Software Patches

June 30, 2015

Features and Improvements

  • Transient Stability: Modification were made when transferring results from transient stability for viewing inside Simulator.
    1. The situation where one end of a transmission line is open is now handled by putting in a fictitious bus shunt so that mismatches which are viewed are meaningful.
    2. We now show islands that exist in stability which would not be permitted to exist in the power flow, such as an island without any load or an island with only 1 bus.
  • Bug Fixes

    • Auxiliary Files (SCRIPT and DATA Sections): Fixed error in script command OpenWithBreakers() which was causing an access violation.
    • Fault Analysis: When applying a single-line to ground fault to a portion of the system that has no zero sequence path to ground, Simulator was not correctly calculating the zero sequence Thevenin impedance (which should have been infinite!). This situation is highly unusual in a real system (as grounding is quite important), but the software will now appropriately return a fault current of zero.
    • Transient Stability: Changed the DC line algebraic equation solution in transient stability. Previously when using algebraic models of the DC network (no model, EPCDC, CDC6, etc...) during the network boundary equation solution, after each Newton iteration, the firing angles would be recalculated to attempt to exactly set the desired Current (or Power) of the DC converters. For algebraic solutions near points where firing angle limits were either hit or backed off, this could make the network boundary solution oscillate and fail. Attempting to get the current/power setting exact is unnecessary though. Instead, we now only modify the firing angle at the beginning of each time step in the simulation and then leave the firing angle constant. This is the same process that is done when actually modeling the dynamics of the DC network and using a transient stability model for the DC converters. Essentially the new algebraic solution is assuming that the firing angle is calculated based on the last time step's AC voltage profile and thus essentially the firing angle control has a 1 time-step delay. This is perfectly appropriate and makes the network boundary solution much more robust and the current/power control is still met exactly withing a couple time steps.