Revision history [back]

For the case when load is 100% and SR (spinning reserve) is, say 10%: [task A] Do a load flow, prepare case for dynamics and run a non-disturbance test to check that the system is stable. Choose a unit to trip and plot the frequency response of the system. Now repeat the gen trip run with less amount of SR by taking out of the dyr file the governor model of another unit to simulate lower SR % conditions. This way, power balance is retained while changing the SR amount. The unit to trip (choose always the same unit, with governor disabled) can be scaled down (with other unit being scale-up) to simulate system frequency response at different amount of MW loss, then do task A before tripping the unit. Now for the case when load is 95% (100-5) and SR is 15% (10 + 5): repeat all the test done previously. Chances are your system does not need task A anymore, because all the models will be initialized for a 'unloading' network condition. These studies require a lot of units to form the sub-group of units providing SR, another sub-group of units to be used during the scaling up, with governor disable, etc. I recommend you use at least the IEEE118 or larger network.