Hacker News Reader

Comment by sandworm101 | Hacker News Reader

What they do not model is the local effect of the cities that will need to be built to service all these panels. 9,000,000 square KM is a gigantic area. Millions of people will be needed to maintain these farms. Those people will need to either live locally or be transported regularly. The study should model how the presence of those people in such a relatively underpopulated area will impact local climate.

1 person per km2 = 9million people. But I'd bet you would need a couple dozen people for every km2 of solar farm, and associate infrastructure, meaning perhaps 100 millions people. And then all the people to provide services to those people ... this is like building a few new countries from scratch.

View on HN

This line of thought interested me so I went and looked for some numbers.

The 10 sqkm 648 MW Kamuthi Power Plant [1] was built in record time in 2016 at a rate of 11 MW/day with a crew of 8500, which comes to about 2 months. This project is many orders of magnitude larger, lets be generous and say there is a 20x economy of scale in this project above Kamuthi: 100 sqkm / mo / 8500 crew. With 100 such crews (a lot) it would still take 3/4 of a century just to install 9M sqkm, and the installation alone would employ ~1M people. Wow.

According to [2], the degradation rate varies between 0.5%-1% per year. Using a generous 0.1% as an estimate for maintenance would put it at 9000 sqkm needing maintenance/replacement every year which according to the estimates above would only require ~10 full-time crews from above. Still ~100k people, a big city for sure, but not millions.

Another thing to consider is that a project of this size would be decades in the future and I imagine it would heavily employ automation, so dropping a couple zeros off the number of people required could be quite feasible.

[1]: https://qz.com/india/854483/worlds-largest-solar-plant-kamut...

[2]: https://medium.com/@solar.dao/everything-you-need-to-know-ab...

1% per year? So some bits of these solar plants will last a century or more? I'd put a 20-year expectancy on everything, to cover both degredation and periodic upgrades. That is 4%+ per year. Nothing functional lasts for a century in the desert.

That figure is for the power generated only. Exposing the panels to the elements is likely to cut the life of some of them short.

But all in all it's essentially a rock generating electricity in certain circumstances. Not much to go wrong here.

Solar utility installations in California that I’m familiar with employ fewer than 1 person per square km.

And how do they keep all those panels clean? Who maintains the transformers when they break? Who swaps out the panels as they reach their life expectancy? It isn't just the night watchman, but the entire community of people involved in any large outdoor installation.

Like any utility, you don't need a guy just hanging around waiting for the transformers to explode. You truck him in every 50 years when that actually happens. It's time domain multiplexing, essentially.

The cleaning is largely a solved problem too.. You don't need 50 $10k/year people with squeegees, you just need one person in a $50k truck.

https://www.alibaba.com/product-detail/China-Road-tunnel-wal...

If there’s one job that robots are certainly going to take, it is solar panel squeegee guy.

And how do they keep all those panels clean?

Robots, like these from Ecoppia:

https://www.ecoppia.com/wp-content/uploads/2016/08/Product_D...

https://www.youtube.com/watch?v=VIH-C29N82o

Do they have problems with security? Try putting down any large piece of infrastructure in Africa - and careful that it doesn't get carried away.

Reply on news.ycombinator.com