If It Can’t Stand the Heat, You Need to Upgrade the Kitchen:
Repairing the Northeast Corridor’s Electrical Systems
The power systems on much of the Northeast Corridor are outdated and in need of replacement (Image by JP Mueller).
Over the past month, the Northeast Corridor (NEC) has experienced a number of major disruptions, the latest of which put it into such a state of meltdown that it was covered in the New York Times and numerous other news outlets. Train service on both the NEC and the Empire Corridor to Albany and beyond was extensively delayed, in some cases by hours. Collectively, these delays cost riders millions of hours of lost time both on the NEC and on other transit services slammed by an influx of stranded passengers. They also caused untold economic damages, impacted the travel plans of hundreds of thousands of riders, and further degraded the image of transit in the minds of many passengers. There are around 1,000 trains a day operating on the NEC in or near New York; delays like this have very high costs. Travelers—whether intercity riders or commuters—deserve service they can depend on.
Much of the blame for the service cancellation has stemmed from the heat: the disruptions were caused by hot summer weather of about 90 °F, which made the catenary that supplies trains with power vulnerable. While the weather was hot, it was still very much within normal for the summer in New York. 90 °F days are as common during the summer as below-freezing days are during the winter. Instead, the real problem is that the overhead wires and other electrical systems that serve these busy lines are still based on designs from the Great Depression.
All the trains that operate south of Penn Station and into New Jersey are electrically powered from overhead wire, as is the global standard for mainline rail propulsion; for some more background, see ETA’s electrification report. It is critical that the catenary be modernized as soon as possible, with as little disruption as possible for the hundreds of thousands of riders who pass daily underneath it.
To their credit, Amtrak and New Jersey Transit (NJT) have both taken accountability for and put forth a general plan to address the catenary system’s reliability. To do so, however, they need the funding, plans, and political support to rebuild the NEC’s catenary to be heat resistant. New Yorkers and all travelers on the NEC deserve better. ETA strongly recommends that the region immediately begin a program of catenary replacement and electrical upgrades. All told, a renewed focus on good maintenance practices and reliability is key to maintaining high quality service. With these system upgrades, major meltdowns due to normal summer weather will be a thing of the past.
Background
The catenary south of GATE interlocking, located in Queens on the approach to the Hell Gate Bridge, is uniquely old and unsuited to modern operations. In contrast, the catenary from this point north to New Haven not only used different New Haven Railroad designs, but has been or is currently being upgraded by Amtrak and Metro-North, while the catenary north of New Haven was newly built for the Acela in the late 1990s by Amtrak to modern European-derived designs. For that reason, traction power past the northern border of New York City is generally reliable.
The location of GATE interlocking along with the stations to be added for Metro-North’s Penn Station Access (Source: MTA, edited by ETA)
The catenary south of the interlocking to Washington, DC was originally designed by the Pennsylvania Railroad during the Great Depression, and has remained largely unchanged in the near century since. There have been some limited steps towards modernization, including short sections near Secaucus Junction, the new Portal North Bridge, and the New Jersey High Speed Rail Improvement Program area of focus from Trenton to New Brunswick. Although some of these rebuilds allowed for increased speeds, they did not improve the system’s mechanical reliability. Most sections still use designs that are fundamentally from the 1930s, lacking the substantial improvements made in reliability globally over the past nine decades. It is these sections that have caused the several systemwide meltdowns seen in the past few weeks.
The system also has other old equipment on the side of the lines such as circuit breakers and transformers that fail at relatively high rates, and are in need of replacement.
Thankfully, any reconstruction project should be free of significant legal hurdles as, under federal law, catenary replacement qualifies for a categorical exclusion from environmental review.
The Heat: Not Unusual
The infrastructure on the NEC has long struggled with summer heat. None of this is new, however, it should not be acceptable given the prevalence of these temperatures.
The weather that precipitated the latest series of issues were not at all unusual for a New York summer. The highs were 90 °F, not much higher than the 1991-2020 daily July high of 85 °F. During the last few weeks, the average highest temperature reached was around 90 °F, which regularly occurs during July.
We can and should expect our infrastructure to deal with normal weather variations. Consider the fact that the difference between 85 °F and 90 °F is slightly smaller than the difference between the daily January high, 39 °F, and the freezing point. Most winter days have above-freezing daytime highs leading to snowmelt. Many days, however, do not, and therefore snow accumulates on the ground every winter. Because this happens frequently, New York City cannot afford to shut down, and so it is well-prepared for snow accumulation. Compare this to a warmer city like Washington, DC (January high: 45 °F) where snowfall is significantly less likely, and therefore significantly more likely to cause major disruption.
Vital transportation systems need to stay reliable during normal weather events. The catenary used on the NEC in and south of New York, however, is based on obsolete technology, limiting both the maximum speed of intercity trains and the reliability in the heat, as borne out a recent service stoppage between New Haven and New York due to a faulty circuit breaker.
Catenary Tensioning System
Catenary wire needs to be properly tensioned to function. This can be done in two ways: the simpler one is called variable tension or fixed termination; the other, more advanced one is called constant tension or auto-tension. Most of the catenary in the United States has variable tension. While variable tension systems are simpler to build, they can only operate within a specific temperature range, as metal expands in the heat and contracts in the cold; constant tension installations resolve this problem. Another problem of variable-tension catenary is that it limits trains to 135 mph in its most common configuration, with a few recently upgraded sections allowing slightly faster running. This limitation prevents pressure waves that moving trains create in variable tension wire from damaging the catenary or causing a train to lose power. On much of the NEC, including the entire section north of GATE and more recently part of the New Brunswick-Trenton section, Amtrak and Metro-North have installed constant-tension catenary, and there have been proposals to outfit the rest of the corridor with it, in large part because it mitigates the pressure waves. Most railroads that use variable tension catenary only run trains under it at much slower speeds than are possible on the Northeast Corridor, comparable to those at which NJ Transit diesel trains run on outlying lines.
Variable-tension catenary involves fixing wires to every support structure. The following figure, taken from Garry Keenor’s Overhead Line Electrification for Railways 6th Edition, illustrates the view from the track side, with the bottom line representing the contact wire:
Between each anchor point, the wire sags in the heat and rises (hogs) in the cold. What’s more, metal catenary wire often ends up considerably hotter than the air. The resultant deformation increases the risk of pantographs felling wires or becoming damaged themselves, both of which require major repairs that stop service. In London, summer heat reaching about 90 °F in 2011 led the catenary wire to record 167 °F temperatures, leading to train cancellations as the wire lost tension. This episode led to the installation of constant-tension catenary, as shown in the figure below, taken from Keenor’s book as well:
Constant-tension catenary only anchors each stretch of wire at its midpoint. The wire slides along all the other supports as it expands and contracts. Weights or spring tensioners at each end maintain the wire’s tension. Spring tensioners are usually preferable for lines with three or more tracks, as they avoid the need to run wires for one track across those of another to reach the side of the right-of-way, where weights are usually located. While constant tension catenary requires more complex engineering, the wire’s performance does not suffer in the heat or cold, and the maximum speed is therefore higher. Constant-tension catenary is not something at all unique to high-speed rail: recent installations have sold in countries with no trains faster than 125 mph.
ETA strongly recommends that the region quickly develop a plan to convert all of its catenary to constant-tension, including that installed on non-high speed lines. Compared to the current best possible travel times, constant-tension catenary saves intercity travelers 10 minutes between New York and Washington; accounting for padding that would likely not be necessary with a modern traction system, it saves more. Constant-tension catenary is thus a necessary infrastructure upgrade and should have been completed long ago.
Headspans, Gantries, and Cantilevers
There are three ways to suspend catenary wire above track: headspans, gantries (also known as portals), and cantilevers.
.jpg){kind=link}
Source: Kara Fischer, ETA
Headspans and gantries differ in that headspans suspend the catenary from flexible wires, whereas gantries do so from fixed metal structures. Gantries have the ability to provide mechanical independence: if one track’s catenary wire fails, it doesn’t impact the other tracks’ power supply, and it’s possible to do maintenance on one track’s wire without disturbing the others. While simpler to install, than gantries or cantilevers, headspans can not furnish mechanical independence. They are not reliable enough for this reason. Cantilevers can be installed either to the side of tracks or between a pair of adjacent tracks, suspending wires above each on independent hangers; this provides mechanical independence, like gantries, at the cost of requiring more spacing between tracks than the other options.
Bay Area advocate Clem Tillier has argued against headspans for Caltrain and California High-Speed Rail electrification and in favor of cantilevers; importantly, the largely two-track Caltrain corridor has room for cantilevers between tracks, but the NEC does not.
We recommend a combination of single and twin track cantilevers as well as gantries for the NEC. Depending on the soil conditions, track outage availability, and number of tracks, Amtrak and NJT would need to decide which of these solutions make the most sense in each area. This video from the Permanent Way Institution describes some of the various tradeoffs of the different styles of structures we are recommending. British engineer Noel Dolphin has written about a method of converting headspans to gantries that minimizes multi-track outages.
Power System Controls
Upgrading the fault handling systems such as switches and circuit breakers on the south end of the NEC with modern equipment would allow quick isolation of electrical faults and permit restoration of service on unaffected tracks within seconds.
There is also the issue of voltage and total power delivered. The current voltage from New Haven south, 12 kV, is less than the standard 25 kV, and on the margin, this reduces the ability of high-speed trains to run at full power, at the cost of requiring a few inches of extra static clearance. Amtrak has occasionally planned converting the catenary power to 25 kV; we advise it to ensure that any infrastructure changes are compatible with future conversion to 25 kV. For example, engineers should ensure all insulators, both those located in switchyards and those above tracks, are rated for 25 kV. Additionally, Amtrak should add additional circuit breakers to the 138 kV transmission network on the NEC to provide better fault handling capabilities, as the network has very few circuit breakers.
Maintenance Practices
Until Amtrak can renew the catenary system south of GATE interlocking, it is vital for the agency to follow-through on its recent announcement and use its automated catenary inspection trains and catenary maintenance cars to inspect the system far more frequently, especially during periods of hot weather. Additionally, NJT should also increase the pace of its pantograph inspections and aggressively repair or replace defective ones.
Brush and tree pruning projects are also critical, as during periods of high winds, and other storms, trees, and branches can hit or fall on the catenary or fall on it causing outages and extensive damage.
Making the Most of Weekend Closures
Amtrak and NJT currently schedule NEC service through the North River Tunnels to permit single tracking overnight and on weekends, providing a weekly 55-hour maintenance window for one tube at a time. The tunnels’ two tracks are in two separate bores, and thus maintenance on one track does not disturb the other, unlike on systems where two tracks run closely together such as on the surface, or on most of the New York City Subway.
Unfortunately, Amtrak frequently does not make use of these weekly maintenance windows. Reporting by Michael Aronson at the New York Daily News has found that a full weekend closure is used only once in three months. During 12 weekends out of 13, maintenance work is not actually conducted.
To improve reliability, Amtrak must step up its maintenance tempo and use the existing windows every week it can. In 2019, the Gateway Program commissioned an independent review by London Bridge Associates, which determined that Amtrak could entirely overhaul the North River Tunnels on nights and weekends in approximately 31 months each, or slightly more than five years in total.
The overhead power, physical track, and signaling systems of the North River Tunnels have significantly deteriorated, leading to significant delays. By increasing the speed of repairs, Amtrak would not only increase the reliability of the system for the riders and travelers of today, but would also allow a full, four-track route into Penn Station to open as soon as the Gateway Program’s new Hudson Tunnel Project opens in the 2030s, rather than years later as currently planned.
Hoboken, Grand Central, and Redundancy
In the meantime, there is already considerable redundancy on the NEC. By using what exists smartly and planning a moderate amount of future capital funding, agencies can give passengers options to get around around further failures in (or leading to) either the North or East River Tunnels without the hassle and expense of emergency rail replacement buses, taxis, or rideshare apps.
Grand Central and Metro-North
Trains from points north of New York can go into either Penn Station or Grand Central. All intercity trains currently use Penn Station, as Grand Central’s third rail electrification is not compatible with Amtrak’s intercity equipment without making changes to the third rail power setup in the shop. In the event of significant service interruptions, Amtrak should work with Metro-North to allow passengers from points north to connect onto commuter service to finish their trip into or out of New York. Such a service would be best served by having Amtrak trains continue not only to New Haven, as they did last weekend and previous weeks, but to Stamford or even New Rochelle, as well as permitting ticketed passengers to travel on Metro-North for free, just as they permitted passengers to cancel their heavily delayed bookings for free. Stamford is a better transfer point than New Haven, since Metro-North serves it at higher frequency, and also provides fairly fast, non-stop trains to Manhattan, whereas off-peak commuter trains serving New Haven make all (or at least most) local stops all the way down to Stamford.
In the near future, Metro-North will open Penn Station Access (PSA), currently scheduled for late 2027. This project is primarily designed to improve access from the Bronx to Manhattan from four new stations, such as one at the currently underserved Co-Op City. The southernmost station that will open as part of PSA, Hunts Point, is adjacent to the 6 train on the subway, giving passengers another transfer option in case of disruption.
The MTA should also reinvestigate building another station farther south, at the intersection of the line with the N and W trains at Astoria-Ditmars. Such a station was studied and rejected in the original PSA planning, partly due to construction difficulties, but also partly due to an assumption of low frequencies and high fares, which would make it uncompetitive with the subway. However, as ETA documented in our commuter rail modernization report, an infill station at Astoria would serve to connect northwest Queens to a regional transportation network. It would also provide another valuable, redundant option for long-distance travelers from the North in the event of disruptions in the East River Tunnels.
Going forward, ETA also recommends that the MTA study and complete initial design for the installation of catenary all the way to Grand Central, so that trains can divert there in the event of major work on the Hell Gate Line, or grants become available to install it to increase redundancy. While there are tight clearances within the Park Avenue Tunnel, it does accommodate Amtrak’s Northeast Regional and Acela equipment. Amtrak already sometimes uses Grand Central for special events such as National Train Day or disruptions in the Empire Corridor to Albany, using diesel or dual-mode locomotives. With catenary, commuter trains could run into Grand Central without needing third-rail shoes, making it easier to buy new trains for the line, and Amtrak could run directly into Grand Central in cases of Penn Station disruption. Finally, to allow this to happen, all Amtrak crews operating in Metro-North territory should be certified to operate into Grand Central station if necessary on short notice.
Hoboken and NJT
Trains entering New York from the south have no direct alternative to Penn Station, unlike ones from the north, which have Grand Central. However, both Newark Penn Station and Hoboken are close enough to the city to allow PATH trains to take passengers into and out of Manhattan.
Currently, all Amtrak trains in revenue service on the NEC can run to Hoboken Terminal. NJT is retiring the last of its cars that can not make those runs, the Arrow III stock. NJT built the Waterfront Connection in 1991, which permits NEC trains to divert to Hoboken, shortly before the much better-used Kearny Connection, which permits the Morris and Essex (M&E) Lines, which previously could only serve Hoboken, to go directly into Penn Station. Unsurprisingly, the Kearny Connection sees much more ample use, whereas in normal service, only a single train per day uses the Waterfront Connection, a Raritan Valley Line diesel.
But while the Waterfront Connection could provide NEC redundancy, there are some problems that limit the ability of Amtrak and NJT to use Hoboken as an alternative to Penn Station during disruptions. First, Hoboken only has low platforms. While all NJT trains can handle them, as can Amtrak Regional trains which serve low platforms at some lower-use stations in Maryland and Connecticut, the Acela can only use high-platform stations. Until Hoboken is rebuilt with high platforms, during disruptions Acela trains would have to terminate at Newark. Thankfully, a project to add high platforms is already in the contracting process as part of the Hoboken upgrade plan. As at Grand Central, Amtrak crews would also need to be certified to operate into Hoboken.
The Waterfront Connection is currently single-track, as demand during normal operations is low. Given this light usage, trains could platoon through this short section of single-track during disruptions, but a second track is necessary to provide enough capacity to relieve Penn Station. NJT’s capital plan includes a potential double-tracking project, for $474 million. The organization should re-evaluate the cost of this project: in the 1990s, the Waterfront Connection cost $16 million, and adding a second track does not require more scope than building the first, nor have construction costs increased 30-fold since 1991.
Teaming Up, During Disruptions and Every Day
Recent reporting suggests that Amtrak and NJT are exploring the use of NJT Electric Traction Personnel during outages to speed up repair times. We encourage Amtrak, SEPTA, NJT, and Metro-North, along with their various unions, to develop a labor compatibility program, and any necessary legal agreements, training, and financial agreements necessary to permit not only speeding up construction of critical projects using all available resources from the transit agencies in the region, but also ensuring that the closest certified personnel are able to rectify any outages.
Conclusion
Maintenance and operations on the Northeast Corridor can be a complicated affair, as different sections are owned by different organizations, and the line is used by many different operators. Ultimately, however, riders don’t know or care who is responsible for the tracks their train is using, they simply want to get to their destinations. To that end, every operator using the NEC needs to do their best to coordinate not only on implementing these improvements quickly, but to resolve incidents quickly when they happen.
It should go without saying that massive failures like those seen recently on the NEC are unacceptable—especially when they are caused by weather that is normal for the region. These major outages delay or strand hundreds of thousands of travelers, give transit a bad name and cost individuals and the economy at large over $100 million a day.
Amtrak and NJT are taking responsibility for these failures, which is a good first step. Ultimately, however, as they have stated, these agencies need to make necessary upgrades to their electrical systems. When built and maintained to the right, modern standards, electric trains are perhaps the most reliable mode of transit on the planet today. The New York region can no longer accept Great Depression-era wire designs, nor should it have to wait for a new set of Hudson tunnels to open for improvements to begin. It is long past time for us to upgrade our catenary and other related electrical systems and make disasters like these a thing of the past.