constant current loads in dSbus_dV about gridcal HOT 4 CLOSED

ok :) that is actually that made me a confusion and I was no more sure what was correct. At least we found a tiny bug :)
thanks a lot once more...your software is really great, I'll definitely continues using it and may come back to you with some questions.
best regards

from gridcal.

Hi, thanks for the compliments!

I have just double checked, and the implementation makes sense as it is now.

To clarify the signs:

• The current in a load in the interface is positive for load effect and negative for generation effect. This means that if I create a load of 10kA it should make the voltage drop, and if I enter a load of -10kA it should make the voltage raise.

• Now under the hood, both power (S) and current (I) are summed into the Sbus and Ibus arrays with negative sign, hence my 10kA load would be -10kA / (Vnominal * sqrt(3)) at the Ibus vector

• The Ibus vector (with presumably negative values for loads in it) is passed to the power flow algorithms.
  Now in the Jacobian, where that formula belongs: Ibus = Ybus * V - I . The Ibus from the Jacobian formula is expected to be positive, that is why I add the currents injection vector with negative sign.

I hope It is clear. Nevertheless, if you think there's an error, please let me know to fix it. My reasoning is that at the moment, the results make sense for the used signs.

from gridcal.

Hi, thank you for so detailed explanation!
ok, minus sign in Ibus = Ybus * V - I is now clear, but still I don't get why in case of non-sparse Ymatrix you have + sign here ?

from gridcal.

A bug probably, I'll fix it now.

That part is a remaining from the MatPower code. Ybus is always sparse in GridCal

from gridcal.