Grounding Grid 3

[burntcoil]

Hi,

One of our client is looking to check the effectiveness of its 30 year old grounding grid. Its basically a gas plant having multiple substations, power from multiple sources with total power demand over 150MW. What is(are) the best way(s) to analyse the grid integrity?

Thanks

 

[wbd]

There are ground grid testers on the market or you can higher a company that does ground grid testing to test the integrity of the ground grids. It it is an older ground grid that had mechanical connections instead of exothermic weld connections, I would not be surprised to find issues.

Additionally if the fault currents have increased since the grid was designed and installed, the step and touch potentials may be unacceptable.

 

[krisys]

wbd, if the ground rods are corroded, you cannot identify from the earth tester. It seems burntcoil's client may be interested to know, whether the earth grid conductors and rods are corroded or in good shape.

In my opinion, it may need digging out few rods and grid conductors on a sample basis and inspect the condition of the rods and the grid conductor. Then qualify or disqualify the installation, based on the physical condition of installation.

 

[Skogsgurra]

Ground grid testing and verification is a challenge.

We have had cases where ground clamps have been used and where measured values were in the 500 - 1200 milliohms region. These measurements had been used as reference data for furthermeasurements. Then, there was a series of violent bearing failures in pumps and blowers and we were asked to have a look at what had happened.

There had been a ground fault in the plant just before the bearing failures and we suspected that the ground grid had defects. A resistance calculation showed that the resistance between grid meshes should be around 15 milliohms instead of the 500+ milliohms that had been measured for years.

The reason for the discrepancy is that ground clamps work with kHz instead of 50 or 60 Hz. At 2 kHz, the L/R in a ground grid is much too long and what the ground clamp shows is almost pure reactance instead of resistance.

We did it another way and found several corroded meshes. The method is described here:

http://www.gke.org/presentationer/files/Low ohms ground grid measurement.pdf

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[waross]

Great presentation Gunnar.
I may have missed this point,and please accept profuse apologies if I have:
I will value your comments on the following.
I accept that the DC resistance is a very good indication of the integrity of the grid connections.
However, fault currents are generally 50Hz. or 60Hz. The impedance at the plant frequency as well as the resistance determines the fault current in real world conditions.
When testing do you also determine the impedance at the working frequency as an indication of the ground connection's performance under real world conditions?
An impedance measurement may indicate not only deterioration but bad design (eg: too long of a single grounding conductor from the protected equipment to the main grid.) And poor workmanship (eg: rather than cut a cable to the proper length at the installation site, the cable is cut long and the excess disposed of by burying several coils in the ditch).
I am thinking of the perspective of a non technical loss prevention manager who hires us to check his grid. The grid resistance is fine and we sign off on it. Then there is a fault with damage and financial losses.
Our answer;
"The resistance is great but the impedance was too high. Sorry about the ground potential rise destroying all that expensive instrumentation."
If these points are suggested at a meeting to plan ground grid integrity testing at a large plant, how should we respond?
Yours
Bill


Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[Skogsgurra]

Bill,
The short answer to your question is that it is the impedance, not the resistance, at the selected frequency that is being shown in the box under the heading "IMPEDANS". The frequency can be selected to be anything from 20 - 30 Hz up to 20 Hz, but IRL it is difficult to reach more than around 7 kHz because the transformers don't work very well at higher frequencies.*

The frequency is 1.009 kHz in this measurement. Read from the lower green field.

The applied voltage is 175.3 mV (that doesn't create any explosions) and the resulting current is 8.08 A. Both are shown in (true) RMS. The crest factor is 1.414 for voltage and 1.440 for current. That is as close to sqrt(2), which is the ideal crest factor for a clean sinewave, as you can hope to get.

We use this set-up mostly to find defects in old ground grids and that is why the example shows 1 kHz. At 1 kHz, the ear is very sensitive and it is easy to trace the buried conductors with a simple search coil and headphones.

For actual 50 Hz impedance measurements, we set f at 50 Hz and the impedance at that frequency will usually then be a lot lower. Like around 15 milliohms.

The main reason why we did this write-up is that so many guys told us to use a ground clamp instead. Our results were also questioned. Their clamps showed several hundred milliohms - which they believed in. Our 15 - 20 milliohms were a lot closer to the expected (calculated) mesh impedance. It was only when we explained the physics and demonstrated the difference when measuring at different frequencies that we were able to do the measurements in, what we think is, the correct way.


*We are looking at Hi-Lite 100 micron plate cores and they will certainly take us a lot higher. But that is for very special measurements and we are satisfied with the 20 - 1000 Hz range that we use mostly.


Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[waross]

Thank you for taking the time to explain this, Gunnar.
Yours
Bill

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[Skogsgurra]

Apparently, quite a few people read these threads. One guy phoned me yesterday evening to discuss the matter at some depth. We came to some sort of conclusion, which can be split into several aspects.
There are legal, economical, "afraidness", technical, safety and probably a few more aspects.

Legal: Step and touch voltages. Tests shall be carried out regularly and there are well defined procedures for this. Plant protection is also one of the legal aspects.

Economical: It is not always possible to do all tests, for economical reasons. So more and more tests are skipped and, at the end of the day, there are just a few of them carried out and the rest is "forgotten". It is also quite time-consuming to open every down-conductor and inject test current. So, usually only a few of all meshes are tested. And always the same ones.

"Afraidness": Someone gets concerned that some tests *can* disturb operation of the plant. So "better safe than sorry" - which reduces the number of tests that are actually carried out even more.

Technical: Circumstances like a) Not possible to make a connection b) A cable is not long enough for certain test points c) Can't reach a test point because of heavy traffic d) There's gas present most of the time and "we would rather not..." e) No power outlet available where it is needed.

Safety: Includes most of the reasons above. And Safety is King - isn't it?

We do not test with high test currents. We do not need to and we would probably not get permission to do it either. We do not open any connections. We induce the test signals with transformers that can be opened like clamps and the tests therefore are quick and easy and we test all down-conductors.

Our few hundred millivolts induced test EMF can never ignite anyting. And as soon as impedance gets high, the current drops to safe levels. By using a settable frequency, we can measure resistance and inductance. And the primary value is always impedance.

We run off a battery, so we do not need to have outlets available near the test points. We induce and measure with current transformers and voltage transformers that open to include the tested bars and cables. We never open any connections, so no risk involved.

We do all those measurements that have been excluded for "afraidness", technical or safety reasons (I think that comfort/laziness could be included, too - but I didn't say that). The result is that we do find open connections, corroded meshes and sometimes ground grid wires that have been drilled through or ripped open by excavators. Defects that have gone undetected because of all the fear- safety- and other reasons mentioned above.

A pretty interesting result from a one hour phone conversation, I think.




Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[burntcoil]

Any comments or experience with this technology? Please see the link below

http://www.hoodpd.com/smart-ground.html