, specifically my CFD – the models, simulations & choices that I’m directly responsible for.

CFD has a huge amount to contribute in the search for sustainable solutions to environmental issues. We’re increasing the efficiency of transportation, improving industrial processes, cleaning & boosting power generation and much, much more besides.

There are people reading this email who are literally changing the world with their simulations.

I’ve worked on projects a little bit like that, but they’re not my everyday.

I’m usually making something go faster or working on the next generation of something you might like to buy – not quite so world changing.

Given that my projects aren’t typically that sustainable, what could I do to reduce the environmental impact of my CFD?

And should I even care, it’s only little old me after all?

Reduce, Reuse, Recycle

You’ve probably heard this mantra (or a variation) but is it any use in this context?

Whilst I’d love to reuse & recycle my simulations (& still get paid for them) I’m not sure my clients would be too happy. So we’re left with “reduce” and, as I’ve never known a CFD model to get smaller or simpler over time, so this might not be too easy.

We’re a greedy bunch (at least I am) give me some more compute and I’ll use it. More RAM and I’ll fill it with cells. More storage and I’ll write more files. More time and I might even fire up the DES (I won’t, who am I kidding?).

Reducing doesn’t come naturally to CFD people – so how about not doing things instead?

If you acknowledge that the most sustainable option is often not doing that thing in the first place – not making that trip/flight/purchase – then perhaps we should look at whether we actually need that extra node, that extra terabyte of storage, that extra million cells or even that extra run 🤔

Thinking points

I’m sure you’re already chewing on some of these questions, but these are the ones that easily come to my mind:

• Is there anything we can do to reduce our compute usage?
• How energy efficient is our compute resource & can we improve it?
• Where does our power come from?
• How much data are we storing & do we need it all?
• Do we need to refresh our hardware this year, or can it wait?
• Could/should we offset our activities, how & by how much?
• Should we be sending out hundreds of emails every Saturday? 🙈

Lot’s of questions, lots of thinking & no right answers.

From my perspective, the object of the exercise is to kick-start a train of thought that, hopefully, eventually get us into a position to be able to answer…

what are the environmental impacts of my next CFD run?

Stay in your lane

Like I said, this is an unfinished thought so, if you’re further down this track than me, I’d be keen to hear your conclusions. Is this something that’s on your mind? Have you found any answers or taken any action (on whatever scale)?

I don’t like straying out of my lane, but I feel like this is something worth pursuing.

Can you help me fill in the blanks?

Until next week, stay safe.