IET
Decrease font size
Increase font size
Topic Title: ARC Flash Incident on Ship
Topic Summary: Cause?
Created On: 25 December 2011 08:23 PM
Status: Post and Reply
Linear : Threading : Single : Branch
<< 1 2 Previous Last unread
Search Topic Search Topic
Topic Tools Topic Tools
View similar topics View similar topics
View topic in raw text format. Print this topic.
 31 December 2011 03:51 PM
User is offline View Users Profile Print this message



williamjohn

Posts: 178
Joined: 22 November 2010

Hi Mike
Have you checked the insulation resistance between the primary and secondary windings of the transformer. All the evidence points to a breakdown in this insulation so that the 690 volt breaker was trying to clear a 6.6kv system fault.

Best wishes
John
 31 December 2011 06:21 PM
User is offline View Users Profile Print this message



michaelfivey22

Posts: 27
Joined: 21 January 2004

John,

Just tested it with a 2.5kV megger test and everything is 1500Meg or above.

Have not managed to do a continuity test of the windings as we do not have a ductor on board and cannot get one as it is new years eve.

Would it be possible it could only break down under load?
 31 December 2011 07:38 PM
User is offline View Users Profile Print this message



williamjohn

Posts: 178
Joined: 22 November 2010

Hi Mike
Sorry to hear that the injuries were not minor. I send my sympathies and hope for a speedy recovery.

In answer to your question, I once had a 2000 hp 3.3 kv motor trip on earth fault but it tested 1500 Megohms to earth with a 5 kv Megger. On removing the end covers the fault could be clearly seen but it was only occurring as the windings were bent by the starting current. So to answer your question, yes it is possible for a fault to clear when not energised. However transformer windings are very well braced and movement of the windings is unlikely. It is possible for the connections to the terminals to move under load but it seems unlikely that they would move when energised off load.

There is of course an AC connection between primary and secondary by the capacitance between them. You say the 690 volt neutral is not earthed, so this capacitance connection could lead to an overvoltage on the 690 volt system but this connection could hardly be enough to blow up the 690 volt switch.

I do not think that an open circuit on the primary or secondary winding could lead to the problem that you have had, unless there is a loose end.

Has the system operated satisfactorally in the past or is it a new installation?

I thought I knew the cause, but not now.
Good luck
John
 31 December 2011 09:21 PM
User is offline View Users Profile Print this message



dlane

Posts: 690
Joined: 28 September 2007

Originally posted by: michaelfivey22

"For a 690V system I would have expected initial direct contact but that would then be blown away by the arc as it is initiated and then sustains itself so the initial short may not be detected post event."

Can you please expand on this - the only part in the network we see damage is in the breaker - we can see no signs of damage i the swithboard bus bard apart from the small damage to the back of the downstream breaker cubicle.



Hi Mike,

With regard to your question above, air, dependant upon the ambient conditions has a breakdown of around 3kV/mm so on an LV system there isn't the voltage level to breakdown the air and initiate the arc. You would therefore need a direct phase to phase contact to initiate the arc and cause the blast.

When we carry out arc flash tests we actually initiate the arc by linking the electrodes together using wire. Then after taking some big strides the power is switched to the electrodes, current flows and the wire between them melts but as it breaks up enough heat has been generated to ionise the air, cause conduction and the arc builds up until the power is removed.

So all you need to start an arc in the breaker is a small amount of conductive debris that initially creates the short but as with the arc test when the blast is over you cannot detect an actual short in the circuit.

I think you mentioned that there was an undervoltage trip coil within the unit. Off the top of my head trip, coils are generally short term rated (as in seconds) do you have a permanent voltage to the trip coil which has subsequently burnt it out and due to the comapct nature of the ABB Breaker has caused an internal short and hence initiated the arc? I don't know how the protection works or if the breaker has internal aux contacts that break the supply to the trip coil but it may be worth considering if you cannot find any other fault.

I don't suppose you have a spare breaker that you could open up or any pictures of the internals and take a look at its internal construction to give you some ideas.

Also from your description I take it that the down stream breaker where you have indications of arcing is situated away from the failed breaker and any arcing debris could not have caused any issues?

Is there anything common between the 2 incidents during the energisation period either time wise or with regard to certain switching actions on other systems?

With regard to the transformer, the windings will move on energisation, as John has pointed out they are usually braced well but I have seen bracing break or move due to forces from fault currents that have flowed through the transformer.

The final thing I can think of at the moment is was any testing carried out on the new breaker prior to installation / energising. I know it is remote but new apparatus can fail and you could be a victim of 2 independant failures.


The 3 guys are actually doing well - all out of hospital and 2 of the three requiring skin grafts. They have all been very lucky. Worth noting though that during the investigation we realised that they were not wearing the full arc flash suit but only the visor and gloves with "bib" overalls. As you can imagin the buring was on the upper arms.



I am glad they are on the mend, burn injuries are some of the worst type of injuries we can suffer and skin grafts are not easy to undergo - I hope all goes well for them.

Unfortunately you have hit on one of the problems I have with PPE, getting people to wear it. I have done extensive work on arc flash mitigation and engineered things down to people only needing to wear a balaclava, face visor and gloves along with their normal FR overalls. We even have it in small kitbags for them to carry around when they go an do some switching but I still find some not wearing it.

The arc energy levels at LV can actually be higher than that at HV so if your PPE is only rated for the HV it may be insufficient for the LV system anyway.

Hence my preference is to opt for engineering solutions.

Kind regards

Donald Lane
 02 January 2012 12:12 AM
User is offline View Users Profile Print this message



williamjohn

Posts: 178
Joined: 22 November 2010

Hi Mike

Further thoughts, your link to the ABB breaker specification states 85 kA breaking capacity at 600 volts. The fault clearance rating is root 3 times volts times amps and is therefore 88 MVA. If your transformer is no bigger than 700 kVA and impedance volts no lower than 4%, the maximum prospective fault MVA is 18 MVA (it is probabely much less). Therefore whatever type of fault the breaker is called upon to clear , it should be able to do it easily. This is what I could not understand.

This suggests there must be something wrong with the transformer.

My only suggestion is to disconnect the 690 volt connection on the transformer and fit voltmeters on the terminals between phases and phases to earth. Stand well back, switch on the 6.6 kv breaker and read the voltages. Take care, the voltage to earth may rise to several kv if the voltmeters are disconnected due to the capacitance coupling between the transformer windings.

Sorry but I cannot help more.
John
Further comment added 2 Jan. With all generators running, the prospective fault MVA on the 6.6kV system will be of the order of 180 MVA. If there is a fault between the 6.6 kv system and 690 volt system, then the ABB breakers will surely explode if they try to clear a 180 MVA fault. Could perhaps a connection between the systems have been initiated by one of your voltage spikes?

Edited: 02 January 2012 at 08:39 AM by williamjohn
 02 January 2012 03:09 AM
User is offline View Users Profile Print this message



michaelfivey22

Posts: 27
Joined: 21 January 2004

Guys,

Happy New Year.

Many thanks for your input. I will try and provide the answers to your questions.

The system has worked in the past but only for commissioning and load test.

The UV is long time rated and when the voltage applied the coil is energised and holds the breaker in - this is typical for most UVs I have came across.

A further confusion factor is that the first type of breaker that failed had a thermo magnietic type trip (with no CTs) and the second had a small electronic protection device with CTs. Both different generations of the same breaker I believe - all other specs and frame size identical.

One point worth noting about the installation is that the live side is on the bottom of the breaker and the load side at the top - reverse to what I am used to generally although ABB have confirmed this is ok as the prote3ction is non directional.

This would mean the trip device would have voltage on it all times even when the breaker is open and would be susceptible to any issues with the supply.

The alarming similarity between both incidents is that before each time the breaker operated the contacts were verified OPEN. The board was live and both times a megger reading checked the downstream cables for a fault. Each of the injured partied have reported barely touching the breaker before it exploded. The first instance the breaker was found to be tripped and trying to be reset and the second time the newly installed breaker was trying to be closed. For this reason I am starting to get away from the idea of a downstream fault. The condition of the breaker contacts support this.

The breaker with the small burning o the back is in a different compartment so well removed from the actual site of the failure.

We have a fault level study and the rating of the breaker - 70kA is good. The closing capacity is something like 143kA off the top of my head and the max assymetrical fault is 130kA or there abouts. So I am satisfied the breaker should be able to close onto a fault at that part of the network. If we were closing onto a fault that did this we would have found a full on bolted fault by now.

Analysing the data from the data logger in the automation system we can see some very strange things happening with the bus voltage on the 6kv network - some spikes are evident but we will need to get this verified by the automation and power system supplier.

The final theory I have for just now is that adjacent to the damaged breaker we have a crane feeder breaker. On this crane is a large harmonic filter. The electrician reports finding fuses pulled from this unit a few days prior to the first incident and replacing them. For those of you who are unfamiliar with ships they tend to run an island power network which consists of more than one generator in parallel. Switching in and out capacitive loads can cause real issues with reactive power and if large enough (and the AVRs and old / rubbish) then also system voltage as they cannot react quick enough to compensate for the change in load impedance.

It is possible that switching in and out this capacitor bank is causing havoc with the system voltage and has burnt out the trip device on both occasions. So when the breaker is operated it has subsequently blown up.

Reading through the reports of the system form commissioning it appears that system voltage spikes have been reported in the past - as you will no doubt know overvoltage spikes are hard to produce on a system running parallel generators unless we are switching some kind of massive reactive load in and out. A run away AVR for example would just affect the KVAr load sharing and cause another AVR to sit back.

We will investigate this tomorrow.

Does anyone else have any thoughts on this theory?

Thanks again for all the invaluable help.

Mike
 02 January 2012 10:42 PM
User is offline View Users Profile Print this message



dlane

Posts: 690
Joined: 28 September 2007

Hi Mike,

Happy New Year to you too.

From your response we can at the moment rule out the trip circuit and any down stream circuit as it seems that the breaker was never actually closed on eother occasion.

I have come across overloads connected to the supply before the breaker on numerous occasions and never had any issues with them but I agree with you that it leaves the unit susceptible to any voltage variations which brings us back to Arthur's origianl question of the rating of the trip unit phase to phase and phase to earth.

I would have thought the unit would have had a withstand voltage level but couldn't find it in the specifications unless it is part of one of the standards mentioned.

The crane filter, I guess that there is no documentation to say why the fuses were removed originally or who did it? I am agree with your explanation of how it can cause voltage spikes but what would be the duration of these spikes and what level, would it take it beyond the voltage rating of the trip unit?

I would though have expected the trip unit to have flashed over at the highest peak voltage so unless you have been unlucky enough to have electricians at the breaker during each highest voltage peak I would have thought the flashovers would have occurred as the filter was switched as a one off.

I suppose it would be possible for the trip units to be damaged by high voltage levels to the extent that they are brittle and pressure on the operating mechanism may have caused them to break and cause a short circuit. But we are now talking about the failure ot a thermal protective trip and an electronic trip which has removed a common element to both incidents.

Is this breaker unique amongst the installation or are there other breakers that can be inspected as if this one has failed due to common high voltage you would expect to find damage on other trip units etc.

Personally, unless you are 100% confident of the condition of the transformer I would not recommend its energisation, would it be possible to get someone like DobleTest on board to carry out high voltage testing on the windings?

I think that you are going to be reliant upon ABB to carry out some root cause analysis on the breaker before you are going to get any answers. With the limited knowledge I have of ships systems I can't really suggest any more avenues for you to investigate other than the ones you have come up with yourself.

Kind regards

Donald Lane
 05 January 2012 12:08 PM
User is offline View Users Profile Print this message



williamjohn

Posts: 178
Joined: 22 November 2010

Hi Mike

There are postings about non earthed systems in the ENERGY section.

You say the protection unit has been installed upstream of the ABB breaker. Thus a fault on the protection unit has to be cleared by the 6.6 kv breaker with possibly higher setting and longer clearance time.

The ABB beaker does not appear to be metalclad. Therefore an internal fault is not contained within a metal enclosure. Furthermore, touching the operating lever will alter the capacitance to earth. If there is corona within the breaker, this may alter the path of the corona and perhaps trigger a fault.

As the 690 volt system neutral is not tied to earth, there may be a high voltage to earth on the 690 volt system due to capacitance coupling to a high voltage conductor. Although this would be low power, it might trigger a discharge leading to the high power faults in the protection unit that you have experienced.

Good luck
John
 10 January 2012 12:22 PM
User is offline View Users Profile Print this message



cannella

Posts: 4
Joined: 18 January 2003

Hi Everyone and Mike.
Any unusual noise coming from the switchboard should be treated as if the board is in fault. The switchboard should have been isolated there and then upstream. To continue and reset the breaker could have been fatal. I sincerely hope the men are doing ok and are being looked after.

Tracking and arcing is easy to detect using IR/other detection equipment and this would have identified a hot spot possibly.
I assume you already conduct maintenance using IR on the switchboards while the ship is at sea and at peak load periods?

There could be many reasons for the failure other than component failure. Moisture, dust, vermin, structural vibration, water or sabotage.

The breaker in question in the first incident should have been removed and fully examined possibly by ABB. Did you strip it down completely? Any photos?

Conducting a forensic analysis of the switchboard would have taken place if there was a fatality (UK registered Ship); therefore a similar examination should really have taken place.
To repeat the fault second time and expose personnel is very unfortunate.
I suggest possibly looking outside the obvious for the cause. But if the cause is not found the switchboard should be taken out of service/replaced.
 13 May 2012 07:38 PM
User is offline View Users Profile Print this message



bgnath13

Posts: 1
Joined: 13 May 2010

Hi Mike,
We had an arc fault in our 690V Switchboard at the input side on outgoing Tmax MCCB feeder. No operation was carried out during that time and MCCB was on. Investigation has just started and the causes of the flash over are unknown. While going thru the internet for some similar incident, I stumbled up on your post. May be this is an unrelated case, but just curious to know the findings of your investigation and ABB's root cause failure analysis.
Thanks,
Gobi
IET » Other and general engineering discussions » ARC Flash Incident on Ship

<< 1 2 Previous Last unread
Topic Tools Topic Tools
Statistics

See Also:



FuseTalk Standard Edition v3.2 - © 1999-2014 FuseTalk Inc. All rights reserved.