Acela 21 (Avelia Liberty) development, testing and deployment (2018 - 1Q 2024)

[AmtrakMaineiac]

Now add management, pointy haired or not, purchasing and contract agents with checkbox forms, risk-averse lawyers, and the like, and money, and technicians and engineers. That's my guess.

Also don't forget politicians with itchy ribbon cutting fingers.

Out of curiosity, are the US crashworthiness standards that much more onerous than those in Europe? I often see those standards cited as one of the reasons why Amtrak cannot buy train sets that are closer to their off the shelf European counterparts.

One would think that, given Europe‘s extensive history with HDR, they would have developed sensible crashworthiness standards. It certainly appears to the casual observer that their standards have, by and large, kept European passengers safe for decades. Does anyone have a view on whether our more stringent standards provide meaningful incremental safety benefits? Or is it more of a not invented here thing?

Of course Europe doesn't run the kinds of heavy freight trains we have which is a factor.

 

[Dutchrailnut]

not just freight but average locomotive een if ballasted in Europe weighs about 80 Metric ton. compare that to any American Locomotive or like a Genesis at 275 000 Lbs. (125 metric ton) and you find its about equal to Smart vs a F150 pickup truck

 

[Dutchrailnut]

meanwhile in upstate NY

 

[jiml]

meanwhile in upstate NY

What a beautiful shot.

 

[eldomtom2]

Out of curiosity, are the US crashworthiness standards that much more onerous than those in Europe? I often see those stamdards cited as one of the reasons why Amtrak cannot buy train sets that are closer to their off the shelf European counterparts.

I believe the primary difference is that US crashworthiness place/placed much greater emphasis on buff strength, with much higher requirements in that area compared to Europe, where there is more of an emphasis on things like crumple zones.

 

[GDRRiley]

I believe the primary difference is that US crashworthiness place/placed much greater emphasis on buff strength, with much higher requirements in that area compared to Europe, where there is more of an emphasis on things like crumple zones.

While thats still an option (T1) we now also allow lightly modified euro spec as T1 Alternative and T3 for high speed trains

 

[jis]

While thats still an option (T1) we now also allow lightly modified euro spec as T1 Alternative and T3 for high speed trains

Additionally, it should be noted that the safety cage protected by CEMs in T1 Alternate has the same buff deformity (or lack thereof) requirement as for the entire body of the car in T1. T3 is similarly related to T2 IIRC.

 

[Acela150]

Monday night (9/11) KP2022 was out with one of the new sets.

Tonight (9/13) KP2022 is taking the First Production set (PC 2108/2109) back to Hornell. No idea if another will replace it in Philly.

 

[AmtrakBlue]

Monday night (9/11) KP2022 was out with one of the new sets.

Tonight (9/13) KP2022 is taking the First Production set (PC 2108/2109) back to Hornell. No idea if another will replace it.

I think I read TS9 will be sent to Philly

 

[Acela150]

I think I read TS9 will be sent to Philly

You were correct!

TS9 with PC's 2150 and 2151 came down.

 

[Fenway]

WCVB-TV Boston

 

[GDRRiley]

https://amtrakoig.gov/sites/default/files/reports/OIG-A-2023-013 (REDACTED)_0.pdf

Trainset designs have not yet met FRA requirements. Federal regulations require the company to submit to FRA trainset performance predictions from a computer model showing that the model is valid and the trainsets are safe to proceed with required testing.2 Although the company must submit these predictions to FRA for approval, the vendor is responsible for developing and validating the model. Contrary to leading practices,3 however, the vendor started producing serial trainsets4 before it validated the model. Now, three years after starting serial production, the vendor has built more than half of all units, but has not yet validated the model, which is the first step in a multi‐ step regulatory testing process. Company officials told us they did not object to the vendor proceeding into production because, among other things, they were overly optimistic about the vendor’s progress and believed the vendor’s assurances that validation was close. Without a validated model, FRA will not let the company move forward with the rest of its required testing. Until the company completes testing, the trainsets cannot operate in revenue service—let alone at the advertised speed of 160 miles per hour.

Trainsets have defects, and the schedule for addressing them is incomplete. As of July 2023, the vendor has produced 12 of 28 serial trainsets and 22 of 28 café cars, all of which have defects—some that require structural and design modifications and others that require sealant, drainage, or corrosion corrections. Although some defects are expected when producing a new trainset, the vendor’s schedule for addressing them is incomplete, and the company has not been successful in obtaining more complete scheduling data from the vendor. Without more complete information, the company cannot verify whether remediating the defects will impact the overall program schedule and the revenue service launch

P2-3

From August 2016 through July 2023, the vendor produced at least 14 different iterations of the trainset model, and none have met FRA requirements. Vendor officials told us the models have not met requirements because, in their view, 1) track conditions on the NEC are challenging, and 2) trainset modeling compliance guidance is ambiguous. the vendor cited variances in company‐provided data about NEC track conditions—commonly called track geometry—as a contributing factor to its ongoing challenges.19 The vendor, however, was a partner in the consortium that produced the legacy fleet in the early 2000s and has been performing maintenance on them for more than 20 years; therefore, it should be familiar with the NEC’s track geometry and how the legacy fleet—which has a tilting system—performs on it.20 Moreover, officials from the company, FRA, and DOT said that the company has consistently provided the vendor with track geometry data. Further, a subject matter expert at DOT told us it is common for track geometry data to have variances, and that any variances in different sets of the track geometry data collected at different points in time would not materially affect model validation testing outcomes. Ultimately, the contract specifies that the vendor must comply with the modeling regulation and that the New Acela trainsets will operate on the NEC.

 Vendor’s compliance with FRA guidance. Vendor officials told us the other reason it has not yet met FRA requirements is because they believe that the regulation governing model validation is ambiguous about what is ultimately needed. Vendor officials also told us the FRA requirements related to computer modeling are significantly different than how they verify trainset safety in Europe where the vendor is based. FRA and DOT officials said they have been working closely with the company and the vendor to answer questions and help them comply with the requirements, including demonstrating a step‐by‐step process on how to comply. For example, as early as October 2017, the DOT subject matter expert outlined in detail to the company and the vendor the steps they needed to take to comply with the requirements. Additionally, company officials said that as early as July 2021 and many times since then, they offered to assist the vendor in its model validation efforts. For example, company officials said they offered to troubleshoot challenges in developing the computer model itself within the software the vendor uses, including an in‐person session in Montreal, Canada, in fall 2021. In January 2023, the vendor’s Chief Executive Officer, the company’s Chief Executive Officer, and other executives began meeting regularly to discuss the model validation process and other issues related to maintaining the legacy fleet. Despite these offers to assist the vendor and the recent executive‐level discussions, the vendor has denied the company and DOT subject matter experts direct access to the trainset model input files, according to program documents and officials from the company, FRA, and DOT. During a March 2023 interview, vendor officials told us the contract does not require them to share the trainset model. Vendor officials said that the input files were their intellectual property and sharing them would risk disclosing proprietary information.

pg 13-14
This report is not pulling any punches, Alstom is completely failing to deliver the product. I'd like to see Alstom be forced to pay for the costs of continued upkeep of the legacy fleet and provide several trainset as compensation for how late the delivery is.

 

[AmtrakMaineiac]

Looks like this is turning into a real dumpster fire. They can't even start the real testing yet until the computer model is validated. Then there are multiple defects that need to be resolved, since they went ahead and built several trains even though testing had not begun. The plan that they can solve all this in time to put these trains into service starting in mid 2024 is looking very optimistic.

 

[MARC Rider]

Now that they have actual trainsets that can be used as test articles that can be modified as needed, please explain why they need to rely on computer models. OK, it's in the contract, but why did the contract require them? Also sounds like Amtrak made a mistake by allowing them to build additional trainsets before the models were validated.

 

[jis]

Now that they have actual trainsets that can be used as test articles that can be modified as needed, please explain why they need to rely on computer models. OK, it's in the contract, but why did the contract require them? Also sounds like Amtrak made a mistake by allowing them to build additional trainsets before the models were validated.

I am almost certain that you need to ask FRA, Alstom and Amtrak to get the real answer and I am sure none of them will feel obliged to provide them at this time. It is almost certain that no on here one AU really knows what is going on. Heck we don't even know what the test and acceptance protocol is that FRA, Alstom and Amtrak agreed to, let alone what broke exactly. We vaguely know that there is a problem with tracking model and perhaps even performance. From what I have heard they could run them at less than 100mph . But they have to resolve stability issue above that. Eventually I am sure the OIG will produce another scathing report baring all at some point.

As an aside, riding anything above 125mph on the NEC has always seemed to me like riding a roller coaster of sorts when compared to rock solid steadiness everywhere else in the world. Maybe we should just stick to 125mph on the NEC and stop wasting money on trying to put adequate number of layers of lipstick on the pig and get on with life?

 

[GDRRiley]

Looks like this is turning into a real dumpster fire. They can't even start the real testing yet until the computer model is validated. Then there are multiple defects that need to be resolved, since they went ahead and built several trains even though testing had not begun. The plan that they can solve all this in time to put these trains into service starting in mid 2024 is looking very optimistic.

Its been a dumpster fire for a long time now, we just know the full extent

Now that they have actual trainsets that can be used as test articles that can be modified as needed, please explain why they need to rely on computer models. OK, it's in the contract, but why did the contract require them? Also sounds like Amtrak made a mistake by allowing them to build additional trainsets before the models were validated.

the best way to understand how the trucks are interacting and will over the entire lifespan of components being worn is via computer modelling. FRA requires it, they knew they had to do it from the start and have attempted but failed 14 times so they just don't want to do it anymore. They've also refused FRA, amtrak and other experts help which has been offered sense 2017

 

[west point]

Over 27,000 items to be fixed on 6 train sets? Cannot even wrap my arms around that number.

 

[Septa9739]

As an aside, riding anything above 125mph on the NEC has always seemed to me like riding a roller coaster of sorts when compared to rock solid steadiness everywhere else in the world. Maybe we should just stick to 125mph on the NEC and stop wasting money on trying to put adequate number of layers of lipstick on the pig and get on with life?

I agree with this. The USA have wasted billion after billion on track and signal upgrades, paid for increased maintenance, wasted billions on rolling stock that apparently can’t roll, so that we can have a high speed rail system akin to, I am often told, non industrialized former Soviet protectorates in Asia (nothing against them). Note WAS-NYP trip times are still comparable to 1960‘s schedules. It would have been better all along if we prioritized removing slow orders, straightening curves like Frankford Junction, that slow all freight, commuter, and intercity service, and addressing resilience issues than chasing that one minute savings for getting to the magic 160 for 20 miles. When Gunn fixed the Harrisburg line, it garnered no big headlines and set no records. What it did do is leave PA with a competitive, reliable service that gets people where they need to be safely, cheapishly, and punctually. If Acela 21 were about applying that same principle to premium service on the NEC, we’d be much better off.

 

[blueman271]

Nobody botches procurement contracts quite like Amtrak. Understand this isn’t entirely their fault but my god are they snake bitten.

 

[west point]

Many if not all these track problems can be traced to Amtrak not performing proper undercutting of the tracks over the years, There is so much junk such as soil, vegetation, seeds, fines, etc that make the effort to stabilize the tracks almost impossible. Instead Amtrak has to surface the tracks much more often. Amtrak did not even have a undercutter that could perform work on a track without closing adjacent tracks. That I believe has been corrected. Have no idea how deep the undercutting needs to be to get stability much better?

 

[user 13416]

I agree with this. The USA have wasted billion after billion on track and signal upgrades, paid for increased maintenance, wasted billions on rolling stock that apparently can’t roll, so that we can have a high speed rail system akin to, I am often told, non industrialized former Soviet protectorates in Asia (nothing against them). Note WAS-NYP trip times are still comparable to 1960‘s schedules. It would have been better all along if we prioritized removing slow orders, straightening curves like Frankford Junction, that slow all freight, commuter, and intercity service, and addressing resilience issues than chasing that one minute savings for getting to the magic 160 for 20 miles. When Gunn fixed the Harrisburg line, it garnered no big headlines and set no records. What it did do is leave PA with a competitive, reliable service that gets people where they need to be safely, cheapishly, and punctually. If Acela 21 were about applying that same principle to premium service on the NEC, we’d be much better off.

Seconded. I think there's been too much of a push to "catch up" to European rail service which works very well for Europe but can't necessarily be shoehorned into the reality of the US rail network. We have to work with what we have, and with decisions that were made 200 years ago.

 

[GDRRiley]

Over 27,000 items to be fixed on 6 train sets? Cannot even wrap my arms around that number.

on 12 not 6 trainsets, and in some cases they've made nearly all the cars like 22/28 cafe cars

I really hope alstom as a result is forced to provide a few extra trains, longer and better support plus pay for the cost to maintain the legacy fleet longer

I agree with this. The USA have wasted billion after billion on track and signal upgrades, paid for increased maintenance, wasted billions on rolling stock that apparently can’t roll, so that we can have a high speed rail system akin to, I am often told, non industrialized former Soviet protectorates in Asia (nothing against them). Note WAS-NYP trip times are still comparable to 1960‘s schedules. It would have been better all along if we prioritized removing slow orders, straightening curves like Frankford Junction, that slow all freight, commuter, and intercity service, and addressing resilience issues than chasing that one minute savings for getting to the magic 160 for 20 miles. When Gunn fixed the Harrisburg line, it garnered no big headlines and set no records. What it did do is leave PA with a competitive, reliable service that gets people where they need to be safely, cheapishly, and punctually. If Acela 21 were about applying that same principle to premium service on the NEC, we’d be much better off.

The constant tension wire replacement has been overdue, it should have been fixed decades ago the existing system is coming up on 100 years old now.

Seconded. I think there's been too much of a push to "catch up" to European rail service which works very well for Europe but can't necessarily be shoehorned into the reality of the US rail network. We have to work with what we have, and with decisions that were made 200 years ago.

Working with what we have will never get us a good enough rail network. We use to do major realignment projects quite often till the 40s and 50s, we built the entire interstate highway system in 70 years. We must upgrade and build new lines.
Some railroads clearly care enough to make upgrades work, BNSF is tipple tracking + grade separating more and more of their southern transcon while LA-Bartow is going to have 3-4 main tracks

 

[user 13416]

Working with what we have will never get us a good enough rail network. We use to do major realignment projects quite often till the 40s and 50s, we built the entire interstate highway system in 70 years. We must upgrade and build new lines.
Some railroads clearly care enough to make upgrades work, BNSF is tipple tracking + grade separating more and more of their southern transcon while LA-Bartow is going to have 3-4 main tracks

By "work with what we have" I mean that rail upgrades don't happen overnight, and those upgrades have to be backward-compatible so today's trains can run on both old and new infrastructure, and that new trains also have to be able to run on yet-to-be-upgraded tracks. That's why we can't just buy European off-the-shelf equipment. It has to exist on our rail network that is a cobbled-together mess of brand new, decades-old, and pre-Civil War rights-of-way.

 

[Septa9739]

I really hope alstom as a result is forced to provide a few extra trains, longer and better support plus pay for the cost to maintain the legacy fleet longer

The constant tension wire replacement has been overdue, it should have been fixed decades ago the existing system is coming up on 100 years old now.

Working with what we have will never get us a good enough rail network. We use to do major realignment projects quite often till the 40s and 50s, we built the entire interstate highway system in 70 years. We must upgrade and build new lines.
Some railroads clearly care enough to make upgrades work, BNSF is tipple tracking + grade separating more and more of their southern transcon while LA-Bartow is going to have 3-4 main tracks

The consistent issue with this line of thinking is that sheer idealism can’t send trains down the tracks and that’s what Acela 21 was all about. Why would/by what mechanism would Alstom just give Amtrak more trains and service when they are already certainly incurring nasty losses on this project. Caveat emptor. We’ll be lucky if they don’t pull a Midwest bilevel, say “sorry we can’t help you” and cut bait.
Yes the power system is old, but with what money should it have been fixed back in the days when Amtrak was leasing rail cars to keep solvent? I don’t deny that work was needed, but how much off that half billion was truly SOGR? How much was for signal retiming? How much cheaper might it have been to run new variable tension? I don’t know, but it’s worth an ask, given we saved ONE minute for a MINORITY of passengers.
Working with what we have may never get us to a world class rail system again, but insane $100,000,000,000+ pitches to Congress won’t get us anything. We used to do major realignment projects when tracks became congested or the market indicated it, and we continue to based on those criteria, and the hard reality is that there is zero market indication or political will to justify massive eminent domain for premium trains for the 1% in the densest and most valuable land in the country. BNSF is upgrading the transcon because capacity issues limit their ability to make money. If we were running jammed 15 car trains up and down the Corridor every 6 minutes, that might be different. But for Now there’s plenty of room along the ex-PRR for most of the way, choke points are actively being addressed. When the PRR fills up, they can restore and wire up the last bit of the ex-RDG and run trains on that. The same with the Shoreline and Inland. If Amtrak had spent the last 30 years applying Gunn’s incrementalism over trying to show the Europeans that “we can be cool too” and face planting every time with more and more tax money we would be in a much better place rather than trying to figure out why our billion dollar new toys don’t work. The cost benefit between a 225mph new ROW from WAS-NYP vs improved conventional service just doesn’t indicate construction right now.

 

[GDRRiley]

The consistent issue with this line of thinking is that sheer idealism can’t send trains down the tracks and that’s what Acela 21 was all about. Why would/by what mechanism would Alstom just give Amtrak more trains and service when they are already certainly incurring nasty losses on this project. Caveat emptor. We’ll be lucky if they don’t pull a Midwest bilevel, say “sorry we can’t help you” and cut bait.

Its a pretty common compensation tool instead of giving straight cash for example SBB RABe 514 fleet was late and had issues or Septa getting a ALP-44 for the N5 train issues.

Yes the power system is old, but with what money should it have been fixed back in the days when Amtrak was leasing rail cars to keep solvent? I don’t deny that work was needed, but how much off that half billion was truly SOGR? How much was for signal retiming? How much cheaper might it have been to run new variable tension? I don’t know, but it’s worth an ask, given we saved ONE minute for a MINORITY of passengers.

They could have done it when they brought out the Acela they already had a team put together.