Train wheels friction heating track?

[contrex]

There was, just now, a professor of sustainable transport was speaking on Channel 4 News and she said that train companies have to be mindful that trains heat the rails by friction. I've never heard of that. Is it a thing? If so, is ir running or braking or a mixture? I have heard of magnetic brakes heating rails.

 

[najaB]

There was, just now, a professor of sustainable transport was speaking on Channel 4 News and she said that train companies have to be mindful that trains heat the rails by friction. I've never heard of that. Is it a thing? If so, is ir running or braking or a mixture? I have heard of magnetic brakes heating rails.

Well from first principles must be friction between wheel and rail or the train wouldn't be able to accelerate (positive or negative). However one of the defining characteristics of railways is that steel on steel rolling friction is incredibly low.

So I'd say that she was technically correct. Which, as everyone knows, is the best kind of correct.

 

[contrex]

Well from first principles must be friction between wheel and rail or the train wouldn't be able to accelerate (positive or negative). However one of the defining characteristics of railways is that steel on steel rolling friction is incredibly low.

So I'd say that she was technically correct. Which, as everyone knows, is the best kind of correct.

Agreed! But is it really something that operators must consider during very hot weather, as a factor in rails buckling? On a slightly related topic, I once saw a Roger Ford article in MR which mentioned the anti-wheelslip system of the Class 60 locomotives; he said the control was so fine when starting a heavy train that he could hear 'creep scream' from the wheel/rail interface. I haven't been able to find any references online about that.

 

[edwin_m]

If we're into technical correctness, heat can only appear when work is being done, and work is force multiplied by distance in the direction of the force.

So there is simple friction between the wheel and the rail but it only generates heat if the wheel is sliding, which obviously it normally isn't.

They may be talking about the rolling resistance effect, which arises because the wheel and the rail are slightly compressed at the contact patch and then spring back as the wheel travels onwards. That depression is a pretty big force over an extremely small distance, and I have absolutely no idea how much heat it would generate. It's the same reason road tyres get hot, but with metal being less deformable it's a much smaller effect on rails, which is the main reason trains are so energy-efficient.

Heat in the wheel, particularly from tread brakes, may also be transferred to the rail, and eddy current or magnetic brakes will heat the rail directly but we don't have those on the UK main line.

 

[devon_belle]

I guess because the wheels are conical and the contact patch is not infinitesimal that will result in parts of the area sliding forwards/backwards relative to the rail on curves, increasing friction. Wheels/rails will also be very slightly deformed on contact. High impact areas such as pointwork may also be subject to heating due to the forces when the wheel impacts them? I imagine all these sources of heat are tiny and very quickly dissipated throughout the rail cross section.

Would tread-braked wheels transfer some heat to the rails during/after braking? Does heat due to rolling resistance pass through the wheels at all?

Agree with the above, technically correct indeed!

 

[LYradial]

Whatever force is applied to the rail by the wheel must be met by an equal and opposite force from rail to wheel so the wheels must get very hot,

 

[Richard Scott]

Whatever force is applied to the rail by the wheel must be met by an equal and opposite force from rail to wheel so the wheels must get very hot,

The existence of a force does not imply massive heat generated, as pointed out above when work is done changing one form of energy into another then heat will be dissipated but may not be in huge amounts.
Assume you're implying Newton's 3rd Law? If you walk down the street then it applies but you're not generating huge amounts of heat.
There will be a tiny amount of slip between the wheel and the rail when moving, if there wasn't then the wheel nor rail would ever wear out!

 

[Crithylum]

Assuming a mass of 400 tonnes and a coefficient of rolling resistance of 0.0015, that is roughly 6,000 J/m. With a rail mass of 60 kg/m, that is 50 J/kg (two rails). With a specific heat capacity of roughly 500 J/kg k, that is a temperature increase of 0.1 degrees C every time a 400 tonne train passes over, assuming all the energy goes into the rail.

 

[contrex]

Assuming a mass of 400 tonnes and a coefficient of rolling resistance of 0.0015, that is roughly 6,000 J/m. With a rail mass of 60 kg/m, that is 50 J/kg (two rails). With a specific heat capacity of roughly 500 J/kg k, that is a temperature increase of 0.1 degrees C every time a 400 tonne train passes over, assuming all the energy goes into the rail.

These are the sort of ballpark figures I was thinking.

 

[Chris Butler]

I think it was pretty obvious that she didn't have a clue what she was talking about.

 

[IanXC]

I think it was pretty obvious that she didn't have a clue what she was talking about.

I came here to say this.

For a factual answer to the question see this page:

Hot weather and the railway - Network Rail

On the railway, hot summers can be as challenging as freezing winters. Find out what we do to keep you moving in hot weather.

www.networkrail.co.uk

 

[edwin_m]

I came here to say this.

For a factual answer to the question see this page:

Hot weather and the railway - Network Rail

On the railway, hot summers can be as challenging as freezing winters. Find out what we do to keep you moving in hot weather.

www.networkrail.co.uk

Are there still significant amounts of fixed OLE out there? I thought the last major area was the Great Eastern until replaced by auto-tensioned equipment at least a decade ago.

 

[contrex]

I think it was pretty obvious that she didn't have a clue what she was talking about.

That is absolutely the impression I had.

 

[Nottingham59]

Assuming a mass of 400 tonnes and a coefficient of rolling resistance of 0.0015, that is roughly 6,000 J/m. With a rail mass of 60 kg/m, that is 50 J/kg (two rails). With a specific heat capacity of roughly 500 J/kg k, that is a temperature increase of 0.1 degrees C every time a 400 tonne train passes over, assuming all the energy goes into the rail.

That assumes a train is coasting along horizontal rails. Under braking or acceleration, the heat generated by friction will be higher. And if anti-wheelslip is in operation, the friction power generated will at least be an order of magnitude higher.

A locomotive hauling a fully loaded 1800t intermodal up Shap at full power with the wheelslip protection actively seeking out that very last bit of grip will have its wheel rims constantly sliding with respect to the rail head. I don't know how hot the rails will be after such a train has passed, but I'm sure the heating will be much more than 0.1C

 

[edwin_m]

That assumes a train is coasting along horizontal rails. Under braking or acceleration, the heat generated by friction will be higher. And if anti-wheelslip is in operation, the friction power generated will at least be an order of magnitude higher.

A locomotive hauling a fully loaded 1800t intermodal up Shap at full power with the wheelslip protection actively seeking out that very last bit of grip will have its wheel rims constantly sliding with respect to the rail head. I don't know how hot the rails will be after such a train has passed, but I'm sure the heating will be much more than 0.1C

If creep control is active or the wheel is spinning or sliding, then work is being done in the direction of the friction force, so yes heat will be generated in the wheel and rail.

If neither of these applies then wheel-rail friction does not generate heat, because there is no work being done in the direction of the friction force (which is along the direction of movement). Heat is generated by rolling resistance as discussed in earlier posts.

 

[8ace]

Won't some of the energy created by friction be dissipated as sound energy too?

 

[Chris Butler]

I've asked her. She confirms that that's not her expertise. She actually is, and was introduced by C4 as, a professor of sustainable and resilient CITIES, not transport. She wasn't expecting the question and got interrupted in her answer.

Regardless of that, I don't think the comments above about the frictional forces being along the rail and so no work being done are correct.

They are conceptually correct for idealised/theoretical rigid smooth bodies, but in that case there wouldn't be any friction. Friction is caused by either/both:-

1. the fact that at the microscopic level neither the rail nor the wheel are smooth, and
2. constant making and breaking of electromagnetic bonds between rail and wheel,

both of which result in forces which disturb both surfaces at a molecular level causing the atoms to move quicker and so increase temerature.

More practically, we all know that friction causes heat. For example the friction from wheel slip can melt rails.

Finally, a coasting train will be creating negligible amounts of friction, but it will be compressing the rail and the resulting (plastic) deformation of the rail will generate some heat even if friction is ignored. How much, I have no idea.

 

[al78]

There will be some heat generated through friction but I'd be surprised if it was significant compared to the effect of the ground temperature from solar heating and direct sunlight. It should be possible to do a back of the envelope calculation using an idealised model of a train to come up with some orders of magnitude estimate.

== Doublepost prevention - post automatically merged: 24 Jun 2026 ==

I put this to ChatGPT, it did some back of the envelope calculations comparing an estimate order of magnitude of energy transfer to a 1 meter section of rails by friction under a train moving at a constant speed, with the amount of energy transferred through direct sunlight during a hot summer day. It assumed a 400 tonne passenger train moving at 30 m/s, a wheel on rail rolling resistance coefficient of 0.001, a specific heat capacity for steel rail as 500 J/kg/K and a rail weight of 60kg/metre.

It came up with 4 KJ per metre for the train passage.

Under strong summer sun with a solar input of 800-1000 W/sqm, two rails of around 0.3-0.4 m^2 per meter of rail thus around 300 W/meter of rail absorbed for the pair of rails, over one hour that equates to a little over 1000KJ per meter per hour. That is why rail temperature can be 20C or more than the air temperature.

Even with a train every few minutes, frictional heat is very small compared to solar heating.

It does say that in specific circumstances, friction generated heat is important, examples being heavy braking, tight curves and freight traffic.

 

[edwin_m]

Regardless of that, I don't think the comments above about the frictional forces being along the rail and so no work being done are correct.

They are conceptually correct for idealised/theoretical rigid smooth bodies, but in that case there wouldn't be any friction. Friction is caused by either/both:-

1. the fact that at the microscopic level neither the rail nor the wheel are smooth, and
2. constant making and breaking of electromagnetic bonds between rail and wheel,

both of which result in forces which disturb both surfaces at a molecular level causing the atoms to move quicker and so increase temerature.

More practically, we all know that friction causes heat. For example the friction from wheel slip can melt rails.

Finally, a coasting train will be creating negligible amounts of friction, but it will be compressing the rail and the resulting (plastic) deformation of the rail will generate some heat even if friction is ignored. How much, I have no idea.

It's precisely the microscopic roughness of the surface that causes the frictional force that allows the train to accelerate and brake. But this "simple" friction only starts generating heat if there is sliding involved. As the wheel turns, its surface goes down onto the rail and lifts off again, both of which are near enough vertical so the microscopic peaks and troughs do not slide against each other.

There is indeed a compressing effect, which has already been discussed and quantified earlier in the thread.

 

[Chris Butler]

It's precisely the microscopic roughness of the surface that causes the frictional force that allows the train to accelerate and brake. But this "simple" friction only starts generating heat if there is sliding involved. As the wheel turns, its surface goes down onto the rail and lifts off again, both of which are near enough vertical so the microscopic peaks and troughs do not slide against each other.

There is indeed a compressing effect, which has already been discussed and quantified earlier in the thread.

I don't think that's right. However, nothing below makes any difference to the conclusion of this thread, namely that the heating effect of friction in normal operation is tiny compared to the thermal gain.

There are three things:-

1. When two uneven surfaces contact each other vertically some contact points (like the one in red in the picture) are, as you say, impacted only by a vertical force. Many more (like the one in blue) are impacted by forces that are not vertical.
2. The wheels have flanges. The flanges are always sliding, even when the train is not sliding.
3. The bodies involved are not rigid, they compress, they slide, they fracture and they shear. They may even melt. The effect is that even if they are impacted by a vertical force externally, the forces within the materials impacted (including rail, sleeper and ballast) are not all vertical. Work is done within the materials involved and heat (and sound) generated. For example, occasionally the result of a train wheel pressing down vertically on the track is a piece of ballast flying horizontally out of the track formation.

Two other minor points.

• The rolling resistance discussed above is the sum of all the effects, not just compression/deformation.
• Sliding does not require braking or acceleration. Air resistance, friction and the gradient of the track need to be overcome and so involve sliding. It is a coasting train that involves no sliding.

 

[edwin_m]

When two uneven surfaces contact each other vertically some contact points (like the one in red in the picture) are, as you say, impacted only by a vertical force. Many more (like the one in blue) are impacted by forces that are not vertical.

Agreed. However, if there is no sliding, the horizontal forces do no work, because work is force multiplied by distance in the direction of the force. So no energy is lost to friction. If you lean against an object that doesn't move, you are not doing any work. You do work if you push it along.

The wheels have flanges. The flanges are always sliding, even when the train is not sliding.

The flanges are not in contact with the rail except on relatively tight curves or if the vehicle's dymamic behaviour is less than optimal. Otherwise the wheelset adjusts itself side to side so that the increased radius nearer to the flange is in contact on the outer rail of a curve, so that both wheels are in rolling contact with no sliding involved at least in the direction parallel to the rail. I guess the actual lateral movement of the wheelset may involve some work being done.

The bodies involved are not rigid, they compress, they slide, they fracture and they shear. They may even melt. The effect is that even if they are impacted by a vertical force externally, the forces within the materials impacted (including rail, sleeper and ballast) are not all vertical. Work is done within the materials involved and heat (and sound) generated. For example, occasionally the result of a train wheel pressing down vertically on the track is a piece of ballast flying horizontally out of the track formation.

I agree with this too. These forces have been discussed upthread, but the heat arises from deformation of the wheel and rail not from friction - similar to repeatedly compressing a spring which will get hot.

A piece of ballast flying horizontally must be the result of the rail contacting it at an angle so that part of the weight force of a passing train passes through that interface at right angles to the plane of contact (think of it as a wedge). The horizontal component of that force may then accelerate the ballast horizontally, and is balanced by a horizontal reactive force through the rail and into the fastenings. Work is done by the rail moving downwards towards the stone.

 

[cornishjohn]

That is a temperature increase of 0.1 degrees C

I'm surprised it is that much actually. In an elavated temperature of (say) 5 degC, that is 2%!

Won't some of the energy created by friction be dissipated as sound energy too?

This is even more technically correct than the temperature assertion. I can find a moderately respectable web source that claims the noise energy generated by a "subway car" is of the order of 0.01 W/m^2, although it is not clear what this relates to. How much of the noise in a typical train I wonder is from engine/motor and wind turbulence.

 

[Crithylum]

I'm surprised it is that much actually. In an elavated temperature of (say) 5 degC, that is 2%!

For a rail at 50 C, it is 0.03% in absolute terms (based off absolute temperature)

 

[eoff]

Wow, unless my web browser search is not working, we got through 23 posts without any mention of static and dynamic friction.

 

[ac6000cw]

Agreed! But is it really something that operators must consider during very hot weather, as a factor in rails buckling? On a slightly related topic, I once saw a Roger Ford article in MR which mentioned the anti-wheelslip system of the Class 60 locomotives; he said the control was so fine when starting a heavy train that he could hear 'creep scream' from the wheel/rail interface. I haven't been able to find any references online about that.

If you haven't heard controlled wheelslip/creep in action (albeit on a different continent and locos with 30+ tonne axle loads), a few examples from my videos:

Wet rails, climbing a 1.4% gradient in the Appalachian mountains, head end locos with DC traction drive at 04:13 and, for comparison, followed by head end locos with AC traction drive at 06:34 (the tail end locos are AC drive in both cases, but aren't noticeably in 'creep' as the rails have been cleaned by the passage of several hundred wheels in between):

Dry rails, climbing a 1% gradient past Lake Louise, tail end loco (out of 4 distributed along the train) with AC traction drive:

BTW, if you think it sounds loud in the video, in reality it can be teeth-on-edge, fingers-in-ears loud when you're trackside...

 

[edwin_m]

Wow, unless my web browser search is not working, we got through 23 posts without any mention of static and dynamic friction.

Implied by mention of wheel creep I think. A small amount of relative movement increases the maximum available frictional force, but it rapidly falls off as relative speed increases. Hence creep control systems need to manage the wheel rotation very precisely.

 

[Lucan]

Former wheel/rail interface specialist here. I worked at the Railway Technical Centre at Derby while the APT was being developed.

While I have not heard the actual words that the "professor of sustainable transport" said, the idea that we should be mindful that trains heat the rails by friction is complete and utter BS. Maybe they think that trains brake by skidding to a halt with locked wheels, or that brake blocks act on the rail. Even then, it might only be a problem on the LU at Oxford Circus in the rush hour.