Protection schemes: the good, the bad and the ugly
CEE presents some case studies of when power system protection won’t work as intended, or looks like it should not work but does. There will always be instances where some compromise is necessary; we hope that these examples will help you to find the best solution for your application.
Case 1: Attack of the Generators
SLD grid and generator
An 11kV generator has been added to a network with a grid connection. The generator cannot support the whole network on its own, but it is designed to run in parallel with the grid power supply. This usually happens when renewable energy schemes are implemented. The generator’s protection relays include all the necessary directional, out-of-step and overload protection but the overcurrent relay protecting the cable from the generator to the 11kV bus seems to be set a bit low (see TCC, right)
This TCC doesn’t look right. It seems to be telling me that the generator overcurrent relay (Incomer02) will trip faster than the feeder overcurrent protection (Feeder01) for any fault current under 700A. But wait – when that happens the generator won’t be supplying all the fault current on its own! I can make use of PTW’s Branch Fault Current feature to adjust the TCC based on the relative contributions of the generator and the grid (see TCC, left):

Now it’s obvious that the relay protecting Feeder01 will trip first because the generator contributes much less fault current than the grid. I should be careful though – when the grid connection is offline the Feeder01 relay won’t trip as quickly because the fault current is only coming from the generator! I should make use of the NP900’s multiple setting groups so that the Feeder01 relay is more sensitive when the grid is out of service.
Case 2: The Clustered Cable Caper
A simple-looking protection scheme with a dark secret.

The motor draws so much current that it needs three cables per phase just to meet the demand. I use CAPTOR to check the TCC; the cable damage curve is above the motor starting current but the circuit breaker is set above it!

If there’s a fault the cable would be damaged before the protection isolates the problem – the cable current capacity looks low too.
Of course, the whole reason I put in so many cables was to carry enough current for the motor; the data on the TCC is for only one cable per phase so I need to take account of this somehow.
I can always rely on PTW and this time is no exception. I set the cable plot to “parallel” and now I can see that with all the cables working together the normal running current will be fine (TCC diagram, left). The damage curve is to the right of the circuit breaker curve so the cables will be protected during a fault too.

There’s a snag though; if just one of those cables is accidentally disconnected we’re back to square one. Maybe I need the NP900’s “cable end protection” feature to raise the alarm if one of the cable terminations is loose or disconnected at the source?
Case 3: Troublesome Transformer Trip
star-delta and delta-star transformer
A 33/11kV transformer with a delta-wound secondary has been used to power a second transformer with a delta-wound primary winding. There’s no neutral conductor at 11kV and no earth; just the three phases. There’s just one hitch – transformer 2 won’t power up. Every test possible has been run on it and it passed with flying colours, but the up-stream protection trips every time power is applied.

Because there’s no earth conductor, the 11kV protection relay has been set with a really sensitive pick-up on the CBCT. I’m sure that there’s no earth fault but there must be a current imbalance coming from somewhere. I remember reading back in the April newsletter that third-order harmonics can cause current imbalance and then it hits me – like all transformers, Transformer 2 will have high third-order harmonics in its magnetisation current! 

This time around I had to calculate the charging current using the cable and transformer capacitance I = 2√3 Vπf (CTx1+CTx2+Ccable) and set the CBCT relay above this level. If I’d installed an NP900 relay I could just have set the unbalance protection to ignore third harmonic currents; I could even have used the disturbance recorder to see what the problem was in the first place! 
Case 4: The Mystery of the Missing Current
So just when I think that two delta windings connected together is bad enough, there’s a case where two transformers with star-wound secondaries work in parallel to feed a switchboard. The switchboard had a phase-to-earth fault but only one of the core-balance relays on the incomers picked up and I can’t figure out why.
Whoever designed the system only earthed the neutral on one of the transformers (the one on the right, in this case). Maybe they were afraid of circulating currents; we’ll never know. A PTW model seems like my best shot at cracking the case and it confirms exactly what was measured: Neutral current flows in only one of the incomers. Things get stranger when I look at the primary windings; both transformers have identical input currents, meaning that they must both contribute equally to the fault! 

If in doubt, draw it out. The PTW result clearly shows that the currents in each phase of the transformer secondary windings are either in phase or 180o out of phase (effectively, they travel in opposite directions). This makes it really easy to see which currents add together and which ones cancel out. Suddenly it all makes sense; the earthed neutral of the right-hand transformer becomes the return path for the current from the phase-to-earth fault. The currents in the left-hand transformer sum to zero at the neutral point, so no current flows in the CBCT.  
Fault currents
Maybe I need to set up an inter-trip to make sure both incomers are tripped by an earth fault detected on the right-hand side?