Pakistan System Blackout

System demand before collapse was 11,683 MW.
Took about 40s from when the oscillations were detected to system collapse.

Economic Merit Order dictates to run cheaper nuclear and coal-fired generation in the South and send 4100 MW towards the demand in the North via a combination of HVDC and HVAC tie-lines.

Two important 500 kV circuits switched off overnight for voltage control and due to foggy weather (probably to reduce corona losses!?), the circuits have not been returned to service before the morning demand pick-up.

500 MW of wind in the South curtailed overnight, due to downward regulation issues on the conventional plant, was released uncontrolled increasing the South to North transfer to above 4600 MW which caused overloads on the 500 kV corridors between North and South and depressed the voltage profile. Lifting the wind curtailment caused the frequency to reach 50.7 Hz which prompted operator actions. However, this actions were taken on synchronous machines in the North.

Soon after, oscillations have developed the parallel HVDC tie-line started to experience commutation failures, temporarily dumping the 2400 MW transfer on to HVAC path. Many 500 kV circuits tripped in quick succession due to power swings (distance protection Z1) at one end only with the remote end tripping on overvoltage (>1.3pu at the end connected, not known what the voltage magnitude was at the open end), quickly followed by all generators disconnecting on frequency protection functions.

The restoration took about 20 hours, due to unavailability of capable plant in the Central Region of the country and lack of blackstart testing and lack of SCADA and capability to remotely operate switchgear. Islands collapsed several times during restoration, one particular power station collapsed the island 9 times due to large frequency swings, the investigation suggests that the units might not have been operating in frequency control mode.

The reasons for the system collapse are related to lack of situational awareness and operators' errors. No SCADA EMS with contingency analysis to highlight the system was operating in a topology conducive for development of instability phenomena. Counter-intuitive actions taken by the operators, first to release fast ramping wind output on the already highly loaded HVAC tie-lines between South and North then to try and deal with the resulting overfrequency by reducing output and shutting down synchronous generation in the North which exacerbated the overload on the HVAC corridor and also depleted the system of reactive power resources and means of fast voltage control.