(Haven’t figured out how to disable the automatic text formatting, haven’t searched either though. I’m an old school guy who likes manually controlled things… where no process control is required 
)
Let’s assume that the data of the PR is correct, though usually generators are specified in kVA and engines in kW:
Installed, total: (DG1 4400 kW) + (DG2 4000 kW) + (DG3 4000 kW) + (DG4 4400 kW) = 16800 kW = 16.8 MVA.
Running only 3 out of 4 DG’s (keeping one as spare):
Available, total: 12400 to 12800 kW (12.4 to 12.8 MW).
Loads:
-
Base load: Not mentioned in the PR but let’s assume that when covered by a single 4000 kW there is still some margin.
-
Bow Thruster (BT): Up to 3000 kW, possibly somewhat more if the power refers to the availble power at the shaft of the motor ((torque in [Nm] x (angular speed in [Nm]), sorry I’m a SI guy.
The effectively required power will vary depending on how the BT is used. The required current to start the variable pitch propeller BT depends on how it is started, no details are known. It’s certainly not a VSD, too expensive here. The local BT switchboard is supplied by the HV BT breaker but that breaker does not start/stop the BT motor directly. -
Up to 1400 reefer containers (see Hogsnort’s contribution): About 6100 kW (6.1 MW).
That load can very widely vary as it depends on the amount of reefer containers, their type and cargo, ambient temperature,…
Also for a limited time, for example when the bow thruster is used and there’s not enough power available reefer containers can be powered sequentially if required as there are several smaller HV 6.6 kV/LV 0.44 kV transformers (not shown in diagram), maybe around 2 MVA ea. or so.
(Sorry for the erroneous automatic text formatting!)
Total 1 to 3: 4000 + 3000 + 6100 = 13100 kW.
Available: 12400 to 12800 kW (running 3 out of 4 DG’s).
Now, in real life the base load will likely be below 4 MW but here I’m just guessing, most of that power is used by pumps, compressors and some blowers/fans as high temperature heat is provided by the boiler amd exhaust gas heat recovery. The only temporarily required BT will not always require the full 3 MW (or slightly more) and the reefer containers won’t either always require 6.1 MW.
Overall I don’t expect that there should be any critical overload problems even if one DG is not available, e.g. due to maintenance.
Also, the 6.1 MW for reefer container is probably theoretical. How many of the 1400 available sockets will ever be used at the same time?
Having enough power available problems could still occurs if loads are not balanced correctly, i.e. no busbar nor breaker should be overloaded.
As only one single main service transformer, either TR1 or TR2 but never both (started in the PR (NTSB Preliminary Report)), is in service at any time, the LV BUS tie breaker LVR is only useful to isolate one half of the LV main distribution, maybe for maintenance, but is otherwise useless and represents a point of failure. Simple removable copper plate links would be both less expensive and more reliable than a breaker if it is strictly prohibited to operat both TR1 and TR2 at the same time, e.g. due to insufficient short circuit rating of the LV BUS busbars .
Therefore if it assumed that LV BUS tie breaker LVR remains permanently closed, load cannot be balanced as there the LV BUS is always operated as single bus (cannot be split as LVR is kept close, it it would open for any reason one segment of the LV BUS will no longer be powered, meaning that roughly half of the LV (Low Voltage, 440 V) electrical equipment of the ship will no longer be powered (the reefer containers not being considered as equipment),
For the 6600 V HV BUS it’s different as it also supplies the BT (Bow Thruster) and the various not shown 6000 V/440 V transformers supplying the sockets for the reefer containers.
Not enough running DG’s can obviously lead to an overload condition but even with a half-baked power management system that shouldn’t be a problem. The only load which could possibly lead to a sudden overload is launching the bow thruster but we dont know how high the current peak to start the motor is as it depends on the type of motor starter used, also the BT is started with propeller blades in neutral position.
Othe sudden high load changes would be the result or human errors and/or a problem related to the Power Management System (PMS), like e.g. opening inadvertently the HV BUS tied breaker LVR leading.
For each breaker which is to be closed the PMS knows the maximum expected load and should not close it before enough power is available, i.e. starting and bringing online an additional DG (Diesel Generator) automatically.
Also if implemented wisely, the transformer protection can allow a fairly high short duration overload but that depends on the load history of the previous hours, if the transformer already it hot the short term overload is low, if it’s cold it can be quite heavily overloaded during a short time. Ideally large service transformers should be protected by advanced specialized transformer protection relays which feature a firmware specially optimized for the protection of transformer, there are other specialized protection relays for large generators, motors, busbars, etc. The hardware is often the same as basically various currents and voltage need to me measured but the whole processing is fully digital and there are digital outputs to control the opening and, where applicable, the closing coils of the breakers. The more a transformer is expensive or critical, the more advanced protections are. Large generators and large transformer are protected by redundant protection relays where a failed control device can be replaced without interruption. Oil rigs and vessels requiring dynamic positioning feature way more reliable power supply designs, they ressemble more what’s common in critical industrial plants and power generation. Even waste incineration plants feature very well designed power distributions as a full blackout can cause millions USD of damage within a very short time (during a large storm in my country many waste incineration plants lost external power, was a great opportunity to check blackout operating procedures). But I digress again!
Although I’d have expected to have 3 DG’s running during departure as simple principle of caution regardless of the required power there is no obvious indication that even with only 2 DG’s there should have been any problem.
With DG3 (4000 kW) and DG4 (4400 kW) running as the incident voyage began there were 8400 kW available. Now let’s suppose that he discussed 6100 kW reefer container power was required, which is mostly purely hypothetical, that would lead to an overload of 1700 kW (assuming a base load of 4000 kW) which would require a 3rd DG. Now if there would have been a significant reefer container load, at least 3 DG’s would have been running from the beginning and even if the reefer container load would have increased for any reason en route, which is also purely hypothetical, the transformers would have been brought online sequentially and only after enough DG’s would be connected to the HV BUS.
Only a poor power management system would allow an overload leading to tripping breakers. Running DG3 and DG4 and enough reefer container load, using the BT could lead to an overload but logically as the BT is prioritary, enough transformers supplying reefer containers should have been disconneced automatically as part of the load shedding sequencing procedure, or maybe the power of BT should have been limited and with a pushbutton one can request priority which sheds load to free enough power to use the BT without restrictions (but unlike for the ME, there’s probably no such button for the BT).
The power management can be handled either by specialized controllers or by common PLC’s with corresponding marine approval which are interfaced with specialized measurement devices and/or retrieving current, voltage and power data from various devices like protection relays, genset controllers, etc. There is also specialized substation control hardware but it is expensive and proprietary. Redundant PLC’s would be a good idea, power management could be handled easily as the critical functions would still be handled by dedicated specialized electronic devices.
(Just noticed that I replied in the wrong topic, I’ll try to repost in the other topic and delete here if it is possible.)