How can I model the arc voltage characteristic of a circuit breaker?
In EMT simulations the circuit-breaker is modelled as an ideal switching i.e. interrupting at current zero and exercising no influence on the flow of the current. The arc voltage of a practical circuit-breaker -representing the nonlinear resistance- can reduce the DC time constant of the short-circuit current and hence force current zeros and reduce the fault clearing time.
How can this effect be taken into account in PowerFactory?
The arc voltage of the circuit-breaker can be modelled as a nonlinear resistance. Therefore we connect a resistor (series impedance)in series with the circuit-breaker. Both the real and imaginary part of the series impedance, and hence the resistance, can be controlled now externally from a DSL model.
The DSL model should then account for the following functions:
- Control the nonlinear characteristic of the series resistor according to the arc voltage characteristic. It means then, the output of the model shall be the resistance. Input signal is the current through the resistor, i.e. through the circuit breaker. As for the initial condition we can assume that the circuit breaker is closed, we initial arc voltage will be zero (contacts closed) and therefore the resistor can be initialized to zero. Alternatively it can be also initialized to the resistance of the contacts of the circuit breaker (some mOhms, when the contacts are closed).
- Based on a trigger (eg. boolean variable iopen=0/1), the model starts controlling the resistor according to the equation of the arc voltage characteristic. Hence it starts influencing the flow of the current. The trigger is set by a parameter event in the simulation.
- At the same time, once the model has been triggered, it automatically generates a switch event on the associated circuit-breaker, which will finally interrupt the current by the next current zero-crossing.
Finally, in the list of simulation events we do not define a switch event to open the circuit breaker directly (as we would normally do) but we create a parameter event for the corresponding variable (trigger) in the model. The model does the rest.
For further reading refer to: "Delayed current zeros due to out-of-phase synchronizing", Koeppl Canay, Braun. IEEE Transactions on Energy Conversion, PE-596-EC-0-05(1), May 1997.