Utilization of Frequency-Domain Information of Barkhausen Noise Signal in Quantitative Prediction of Material Properties

This paper describes different approaches for utilizing frequency-domain information of Barkhausen noise (BN) signal in quantitative prediction of material properties. Different approaches include the calculation of power spectral density (PSD), moving window PSD for obtaining the BN profile and the utilization of the BN spectrum. The PSD value is calculated directly from the BN signal and is related to overall Barkhausen activity. The application of moving PSD filtering produces the so called BN profile. Peak height, position and width of the profile can then be calculated and used. Spectrum gives the signal power as the function of frequency. This information can be related to material properties and thus utilized in predictions. This paper mainly discusses the different approaches but also presents some results which show that frequency-domain information can be important. It is common in the literature that only time-domain properties of the BN signal are used. However, prediction of material properties is case-dependent and thus it may be beneficial to use also frequency-domain information as shown in this paper.