Power Supply Solutions

Posted: 26 September 2007 | | No comments yet

The dimensioning of power supply systems for railways is complex. To obtain a power supply that fulfils the traffic demands over its lifespan, improving or upgrading of the power system cannot be avoided. The many planned projects for the northern transalpine line Lyon – St Jean de Maurienne and the southern alpine corridor call for a study to dimension and assess their power supply in order to weigh investment proposals.

The dimensioning of power supply systems for railways is complex. To obtain a power supply that fulfils the traffic demands over its lifespan, improving or upgrading of the power system cannot be avoided. The many planned projects for the northern transalpine line Lyon – St Jean de Maurienne and the southern alpine corridor call for a study to dimension and assess their power supply in order to weigh investment proposals.

The dimensioning of power supply systems for railways is complex. To obtain a power supply that fulfils the traffic demands over its lifespan, improving or upgrading of the power system cannot be avoided. The many planned projects for the northern transalpine line Lyon – St Jean de Maurienne and the southern alpine corridor call for a study to dimension and assess their power supply in order to weigh investment proposals.

As part of the new rail link between Lyon and Turin, the 52km-long Franco – Italian base tunnel underneath the Alps will be built. The route from Lyon to the base tunnel crosses the Alpine Corridor, a North/South corridor linking Geneva-Chambéry-Grenoble. Due to the performance required in terms of both capacity and speed, RFF (Réseau Ferré de France) is planning two separate lines between Lyon and the Alpine corridor. One is a 78km-long high-speed line for passenger trains running between Lyon (St Exupéry) and the north of Chambéry, which will cross the Dullin and l’Epine massifs via two successive tunnels. The other is a primarily new line for freight trains between Lyon (St Exupéry, where it connects to the Lyon Bypass line) and the Alpine Corridor, to the south of Montmélian, which will cross the Chartreuse massif via a tunnel. To the east of the Alpine corridor, passenger transport requirements are such that freight and passenger traffic can be grouped together on a single new line which will join St Jean de Maurienne via tunnels through the Belledonne and Glandon/Rocheray massifs, and will subsequently continue on towards Turin through the Franco – Italian base tunnel. Moreover, there is talk of establishing a short-cut between the lines Lyon – Grenoble and Valence – Moirans by 2020 in order to enhance the conditions of the line leading to Grenoble, a project referred to as Shunt de Rives.

Although the tunnel to Turin will (partly) open to traffic by 2020, not all the planned routes leading from Lyon to the Franco – Italian base tunnel will have been executed by then. The first part of the route, from the airport Lyon St Exupéry to Avressieux, will be mixed high-speed and freight trains. Though the next section should develop into a separate freight and passenger route, the work will not have been completed by this date. This means that until at least 2020 freight and passenger trains will share the same tracks up to Avressieux, with a maximum speed of 220 km/h, from which point onwards passenger services and freight services follow their own route. The last section of the route towards St Jean de Maurienne will also not be completed before 2020, which means that all trains in the direction of Turin must use the existing “Ligne de Maurienne”.

Though not all the lines of the corridor are currently electrified, this should be the case by 2020. The lines Lyon – Grenoble and St André le Gaz – Chambéry are already electrified with a 25 kVAC autotransformer system. The new lines between Lyon and the Alpine Corridor will be fed by a 25 kVAC autotransformer system for the passenger service route and partly by a 25 kVAC system for the freight trains route. For this purpose three substations are planned. The line Valence – Moirans will have been equipped with double track, except for a section with single track between Romans Bourg de Péage and St Marcellin, and 25 kVAC power supply by 2013. The feed will come from existing 25kVAC substations. Similarly, the line Grenoble – Montmélian will be electrified to a 25kVAC autotransformer system using a new substation located close to the national electricity grid, which will also be used by the freight line between Lyon and the Alpine Corridor. Conversely, the line Chambéry – St Jean de Maurienne will keep its 1500 VDc electrification in the future, despite earlier studies showing that this power supply has already been stretched to its maximum capacity.

A wide variety of studies

The planned developments for the coming years are divided over many projects in this region. The various effects of electrification of the old and new lines have been detailed in over 10 studies. However, an overall study of how all these projects and electrifications interact was not yet available. Moreover, RFF planned on putting the contracts for some of these developments on the market by the third quarter of 2007, yet essential information was still missing. Therefore, RFF hired ARCADIS to do an overall study, analysing this complex rail network situation and the consequences of adapting and modernising the electrification of the various lines in this region. More specifically, the purpose of this study was to evaluate the number of substations, which are needed to meet the required performance levels in terms of robustness, reliability, quality, dimensioning, life cycle, and installation costs, both for the 2013 and 2020 horizons.

Approach of overall power supply study

There are two methods to perform a study of the capability of the power supply for future scenarios:

  1. A qualitative approach: can be used in cases when one is searching for an indication about whether the power supply can fulfil the (future) demands of increased traffic. For instance, if the power supply network currently has already reached its capacity limit, it will not be able to accommodate a great increase of traffic.
  2. A quantitative approach: is more appropriate for giving insight into the different components of the functioning of the power supply network in terms of energy consumption and determining which measures can help respond to future needs. An added advantage to this method is that it allows the possibility of adding specific improvements to the power system.

We have used a combination of these two approaches in the overall study to fulfil both the demands from the national electricity grid and the trains using this network.

The European standard EN50388 provides practical criteria for the quantitative analysis of power supply systems of existing and new railway connections, both 1500 VDC and 25 kVAC. Based on these criteria, a first scan of the network will demonstrate the performance of the substations and the quality index of the complete power system by means of the mean useful voltage. This voltage must be obtained by simulation in a dynamic load flow.

Xandra simulation tool

To carry out the quantitative analysis of the dynamic load flow calculations we have used the simulation tool Xandra. With this tool, developed in-house by ARCADIS, networks can be modelled, thus allowing the analysis of exploitation scenarios and the optimisation of traffic flows. Also it permits the integration of modifications to, amongst others, the infrastructure, the signalling, the power supply, and the rolling stock, which makes it a perfect tool for calculating the effects of future improvements to the service in long-term studies.

To build a model in Xandra the following information is generally required:

  • The tracks and their track characteristics as well as height profile. This includes the line’s integration in the larger network with its connecting branches and local constraints.
  • The rolling stock and characteristics of the trains, as well as the timetables they run.
  • The power supply system and its characteristics such as the location and type of the substations and the cable connections.
  • The security and signalling systems.

This simulation tool enables exploitation studies based on regular delay-free timetables or a stochastic approach of delay/traffic interference with timetables based on Monte Carlo variability. However, this tool also allows load flow calculations of the electrical power supply based on mutual coupling of electrical wires using ATP-EMTP (EMTP line models using Carson formulas) with a return feed back on the mechanical power of the train. Xandra can combine both direct current (DC) and alternating current (AC) lines in one network simulation, making it perfectly suitable for this overall study combining 1500 VDC and 25 kVAC lines.

The modelled network and scenarios

As the power demand of a rail network is dependent on the number of trains, the type of trains, the timetable of these trains and the type of infrastructure in the area-specific surroundings, all these aspects were entered into the Xandra model.

Considering the lines are located in mountainous terrain and slopes play an important part to acceleration and deceleration of trains, all slopes exceeding 6% and all arcs of less than 250 metres were included in the height and length profiles of the model. The power supply and traction designs were copied from the various studies and included in the model as well. Next, a timetable for the entire area was construed from RFF documentation. For all the rolling stock in the model, the characteristics, ranging from tractive effort to weight and nominal deceleration, have been integrated.

In order to be able to study the situation in stages and determine the robustness of the traction network, two scenarios were studied. For both of these, the characteristics of the infrastructure, traction system and estimated traffic have been simulated. The first scenario comprised the reference situation, the 2013 horizon. By this time, the connection of the southern Alpine corridor (Sillon Alpin Sud) to the LGV Méditerranée at the Valence TGV Station will have been constructed, as well as a third track at Gières, and a capacity increase by doubling sections and electrification of the full line Valence – Moirans, electrification and modernisation of the section Gières – Montmélian of the line Grenoble – Montmélian. The results from this phase will give indication of the possible reserves and potentials, and of specifications for new substations and new installations to be provided by RFF. Based on theses data, RTE will be able to check if the existing national grid could cope to the demand.

The second perspective is the situation of 2020. In addition to the parameters set in the 2013 phase, the Shunt de Rives should have been taken in service and the first phase of the route Lyon-Turin built (mixed traffic from Lyon to Avressieux, freight tunnel consisting of a monotube under Chartreuse, connection to the line Grenoble – Momtmélian, and passenger tunnel underneath Dullin – L’Epine).

By means of the simulation of these scenarios, the mean useful tension of the pantograph could be determined in order to guarantee the optimum performance of the rolling stock. The average power delivered by the substations was calculated for various time periods to ensure the substations would be dimensioned correctly. Finally, the unbalance in the substations was investigated for the connection to the national electricity grid.


With Xandra the rail network and corresponding traction system could be modelled as a function of the future traffic. This way, it could be established where the traction system was dimensioned correctly, which sections performed less, and where (and which) improvements were necessary to remedy incorrectly dimensioned sections. In general, the simulations provided RFF with confidence on the correctness of various line-specific SNCF studies. Not only did the overall study confirm SNCF’s earlier conclusions, it also showed that the electrification or construction of one line will not be detrimental to another project. As a result, the existing 1500 VDC Chambéry – St Jean de Maurienne line can remain to exist at its current maximum used capacity until 2013, though it will not manage the 2020 timetable. Three additional 25 kVAC substations in 2020 would suffice to provide the new lines with proficient power. These new RFF substations can be fed by the nearby 225 kV and 400 kV public utility lines. Alternatively, instead of adding a new station at Le Chaffard, which was proposed in former studies of the line between Lyon and the Alpine Corridor, new groups can also be added to the RFF substation at Grenay.

Food for thought

However, the simulation results have also provided some food for thought. Generally, a 25 kVAC electrified line and its substations will be connected to a high voltage network of over 100 kVAC. Then only the matter of unbalance can pose a problem to such a connection. Unbalance refers to the fact that a three-phased electricity supply over three branches must be more or less equal in time and size, meaning that trains at the different branches must demand simultaneously more or less equal amount of power. The maximum allowed value of unbalance of a substation feeding these three branches as accepted by RTE is 1.8% for a mean of 10 minutes.

The Perelle substation is connected to a 63 kVAC network with already a heavy load on it, making things a bit more delicate than with the more common connections. This is the reason we focused on the unbalance for this substation. In the 2013 situation with electrification of the section St Marcel les Valence – Moirans an unbalance in the order of 1.2% will be expected at the Perelle substation, which means this will not pose any problem to the running of trains on the line. However, the Perelle substation is also located in a network with already a heavy load on it, which means that supplying it with more groups from the existing national grid is likely to be difficult and would require the building of a new connection line. Then again, the simulation showed that the real load on the network, from the demand the trains put on it, will be less than the proposed, additional load of the 2 x 30MVA transformers. This knowledge could prove to be beneficial in deliberating whether to maintain the current substation as it is or to create a new RFF substation with a higher connection value at a location closer to a 225kV line.

Optimising investments

Finally, ARCADIS provided an alternative for the 2020 situation, in which the 1500 VDC power supply of the Chambéry – St Jean de Maurienne line is replaced by a 25 kVAC solution. This will have two distinct advantages. Firstly, whereas freight trains are allowed 100 to 120 km/h, they will barely manage 70 km/h climbing the steep slopes on this line due to shortage of available power provided by 1500 VDC. The 2020 timetable will be seriously disrupted by these low speeds, with trains waiting for one another. Even upgrading the existing substations and adding 5 or 6 new 1500 VDC substations to the line will not be able to relieve the fact that trains will run with limited power on these slopes. Yet with a 25 kVAC feed, this higher speed and hence a working timetable is feasible. Secondly, the alternative will require fewer substations. The substations used for feeding the Lyon – Turin route (RFF Ste Hélène du Lac and the new RFF St Jean de Maurienne) can be used to replace the complete 1500 VDC power supply between Aix les Bains and Modane and the section of the line from St Pierre d’Albigny to Albertville.

Thus, the overall study allows RFF to consider the various phases in the intricate compound of projects and to distribute investments in the power supply network according to its realistic needs, that is to say to concentrate its investments on where the performances are least in relation to expected traffic.

About the authors

Remco Paulussen is an MSc in Electrical Engineering and has been working at ARCADIS since 1998. Being involved in the development of Xandra he has used it to realise capacity studies, treat saturation problems at critical points, analyse power supply networks, and dimension substations.

Maurice Debrand, engineer SNCF, joined Réseau Ferré de France in 2001 as a Project Manager of the HSL Lyon Sillon Alpin, Area Manager for Rhône Isère Avant-Pays Savoyard and since 2005 he is also managing the modernisation of the southern part of the Sillon Alpin Valence – Grenoble – Montmélian.

Margo van Vliet MA gained her degrees from both Groningen and Utrecht University. Since 2001 she has been working for ARCADIS as a Senior Consultant, being involved in management of Xandra projects and general project management.

Mikaël Beck obtained his MS degrees from both Supélec Engineering School (France) and University of Wollongong (Australia). He has been working at ARCADIS since 2004.