Answers are grounded in this paper and cite their sources.
Aiming at the problem that hidden water accumulation inside multi-segment steel tube lightning rods easily induces internal corrosion and degrades lightning protection performance, a dual-excitation frequency-domain spectrum transformer (DEFST) is proposed for joint identification of the axial distribution and filling level of accumulated water. Based on the modulation mechanism of local trapped water on the structural dynamic characteristics of hollow steel tubes, full-band amplitude spectra collected from two axial excitation positions are adopted as direct inputs to build an end-to-end identification framework. A frequency patch embedding strategy is designed to convert continuous spectral curves into sequence tokens, and a dual-branch self-attention encoder captures the coupled full-band spectral variations caused by water accumulation. Furthermore, a cross-excitation attention fusion module is constructed to mine the spatial sensitivity information embedded in paired spectra, realizing synchronous axial distribution and filling level quantification of accumulated water. Experimental verification is carried out on a 1:4 scaled steel tube lightning rod test rig covering nine combined water distribution conditions. The results indicate that the proposed method achieves an overall identification accuracy of 92.59% for the nine working conditions. Ablation experiments verify the positive contribution of each core module. The proposed method requires no structural modification to lightning rods, and has potential application value for refined non-destructive inspection of substation lightning rods after further field verification.
No takes yet. Share an insight, caveat, or question.
Wan et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: