Back to the future – Mainline electrification in Australia?

Recent issues of UK rail magazines have discussed the merits of extending conventional (medium speed up to 200km/h) rail services along the remaining non-electrified main lines in that country. Some of these have recently been announced, including in Scotland, while other iconic possibilities include the Great Western Railway territory, a famous home of fast but conventional trains and deemed for many years, too low density for electrification.

The UK is one of the rail systems in Europe at the low end of mainline electrification reach. While some of the important lengths of the country have been spanned since the 1970s and 1980s, such as both mainlines to Scotland, and the shorter network towards the south east on the third rail system since the 1930s, the UK has manifested some of the anti-electrification habits that we see in this country. While the 1970s gave everyone a shock on energy security and price issues, the UK felt somewhat insulated, for a time, by its oil deposits in the North Sea.

An immediate focus on the developing a medium to high speed option in the Intercity 125 (now class 43), the precursor to the Australian XPT, responded to the demand for a faster intercity service that forestalled, at least for a few years, the possibility of electrification on major routes to the North and West. With a passenger:freight traffic ratio favouring passenger, the UK had no major drive to reduce fuel costs for rail freight hauls, also tending not to push for electrification.

Notwithstanding this, the mainlines to Glasgow, Liverpool and Manchester were electrified in the 1970s, while the north eastern mainlines to Leeds and Edinburgh with some local Scottish routes were electrified in the 1980s. With electrification came some limited attempts to increase mainline maximum speeds above what the class 43 had been rated at, with the north eastern mainline rerated at 225km, grazing the lower limit of what some would call high speed rail. However, this was not feasible on the west coast route which led to attempts to reduce transit time using tilting trains. The first such attempt in the 1970s was a failure, however, tilting trains have since found a home in the UK.

The only challenge to date from high speed rail has been in the SE, linking to France and the extensive continental high speed network. The technology and operating practices for this service have been unapologetically and explicitly Continental. It was not felt wise to be developing a home-grown template for high speed services, such as occurred in Mainland Europe, with separate seeds for high speed in France, Germany, Italy and Spain. Conventional rail electrification in the UK is probably at its technological limits, but not its limits of extent, hence the recent debate and announcements.

In Australia there have been roughly three phases (but not three-phase, as such) to mainline electrification. The first was largely abortive, with planned developments to Newcastle not proceeding in the 1920s. Suburban railways in Sydney and Melbourne were electrified. The principal driver was to reduce the heavy loco operational and maintenance costs associated with steam operation, and increase speeds at the margin, and traffic density, within the suburban passenger area.

The Melbourne network was more extensive and took electrification a long way into the rural hinterland but generally not on the mainline network.

Electric locomotives were built, but only used on local shunting trips. St Albans was the extent on the Bendigo line however no attempt was made to use electric locomotives to haul goods or passengers to the limit and steam beyond, as occurred in Wellington. The same applied to Dandenong.

Broadmeadows, which might also have benefitted from electric haulage (at the top of a long grade) was probably also too close to Melbourne to warrant such a change – and a bypass freight line was built that reduced the consequences of the adverse grade.

Electric motive power was swapped at the boundaries of the electrified network at Lilydale and Frankston. These services were very low density and rural, and offered little practical benefit in swapping motive power. Services to the ends of the low density rural network at Hurstbridge and Upper Ferntree Gully were delivered, in all likelihood to remove the requirement for steam locomotives based at the extremities.

Sydney had a much more limited electric network. Hornsby and Sutherland, like Broadmeadows, sat at the top of a heavy grade and might have benefitted from electric haulage, however, no serious attempt was made to do so.

The other extremities were purely suburban and on easier grades. None was far enough from Sydney to provide any benefit from mainline haulage, as the swap would occur to close to the city centre.

The principal benefit in Sydney’s case of electrification was to enable an earlier start on a city underground than had been delivered elsewhere. As the 1920s came to a close, and economic decline set in, the Sydney and Melbourne networks stood as they would for another 30 years.

The 1950s and 60s saw the second push towards mainline electrification. In both the Victorian and NSW examples, this time the key driver was heavy mainline freights, and the cost of locomotive exchanging was factored into this case. In the NSW case, the extreme mainline grades on the Central Coast and Blue Mountains routes provided the main differential benefit over steam, while in Victoria’s case the expected high traffic densities meritted the work over a dense steam operation.

In both cases passengers also benefitted, although only in NSW was a dedicated set of electric rollingstock provided for the service. This reflected much higher passenger traffic densities on the Sydney lines, while in Melbourne using electric locomotives to haul conventional long distance passenger services sufficed.

The NSW electrifications also provided the opportunity to extend suburban passenger services, in the western case, to Penrith, and in the northern case to the somewhat artificial terminus of Cowan. This pattern continued in the late 1960s as electric coal haulage to Glenlee provided the first electric suburban trains to Campbelltown.

In Melbourne this should also have been possible – except that apart from a few workers trains a short distance past the Dandenong limit, it was not taken up until the mid 1970s.

In some ways the 1950s electrification movement was a false dawn – not waiting to see the benefits of dieselisation, which reduced operating costs by a similar amount without the heavy capital requirement, and in many cases saddling the railways with 30 years of locomotive exchange and slower journey times. Transit times over the electrified section were only incidentally increased.

Only passenger services have endured to this day, driven as they have by urban growth rather than a preference for rail transport. Services beyond Pakenham have been de-electrified. The traffic for which it was electrified has not only reduced, it has completely disappeared. Not only was the electrification technology the wrong tech – but even the traffic was disappearing at the time they thought it was appearing, as people no longer needed brown coal briquettes for domestic and industrial use. Ironically, electricity itself was one of the causes for this displacement of briquette technology, as was natural gas which the government was promoting in parallel.

The 1980s saw the third burst of electrification – sadly, a rehash of the 1950s. This cements for me, how Australian rail planners not only make mistakes, but make the same mistakes again and again. Driven by the same resource boom mentality of the 1950s, the same responses were selected – electrify over conventional rail tracks, swap locomotives at the boundary, which was often a political or operational boundary rather than a natural limit to traffic. Mineral haulage estimates and fuel prices were again exaggerated and not achieved- and the benefits of improved diesel technology pre-empted.

In Queensland at least there was a singular advance, through the use of high voltage AC electrification. This was potentially available in the 1950s however was not quite mature enough in its use in Europe to be available here with few risks. For passenger services especially there was concern about reducing the scale of transforming equipment to fit onto a conventional multiple unit carriage in the early days – issues well resolved by the 1970s.

Queensland also made some attempt to reduce transit times by realignment but this exposed the issue somewhat – that the considerable cost of electrification might have been spent on more realignment, both in Queensland and NSW, and delivered measurable benefits in energy costs (through shorter, quicker, flatter journeys) and other benefits of transit time and reduced loco crewing, without the capital costs of wire and a dedicated fleet of electric rollingstock.

The criticism is particularly damning in NSW. Here, the case for improved passenger services was stronger due to urbanisation. However, lessons from Europe or Japan were not heeded – the availability of dual voltage rollingstock for example, the desirability of major realignments or new routes if real market share in passenger transport were to be achieved. Or the potential to build fast, comfortable and even tilting rollingstock to serve diesel lines (which in one sense already existed with the XPT, misused as it was on long distance slow journeys). A whole generation of 160km/h DMUs such as now exist with the Velocity in Victoria, was a technology available in Europe in the 1980s. Route improvements, such as deviations near Newcastle around the Awaba-Fassifern-Teralba area, or a reinstatement of older routes on the Illawarra, might have been more realistic. In Queensland, no new DMUs have been built since the 2000 class in the 1960s. A belated program of realignments, commenced again in the 2000s, has seen route improvements in Queensland that could have been facilitated 20 years earlier.

Like the 1950s, much of the 1980s has been unwound. Electric traction has disappeared from Emerald (another political and operational outpost, rather than a traffic-mandated one) and the Gladstone Gympie section and the Rocklands to Rockhampton section are down to one electric passenger train a day. The Illawarra network is down to the bare-bones passenger service to Kiama and Port Kembla, and the Newcastle one restricted to the Sydney to Newcastle passenger line.

It would be my contention that if route and transit improvements drive market share and density, and density then drives electrification, we might have had a more virtuous circle of dieselisation leading to route improvement leading to more traffic leading to electrification, rather than the path that was actually followed, which I would call “premature electrification”.

But with the pendulum in the UK, a lag-adopter by European standards, now swinging back to electrification, let’s look at the drivers and challenges for such an approach on conventional Australian mainlines (which I define as routes with general freight and a mix of long distance and local passenger traffic).

Traffic densities are not predicated around distance from capital (nor distance hauled) in Australia. Heavy hauls are seen on the lines north west from Newcastle as far as Narrabri but only one freight a day goes to Bomaderry. You will see a much lower amount of traffic at Virginia, within the suburban area of Adelaide, than you will see at Blackwater in rural Queensland. Even passenger densities are arbitrary, as explains a good service frequency to Lithgow ahead of a poor service frequency to Stony Point, albeit the opportunities to realise density gains are greater on the latter route.

Even the percentage of a traffic’s distance hauled indicates little. For example, using Stony Point as the example, the broad gauge haul from Dynon to Long Island is electrified most of the distance, from the Dynon area to Frankston. As a discrete haul (the load is transhipped) this line could be easily electrified, providing an electric passenger service to Hastings (most of the traffic) and a potentially faster, quieter and less polluting option within the suburbs. Of course, no such service is provided. Even traffics entirely ‘under the wire’ are no longer provided with electric haulage, for example, the Owanyilla woodchips were provided with electric traction to start with, but no longer.

Comparing the UK with Australia we need to examine the following benefits and challenges:
-capital cost
–sovereign debt financing
–construction cost
-operational cost
–fuel or energy
–incidental costs of swapping rollingstock and maintenance
-flexibility of rollingstock deployment
-traffic density
-higher speed of rollingstock
-incidental transit time improvements to route
-improved ‘image’ of electric rail

First, the capital cost is a ‘dead weight’ on any comparison of electric versus non-electric, and puts every other possible improvement ‘behind’ which we saw especially in Queensland. This is made up of two costs – public finance costs and construction costs.

Public finance is subject to a multitude of paradoxes – such as how the UK with a sovereign debt around 100% of GDP is able to offer low interest rates than Australia with only 16% (and with Malcolm Turnbull screaming blue murder about it!). One of the paradoxes is explained by the home country bias: people will lend to their own sovereign at cheaper rates than others, irrespective of the actual risks. That is how Japan has managed to stay in business at all – with Japanese people lending money for free to their governments.

Of course that in turn is answered by two realistic explanations – that people are in fact paying for all the economic stimulation their governments do, to protect their jobs and develop their countries – by subsidising their national borrowings. Secondly, they have some influence, through the political system of their own country on how it is spent and paid back. Equity, rather than debt. No Japanese voter can influence how another country’s government spends or repays.

Unlike Japan, Australia is a debtor nation. It is not the government debt causing interest rates to be above world levels, but private debt. Australia’s private economy is seen as too much of a risk, too dependent on low-value-adding industry and not much depth. And for Australian governments two risks: having to effectively underwrite private debt, but also having limited ability, though the income and consumption tax system, to pay off loans quickly, because our income and consumption tax bases are low by developed world standards. Hence worthwhile government expenditure is deferred, while unsustainable investments in housing construction and company acquisition (reducing the pool of economic talent further) are the preference.

This would explain Australian government’s fetish with being the ‘model debtor’ to the world – to make up for the private sector’s failure to generate sufficient wealth within our borders. A perverse sort of “crowding out” where private sector demand crowds out the public sector.

The long and the short of it is, a total project cost over 30 years at 6% is going to cost Australia far more to finance than the UK, dollar for dollar (pound for pound).

Looking at construction costs, this is a hobbyhorse of great age. The expertise and some of the equipment for electrification is highly traded and mobile. However, at the risk of overlapping with the sovereign debt argument, the real consideration is not traded prices but purchasing power parity (PPP) – what does it cost in terms of local opportunities foregone to pay for these projects, not the actual price.

While currencies move dramatically around conditions and whims (for example, the AUD:GBP rate has gone from 3:1 to 2:1 in a very short time). For actual traded goods, this can provide windfall gains and losses very quickly. For example, a UK electrification contractor might find a move to the Australian market very lucrative when reported in their own currency at present. It might change if we return to earlier levels. However, PPP rates do not change markedly over time because the economic fundamentals and the scarcity costs of all the different goods and services WITHIN the domestic economy do not change quickly. So the general labour cost of electrification workers in Australia is lower than the UK and will remain that way for some time. However, specialist workers are in high demand and Australia does not produce enough for itself – we pay over the odds for these people.

The real cost drivers are the lack of constant work (boosting the fixed costs of design and project establishment, equipment acquisition and recruitment), variable costs not able to be spread over large projects (long km lengths), and our sovereign debt rates.

Looking at fuel/energy costs it may actually favour diesel in the UK, as the electricity costs are higher than here, while diesel is actively traded around the world. I’m unclear on UK taxation of fuels as a business input, so won’t say anything more about it.

From our point of view, any differentials in prices in Australia must be artificial, we are a net energy exporter albeit an importer of petroleum products. I do not see energy price or security issues affecting the equation either way – it would come down to a simple fuel efficiency equation – does the generation of electricity and its transmission to the point of use, being cheaper than supplying that energy as liquid fuel, overcome the cost difference between a locomotive generating its own electricity, and the locomotive which doesn’t and therefore relies on lineside transmission and delivery equipment.

Both energy sources have extensive and established supply chains – liquid fuels via private supply and delivery, electricity via transmission networks. The main difference is the railways have to supply a ‘to the railhead’ electricity supply system, whereas a tank or 2 of distillate at a few depots is all that it takes to keep a diesel locomotive fuelled – the supplier does the rest.

Incidental costs including loco exchange and extra maintenance can be a killer. To minimise the amount of loco swapping – the electrification scheme needs to reach the limits of the traffic, not the limits of a political region or former rail operational base.

NSW readers will recall what happened with Lithgow. The scheme was supposed to reach Wallerawang and the mines in that area, so that end to end haulage (mine to power station or export loader) could take place with a single set of locomotives. However when the cost of the entire scheme got excessive they terminated it at Lithgow, a rail depot that had been located not because of its coal deposits or traffic, but because it marked the end of a difficult section of rail. As well it had become a large town in its own right – making reaching there a political imperative as much as an operational one – but ignoring the actual traffic to be hauled.

For the saving of maybe 5% of the total project cost by truncating the service at Lithgow/Bowenfels (which obviated a tunnel enlargement) the line was condemned to 30 years of needless locomotive exchange to reach destinations only 5% extra distance beyond from the end of the wire.

Too many other examples abounded in Australian electrification – traffics out of unwired sidings or branches which undermined the whole rationale for the electrification. Metropolitan Colliery, Sulphide works, Newstan/Eraring/Vales Point, Maryvale paper, Holmview cattle siding and the Ebenezer Colliery would all be a good place to start. They were unable to even wire to the gates of Goninan in Broadmeadow.

And its never clear just how many sidings are required to be wired. Most yards have a siding or crossover that’s hardly used, but is kept just in case. But unwired, it becomes unusable without a shunting loco kept on site, another cost.

And keeping diesels ‘around’ increases the likelihood they will be used, and the electrification scheme becomes steadily redundant. In Victoria, for example, they turned off the Latrobe Valley electrification every Sunday. With low service levels and a more efficient freight service actually provided on Sunday than during the week (for example, a Sunday Bairnsdale train could run right through without loco change) – the incentive to keep the wired service was reduced.

This also points to flexibility, another opportunity cost of electrification. This is fairly self-evident, but has become all the more so with generic body designs for passenger rollingstock, where the power unit is the only difference between diesel and electric. Think of the Rockhampton and Cairns Tilt Trains, for example.

Freight haulage today will depend on how well electrification maximises the heavy hauls, and the extent to which overhead wiring limits payload space, for example, trailer on flat car or double stacked containers. It is possible to install traction wires at heights that allow for double stacking – it has not been seen as important in Australia to date to do this – and many of the clearance adjustments undertaken in electrification schemes have been ‘barely enough’ to squeeze a pantograph down above the roofline of a conventional locomotive or wagon/carriage body.

Heavy haulage should be the domain of the electric locomotive. It was troubling that the 86 class running on low voltage DC electrification, though notionally rated a more powerful locomotive than the 81 class diesel of the same generation of purchases, was in practice limited to the capacity of the supply system installed. When 2-locomotive coal trains of 30 odd wagons was the norm, the difference was not evident. When trains double that length started to appear, the 4×86 combination was trumped by a 3×81 combination – a 33% improvement in locomotive productivity. As I understand it, high voltage AC locomotives tend to avoid this power drop.

Traffic density is now the key determinant from the need side of the equation. Tractive effort, which used to give electric locomotives a clear lead over steam, is no longer such an advantage against diesel locomotives. Higher speed (facilitated by lighter weight and track impact at speed) only favours electric at the top of the speed range (typically above 220km/h). Underground operation favours electric but with the cost of underground rail many times greater than above ground, any savings from the use electric traction will be marginal to the overall decision.

It is no surprise density is a key factor in a renaissance of mainline electrification in the UK. We are not talking population density, though that is part of it. We are talking traffic density, especially what I would call general intercity traffic – journeys from settlements of all sizes to others for business, for visiting friends and relatives and so on, not merely for the daily work journey or for welfare.

Because this traffic can be very non-specific (not focussed on a handful of large centres like in Australia) it is not easily amenable to replacement by grand high speed rail schemes, as in France. It is closer to the German example.

Any move to a renaissance of long distance passenger rail in Australia would more closely mirror the French example – justified by a few large cities, and the needs of smaller settlements en route or elsewhere would be irrelevant. It is hard to imagine for example, a journey from Wagga to Bathurst being of any great import to a decision to improve passenger rail in Australia. In Germany, however, two cities of 50,000 would be just the market for the sorts of journeys that network does well.

Mainline electrification has seen what I would call incidental improvements in transit speed. This is distinguished from design improvements, which might involve faster locomotives, rollingstock, track or signalling. Some of this can be regarded as consequent on catching up on deferred maintenance.’

It was noticeable that apart from the presence of overhead masts and wire, the 1950s and 1980s electrification schemes showed precious little evidence of the whole rail line being upgraded to meet modern requirements. The track and signalling often was unchanged from before electrification, except in some cases were track recircuiting required the signalling to be modernised.

Incidental improvements in transit time also came from the greater power on the hills, and quicker recovery from stops. OS Nock, in his seminal Railways of Australia, gave the example of the U-Sets in the Blue Mountains running the same timetable up the hill and down. This was not the case with the steam-hauled local trains they replaced!

Higher speeds should be possible from rolling stock, given two advantages of electrification. Opportunities to distribute power throughout the train, reducing axle loads from locomotives. Electrification allows EMU designs to be developed where high power to weight is achieved, compared with diesel equivalents. Higher torque allows power to be picked up quickly and applied to moving the train to cruising speed. Electric locomotives can also provide peak bursts of power, in excess of the continuous power rating, that can be used to start a train rolling to high speed, with lower power required to keep it at high speeds. This mechanism allows TGVs to develop 5 digit kw output for a short time to reach 300km/h, but not have to sustain it as the train has considerable momentum.

The case for electrification is often made by pointing the elusive preference people have for it. I would distill this into a few discrete factors: Electrification is often accompanied by new rollingstock and station upgrades, hence an association in the public mind with it being more desirable.

Because electrification requires higher traffic densities, a plan to increase traffic densities may occur concurrently with electrification, and the two become conflated. For example, post electrification, Nambour, Newcastle, Wollongong, Kiama and countless other places have better services. However, Ballarat and Bendigo also have considerably better services, as does Traralgon, which has more services now than it did when wired. Maitland and Nowra manage with excellent (for their size) diesel passenger services than are unlikely to see electrification.

And electrification may not presage the way to better service. Sunbury is not on the line to Geelong, but Sunbury definitely points the way to Geelong. One is currently unhappy and the other definitely would be. Because electrification has become conflated, in Victorian minds, with an inferior quality of service, with uncomfortable rollingstock, delays, strikes, lack of staff presence and social undesirables. Of course overhead wires and substations have nothing to do with this; but just as a good image can become conflated, so can a bad one. Some rail enthusiasts, who passionately advocate for electric services to places like Geelong, fail to see this.

Finally the environmental argument should win it for electrification, but often doesn’t. Coal fired, especially brown coal fired electricity might only move the pollution from city to powerstation. Renewable sources will clean up the emissions, however an empty train is still wasteful, taking renewable power in the grid from other industries who might have had a better use for it.

The most environmentally friendly thing a train can do is be full. A full train represents car or truck journeys not made, and improves at the margin, ie the marginal wagon or carriage will represent less of an impost on the environment than the ones before it, but draw cars and trucks away that had a constant rate of damage. But an empty train is wasteful. Socially, environmentally and economically.

So how and in what circumstances can Australia benefit from mainline rail electrification?

Given the limitations I’ve mentioned above, some key points are:

-the scope of the electrification must fit around a traffic, not a set of political or operational end points. Ending schemes at the modern equivalents of Lithgow or Rockhampton is doomed. The synergies of Glenlee for freight, and Campbelltown for passengers are the right way to go – the Ebenezer/Rosewood scheme the wrong way to go.

-if the traffic is high value, high volume passenger traffic, then possible mainline electrification is not the way to go, but to build a whole new line, which in probability will also be electrified. Mandurah and Bunbury are pointing the way. If Bunbury is worth having a passenger service, it will be worth having a high speed one, with new fast corridor, dedicated stock and electrification. Not a tacked-on arrangement on the existing route, dodging the slower freights. If it is not worth a new route, it is probably not worth doing.

-traffic density rather than haulage capability or speed are the most realistic options in Australia. If speed becomes make or break, as in high speed intercity rail, then electrification will be favoured. If it isn’t make or break, for example, in the case of Geelong or Sunbury, where rail is already faster than road on point to point (but not door to door) then electrification will make no difference. If however a 220km/h diesel train from Sydney to Canberra was simply not able to beat an aircraft end-to-end, but a 300km/h electric train could – then the case would be clear.

And to justify electrification would require the sorts of suburban densities we see now – out on the mainlines. We may yet see them in some places, for example, between Sydney and Brisbane, or between Sydney and Narrabri/Ulan. If Narrabri/Ulan electrification fortuitously provided a suburban service to Maitland that only required the short section to Telarah wired, then great. However, I can’t see that benefit without coal haulage driving it.

A proper carbon price might skew some of the decisions in favour of electrification, as with peak oil. However, to build the case for rail being a good environmental citizen, it needs to focus on core efficiencies – relieving congestion in the cities, including going underground if need be; hauling the big hauls on the rural mainlines; and possibly providing an alternative to air travel between the closer large cities on the coast. That way, it will kick the big goals against the other modes and develop a market share large enough to justify the electrification that will make a virtuous circle complete.

This entry was posted on Saturday, September 5th, 2009 at 6:59 pm and is filed under Economics, Planning and Operation, Politics and History. You can follow any responses to this entry through the RSS 2.0 feed.