NTSB Preliminary Report M/V Dali

Professional Mariner Forum
265

Air does not get in. Just some random ramblings about atmospheric pressure pushing into ‘too low oil pressure’, and it goes on.
But I do feel bad for calling out this (think) young lad who seems to be an instrument guy and says stuff on all other disciplines with great deal of confidence. (Sercos - sorry!)

There are very experienced engineers and master mariners here who are just interested to know what happened with the sole intention of preventing these mishaps. (lesson learnt to implement changes as appropriate or required)

So, you are more than welcome to air your views - but suggest you please provide comments based on your education, training and experience.

266

I was a bit suspicious but did not have the technical knowledge to debunk anything.

Thanks again you guys :+1:

267

Has one person yet speculated that the motor starter for the #3 steering HPU was possibly not in autorun mode requiring it to be manually started from in the steering flat?

Such stupidly simple causes are quite possible and much more likely that all these other theories being bandied about.

268

Is this possible?

As far as I remember from being a ship driver we had to start steering pumps from the bridge.

269

The steering gear pumps are supposed to start (if running) automatically when power is restored after a power failure. ABS rules … Class NK should be similar being an IACS member. At least the SG pump 3 should have started as soon as the EDG came on line. Maybe the others required a start from the steering stand.

The NTSB PR is so lacking in info and in places downright amateurish such as their statement below … ‘could be configured …’??. Ocean going steel vessels are required to have an emergency source of power. If it is an EDG, it SHALL start and come on-line within 45 secs and carry all the load connected to the e-bus.
Quote: “Typical for oceangoing vessels, the Dali had an emergency diesel generator (in addition to the four generators) that could be configured to automatically start and connect to the emergency bus if normal electrical power and lighting were lost.” Unquote.

Almost looks like an intern is writing the draft and it passed through the checks.

PS: Remember the intern from NTSB who played a prank and called the local TV station and advised that he had the names of the flying crew on the Korean Asiana Boeing 777 flight that crashed in SFO airport? Sumting Wong, Wee tu lo, etc.

270

Straight from SOLAS:

5 Main and auxiliary steering gear power units shall be:
.1 arranged to restart automatically when power is restored after a power failure; and
.2 capable of being brought into operation from a position on the navigating bridge. In the event of a power failure to any one of the steering gear power units, an audible and visual alarm shall be given on the navigating bridge.

I was wondering in a previous post, when they lost steering control they should have send an engineer down in the steering room to regain control. Emergency steering should be practiced every 3 months I believe. Of course, easy to talk now.

271

So, if the pump which is fed from the Emergency Switchboard was not running before the blackout, it could not have restarted by itself, it had to be started manually from the bridge after the Emergency Switchboard became live. It all depends on the actual arrangement of the controls.

272

General consensus or assumption.
The Dali has Hydraulic ram steering with three steering motors. Not Hydraulic Vane steering.

The time from midship to hard over. With three pumps , I can’t recall Requirement .Just seam’s to take a heck of a long time when suddenly required.
On only one pump the heck of a long time turns into an eternity when suddenly required.
I don’t see the effect of port rudder on her track. Just head falling off to starboard probably due to a combination of multiple outside influence.

273

But I still think there was enough time even with reduced speed on the steering pumps and lack of drive from the propeller for the ship to have recovered from a slight starboard turn at the time of the blackout at the speed through the water almost 9 kts .

I am fairly sure in my mind that the bridge crew never recovered rudder control.

Either that or they were totally incompetent.

I have never worked with American pilots but I have with many many Indian bridge crews.

I refuse to believe that the American pilots were so useless or indeed that the Indian bridge or engine room crew were equally stupid.

So once again I defer to our engineers who actually understand the complexity of the power distribution system and will hopefully dig deep and find the cause of this disaster.

Maybe in spite of the NTSB report.

I hope I am wrong

274

@ retdmarineengineer

Please tell me exactly where I’m wrong excepted some SOLAS details which I didn’t check thoroughly. I try to be technically accurate.

I didn’t say that massive amount of air will enter the hydraulic system, I only meant that if, like when pulling back the piston of syringe with a clogged needle, some air will enter after some time and it does not need much to cause problems. Even a clogged hydraulic oil tank breather can cause problems when the pressure inside the tank from which a pump is fed drops too far below the atmospheric pressure.

The local control system of the steering gear as well as power for Pump #3 (and one small pilot pump) is supplied by the EBUS, the other 2 main pumps are supplied by the LV BUS.
When only the supply from the EBUS remains, regardless if the EBUS is supplied by the EG or the LV BUS, the steering gear enters automatically the emergency power mode where only Pump #3 is operated (though I’ve serious doubts about the RPM of that pump being lower when in emergency power mode as it would make absolutely no sense with a variable pump).

There is no need for any crew member to attend the steering gear. There are additional emergency modes, one electrical with local manual control and one sort of last resort mode where control valves (i.e. the 1st control stage) are operated manually instead of electrically but that requires some manual preparations at the hydraulics controls which take a couple of minutes. I didn’t check if the local gyrocompass repeater is optional or mandatory (the bridge intercom is mandatory).
Such emergency manual controls are also common in other domains when it’s necessary to control hydraulics by hand, typically the first pilot stage is operated manually by pushing the valve spool from end of the control solenoids, sometimes the’s just a rubber cap which remains in place, sometimes a screw-cap preventing inadvertent operation must be removed. Just as example, details vary.

Some steering gear faults require manual valves to be operated locally, especially to unlock a locked steering gear. Locking conditions as well as oil level alarms are transmitted to the bridge together with other indications like e.g. tripped motor overload protections.

The PR doesn’t mention anything about steering gear faults. The steering gear stopped operating as the 1st blackout ocurred and emergency power mode was most probably enabled as the EG came online and powered the EBUS (not after about 1h26mn39s EDT, real time, not video timestamp) from that moment the steering gear did probably operate in emergency power mode without further issues. I assume it as the PR does not mention anything contradictory.

As far I can interpret the PR there was no technical issue with the steering gear itself, here meant that it performed as expected when supplied only by the EBUS (i.e. angular speed is reduced but the full torque remains available as pressure relief setpoints remain unchanged).

For steering gear pump there are typically a couple of pilot lights for statuses and alarms and controls to start and stop each pump.

275

As far as I can recall SOLAS does not require a gyro repeater in the steering gear room. All ships that were built this century have a repeater and it was checked for alignment as part of the pre departure check list.

276

@ Hogsnort:
Thanks. I wasn’t sure.

BTW I saw a picture of a large hydraulic ram-type steering gear with a wall-mounted electrical room heater not far awy. If there’s a high pressure hydraulic oil leak, its spray can very easily ignite if a hot enough surface can be reached.

.
I’d also like an expert in oil hydraulics to confirm that what I wrote is wrong, retdmarineengineer and Steamer discredited me without even mentioning exactly where I was wrong.

I welcome constructive criticism, absolutely no problem, I made some mistakes, mostly referring to regulation details which I don’t know in detail though I try to write reasonably accurately when it comes to technical issues.

It’s also important to remember that ultimately some very small details may make a difference.

Referring to a 4-ram steering gear, without the detailed “as built” hydraulic diagram, and possibly also including some component datasheets and the electrical diagrams, we can’t know exactly if there could have been issues.

Overall I’d expect the steering gear to revert to regular operation when power from the breakers supplied by the LV BUS is restored.

I suppose that Pump #1 and Pump #2 incl. their correspondig auxiliary pump (called servo pump though it’s not a true servo-hydraulic system as there are no servovalves) are supplied by separate dedicated breakers directly fed by the busbars.

Critical loads are usually fed directly from main busbars (LV BUS) by breakers with a high short circuit rating (Icu).
Marine low votage main distribution switchgear can feature large busbars for very high currents (> 6300 A) but prospective short circuit currents are not extreme because the short circuit current the generators can provide is low, it must even be verified that at least 3 times the rated current can be sustained during 2 seconds in order to allow downstream overcurrent protections to trip reliably.
If the available short circuit current is too low, some downstream protections may not clear the fault as expected (there’s a maximal delay to respect to limit the let-through energy).

To illustrate the idea: A 16 A socket can be fed directly from a 10’000 A busbar, it’s just that the breaker will be more expensive and larger than the one in a private home, the rationale being that the selectivity will be at the highest possible level, i.e. there are no other additional intermediate breakers.

A common power distribution is somewhat like a tree where each ramification is protected by a breaker or a fuse. If the trunk is gone all power is lost but if a breaker only protects a leaf only that single leaf will be affected.
In a highly selective design, visually speaking, all leaves would be connected directly to the trunk and therefore any breaker can trip any time without affecting any other circuit.
Hope this explanation makes sense as it’s a major concept when it comes to reliable power distribution. Selectivity issues have lead to all sorts of incidents.

277

(Had a connection problem, hope it won’t end as accidental double post.)

Referring to an earlier message: Cavitation as such won’t happen in the cylinder, it can especially damage pumps. If the absolute pressure of the hydraulic fluid even if only very locally and very shortly drops below some limits, small voids like vacuum bubbles can form and than implode which can cause serious damages. Due to physical laws the vacuum ist not an absolute vacuum but I leave that to the experts.

When designing hydraulic systems among the basic rules to follow is to keep all pressures and all temperatures within the applicable low and high limits and to keep the oil clean.
Sneaky can be some issues related to important temperature changes which can lead to overpressures or air entering the circuit as sealing is never perfect.

.
I expect regulations to require that both the High Voltage side breaker and the Low Voltage side breaker of a transformer always trip together, regardless of the cause of the breaker opening.
I.e. if HR1 opens, LR1 will open too and vice versa (also applies by analogy to HR2 and LR2).

If HR2 or LR2 tripped to clear a detected fault, the protection function which issued the trip command should normally be displayed and possibly even be logged by the Power Management System (PMS), it depends how the protection relays are interfaced. Often critical interfacing is handled by discrete signals (e.g. 24 V DC digital I/Os) and additional non-vital communication is handled by bus communication but as said, we’d need to know further details about the exact hardware architecture of the PMS.

Among other possibilities would be human error though in my opinion not very likely but of course not fully excluded as well as a spurious trip which cause may be found… or not.

Lots of causes can initiate an unwanted (spurious) breaker trip and it’s not always possible to find out the exact cause, could for example be some shock in a specific direction with a small spring which has become too weak but still won’t cause spurious trips otherwise.

Though overall I rather suspect an issue with the protection and other digital relays as well as the PMS.
A spurious breaker trip may happen but here first the breakers protecting TR1 tripped and than there was a double generator breaker trip (DGR3 and DGR4). Also we don’t know if there are specific busbar protections which, if present, can also issue breaker trip commands.

Referring to the ICMS (PR page 10) I’m not sure at all that it’s fast enough and networked in a way that all event timestamps remain coherent for remote modules. Also we don’t know if it’s partially interfaced digitally (bus communication) or only discretely (using I/O’s, Inputs/Outputs).

When referring to the PMS and digital protections I also include genset and bay controllers.

For know I ignore if the breaker trips of the incident voyage are related to what happened the day before but I wouldn’t be surprised.

Purely intuitively I’d tend to believe that the whole suite of electrical incidents could be related to some misconfiguration of the PMS and/or other issues with devices which can issue breaker trip commands. That said, of course it’s also possible that there were real faults which had to be cleared by automatically opening the corresponding breakers.

Why TR2 was not reconnected automatically by the PMS immediately after TR1 was disconnectd is unclear, especially as TR2 was brought back online manually. Also, after DG3 and DG4 tripped (DGR3 and DGR4 tripped), once DG2 which had started automatically and got also reconnected automatically to the HV BUS as DGR2 closed automatically, it’s all but normal that neither TR1 nor TR2 was brough online automatically by the PMS.
Also the fact that DGR3 and DGR4 tripped simultaneously (or nearly?) clearly shows that two distinct trip orders have been issues, a dual purely spurious trip without any common cause would be extremely highly improbable, i.e. “close to impossible”.

It’s unclear if the engines of DG3 and DG4 were shut down at any time, if not there was probably always enough power available to run the ME (Main Engine).

It will be interesting to find out if a ME can ride through a short blackout and if so how long that grace period can be. If well designed I don’t see why a ME couldn’t ride during a couple of seconds through a blackout especially also that there’s some propeller windmilling.

278

Here’s a timeline.

0109 Pilot orders course 141°

0125 1st blackout 0.6 miles from key bridge, about 3 shiplengths
-VDR no data, only voice
- All three steering pumps stop
- Heading 141.7° COG 140.8° SOG 9 kts Rudder amidships

0126:02 VR data resumes (EDG?)
-Heading 144.3° COG 142.7° SOG 8.6 kts

0126:13
-Pilot orders 20 degrees port rudder:

0127:01 Second blackout 0.2 miles from Key Bridge (EDG running)
-Pilot orders port anchor dropped

0127:23
- Pilot orders hard port rudder

0127:32 Crew restores power, about 30 seconds after second blackout

0129:10 Dali’s starboard bow strikes pier of Key Bridge at 6.5 kts.

279

100% correct. I checked the 2004 and 2010 consolidated editions and wording is identical.

So what this means is that aside from SG pump 3 that should have started with the EDG online, the main pumps should also have come on as soon as power was restored - pilot orders 20 deg port rudder seconds later. Power remains on for a little more than a minute. More than enough to turn the rudder 35-30 twice over. (Reqt is 35-30 in 28 secs).

And yet, no discernible change in heading. SG damaged or useless without the ME. I suspect the former.

PS: No mention of any engineer heading to SG flat in the PR. Probably only 2 engineers on watch and with the blackout they must have had their hands full.

280

All three pumps were on line per the report (page 7). Suspect with this arrangement 2 pumps would be required to meet Solas of 35-30 in 28s. With 3 pumps online, speed should be better.

Timeline above from KC is useful to see the duration of main power availability.

281

There he goes again. What auxiliary or servo pumps?? Where did you get this?

Really??

10,000 amp busbar? Most marine LV switchboards with total generating capacity of just 2000kW will be rated at 100kA short circuit rating. (100,000 amps)

282

I am curious, the stuff you write, does it come from your own experience? It seems that you have sailed onboard Dali and you have access to her drawings/manuals.

283

It seems unlikely that the steering would fail and the report would make no mention of even the possibly.

If it did in fact fail the first step would be to switch to NFU (Non-follow up) before sending someone to the steering gear room.

284

I agree. I guess that the bridge voice recordings could shed some light on this issue. Was the bridge team aware that the EDG was running? Did they try to start the steering gear pumps manually? Did they ask engine room to check the steering gear locally?

Another intriguing issue is with the VDR:

At 0126:02, the VDR, which had stopped recording vessel system data when the blackout occurred, resumed recording the data. The VDR audio recording had not been affected by the blackout.

Usually most of the systems transmitting data to the VDR are supplied from UPS, why the data recording stopped when blackout occurred? I am just speculating here, but if the integrated control and monitoring system (ICMS) stopped sending data to the VDR, it could indicate that it’s UPS malfunctioned and could not supply power resulting in an uncontrolled shutdown of the ICMS. And the ICMS usually integrates the Power Management System, which is responsible with reclosing the breakers and sequentially restarting all the auxiliary pumps. Maybe this was the reason the transformer breakers had to be closed manually.