Update: high-speed rail Novi Sad – Belgrade

Proposals for a high-speed line between Novi Sad and Belgrade were posted here several years ago, when official plans were still incomplete. By now, the project is nearing completion, so there is no point in proposing an alternative alignment. Most of it follows the old line anyway. The official project is less ambitious than proposed here, with lower capacity and speeds, but still a significant improvement by Serbian standards. The line was always intended as part of a longer Budapest – Belgrade route, and work has started on the rest of that route, slowly. The line in Belgrade itself still needs improvement, and the new central station is still unfinished, despite decades of work.

There are possible alternative routes between Budapest and Novi Sad, but not between Novi Sad and Belgrade. Both cities are located at strategic river crossings, and the geography determined the alignment of the railway line in 1883.

---

Novi Sad is located north of the Fruška Gora mountain range. The range runs east-west, it is about 80 km long, and the Danube flows east-west to avoid it. The range has a ‘promontory’ on the northern side, and the Danube also bends around this promontory. Exactly at its northern point is the fortress of Petrovaradin, and on the opposite bank is the modern city of Novi Sad. The line from Budapest crosses the Danube here.

Originally the line used the shortest river crossing, and ran under the fortress in tunnel. From there, a straight line to Belgrade would require an additional tunnel under the Fruška Gora range, so the line turned south-east, to run alongside the Danube. The Novi Sad station, and the Danube rail bridge, were later moved further east, so the line no longer needs ridge tunnels. The old line had another tunnel further south, at Čortanovci. There, the line turns away from the river, and climbs up to the plateau south of the Fruška Gora. From there, it is almost a straight line to Zemun, at the edge of the Belgrade agglomeration.

The climb from the Danube at Čortanovci was the worst section of the old line, and it has been completely replaced, by a new viaduct/tunnel combination. The bridge at Novi Sad (Žeželjev Most) had been replaced earlier, after NATO bombing in 1999. South of Čortanovci, no major work was needed – just track renewal, and realignment of some curves. However, the most that Serbia can afford was a mixed-traffic double-track railway, designed for 200 km/h, at least on the middle section. The capacity is insufficient for true demand at a European level, and in some parts of Europe new high-speed lines are faster – 300 km/h.

The most significant realignment on the new line is through the Čortanovci tunnel, and on to Beška. The two images below show the variants proposed here earlier, and the line as built.

---

---

The earlier post also proposed a new tunnel from Zemun into western Belgrade (Novi Beograd). The line drops 15 m here, from the plateau to the Danube/Sava flood plains, and technically, a new tunnel is no problem, so it can be added later. However, the earlier proposal also assumed, that the HSL would terminate near the old Belgrade central station. That station has now been closed, and the site cleared for redevelopment. The only option is the badly-designed new central station at Prokop, Beograd Centar. It was planned decades ago, for internal Yugoslav train services, not for European high-speed networks.

In theory, Belgrade is the logical interchange station, between Central Europe and the Balkan Peninsula. That is irrelevant at present, because services south of Belgrade are so bad – infrequent trains, on low-speed single-track lines. A first step is the upgrading of the Belgrade to Niš line for 200 km/h, which is at the planning stage. But even if there were high-speed lines to Belgrade from all directions, Beograd Centar in its present form would be inadequate. That is a good reason to consider new bypass lines, such as this northern bypass proposed here earlier, from Batajnica to Smederevo.

---

That would not, however, solve the problem for trains into Belgrade itself. The approach lines to Beograd Centar are of low quality, and shared with local and regional services. There is no space to reverse trains, or store and service them, even for short periods. It might be possible to build additional terminal platforms on the north side of the station. That would also require extra approach tracks, including an additional bridge over the Sava.

And in theory, the low-grade lines into Beograd Centar could all be replaced. The existing tunnels could be duplicated, with additional fast tracks, and the older tunnels reserved for urban and regional services. Similarly, the bridge over Sava could be duplicated, and the life-expired Danube bridge replaced, by a new four-track version. The remaining surface approaches, inside the urban area, could be four-tracked – with some limited demolition. However, even with 100% adequate approach tracks, Beograd Centar would still be badly located, and have insufficient capacity.

Update: Rail Baltica in the region

New and upgraded rail lines, in the Baltic States and Kaliningrad Oblast, were proposed here earlier. The proposals were integrated with each other, and with the older network. One proposal was a north-south high-speed rail line, Kaunas – Riga – Tallinn. Since then, the official Rail Baltica project has finalised its alignment, and it is no longer relevant to propose an alternative. On the other hand, the war in Ukraine has altered the logic of the project. This post reconsiders the earlier proposals, with suggestions for improving the official alignment.

The capacity of the rail network in the region is limited, and passenger services are minimal. Rail transport has low priority in the three Baltic states, Lithuania, Latvia, and Estonia. They don’t cooperate with each other, or with their neighbours Belarus and Russia, and sometimes obstruct cross-border services. That was true long before the war. Lithuania was fined by the European Commission, for deliberately vandalising its own tracks, to block rail freight traffic into Latvia.

Not surprising, then, that the Rail Baltica project suffers from political hostility and interference. The alignment was determined by politics from the start, to avoid the Russian exclave Kaliningrad Oblast, which is the northern half of former East Prussia (Ostpreussen). Another example: the planned high-speed route from Vilnius to Warsaw makes a long detour. There was originally a direct line, but part of it is now in Belarus, and the three states can’t agree on re-opening. That would also require conversion to standard gauge, and the five states also can’t agree on gauge conversion policy. Another example: Latvia insisted on routing Rail Baltica through Riga airport, which will delay travellers between Estonia and Lithuania. These travellers will not be using Riga airport anyway, because all three countries protect their own national airports.

Then came the Ukraine war, which undermined the stated intention, to connect the Baltic states to the rest of the EU. The Rail Baltica alignment passes between Belarus and the Kaliningrad exclave. When it was planned, that was not a problem, because no hostile action was anticipated. But since then, Belarus has allowed Russia to use its territory for an invasion, and it might do that again. The unproblematic corridor from the EU to the Baltic states became a geopolitical flashpoint, the Suwałki Gap – a potential Russian corridor to Kaliningrad. Of course, that would mean war with Poland, but even if it was arranged peacefully, it would still destroy the concept of an internal EU link to the Baltic states.

For the present, it does not matter anyway, because the Polish government never took the line beyond Suwałki seriously. It was left as a single-track non-electrified route, with an indirect alignment via Šeštokai. In practice, the Polish refusal to improve the cross-border link, reduced Rail Baltica to a Kaunas – Riga – Tallinn line, with a few extra trains to Warsaw.

What kind of line?

The Rail Baltica project is not a standard high-speed line either. It will carry freight – serving three multimodal terminals, and the port east of Tallinn (Maardu). High-speed rail lines usually carry no freight, for good reasons. Slower freight trains obstruct fast trains, and they require an appropriate gradient, track, and structures.

The Rail Baltic line will also have regional stations – possibly almost 40. Again there are some examples in other countries, but conditions there are very different. Regional population densities in the Baltic States are very low, and declining even further.

Population decline, 2011-2020 (map from Reddit, no source was given).

That means that the regional stations will serve very few people, too few to fill the trains. Some of the proposed stations are clearly absurd – Pasraučiai village, with 34 inhabitants, Kaisma village (123 inhabitants), and Urge village (144 inhabitants). No specific funding is available for these stations, which will cost about €2 million each, and the small local governments cannot afford them. If there is no external source of funding, then most of these stations will never be built.

Regional trains are only possible, because planned frequencies are very low. Trains between the main stations – Tallinn, Pärnu, Riga, and Kaunas – will run every two hours. A high-speed line cannot be economic with such frequencies – the HSL-2 in England is designed for trains every three minutes. The frequency of regional trains is not fixed yet, but it cannot be very high either, or the trains will be almost empty. This suggests that the ‘high-speed’ Rail Baltica is in fact primarily for freight, which would make it unique among high-speed lines. These freight trains will run at 120 km/h, on a line built for 250 km/h.

Deficiencies of the planned alignment

Kaunas, Riga, and Tallinn are approximately aligned north-south, and the Rail Baltica line generally follows this route. The alignment is not a straight line at local level, however. It curves continuously, partly due to environmental constraints. Most of it does not follow existing infrastructure.

At three points, there are major flaws in the alignment north of Kaunas. After leaving the built-up area of Kaunas, the line turns toward Panevėžys (population 95 000). Instead of entering that city, however, it turns west, to pass though open country. A new north-south line through Panevėžys is difficult, because the existing rail line run east-west. So instead, there will be an interchange station, where the new line crosses the Šiauliai –  Panevėžys line. Experience in France shows that stations in open fields don’t serve the surrounding region, even if they cross an existing rail line.

The design error at Panevėžys can be corrected, by adding a new standard gauge line into the center from the west. It can largely follow the  bypass (A17), and join the existing alignment about 3 km east of the station. It is not a perfect solution, but it would allow a high-speed Panevėžys – Kaunas shuttle service.

The second major design flaw is the approach to Riga. In fact Rail Baltica has a double alignment at Riga, with a line through the city, and a bypass line. However, the new line through Riga will turn west, then follow the ring motorway to the airport, and then turn back toward the central station, alongside the existing Jūrmala line.

It is relatively easy to correct this design error, by building a link to the existing Jelgava – Riga alignment near Tiraine, with extra standard gauge tracks into Riga. The two routes join at Tornakalns, close to the main station. The new link alignment would be about 15-20 km long, with about 8 km alongside existing tracks.

North of Riga station, the planned alignment could be improved by building new standard gauge tracks, alongside the exiting line through Garkalne. This alignment starts with two L-curves, but is almost straight, after passing Jugla.

North of Riga, the planned route roughly follows the coast to Pärnu in Estonia, and from there to Tallinn. The third major design flaw is on the approach to Tallinn. The new line does not follow the logical south-north route, alongside exiting lines, but uses a semi-circular route to enter Tallinn from the east. Apparently, the freight link to the port at Maardu was more important, than passenger access to Tallinn. A possible tunnel to Helsinki cannot justify the diversion: it is not even clear, if a tunnel will run east of Tallinn. In any case, Rail Baltica trains will terminate at a new station at Ülemiste, at the edge of Tallinn. Passengers must change to local trains, to reach the city centre.

There is no easy fix here. It would however be possible to build a new line from the south, following existing tracks, from near Saku. It could terminate at the existing Tallinn central station, and include a connection to Ülemiste, if necessary.

Apart from these major corrections, it might be possible to correct other definities of the the Rail Baltica project, such as the lack of connection with local rail services, and no planning for conversion of Russian gauge railways. The general principle suggested here earlier, was to retain the east-west lines in Latvia and Estonia at Russian gauge, and convert most north-south lines to standard gauge, with most of the Lithuanian network converted.

Line from Warsaw

Gauge conversion is also an issue on the route from Warsaw to Kaunas. The present low-quality compromise uses either mixed-gauge single track, or two separate tracks. This is generally unnecessary, as there is no reason for Russian gauge tracks south of Kazlu Ruda, about 40 km from Kaunas on the Kaliningrad line. In fact, the logical option is to convert the entire Kaliningrad – Kaunas line to standard gauge – as most of it was until 1945. That is politically unacceptable at present, because the line links Russia to a Russian exclave, but this blog ignores political constraints anyway.

If the existing line through Šeštokai was standard gauge, the real problem would be apparent – no political will to improve a low-quality line. At least in Poland – on the Lithuanian side, there are plans for a new link line, passing Marijampolė. That is a logical route, following the main road toward Suwałki. On the Polish side, however, there is certainly no prospect of a new high-speed line from Suwałki to Warsaw, which is the logical extension of any Tallinn – Riga – Kaunas HSL.

An approximate alignment for a Suwałki HSL was posted here earlier, but that post is out of date. In summary, the new HSL would start with a new tunnel in Warsaw, to connect the cross-city main line to Polish line 6, formerly the Warsaw – Vilnius – St. Petersburg line (1862). The HSL would run alongside Line 6, and diverge about 60 km from Warsaw, turning north-northeast toward Łomża. Because the centre is on higher ground, the new line can pass through the city, using the existing terminal station and a new tunnel.

The HSL would bypass Ełk, following the Via Baltica road, but some trains could serve Ełk, by leaving and rejoining the HSL. The HSL would then run north-east to Suwałki, again following the Via Baltica rather than the old rail alignments. The current terminal station in Suwałki (population 69 000) can be replaced by a new through station, using existing rail alignments on the eastern side of the centre.

From Suwałki, the HSL would follow the old main road to the Polish border, where is would join the proposed new line to Kaunas, with one stop at Marijampolė. The new HSL would therefore have only three intermediate stations – Łomża, Suwałki, and Marijampolė.

Berlin – Kaunas route

If we ignore the geopolitical issues, the most logical route from Western Europe to the Baltic States, is via the old German main line. The Preussische Ostbahn to East Prussia extended 600 km, from Berlin Ostbahnhof to Königsberg. East Prussia was divided between Poland and the Soviet Union in 1945, but if that alignment was still German territory, then trains from Berlin to Kaunas would probably run via Königsberg. That is still an option, because a new standard-gauge HSL, from Berlin to Kaunas via Kaliningrad, could connect to the Rail Baltica alignment at Kaunas.

Das Deutsche Reich, Alsace to Memelland

When East Prussia was still German, the trains from Berlin continued to Memel, now Klaipeda. That was the last German city, at the ‘far end’ of the Reich. There was no main rail line from Königsberg to Riga, which was historically better connected to Moscow. (A secondary line was later built via Radviliškis junction, near Šiauliai.) A new high-speed route from Kaliningrad to Riga, via Šiauliai and Jelgava, was proposed here earlier. However, assuming completion of the Rail Baltica line, the route via Kaunas seems a good alternative.

Berlin – Riga HSL

There are other possible routes to the Baltic States. Parallel to the Berlin – Königsberg route, there was also a route Berlin – Posen – Thorn – Allenstein – Insterburg – Tilsit – Memel. Of these seven cities, only one is still in Germany. Upgrading of this line, to the standards of the German Ausbaustrecken, was proposed here earlier.

Restored and upgraded Olsztyn route

The new route would be Berlin – Poznań – Toruń – Olsztyn – Chernyakhovsk – Sovetsk – Klaipeda, connecting to a possible new line to Liepaja. As with the Berlin – Königsberg line, this line could also carry trains to Kaunas, which can join the Rail Baltica line to Riga.

So even if the present route of Rail Baltica has many defects, most could be corrected in the longer term, as part of a European HSL program. Again, this simply ignores the present geopolitical issues, and the self-interest of national governments.

Helsinki tunnel

The Rail Baltic project is often presented as access to an undersea rail tunnel, linking Tallinn to Helsinki. In fact, this tunnel is not a precondition for the project, and not a passenger route to Western Europe. Despite much public interest, there are no serious plans at the moment, for this expensive project.

Several route options have been proposed, some of them by private investors. If there is a single long tunnel, then it might start east of central Tallinn, on the Viims peninsula, which is already served by the Maardu port line. If the tunnel is constructed from island to island, possibly via Naissaar island, then it might start west of Tallinn, even through the total route is longer. Similarly, on the other side, an indirect route to central Helsinki might be easier to construct.

Possible Helsinki – Tallinn tunnel alignments

The route, from central Tallinn to central Helsinki, would be about 100 km long. The absolute minimum is the great circle distance between the two central stations, which is 82 km. That is not a ‘short trip’. The route from Berlin to Helsinki is about 1400 km, beyond the generally accepted range for displacement of air traffic to high-speed rail.

So even with a Tallinn to Helsinki tunnel, Rail Baltica is not a ‘shortcut to western Europe’ from Finland. Most passenger traffic through this tunnel will be Finland – Estonia traffic, just as with the existing ferries. The issue of gauge conversion must be seen in that context: it is not absolutely necessary to convert the entire Finnish rail network. Theoretically, the tunnel should have mixed-gauge tracks for maximum flexibility, but that is not compatible with high speeds – trains cannot be centred under the catenary. So in practice, it might be more logical to use Russian gauge for a tunnel between two Russian-gauge countries. In that case, Ülemiste would be the interchange station. The balance is different for freight, however, because the long route to western Europe is still faster than a ship. It might therefore be preferable to construct a separate single-track mixed-gauge tunnel, for freight trains only.

Update: HSL-Oost van Utrecht naar Arnhem

Over een oostelijke hogesnelheidslijn naar Duitsland, wordt al langer gepraat. Officieel is het project in 2001 afgeblazen. Een overzicht van de eerdere en huidige plannen staat in Herwaardering HSL-Oost door J.C. van Ham en J.H. Baggen, van TUDelft. De HSL zou in bijna alle plannen de bestaande lijn volgen, de Rhijnspoorweg uit 1845.

De Rhijnspoorweg …

Maps by OpenTopoMap, under CC-BY-SA.

Hier werd in 2010 een variant voorgesteld, dat rekening houdt met omlegging van de spoorlijn Utrecht – Arnhem naar De Uithof, en een nieuwe regionale spoorlijn Veenendaal – Ede. Door toenemende bebouwing van de omgeving, zijn die twee voorstellen minder haalbaar, en voor De Uithof is een nieuwe variant voorgesteld. Hieronder wordt gekeken naar ombouw van de bestaande lijn.

Stand van zaken

De lijn Utrecht – Arnhem werd in 1938 geëlektrificeerd, en sindsdien is er niet veel veranderd. De lijnsnelheid (baanvaksnelheid) is 140 km/h. Het is duidelijk dat verbetering van de lijn voor ProRail geen prioriteit heeft. Het tracé door Utrecht is wel verbouwd, met een ongelijkvloerse aftakking bij Station Lunetten, maar van daar tot Station Arnhem is bijna niets verbeterd. De laatste jaren is wel begonnen met het vervangen van overwegen, maar dat gaat niet snel. Langs het traject liggen nog enkele eenvoudige overwegen ten behoeve van landbouw-voertuigen, een bekende risicofactor.

Toegang tot de boerderij Groep 17: inmiddels overbodig want het pand is gesloopt…

Enkele stations worden ook verbouwd, in Driebergen-Zeist met vier sporen, maar daar blijft het bij. Over een HSL-Oost naar Duitsland wordt al zeker 50 jaar gesproken, maar in dit tempo komt het er nooit.

Functie van een HSL-Oost

Over de huidige lijn rijdt elke twee uur een ICE naar Duitsland, en DB voorziet ook geen verhoging van de frequentie. Dat is niet genoeg om een hogesnelheidslijn te rechtvaardigen. De HSL-Oost zoals hier beschreven, is bedoeld voor hogesnelheidstreinen op binnenlandse routes, naast frequente ‘Europese’ diensten. De huidige Intercity zou verdwijnen: Ede-Wageningen, Veenendaal-De Klomp, en Driebergen-Zeist worden bediend door snelle inter-regionale diensten. Er zouden ook meer stoptreinen rijden: de reeds geplande 8 treinen per uur van Randstadspoor rond Utrecht, en een vergelijkbaar agglomeratie-net rond Arnhem en Nijmegen.

Hier werd ook een HSL Amersfoort – Wageningen – Nijmegen voorgesteld. Deze zou het Rhijnspoorweg-tracé volgen tussen km 63 en km 70, ten noorden van Veenendaal.

Die HSL is geen voorwaarde voor een HSL-Oost. Bij aanleg zouden nog meer meer treinen over de lijn Utrecht – Arnhem rijden, dat wel. Het maakt bijvoorbeeld een snelle dienst Amersfoort – Arnhem mogelijk (20 minuten), en een rechtstreekse dienst Utrecht – Nijmegen (25 minuten).

Langs het tracé

Met de verbouwing van Utrecht Centraal, en de spoorlijn door Utrecht, zijn deze in principe geschikt, voor hoogfrequente diensten en hogesnelheidstreinen. Na de ongelijkvloerse kruising bij Lunetten, is de lijn echter tweesporig, en al snel volgt de eerste overweg, Mereveldseweg.

Tussen Utrecht en Station Driebergen-Zeist liggen er vier overwegen. Dat waren er meer: de lijn is dwars door de percelen gebouwd, en boeren kregen een overweg voor eigen gebruik. Daarvan is één overgebleven. In Bunnik heeft ProRail de overweg bij het station wel vervangen, maar niet die op de hoofdweg (Stationsweg). Bij Station Bunnik is geen rekening gehouden met vier sporen, en die zijn er ook niet gepland. Alleen Station Driebergen-Zeist krijgt vier sporen, met een eilandperron voor de stoptreinen, en een onderdoorgang voor de N225. Iets verderop wordt ook de overweg Arnhemse Bovenweg gehandhaafd.

Dit is duidelijk onvoldoende voor hoge snelheden en hoge frequenties. Dit gedeelte moet niet alleen vier sporen krijgen, maar ook beter ingepast worden. Station Bunnik moet verbouwd worden, volgens het model van Driebergen-Zeist: een eilandperron voor de stoptreinen. Ondertunneling in Bunnik zelf is overbodig, omdat de spoorlijn toch de grens vormt tussen wonen en industrie. Ondertunneling is wel een optie tussen de A27 en Bunnik, en vooral tussen Bunnik en Driebergen-Zeist. Tussen de A27 en Bunnik zou ook de voorgestelde tweede lijn naar Amersfoort aftakken, maar dat heeft geen verder gevolgen voor de lijn naar Arnhem.

Nieuwe lijn naar Amersfoort: Oostbroek-tunnel…

Na de overweg Arnhemse Bovenweg, loopt de lijn door de bossen, naast de A12. Hier is voldoende ruimte voor meer sporen. Een HSL naast de bestaande lijn is hier een optie: het is makkelijker aan te leggen, maar vereist extra ongelijkvloerse aansluitingen. Ongeacht de gekozen variant, is het tracé in Maarn ongeschikt voor hoge snelheden. Gelijk met de bouw van de A12 werd de spoorlijn verhoogd, en het station verplaatst (het lag vroeger aan de westkant van de Amersfoortseweg). De snelweg ligt echter pal naast het perron, en de bocht hier moet wel verruimd worden. Verschuiving van de (net verbouwde) A12 is onvermijdelijk. Daarbij moeten de perrons van Station Maarn waarschijnlijk naar het oosten worden verschoven.

In Maarsbergen, is ProRail al bezig met het vervangen van de overweg, door een C-vormige onderdoorgang. De lijn kan hier op maaiveld blijven: er is voldoende ruimte voor vier sporen, en ook voor een heropening van Station Maarsbergen.

Na Maarsbergen volgt (bij km 61) de aansluiting De Haar. In de eerste versie van deze post (2010) werd verondersteld dat alle stoptreinen via Veenendaal zouden rijden, en vervolgens over een nieuwe lijn naar Ede. Omdat deze niet meer haalbaar is, zou de lijn richting Ede gewoon viersporig moeten worden.

Het tracé tussen Maarsbergen en Ede loopt door open agrarisch gebied, met meerdere overwegen. Enkele boerderijen zijn slechts via een overweg te bereiken, maar ze zijn allemaal te bereiken zonder overweg, door een nieuwe toegangsweg aan te leggen. Eén is al gesloopt, wat ook een oplossing is.

Enkele bruggen en onderdoorgangen voor lokaal verkeer blijven nodig. Bij de Laagerfseweg kan de brug over de A12 verlengd worden, bij de Heuvelsesteeg is een onderdoorgang waarschijnlijk beter. De belangrijkste overweg is de N233 aan de rand van Veenendaal, dat op een bocht in de spoorlijn ligt. De bocht zelf ligt in het voormalige Fort aan de Buursteeg, pal naast het nieuwe Bezoekerscentrum Grebbelinie. In de eerste versie (2010) werd hier een viaduct voorgesteld. Inmiddels heeft de gemeente besloten tot de aanleg van een geheel nieuwe onderdoorgang verderop, en de oude overweg wordt vervangen door een fietstunnel. Dat verandert niets aan de scherpe bocht zelf: bij verruiming is de sloop van een deel van het fort onvermijdelijk.

Afhankelijk van de gekozen variant, krijgt Station De Klomp een eilandperron en twee buitensporen, of wordt de HSL naast de bestaande lijn aangelegd. Omdat Station Ede waarschijnlijk het model van Driebergen-Zeist volgt, is eventueel een ongelijkvloerse aansluiting nodig, ten oosten van De Klomp. De spoorlijn is hier al verhoogd.

1500 m na De Klomp begint de bebouwde kom van Ede, eerst alleen bedrijfsterreinen. Binnen Ede worden de vele overwegen geleidelijk vervangen. Het lijkt niet dat de nieuwe onderdoorgangen berekend zijn op extra sporen, maar er is wel voldoende ruimte, ook onder de A30-brug. Bij aanleg van vier sporen door Ede, kunnen ook twee stations worden toegevoegd, één bij het bedrijfsterrein Frankeneng, één bij de nieuwe onderdoorgang Hakselseweg.

Deze stations worden bediend door regionale diensten uit Utrecht, al dan niet als onderdeel van Randstadspoor. Ede is een geschikt eindpunt voor alle stoptreinen uit de richting Utrecht, en eveneens uit de richting Arnhem. Dat betekent dat het NS-model, waarbij stoptreinen door meerdere provincies doorrijden, losgelaten wordt. Treinen uit Amersfoort (Valleilijn) zouden wel doorrijden naar Arnhem, als dat goed te combineren is met een agglomeratie-net Arnhem / Nijmegen. Station Ede moet flink worden verbouwd: aanvullende keersporen, en doorgaande sporen voor de snelle treinen. Eventueel kan de HSL hier op viaduct, om voldoende ruimte vrij te maken.

Ede – Arnhem

Van Ede tot Arnhem loopt de lijn vooral door bossen. Het tracé is bijna kaarsrecht, maar wel met lokale hoogteverschillen tot 20 m. De weinige overwegen zijn makkelijk te vervangen door onderdoorgangen of bruggen. Omdat de oude spoordijken niet ontworpen zijn voor hoge snelheden, is een HSL naast de bestaande sporen daar waarschijnlijk de beste oplossing. (Mogelijk kunnen ook de bestaande spoordijken door viaducten worden vervangen). De HSL is dan alleen voor de snelste treinen, de oude sporen voor goederen, stoptreinen en sneltreinen.

In Wolfheze, waar de lijn op maaiveld ligt, is een onderdoorgang nodig. Voor extra sporen is hier voldoende ruimte. Van daar tot Arnhem is de lijn al kruisingsvrij, en ook vrij van bebouwing naast de sporen.

1500 m voor Station Arnhem, ligt de aansluiting met de lijn uit Nijmegen. De nieuwe dive-under hier, is ontworpen voor het kruisingsvrij keren van de treinen van/naar Nijmegen. Het keren zou juist overbodig worden, met de aanleg van de hier voorgestelde HSL Wageningen – Nijmegen. Maar ook zonder deze HSL, is de dive-under van beperkte waarde, als hier een viersporig lijn moet aansluiten op de perrons, en zeker als de lijn uit Nijmegen ook viersporig wordt. De aansluiting moet dus volledig verbouwd worden, om alle treinen kruisingsvrij naar de perrons te leiden.

De echte problemen in Arnhem beginnen aan de andere kant van het station. Het verbouwde station, met vier eilandperrons, kan zelf alle toekomstige verkeersstromen verwerken, ook frequente hogesnelheidstreinen richting Oberhausen. De tweesporige lijn langs de Arnhemse binnenstad vormt echter een flessenhals: aanleg van extra sporen wordt al jaren tegengehouden door de gemeente. Uitbreiding is ongetwijfeld een voorwaarde voor meer treinen richting Zwolle, Twente, de Achterhoek, en Emmerich, maar deze routes worden hier niet verder besproken.

Van Utrecht Centraal naar Arnhem is het slechts 57 km, en met verruiming van alle zeven bochten, en een lijnsnelheid gedeeltelijk boven 200 km/h, moet een reistijd van rond 20 minuten haalbaar zijn.

Hyperloop Schiphol Airport – Lelystad

The Dutch government has proposed a Hyperloop test track in the Netherlands. It might also be incorporated in a future commercial line, between Schiphol Airport and Lelystad Airport. This post looks at the constraints and alternatives. At this blog, posts on infrastructure inside the Netherlands are normally in Dutch. This one is in English, because of international interest in Hyperloop projects. (For the same reason, the official report is also in English.)

Amsterdam and Flevoland: click to enlarge…

Maps by OpenTopoMap, under CC-BY-SA.

Is this a realistic plan? The short answer is no. Even if the technology was available tomorrow, the traffic on this specific route is unlikely to justify the investment. The main constraint is the limited function of the proposed line, as a pure airport-to-airport link. That is explained in more detail below.

However, that probably does not matter anyway, since the entire project may be no more than an election gimmick. The Netherlands has a history of ‘electioneering by infrastructure’ – voters are promised new rail lines, new stations, new metro lines, and new tram lines. In this case, the government is also suggesting that the Hyperloop could be extended to the north of the country. This too is a tradition: politicians repeatedly propose a northern high-speed line, but it never gets built. The current Hyperloop proposal is timed to coincide with campaigning for municipal elections, in early 2018.

The only thing that might get built is a section of test track, east of Almere. For that, there are no constraints except the finance. There is an alignment available (a reservation for a road which was never built), and it runs through open flat farmland. You can find it as ‘Vogelweg Zeewolde’ on Google Street View. The official report claims that the test track could be incorporated in the later commercial line, but it seems to be in the wrong place for that.

Function of the proposed commercial line

The proposed Hyperloop would connect Amsterdam Airport Schiphol with Lelystad Airport, which is currently being expanded. Schiphol Airport is very close to Amsterdam: only 9 km from city centre to runway. That is a great advantage for passengers, but an airport in a densely populated urban region brings many problems: congestion, development pressure, and environmental damage. Relocation to an artificial island in the North Sea has been suggested for decades. The prohibitive cost of that option led successive governments to search for alternatives on land.

Some traffic was relocated to regional airports, but Schiphol continued to grow. The government concluded that a relief airport was needed in proximity to Schiphol, and Lelystad is the only candidate. It is already owned by Schiphol Airport, so there are no problems in coordinating strategy.

Lelystad Airport is located in open farmland in the Province of Flevoland. The land was reclaimed from the sea, which has a specific advantage: there are no historic towns and heritage sites to obstruct expansion. Since this is a fully planned landscape, there is no urban sprawl either. The airport is located about 4 km outside the city of Lelystad, a new town with 77 000 inhabitants.

The plan is for an expanded Lelystad Airport to take the low-cost carriers and summer tourist traffic. Schiphol would keep the higher-value long-haul and transfer traffic. However, there is a problem with this strategy: no tourists want to go to Flevoland. Most of them want to go to Amsterdam, and for that reason a rail link to Lelystad airport has been discussed, although there are no concrete plans. The existing rail line Amsterdam – Lelystad – Zwolle passes 5300 m to the north of the terminal.

The proposed Hyperloop would link the two airports directly, with no intermediate stops. That in itself has major consequences for the probable traffic. It also matters exactly where the stations are located. The line might be a airside-to-airside link, which would create a ‘single airport’, from the passengers’ perspective. On the other hand, all non-airside passengers will face airport check-in and security procedures, which would cancel the advantage of high speed. If the line is not an airside-to-airside link, transfer passengers must check out, and then check in again. Whatever the arrangement, the line will serve the two airports, and nowhere else.

Most importantly, it will not serve Amsterdam. Anyone in Amsterdam who wants to use this Hyperloop will first go to Schiphol Airport. Unless their destination is Lelystad Airport, they will need to change mode again, at the end of the Hyperloop. Given the relatively short distances, Hyperloop cannot offer faster journey times if passengers must detour. More on this below, where the line is compared with existing and improved railways.

Possible alignment

The two airports are 55 km apart in a straight line, terminal to terminal. At Lelystad Airport itself, the choice of alignment is open: it can exit the airport either north or west, or possibly on the south side. The proposed Hyperloop test line is near the airport, but is not aligned with the future terminal. An extra curve, and an extra line along the N302 road, would be required to incorporate it into the commercial line.

At the western end of the proposed test track, there is no self-evident route toward Schiphol. If you simply extend the line of the test track, it points at Rotterdam. The Hyperloop line to Schiphol must therefore turn further north, and that means it must pass through the city of Almere. Even as it turns, it will cross planned new housing developments. At some point it must also cross the lake, which separates reclaimed Flevoland from the older land to the west.

These constraints imply, that the Hyperloop  must follow existing infrastructure around Almere. Notably, there are very limited options for bridges across the lake. The government report suggests, that the line will follow the motorways A6, A1 and A9. That rules out the Vogelweg alignment, which is 4 km south of the A6.  (In other words, the report contradicts its own claim, that the test track can be incorporated in the later commercial line.)

The Hyperloop will probably cross the lake alongside the existing rail and motorway bridges, at Muiderberg. There is very little flexibility here, since there are urban areas on the other side: Naarden, Bussum and Huizen. It is an upper-income area, with expensive houses, protected landscapes, and nature reserves. The area is also unsuitable for tunnelling: it is a low ridge of relatively loose sand and gravel.

However, following the motorways brings new problems, as shown by initial studies for a Hyperloop in California. Very few motorways run in straight lines, and their curvature is too sharp for the Hyperloop, running at over 1000 km/h. Instead of running neatly alongside the Californian freeways, or in their median strip, the Hyperloop alignment would need to cross the freeway often, to maintain its own curvature. There are some straight motorways in the Netherlands, but the motorway through Almere is curved.

The problems are worse on the other side of the lake bridge, heading toward Schiphol Airport. The bridge is aligned southwestwards, again toward Rotterdam and not toward Schiphol. The existing motorway turns 60 degrees westward after the bridge, from the A6 to the A1. It then turns 50 degrees west-southwest, to pass through suburban Amsterdam, where it is numbered A9. The A9 then turns 90 degrees to run alongside the A2, and then 120 degrees, to run west-northwest to Schiphol. There is no way a Hyperloop line can follow all these bends (in blue on the map)

There is an alternative: the unbuilt alignment of the A9, which would have run in a wide curve south of the town of Weesp (shown in green). A motorway on this alignment was planned for decades, but rejected on environmental grounds. A tunnel was subsequently rejected on cost grounds. It is possible that the Hyperloop could descend into tunnel from the bridge at Muiderberg, and follow this alignment. That would however cancel one of the claimed advantages of Hyperloop, namely that it needs no expensive tunnels.

The A9 motorway continues through Amstelveen, a suburban municipality. Here too there are curves, but the main problem is limited space. The motorway is jammed in between housing, and some demolition would be required to accommodate the Hyperloop.

After Amstelveen, the motorway curves around Schiphol airport, and then passes under a runway, to reach the terminal zone. Again, the curves in this route are too sharp for Hyperloop, and in any case a direct alignment would be shorter. A possible route into the terminal area is a tunnel under the two eastern runways, about 5 km long.

Assuming it exited Lelystad Airport on the west, this alignment could be about 60 km long, airport to airport. That is probably close to the shortest possible surface alignment, but it does mean that the Hyperloop will definitely avoid Amsterdam. The consequences for travellers are evident, by comparing the Hyperloop with the existing railway.

Schiphol – Lelystad by train

Historically there was no railway through Flevoland: it was under the sea. The land was reclaimed mainly to grow crops, but planners did realise that it offered a faster rail route, from Amsterdam to the northern Netherlands. After long delays, the railway through Flevoland was completed in 2012.

Starting from Schiphol Airport Station, which is under the terminal concourse, a four-track line runs toward Amsterdam. It then splits into two routes. One runs to Central Station, and the other to the relatively new southern station, Amsterdam-Zuid. At Weesp, to the east of Amsterdam, the two routes rejoin.

After Weesp, the line is called the Flevo Line: it crosses the lake into Flevoland, alongside the A6 motorway. It passes through Almere, and continues to Lelystad Station. There is no branch to Lelystad Airport at present, so passengers use a local bus. (The Intercity trains continue to Zwolle, Groningen and Leeuwarden).

A branch to Lelystad Airport was proposed here earlier (in Dutch). Operationally, the simplest option is a 7 km north-south branch from Lelystad Station. It is shown in blue on the map (north is on the left). The branch would carry a shuttle train service, which would take about 6 minutes. Existing Intercity services through Lelystad would be unaffected.

This variant could be combined with a west-to-south curve, which would allow trains from Amsterdam to run directly to the airport (shown in yellow). It could be used by a dedicated service between the two airports.

If a branch to the airport was built, without further improvements, journey time between the two airports would probably be 44 minutes. That is 2 minutes more than the present Schiphol – Lelystad time. Obviously, the Hyperloop would be much faster, under 10 minutes if the promised speeds are implemented. On the other hand, the existing train service stops in Amsterdam, and allows interchange with other Intercity services, for instance to Amersfoort and Alkmaar. Connections will also be  improved by the new north-south metro line (2018), which will improve access to the city centre. In Almere there is interchange with regional services to Zwolle and Hilversum. In any case, trains through Schiphol to Lelystad already serve other cities, such as The Hague.

The proposed Hyperloop will not offer any intermediate connections at all. It will offer no through service. It will not take you anywhere, except its two terminal stations, Schiphol Airport and Lelystad Airport. Using this Hyperloop route for longer journeys will carry specific time penalties: exit first mode, transfer, check-in procedure, exit Hyperloop, and transfer to third mode.

Upgraded and new rail lines

The estimated 44-minute train time between two airport stations could be cut, by adding additional rail infrastructure. Some is already under construction: four tracks on the southern route through Amsterdam. Increased line capacity reduces scheduling and operational conflicts, between slow and fast trains. It does not however increase the line speed as such. Both routes through Amsterdam have maximum speeds of 120 km/h, with 80 km/h through some stations and junctions. The rest of the route to Lelystad has line speeds of 130 or 140 km/h.

The existing rail line east of Almere can probably be upgraded for higher speeds, 160 km/h or perhaps 180 km/h. That would cut journey times for a potential airport-to-airport shuttle train, although not spectacularly. A much more substantial improvement was proposed here earlier: a new high-speed line through Flevoland, the ‘Flevo-HSL’. It could also serve as the main route to Berlin, with more high-speed lines (HSL) east of Zwolle. The proposed alignment runs close to the A6 motorway, before turning east towards Zwolle. The Flevo-HSL alignment is notable for passing through Almere in a straight line: it does this by following a high-voltage power line. Trains would not stop in Almere, and the line would bypass Lelystad.

This Flevo-HSL is also compatible with extension to the north-east, to create the long-planned high-speed route to the northern Netherlands. In principle, this proposed Flevo-HSL could run through Lelystad Airport. The question is, whether the airport is important enough to warrant this route variant, and why every high-speed train to Berlin or Groningen should stop there. If only some trains would stop, then a branch line seems preferable.

An additional proposal, which only makes sense in combination with the Flevo-HSL, is a fast exit line from Amsterdam to the Muiderberg bridge, alongside the A1 motorway. This would connect to both routes from Schiphol through Amsterdam. These three proposals – fast exit line, Flevo-HSL, and high-speed airport branch – could cut 15-20 minutes off the 44-minute journey time, airport to airport. That further limits the relative advantage of a Hyperloop.

An additional rail route to Lelystad Airport can be created by extending the proposed airport rail branch south, along the N302 road. This new line would connect Lelystad to Harderwijk, on the old main line to Zwolle.

This would cut journey times from Amersfoort and Utrecht. The proposed Hyperloop can offer no such regional connections, unless it is extended well beyond the original proposal. (And they would have to figure out, how it would stop at intermediate stations).

Inflexible infrastructure

An isolated non-compatible line is inflexible, compared to an existing railway network. That was the main reason to abandon two Maglev proposals in Germany, the Ruhr Express and the Munich Airport Link. And in fact a Maglev train can stop at intermediate stations, but a Hyperloop can not. Vehicles must leave the running line, slow down, and enter an airlock.

Using existing rail technology, Lelystad Airport can be connected to the ‘legacy network’ in three directions. With these new links, trains through Lelystad Airport could realistically serve about 20 stations, without a change of train. Hyperloop, by definition, is incompatible with all existing transport systems, and therefore has no network synergy.

Single lines with new technology make more sense, where there is no ‘legacy network’ – for instance in Abu Dhabi. But what about extending the one Hyperloop to other places? Again this is only rational, if there is no ‘legacy network’. Replacement of the entire Dutch railway network by Hyperloop lines would not make sense, even if someone else paid for it. The Hyperloop is intended to combine ‘low cost’ infrastructure with high speed, and that led to the choice of small vehicles in small tunnels. Inherently, Hyperloop cannot match the capacity of modern heavy rail systems.

Low capacity might be appropriate for airport-to-airport transfers, since this a limited subset of total passenger travel. Even is capacity is sufficient, however, a dedicated airport-to-airport line, with incompatible technology, can not be effectively incorporated into the rail network

Like any isolated line, the selected route is also vulnerable to changes in travel patterns. In this case, the proposed Hyperloop has a very specific function, determined by a specific airport strategy in the Netherlands. That strategy can change. The Schiphol company might lose control of Lelystad, possibly due to European Commission anti-monopoly policies. If the airports split and become competitors, Schiphol might try to grab all traffic, and Lelystad Airport might simply close. The government might also revive old plans for a ‘second Schiphol’, for instance at Dinteloord, between Rotterdam and Antwerp. Outbound air passengers might switch to airports across the border in Germany, such as Weeze and Düsseldorf. All such changes could sharply reduce travel on the dedicated Schiphol-to-Lelystad link, especially if it was an airside-to-airside link.

And finally…

There is no Hyperloop anyway…

The Dutch government has made its plans, or at least its suggestions, on the assumption that Hyperloop wil soon be available for passenger traffic. That is wholly unrealistic. Hyperloop is essentially an unproven technology, with research in the initial stages. The technology has already been significantly altered since the earliest concepts, in response to major technical obstacles. That will probably happen again, before a passenger-carrying prototype starts testing. There are known major issues, such as maintaining a near-vacuum tunnel, high acceleration, uncomfortable ride, noise, and passenger safety and evacuation on failure. There are known operational constants, such as complex switches, and unloading via airlock. The infrastructure, a long metal cylinder on pylons, is untested at 100% scale, and its construction costs are unknown.

All these issues are well documented and well commented. Hyperloop proponents are usually emotionally committed to what they see as a utopian technology, and tend to dismiss criticism. All technical problems, they believe, will be overcome. The heroic status attributed to Elon Musk has a significant role here: if he says it can be done, then his admirers think it will be done. That is not, however, a rational basis for infrastructure planning, in the Netherlands or elsewhere.