Network stability is traditionally measured by using synchrophasor data. Synchrophasors are typically produced at an exact rate of 50Hz. To satisfy the Nyquist criteria it is therefore only possible to detect oscillation frequencies up to 25Hz (50Hz/2). But for all practical purposes, you can only accurately quantify oscillations up to 12.5Hz (50Hz/4).
With a revolution underway in the renewable power generation, the need for new, more immediate insight into harmonic performance is now an essential for power grid effective power grid management.
Many IBR resources are small compared to the grid and you might find numerous sources in close proximity to each other – interacting with each other. Oscillation frequencies at higher than 12.5Hz is therefore not uncommon. In one test case recorded by VECTO at Wemen in Australia, the network oscillates at frequencies from 18Hz up to 21Hz.
Although its presence could be detected by using synchrophasor data, it is almost impossible to study both amplitude and phase performance. The relative phase performance of the oscillation recorded at different locations can be used to identify the contributing sources involved in the oscillation event.
Do you have a power monitoring problem that accurate, real-time data might solve?
CT LAB came up with the idea to record a 1-sec GPS synchronised high speed data frame. We then do an FFT on the data. The 1-sec window produces a frequency spectrum with 1Hz resolution. We only utilise the bottom 60 oscillation frequencies Spectrum. This 1-sec interval Synchronous Sub Harmonic Spectrum (DC-60Hz) can be stored to disk.
At the same time we also records a 10-sec window of GPS synchronised data. This time the FFT produces a frequency spectrum with 0.1Hz resolution. On this spectrum we allow the user to choose a spectrum of 60 frequencies with 0.1Hz resolution (6Hz). This 10-sec interval Synchronous Sub Harmonic Spectrum (DC-60Hz) can be stored to disk.
Each frequency component in both of these spectra can be analysed afterwards to understand both static and dynamic behaviour. It must be noted that both spectra produce data at a very high rate.
To combat this, we have derived two 10-min interval profiles from the 1Hz spectrum. The amplitude profile obtains the amplitude of the highest oscillation frequency within the 10-min interval. The frequency profile contains the frequency value of that highest oscillation component.
These trends produces very little data and can be used on each of the field installed devices to permanently monitor and detect the presence of sub harmonic oscillations and its frequency.
A monitor was built to identify Sub Synchronous Oscillation Events and to record diagnostic data with pre-and post information.