Study On Fast And Accurate Spectrum Analyzing Algorithm For Power System Harmonic Meter

The increasing harmonic contamination in recent years has been a severe problem to the quality and safe operation of the power system. Thus the studies on how to implement on line harmonic monitoring became more attractive topic. This paper introduces the architecture and working principle of the 80C196 microprocessor based harmonic meter, model SL-11M1. The most important technique aspect that the meter involved is how to get the spectrum analyzing fast and accurate enough with meter's very limited system resources and so to make the real time harmonic monitoring possible. For that reason, the commonly used FFT algorithm was first derived, the theory and the programming approach of the FFT based chirp z algorithm was then discussed in detail. The principle and methods to reduce the spectrum leakage by synchronizing the data sampling with the signal frequency was too given. Based on the traditional frequency measuring and tracing technique, two software algorithm that will help getting even better spectrum accuracy were presented. One is to have the non-synchronized sampled data synchronized by using an adaptive data length for spectrum analyzing, another is to use the linear least square estimation for frequency calibration.The system implementation, such as the programming environment, the flow diagram for the major algorithm and the system testing and emulation techniques are briefly stated.