Sunday, May 05, 2019

Melbourne's single track rail bottlenecks

Most parts of our rail network have at least two tracks. At least one for each direction.  Then trains can pass in opposite directions without points having to be continually switched to allow access to the only track. That's how it should be in any reliable urban rail system.

However there are also single track sections. Some are within 20km of the city. With other constraints, single tracks make it harder to timetable frequent trains at even intervals, leading to overcrowding on some services.

Reliability is also affected. The train network may still operate but a small delay becomes major if single track bottlenecks cause trains to be held back. This is because safe working rules and the signalling system enforce spacing between trains. Trains may be out of position and recovery can take some time. In the interim, because trains cannot pass they may need to terminate early, bypass your station or not run at all. 

Thick lines on the interactive map below shows single line sections of track on the Melbourne train network. Click on top left to switch on and off rail network and single line sections. Click on single line sections for more details on what may happen to your train if service alterations are necessary. 

Single track sections are appropriate for a quiet rural railway. However they have no place on  modern frequent urban systems. Provided the land is there track duplication can be relatively cheap. Past governments have been slow to duplicate as frequency(*) and reliability don't seem to be highly valued. Freeways, road-rail grade separations, rail electrification, additional commuter parking and even projects with no transport system benefit (eg rebuilds of existing stations) are all seen as more tangible, exciting and vote-winning.  Nevertheless we have had some recent duplications, such as between Heidelberg and Rosanna on the Hurstbridge line. 

What's the outlook for our remaining single track sections? 

Before being re-elected state Labor promised to duplicate the Cranbourne line south-east of Melbourne.  Cranbourne is the only single track section that serves an outer growth area. Growth pressures on it are acute. Especially when failures have knock-on effects on Melbourne's busiest rail corridor, including the proposed Metro tunnel.  

Duplication between Greensborough and Montmorency and around Diamond Creek on the Hurstbridge line is also now government policy. Some form of duplication was promised by both parties as the seat of Eltham was marginal going into the 2018 state election. 

While not growth areas, both the Lilydale and Belgrave lines attract commuters from a wide catchment. The latter has substantial growth potential if feeder buses from Rowville and Lysterfield were improved. 

Areas along the Upfield and Altona lines are undergoing densification and gentrification. This is likely to increase the number of city commuters. A fully duplicated Upfield line has potential to relieve both the over-burdened Craigieburn line and the Sydney Rd tram. The Altona line duplication's benefits are more local but a reduction in the number of people driving to Newport or Laverton in search of better service may offer some relief to the busy Werribee line.  Upfield and Altona both have local activist groups advocating duplication. Their job is made harder by both being safe seats (unlike the marginals on the Hurstbridge line). 

In the meantime, if you want a reliable train service, avoid living near single track sections.

(*) Provider attitudes to frequency vary. Some regard it as something only to be increased when crowding becomes acute and extra capacity is needed to meet demand. Whereas more progressive decision makers with a network vision invest in frequency to lead patronage growth. In the last few years Melbourne has lagged on frequency while Sydney has led. Attitudes towards frequency are one of the key differences between reactive and progressive public transport service planning and marketing. 


Andrew said...

Duplicating the Altona line would be very expensive and tricky.

The line beyond Kororoit Creek was originally constructed by a private company to serve their major subdivision. They allocated the absolute minimum width of land for the railway line, and flanked it beyond what is now Seaholme with two subwidth roads.

For most of the distance between Seaholme and nearly Westona there is not the railway reservation to put a second line. Nor is there even road width on either side to steal to increase the rail reservation. Nor can you close one of the roads as they provide property access.

Beyond Lower Gully the topography is not friendly to duplicate the remainder of the Belgrave line. Beyond the EIS and local residents there is nothing tricky, but it would be expensive.

There would not be the traffic beyond Eltham to justify any duplication.

Simbera said...

Single track sections definitely have no place on a modern urban metro system. I would argue they have no place on a modern regional commuter system either (eg to Ballarat) though certainly they're acceptable on longer-distance regional services (eg to Maryborough)

Anonymous said...

Given the record of the government, I wouldn't be surprised if they went through, duplicated the lines, then ran no additional services.

Mike said...

The easiest way to duplicate the Altona to Westona section would be with Skyrail. But, good luck getting any local residents to agree to that.

P said...

Newport to Seaholme could be duplicated, and Westona to Laverton. That would leave just a three station bottleneck, about 1/3 what it is now.

But I doubt delays on the Altona loop are the reason for most bypasses. Even if it was fully duplicated, it's still much quicker to send the train direct to Newport, so they'll keep doing it when they're running late.

Robbie said...

Duplication would be useful, but it's not always necessary. A lot of lines in Switzerland have single track sections, often with constraints that make duplication really difficult (e.g. lengthy tunnels). Route S4 of the Z├╝rich S-Bahn is mostly single-track, with lots of passing loops, and a rather unusual constraint at the city end where the double track is shared with another route that uses a different electrical system, and yet they are still able to get a ten-minute service frequency during peak.

If we optimised and modelled our railway networks like the Swiss do, we might be able to get more performance out of it without having to build expensive new lines.

Peter Parker said...

Robbie - No doubt we can do better. I think their approach is you design the service and let that drive the infrastructure. So if an operating plan calls for a specific timetable that requires a passing loop in a certain spot you build that. Passing loops work if the service is reliable and lines are isolated.

Unfortunately neither holds true here - our network has daily signal failures, point failures, power faults and train problems. And it can't help that Altona is connected to Frankston - a long line that's near the bottom regarding reliability.

So overall the recipe probably is duplicate everywhere you can, reducing the bottlenecks as much as possible, avoiding where you can't, and step up preventative maintenance (to make things better rather than just keep them sort of going).

Malcolm M said...

Google Earth indicates there is no track that can move a Laverton train from the terminating platform (Platform 3) onto the Altona bypass track. This simple piece of infrastructure would allow the Altona Loop to be run in peak direction during the morning and afternoon peak and return via the bypass track. During the interpeak it could be a shuttle connecting at Newport, as at present, so any delays have a minimal effect on the rest of the network.

Nocknock said...

Upwey to Belgrave will be very hard to duplicate as Tecoma station is inside a valley under a road bridge.

Nocknock said...
This comment has been removed by the author.
Nocknock said...

Funny you did not mention the Alamein Line single track. It may be the last priority to duplicate.