Preprocessing for characteristic mode tracking based on the correlation

Abstract

Mode swapping in characteristic mode analysis (CMA) affects consistent interpretation of the modal data. Mode tracking based on the eigenvector correlation can effectively determine the correct ordering of modes at each sampling frequency. However, there has been a lack of corresponding preprocessing for this mode-tracking algorithm. For example, the applicable frequency band cannot be determined in advance. In this paper, cosine similarity is used to calculate the correlation between modes of adjacent frequencies. It is then proposed that the 3/2 proportional relationship between the maximum size of the structure and the minimum operating wavelength can be used to choose the reliable operating band range of the algorithm. Meanwhile, it is recommended that the meshing side length should be less than one-tenth of a wavelength, and the adjacent sampling frequency interval should be less than the relative bandwidth of 10%. Through this preprocessing, correlation-based algorithms can efficiently implement mode tracking for open conductors.

A fast multi‐structural tracking method for characteristic modes with the ability to identify and amend errors

“…There are two approaches for adjusting frequency samples. In the first one, the frequency step size remains fixed during the process [11, 12]. In this approach, frequency samples are defined by $${f}_{\mathrm{i}}={f}_{\mathrm{s}\mathrm{t}\mathrm{a}\mathrm{r}\mathrm{t}}+(\mathrm{i}-1)\times \mathrm{S}\mathrm{T}\mathrm{P}$$.…”

“…There are two approaches for adjusting frequency samples. In the first one, the frequency step size remains fixed during the process [11,12]. In this approach, frequency samples are defined by…”

“…But working with σ mn is hard because it is a continuous parameter within (−∞; þ∞) range. Thus, the TMA assigns a 2-bit code according to (11) for a better exhibition of the status of n th mode with other modes.…”

“…Indeed, the correlation between a number of TCM has been considered as a decision criterion for mapping the same physical modes. The correlation-based techniques can be separated into three groups: (a) the eigenvector correlation-based method [11][12][13][14], (b) the far-field pattern correlation-based method [15] and (c) the surface current correlation-based method [16].…”

“…Although comprehensive efforts have been devoted to introducing an authentic tracking procedure, simultaneous achievement of high accuracy, low computational cost and low sensitivity on frequency step size and mesh density is still difficult. In addition, dealing with undesired crossing and crossing avoidance is still a big challenge in many MTAs [6,[11][12][13][14]. However, there are some fast admirable algorithms based on theoretical calculations that have overcome many problems in modal tracking [9,21].…”