Industrial Solar Heat Hits 1000˚C

An image of an orange, translucent glowing quartz rod. Thermocouples can be seen at intervals along the rod looking in.

While electricity generation has been the star of the energy transition show, about half of the world’s energy consumption is to make heat. Many industrial processes rely on fossil fuels to reach high temps right now, but researchers at ETH Zurich have found a new way to crank up the heat with a solar thermal trap. [via SciTechDaily]

Heating water for showers or radiant floor systems in homes is old hat now, but industrial application of solar power has been few and far between. Part of the issue has been achieving high enough temperatures. Opaque absorbers can only ever get as hot as the incident surface where the sun hits them, but some translucent materials, like quartz can form thermal traps.

In a thermal trap, “it is possible to achieve temperatures that are higher in the bulk of the material than at the surface exposed to solar radiation.” In the study, the researchers were able to get a 450˚C surface to produce 1,050˚C interior temperature in the 300 mm long quartz rod. The system does rely on concentrated solar power, 135 suns-worth for this study, but mirror and lens systems for solar concentration already exist due to the aforementioned electrical power generation.

This isn’t the only time we’ve seen someone smelting on sunlight alone, and you can always do it less directly by using a hydrogen intermediary. If you’re wanting a more domestic-level of heat, why not try the wind if the sun doesn’t shine much in your neighborhood?

17 thoughts on “Industrial Solar Heat Hits 1000˚C

  1. One of the problems with concentrated solar power is the inconsistent heat input, and what that does to the generator system. Plants like the Ivanpah Solar Power Facility under-performed badly because it takes time to heat up the collectors and fluids.

    Unlike photovoltaics that generate whenever the sun is visible, a CSP system on an industrial scale is so slow that it more or less sees the average amount of sunlight over several hours, so a partially cloudy day is like putting a dimmer switch on the sun. It just can’t react fast enough, so the temperature remains low, so the turbine efficiency becomes low, or it doesn’t even start. The Ivanpah facility ended up making half as much power as projected by the amount of sunlight they were getting, and burning tons of natural gas just to keep the boilers hot.

    The same problems arise in trying to get industrial heat out of solar collectors: on most days you don’t have enough un-interrupted sunshine to get the collector up to the desired temperature, for long enough to actually use it, so you either have to deal with lower operating temperatures or have no use of it at all.

    1. That is making the assumption you didn’t build your sunlight capturing array bigger than bare minimum or don’t care how it gets, hotter the better at all times. Neither of which is likely to really be true now.

      It is relatively easy to occlude, defocus parts of or actively transfer excess heat from a large array that actually catches more sun than you need on a really good day. So if you build your sun capture system large enough to grab a ballpark of 133% of your industrial processes actual sustained energy demand under idea weather conditions you now have plenty of overhead, which both lets you bring the system back up to temperature in the morning quickly* and keep the energy input at the level you actually want most days. And as the sun capture part is pretty darn cheap to build if you have the land area, which most industrial sites will there isn’t much downside to building it bigger and cutting your energy input costs.

      *assuming you didn’t have a night shift working with alternative energy sources so the temperature is maintained anyway.

      1. >That is making the assumption you didn’t build your sunlight capturing array bigger than bare minimum

        When it’s cloudy and the sun gets covered, the whole array goes dark anyways and the collector quickly loses temperature. The extra mirrors do you no good there. When the sun comes out again, you can’t speed it up by concentrating more sunlight, because the collector would overheat and melt.

        Of course they tried to do exactly as you suggest. The Ivanpah facility has a gross capacity of 392 megawatts if you count up all the mirrors, while the turbine is 123 MW, so you got 219% over-provisions and it still doesn’t work properly.

        1. Actually, the 392 MW was the gross total of all the turbines, not the heliostats.

          The total average insolation that could be collected by the heliostats is 6,621,720 MWh per year or about 755 MW of solar thermal power, or roughly 1.5 GW over the average day considering that you get no power at night. The actual peak thermal power you could have is just over 2.1 GW.

        2. Doesn’t matter if the sun is covered a bit, its still getting quite a lot of energy to the ground level. At least on almost every day, as weather bad enough to really really block out the sun practically never happens.

          So if you are collecting and focusing a larger area – its gone from 1000 watt per square meter to perhaps 700, which is no big deal at all if your initial capture area is deliberately a bit oversized to your needs, and you actually only need say 500 watt sustained output from each square meter… And unlike Photo Voltaic solar panels where your capture efficiency is on a curve relative to light intensity and angle of incidence and was never that efficient the focusing mirror/lens type concepts really doesn’t much care, and the efficiency tends to be fantastic.

          (Pretty arbitrary numbers as the cost of a big mirror or lens for an industrial process amounts to zip and off the top of my head I can’t actually recall the expected sun intensity ballpark, think I’ve underestimated a bit. But as it is the space to actually do the capture that might meaningfully cost, and most industrial site are already large enough that isn’t much restriction…)

    2. You can also run an electrical furnace from PV panels. So it’s solar powered, almost stable, heat generator. The R&D here shows another path, they don’t say it’ll take over the world right now. Since the efficiency of a PV panel is, at best 20%, while the thermal collection is close to 100% but with a lower availability, you might substitute a cloudy day production by running from PV’s based electrical heater. On sunny day, you’ll sell your electricity since you don’t need it.

      1. Electric heating is tricky when it comes to high grade process heat, because the resistance wires you use start to melt at the temperatures you want – and Tungsten is expensive.

        That’s why e.g. electric iron smelters use arc heating.

          1. If seeking an alternative electric heating to resistive means, Id take induction over co2 lasers any day. Lasers are great at heating very precise spots, Not so much for general thermal transfer really

        1. I mean, it depends; it can still be simpler than that. SiC globars and some various alloys go up to 1600C or there’s MoSi to 1850C. The alumina you may be using to insulate things melts starting around 2000, it’s not exactly a low temperature. Of course something cheaper is carbon/graphite for resistance and you can always cover it in something refractory or put it in a vacuum to reach higher temperatures without vaporizing so fast.

          Eventually you do end up reinventing the light bulb, but on the other hand we have lots of experience in making resistive light bulbs very cheaply, even if we have to make sure to avoid calling them light bulbs instead of heating elements to sell them nowadays. The most economical of those may be reasonable at some scales despite the tungsten. Maybe somebody will reinvent the Nernst lamp in searching for a DIY option.

          Although it doesn’t have much bearing on industrial heat, it could be fun to see how long pencil leads could be heated in air if first coated in the right thing…

      2. >On sunny day, you’ll sell your electricity since you don’t need it.

        With other solar collectors and PV on the grid, neither does anybody else. That’s the problem with solar. When it’s on, it’s on for everybody in a thousand mile radius, and the “virtual battery” where you simply shove the power onto the grid doesn’t work.

        See, California, “Duck Curve”.

        1. You can’t have it both ways Dude and claim above that you are getting nothing because the sun is covered, and yet nobody else will need the solar power from the places where it isn’t. Weather patterns are just not that large (in general anyway) – its overcast and wet here and less than a mile down the round its brilliant sunshine all day isn’t exactly unusual… And the grids can and really should manage to very efficiently ship power a very very long way, unless no new infrastructure has ever been built since the 1960’s or something stupid…

          While there certainly can be some times nobody else will want the power in a sane radius to deliver it, especially if the contracts with the power company mean the consumers never get it at a discounted rate to match the current supply costs – as that means there is no incentive for the industrial users like an electric casting foundry to react to the supply price and do all the melts they can handle on the cheap days and focus on the mould prep on the expensive…

        2. You should watch the gridpoint-by-gridpoint numbers from during the eclipse. While it’s not THAT different from whenever a stormfront is moving through, the much sharper-edged shadow made for some really impressive ramp-down/ramp-up graphs for each solar site.

          ERCOT runs most of their data live, and with copious after-the-fact archives, which has been used to make some fairly interesting realtime renders of weather events via power-generation and power-demand changes, for example.

          I don’t have access to such detailed California data, but in Texas, clouds rarely have as big of a grid-wide effect as you’d predict from observations at any one facility.

          There’s also one other advantage for solar in Texas. A lot of the grid runs are based on layouts from the 1940s, and power demand for the average house is MASSIVELY higher than it was. This often means many miles of feeder lines between the nearest generator facility and that new subdivision with all its HVAC load. Currents are kept within the safety margin, of course, but they often cannot be kept below the efficiency knee. Placing even a couple of acres of solar near the subdivision can offload most of the HVAC from the grid, saving significantly more power than the array itself generates. And, in Texas, there’s plenty of space that’s not practical to zone as commercial or residential, and is too close to the new subdivision to really support agriculture or dairy. Sticking a few acres of panels on it is both cheap and beneficial to the distribution of the grid load. When the sun isn’t out, the cooling demand is greatly reduced, as well…

          It works well enough in Texas, although most of the factors would be different in other regions.

  2. I lived abroad in a climate similar to Florida in heat and humidity, and we used cheap solar to heat water; subsidized by electricity when the sun was not available. This alone saved us a great deal of money in the long run. Why is solar water heating nowhere to be seen in Florida?

  3. I don’t understand how this is different than simply providing a insulating the opaque absorber’s incident surface from conduction/convection. For example, common solar water heaters.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.