LV generators on same bus - One neutral grounding?

[digitrex]

I am studying reliability of an old installation and I am asking the client a question on why all four 415V generator sets have their neutrals solidly grounded(3 sets in normal operation).

Can you give me opinions on whether there is a way to mitigate the flow of unbalanced current through the neurtral groundings, or I must advise the client to maintain only one grounded neutral point?

 

[waross]

The preferred way to connect the sets is to connect the generator neutrals to the neutral bus. There should then be one connection from the neutral bus to the ground grid.
That said, I have encountered old installations where the neutrals were individually grounded. The load was two transformer banks with delta connected primaries. There was no system neutral. The generators were a mix of two 600 kw star wound sets with grounded star points and three 350 kw delta wound sets with no grounding. In normal operation one of the 600 sets was usually on line and providing a ground. The ground would only pass current in the event of a ground fault in one of the transformers or one of the generators at which time it would hopefully pass sufficient current to trip a generator breaker.

It was very unlikely that a 350 kw generator would be running alone.
If a 350 kw was running parallel with a 600 kw set and there was a ground fault, the breaker on the 600 kw set would trip. At this point there would no longer be a ground path but the load would be too much for the 350 set, and it was a race between the breaker tripping on overload and the set slowing down to the point that the underfrequency relay shunt tripped the breaker.
It was not the best arrangement, but I inherited it and we were lucky. In the life of the plant, we had no incidents related to the grounding issues, and the old sets have now been replaced.
Re unbalanced currents. With delta connected loads, you should not have any ground or star point currents.
An exception is if the generator winding pattern is not identical. This can result in the generators generating slightly different waveforms. In this case, the most common result is small currents between the star points.
If you have line to neutral loads, the star points must be connected to a neutral bus and one and only one connection made from the neutral bus to the ground grid.

If your system has been in operation for some time;
If you have only delta connected loads and no line to neutral loads of any kind.
If there is no objectional current over the star point connection to ground. (A 60 hz. current may be an indication that some thing is failing to ground. A low level current at 240 hz. or higher is usually a sign of mismatched wave forms.)
If you are not subject to stringent codes.
Leave the system as it is, but inform the owners that the system is in non-compliance with current practice and codes but is servicable. Include a rider that any future modifications may render the system unsafe.

If you find any loads connected phase to ground or phase to neutral.
If you wish to monitor ground curents on the system.
If you must adhere to codes that mandate one single connection from neutral to ground.
Then instal or designate a neutral bus and cable the star points to the connom neutral bus and thence to ground with one connection only.

What I visualize, and what I inherited was a plant with an adequate ground grid. Each generator frame is connected to the ground grid independently. The generator star points are then connected to the generator frames.

If your star points are cabled back to the control room and connected to a common bus, then this may be a neutral bus or a ground bus. If the bus is or can be insulated from ground it may then be designated as a neutral bus and the one connection made from this bus to the ground grid. Note, the "one connection' may be parallel cables as long as they originate on the same bus and terminate at the same place on the ground grid. Alternatively, the connection may be an impedance rather than direct cables.
respectfully

 

[digitrex]

Waross, thank you for the detailed information and story.

I have more findings on the installation:

1. All generator neutrals are brought to the switchboard and connected to the earth bar there. However, the earth bar is not isolatable because all other components in that switchboard, including the switchboard frame, are also connected to it.

2. For each generator, there is an isolation switch between the neutral and earth bar. A CT is at the grounding is provided for earth fault protection relay (51N). The isolation switch has an auxiliary contact which is used for shutting down the generator upon opening of the isol switch.

4. There are a mixture of 3-phase loads (motors & UPS), and 1-phase loads (lightings, office equipment etc).

5. A readout from the digital protection relay of current at the grounding CT indicated a current of 30A flowing through the ground.

6. There is a new recommendation for generator earth fault protection setting to be lowered down to 100A (it was previously set at 250A).

I foresee a problem if there is a downstream earth fault that may escalate the neutral grounding currents and it will trip the generator.

 

[waross]

Hi;
I'm a little short of time right now.
I'll get back to you as soon as I can. Reading quickly, it appears that you are measuring a mixture of neutral currents and ground currents with your ground fault sensing.
What are the sets rated at? KVA, maximum allowable current?
respectfully

 

[digitrex]

No, I am very sure that it is measuring the ground current because the CT is located just before the earth bar and after the T-off from neutral.

The gensets are all rated at 300kVA, rated current 417A.

 

[waross]

Hi digitrex;
I may have made an assumption when I asked a question.
Do you have any single phase loads that are connected from phase to neutral?
Do you have one ground measuring ct per machine, or one ct for the whole system?
You mentioned a "T-off from the neutral".
Do you have a neutral bus as well as the ground bus? Can you explain a little more on this point please?
Is this a stand-by installation or the prime power?
Thank you.
respectfully

 

[digitrex]

Q: Do you have any single phase loads that are connected from phase to neutral?
Ans: Yes, see Item 4 above.

Q: Do you have one ground measuring ct per machine, or one ct for the whole system?
Ans: One ground measuring CT per machine. The CT is mounted between the isol switch and the earth bar.

You mentioned a "T-off from the neutral".
Q: Do you have a neutral bus as well as the ground bus? Can you explain a little more on this point please?
Ans: The switchboard has both neutral bus and ground bus. Let me give you a clearer picture: 2 cables from generator neutral to the switchboard, one connected to Neutral bar, the other connected to the earth bar through an isol switch. CT is mounted on copper bar linking isol switch to earth bar. Somehow I found another grounding cable from terminal of neutral bar to the isol switch. This configuration is the same for the other 3 generators.

Q: Is this a stand-by installation or the prime power?
Ans: The installation is primary power, on an offshore platform.

 

[waross]

Hi digitrex;
An explanation; single phase loads may be connected phase to neutral or phase to phase. Different practices are used in different parts of the world. It is important to know that the single phase loads are connected phase to neutral rather than phase to phase. Thank you for your reply.

I hope that I don't offend you with simple questions but it is important that I don't make any assumtions. It is important to verify my understanding.

I will call the generator neutral the generator star point. This is at the generator.
I will refer to the system neutral bus and the system ground bus. These are in the switchboard.


As I understand your connections;
1: There is a cable from the generator star point to the system neutral bus.

2: There is a cable from the generator star point to a disconnect switch, from there to a CT and from there to the system ground bus.

I understand that you have found another cable from the system neutral bus to the ground isolating switch. Is this a third cable?

I get the feeling that this is a fairly old installation. Is this correct.

Have any of the generator been rewound at any time?

Is the single phase load on one phase or is it distributed on the three phases?

Have the loads been checked for grounds?

Is there a posibility that a load has been connected from phase to ground rather than phase to neutral?

It is common for all the cables used to connect a generator to be the same type. If cables are added at a later time they are often not identical to the original cables. Is there any indication that either the neutral cables or the ground cables were added after the original installation?

From the information I have so far, it appears that the neutral/grounding arrangement has been changed improperly.
Subject to further information, I suspect that you have one type of grounding/neutral system superimposed on top of another type of system.

It would appear that the generator star points are all connected to the system neutral bus.
This will result in any current between the neutral and ground being being shared by all the CTs but not equally.

The division of current will be inversely proportional to the impedance. In a perfect world, the impedance will be proportional to cable lengths. In the real world, ageing connections and oxidation on switch blades and connections may increase impedance.

In case of a fault in the system, the impedance from the system ground bus in the switchboard to the individual generators will be the major factor in the division of current.

For a fault in a generator, the fault current in the CT of the affected unit will be dependant on the impedance of the ground path (through the generator frame and steel platform frame to the ground bar at the switchboard.) and the cable from the ground bus to the CT and thence to the ground isolation switch at the generator and back to the generator star point.
The other CTs will see a parallel but lesser ground current that will pass from the system ground bus to the CTs and thence to the isolation switches and generator star points of the unaffected generators. The current will then pass back on the neutral cables and recombine on the system neutral bus. From there the ground current will pass on the neutral cable of the affected machine to the generator star point of the faulted generator.
If ther is a neutral bonding jumper in the switchboard part of the fault current will flow throught the CT of the affected machine and part will be shunted directly to the machine via the neutral cable.

Again, subject to further information, it appears that the ground cables and the neutral cables are sharing the neutral currents and the ground currents (if there are any actual ground currents).
Thank you for the answers to my questions so far.
Respectfully

 

[digitrex]

Thank you waross. You are very precise in getting details and I really appreciate it.

Now, answers to your questions:

1: There is a cable from the generator star point to the system neutral bus.
ANS: Yes

2: There is a cable from the generator star point to a disconnect switch, from there to a CT and from there to the system ground bus.
ANS: Yes

I understand that you have found another cable from the system neutral bus to the ground isolating switch. Is this a third cable?
ANS: Yes, this cable is about 2ft long and I believe this cable was added in.

I get the feeling that this is a fairly old installation. Is this correct.
ANS: Yes, 2 gen set were more than 20 years, and 2 gensets about 6years old. The switchboard was 6 years old, an upgrade to the old installation.

Have any of the generator been rewound at any time?
ANS: Not sure.

Is the single phase load on one phase or is it distributed on the three phases?
ANS: All the single phase loads are distributed on the 3 phases. They are all connected phase to neutral.

Have the loads been checked for grounds?
ANS: Downstream feeders have their own earth leakage detection.

Is there a posibility that a load has been connected from phase to ground rather than phase to neutral?
ANS: I don't think so. The ground is the platform's metal structure. All cables to load have neutral connections.

It is common for all the cables used to connect a generator to be the same type. If cables are added at a later time they are often not identical to the original cables. Is there any indication that either the neutral cables or the ground cables were added after the original installation?
ANS: The neutral cables are black, armoured cables, about 100m from 2 of the older generator stars, 30m from the other 2 generator stars. The 2ft added cables are earth cables (green/yellow in color)

I think I can guess what had happened in the past. Would like to hear more adise from you. Thanks.

 

[waross]

This is starting to remind me of a pumping station that we once installed. We got the strong impression that the design engineer was inexperienced with this type of installation.
The service was 480 volts with a solidly grounded wye point.
For some unknown reason an artificial neutral was added to the switchboard. (Probably a zig-zag transformer but I didn't investigate further.)
This was of course, completely superflouous on a solidly grounded system.
To this was added ground detector lights. These of course would never indicate a ground. They may be useful to indicate a missing phase.
Then there were the ground detector voltmeters. Same comments as the ground detector lights.
There were also phase to phase voltmeters.

Back to your system.
We haven't yet discussed the currents that I understand you are measuring on the ground cables. I wanted to get a clear understanding of your system connections before going into that.

Some comments before investigating the ground currents.
It appears as if a three wire, three phase system with ground fault monitoring has been combined with a four wire three phase system.
The action of the the ground protection will be indeterminate.

My understanding of the justification for monitoring earth fault currents is to trip a faulted generator off line before an internal ground fault damages the stator as well as the winding.
This is more important with larger machines. With 300 kw machines the expense of the protection may well have exceeded the cost of a replacement generator end.
The ground fault typically operates after a winding has failed to prevent damage to the iron, but at this time, the winding is already in need of repair, usually a rewind.
I had a conversation with the sales manager of a Caterpilar dealership a few years ago in regards to rewinds. He told me that the policy of their branch was to replace a generator end with a new unit rather than rewinding. We were discussing a 1600 kw set. In this case, damage to the iron is unimportant if policy is to replace with new.
I have seen 350 kw sets and 600 kw sets succsesfully rewound in other environments. Fortunately the burn-outs did not damage the stators. The units did not have ground fault protection.

In your installation, the overlap of the ground fault protection and the neutral connections may have been done when the switchboard was upgraded, but what is in place now is more important than when it happened.

What are the options and implications?
1. Do nothing. Report that the ground fault protection is compromised by the neutral connections. The grounding protection system is probably seeing a combination of an indeterminate part of both the neutral currents and any ground currents.

2. Change the system to a solidly grounded system. To do this, I would use one of the exsisting cables to ground the frame of the machine. I would use the other cable to connect the generator star point to the system neutral bus. I would then make one connection from the system neutral bus to the system ground bus. I understand that the connection from the generator star point to the system neutral bus is already in place. I would leave it as is.
I would then remove the other connection from the generator star point to the isolating switch. I would also remove the isolating switch and the short grounding jumper. I would connect the cable from the switch to the system ground bus, to the generator frame for supplementary grounding. (It's there, why not use it?)
It is debatable whether or not to monitor current from the neutral to ground on the jumper. This is not to say that it shouldn't be done but rather it should be discussed.
If the ground current is monitored it is again a matter for discussion what to do with the information. Do you want to shut the system down, or merely send a warning alarm?
You may choose to add window type CTs to this option. If you pass all the generator conductors including the neutral conductor through the window you will be able to pass neutral currents, but discriminate ground currents. You will be able to use the existing relays.

3. Restore the ground current protection and use a delta/wye transformer to supply the single phase loads.
To do this, I would remove the neutral cable connection from the transformer star point. I would also remove the short jumper. This would leave the generator star point with only one conductor connecting it to the isolating switch and from there to the CT and thence to the system ground bus. The original ground protection system should then function properly.
I would then reconnect the neutral cable from the system neutral bus to the system ground bus and use the other end as a supplementary ground to the generator frame. (It's there, why not use it?)
I would then instal a delta/wye transformer of sufficient capacity to supply the single phase loads. The existing system neutral bar will probably be suitable for use as a system neutral with the transformer. Again, one connection from the neutral to the ground.

I can't honestly say which of these options I would recomend. It depends to some extent on the wishes of the customer. How is the budet? Are the CTs and the 51n relays in good shape or must one or more be repaired, recalibrated or replaced? Does the customer have company standards that mandate a particular type of connection and/or protection.
I would tend to the solidly grounded system with the addition of window type CTs. to maintain ground protection. You have the 51n relays (It's there, why not use it?) Some discussion as to co-ordination and whether to trip or alarm.

I would not suggest ground current protection for 300 kw generators because of the relatively small size, but if the customer particularly wanted it I would agree. In this case having the equipment already installed is a factor.

Re the ground currents;
Do you have a DMM with frequency capability? Try to measure the frequency of the ground currents. If you have ground fault protection functioning on all loads, it is unlikely that you have any ground current. (If the ground fault settings on the loads are high enough to sum up to the current that you are seeing, it may be ground leakage.)
However, I suspect that the current on the ground is either neutral current or circulating current.
Try for a frequency reading.
repectfully

 

[electricpete]

Can someone explain to me what would be wrong with connecting both generators individually to the ground grid?

(as long as ground grid is good, it seems like the same configuration electrically)

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[ScottyUK]

Hi Pete,

Assume you have two generators operating in parallel. In the scheme you envisaged with the neutral point of each generator connected to earth, a parallel path is formed by the neutral conductors and the earth connections and the earth grid. It might help to draw the connections out as a mesh of interconencted impedances to see where the parallel path is formed.

Any imbalance current which would normally flow in the neutral conductor will divide between the neutral conductors and the parallel path through the earthing system. It can play havoc with differential relaying, and a current-carrying earth connection is prohibited, at least by the UK regulations.


----------------------------------

I don't suffer from insanity. I enjoy it...

 

[ruggedscot]

Ive seen the following :-

three large generators two smaller ones - all linked to a common bus, the neutrals are switched on all five gensets so that the first to start and go to bus becomes the main set and this set has the neutral close, All the others operate with the neutral breaker open but closed onto bus. It works fine and you know we dont get any circulating currents. Obviously the neutral is fully rated for the load we supply and one set can provide for this.....

kinda neat to see and impressive when it starts in anger the sets all start and switch to bus and then the load is applied and you hear the engine note change with the load and the turbos spool up. The revs obviously staying the same.

Rugged

 

[waross]

Hi electricpete;
ScottyUK has said it quite well. Thanks Scotty.
Hi ruggedscot;
Those circulating currents may still be there, phase to phase rather than phase to neutral, but masked by the load currents.
respectfully