Numerical aspects of determination of natural frequencies of a power transmission line cable equipped with in-line fittings

• Published in: Engineering structures Vol. 160; pp. 510 - 518
• Main Authors: Wolf, Hinko, Singer, Sanja, Pustaić, Dragan, Alujević, Neven
• Format: Journal Article
• Language: English
• Published: Kidlington Elsevier Ltd 01-04-2018; Elsevier BV
• Subjects: Aeolian vibrations; Aircraft vibration; Dampers; Design engineering; Eigenvalues; Eigenvectors; Electric power; Electric power lines; Electric power transmission; Energy balance; Energy transmission; Fittings; Power lines; Resonant frequencies; Singular value decomposition; Singularity detection; SVD; Transmission line; Transmission lines; Vibration; Vibration analysis; Vibration isolators; SVD; Singularity detection; Aeolian vibrations; Transmission line
• DOI: 10.1016/j.engstruct.2018.01.043
• EISSN: 1873-7323
• ISSN: 0141-0296

•Aeolian vibration response of electric power transmission lines is analysed.•The use of the Modified Energy Balance Method (MEBM) is considered for the analysis.•Numerical problems associated with the implementation of the MEBM are identified.•A new efficient numerical procedure is developed in order to avoid these problems.•Examples are presented which illustrate advantages of the proposed procedure.
In the analysis of Aeolian vibration response of electric power transmission lines, the modified energy balance method (MEBM) is often used. The first and crucial step in applying the MEBM is an accurate determination of the natural frequencies and modes of the system, or in other words, the eigenvalues and eigenvectors of the system matrix. In this paper an efficient numerical procedure that searches for frequency parameters s that make the system matrix J(s) of transmission line cable singular is considered. A realistic case where the cable is equipped with in-line fittings such as Stockbridge vibration dampers and aircraft warning spheres is taken into account. It is shown that the rank of the considered system matrix J(s) is either full, or the full rank minus one. From a numerical point of view, this is an important property of the system matrix since multiple eigenvalues without a full set of eigenvectors, i.e., with Jordan blocks of order greater than 1, are very sensitive to small perturbations. The developed numerical procedure, which is easily parallelizable, consists of the hybrid minimization method paired with the Singular Value Decomposition (SVD) for the detection of the singularity of the matrix. Presented numerical examples illustrate advantages of the numerical procedure proposed.