Ondol: Korean Underfloor Heating

One of the many aspects of the modern world we often take for granted is the very technology that keeps our accommodation at a habitable temperature. Examples of this include centralized heating systems using hot-water circulation, or blown air ducted to multiple rooms from a central furnace. Certainly in Europe, once the Romans shipped out, and before the industrial revolution, we were pretty cold unless someone lit a fire in the room. Every room. But not in Korea. The Ondol heating principles have been used constantly from about 5000 BC to only a few decades ago, keeping your average Korean countryman nice and toasty.

Having said that, the sophistication has improved a bit. Initially, the idea was to simply heat up a bunch of rocks in the fire, and bring them indoors, but Ondol quickly became part of the building itself. As will be seen from the video embedded below, the house sits on top of an elaborate double stack of serpentine channels, that circulate the hot combustion products from the furnace as thoroughly as possible, slowing down the gases and allowing their heat to transfer into the structure of the floor, and then radiate into the space above. It does bear more than a passing resemblance to the Roman hypocaust system, ruined examples of which can be found all over the UK and Europe. The skill demonstrated in the video is considerable, but must surely be an expensive build reserved for the most culturally aware Koreans who wish to live in simpler (and less hectic) locations in their country.

Maybe for the vast majority of us, this kind of thing is not viable, and we’re more likely to benefit from a more centralized approach, perhaps using waste heat from data centers or geothermal activity. (See: Iceland)

Thanks to [Keith] for the tip!

49 thoughts on “Ondol: Korean Underfloor Heating

  1. The trick is to get the airflow even. One modern method uses a small number of cast block types that work together in a 3D Truchet tile like method to allow you to build an airflow tree, actually is is more like the vasculayout of the lungs but you get the idea. Did the artwork for it a few decades back, saved it in an old Adobe .ai format that nothing FOSS seems to want to open now. :-(

      1. I only use FOSS these days and I rarely trust online services even if they are “free”, they are mostly not worth the privacy and security risk. If I must I will redo the work in OpenSCAD or Blender 3D’s geometry nodes, perhaps even FreeCAD, then I’d have something parametric and saved in an open file format.

          1. I disagree, 99% of users don’t have the mastery to exceed what FOSS offers. Proprietary software wins in edge cases, but with most FOSS scriptable in python even those cases can often be covered with some customisation.

  2. This style of heating is still used in the ethnic Korean part of China, specifically in villages. The house is heated when cooking meals and usually stays warmish though the day and night. The target is 20c warmer than the outside, otherwise you will “get sick”.

    Koreans also traditionally do everything on the floor. Eat, sit put out the sleeping mats in the same place. Once you realize that is where the heat is it makes more sense. South Korea still uses heats the floor, just with gas and running hotwater through the floor.

    1. Is there ever an issue with cleaning out the vents just hypothetically. Issue with a lot of European style things is build up of ash and such that has to be regularly cleaned

  3. That’s the same principle that the “gloria” system used in spain since medieval times, and the “gloria” system is also an evolution of roman hipocaust system.

  4. Korean heated floors offer winter comfort in a land that has truly frigid winters. You won’t find them in large cities, but they are still used in some rural locations. They are especially delightful for certain winter time activities performed on a futon.

    Considering their cultural similarity, I was surprised that a similar heating technique is not used in Japan. I once encountered hot spring water plumbed thru a floor at a rural Japanese onsen, but this was very unusual.

    If anyone has seen heated floors in Japan, please share your experience.

    1. Afaik, traditional Japanese construction is wood, because earthquakes and also is portable. Instead, they wore the equivalent of robes (no sash) and sat with the bottom opening beneath a heated table (kotatsu). This way the hot air would travel up the robes, out the neck. Modern versions are electric (resistive) instead of a pan of coals.

  5. The article claims about the Goguryeo wall paintings, dated from 1st century CE onwards. Then it notes that “heating an entire room first emerged in the mid-13th century”. CE that is. The next article notes UNggi, without pipes, kind of hypocaust, dated by radiocarbon at 1000bce.
    Where is the 5000BCE claim originating from?

    It seems that the earliest known systems of unrfloor pipes and heating seem were a prior art, maybe from Europe, like in the Aegean sea, in Akrotiri and Knossos sites, at about 2000BCE, that is 4000 years from today.

  6. Does anyone know how you clean it? A western house with a fireplace normally needs the chimney cleaned every once in a while or it becomes a fire hazard with the build up of soot. Wouldn’t the same apply here?

  7. The channels seem too small to go into for cleaning. How do you deal with the soot building up over time? (Sorry if it is in the video, I’m currently on a data-restricted internet connection, so plain-text preferred)

    1. I don’t think the deep channels can be cleaned. There are multiple outlet flues which I guess can be cleaned with a rod or snake.

      source is this PDF I found:
      Ondol Radiant Heat In Korea By Warren Viessman

    2. In traditional heat storage wood-burning devices (Korean ondol, Roman hypocaust, European masonry stove, etc) the goal is to burn the wood hot and fast and completely, and absorb the heat after the wood and combustion gasses react completely to CO2 and H2O. Creosote (the flammable deposit in western chimneys) is a product of incomplete combustion-burning without enough oxygen, followed by contact on a cold surface.

      Fireplaces were used as a fast and/or simple construction to provide heat for people and for cooking, and a hot, fast fire that completely combusts is difficult to maintain for long slow cooking without constant tending. The North American transition to cast iron stoves occurred because they added convenience, and where it was cold, generally wood supplies were plentiful. It doesn’t take more than a generation or so for the unwritten knowledge of efficient fire burning and heating practices to disappear, when “everybody” seems to be using the latest technological marvel.

      Have a look at this set of threads about modern DIY rocket mass heaters:

      1. Ah, so the knowledge of how to get a complete combustion replaces the yearly cleanout — for me the surprising part is, that this actually works out.
        Thanks for the link, seems to be quite a rabbit hole, so I’ll take a look into it when I get some consecutive hours of reading time tomorrow[tm] :)

      2. From what i saw, that house had no chimney, no chimney = bad draft = creosote. Also the thick smoke billowing out from the exhaust outlet suggests incomplete combustion. That house will develop soot, no doubts about it.

    3. The heat routing in the video is a very extreme example, with only efficacy in mind. Most of real-life ondols are straightforward, often just parallel ducts, therefore more clean-able and less prone to soot buildup.
      Nonetheless, once in a while (they say once in 3~4 years), you have to break up the floor, and remove soot from the ducts and the covering slabs. Lucky I don’t live in one of those houses ;-)

    1. Interesting article. According to DOI:10.1007/s10963-020-09144-2 “baked earth” from 5000bc is mud and straw, a kind terra-cota, not a heating element. The pdf proposed says that from 2800bc onwards “chinese writings” refer to heated floors. But the oldest Chinese text is dated to ~1200bc. The Ningxia Hui tombs, refered in the pdf, are from ~ 1000bc. And do they mention any heating system? A finding of a hypocaust/ondol system occurs for the ~1000bc period, in Alaska, 5000km away.

      Then again, i cannot spot texts or findings on heating pipes/channels in the area and the period.

      1. Yes, I can’t help that people sometimes barely scan a document and pull out the most extreme factoids. Reading the linked to doc clearly shows the 5000 year number is not for heat flooring. I think the fact that there were civilizations with this kind of technology even 2000 years ago to be pretty amazing. We humans are a pretty resourceful ilk.

    1. All wood heating stoves in scandinavia/fennoscandia use recirculation to extract the most of the heat from the fuel. A rocket stove has no heating benefits beyond what little heat can be stored in its construction.

    2. There is no recirculation back to the combustion chamber, only between the first and second “levels”. think of it more as a 3d labyrinth for the air between the firebox and the chimney. I believe it’s more about balancing the system (consistent path length from fire to chimney) in a rectangular layout when they are more accustomed to building square.

  8. So.. I’ve been wondering about this since hearing about Roman hypocaust systems. I guess my question applies to this too. What is it actually like inside such a room? Warm floor.. sure.. but is it smoky? As someone with a smoke allergy would I be uncomfortable in such a place? Just curious.

    1. No smoke or odor at all, unless the floor is cracked. In that case, there’s occasions where people have been asphyxiated. Koreans are used to the cold, similar to say Finn’s, and the construction techniques are impressive. They have these 8 inch cylinder charcoal blocks that they replace about every 6 to 8 hours. It makes a one room hooch toasty in 0 degree weather, but the water source used to be outside, and the bathroom, as it was, wasn’t part of the underground heating, so that sucked.

  9. The really wavy masonry in this makes it difficult for me to watch. A string line and some care will result in perfectly straight runs and flat floors. I sure hope this wasn’t an actual build for a family and just a demonstration for “TV”. Masons generally take pride in straight, level, and plumb work, this isn’t any of those.

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.