Hacker News Reader

Comment by brudgers | Hacker News Reader

The supports in the picture are massive and don't look cheap.

They look like ordinary “bents” for prefabricated metal buildings which are about the cheapest kind of building per square foot. Made out of steel, they are typically sold at a cost per ton like other steel shapes.

The cost of the scheme is in maintenance. No ground level access. No easy way to service panels in the middle of the roof. And lots of electrical current if you need to do something during day light because solar panels don’t turn off.

So I agree it is a dumb idea but not because of the cost of steel but for the lack of modularity and abundance of inaccessibility.

View on HN

Solar panels are cheap these days; wouldn’t be terribly surprised if they take the data center approach (if a few machines die, well, who cares, deal with it next time a major overhaul is required).

I would expect a middle ground. They collect statistics from the inverters and when one doesn't report, or reports lower than expected levels they schedule maintenance for next time somebody is there anyway. But yeah, they probably replace a whole string of panels at once. I expect someone will design a crane for this purpose so the whole string can be done quickly - most of the work is at ground level.

I'd also expect that they have a crew to wash them off once in a while (I'm guessing yearly), but this is cheap labor not electricians.

> No easy way to service panels in the middle of the roof

Cherry pickers like they’re used for every other type of infrastructure.

Those cherry pickers are used to access things above them, and a little to the side.

In this case, it is a ton to the side, and only a little above. And the cherry picker would be on dirt, which is not a very stable platform

Not going to work.

Construction already uses cranes for longer spans than this. They already know how to brace them on dirt as well. There is nothing large about this - they need slightly larger machines than you would expect at first because they can't be right next to the ditch, but nothing that construction doesn't already deal with in other situations.

Cranes are expensive, even small ones. And you need a bigger crane quickly when you have a horizontal reach.

And keep in mind all that reach is over live electrical panels.

Cranes are expensive yes. However they are common, and the needed horizontal reach isn't that much. (I've been in construction where we needed cranes that big - generally a house will use a larger crane than a commercial building - the house a large crane can reach the back corner and so you don't need to move it, commercial you have to move the crane anyway so you save with a cheaper one). Note that I'm not talking about tower cranes which are very big but a whole different beast that I don't know.

I don't know the details of their plan, but I wouldn't design the reach over live electrical panels myself. The obvious thing is not lift individual panels, it is to lift a whole string of panels already wired up and connected to the frame.You have somebody on each end (which is on land) to put the bolts in, connect the last wires (which are live, but nothing electricians don't work on live all the time). Even if the worst disaster happens - nobody is ever under something falling and the live wires fall to the ground without harming anyone.

The above is not difficult. Well there are a few more details that need to be figured out, but they all fit in the above plan and are obvious to anyone who has been in construction.

My AEC experience has been that cranes are expensive enough that “everyone” does whatever they can to avoid needing them and when they need one they rent one.

Because cranes are expensive crane operators are expensive and crane insurance is high (because cranes kill people), crane companies only work economically because crane companies can stay busy.

But when you rent, your schedule is dependent on every other crane customer’s schedule. You get a crane only when one is available and odds are you are not the crane company’s best customer nor the customer with the deepest pockets to pay priority service rates.

In my AEC experience cranes get “surprisingly bigger” as horizontal reach increases…levers work that way. By “surprisingly bigger” I mean without actual experience with crane tables…my first AEC job involved cranes and prestressed concrete in a business culture where employees had been killed with cranes. And I spent time in the yard where driveable gantries were handling big chunks of concrete “overhead”…by “overhead” I mean you could wind up under twenty kips if you weren’t careful and Carl didn’t know you were there.

You seem to be trying to refute me, but everything you said exactly matches my experience and what I'm trying to say. Your project managers and design should be such that you minimize your crane costs, and when it comes you get as much done as you can. Even though a crane is expensive, when used well their portion of the cost is tiny.

There are enough canals in CA that I would expect to be an important customer of the cranes. You don't get a crane on demand, but you can always plan such that when you get the crane there is plenty of work.

Project management is a construction function.

My comment way up thread criticized serviceability.

Electrical infrastructure needs to be serviceable during disaster recovery. You can plan for it can’t schedule for it.

And the worst time to need a crane is for disaster recovery.

Yes, you can just use a crane but “just” is doing a lot of heavy lifting.

I don't understand what problem you're worrying about though. Others have already brought up the data center example where Google often will leave a server rack in place for five years even if some of the individual computers in it have failed. That is, it's easy to replace an entire rack at once than to replace just parts.

Similarly, for a system like this, I would expect they would build in components, let's call them racks of panels, however many that are connected together. If one of these so-called racks fails, leave it up until whenever and then replace the whole thing.

Possibly you'd have to remove a bunch of panels from the canal if the support's collapsed, but you should be designing them for the highest possible wind roll, but so that's unlikely to happen. Well, earthquakes are a problem, and I'm not sure how much you can design for that, but once again, you design for that. In any case, if something happens in a disaster, well, you may have to remove some panels from the water, but that's a labor job that needs a little bit of safety training. But otherwise, no crane. You just remove it and throw it away. You come back with the cranes later to repair the system back to where it should be. I'd be thinking using a tow truck for disaster recovery, not a crane.

Public utilities are regulated.

Their capacity is sized for peak demand.

Their infrastructure is financed through bonds.

Alphabet is an advertising company.

And you think that’s cheap or easy in the middle of nowhere?

Almost anything is possible with enough money.

It is routine all over. Construction is always a game of moving people and equipment where the work is. There are a lot of people in construction who never work in the same city for more than a year, they just live in a local hotel and move the crane as needed.

I suspect this is cheaper than many projects just because the work is consistent - once you have the canals covered you keep a crew around to repair/replace them. That is you will need one crew full time within 2 hours of some small town so you hire someone who is willing to live in that small town (they get to go home to their family every night unlike the construction crew above working in a different city every year)

Maintenance, not construction. Can you not read?

Makes no difference. You schedule the crane when you need it. Just make sure there's enough to do that it's worthwhile for the crane to come out for the day's work.

Ho boy, the finance boys must love you.

What day is the earthquake?

Yes and this project is an example.

Jordan Harris, once a Virgin Records executive who had a hand in turning acts like Paula Abdul and The Smashing Pumpkins into household names, now serves as co-founder and CEO of Solar Aquagrid. He says the $20 million Project Nexus has done much more than serve as simple proof of concept.

https://solarbuildermag.com/projects/california-completes-wa...

A band promoter? Boy, I’m sold.

>> And lots of electrical current if you need to do something during day light because solar panels don’t turn off.

No. Lots of voltage (potential) but current only if you tap it. Panels will charge to a slightly high voltage state but do not become a problem.

    > The cost of the scheme is in maintenance. No ground level access. No easy way to service panels in the middle of the roof. And lots of electrical current if you need to do something during day light because solar panels don’t turn off.

This is classic HN "armchair critic" comment. Yeah, I'm sure the designers didn't think about any of these issues. If they need to clean the panels, they can use a cherry picker. If they need to service the electrical components underneath, they can wait until the canal is drained (usually once per year for cleaning), and, again, use a cherry picker.

For clarity I was using “service” in a technical sense.

Ground mounted panels can be serviced from a bucket truck right after a hail storm or other disaster. Disaster recovery is a critical design criterion for electrical infrastructure and the bucket truck is the standard electric utility service vehicle. Electrical utilities depend on each other and dispatch their fleets of bucket trucks across the country to aid in disaster recovery. If you drive much across the country, you might come across convoys of bucket trucks deploying down an interstate.

Which is to say cherry pickers don’t easily scale to handle the kind of critical events that need to be handled with electric utility infrastructure (and as an aside, in a pinch ground mounted solar panels can probably be serviced from a half ton pickup).

For context, in a former life I was an architect and one of my clients was an electrical co-operative. I also know a lineman socially.

You can read more about this project and the players in this article and draw your own conclusions:

https://solarbuildermag.com/projects/california-completes-wa...

Having to rent a cherry picker and being able to do maintenance only during a limited time per year significantly increases the cost of maintenance compared to an installation you can trivially access from ground level year-round. Canals being long and narrow don't exactly help either.

Sure, it can be done, but strictly from the PV perspective it doesn't make any economic sense compared to a regular solar farm when there's still plenty of land available. Similarly, from a "preventing evaporation" perspective going for a plain corrugated metal roof might not be much cheaper in installation, but it will be significantly easier to maintain.

Just because two thing can technically be combined, doesn't mean it makes sense to do so.

This is a classic HN “armchair critic” comment. Running a cherry picker team along a whole canal is far more expensive than maintenance would be on a field of solar panels.

With a pickup truck.

Except getting permission to build your "field of solar panels" costs X times as much as a local/state controlled area to build. Is that you, Brandolini calling?

I don’t even trust California to remember where they put these solar panels

While yes, these are cheap for steel structures, calling them cheap overall is silly. They literally cost more than the canal under them to make (assuming basic earthmoving and concrete formwork), especially when you include the cost for the footers.

As a way to save water, maybe. But it makes no sense as a way to make electricity. Someone else did the math, and it was approx $10/watt, which is silly.

The steel and concrete have the same materials cost as steel and concrete for ground mounted panels.

The large scale members usually mean there is less labor cost and mobilization time, and hence overhead (insurance, fencing, trailer, etc.). Like I said pre-engineered metal buildings are about the cheapest cost per square foot you can get at scale.

This is simply not true. I-beam's 4 ft deep, arched over a 100 ft wide canal would not be the same cost as a ground Mount system.

No commercial solar farms use large members to save on labor costs.

Standard ground mount equipment also uses pre-engineered metal, just of much lighter weight due to supporting a small fraction of the load.

The raw material cost for steel is by the ton. Barring unusual circumstances. And metal bents are very very structurally efficient.

Incidentally, they are not “I beams” and technically I-Beams are Wide Flange Shapes and used vertically as well as horizontally, and designated with as W(n)x(m) where W means wide flange, n is the depth in inches, and m is the weight in pounds per lineal foot. W shapes are rolled in rolling mills.

Bents are fabricated from steel plates to a design specific to the structure.

Reply on news.ycombinator.com