Why the London Underground is so hot

[IvanV]

A HuffPost article came into my feed talking about the excessive temperatures in the Tube in recent days. Various stations in central London are currently 33C. The article claims that the main source of the heat is train braking, which turns the kinetic energy of the tube train into heat. A commonly mentioned potential source is all the people in there, radiating heat from their bodies. But apparently some calculation easily shows taht train kinetic energy greatly exceeds warm bodies as a heat source. If braking is the main problem, then implementing regenerative braking - increasingly common on trains - should make a big difference.

When the deep tube lines were new, I think they were about 11C down in the deep lines. That's the typical ground temperature at that depth in the London clay, remote from heat-producing operations. But that London clay is very heat retentive. And so it has heated up over the decades, as a result of emission of heat from the tube operations. And that heat moves away very slowly, and so accumulates over time as operations continue. So it is much hotter down in the tube than when I was a kid. There is no easy way of extracting that heat and shedding it, except in a few fortunately located stations that can easily retrofit a solution, or new stations which have heat extraction designed in.

A few years ago I read an article by a retired engineer called Calvin Burrows. He says the idea that brake it is the main problem is demonstrably wrong. And the proof lies in the fact that the tube is so much hotter in the summer than the winter. The seasonal variation is about 8C. If the heat is so thoroughly locked in down there, and tube braking the main source of it, it would remain much more of an even temperature over the seasons. For that summer heat is not conducting down to the tube from the streets. Or at the least, the presenters of the braking theory need to to explain these seasonal temperature variations, which it seems they have not. And it is the high temperatures in the summer that are the problem. So it is more important to address the source of the seasonal variation than the year to year gain. At least for the moment. Some lines are can be about 24C even in midwinter, and as that increases even the winter temperature will be too much.

Burrows points out that much heat must be entering the tube in the summer from some source, and then leaving it in the winter, to create that 8C seasonal variation. And he says that seasonal source of heat is solar gain from when trains are out on overground sections. The trains get hot when out in the sun. Then they take that heat into the tunnels and radiate it out when in there. And in the winter cold trains enter the tunnels and cool it down.

Whilst something like this seems evident - at least to me - as a source of the seasonal variation, does this explain the long term heating over the years? Barrow does not address what is the net balance of that solar gain/atmospheric cooling from moving hot or cool heavy trains in and out of the tunnels over the course of the year. Maybe this solar gain process just explains the seasonal cycle in temperatures, but not the long term gain. Maybe the braking issue is the large reason for the long term trend over the years.

But as Burrows points out, whatever is the average temperature over the year, the 8C higher temperatures is always going to be a very big problem. So addressing the solar gain problem is necessary, whatever other heat sources also need addressing. Mechanisms to address it include use of reflective coatings on trains, and shading of overground lines - which is apparently often implemented in other country. But leaves on the line is a problem and TfL goes around cutting trees down, rather than encouraging them to shade the lines.

There's a variety of articles setting out Barrow's theory, I think this one from Rail Business Daily is the most detailed.

Any thoughts on this?

[nekomatic]

Any Tube user is familiar with the air currents caused by trains moving along narrow tunnels, which must have the effect of mixing outside air into the air underground over reasonably short timescales - in fact if that mixing didn’t happen then the air underground would become unbreathable. Presumably there’s forced air ventilation in places as well, rather than relying purely on train-driven air exchange. That seems a sufficient cause of the seasonal variation?

Edit: I see the article dismisses that as ‘probably not significant’, and the author is a clever engineer, but I’d still like to see the calculation by which he decided to dismiss it.

[bob sterman]

Burrows points out that much heat must be entering the tube in the summer from some source, and then leaving it in the winter, to create that 8C seasonal variation. And he says that seasonal source of heat is solar gain from when trains are out on overground sections. The trains get hot when out in the sun. Then they take that heat into the tunnels and radiate it out when in there. And in the winter cold trains enter the tunnels and cool it down.

Nice idea but TFL data showed that the hottest deep line in summer 2024 was the Victoria Line. And Victoria Line services are entirely undergound. Trains only go above ground when going to and from the depot.

[dyqik]

Burrows points out that much heat must be entering the tube in the summer from some source, and then leaving it in the winter, to create that 8C seasonal variation. And he says that seasonal source of heat is solar gain from when trains are out on overground sections. The trains get hot when out in the sun. Then they take that heat into the tunnels and radiate it out when in there. And in the winter cold trains enter the tunnels and cool it down.

Nice idea but TFL data showed that the hottest deep line in summer 2024 was the Victoria Line. And Victoria Line services are entirely undergound. Trains only go above ground when going to and from the depot.

Any line by line comparisons is going to have to take a chunk more information into account to disentangle the various causes proposed above. At least the following will need to be controlled for:

Station spacing
Station ventilation
Ridership
Average wait times of passengers on platforms
Train frequency
Train speed and braking energy
Number of trains entering and leaving the underground sections
Temperature of trains entering and leaving the underground sections

[dyqik]

In terms of it getting warmer over time, it wouldn't surprise me if some heat mitigation practices had been reduced over time - e.g. leaving doors to emergency staircases open when it's hot below to increase air circulation. Possibly these were never designed, but used informally.

Fire mitigation in the wake of Kings Cross, and security, are two possible things that might have changed practices.

Another obvious thing that would make it get warmer over time is increased ridership and train frequencies, as well as extended operating hours.

[bob sterman]

Burrows points out that much heat must be entering the tube in the summer from some source, and then leaving it in the winter, to create that 8C seasonal variation. And he says that seasonal source of heat is solar gain from when trains are out on overground sections. The trains get hot when out in the sun. Then they take that heat into the tunnels and radiate it out when in there. And in the winter cold trains enter the tunnels and cool it down.

Nice idea but TFL data showed that the hottest deep line in summer 2024 was the Victoria Line. And Victoria Line services are entirely undergound. Trains only go above ground when going to and from the depot.

Any line by line comparisons is going to have to take a chunk more information into account to disentangle the various causes proposed above. At least the following will need to be controlled for:

Station spacing
Station ventilation
Ridership
Average wait times of passengers on platforms
Train frequency
Train speed and braking energy
Number of trains entering and leaving the underground sections
Temperature of trains entering and leaving the underground sections

Indeed that is true.

But it's difficult to see how a model incorporating all of these factors is going to throw out "time spent with carriages getting heated by sun while over ground" as the most important positive predictor of heat underground - when the warmest line is the one where the trains don't go overground during their normal working day.

In fact - as some comments from people at TfL have suggested - quite the opposite may be the case. Time spent by trains over ground may help cool the line because carriages can dissipate heat when overground. This may be the reason the Victoria Line is warmest.

[IvanV]

But it's difficult to see how a model incorporating all of these factors is going to throw out "time spent with carriages getting heated by sun while over ground" as the most important positive predictor of heat underground - when the warmest line is the one where the trains don't go overground during their normal working day.

In fact - as some comments from people at TfL have suggested - quite the opposite may be the case. Time spent by trains over ground may help cool the line because carriages can dissipate heat when overground. This may be the reason the Victoria Line is warmest.

Crispian Jago (youtube 12mins) has just done a video on the subject. He mentions solar gain in passing. He notes that they tried putting reflective coatings on Central Line trains, but it didn't help. The sun still won, he said. It implies that solar gain is evidently an issue. But, as you say, there is also the dissipation effect at cooler times - which includes cooler times of day in the summer - and, as you say, that may be why the Victoria Line is the warmest. But we still need to understand the seasonal temperature differences in the Victoria Line.

The Jago video observes that the tube engineers have been trying to ventilate the tube in various ways for at least a century. And as opportunity has presented itself over time - disused stations, abandoned lift shafts, convenient sites being purchased, the number of ventilation shafts have increased over the years. But the piston effect - trains pushing air in front of them and pulling it behind them - is so strong that it reduces the effect of ventilation, by pushing or pulling the air against the intended ventilation. But it is not without effect. So perhaps the reason that even the Victoria Line heats up in the summer is because a lot of air is brought in from outside to try and cool the tube, and that air is a lot warmer in summer.

The latest experiment to try and cool things down, tested at a disused platform at Holborn, is to install water chilling behind the decor panels on platforms. Apparently quite effective, but very expensive to install and doubtless operate.

Aircon is generally considered implausible for deep tube trains, because of the heat output by the air conditioners into the tunnels. But apparently the new Piccadilly Line trains have something like aircon, and Jago couldn't quite explain it. I looked it up. Apparently you can devise an "air cooling" system which avoids the humidity reduction of standard aircon, and which thus outputs less heat. So the Picc line trains still output additional heat to the surrounding tunnel air when the chilling is operating, but less heat than a standard aircon would. And that is apparently more acceptable.

[Grumble]

Surely you could attach a heat battery to trains and dump the heat into them, then swap them out? There are various domestic ones available, some of which are very compact relying on phase changes to store or release heat. Dump the heat or change the battery out at the depot at night.

[bob sterman]

Surely you could attach a heat battery to trains and dump the heat into them, then swap them out? There are various domestic ones available, some of which are very compact relying on phase changes to store or release heat. Dump the heat or change the battery out at the depot at night.

Replace Oyster Cards with Oyster Bricks. Everyone carries a cold brick onto the tube. It absorbs heat - and then when they get off the heat gets dissippated above ground.

Obviously need to calculate whether there's a net benefit given the increased body heat generated by people lugging bricks around.

[IvanV]

Surely you could attach a heat battery to trains and dump the heat into them, then swap them out? There are various domestic ones available, some of which are very compact relying on phase changes to store or release heat. Dump the heat or change the battery out at the depot at night.

You need some kind of a heat pump to transfer heat into your heat store. You can organise an air conditioner to dump its waste heat into a heat store. But as it requires the waste heat to be organised to be at a higher temperature than is usual, it reduces the efficiency of the air conditioning, and there will still be some lower grade waste heat from it that you can't easily store. Whether that is all worthwhile doing on a train also requires assessing. Trains have limited space for storing bulky equipment. And air conditioning equipment is already sufficiently heavy in relation to trains that it is a material factor in the design of the train and its power draw when accelerating.

And this is only going to be marginal in terms of keeping the temperature of the underground stations down, if in fact much of the heat is coming from train braking as suggested in the HuffPost article I quoted at the start of this.

What you really want is ground source heat pumps taking heat from the clay surrounding the stations, and using it to heat nearby buildings. You'd only run that in the cooler times of year when you have buildings that require heating in the area. But these things happen slowly and over long periods of time, so that would be OK. Iirc from the Crisipan Jago video, I think they managed to actually do something like that around one station, which was a more recent new-build, sending the collected heat to a nearby estate. But retrofitting elsewhere it would be horribly expensive.