Contact of Containership Dali with the Key Bridge and Subsequent Bridge Collapse [pdf] | Hacker News Reader

Contact of Containership Dali with the Key Bridge and Subsequent Bridge Collapse [pdf]

static01.nyt.com

100 pointsmhb89 comments

Interesting. When the disaster first unfolded, a lot of experts came out to say the bridge appeared to lack protective barriers to help absorb the energy from a ship impact. [1]

However, the NTSB says the bridge did have these barriers. There were large protective "dolphins" directly in front of the bridge piers, but the ship managed to careen in between them at an angle, missing them, but still hitting the bridge.

In addition, there were large protective barriers surround the bridge pier that was struck:

> In addition to the dolphins, pier nos. 17 and 18 were each surrounded by a 100-foot-by- 84.5-foot crushable concrete box and timber fender system , as seen in figure 12. These systems were comprised of hollow, thin-walled , concrete box structures attached to the piers. The timber portions of the fender were attached to the outer face of the concrete box and utilized a combination of vertical and horizontal members. Additionally, steel plates were secured at the base of the vertical timber members.

[1] https://www.nytimes.com/2024/03/26/us/baltimore-key-bridge-s...

When the bridge was originally constructed the concept of such protective impact barriers was quite new. The ship is also very different from the comparatively snub-nosed envisioned C7 types, trans-Atlantic cruisers, and barges that would've gone through when the bridge was designed in 1972. Cargo ships entering the harbour used to have wider bows with a shallower draft of thirty to thirty eight feet with a very shallow prow overhang because they were built for Atlantic voyages only.

There's few barges going through the harbour these days, and cargo ships now are global and mostly built to the variety of post-Panamax standards making them extremely long with a very deep draft of forty five feet or more. That means the spacing and height originally thought to be needed to deflect barges and C7s was far too broad and short for the Neo Panamax II class MV Dali or any other modern cargo ship, and the deep draft that required a narrow bow and long prow allowed the bow to slide right in between the barriers and then ride up until it hit the bridge support. Essentially the nose had an overhang long enough that it hit the bridge support long before the rest of the ship hit the protective barrier.

That sounds analogous to what happens in software when X is designed for one thing, then a bunch of interfacing components Y & Z are changed but X isn't updated to accommodate the new edge cases.

I’m no expert but looking at the pictures it is clear that most of the collision took place in the air well above the water, where the prow of the ship hit the top part of pier 17 causing it to collapse onto the bow of the ship. To me, the protective barrier around the pier at the water level should have stood off much farther laterally from the pier to prevent that kind of prow contact - but then it would have blocked much of the navigation channel. Therefore the bridge should have been torn down and redesigned to modern ship dimensions long ago.

I think these barriers are designed to protect against much smaller ships.

For big ships, the ship is expected to have sufficient redundant systems and assisting tugs to not crash.

That may be the case, but such assumptions have not been justified for some time (the report says the tugs were let go on entering the channel, which it calls standard procedure.) As I noted elsewhere, more recent structures in the vicinity have been given what looks like much more substantial protection.

What is the point of the tugs if they don't take it through the part where it's most critical to have a tug?

That is a fair question, and I would guess that there are at least two different issues in play.

The first is that the ship probably has little or no ability to steer until it is moving at a few knots (depending on whether it has thrusters; I suspect not.) Once it was moving down the channel, it could steer (assuming the steering gear worked, of course.)

The second is probably that no-one with the relevant authority realized how vulnerable the bridge was to the much larger ships that have come into service since it was built (yet someone apparently realized that the adjacent power transmission towers needed better protection; sometimes people are not good at connecting the dots.)

Oh I am sure someone connected the dots, just that no politician wants to spend $bil on infrastructure if they won’t take the blame personally for an immense failure like this

While it is true that some protective elements existed, it is not as if this collision was a result of an unforeseeable chain of events that found a loophole in an otherwise impenetrable defense. Why were the dolphins placed so far apart, and the crushable barriers so close to the bridge piers? At some point, towers to support electric transmission lines across the channel were erected close to the bridge, and they were surrounded by barriers with a much greater radius (though I do not know if they are enough to stop the MV Dali, let alone the largest ships using the port.)

https://maps.app.goo.gl/FQCsXqomTNHShzP67

Here are two other NY Times articles, one describing a collision in 1980 by a ship half the length and with about an order of magnitude less capacity than the Dali, and another comparing (unfavorably) the protection of the Francis Scott Key bridge with that around others.

https://www.nytimes.com/2024/03/28/us/baltimore-bridge-ship-...

https://www.nytimes.com/interactive/2024/03/27/us/key-bridge...

This is a fantastic example of clear technical writing! The NTSB goes out of their way here to define domain-specific terms (in the footnotes) to ensure this report is digestible by a public audience, which I think is really neat. We need more examples of this type of technical communication from government agencies.

https://www.plainlanguage.gov/ is an underutilized resource within the federal government, but it's evident that the NTSB has made an effort.

If you're interested in supremely detailed breakdowns and animations of accidents caused by safety failures, check out the U.S. Chemical Safety and Hazard Investigation Board's youtube channel. It strokes the brain just right.

https://www.youtube.com/@USCSB

It would seem the real mystery is why did the breakers open.

I have watched a youtube video where a knowledgeable chief engineer who works on comparable ships was very dismissive of the idea that it might be a cyber attack.[1] He was mainly reasoning that the on-board equipment is air gapped from the internet. Which of course it makes an attack more complicated but it is not like the airgap saved the centrifuges in Natanz.

That being said it also does not say that it was a cyber attack. Could be any number of ordinary malfunctions. But i don’t see that the publicly available evidence would exclude the possibility of a cyber attack as of yet. It would be very different if they would have found a damaged shaft in the generator, or a clogged fuel filter. Something hardware-ish.

1: https://youtu.be/9B9znFDwdBI?si=aFSoqVsmrTaaQofY

I don’t see that the publicly available evidence would exclude the possibility of aliens causing the breakers to open either.

We don’t see any UFO on the videos and there are no reports about little grey ones on-board. Seems pretty much exluded by the evidence that they tripped the breakers in person if you ask me. Did they do it remotely or in a delayed manner? I think that leads back to my question.

You are ignoring the existence of underwater UFOs (UUOs).

Those are in fact the real UFOs and the Congress have been misleading the public for decades by having everyone chase some made up flying sausage in the sky, as if we are small children. Just the fact that they created the hoax around Area 51, in the middle of a dessert and far from any water should have told you all you need to know. And don't get me started on the 1961 fiasco in Bay of Pigs where CIA failed to capture the damaged UUO that had floated to land.

But if we just for a minute step away from the obvious connection between JFK Jr, aliens and the Illuminati... have you ever considered that training LLMs on public forum comments is a bad idea?

THAT'S WHAT THEY WANT YOU TO THINK

/s (I think)

But aren't they wearing human skins? Unless the NTSB checked for zippers on the crew, we'll never know who is compromised.

Probably time travelers. Most likely to prevent transport of a specific item or person over the bridge at a future date.

Perhaps a shadowy group striving to put wrong what once went right.

Or just thrill seekers[1].

[1]: https://www.imdb.com/title/tt0204686/

1999 was truly the peak of film making.

Office Space, The Matrix, Fight Club

Arguably all the same movie

Well, the Matrix was about taking red estrogen pills to transition your body to your preferred gender. I'm not sure what the other two were about.

We can't exclude midi-chlorians either, 5/4 was coming up.

DEW's (Directed Energy Weapons) are capable of surging power and more likely. https://climateviewer.com/2018/07/05/directed-energy-weapons...

Breakers are usually tripped when too much power is used. Because the ship was traveling slowly in the harbor using lower power, we need to understand other reasons why a breaker would open.

Failing breakers commonly trip below current limits. Hopefully they will comprehensively test the circuit path and its components in scope. Someone will have to explain why they weren’t replaced prior to failure if root cause.

True. But to have both fail at the same time usually means an actual overcurrent for some reason

Hopefully they get an actual current measure somehow

It would not be unusual for the protection design to have the LV and HV breakers on either side of the transformer linked, so that if one trips the other does too. So you can't necessarily read anything in to the fact that both tripped.

Why link them? Any failure opening one or other breaker is sufficient to stop fault current.

The only reason to even have both breakers is so the transformer can be taken out for maintenance.

The idea is to fully disconnect the faulted element from the power system.

In particular with transformers, you don't want it backfed from the low voltage side if the high voltage side is isolated.

Interesting, makes sense. And yes I think the backfeeding is the problem, you don't want the other generator to start automatically while the other side is still connected

I don't know anything about marine regulations, but on land you generally only have a breaker or fuse on one side of a transformer (usually the "powered" side).

If you're in US, the mains transformer feeding your residential service almost certainly has no overcurrent protection on the secondary, it has a fuse on the primary.

Traveling slowly in a harbor is when the most load is required due to having a bunch of stuff happening during transit. You’re going to have every generator and engine on, all the anchor equipment, line handling equipment, the steering and auxiliary motors will be working overtime, etc.

Engine speed has nothing to do with power generation, these are totally separate systems. It's not like a diesel train with electric motors; on this ship (and most others) the engine is directly coupled to the prop.

The engine requires one of the HV generators to remain online so that oil and cooling pumps operate (and also to bring fresh air in to the engine room for the crew, but those fans can probably run off the emergency genset). Once the HV generators were disconnected, the pumps shut down and the engine stopped. Restarting the engine is a process. Here's how it's done for a small engine (5500hp)[1]. And for something similar to the Dali (55,626hp)[2].

The circuit breakers are a little different than common magnetic household breakers. They will trip in over-current scenarios of course, but also for over- or under-voltage, over- or under-frequency, or if triggered by the ship's SCADA system(s). Breakers tripping on the HV and LV side of the transformer makes me think that they tripped because of something other than over-current.

Three things in the report caught my attention:

1) The crew initially tried resetting the tripped breakers (HR1 and LR1) instead of energizing the alternate set (HR2 and LR2) which they had been operating on for months.

2) The second blackout was caused by the diesel generator breakers (DGR3 and DGR4) tripping. The report doesn't specify whether they began startup procedures for generators 1 and 2 when HR1 and LR1 tripped (probably not, because at this point gensets 1 and 2 were still operating), but it does say that genset 2 was in standby and connected automatically after the second loss of power. With #3 and #4 online, adding #2 to the mix isn't as easy as throwing a switch, it has be to be synchronized to the frequency of the other two first and this takes some time. But once DGR3 and DGR4 opened, DGR2 could be closed immediately and DGR1 wouldn't be far behind (i.e. as soon as generator #1 was synced to generator #2). I expect the state of the generators will be fleshed out more fully in later reports.

3) The report makes no mention of the state of the compressed air tanks. There should have been sufficient compressed air available to attempt several restarts (up to a dozen per regs), so running the compressors wouldn't be necessary. But if there wasn't enough stored compressed air, then no matter what the crew did with the electrical system wouldn't have mattered; they wouldn't have had time to fill the tanks enough to restart the engine. Again, I expect this to be covered in later reports.

My guess is that there was some fault with transformer #1, which caused HR1 and LR1 to trip off. When it was brought back online and they attempted to restart the engine, the combined load of normal ship operations, additional startup equipment (combustion air blowers, etc) along with whatever (possibly intermittent) fault TR1 had overloaded the generators, which then tripped offline. Perhaps instead they should have closed HR2 and LR2, and started an additional generator before attempting an engine start. But that's just a guess. The report doesn't mention any restart attempts, so there may have been other electrical faults that needed to be sorted out first.

Also, having used up some compressed air to start the engine at the dock, they may have been running the compressors to bring the tanks up to the required stored capacity which would have contributed to the load on the electrical system (but only up to the first blackout, as they wouldn't have restarted automatically).

1: https://www.youtube.com/watch?v=J-D9Ka3TM1I 2: https://www.youtube.com/watch?v=5mFmCxOjY_A

> the engine is directly coupled to the prop.

Mildly interesting: to reverse propulsion, the engine is stopped and restarted rotating the opposite direction.

‘Airgapped from the internet’ is very much the default state of technology on a ship. Connecting things to the internet seems like something that would require additional steps.

At the moment perhaps, yes. I will predict that will change as satellite internet becomes ubiquitous.

This should not have been downvoted. I work with the shipping industry and one of our partners just bought a satellite network.

All commercial ships these days have WiFi and satellite uplink is becoming ubiquitous. The tech we are building depends on low speed, low volume bidirectional data streaming and we haven’t yet found a ship in 2024 that doesn’t have much more bandwidth than we need

Here is a video from Chief Makoi on this very topic. He is a highly respected working Chief Marine Engineer.

https://www.youtube.com/watch?v=9B9znFDwdBI [10:36]

ever heard of AIS, e. g. https://www.marinetraffic.com/en/ais/home/ ?

That's a VHF transmission emitted from a boat containing a data packet about its position and speed.

Quite a long way from an internet connection.

But of course I was being flippant - modern commercial ships likely have satellite internet access, onboard networks, and probably at least some equipment on board that accepts over-the-air updates of some kind.

The fact that they had two explained power failures in port, and they made some changes to the configuration point to more likely causes: they made some faulty reconfiguration or the previously dormant breakers were faulty.

The report has at least one of the power outages in port is explained (and says it was the only one). They did change config after it though; but still.

It does give an explanation, but that doesn't mean that they didn't play with other things to diagnose or guess at fixes.

I think a cyber attack seems unlikely, if only because going from tripping those breakers to the bridge collapsing is a hell of a rube goldberg machine. (And if that was only an accident, and their only goal was "make the lights flicker", then that seems a bit of a pointless cyber attack).

They are not telling us everything. NIST has concocted things in the past. The whole ship crew is still held on the ship with all their communication devices confiscated. When/if investigative journalists get access to them, we will learn more. https://www.cbsnews.com/baltimore/video/dali-crew-endures-st...

Good thing the NTSB wrote this report, then. Definitely can't trust those NIST folks, they're plainly nefarious.

Believe everything the government tells you. Do not question.

Wait - they are still stuck on the ship after all this time? Is this a company/government rule? Some kind of maritime law?

Not like I expect any of them to keep their jobs after this, even if they acted impeccably during the failure.

It seems to be pretty common in the maritime industry. There were crews stuck on ships for months during the pandemic.

Having asked these questions myself you can find out through Sal at WHATS GOING ON WITH SHIPPING on YouTube.

The company that acted as salvor, Resolve Marine, publishes detailed explanations of past ship salvage projects: https://resolvemarine.com/services-capabilities/salvage-wrec...

wow i had no idea they had some much generation and distribution going on. That's a lot of power they're moving. That's a lot of things that can go wrong. Even with all the backups, complex systems are complex, and there's obviously a gap in redundancy. Still, better a system failure than a human one.

It was interesting that on the video you can see the ship lose/regain/lose power, and now it makes sense.

The ship left port weighing 112,383 metric tons (ship+cargo). Travelling at it's loaded full speed of 16.8 knots, that comes out to be about 10.6 GigaJoulese or 3 Megawatt-hours of kinetic energy.

TNT has an energy density of about 4.2MJ/kg. So the kinetic energy of the Dali at full speed is equal to about 2.5 metric tons of TNT.

According to the report, the Dali was running at slow speed, which is 10 knots, giving "only" 0.7 tons worth of TNT in kinetic energy.

Granted, ships operate in a relatively low friction environment, so you do not need that much power to get it moving. However, modern shipping vessels are massive.

tnt is the more commonly used metric but to put it in more practical civilian terms that would be around a trunkload of anfo (might need to share the back seat but who uses all that legroom anyway)

Why did this ship have 5-way redundancy in electrical power generation, yet zero redundancy with propulsion?

I understand power is more important, but even so, isn't 5x redundancy overkill? Wouldn't they have been better to switch out 1 of those diesel generators for a 1000 HP propulsive engine, allowing the ship to limp home at 2 knots if the main engine fails (speed is very nonlinear with power)?

It's not 5x redundancy in that sense. There are times when the ship needs all five auxiliary engines. There are five smaller engines because it allows for more efficient utilization.

The main engine, which provides the propulsion, is a two-stroke engine connected directly to the propeller shaft. If the ship needs to reverse the engine is stopped, and started again going the other way. Consequently, there's no way to make use of some sort of fall back propulsion.

Lastly, I doubt a 1000 HP engine is powerful enough to do anything useful. Especially because it would only be needed for emergency maneuvering.

Matters of Fact, still investigating. All good info to answer the public’s curiosity.

I am rather impressed that every crew member tested negative for drugs and alcohol.

I was under the impression that 2 month voyages across the ocean could get rather dull and personal stashes of all kinds of substances were common.

Looks like the whole ship power system was designed to be multi way redundant, but not to provide continuous power.

Makes perfect sense for dealing with issues that occur mid voyage, where you can send someone below deck to mess with circuit breakers for 20 mins to get everything going again.

Doesn't make sense in port where loss of steering or propulsion will lead to a crash.

The report mentions that "Six victims were later recovered by divers".

I can't imagine how terrifying it must be to a diver when he actually finds a dead body underwater. I guess there's no amount of experience that desensitizes you for that encounter.

Note that a bunch of commenters suggested reverse thrust was used after the power failure which caused prop walk which turned the ship into the bridge.

That appears not to be the case.

>The day before the accident, the Dali had experienced two in-port blackouts, prompting the crew to switch the main electrical bus configuration, which was in use when the ship departed.

...

>The NTSB is now investigating the electrical configuration and its potential impacts on the accident voyage.

As a retired Navy ship driver, the whole incident is astounding.

Shipping was foundational in the development of the concept of insurance[1] for precisely the reasons underscored by this tragic fiasco. One surmises that this particular srew-up wasn't the first time these people rolled the dice, but it's likely to be their las.

[1] https://en.wikipedia.org/wiki/History_of_insurance#Medieval_...

Reply on news.ycombinator.com