Industrial Facility Power Quality Issues

[aelectricalman]

I've worked with a lot of power quality issues at a lot of different facilities, but I have never seen this before. I'm hoping someone has had this problem before and can shed some ideas on how to approach finding the culprit, so that I don't waste an incredulous amount of time.

Back-story: Facility frequently burns drives, trips breakers, burns transformers up and they have even melted the wiring on all of their CT's and PT on the power logic's in their gear, causing fires and shutdowns. There are two 1000hp DC motors and three 300hp AC motors among the two 2500KVA pad mount transformers (side by side). The facility is an ungrounded 3 wire delta. The power company after the last transformer burn up put back in two dual transformers capable of being delta or wye. The system is still delta however. The neutral paddle is not being used just FYI.

Measuring the PQ in the 2- 2500 KVA transformers (Reliable Power Meter) and independently in the main gear in the mezz (with a Dranitz and a Fluke 43), we confirmed that there is the following;

Transformer 1 = Phase A - 472 V - L to G
Phase B - 473 V - L to G
Phase C - 2.81 V - L to G

Wait......wait.........

Transformer 2 = Phase A - 2.114V - L to G
Phase B - 481.8V - L to G
Phase C - 481.8 V - L to G


Both XFMR's have a L to L voltage of 474V and 482V respectively.

Notes : There is no paralleling of the XFMR's and there is no connection of the two XFMR's past the metering points (That I know of). Any ideas as to what would show its head like this? I'm open to brainstorming and suggestions! I'm going back in the AM. We are going to look at Z of the equipment and inspect for ground loops.

 

[aelectricalman]

I meant to say there is nothing in common with the xfmrs past the 12.47 pole that I know of, not metering point.

 

[gar]

130805-2218 EDT

Chris:

A number of questions and comments:

1. What is G? Is it a screwdriver stuck in the earth, nearby or far away? A grounding electrode, or system? The building steel? Or an EGC bus in a panel?

2. Is there an extremely low impedance between the Gs of the two transformers?

3. Is phase A of transformer 1 in phase with phase A of transformer 2? Or is phase C of 1 in phase with phase A of 2?

4. Why is one phase of each transformer close to the potential of its G, or both Gs?

5. There is no use to a 3 place resolution on the phases that are essentially grounded. On phase to phase voltages are the measurements really within +/- 0.5 V?

6. What burned up in the current or potential transformers? Did all or just some burn up?

There is a lot more detail that is not present.

.

 

[__dan]

One phase of each service is coupled to ground. Maybe not solidly or impedance grounded, but coupled. It is a bad sign that they are different phases for each service. Same phase would look better, but you would still want to know why.

There is no downstream service tie breaker or tie to emergency generator bus? There is likely some dual fed bus downstream where both sources are present.

Assuming the large drives have delta fed, 3 phase, 3 wire fed, rectifiers at the input, no neutral as indicated - if the drive filter caps, input or output, are internally Y connected and grounded, that gives you a phase coupling to ground.

And any reactive imbalance (aged / failed caps) in the cap banks will seek out its reactive balance somewhere else in the distribution system, that I would say may be the source of any large ground loop currents found. It's a reactive flow over the ground paths between net reactive elements in the system, capacitive and inductive elements finding their balance point. In a facility with large UPS (a DC), I would look for it circulating between the UPS rooms.

Burning up transformers would point to high harmonic loads (nonlinear loads), higher frequency distortion of the 60 cycle sine wave, and high frequency transients (switching events). There could be more than one unrelated problem.

Filter caps are a wear item after 5 to 8 years. That and transient suppression is something to look at.

 

[__dan]

With the floating delta, if you're monitoring I would want to see the floating phase voltage to ground over time. If the phase voltage moves around a lot or goes overvoltage in transients, you may want to consider referencing the service Y secondary to ground intentionally.

 

[Sahib]

It may be arcing ground fault causing over voltages leading to equipment damages in your ungrounded delta system.

Most probably no ground monitoring and alarm indication was installed in your system.

To check and find the location of existing ground faults in your system, portable high resistance grounding equipment available in the market may be used.

 

[aelectricalman]

130805-2218 EDT

Chris:

A number of questions and comments:

1. What is G? Is it a screwdriver stuck in the earth, nearby or far away? A grounding electrode, or system? The building steel? Or an EGC bus in a panel? ........EGC and Earth ground on meter base. Tested at both locations

2. Is there an extremely low impedance between the Gs of the two transformers? ......will investigate.

3. Is phase A of transformer 1 in phase with phase A of transformer 2? Or is phase C of 1 in phase with phase A of 2?.......the anomolies are not on the same phase.

4. Why is one phase of each transformer close to the potential of its G, or both Gs? ......thats the million dollar question.

5. There is no use to a 3 place resolution on the phases that are essentially grounded. On phase to phase voltages are the measurements really within +/- 0.5 V?

6. What burned up in the current or potential transformers? Did all or just some burn up?

There is a lot more detail that is not present.

.

see my response embedded in the message.

 

[aelectricalman]

Seems like phase to ground should not be 480 on the other 2 phases.

 

[Besoeker]

Seems like phase to ground should not be 480 on the other 2 phases.

It would be if it's a corner grounded delta..

 

[aelectricalman]

The power logics in the gear are all burned. There is another gear in the buolding that has 480/277 (in the front of building) that has working power logics. Btw there are no problems in this gear or on the system.

 

[aelectricalman]

It would be if it's a corner grounded delta..

Thats exactly what I told thebpoco rep. He said there is no corner grounded here.

Technically there could be if they hooked it up wrong. Smh.

 

[beanland]

Delta Systems

Delta Systems

I have worked on several ungrounded delta systems. You appear to have two different phase-ground problems (not faults because there is no current.) Because you have no grounding transformer, there is no ground fault current to trip circuit breakers. NEC requires a ground detection system on these facilities. Does it exist?

The problem with an ungrounded system is that the voltage on the ungrounded system can go extremely high due to capacitive coupling through the transformer. I have seen surges/lightning on the POCO lines go right through the transformer capacitance and cause failures on the "ungrounded" system.

Fixes: (1) you need ground detection per NEC. (2) adding a grounding transformer that can provide enough current to trip breakers may aid in locating the phase-ground "faults." (3) if this is a 480V system, you may be able to mid-side ground (480/240V) to get fault current. (4) if this is a 480Y/277V transformer with the X0 disconnected, you could resistance ground the X0 to give limited fault current.

Good luck.

 

[GoldDigger]

Thats exactly what I told thebpoco rep. He said there is no corner grounded here.

Technically there could be if they hooked it up wrong. Smh.

You need to find the actual impedance to ground of the low phases
It could be a POCO issue, an accidental bond or a fault in connected equipment. You can tell the whether it is a solid ground, a resistive. ground or a parasitic ground by connecting a test load from one of the high phases to ground. If it really is an intentional ungrounded delta, there must be a ground detector for safety and compliance.

 

[al warner]

The problem is these are ungounded 480V delta systems.

The problem is these are ungounded 480V delta systems.

The power logics in the gear are all burned. There is another gear in the buolding that has 480/277 (in the front of building) that has working power logics. Btw there are no problems in this gear or on the system.

The issues and conditions above and in your original post are classic reasons that an ungrounded system is not recommended practice by most knowledgeable electrical engineers.


The RPM data clearly shows the two "floating systems" each have a phase fault to ground or partial fault to ground, albeit on different phases. This is a disaster waiting to happen......... the next phase to ground fault on one of the other phases will likely recreate the same or similar damage. Being they are ungrounded, its very easy to generate standing waves in the circuit conductors that can easily exceed 2-3 times the 480V rating or even more. The NEC may permit ungrounded systems, but it also doesn't tell us they are stupid! If you want to blow up equipment use ungrounded systems.

Be very, very careful! I would recommend they shut down both of the systems and install a 10A HRG with pulser capability on each transformer secondary - HRG corner grounded with appropriate monitoring etc.. That will stabilize the voltages and also allow them to find the existing faults.

Good luck!

 

[aelectricalman]

With the floating delta, if you're monitoring I would want to see the floating phase voltage to ground over time. If the phase voltage moves around a lot or goes overvoltage in transients, you may want to consider referencing the service Y secondary to ground intentionally.

Thats a good idea. Something I just found through keeping my ear to the street is that the facility was originally a delta system. Then the poco replaced transformers when the two failed with dual wye or delta type and actually hooked them up wye. Then through the years the facility was having some issue, then attributed them to the wye system (having hooked all of their new drives up in a wye config..) . Then the poco comes back out at facilities request and changes back to a delta. Do you see the problem here.... I suspect that what was happening is all these drives that were hooked up wye config have burned up and been replaced with a delta config.

 

[aelectricalman]

Update: Thanks very much for all the help, everyone. So far I have found that the system infact was never at any point converted to a wye. Forget I said that. Basically the pole mount ungrounded delta transformers burned up. The utility could not replace them with pole mounts so they went to pad mounts. The transformers they used are for delta or wye. They were configured for delta and made to be ungrounded delta. The n0 bushing is not referenced to ground and of course there are no grounds attached.

Upon further investigation the system is operating like a corner grounded delta. The utility company says that something in the facility is causing this and they do not provide corner grounded deltas in their tariff structure. The reference voltage to ground is on the C phase on one xfmr and A phase on the other and this was verified by phase rotation meter at both meter sockets and xfmrs.

All drives that we inspected had delta /wye iso transformers on them for clean power and for ground reference for drives. One of the 1000HP drives was fed from a delta/wye xfmr but had the 750 kcmil ground cut out of it. It was still referencing 266V to ground but had 40 ohms on the ground. So they were operating with a very poor ground. Oh fyi, the iso xfmr is 460Y/266 .

All of the power logics were burned up. 6 of them from two secondary xfmrs. There was one ground fault limiting device on 1 of 3 parallel gears on each xfmr. My assumption is that will take care of the entire xfmr if installed properly. Please refute if not accurate. PQ studies will be in monday on both xfmrs.

In one the most problematic secondary xfmrs, the facility uses the rigid conduits going to the gear to bond and ground the system, in leui of pulling a ground through the conduit. That is to be expected since it is designed to be an ungrounded Delta. Then someone has come in later and run a 750 kcmil ground from the main gear to earth, out next to the xfmr. I found that the resistance on these grounds are acceptable but its not supposed to be a grounded system. Have I stumbled onto something or am I barking up the wrong tree? Could this somehow cause a corner ground?

 

[aelectricalman]

Thats a good idea. Something I just found through keeping my ear to the street is that the facility was originally a delta system. Then the poco replaced transformers when the two failed with dual wye or delta type and actually hooked them up wye. Then through the years the facility was having some issue, then attributed them to the wye system (having hooked all of their new drives up in a wye config..) . Then the poco comes back out at facilities request and changes back to a delta. Do you see the problem here.... I suspect that what was happening is all these drives that were hooked up wye config have burned up and been replaced with a delta config.

Please disregard this. This information is incorrect.

 

[aelectricalman]

Update and need of clarification

Update and need of clarification

Assuming the large drives have delta fed, 3 phase, 3 wire fed, rectifiers at the input, no neutral as indicated - if the drive filter caps, input or output, are internally Y connected and grounded, that gives you a phase coupling to ground.

Very interesting because the secondary pad mount XFMR's on both of the services are delta ungrounded. Then, the delta is transformed in the building (with an iso delta/wye). So, at the drive itself, I am able to read...

465V L -L on all three phases
266V L -G on all three phases

But the interesting part is that the only EGC or other ground that is feeding the drive is cut off in the conduit. It does not have a good ground on the 1000hp equipment. Somehow the grounding at the drive is strong enough to pick up 266V L-G but the resistance is 35ohms. They are picking up a little bit of ground from the building steel on the load side of the drive as the conduits pass through down to the motor. All the bonding is in place but it does not have a solid path to ground, (not that we need it).

The interesting problem that I am facing is.....how does a stepped down wye create a corner grounded delta back in between the secondary XFMR and the Iso XFMR. Wouldn't the phase couple to ground be on the secondary side of the wye system, not before the iso XFMR side? Maybe I don't understand enough about the inductive nature of whats going on in the iso XFMR.

 

[__dan]

The interesting problem that I am facing is.....how does a stepped down wye create a corner grounded delta back in between the secondary XFMR and the Iso XFMR. Wouldn't the phase couple to ground be on the secondary side of the wye system, not before the iso XFMR side? Maybe I don't understand enough about the inductive nature of whats going on in the iso XFMR.

Coupling of the service supplied floating delta is happening, but I don't think it would be caused by the delta to Y, iso xfmr, at the drives. The preferred phase through the transformer would be changing all the time, reading as a voltage to ground. I'm looking at Golddigger's post where he suggests a test impedance could be connected phase to ground to test the impedance of the other phase to ground coupling. If the delta is truly floating, a small impedance should be able to pull it to ground. If the other phase reference to ground is bigger, you would want to know what is causing it. And tested very carefully as a phase to ground fault or defective equipment would be one of the suspects.

But the interesting part is that the only EGC or other ground that is feeding the drive is cut off in the conduit. It does not have a good ground on the 1000hp equipment. Somehow the grounding at the drive is strong enough to pick up 266V L-G but the resistance is 35ohms. They are picking up a little bit of ground from the building steel on the load side of the drive as the conduits pass through down to the motor. All the bonding is in place but it does not have a solid path to ground, (not that we need it).

If the EGC to the drive was cut, I would be afraid it was cut intentionally. I would ask the maintenance staff this, if they knew it was cut and why. If they cut the egc because of observed high current, that current is still flowing, but over the conduits and building steel where they cannot see it heating or measure it. I would guess again that if there's high current on the egc, it is a reactive flow from a reactive imbalance in the drive. Something that may be native to the drive but could indicate a problem.

If the drive iso xfmr Y secondaries are solidly grounded, I would want to see an amp clamp reading of the system bonding jumper to see any imbalance current flowing off the neutral to ground. Not "to ground", but off the neutral onto the facility bonded metal structure and primary side egc.

If you see something that doesn't look right, the drive xfmr Y secondary having a high impedance to ground, I would go to the drive manufacturer and ask them what the neutral is required by them to be, solidly grounded or other. I'm thinking the drive iso xfmr's are equipment supplied or specified by the drive manufacturer. I would confirm this and confirm the condition of the neutral, if it has neutral imbalance current, neutral to ground current, and if the quantity is more than allowed by the drive manufacturer, possibly indicating a problem.

 

[GoldDigger]

The interesting problem that I am facing is.....how does a stepped down wye create a corner grounded delta back in between the secondary XFMR and the Iso XFMR. Wouldn't the phase couple to ground be on the secondary side of the wye system, not before the iso XFMR side? Maybe I don't understand enough about the inductive nature of whats going on in the iso XFMR.

If you leave out reactive (both inductive and capacitive) coupling, you can create an ungrounded delta from an isolated wye secondary just be not grounding any phase or the neutral wire.
Or you can create a corner grounded delta by grounding only one phase terminal and letting the center point float.

If you create an ungrounded system, then any small combination of unbalanced fault or parasitic reactance to ground favoring any single phase can move the whole delta in that direction.