Detecting DC Leakage in Grounded Systems

If the device's function were to be alarm-only, then it is only going to display what is already known, i.e. that there are earth leaks on the system.

If the device's function were to be alarm and disconnect, then the system will display a marked reluctance to operate.

Wouldn't curing the problem(s) be a more worthwhile investment, perhaps?

As I said earlier that the DC system in our plant has become vast distributed, there are around 50 DC breakers each of which caters to a number of feeders. From some of the feeders the DC control wires are running for almost 500 meters. We have tried on a few occassions to switch off the DC Breakers one by one to trace out the source of positive grounding but we have not got any fruitful results. In order to eliminate mal-operations of breaker/relay DC cicuits the next idea is to monitor the level of leakage in the negative side in each circuit. In the idle condition of a circuit there should not be any current flowing from the Minus 250V negative side to the 0 positive side. If there were any intermittent path then there would be some amount of leakage current flowing. Low levels of leakage would be insufficient to operate relay coils, but at higher levels the relay would operate thereby tripping the entire circuit.

In the recent past we had the station bus bar differential lockout relay mal operate twice and now it has been kept out of service.

Any suggestions????

<...on a few occassions to switch off the DC Breakers one by one to trace out the source of positive grounding...Any suggestions...>

Yes. One. Try harder. Complete it before someone gets hurt/killed.

<...the station bus bar differential lockout relay mal operate twice and now it has been kept out of service...>

So an installed safety protection system operates twice to indicate that something is amiss, and the action is to isolate it and carry on regardless? What kind of sanity is that (rhetorical question - NNTR)?

On further reflection, it seems the installation has one of these already. It's just that, having operated twice, someone has isolated it.

One might predict a future where additional equipment is added to the installation, which operates twice to detect and disconnect a fault, and some miscreant isolating that as well.

Where would it all end?

One of the benefits of an ungrounded distribution system, whether on a DC or AC system, is that a failure of one of the ungrounded conductors to ground will not cause an interruption in the system. That is why they are often times specified for "high up time" systems such as sewage pumping systems and other systems where downtime has to be avoided.

But the corresponding danger is that now, your bonding system is at an elevated potential relative to the rest of the distribution system.

Having said that... one of the core responsibilities involved in using an ungrounded system is to trace, find and repair the wiring system when ground faults do happen.

If one conductor has failed and introduced a ground fault, it is almost a certainty that the other ungrounded conductor in the faulty portion of the distribution will eventually also fault to ground.

When that happens, the full fault current available to the system will be dumped into the fault. A failure to trace, find and repair your system is exposing your system to a potential much greater failure.

I hope your overcurrent devices are properly coordinated. Take the advise given by PW Slack, and fix your system, then worry about monitoring gear.

"I hope your overcurrent devices are properly coordinated" . Overcurrent devices in DC distribution system??? We are using a 250V DC system for control of switchgear circuits. There are no overcurrent devices in the DC system.

"So an installed safety protection system operates twice to indicate that something is amiss, and the action is to isolate it and carry on regardless? What kind of sanity is that". When around 750MVA of power flowing through the station bus bar trips all of a sudden due to no real fault on the AC side but due to mal-operation of the relay just because of grounded DC supply, the easiest option is to switch off the differential relay till the time the DC grounding is detected and isolated.

You said in an earlier post... "around 50 DC breakers each". Now you are saying there are no protective devices on a 250 volt distributed DC system? Is that correct?

If so... ground faults are the least of your problems. Having ground faults on a distribution system like this WITH NO OVERCURRENT DEVICES is a pretty good pre-cursor to a rather large fire and excellent weenie roast.

As to the second parragraph you quoted in your reply, that statement was not from me but from another member, as such, I will not comment on it.

<...There are no overcurrent devices in the DC system...the easiest option is to switch off the differential relay till the time the DC grounding is detected and isolated...>

Run for the hills! Flee!

My systems are monitored through what we refer to as the Current Protection Panel. Once a day we push a button and check the leakage current between station ground and the DC ground for each of our power supply systems. These systems range up to near 8000 -VDC and we run about 2.5 mA - 3 mA leakage on each.

The only regular point of exposure for the separate grounding systems is protected through interlocks to ensure safety of the maintenance personnel.

Our panel was designed specifically for our systems, but if you have a single point of contact for each of the DC grounds you could install a simple monitor there. It sounds, though, as if it is too late for your plant.

You could run a full time system to monitor the leakage, but that would be overkill in my opinion. A simple meter that can be checked on daily rounds should suffice.

I could not resist asking/may not be correct to ask...

Would a meggar test reveal the problematic area?

Gajanan Phadte

@olehwi - Thanks for your answer!! You have provided a solid lead for me to work on. Thank you. Highly appreciated. And please do me favour by providing a copy of the CSA Report 142T259. I will be obliged.

email address removed. - cr4 admin

Thank you . . . this was remarkable . . . I always knew there was another name for "admin".

This is one of those times when, rather than persisting with study, urgent action is required as previously indicated.

Umm...yeah...it is....

I just did it. If you want to come on over I'll do it again...

I can do it all day long. In your own post you state "...or grounded with a resistor"...of course, using a shunted meter is essentially placing a resistor in the circuit?

So, do you mean really UNgrounded or just grounded in a different manner?

I mean grounded...I don't use a resistive ground, unless you count all of that electronickey stuff in the path between the power supply and the ground field...just sayin'

nice troubleshooting method !

Thanks

If you can turn off all your loads, there is a simple way to locate the general area of your fault. I'm assuming that your distribution gear has exposed accessible bus bars. Shut off all your loads, then using a scope or sensitive meter, place your positive and negative probes on the same bus bar as far apart as possible. You should get a polarity indication on the meter, even if the voltage is too small to actually read, if there is any current on that bar. This will tell you which circuit has the fault. Get the probes placed in polarity and you can tell which direction the current is going as well if needed. I have a device that has a highly sensitive amplified probe set with both probes in one tip that i have had for years. Unfortunately it is no longer being made and the maker has gone out of business. It gives you very accurate current and direction info by analyzing the voltage on the probe and the resistance between the tips. Good luck! I hope this helps.

"If you can turn off all your loads" . I cannot turn off all my loads. I have extremely critical loads like Generator DC Seal system along with the the entire control and protection of a large number of equipments.

Has anyone ever used one of the Bender, Germany products?? They claim that theirinsulation monitor and fault location system is very accurate and reliable solution in presence of today's non-linear loads and industrial noises as it conform to IEC standards and qualified for EMC????? Anyone???

If these are critical loads then you should have back up systems. Use them! Shut off the primary circuits and isolate them. Then use the tricks i suggested. Another idea. Use a DC clamp on ammeter. Measure the current at the load end of your cable, then measure it at the disconnecting means or OCPD. When you find a significant difference you've found your fault. Measure the voltage to ground at both ends in series with a 100k resister. Now you can use ohms law and thevinins theorem to calculate the approximate position of your fault along the cable. Assuming of course that you have the manufacturers specs of the resistance per foot of your cable. You should be able to get within a few feet of where the fault is!