The London Underground Is Too Hot, But It’s Not An Easy Fix

The London Underground is an iconic piece of Victorian era engineering. What started in 1863 quickly became a core piece of infrastructure that would define the modern character of the British capital. It’s grown and changed immensely in the many years that have passed. Sadly, increasing patronage and more trains have created problems that the original designers never envisaged.

Deep in those London tunnels lies an engineering challenge. The Tube is literally cooking itself. Every day, millions of commuters descend into a network of tunnels that have been absorbing heat since the reign of Queen Victoria. Those clay-lined tubes have been soaking up excess thermal energy like a giant underground radiator, and now they’re giving it back with interest. The tunnels are simply too hot, and cooling them down is inordinately difficult.

The Perfect Storm of Thermal Chaos

The Tube’s heat problem isn’t just about one thing gone wrong – it’s about everything gone wrong at once. When Victorian engineers designed these tunnels, cooling wasn’t a major consideration. The tight, compact tunnels were built deep, nestled in the clay beneath London. In the early days, temperatures in the Underground were considered comfortably low.

“The Underground’s the only spot for comfort when the days are hot; it is cooler below.” – London Underground poster, 1926

Originally, the clay surrounding the tunnels sat at around 14°C, acting as a heat sink for the network. However, over the years, with more trains coming and going and more heat pouring in, the temperature has risen. It now typically sits anywhere from 19° to 26 °C. That’s just the earth around the tunnels, though. Air temperatures are worse—hitting as high as 47°C during a 2006 heatwave. The problem has been a continual bugbear of the beloved Tube, with concerns that future heatwaves could see temperatures rise ever higher.

Victoria and Central have been the hottest lines in recent years, according to TfL data.

The problem varies depending on which part of the Tube you’re on; some lines are worse than others. The Central Line is worthy of the nickname “The Central Heat Line”, with temperatures historically reaching 35°C. That’s not just uncomfortable – it’s approaching the limit of what the human body can handle efficiently. Much of this is due to the tunnel’s design. Opened in 1900, it featured two compact tunnels buried over 20 meters underground with minimal space for ventilation. It’s one of the so-called “deep-level” lines on the Underground network. Meanwhile, the Victoria line hit 31°C at its peak in 2023, and actually overtook the Central line as the hottest line, recording an average temperature of 28°C last year. Contrast that with the newer Jubilee line, which recorded an average temperature of just 22°C—far more comfortable.

To understand the problem, we need to know where the heat is coming from. A breakdown of heat sources was provided by Rail Engineering in 2007. Trains using their brakes, converting kinetic energy to heat, contributed 38% of the total heat input to the underground. The rest was put down to mechanical sources (22%) and the drivetrain (16%)—because those big electric motors get hot in operation.

TfL at times has to remind customers that the Underground is warm even when it’s cold outside.

The rest of the heat came from a variety of sources, with train auxiliary equipment and tunnel support systems making up 13% and 4% respectively. The human factor can’t be ignored—each passenger is basically a 100-watt heater on legs. Multiply that by the millions of commuters that pass through each day, and you can see the scale of the problem. Indeed, passengers contributed the final 7% of heat generation in the Tube system. Of all the heat generated in the Tube, 79% passed into the tunnel walls, with 11% going into the tunnel itself. The remainder—just 10%—was removed via ventilation.

While the Tube had been slowly getting hotter for some time, the problem really started coming to a head in the mid-2000s, particularly when the European heatwave hit in 2006. Solutions were demanded, but the Underground wasn’t going to make it easy. The oldest parts of the network presented the greatest challenges, with precious little space to fit additional equipment for cooling. Many lines were simply too tight to allow for air conditioners to be retrofitted to existing trains, for example. Even if they were fitted, there would be a further problem of how to remove the additional waste heat generated from the tunnels, which were already too tight to ventilate effectively.

Victoria Station has had plenty of attention in recent decades, with TfL installing new cooling systems. Credit: Oxyman, GNU Free Documentation License

The quagmire had even prompted then-Mayor Ken Livingstone to put forth a £100,000 bounty for anyone that could solve the problem.  However, it went unawarded. Despite over 3,500 proposals, the Underground authorities found only unworkable or unaffordable solutions, or ones they were already considering.

As you might expect, the problem hasn’t just gone away. Indeed, British media have begun regularly putting out articles on the hottest tube lines each year, as well as updates on what is to be done. Looking ahead, climate change is only going to make this underground sauna more challenging to manage. TfL’s engineers are in a race against time and physics, trying to cool a system that was never designed to be cooled.

Transport for London’s engineers haven’t taking this lying down, however. In recent decades, they’ve thrown a range of complicated solutions at this difficult problem. Victoria Station saw major upgrades, with the successful trial of a groundwater-based cooling system and heavily-upgraded ventilation. On the toasty Central line, engineers realized there was an additional heat input into the system. Trains travelled above ground for part of their route, which would see them heat up in the sun and then bring that energy underground. Countermeasures included installing reflective material on train roofs and solar-reducing films on the windows.

Trials of a new panel-based cooling system have also taken place in recent years at the disused Holborn station, with TfL considering a rollout to various stations after successful trials. The system involves circulating cold water through a curved metal structure. Air is chilled by blowing it through the curved panels and into the station. The system is designed specifically to operate in stations on the deep parts of the Tube network, with an eye to keeping maintenance and operation of the system as practical as possible.

Subsurface lines have been running S-Stock trains, which feature full air conditioning to keep passengers comfortable. Credit: (c) Transport for London

Some Tube lines have been lucky enough to get air-conditioned trains, too. These can be found on the Circle, District, Hammersmith & City, and Metropolitan lines. The modern S-Stock trains run largely on the less-deep sub-surface Tube lines, where it’s possible to easily deal with the hot exhaust of the air conditioning systems. These trains also have regenerative brakes, which turn some kinetic energy back into electricity to feed into the tube network. This cuts the amount of kinetic energy turned into heat, which aids in keeping the network cooler.

The Picadilly line is due to gain air conditioning in 2025, when it abandons its 1973 Stock trains for newer models. Other lines will have to wait longer. Central Line is slated to receive new air-conditioned trains in early 2030, which will replace the aging 1992 Stock models operating on that line. Bakerloo, Waterloo and City, and Jubilee lines are slated to receive upgraded trains “within the next 20 years” according to a Transport for London statement late last year.

The Picadilly line will see its aging trains replaced with newer air-conditioned models starting in 2025.

Air conditioned trains will help to some degree by cooling passengers on the move. However, those air conditioners will necessarily pump heat out of carriages and straight into the tunnels the trains are travelling through, plus some waste heat to boot. That heat will have to be dealt with one way or another, lest the network heat up further. There’s also the problem that passengers on platforms will still be exposed to high temperatures. Ultimately, both the stations and the trains need to be brought down to reasonable temperature levels. Ideally, the tunnels would be, too, in order to protect any customers that get stuck in a tunnel on a broken-down service. TfL also needs to find a way to bring temperatures under control if it wants to increase services. More trains means more heat going into the system—so it’s important to find a way to pull more heat out, too.

Overall, the problem is still a long way from being solved. The fact is that the London Underground has 11 lines, 272 stations, and more than 4,000 trains. Upgrading all of those at once simply isn’t economically viable. Instead, it appears that Transport for London will keep chipping away at the issue, bit by bit, over the years to come. Ideally, this will outpace any increases in average temperatures brought on by our seemingly-ever-hotter climate. For now, London’s commuters will continue their daily descent into one of the world’s most interesting thermal management case studies. Just remember to bring a bottle of water and some breathable clothing– you’re going to need it.

42 thoughts on “The London Underground Is Too Hot, But It’s Not An Easy Fix

  1. Kinda thinking they’re missing a trick here and wondering if they could cool the tunnels by recycling the heat to re-sell to businesses, homes etc. that are local to the lines…

      1. You’re thinking in 2D the solution is easy in 3D cause when we’re in summer Australia is in winter. Simply install heat pumps to send the energy through to them.

        I wonder if Khan will honour Livingston’s reward?

      2. The ground has enough thermal mass to carry the summer heat into the winter. For this reason I don’t get the idea of putting an air heat exchanger on the front/backyard for a heat pump in a new building. When you are digging up the ground anyway, just place some pipes to pull the air through into the heat exchanger, and start next to the yearly average temperature instead of what it’s now.

    1. You’d still need a heat pump system to move the heat, which is one of the options they have / are considering. But it’s expensive, and it has to fit into some very small spaces.

      London does have a (small) number of housing developments built for district heating though, so it’s a possibility in those places. I know the Jubilee line runs right under such a development at Bermondsey. But in neighbourhoods where each premise already has its own heating, you’d have to build a new hot water main to each building and then subsidize the owners to convert each property.

      One possibility that has been looked at is cooling the tube using water mains. If you heat the water coming out of everyone’s tap by a couple of degrees, then everyone’s boiler uses slightly less energy. So it basically spreads the benefits of district heating over the whole city. But like all the other options, it’s still a big expensive complicated project.

    2. even dumping the heat outside, which is simpler than recycling it and requires far less equipment, doesn’t fit in the space they have. Taking on the bigger challenge that requires more equipment and more space is a non-starter.

      1. well, it comes down to heat transfer. from what i read, the walls and ground are not able to cool the air inside anymore. i think they should reverse the system: attach cooling ribs to the walls and suck out the warm air. the trains will work like a piston, moving the air by the cooling ribs. but anyway you do it, there should be an cold airflow from outside in the winter and it will take a while to lower the temperature of the ground mass surrounding the tubes.

  2. Why couldn’t geothermal cooling be used here if the air temp is significantly warmer than the ground temp? A cooling loop and some air handlers does not seem prohibitively expensive, and I’d imagine there’s not much to worry about drilling through at the bottom of a subway station. Especially considering air conditioned trains are only going to increase the air temperature on the platform.

    1. Geothermal cooling is exactly what they were using when the tunnnels were built. The problem is that you’re rejecting heat into the ground, but the tunnel is also surrounded by that same ground, and after ~30 years they reach equilbrium. You could spread the heat over a wider volume, so it would work for longer, but if you’re spending money to drill then it’s probably better to drill upwards and make more ventilation…

      1. My thought was drilling up for vents is going to be prohibitively expensive at some stations as there is likely sewage/power lines in the ground and buildings on top of a lot of stations. The permitting, surveying, and negotiating with property owners would cost a lot in time and money. This bureaucratic overhead is why extending subways in NYC is so crazy expensive – the digging-out-the-ground part is the small end of the decimal.

        Also, ventilation won’t work well on a hot day when surface air is also miserably hot -> and that’s when riders need cooling the most.

    1. One of the first subway systems used pumped air to push the train. Huge steam-powered bellows occupied both ends of a tubular tunnel containing a cylinder-shaped train car that filled up the tunnel like a cork.

      Parts of it still remain underneath Broadway in NYC. You used to be able to use some discreet access tunnels to get into it… Was the site of some good parties. Not sure if it’s still possible for a civilian to enter today.

      I found an explorer who recently found one of the ends of the line and the engine room:
      https://www.youtube.com/watch?v=0bLJFncU-s4

  3. What about a heat pump/phase change material combo? Pump heat into “hot ice” or similar in huge vats?

    I can’t believe that that pumping hot water up so people could have cheaper (or free) hot water for showering, etc. in summer, and cheaper heating in winter would be that hard to sell the people on..

  4. Just ventilate the tunnels at night with a big fan. Turn the fan off during the day when it’s hot outside. Same thing you do in a house with no A/C. If there aren’t enough openings to accomplish this, then the tunnels are probably unhealthy to be in anyways, and they should start digging ventilation shafts immediately.

    1. I was also thinking to cool things down at night, when there is not traffic.
      If space to install heat pumps (which, as it was said, could transfert heat to houses) is too small, what about embed them on special trains, drive them in the tunnels to absorb heat (and yes, here, using phase change materials can be useful), and release heat outside during the next day.
      Of course, this heat is easier to use during winter, when needed, but even if you can only evacuate heat on colder months, this can work, it will just take several years to slowly cool down the tunnels (it took years to warm them up ^^).
      Selling this heat could help to finance the process.

      1. I like the PCM idea. Instead of special trains, put a special car on the back that is insulated and filled with a PCM. While underground they run heat pumps that sink into the PCM and cool the air around them.

        Then when they go above ground they either swap cars and the PCM car is drained of heat somewhere or they drain in reverse automatically while above ground.

        The biggest question is if we can efficiently use a heat pump on a PCM. Too high of temps will kill the plan.

        One side benefit is that, depending on the PCM, the heat could be recovered and used to generate a bit of electricity to offset losses.

  5. This is a problem dear to my heart as I hate high temperatures wherever they occur.

    The difficulty with the deep lines of the Tube is simply one of accessibility – it is not possible or even safe to run long lines circulating coolant from the tunnels to the surface. Also, the cramped spaces make large installations of anything unsafe as escape routes can be obstructed.

    Simple and brutal physics dictates that the only ways of reducing temperature are a) to extract heat energy from and b) reduce the heat being input to the system.

    To these ends the following must be done:

    Upgrade all extractor fans and air ducts and run them hard and continuously whenever the outside air temperature is lower than the tunnel temperatures. Also install mobile fans in the tunnel mouths to blast cold air into the system when the passenger trains are not running at night. Prioritise this whenever the weather is very cold. This is necessary to start the process of extracting the enormous amount of heat that has over decades leached into the tunnel walls and the earth beyond.

    Clamp down on heat sources in the system by implementing a heat load monitoring panel that examines existing sources and proposed new sources.

    As noted above, the waste heat from friction breaking is the major heat load on the system. Reducing it is technically very difficult. The Jubilee Line extension deals with this problem by using under-platform extraction ducts in concert with the platform edge barriers and sliding doors so that the hot air from the brakes never enters the tunnel system. As noted above this results in a much cooler system. Unfortunately this cannot be retrofitted in the older stations as there is nothing like enough room for the very large air ducts required.

    Regenerative braking looks like a solution until the density of trains is considered. In the Tube, the trains essentially accelerate and decelerate in lockstep because they need to be so close together. Using regenerative braking will therefore require large amounts of energy constantly to be stored and re-released. The mechanisms for this are difficult. Large flywheels could be used, but they would need to be located on the surface as there is no space underground.

    Finally, to address temperature spikes in hot weather, freezer trains could be deployed. These would simply be heat exchangers connected to tanks containing cold water or ice slush for higher heat capacity. They would not carry passengers. These trains would aim to trundle through the system without stopping so they generate minimal braking heat. Any heat they do generate could be absorbed directly from their brakes into their own water tanks. When the water in the tanks is hot enough the freezer trains would leave the system for dedicated facilities where the hot water would be discharged and the ice slush replenished. There would be opportunities to use the hot water creatively in the community.

    None of this would require major installations underground which are a big headache in the Tube. Nor would any existing rolling stock need to be updated or modified to make any of this work.

    LUL could get started on this tomorrow, but they have been dragging their feet.

    Fifteen years ago, I wrote this up in a paper and sent it to LUL. They put it through their simulators, but only analysed the freezer train component without considering the complete programme. They were more concerned with the loss of service because a freezer train takes a slot from a passenger train.

    It would only take a few heat stress-related passenger medical emergencies where the whole line is stopped for the freezer trains to justify themselves on this ground. Apart from this cold analysis there is the ethical issue of exposing vulnerable people to real danger.

    I hope that one day that LUL gets to grips with this heat problem before entire lines have to be suspended, or worse, an actual fatality occurs.

  6. Some things come to mind, although I’m unsure how effective each would be:
    – Automation to reduce needless acceleration/deceleration and slow the trains slightly. This may impact throughput depending on how many trains are active on the line at a time.
    – Intermittent Absorption Cycle attached to the trains. The intermittent coolers could ‘charge’ on the surface, and operate while underground… this would not be effective enough to be a single solution, and I would expect these modules would be placed as additional cars to existing trains. This is an expansion on the ‘keep the train cool’ process they’ve carried out with reflective coatings.
    – Fluid Cooling .. With such a small area to work on, this might not be possible.. but laying a network of uninsulated liquid pipe into tunnels, possibly embedded into the walls/floors may assist in pulling excess heat out.

  7. That’s an interesting article and a complex problem, so probably the best solution involves multiple actions: some to reduce the heat generation and some to remove the heat that is created. For example, if the brakes generate 38% of the heat, using another type of brake, like regenerative brakes, would reduce the heat generated by converting some of the train’s kinetic energy into electrical energy instead of converting it all into heat.

    This, combined with more ventilation and heat pumps, could gradually help to solve the problem.

  8. TfL

    The fart Line? Thin film Laugh? Tube florescence Light?
    Any other random combination?

    I read the entire HaD article – maybe the whole term is in there somewhere but AFAIKT never near “TfL”…. what gives?

    1. Having all of mankind’s knowledge literally at your fingertips means that you no longer have to ask stupid questions on the internet.

      If you google “tfl” the first hit is actually the correct answer. Even if it wasn’t you could quickly find an answer related to the article.

      “The fart line” and your other suggestions never appear in the search, therefore they are unlikely to be correct.

      Bonus: searching for additional information takes less time than writing a comment saying you didn’t know what something meant and that you were incapable of finding out.

  9. Why not geothermal? How far down do they need to drill to get to an aquifer that can sink all that heat and more? A long term goal of co generation wouldn’t be the worst idea either. Have the warmer water feed into a secondary loop and property owners can tap into it for a reduction of their costs to heat their home water system.

  10. Clad the outsides of the trains with radiators that harvest heat out of the air, store that heat in a portable molten salt storage, and release the heat in the atmosphere once the train is back on the surface?

    Or if the train never goes to the surface, maybe it’s possible to exchange the molten salt storage for an ’empty’ new one at certain stations.

    Heat can be harvested and stored, and the train can be the medium that transports it.

    Of course I don’t know much about how to do all this, but to me there seems to be a possibility that it can work.

  11. Since there are probably working drains in the tunnels, why not attach a crate of ice to each carriage? It would take a long time to cool the tunnels, and the efficiency of producing the ice would need to be worked on, but gradually the temperature would drop.

    The ice could be refilled at a central point, and the drains would carry away the waste water at any location.

    If dripping ice cages were not a good thing then they could be containers. The ice would still melt, but the containers could be drained at a controlled location.

    If they clip on and off then the emptying/refilling process could be quick. They could be shaped to fit each class of carriage and tunnel. If a train goes overground then don’t attach ice. Leave the cooling function to the other trains.

    Note: don’t use dry ice.

    1. Good idea overall but a couple of minor details if I may:
      1) Given the crapton of calories water absorbs to even raise by 1 degree it’d be smart to leave water to slosh around until it gets to 20ºC or so that it can keep cooling.
      2) But work the maths of how much ice would be needed to work out the problem in a reasonable time-frame accounting for the added energy to haul ice per carriage plus the inevitable oxidation and leak issues and the economies of installing several ice-making facilities to exchange the water for ice AND making ice-cooled carriages. Maybe it won’t work without a substantial increase in fares therefore making it unpopular with users.

  12. Barring cassette type heat pumps, my money’s on chilled water radiators which are very common in the UK for heating anyway so it’s got to be a mature technology. Should only need cooling in summertime and basically just water pumps needed to work in winter.

    Here, lord Mayor, send me my hundred thousand quid.

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.