Ungrounded systems protected for max safety 1

[Rodmcm]

Assuming we have a power transformer with only 3 output connections, either delta or unearthed star at say 1MVA 400-600V say. What would be the most comprehensive way to ultimately protect the output for safety reasons for
a) a single phase to earth connection , even though it has no return path, but is deemed a no no
b) phase to phase
c) three phase

 

[davidbeach]

I'm not sure what you are asking. Why would you want an ungrounded system anyway?

But assuming an ungrounded system (there are always better, safer, methods), you can protect against single phase to ground faults by looking at the voltage to ground of all three phase conductors; there are several different approaches here from looking at one phase (undervoltage and it is the faulted phase, overvoltage and another phase is faulted) to looking at all three phases using a three-phase relay. A single overvoltage relay looking at the output of a set of broken delta VTs (not open delta) can detect the presence of a ground fault but not identify the phase.

A bolted phase-to-phase fault will be picked up by the phase elements, no problem. A high impedance phase-ground-phase fault can easily be detected by looking at negative sequence overcurrent.

A three-phase fault doesn't care how the system is grounded, the phase elements will pick it up.

 

[Rodmcm]

Sorry Davidbeach but i don't understand. How do you look at one phase 'undervoltage and it is the faulted phase', overvoltage and a phase is faulted'? What are broken delta VTS?

 

[davidbeach]

Voltage on one phase to ground: with no faults your voltage to ground should be 1/sqrt(3) * phase-phase voltage. If the phase being measured has a ground fault that voltage will go down to near zero (actual voltage depends on where the fault is and the voltage drop between the fault and the measuring point). If either of the other phases has a ground fault, the voltage to ground on the phase you are measuring will approach the phase to phase voltage (again the exact value will depend on the nature and location of the fault). Thus you can detect the presence of a ground fault with one VT (and a stabilizing resistor) together with a 27/59N relay.

Broken delta VTs: Three VTs, primary connected grounded-wye, secondary connected in delta. Now, open (break) the connection at one corner of the delta and insert a stabilizing resistor and measure the voltage across that resistor with a 59N relay. The voltage is the sum of the three phase voltages, or 3*V0. The presence of 3V0 greater than the set point of the relay is an indication of the presence of a ground fault.

Better yet, use three VTs connected grounded-wye / grounded-wye into a three phase 27/59 relay (solid state or numeric) and actually know which phase has the fault.

None of these techniques will tell you where your fault is though.

A high-impedance grounded system is far safer, and many are provided with a means of alternating the ground fault current between two values, such as 5 and 10 amps to make tracing the ground fault much easier.

Ungrounded is just bad design.

 

[Rodmcm]

I would agree for a distribution network. However, a HV delta with two unearthed secondaries ( delta and star) are very common on large variable speed drives.
The concern is of course that the tx's can sometimes be mounted well away from the drives.

 

[Skogsgurra]

Bender (

http://www.bender-de.com/wocms.php?siteID=118)

has a complete range of monitoring relays specifically made for this kind of situations. They force a low current alternating DC (yes, I would also call it AC) from system to ground and measure resulting DC voltage and current. As long as the ratio (U/I=R) it is high enough, you get no alarm. It is only when the resistance drops below the set value that you get an alarm.

Have been using them on 500 and 690 V ungrounded systems and they are very reliable. They even compensate for large capacitive currents that often exist in such systems.

Yes David, ungrounded is mostly bad design. But, it has its proponents - and sometimes you even need it. Like in twelve-pulse rectifiers.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[davidbeach]

OK, I'll buy the 12-pulse (or higher order) converters (inverters as well as rectifiers); but I also mentally draw a box around the transformer and converter and label it as a single entity and the ungrounded circuit does not leave that entity. Writing of transformer alone didn't suggest a converter and so I was thinking transformer and distribution circuits of some nature.

 

[Skogsgurra]

I agree fully. Ungrounded is bad. Many paper mills and steel mills that I have worked with started using ungrounded back in the sixties. The rationale was that "We can still produce if we have one ground fault" That didn't work very well because there never seemed to be any time to find that first ground fault when it occured. So most plants were running with a system with a weakly grounded phase and that caused lots of interesting phenomena. A few of the steel works I visit today have started grounding their neutrals. But some transformers have delta secondaries so they can't do that - even if they wanted to. Grounded corner is not considered good practice here.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[davidbeach]

Grounding transformers are always one possibility if you have a delta winding. Lots of Yyd autotransformers used in the bulk power transmission system and if the delta tertiary is used for something, often station services, there is often a grounding transformer installed to provide ground reference. The grounding transformer could be a zig-zag transformer or it could be a grounded-wye/delta transformer with nothing connected to the delta side. The grounding transformer will often have a resistance in the neutral-ground connection to create an impedance grounded system and limit the fault current (I'm sure you know this already Gunnar, but the rest of the audience might not).

The impedance grounded system can have the resistance sized such that the maximum fault current is low enough for continued operation while the first fault is tracked down and corrected while eliminating the devastating overvoltages that can be associated with faults on ungrounded systems.

The transformer/converter pairs are left ungrounded because grounding the transformers would create a fault. Systems left ungrounded so that they can continue to operate during a ground fault are what should be impedance grounded rather than ungrounded.