CCVT Spark Gap 2

[System Protection]

I ran across this paper on dealing with CCVT transients.

[URL unfurl="true"]https://na.eventscloud.com/file_uploads/96d444004fb61dbb6c16cf985e94906d_PSE_CCVT_Transient_Investigation_2016.pdf[/url]

They saw a decrease in magnitude of the transients at the test switches after replacing the spark gaps.

Has anyone else noticed this? Should spark gaps be maintained when PLC is not in use?

I'm sure proper shielding would make the most difference here. Interesting though.

 

[FacEngrPE]

Attached "CCVT Transients Revisited, Roger A. Hedding ABB Inc."

I have seen the results of a PT explosion on a 23kV ungrounded circuit with the PT's connectecd in a grounded "Y". We concluded the problem was likely related to ferroresonance. Per the attached paper most, perhaps all CCVT's include ferroresonant damping, but it's effectiveness will relate to how well the damping is matched to the available resonant nodes in the system (including between the system and the CCVT).

Searching

https://duckduckgo.com/?q=potential+transformer+ferroresonance&t=lm&ia=web

pulls in a bunch of references that may provide useful.

 

[cranky108]

Most CCVT's with the carrier accessory have a grounding switch, which should be grounded when not in use. This, there should be no electrical degradation of the gap. That said, the gap and coil should be inspected and maintained prior to being put back is service, as they can accumulate dust, dirt, and corrosion.
Same thing with the voltage accessory. Ground it when it is not in use.

 

[Kiribanda]

If the VTs used on the 23kV ungrounded systems are not rated for the full line line voltage of 23kV, the VT will go into saturation due to the overvoltages
during a single line to ground fault on the 23kV system.

 

[System Protection]

@cranky108

They must be referring to the protective gap for the reactor in the paper and not for the PLC carrier. I think I just got that part confused. I'm looking at the CVO manual and the PLC gap is indeed in parallel with the Carrier Ground Switch.

 

[scottf]

Interesting paper, thanks for sharing. I knew of some of the problems PSE was having, but didn't know the details.

We (Ritz) discuss this issue in an older Technical Bulletin, as well as in our instruction manual (section 4.3 of the attached). It appears PSE adopted a lot of the recommendations listed, including grounding the neutral of the secondary windings in the control house, grounding both ends of the coax shield, running parallel conductors for shielding, etc... Only point they differed on what PVC conduit versus metal conduit.

I don't follow the comments on the spark gaps. As noted above, CVTs can have 2 sets of spark gaps (1) the P1-P2 gap that protects the compensating reactor from voltage build-up during switching due to the charging current of the capacitor divider and (2) the spark gap that protects the drain coil (if CVT is provided with carrier accessories) from the same voltage build-up during switch. I don't believe these gaps would play a significant role in transients imposed on the secondary cabling as the paper seems to indicate.

@cranky - there is no need to ground the carrier ground switch when carrier accessories/circuit is not in use. The drain coil provides 60Hz ground reference for the divider. The carrier ground switch is there to provide a means to ground out the carrier circuit if work is being done on the carrier equipment with the line energized.

The paper doesn't mention if any testing was done to confirm that the surge protection device placed across the secondary winding does not interfere with the ferroresonance suppression circuit of the CVT. We caution against placing non-linear burdens on CVTs. Hopefully, they tested this device at the CVT manufacture's factory to confirm that ferroresonance suppression circuit still functions correctly.

 

[scottf]

Forgot he other bulletin....

 

[System Protection]

Excellent info Scottf.

I was hoping you would show up.

I've witnessed the improvement grounding the shield at BOTH ends makes for CCVT transients with an oscilloscope.

Issue is the IEEE 525 recommends running bare copper conductor (usually 2/0 or 4/0) in parallel to cables with shields grounded at both ends in order to protect the shield during ground faults. This is not a quick retrofit.

I've been exploring other possible contributing factors as I've got some transient issues to solve. Thats how I ended up on spark gaps.

How many utilities routinely test/replace these? I'm tempted to replace it and get the scope back out to see if it makes any difference.

 

[FacEngrPE]

VTs used on the 23kV ungrounded systems are not rated for the full line line voltage of 23kV

I am aware of this risk.

 

[scottf]

System Protection

I understand your comments in grounding shield at both ends.The induced transients is a fairly rare issue and with the customers I know that have the issue, I believe they've had good success with grounding on both sides and I don't think they out in add parallel conductors, but I don't know that for sure.

As for the spark gap, it depends on what kind of gap you have. I'll preface the next comments with...I'm just giving my opinion here and not attempting to sell. The Ritz designs (GE current type OTCF and Ritz type CVO) use air gaps that are visible and easily inspected. The OTCF used spark disks and the CVO design used bolts with rounded heads. There is no real failure mode unless they've been subjected to continuous sparking, which would be a result of some other component failure. So...with our design I'm not aware that anyone changes them out.

The Trench design uses a gas filled gap that resembles a small fuse. The state of this gap is not easily visible. I have heard of the need to change those gaps out. With that said, it's still not clear to me how the spark gap would be causing induced transients.

 

[System Protection]

It is indeed a gas filled gap. (Trench)

It's odd as it's only a 115kV station. Getting 1MHZ 1200V+ transients on other single phase CVT circuits when energizing the aux bus CCVT via Tie Breaker. I imagine energizing via a disconnect would be even worse.

Cable is shielded but shield tails are way too long and only grounded at one end. I'll start eliminating variables one by one. At least until effort surpasses the benefit.

 

[scottf]

How is the quality of the grounding at the station?

In our experience these induced transients seem to happen in locations with poor grounding conditions/soul resistivity issues.

 

[cranky108]

I don't download papers as a general rule. But, between the CCVT and the tuner in a carrier circuit, that is the high impedance point, and capacitance at this point to ground will degrade your signal. So I am wondering if that is why a plastic conduit might be recommended. Years ago we used standoff insulators. It also has a potential for a high voltage.

Speaking of grounding switch, it was recommended that unused carrier ground switches be grounded because of the potential for high voltages to develop across the coil, or in some designs without coils (why someone would do this is beyond me).

Most tuners will also have a coil and spark gap. The oldest designs I recall had a Champion spark plug. But the spark gaps can over time become dirty, or may need adjustment (if they ever were adjusted in the first place). It is the flashover of these gaps that cause carrier holes.

 

[System Protection]

The ground grid was re-studied not too long ago.

The soil resistivity was averaged around 600 ohm-meters but was tested when the ground was wet.

I'd love to just ground the shields at both ends but the portion in IEEE 525 about using a ground conductor parallel to the shielded cable makes me hesitant to do so. Not sure how many utilities follow this practice. At a previous utility we did not specifically run a parallel conductor but the shields we used were much thicker than typical.

 

[scottf]

Cranky-

In former times when we were in most aspects of carrier system (CCVTs, line traps, and line tuners) we would recommend that the carrier cable from the CVT to the line tuner be 5kV stranded cable and it be run on stand-off insulators. If run in conduit, it should be plastic and not metallic due to the capacitance impact you mentioned...and it still needed to be set-off from the structure on the run down to the line tuner.

System Protection-

I don't have very good insight into if other utilities run parallel conductors in these cases. As I mentioned above, in my `29 yrs of experience with CCVTs, I can only recall a handful of customers that reported issues with secondary induced transients.

 

[cranky108]

If you are talking about cables between the tuner and the control shack, usually there is a heavy copper ground in the cable trench that follows all the cables, and is connected to the ground grid here and there. So why the concern?

 

[System Protection]

@Cranky108

Everything is run in conduit and I'm not aware of any ground conductor ran along side (or in) the conduit.

If both ends of the shield were to be grounded there would still be continuity to the ground grid on each end.

That is not quite the same as what is recommended though.