The construction of underground spaces, including tunnels and railway stations, presents civil and structural engineering challenges that are likely to be familiar to regular readers, as will be the need to provide fire and life safety provisions such as escape walkways, shafts and tunnel ventilation equipment. However, the management of heat is also becoming an increasingly important design consideration.

Heat management is an emerging consideration for several reasons. For existing railways, trends towards larger metropolitan populations and a shift away from road transport puts extra demands on metro and commuter rail. A logical solution to this is to increase train frequency, which can require more trains to allow more capacity and higher train speeds to reduce the amount of time people spend travelling. Although new trains are often more efficient than those they replace, there are some basic laws of physics that will always hold true. The consequence of more numerous and faster trains is that more power is needed to feed the railway. Eventually, all this power ends up as heat which must be managed.

Heat can have several impacts on a railway. Perhaps the most obvious is on thermal comfort where passengers might begin to feel uncomfortable in stations, on platforms and in trains, particularly if the trains are not air conditioned. Even if the trains are air conditioned, tunnels might get hot enough to risk the air conditioning cutting out if the train stops in the tunnel for an extended duration, in which case thermal safety starts to become a consideration1. Other factors, such as the negative impact that heat can have on equipment reliability and worker performance, must be considered. Railways are not the only kind of tunnels impacted by heat generation. Electrical losses in cable tunnels, used by electricity providers to distribute energy around cities, can result in tunnel heating that could compromise the electrical resistance of the insulation within the cables and cause circuit failures, not to mention make the tunnels too hot for maintainers.

For these reasons, heat must be managed. If heat is treated as a pollutant then, like all pollutants, it’s usually best to reduce it at the source. This can be achieved by improving energy efficiency. Whilst railways are a relatively low-energy intensity mode of travel, improvements in energy efficiency can often still be gained. Measures that have been implemented in railways include designing tunnel gradients to minimise braking and acceleration effort (so-called ‘hump track profiles’), designing for high regenerative braking and feeding back regenerated power to the electricity supplier, coasting, recovery time management, optimisation of traction package efficiency and reductions in train mass. If the added cost of removing the heat is accounted for, the financial case for such energy efficiency can be improved.

When all reasonable efforts have been expended on minimising heat input, cooling may still be required. This can take many forms, perhaps the simplest being the use of ventilation, which expels warm air from the railway and replaces it with cooler air. This ventilation can be passive, using shafts that use the train piston effect as the driving force, or active–using fans. Mechanical cooling can also be used, involving the circulation of some form of coolant around a heat exchanger, which exchanges heat with the tunnel or station air. Coolants can be from natural sources, such as groundwater from an aquifer, or artificial sources such as air or water-cooler chillers that use refrigerants.

The tunnel ventilation team at WSP has previously investigated the effect of using Borehole Cooling on London Underground infrastructure2. A saturated chalk aquifer provides a large reservoir of water, maintaining a temperature of around 13°C for most of the year, which can be pumped to a cooling coil within the station. After distributing the water through a network of pipes around the station, the water is re-injected back into the aquifer at around 23°C.

Whatever method of heat removal or cooling is adopted, a logical question concerning sustainability arises: can some of this removed heat be reused? The UK aims to reduce its greenhouse gas emissions from a high of 800Mt CO2 equivalent in 1990 to 350Mt of CO2 equivalent in 2035 (Figure 1), with a large portion of this reduction to come from the widescale adoption of renewable and sustainable energy sources. Could waste heat tunnels be part of the picture in sustainable heating?

How much heat can be recovered and what can it be used for?

Tunnel ventilation engineers generally use one-dimensional modelling techniques to predict and analyse the effects of heat output from trains in tunnels and stations. The size and geometrical complexity of tunnel networks means that 3D airflow analysis would be prohibitively expensive in terms of money and time. One-dimension models allow for a sufficiently detailed output of the predicted tunnel airflow and temperature in a relatively short period of time, reducing the time from concept to construction.

Networks of tunnels, cross passages and stations can be represented in software such as Subway Ventilation Simulation (SVS) by tunnel, cross passage, platform and concourse segments, joined by junction nodes. The segment airflow and temperature can be calculated for a prescribed train timetable or cable heat load, considering the effects of tunnel gradient, cross sectional area and pressure losses due to tunnel geometry. The engineer can then use the model to design ventilation strategies that results in an environment that meets or exceeds the thermal comfort standards required by the project. Hour-by-hour temperature predictions can be made in any given year, depending on the analysis software used.

Tunnel temperatures vary depending on the distance to the portal or end of the tunnel, how much heat is released into the tunnel and what other measures exist such as bypass shafts and forced ventilation shafts. Tunnel temperatures typically vary in an approximately sinusoidal pattern over the course of a year and the following might be expected as representative temperatures:

  • Metro without platform screen doors: between 23°C (winter) and 30°C (summer)
  • Metro with platform screen doors/end of a long rail tunnel/ end of a cable tunnel: between 28°C (winter) and to 35°C (summer)
  • The heat that can be recovered is influenced by the temperatures of the air within the tunnels and the technology that is used to recover it, but it can be expected to be within one or two degrees centigrade of the tunnel air temperature.

Regardless of the predicted temperature, there will be uncertainty in the predictions that might affect the commercial viability of any heat recovery scheme. These uncertainties are not normally the fault of the engineer, but more a result of large uncertainties in input parameters, such as how many trains the railway will operate in practice, ground conditions and the tunnel water ingress rates. Engineers might take conservative assumptions on these inputs so as not to undersize any cooling systems, which may be exceptionally expensive to retrofit, but this could have the side effect of grossly over estimating the viability of any heat recovery scheme. These uncertainties should therefore be examined before assessing any heat recovery scheme.

The most obvious use for the waste heat is domestic or industrial heating. A traditional domestic heating system would require a water temperature of about 75°C in the radiators, which can be reduced to 55°C if the radiators are made larger to suit a condensing boiler. For water generated from waste tunnel heat, at roughly 25°C, the radiators would need to be enormous; even an under-floor heating system would need warmer temperatures. Therefore, for practical usage, this lower grade heat needs to be stepped-up, which is achieved using a heat pump. If, for example, the heat pumps are immersed in 23°C tunnel air in winter, as opposed to an outside ambient temperature of 5°C, then each kilowatt of heating output from the heat pump would need only 0.13kW of input electricity compared to 0.48kW of input electricity at 5°C ambient temperature.

Heat pumps can generate water at up to 80°C, which can then be connected into either a single building’s heating system or a district heating system. Whatever the heat pump connects into, some form of storage will probably be needed, since the heat demand and heat delivery profiles may not always match in any given hour of the day.

How do you recover the heat

It has been estimated that a typical railway tunnel could contain 4,500GJ of heat energy per km of tunnel3, with higher temperatures expected at the exit portal (Figure 2). There are several methods of capturing this heat but they fall into two categories; those that use intermediary heat exchanges and fluids and those that directly expose any heat pump to the tunnel air.

Intermediate heat exchangers are typically used when mechanical cooling of the railway or cable tunnel is needed to make it comfortable and safe. The financial case or need for the heat exchangers is therefore predominantly justified by their importance to the railway or tunnel operations. In such instances extending the functionality to also recover waste heat in winter could be expected to result in a financiallyviable heat recovery scheme.

A simple cooling technology using intermediate heat exchangers was employed on the Channel Tunnel. Hair-pin pipe loops are provided within the tunnels and cool water pumped through these pipes, which then exchange heat with the tunnel air. The warmed water then passes through chillers located at the portals where the water is cooled again and passed back through the pipe loop. Such a technology might also be well-suited to tunnel heat recovery. A disadvantage of such intermediate heat exchanger technologies for heat recovery is that the frictional resistance of very long pipes can be high and the electrical energy used to pump the water around forms a parasitic loss. Whilst all of this frictional loss turns into heat, which is arguably good for heat recovery, it is not a particularly efficient way of generating heat and thus this can erode the energy savings associated with any heat recovery

More recent developments with tunnel heat recovery include adaptation of segmentally-lined tunnels into large heat exchangers. Polyethylene heat exchanger pipe is tied to the rebar reinforcement cage of the precast segment before the concrete is poured. The pipe tails within each segment are joined together at the joints of the segments to form loops of pipe within the tunnel liner. These loops are connected to a main flow and return pipe which creates a very large heat exchange surface with a moderate frictional resistance. The concept is analogous to an under-floor heating system operating in exhaust. Such technologies have been demonstrated in tunnels on a small-to-moderate scale but not, so far, on a major scale.

Typically, however, the intermediate heat exchanger cooling technology would comprise of a simple, lowresistance finned coil, filters and fans provided as part of an air handling unit delivering cooled air to the platforms. Examples of this can be found in railways around the world and it is potentially viable to modify such systems to be suitable for heat recovery in winter, again subject to verifying that any parasitic energy losses associated with the fans, and potential maintenance costs of changing any filters, does not erode the case for any heat recovery.

There are examples where intermediate heat exchangers installed solely for heat recovery may be viable. A simple coil in the path of an exhaust air shaft, for example, may be viable subject to it having low frictional resistance. However, for heat recovery-only systems perhaps the most viable option is the installation of air-source heat exchanges directly into the warm air stream that is expelled from the railway4,5. London Underground have at least one example of such a technique where space was made in the upper plenum of an exhaust ventilation shaft for the heat pumps.

More recently a heat recovery method was proposed by WSP for HS2 at Victoria Road crossover box, which is a proposed 120m-long open crossover between the Old Oak Common and Northolt tunnels6. Warm air leaving the Northolt tunnel would rise from the box and be ingested by air source heat pumps located around the edge of the box at ground level. The heat pumps, being at ground level, would be provided by any developer and be accessible to them without impact on the railway operations.

Understanding the airflows and temperatures leaving the tunnels was important in estimating the performance of the heat recovery system. A coupled 1D-3D model (Figure 3) was developed to determine the amount of heat expected during a 10-year period after the start of railway operations. Data from Subway Ventilation Simulator models for the Old Oak Common and Northolt tunnels were used as boundary condition inputs for a Computational Fluid Dynamics model of Victoria Road Crossover Box and the heat pumps. Seasonal variations in the predicted heat demand and supply were balanced out by using a Thermal Energy Storage system. Thermal Energy Storage systems allow for the shifting of heat demand and production, so that the heat demand and supply do not necessarily have to balance at the same time. Large Thermal Energy Storage stores can allow for inter-seasonal storage of heat; using the energy from the previous summer to heat in the winter.

Environmentally, the effect of using the air source heat pumps to supply heat to the District Heat Network was predicted to have the potential to reduce the CO2 emissions by 22% over one year of operation, compared to supplying the same level of heat with a commercial gas boiler. The financial case for using the air source heat pumps can also be promising. For the example cited, an energy saving of approximately GBP 16,000 (USD 20,000) per year was predicted when supplying a District Heat Network of 500 homes.

Optimising the control logic of the heat pumps, so that they are only active if there is capacity in the Thermal Energy Storage system, yields an increase in the cost saving per year of 28% (saving 125,000 kWh of wasted heat per year).


Waste heat from tunnels is typically low grade and hence challenging to make immediately useful for heating adjacent spaces. However, if the waste heat is stepped-up in grade by means of a heat pump, then it may provide some useful heat at a lower energy intensity than gas boilers or conventional air source heat pumps.

The value of the energy saved from tunnel waste heat may not be high enough to justify significant investments in heat recovery systems. Therefore, the skill in waste heat recovery from tunnels lies in finding ways to recover the heat economically in terms of capital costs, ongoing maintenance costs and energy usage in recovering the heat in the first place.

Where an existing railway or cable tunnel requires a cooling system to be installed then such systems may potentially be reconfigured to allow heat to be recovered in winter, in which case heat recovery might be quite viable. If there are no cooling system then other attractive techniques may be to use commercial ‘off the shelf’ heat pumps installed in warm air that is otherwise released to atmosphere. These might, for example, be located in the plenums of exhaust air ventilation shafts or close to tunnel portals.

Computer modelling of the railway or cable tunnel is important tool in understanding the quantity and annual performance of heat recovery and give an indication of the need for and effectiveness of any associated thermal energy storage.

However, as with all computer modelling, the inherent uncertainty in the calculations must be accounted for in any analysis.