Identification of electromagnetic interferences based on adaptive sparsest time — Frequency analysis

Electromagnetic interferences measurement system is a multi-input single output system. The signal obtained by the measurement is a mixture of multiple interference signals. In the absence of prior knowledge of the device under test, it is difficult to provide enough support for interference diagnosis and subsequent rectification only using standard measurements in frequency-domain. This paper presents an improved method using adaptive sparse time-frequency analysis (ASTFA) method to achieve the time-domain separation of the electromagnetic interference (EMI) signal. First, the method uses ensemble empirical mode decomposition (EEMD) method to decompose the EMI signal into several signal components. Then the average frequency of each component is taken to estimate the initial phase of the ASTFA method to avoid the inaccurate decomposition results caused by the improper guess of initial phase. The proposed method only requires a time-domain measurement of the EMI. A simulation example of the EMI decomposition was given to verify the effectiveness and accuracy of the proposed method.