A novel DFT-based method for spectral analysis under time-varying frequency conditions

• We propose a simple technique for reducing the errors induced by asynchronous sampling in spectral analysis of electrical power signals.
• It is presented a comprehensive analysis of leakage effect in DFT (Discrete Fourier Transform) result.
• The method is quite simple and is based on mean values of Fourier coefficients, provided by moving average filters.
• The method is suitable for real-time tracking harmonics of time-varying frequencies as well as inter-harmonic components.

This paper presents a novel DFT-based method for spectral analysis in power systems. The main objective is to provide accurate estimates for harmonics/inter-harmonics containing time-varying frequencies. The method addresses the problem of DFT errors when the sampling process is asynchronous, i.e. the sampling frequency is not an integer multiple of the fundamental frequency. The oscillations of the phase angle in the DFT result are attenuated using simple filters to provide frequency estimations for each component of the input signal. The corrections of the input signal parameters are made by using the DFT frequency response equations without the need of adjusting the sampling rate. The individual nature of the correction process enables interharmonic components to be tracked. Simulation results show the effectiveness of the method despite its simplicity, if compared with other methodologies.