MOST EMUs in Japan now use regenerative braking for normal service applications from maximum speed to a complete stop. But regeneration has inferior characteristics compared to rheostatic braking, and the difference is magnified when the overhead line or conductor rail is not receptive.
To cope with these problems, on-board and wayside energy storage has been keenly debated. Development of devices combining higher power and energy density with acceptable losses and longer life such as lithium-ion or nickel-hydrogen batteries and electric double-layer capacitors promise more efficient train operation in the near future.
Almost all passenger trains in Japan are electric or diesel multiple-units. Today, only sleeping car trains and special excursions are hauled by a locomotive, though EMUs are even used for some sleepers and freight.
Aside from regenerative braking for steep and long down gradients in the 1930s, regenerative braking became popular for normal service braking even in the DC traction era of the late-1960s, when field chopper control was introduced by many private railways. Metros followed in the 1970s when armature chopper control was introduced.
In the 1980s, AC traction motors replaced DC motors for almost all EMUs and regenerative braking capability became inherent in the power control circuits. Except for special circumstances, dynamic braking where energy is dissipated in resistance grids was abandoned because of the additional weight, cost and potential danger of fire.
The exception, as in the old days, was the use of regenerative braking on long down gradients where prolonged use of friction brakes can be dangerous in this case, switched or blended regenerative/rheostatic braking is used.
Two problems to be solved
Ever since AC-motored EMUs with regenerative braking capability were introduced there have been two big problems: occasional regeneration failure due to lack of line receptivity, and inferior braking characteristics at the higher speeds when compared to dynamic braking. These problems are now being addressed by introducing on-board and/or wayside energy storage devices.
Another minor problem relates to the changeover from regenerative to friction braking just before the train actually stops. This has already been solved, and since the 1990s all EMUs introduced throughout Japan have been able to use regeneration to bring the wheels to a complete stop. Friction brakes are only used as a back-up, or for preventing a stationary train from rolling away.
Once on-board energy storage has been accepted as a means of conserving energy, it can also be applied to solving less critical problems of a social nature. It can, for example, avoid the erection of overhead wires in scenically sensitive or historic urban areas, especially for light rail.
On tunnelled metros where the stations are shallow but the line between them dips down, which is good way of saving energy, a failure of the power supply can leave a train stranded between stations. On-board storage could get the train into a station for evacuation, if necessary.
In a country like Japan, where most important routes are electrified but there are still many non-electrified branch lines, through operation from long electrified trunk lines on to a relatively short non-electrified section has been required socially to a greater extent than before in order to increase market share for railways.
Improving line receptivity
A normal DC substation using diode rectifiers cannot pass regenerated power back into the high voltage AC national power network, so it can only be consumed by other accelerating trains in the same feeding section. Even if enough trains are accelerating in total, where there is a long distance between the source and the load the pantograph voltage tends to be too high because of the line resistance.
In this case, the regenerating current itself has to be limited so as to avoid excessive voltage on the regenerating train. Countermeasures to improve this situation include: lower circuit resistance of the DC feeders additional power-consuming equipment in the substation or elsewhere along the trackside an additional inverter in parallel with the rectifier enabling surplus power to be fed back into AC power network and two-way power converters instead of conventional rectifiers.
In place of power-consuming apparatus such as resistor banks, various attempts to save energy by storing the surplus have been tried.
Japanese National Railways tried a chopper-controlled lead-acid battery, which was abandoned later, and Keihin Electric Express Railway Co installed two types of flywheel motor-generator sets, one of which is still in use.
More recently, JR West developed a new type of line voltage compensator using a lithium-ion battery with its own charge/discharge controller. This is placed between adjacent substations which are far apart.
After trials on the Katamachi and Tokaido lines, this voltage compensator battery went into operation at Shin-Hikida in October 2006, coinciding with conversion from 20 kV 60 Hz to 1·5 kV DC of the section between Tsuruga and Nagahama comprising 38 km of the Hokuriku Line and 2 km of the Kosei Line.
To reduce the feeding circuit resistance of double-track DC lines, unified electric feeding of down and up lines was uniquely used for one line by Hankyu Railway in the past, and other energy-conscious railways have since followed this example since regenerative braking was introduced.
Additional inverters or chopper-controlled resistors at some substations have been introduced mainly on lines with long and steep gradients.
Two-way power converters using pulse-width modulated (PWM) insulated gate bipolar transistors (IGBT) at DC traction substations are at the moment unique to Tsukuba Express (RG 6.03 p391), although this technology has proved very popular on AC trains since it was introduced for Series 300 on the Tokaido Shinkansen in 1991, using gate turn-off thyristors at the time of introduction.
Regeneration from higher speed
Modern DC trains using inverter-fed AC traction motors have similar tractive effort and speed characteristics in powering and braking modes. The quantitative difference between them arises from two factors: the voltage drop between substation and pantograph, and mechanical and electrical losses.
On 1·5 kV DC suburban and urban railways, typical voltage drop is around 18% to 24%, which gives around a 20% difference in voltage at the pantograph, assuming the substation voltage remains unchanged. Comparing powering with regeneration, and assuming a 20% difference in pantograph voltage due to line resistance which is negative in powering and positive in braking, the voltage drop is also around 10% when seen at the wheel rim since the efficiency of the inverter, traction motors and traction gear taken together is a little less than 95%.
So braking power can be around 30% greater than that of powering, which is far inferior to dynamic braking where power typically exceeds 2·5 times that of powering. This is achieved quite easily by exploiting the over-voltage capability of the traction motors without increasing the current.
In regeneration mode this capability cannot be used because the motor voltage is restricted by the catenary voltage, while dynamic braking is independent of the catenary. To cope with this situation, the two difficulties should be eliminated simultaneously by achieving over-voltage at the traction motor without much increase of pantograph voltage, and to reduce over-current from pantograph to the feeding system, which often loses receptivity even at the present regenerating current.
On-board energy storage devices (ESD) such as lithium-ion (Li-ion) or nickel hydrogen (Ni-H) batteries, or an electric double-layer capacitor (EDLC) can be used for this purpose. If the pantograph voltage and voltage across the ESD is added, the motor over-voltage capability can be used while power equal to the voltage over the ESD multiplied by the current is stored or discharged to/from the ESD, which means pantograph current remains as before.
Note that simple series connection of the voltage sources cannot be used if loss of receptivity takes place, which is the case with a normal feeding system. Instead of simple series connection, controlled series or parallel connection of the two sources is required so that receptive power is regenerated to catenary while the rest is stored in the on-board ESD.
If simple series connection of the two sources is required, the feeding system must be guaranteed to accept the maximum regenerating current at less than the maximum allowable pantograph voltage. On a 1·5 kV DC line carrying longer high performance trains with 6 to 15 car trains, this appears to be impractical.
More recent proposals and trials
'Pure electric braking' was proposed by the author in the mid-1990s, and after successful trials on Shin-Keisei Electric Railway in 1997, this technology is now very popular for electric braking down to a complete stop.
A Li-ion 'battery tram' has been developed by the Railway Technical Research Institute. Although this development is widely known for its ability to run without catenary, hybrid traction for energy conservation was an equally important objective.
At the moment, the Li-ion battery is the best device for high density energy storage per unit mass, but there are two problems still to be solved before practical widespread application. Power (as opposed to energy) density needs to be improved, and battery life in relation to the practical depth of charging and discharging has to be extended. Li-ion storage has been tried on several railways as well as inside RTRI. This system is not focused on regeneration from high speed but the two purposes mentioned above.
Since 2005 JR Central has been trialling EDLCs on one motor car of a Series 313 suburban EMU. The main purpose is to store surplus regenerated energy when the line is not receptive. Practical data has been collected, and optimal charging/discharging control has been studied.
EDLCs have much higher power density as well as a much longer life at the practical depth of charging than a Li-ion battery, but the energy density is inferior to Li-ion at the present stage of development. EDLC development is now being keenly pursued by several manufacturers, and the relationship between power density and energy density, together with internal resistance, has been gradually understood. Many application engineers consider its future for on-board energy storage to be quite promising.
Hybrid diesel/battery traction
JR East began development in 2000 of a hybrid traction system using a diesel-alternator and battery by making a test vehicle called the New Energy Train. Unlike the conventional electric transmission of a diesel railcar, this hybrid can utilise regenerative braking effectively, and the diesel engine can operate in a much improved regime of higher efficiency with less harmful exhaust gas and particulate emissions.
This hybrid system has cut fuel consumption by around 20% with much less environmental damage. Based upon the successful test in 2003-04, JR East will introduce a practical diesel hybrid railcar of Series E200 this summer on the Koumi Line. Although this line is famous for its steep and long gradients, E200 will be used mainly on the level section to the north of Koumi.
The original NE Train has now been converted into another test vehicle, this time of fuel cell hybrid traction. In place of the diesel-alternator and fuel tank, fuel cells developed for electric road vehicles and a high pressure 35 MPa hydrogen tank has been installed. Testing commenced in April this year.
Near future technology
Under a project known as 'Swimo', Kawasaki Heavy Industry is developing a Ni-H battery called 'Gigacell' for on-board energy storage, which it intends to function without a voltage controller.
Tsukuba Express, whose DC substations have two-way power flow with zero voltage regulation as referred to above, is interested in developing high speed regenerative braking in order to reduce the wear of brake pads. The company launched a study in May this year in co-operation with several universities, including the University of Tokyo and Kogakuin University.
The first stage will see maximum use of existing equipment studied, and the output voltage of the PWM converter substation will be raised optimally to about 1·6 kV from the existing 1·5 kV.
Then, using higher acceleration made possible by the increased pantograph voltage, the braking pattern in the high speed region will be rearranged so that the running time between stations remains constant, but regenerated energy is much improved. Thanks to Automatic Train Operation, the modification of running patterns will be easier than would be the case with manual driving.
For the second stage, the study team may consider introducing some kind of high speed regeneration system. Whether or not this will require on-board energy storage is not clear at present.
- Regenerative braking has been a long-standing feature of Japan's railways, where most services are worked by EMUs. Attention is now turning to energy storage technology to overcome receptivity issues Photo: Kazumiki Miura
- Fig 1. Characteristics of high-speed regeneration using an energy storage device to provide extra power when accelerating and absorb extra power that cannot be returned to the line during braking. The thin green lines show the original tractive effort - speed characteristics without ESD
- JR Central is evaluating the performance of electric double-layer capacitors (inset below) on a Series 313 suburban EMU. EDLCs offer far greater power density than Li-ion batteries, and several manufacturers are exploring the technology
- JR East's New Energy Train has been converted from diesel-alternator/battery hybrid operation to hydrogen fuel cell power, harnessing technology that is under development in the automotive sector
- Kawasaki Heavy Industries is using this test car to trial its Gigacell concept. It hopes to produce a full-scale LRV intended to run without catenary Photos: Kawasaki Heavy Industries