Sunday, November 5, 2023

Cross-Chunnel Comparisons

The High Speed 2 fiasco in the UK has prompted me to update my map of high-speed rail in France.  Obviously the two countries are frequently compared, and here's more in that vein in relation to HSR.

1) Population density is much lower in France than in the UK.  This can be seen by simply inspecting the existing and proposed lines in either country in satellite view, and it is backed by data.  The island of Great Britain is 80,823 sq mi (209,331 sq km) in area and had a population of approximately 64,500,000 in 2021, for an average population density of 798 residents/sq mi (308 residents/sq km).  Continental France is 206,665 sq mi (535,261 sq km) in area and had a population of approximately 65,500,00 in 2022, for a population density of 317 residents/sq mi (122 residents/sq km).  Great Britain is approximately 2.51 times as dense as continental France, and continental France is 39.8% as dense as Great Britain.

2) France is somewhat better at containing urban sprawl than the UK.  This is a qualitative observation on my part, and it isn't uniform across either country.(*)  Areas of France such as the wider region around Lille or the Mediterranean coast from Marseilles to the Italian border are very sprawled out, with little countryside between developed areas except where the geography makes building impossible.  And the UK has compact, relatively contained cities and towns, such as Norwich and Yeovil.  But, overall, I think sprawl is less of a problem in France.

3) The combination of the previous two points means that the French routing strategy of connecting HSR lines with existing urban stations via "off-ramps" to legacy lines works quite well.  The existing approaches to the central stations are fairly direct, and the off-ramps can be built fairly close to the cities.  France generally has avoided building new approaches or stations within existing urban areas.  Lille is the only exception so far.  In contrast, the UK has decided to heavily rebuild existing stations and build new approaches in Birmingham, London (completion uncertain), Manchester (recently canceled), and Leeds (canceled a few years ago).  The prices of those stations has become outrageous and are part of the reason HS2 has been so severely curtailed.

4) France has built a number of stations directly on new LGVs.  Some have been built because there is no classic line near the new HSR line, some because there is only a classic line on one side of a city, and some because the nearest city is too small to serve with off-ramps.  Some stations are connected to the nearest city with legacy rail or a tram system, but others are fundamentally car-oriented, with only limited bus service. Overall, I think the success of these stations is mixed.  A few stations serve an attraction directly, such as the CDG station and the Disneyland station, and these are more successful in terms of passenger counts.  However, the very rural "beetroot" stations are relatively cheap to build, so they may be successful financially.  In the UK, the inline stations at Old Oak Common and Birmingham Interchange are fairly complex and expensive, though they should have high utilization.

5) Paris is semi-central to the country, whereas London is closer to being at one end of the country.  France needs to build a lot more track-miles to fully connect the country, but as explained above, most of those miles were or will be much cheaper to build.  But the main advantage to being central is that the rest of the country does not connect to the capital through just one link that risks being saturated before demand is completely met.  Each link to Paris can absorb a large number of passengers coming from a smaller number cities.  In the UK, the rest of the country will be funneled through one link.

6) On the purely practical side, France has completed 1,699 miles (2,735 km) of HSR track starting in 1981.  The UK has completed only 70 miles (113 km) starting in 2003.  France simply has vastly more experience at everything related to HSR, including track design and construction, rolling stock design and construction, and operations.

7) Both of my maps are fantasy maps.  However, the French one is much more realistic because my suggestions use the same routing principles as the French have been using for decades, and the resulting alignments have been much cheaper to build than HS2 in the UK.

All of the above raises the question: is there any way for the UK to "catch up" with France?  No, not directly.  The UK will never need to build as many HSR track-miles as France.  Nor will it ever be able to build as cheaply as France given the substantially different human geography of the two countries.  The UK could and should learn a fair amount from France, but the former country should also look to places like Japan, South Korea, and Germany for lessons.   Those countries have population densities of 840, 1,340, and 600 residents/sq mi (326, 516 and 233 residents/sq km) and so have more experience in building around existing populations.  The UK also needs to look inward and try to understand why HS2 turned out to be so expensive.  If the UK adapts, it might be able to catch up with France qualitatively, connecting a large portion of its population via HSR, but the UK will need to change a lot to do so.

(*) The relative sprawl of each country probably has been analyzed quantitatively, but since I am an amateur, I don't feel the need to track the studies down.

Friday, October 20, 2023

Fantasy Isle

As a complement to the previous post, I have developed a map of high speed rail in the United Kingdom.  It is a combination of existing lines, lines under construction, projects canceled by the Johnson and Sunak governments, and my suggestions for lines the UK should build in the future.  Those categories are available as separate layers in the map.  A few notes on my suggestions:

1) It is a fantasy map.  A previous version of the map was somewhat more realistic, but I see no point attempting to be serious when the actual UK government has demonstrated it has no interest in a national system.  The Sunak government has even taken a scorched earth approach to the canceled segments of HS2 by authorizing the sale of previously purchased properties.  It has no interest in any future HSR projects, and any Labour government will find restarting the previous program difficult for practical and financial reasons.

2) My system nonetheless could probably built, though not economically.  A few of the lines might be close to economic viability if construction costs are reasonable.  "Reasonable" in this case would be £100M/mile and £1,000M per station.  HS2 has blown way past those numbers to £396M/mile and £4-6B for stations in London and Manchester (now canceled).

3) I have chosen to tunnel the lines into city centers in order to maximize train speed.  Long approaches to existing stations on existing lines would reduce the average speed along those lines.  Building new approaches would be very expensive in dense urban areas, and very disruptive to the existing urban fabric.  I have chosen tunnels for those reasons and because, as stated above, I don't have to be serious here.

4) A lot of England is basically suburban or exurban, similar to New Jersey, especially in the Midlands.  Cities and towns all connect, with very little actual countryside between them.  That is part of the reason the per-mile cost of HS2 exploded.

5) Stoke-on-Trent is a polycentric mess.  The existing station isn't near the center of the biggest constituent city, Hanley.  I have chosen to put a tunneled station under the center of Hanley because more bus routes run through there, but either location might work.

6) I've spaced out most of the stations, but Sunderland is an odd case.  It is very close to Newcastle, and the two are connected by existing rail and a joint metro system that partially runs in the rail right-of-way.   However, as it is near the end of branch, adding one stop wouldn't impact as many passengers as adding a stop between, say Stoke-on-Trent and Manchester at Stockport.  And I feel the benefits of having a station in economically depressed area would be significant.

7) I have not included a station at Heathrow or any other airport (apart from the under-construction station near Birmingham Airport).  The French TGV system does include stops at some airports, but they are generally ones that are outside of built-up areas, such as CDG, making routing to them easy.  Heathrow was well away from built-up areas when it was first constructed, but it has been swallowed by London's sprawl, and constructing a HSR link to it today would be very expensive.  Nor is there a need to build one.  HSR systems are replacements for short-haul flights, not glorified airport shuttles.  With a fully-built UK HSR system, nearly all major English cities would be within two hours of London.  Even the Scottish cities would be well under four hours away, making city center-to-city center times competitive with airlines.

8) I have included some projects such as a second Severn tunnel and a second Chunnel that aren't technically high-speed lines, but which I think will be needed as part of the overall rail system in the future.  The existing Severn Tunnel is already 137 years old, and a replacement will be necessary at some point.

9) The HS1-HS2 link - which was initially proposed as part of HS2 but rejected early on - won't be needed until there is more capacity across the Channel.  Then the limiting factor would be HS1.  The link might be technically impossible depending on how HS2 is constructed.

10) The 2100+ routes are pretty speculative.  Who knows what the state of the world will be by then, and there's a good chance that anything planned now will need radical changes.

That's enough scattered thoughts.  Enjoy the map.

Friday, October 13, 2023

High Speed Halt

In a move that is both eminently reasonable and entirely shortsighted, the UK PM Rishi Sunak canceled the northern section of High Speed 2, a high-speed rail project extending from London to several northern cities.  Specifically, he canceled the section from just north of Birmingham to central Manchester.  The reason stated was the rapidly increasing cost of the project, which had grown from estimates in 2010 of £17B for a ~335 mile-long network to £98B in 2023 for a ~230 mile network.  The network is now planned to be only around 130 miles in length, with a currently estimated cost of about £62B.  The responsibility for completing the final section in the south, from Old Oak Common in the London suburbs to the terminus at Euston Station in central London, has been removed from the purview of HS2 Ltd, the entity that had been supervising the construction.  Work on the section had already been paused earlier in 2023, and will supposedly resume, but the timeline has not yet been clarified.

This move is reasonable at one level, as the explosion in cost from the earliest estimates to the latest figures published this past summer means that the utility to the country in a standard cost-benefit analysis has declined significantly.  There is, of course, little point in building public infrastructure that is so expensive that the cost of servicing the debt undertaken to build the project outweighs the economic gains from whatever service the infrastructure provides.  That applies to water, sewer, road, and airport projects as well as rail projects.  The UK government, which has been run either in part or completely by the Tories since 2010, has been unable to control the costs of the HS2 project, and has decided to limit the project to the section where substantial construction has already been done.

However, the statements by the Sunak government on the changes to HS2 don't explain how it plans to get the costs for the parts of HS2 that (currently) survive under control.  There is no reason anyone should believe that current estimates for the project will turn out to be any more accurate than previous ones.  The Sunak government claims that up to £36B that was going to spent on HS2 from Birmingham to Manchester will be invested in other projects.  I expect that number to be both reduced by further cost increases in HS2 and by future budget issues that will "require" diverting the previously diverted funds to existing programs.

It should be no surprise to anyone that the Conservatives have failed to control their own project.  Fundamentally, the UK Conservative Party - along with many conservative parties elsewhere - does not believe that government can do anything well, and that the private sector is inherently more efficient at everything.  This has been proven wrong over and over again.  The private sector is not reliably competent at delivering complex public works in many countries, including the UK.  But the Tories outsourced not only the actual implementation of HS2, but also a lot of the design and high-level management of the project, to entities that are not working for the public good, but for private profits.  Some of the cost increases have been exogenous (inflation, COVID), but I suspect a lot have been due to complications that come from not having the incentive to make accurate estimates or figure out the least costly way of doing a particular task.

Despite the questionable of utility of HS2 at the current cost, cancelling the northern sections is nonetheless short-sighted.  HS2, at least on the western side of the Pennines, had two purposes. First, it would reduce the travel time between Manchester, Birmingham, and London.  Reduced travel time would increase economic activity in areas affect by the project.  Second, it would free capacity on the West Coast Main Line for local and freight services.  In train operations, mixing high- and low-speed trains reduces the number of trains that can run per hour significantly.  With the high(er)-speed intercity passenger trains removed from the WCML, the number of other trains that could run would increase by more than the number of high(er)-speed trains removed.  But the need for increased capacity on the WCML north of Birmingham will not disappear because HS2 P2b was canceled, nor will Manchester get any closer to the south, where most economic growth in the country has occurred in the past few decades.

The response of the Sunak government should have been to slow the construction down while it examined the root causes of the cost increases and developed in-house government expertise to manage the future portions of HS2.  Stopping and starting construction disrupts supply chains and employment pipelines, as does long gaps between major projects.  High Speed 1, aka the Channel Tunnel Rail Link, was completed in 2007; HS2 construction did not begin until 2020.  Both institutions and individuals will naturally lose competency over periods far shorter than 12 years.  And when there are future projects in the pipeline, companies are less likely to try to milk current projects, as they have incentive to do a good job in order to get the next one.  Now there is no clear future after 2033, when HS2 is (currently) projected to be completed.  Though project names are being thrown about by the Sunak government, they are largely just old plans that haven't gone anywhere, or minor if necessary upgrades.  The UK needs to commit to a national HSR system that it will build out over the next 75-100 years in order to signal to companies and individuals that it is worthwhile to make long-term invests in education and equipment.  So far the Sunak government hasn't done that, and probably won't given that it has decided that pro-car and anti-environment rhetoric is a useful electoral tactic.  What a Starmer government might do is unclear at this point, and unimportant unless Labour actually wins the next election.

Slowing construction of HS2 would have a downside, of course, which is that financing would have to be carried for an even longer time before farebox revenue and economic benefits start to cover debt repayment.  Nor is there any guarantee that much could be done by the government at this point.  A lot of the increases could be exogenous, or they could be political, meaning that the specific details of construction have been (and continue to be) altered to buy off specific opponents.  It may be that too many bad decisions are locked in, and nothing can be done.  But the Sunak government should try, and then should apply those lessons to future projects.

As a distant observer I don't know exactly how the UK returns to a point where it can build rail projects at a reasonable and predictable cost.  But I believe it should.  Otherwise, the country have difficulty increasing productivity as congestion will continue to increase as the population rises.  The UK's recent experience is in contrast to countries like France, Spain, and Italy, where national HSR systems are moving forward at close to estimated costs.  But those countries have been building more regularly, and have been more judicious in their outsourcing.  Whatever the problem in the UK ultimately is, I hope they identify and resolve it soon.

(1) A side benefit of completing the line between Manchester and London would have been the reduction of flights between Manchester and London area airports, some of which are runway slot-limited.  However, I don't believe it was part of the project's cost-benefit analysis.


Wednesday, August 2, 2023

Red-Hot Mapping

In something of a coincidence with the event described in the previous post, I have created a new map which covers nuclear power reactors in North America.  Most of these are in the US, of course, as America is the global leader in nuclear power generation.  But Canada has its own independent nuclear industry centered in Ontario, which lacks Quebec's massive hydropower potential, and I've included those plants.  Mexico also has one lone plant, which it imported from America.

In the period after the 2011 Tohoku earthquake and the subsequent Fukushima Daiichi incident, I published an number of posts on nuclear power.  I've largely lost interest since then because the industry seemed destined to become moridbund, which largely happened.  But in the past few years there has been an increasing amount of hype around "small modular reactors," which promise to solve the problem of the massive up-front costs required to build large reactors such as the AP1000 units in Georgia or the EPR units at Flamanville and Olkiluoto.  On the map I have noted 8 different SMR projects from 7 different vendors as being reasonably firm, given the amount of subsidies flowing to them at this point.  The currently authorized subsidies won't be enough to get the projects to completion, but for now I am assuming the subsidies will continue to flow.  But past history indicates that costs will escalate rapidly, and politicians will balk at some point, most likely after considerable concrete has been poured.

Time will tell, of course, but until then enjoy the new map.

First of a Kind or Dying Gasp?

On July 31, 2023, Unit 3 of the Vogtle Nuclear Power Plant entered commercial service.  It is the first new reactor to be brought online in the United States since 2016, and only the second since 1996.  It is a Westinghouse AP1000, which is a two-loop (of primary coolant) design derived from the Combustion Engineering System 80.  It is classified as a Generation III+ reactor, which is a somewhat arbitrary marketing designation used by the nuclear industry to indicate the design has some passive elements in its safety systems.  An additional reactor of the same design is under construction at the Vogtle plant, and should become operational within a year.

The probability of Vogtle 3 and 4 being safe reactors is quite high.  The fundamental design is not new, and China has operated four AP1000 reactors since 2019 without incident.*  The Georgia plant is located inland, so there is no risk of tsunamis flooding the site.  Georgia Power and the American nuclear industry as a whole have a lot of operational experience at this point - far more than when Three Mile Island Unit 2 was operated incorrectly, causing a major accident in 1979.  The first large power reactor in America went online in 1968; the industry has an additional 44 years of experience since the TMI meltdown.

What the nuclear industry has not improved since 1979 is the cost of constructing reactors.  I am not going go down the rabbit hole of attempting to compare costs between different sources of electricity in this post, but it is very safe to say that Vogtle 3 was not cheap to build, and neither Unit 3 or 4 were as cheap as was promised by Westinghouse back in 2013 when construction started.  Why costs escalated so much is hotly debated, but Westinghouse and its subcontractors are responsible for a good portion of the blame.  The nuclear industry also has to compete with other power generation technologies, and the shale gas "revolution" made the economic case for nuclear power difficult even before the cost overruns at Vogtle and Summer were known.

The cost of large plants is why the nuclear industry has shifted to promoting "small modular reactors" - some of which aren't all that small, and none of which will be any cheaper than large reactors if the companies developing them don't get enough orders to justify building the factories that would enable serial construction using transportable modules.  However, the SMR vendors have yet to prove themselves to be failures, so the hype surrounding them continues.  This contrasts sharply with the prospects for new large reactors.  The Nuclear Regulatory Commission website indicates there are active combined licenses for six more large reactors, but none of those are under construction, and six previous licenses have been withdrawn by the license holders.  Two of those licenses were for AP1000 reactors at the V.C. Summer Nuclear Generating Station in South Carolina, which were so far over budget that the power company building them canceled them in 2017.  So Vogtle Unit 4 is likely to be the last large reactor to go critical in American in the next 15 years, at the very least.

Below is a chart showing all of the commercial power reactors built in America sorted by date of commercial operation.  It does not include the 43 reactors that were canceled after construction was started, nor a number of small reactors that were not intended for commercial operations despite being connected to the local grid.  Despite a number of announcements, there are no SMRs under construction as of today (2023-08-02).

US Power Reactors 1954-2023

* That we know of, of course.  But there has certainly been no major accidents, as those are subject to detection by remote monitoring.

Sunday, January 8, 2023

Say No to Mars on Drugs, Bitches

I've long been a skeptic of the idea of sending humans to Mars before we can determine whether or not there is a functioning biosphere on the planet.  This recent post articulates my thoughts better than I ever could.  Some choice quotes:

At NASA, the faith takes the form of a cargo cult. The agency has persuaded itself that re-enacting the Moon landings with enough fidelity will reward them with a trip to Mars, bringing back the limitless budgets, uncomplicated patriotism, and rapt public attention of the early sixties. They send up their rockets with the same touching faith that keeps Amtrak hauling empty dining cars across the prairie, dreaming of the golden age of rail.

I would compare keeping primates alive in spacecraft to trying to build a jet engine out of raisins. Both are colossal engineering problems, possibly the hardest ever attempted, but it does not follow that they are problems worth solving. In both cases, the difficulty flows from a very specific design constraint, and it’s worth revisiting that constraint one or ten times before starting to perform miracles of engineering.
Even the astronaut corps recognizes that exploring Mars and keeping it pristine are irreconcilable activities, like trying to drill for oil in a cleanroom. The problem goes beyond practical questions like how to store 17 months of astronaut shit and gets to the crux of the matter: why is bringing a leaky, bacteria-filled terrarium to Mars step one in our search for Martian life? What incredible ability do astronauts have that justifies taking this risk?

And it is hard to overlook that the $93 billion NASA has already spent through 2025 to not land anyone on the Moon would be enough to send probes to every world in the solar system, including moons we know have oceans of liquid water and two entire planetary systems that haven’t been visited since Voyager 2 gave them a quick once-over in the 1980’s.

The difficult and unglamorous problems of a Mars mission—how do you wash your socks? What is there to eat?— get no love from Elon. Once you get beyond “rocket factory go brrrrr,” there is no plan, just a familiar fog of Musky woo. The Mars rockets will refuel from autonomous robot factories powered by sunlight. Their crews will be shielded from radiation by some form of electromagnetic handwaving. Life support, the hardest practical problem in space travel, “is actually quite easy”. And of course Musk dismisses the problem of microbial contamination (which I can’t emphasize enough is governed by international treaty) as both inevitable and no big deal.

The arguments against any country or company sending a mission to Mars are overwhelming.  But the subject has become an entirely emotional discussion, like the efficacy of Ivermectin in treating COVID, though with less immediate harm.  Budgetary realities will probably push the problem out well into the 2040s at least, by which time a bit more reality will hopefully intrude and the whole project will be scrapped in favor of more robotics.

2022: At Least It Wasn't 2021, Which Wasn't 2020

2022 was an improvement for most of the world over a bad 2021, which was better than the horrid year of 2020.  But it still wasn't great.

On the positive side:

  • The COVID pandemic subsided to an elevated (relative to flu) endemic level in most of the world, though deaths remain higher than would be the case if vaccination hadn't been politicized in certain countries and nationalized in others.
  • Democrats maintained control of the Senate, meaning Biden will be able to continue making quality judicial appointments.
  • The fascist-sounding COVID promoter Bolsonaro was defeated in Brazil, and he did not attempt to overturn the results by force.
  • The Inflation Reduction Act is not a perfect climate bill, but it was an important step forward.
  • JWST came online and produced spectacular images.

On the negative side:

  • Putin ordered the Russian military to invaded Ukraine for no reason other than he is a bitter old man who remains traumatized by the fall of the Soviet Union 31 years ago.  Tens of thousands of Ukrainians have been killed and millions more displaced as a result.
  • Republicans took control of the House, which will lead to a dysfunctional US government for the next two years.
  • The Supreme Court effectively overturned Roe v. Wade, ending legal abortion and degrading women's health in many states.
  • A petty man-child took over Twitter, a highly influential social media site, and proceed to fire thousands and undermine the communities that depend on the site.

On the worrisome but unresolved side:

  • Central bank interest rate hikes in response to elevated inflation, led by the US Federal Reserve, may help trigger a recession in 2023.
  • Republicans are still threatening to use the debt limit to hold the country hostage.
  • China's overly restrictive COVID policies were dramatically and impetuously near the end of the year by dictator-for-life Xi, which will probably lead to a large surge in deaths and continued disruptions in the world's second-largest economy.