Facebook Page
TGVweb - TGV Research Overview

 TGV Research Overview

TGVweb > TGV Research > Overview

Serve this page from: California, USA / Pisa, Italy

This is a summary of the research and development currently going into the TGV program. Much of it revolves around the TGV NG (Nouvelle Génération) initiative, the next step in French high speed train technology. The new research was initiated by the high speed test runs of 1989 and 1990, which showed that it would be worthwile to explore the possibility of running trains at speeds between 350 km/h (218 mph) and 400 km/h (249 mph). The official project, led by SNCF and GEC-Alsthom (now ALSTOM, the main contractor for the TGV), was kicked off on 31 May 1990, just two weeks after the world record run.

The cost of the first five years of the program was 535 million French francs (about 100 million US dollars), of which 55% was provided by GEC-Alsthom (now ALSTOM), 28% by the French government, and 17% by SNCF. The aim of the program was to have a prototype TGV NG power car on the rails by the year 2000; it was to be built as part of the TGV Duplex build, in the form of a spare power car. ALSTOM suspended development on the TGV NG in 1999, concentrating instead on a new EMU trainset design with distributed power rather than dedicated power cars. This train is expected to incorporate the research that went into the TGV NG and will have the same operating speed of 360 km/h.

This and other research projects (tilt TGVs and short TGVs) aim to remove several key technological and operational barriers to higher speeds and especially to lower operational costs. Some key points are detailed below.

Skip directly to any of these topics:

...Or see photos of some research activities, past and present.


Current TGV trainsets use three brake systems: disks, dynamic brakes on powered axles, and in some cases tread brakes for emergencies. For speeds of 350 km/h (218 mph) and more, these conventional brakes lose their ability to stop a train in a reasonable distance and cannot perform a safe emergency stop. The basic problem is simple: kinetic energy, which must be dissipated as heat in the braking system, grows as the square of speed. Braking is therefore a major technological hurdle, which must be overcome to safely attain higher speeds in revenue service.

Some preliminary ideas are applied to the TGV Duplex, the bilevel TGV introduced at the end of 1995. These constitute a first step towards higher performance braking. For this third generation TGV, dynamic braking performance is improved from 24 kN (5400 lb) per powered truck to 30 kN (6800 lb), as measured in retentive effort at the rail. Brake disks have added thermal capacity, and the tread brakes on the power cars are replaced by disks located directly on the wheels. Only one brake disk is used on these powered axles, because two of them (one on each side) add weight without increasing braking capacity. This is indicative of how close these brakes are to the limits of the wheel/rail interface. Overall, these changes do not represent major technological innovations, but push the envelope of existing technologies.

For the TGV Nouvelle Génération, an entirely new brake system is needed. Adhesion at high speeds is insufficient to perform quick stops using wheels alone. The wheels tend to skid, and the brakes can overheat and wear quickly, increasing maintenance costs. Therefore, it is necessary to use a system that bypasses the wheel/rail interface. This system, the magnetic induction brake, is based on existing magnetic brake technology which was originally explored using the Zébulon test vehicle (see history). Magnetic induction brakes dissipate kinetic energy of the train as heat in the rail, by way of induced eddy currents. They are only effective above about 220 km/h (137 mph), because they put so much power to the rail that thermal damage to the railhead would result at lower speeds. There are several technical concerns with magnetic induction brakes. Successive brake applications by several trains over the same stretch of track could overheat the rail, resulting in operational restrictions. In addition, they apply upward forces on the track, which is not designed for such loads.

A prototype of a magnetic induction brake was built and underwent testing on a TGV Réseau trainset; it provided up to 16% of the braking effort. In stowed position, the brake shoe (mounted under the power car's truck sideframes) rides 10 cm (4 in) above the rail; when applied, it skims a few cm above the rail without actually touching it. Currents generated by the traction motors create a magnetic field, and the motion of the train causes circular currents to flow inside the rail. These currents produce a retarding force on the train, and are turned into heat by the internal resistance of the rail. (This is Faraday's law in action!)

The disk brakes are also in need of supplemental thermal capacity to sustain longer applications; this can be achieved through the use of new materials. A carbon disk/carbon pad architecture is under consideration, although such systems experience large variations in effectiveness and wear, in addition to being prone to oxydation problems. Another design is inspired by the disk brakes of airliners, using a rotor disk sandwiched between two stator disks by hydraulic pressure. This runs into problems of braking effectiveness as a function of speed, requiring sophisticated active control of another league than the simple antilock system used today. It also poses problems for maintenance, requiring disassembly for inspection. Yet another direction of research is a high-pressure carbon/carbon brake, using contact pressures an order of magnitude higher than in other designs. This technology is touted to increase the thermal capacity of each disk from the current 18.5 MJ to 45 MJ, while decreasing the weight of each truck by 500 kg (1100 lb). Finally, the "furthest out" designs call on ceramic/ceramic brake systems, which are quite far from being applicable in service.

The dynamic brakes for the TGV NG are once again beefed up to 13 MW (17,000 hp) per trainset, a requirement which must be maintained even in the absence of another load on the catenary. (Regenerative brakes, which feed power back into the catenary, can only do so if that power is being drawn somewhere else; they do not work when the overhead power is out.) The resistor grids and fans must thus be able to handle the increased thermal load.

Traction Equipment

Traction for the TGV NG is to come from asychronous 3-phase AC induction motors. This choice is motivated by simplicity of maintenance, very high power to weight ratio (around 1 kW/kg), and higher RPM limit. The 2 x 6000 kW traction package, 40% more powerful than 2nd generation units, requires that a total of 6 powered trucks be used instead of the usual 4, so that sufficient tractive effort can be put to the rail despite the low (17 metric tons) axle load. The extra two powered trucks are located under the first and last trailer, immediately adjacent to the power cars, as used for the TGV Sud-Est or Eurostar. The trainset is designed to be capable of starting on a 4% (40 in 1000) grade with two traction motors shut down. Cooling of the traction equipment must remain effective at ambient temperatures up to 45 degrees C (113 degrees F), using an environmentally friendly liquid coolant (FC72 by 3M) for the semiconductors rather than freon. Traction motors are individually controlled, whereas they were previously controlled in pairs using one inverter. Reactive currents generated by a trainset must be under 1 amp, to avoid perturbing the cab signal and communications channels in the rails; this is in comparison to the 15 amps of a TGV Atlantique trainset. Overall, there is nothing radically new in the traction equipment; the big challenge lies with satisfying a very large power and tractive effort requirement, within drastic axle load constraints.

Cutting Down Weight

Axle loads become a critical constraint at high speeds, in order for maintenance costs of track and train to be reasonable. The current axle load limit is 17 metric tons, and there is a possibility that this will be reduced to 16 tons for the faster TGV NG. Meanwhile, it is desirable to retain an articulated design (with two axles per trailer), a bilevel seating arrangement, and a host of other requirements that would tend to increase the axle load of a trainset. It turns out that keeping weight down is one of the biggest challenges for the design of the TGV NG or a potential tilting TGV.

The first approach to reducing weight is new materials. For the TGV Duplex car bodies, aluminum is used. The trailers are a monocoque design assembled out of extrusions, yielding a weight reduction of 20% over an equivalent steel structure. The frame of the power cars is made of high tensile strength steel, as in the TGV Atlantique units, for a weight reduction of 10% over lower grade steel. Stainless steel could also make an entry into the new trainsets, as well as composites. Composite materials are not used on the TGV Duplex's main structural components for reasons of cost, and also because the technology was not deemed sufficiently mature. Future TGV generations, however, could be built with a composite main structure assembled with glue. There is a research effort to explore the resistance of composite materials to the wear and tear encountered over 30 years of high-speed operation. Other weight reductions are achieved by using better paints, electrical wires with thinner insulation, and many other small measures that become significant when added together.

The second way to cut weight is by using the least possible material to fulfill structural requirements, or optimization. This has become a worthwile pursuit with the advent of extensive computer finite-element analysis. The TGV Duplex is the first to really benifit from these techniques, and the TGV NG will follow suit. Several areas yield weight improvements. The connection between trailers has been completely redesigned, and is now attached to the trailer bodies rather than the truck and suspension assembly. This allows a substantial reduction in the weight of the secondary suspension, with a 400 kg (880 lb) savings. The Y237B truck used on today's TGV Atlantique has been redesigned, saving 200 kg (440 lb). An aluminum version was tried, but is not ready to be integrated into the TGV NG. The interior, as with commercial transport airplanes, is designed to be feather weight. Seats have been entirely redesigned for the TGV Duplex, with each one going from 26 kg (57 lb) to 14 kg (31 lb), yielding a full 1000 kg (2200 lb) reduction for each trailer.


Noise is another major concern of high speed rail technology, not only for the passengers but also for those living near high speed tracks. Research focusses on identifying the sources of noise at speeds in excess of 350 km/h (218 mph) and finding ways to deal with them. (See acoustic noise measurement.)

The interior noise level of the new TGVs, in particular, suffers from all the weight-reducing measures described above. In the new bilevel trailers, new methods are needed to reduce noise, especially on the lower level, which sits much closer to the track than before. The solution adopted is to isolate the interior from the structure, using flexible blocks as well as a sound-deadening composite laminate.

Outside of the train, aerodynamic noise begins to dominate wheel noise at high speeds. Better aerodynamics not only reduce the emitted noise, but also moderate the energy consumption of the train. Field measurements via laser velocimetry, wind tunnel tests, as well as computer models have been used to investigate the flow characteristics in the vicinity of the carbodies. Much attention centers around the nose profile, and shrouding of the trucks; data is being shared with Japanese researchers who are pursuing similar matters. While the TGV NG power car looks much like a 3rd generation power car, exterior shapes for future designs are bieng further optimized. The fifth generation TGV, currently under the form of the MX100 research project, will have a radically streamlined nose profile. This is because at the speeds envisioned, aerodynamic drag dwarfs every other source of resistance to motion; improving the aerodynamic characteristics of the train therefore has a big payoff.

Truck shrouds can reduce the radiated rolling noise, but have had limited success so far. The reduction is not so great, and the shrouds cause brake overheating and grime or snow accumulation.

Other points of interest are the study of the interaction of the train with nearby obstacles, such as in a tunnel. Special shapes can reduce the strength of pressure waves generated by the train, thus improving passenger comfort. An early result of this research was the profile of Eurostar's nose, which is optimized for running in the Channel tunnel.

In addition to pressure sealing the trainsets, a practice begun with the TGV Réseau, active pressure compensation (not full-fledged pressurization) is being investigated.

Trackside noise reduction has also been a focus, with passive and active solutions. Acoustic walls can be built (an have been built in many places) to shield noise-sensitive areas from the track, yielding reductions of overall train noise on the order of 10 to 15 dB. Active systems using loudspeakers embedded in the walls are also under consideration.


Following recent trends in the auto industry, there is significant effort going into passive security of TGV trainsets. As it stands, the TGV architecture already behaves quite well in accidents, as demonstrated in the December 1993 high speed derailment of a TGV Réseau trainset, and various other encounters with large, intruding objects on the track. A particularly pressing motivation for this research was the aftermath of a crash on 23 September 1988, when TGV Sud-Est trainset 70 rammed into a low-boy trailer stuck on a grade crossing, at a speed of 105 km/h (65 mph). The train did not derail, but the engineer died in the incident. It was found that about 10% of the crash energy had been absorbed in deformations of the first trailer; at higher speeds, serious injury to passengers in this first trailer might have resulted.

Using a new software for railroad crash simulation, called 'Pamcrash', the TGV Duplex structure was optimized on a supercomputer. The trailers have extremely rigid bodies with deformable, energy absorbing crush zones at the ends. The coupling between the power cars and trailers has also been reviewed; it still uses screw-link couplers with buffers, but the buffers have structural fuses built in so that they fold away under crash loads. The car ends have been reviewed to prevent the power car from climbing onto the first trailer. For multiple unit operation, the Scharfenberg couplers in the noses of the trainset have been designed to collapse under heavy loads, so that two power cars coupled nose to nose can make firm contact with each other to prevent telescoping. This is in addition to the energy absorbing ram shield already mounted in the nose of all TGV power cars to defend the cab cubicle.

To determine the effectiveness of these design changes, full size crash tests are conducted with specially instrumented production vehicle shells. For the TGV Duplex program both an end-trailer and a power car were rammed into a barrier at 40 km/h (25 mph), validating most of the computer model predictions.

To Tilt, or Not To Tilt...

With ongoing efforts by the French government to bring costs under control, the policy of heavy investment in new high speed lines is losing momentum. With already three major trunk lines radiating out of Paris, a potentially cheaper way to extend the benifits of high speed rail is to improve the performance of TGV trains when they travel on standard trackage.

One possible way to achieve this is through tilting the trailers to permit higher speeds in curves without causing discomfort to the passengers. It should be noted that this feature provides no advantages on dedicated high speed lines, where curve radii and superelevation are deliberately designed for high speed operation. But even the advantages on standard trackage are small. The cant deficiency permissible on French tracks (considerably higher than in other European countries) already allows all trains (including TGVs) to practice relatively high speeds in curves. The addition of a tilt system does not change the lateral loads applied to the rail in curves, so the major obstacle to higher speeds becomes the axle loading. This will probably prove too difficult to overcome for high speeds without entirely changing the train's architecture to distribute traction along its entire length, such as done in the Italian ETR 460 Pendolino or the German ICE 3 designs. The only definite advantage of a tilt system would then be better comfort for passengers, and it is not clear that this would justify the investment.

Nevertheless, in 1996 the French government and industry (GEC-Alsthom Transport) revived the idea of a tilting TGV, last investigated in the 1970s for the proposed TGV 002 (see history). In a 170MF ($ 25M) research program, TGV Sud-Est trainset 101 was fitted with tilting equipment in early 1998 to begin testing at a maximum speed of 220 km/h. This also allowed the government to scale back its high speed rail plans without ruffling too many political feathers, and provided Alstom with a valuable product for export, to compete with the respective tilt systems of ADtranz, Fiat, and Bombardier on export markets (Alstom has since acquired the rail subsidiary of Fiat).

For more information about the P-01 Tilting TGV Demonstrator, see here. As of 2000 no further development is expected on this project.

Short TGVs

It has been found desirable to increase the operational flexibility of TGV trainsets by making them shorter, with only four trailers instead of eight or ten or more. After running coupled together on a high speed line, the trainsets can separate and continue to different destinations. This can extend the benifits of TGV service to places with insufficient traffic to justify service with full-size trainsets.

The idea is to adapt the TGV Duplex design, yielding a 235 seat trainset with one power car and four trailers. The new element in this design is an articulated driving trailer, which would fulfill the dual function of bar trailer and driving cab. Its nose would be indistinguishable from that of the power car and would contain similar crash protection. The major obstacle to this design is the presence of passengers in the first vehicle of the train, which could compromise safety in the event of a collision. This concern was mitigated with recent advances in rail vehicle crashworthiness, as well as the equivalent German practice.

Last Modified: November 2000

TGVweb > TGV Research > Overview

Back to Top