To achieve carbon neutrality, the integration of renewable energy sources is expected to expand significantly. However, there are concerns that this expansion may deteriorate the supply-demand balance of power grids and lead to a decline in grid frequency stability. To address this issue, adding fast frequency response (FFR) control to grid-following (GFL) inverters has attracted much attention recently. Since GFL inverters are simple in configuration and low-cost, FFR-type GFL inverters serve as a readily deployable measure for grid frequency stabilization.
FFR achieves a faster response than the conventional governor control of synchronous generators by controlling power output of battery energy storage systems based upon measured frequency values. While FFR performance is directly linked to measurement accuracy, the existing standards are limited to accuracy requirements under steady-state conditions. Consequently, the dynamic input-output characteristics of frequency meters remain unclear. These include the tracking capability and errors of the output signal in response to sudden changes in the input signal, such as amplitude, phase, and frequency variations, and evaluation indices for these characteristics have not been established.
This paper evaluates the dynamic input-output characteristics of frequency meters based on experimental results. To evaluate these characteristics on a Bode plot, the three-phase voltage waveform, which serves as the input signal to the frequency meter, was slightly amplitude-modulated by another low-frequency sinusoidal wave. Specifically, the input signal v(t)=(V0+X*sin(2pi*delta_f*t))*sin(2pi*f0*t) was applied, and the output signal, grid frequency f(t)=f0+Y*sin(2pi*delta_f*t+theta), was measured. From these measurements, the gain Y/X and phase shift theta of the specified low-frequency sinusoidal wave superimposed on the nominal frequency were determined.
In the experiments, a Typhoon HIL simulator implementing a grid disturbance model and the power meter were used. Three-phase voltages superimposed with amplitude-modulated signals (amplitude: 1–10%, frequency: 1–10 Hz) generated by the Typhoon HIL were amplified to 100 Vrms using a bipolar power supply and fed into the frequency meter. The measured data and RMS values output from the meter synchronously were acquired via Ethernet communication using TwinCAT control software.
The experimental results showed Bode plots representing the gain and phase characteristics of the meter input and output, using the modulation frequency range of 1 to 10 Hz as a parameter. Since the gain and phase shift between the input and output are not constant relative to the modulation frequency, these results clearly suggest a non-trivial transfer function within the frequency meter. To improve FFR control performance of GFL inverters, the design of the FFR controller should include these dynamic characteristics of the frequency meter.
The full paper will describe the results of identifying the transfer function between frequency meter input and output. Furthermore, by incorporating the obtained transfer function into the GFL inverter model, the stability of the FFR control system is evaluated under different short-circuit ratio conditions using Nyquist plots and pole maps.