Study On DSP Based CZT Algorithms For On Line Harmonic Analyzing

Recent years, with the economy growth, the needs for electric power are increasing. While at the same time, the power system harmonic contamination created by non-linear loads became more serious. To maintain power quality and protect electric equipment, the issue of how to solve the contamination problem has been a common concern, which involves the studies on how to monitor and measure harmonics on-line and to develop new algorithms for fast and accurate harmonic spectrums analyzing. The paper was based on the "Electric Power Harmonic on-line Monitoring System" research project, and will focus on the improvements and applications of CZT algorithms for fast and accurate harmonic analyzing in the DSP based monitoring system.Firstly presented in the paper were the background information of power system harmonic monitoring and analyzing, and the advantages and disadvantages of the popular DFT and FFT algorithms for spectrum calculation. Then the FFT based algorithm of chirp-z transform (CZT) was introduced. The CZT is well suited for solving selected narrow band spectrum with adjustable high resolution, while the computing is time consuming. Thus three improved CZT algorithms for packed real signal, sectioned spectrum modulation and sliding window method were derived to reduce calculation time and speed up the spectrum analyzing for different applications. The algorithms and their programming on DSP platform were discussed in detail. The calculation time and computing speed for various cases were listed for comparison. The flow diagrams and the DSP codes of typical applications were also given.Those improved CZT algorithms were implemented in the DSP based harmonic on-line monitoring and analyzing system and worked well due to their high resolution and fast computing speed.