| Frequency is a fundamental parameter in power system for control, protection and monitoring applications. Three DSP solutions for frequency measurement are proposed in this thesis. The first one is a zero-crossing method that is unproved for metering accuracy by incorporating a simplified cubic interpolation algorithm. The frequency tracking speed is also improved by a supplementary three-point method. The second approach uses a Gauss-Newton method to estimate the frequency as a model parameter. Each Gauss-Newton updating step is supervised by a recursive digital Fourier transform algorithm and a zero-crossing algorithm so that fast convergence can be achieved. In the third approach, a signal demodulation method is used to compute the frequency deviation through a process of signal modulation and demodulation. A Chebyshev type II low-pass filter for signal demodulation is designed through comparison tests. The three solutions are verified by extensive simulation studies. |
