A Method for Improving the Accuracy of Natural Frequency Measurement Using In-the-loop Computing

Abstract

The method presented in the paper is based on in-the-loop computing applied for impulse response to obtain a spectrum with a much higher frequency resolution than using FFT. Then, higher spectrum frequency resolution results in greater accuracy in estimation of natural frequencies. The frequency resolution of estimated spectrum in this method is completely independent of the length of impulse response and, by extension, the method eliminates the problem of spectral resolution limitation using FFT due to finite length of recorded signals. This fact is very useful and is the main advantage of the proposed method. The results of the method have been shown and compared in quantitative terms to natural frequencies estimated using classical FFT with zero-padding as reference method.