Scholarlink

Purpose <p>Conventional techniques often yield substantial errors in pole identification, particularly when dealing with closely spaced modes. Furthermore, the presence of non-periodic frequency components in short-duration earthquake records increases spectral dispersion, particularly in higher modes. Additionally, violations of FFT assumptions for discrete, non-periodic signals result in spectral leakage across all structural modes.</p> Methods <p>To address these shortcomings, a novel method termed Power Spectral Density Transmissibility–Complex Frequency (PSDT-CF) is introduced. The method utilizes forward and backward frequency shifts via the z-transform to generate modified transmissibility functions. This procedure suppresses spectral leakage in the vicinity of the poles and enhances modal discrimination in systems with closely spaced modes. The proposed approach is validated through the dynamic analysis of the Koyna Dam subjected to three distinct excitation records.</p> Results <p>Based on noise-contaminated structural response data, the results indicate that for frequency shifts exceeding 120, the maximum error in estimating the natural frequency of the fourth mode is limited to 2%, while the corresponding maximum error in the estimation of the damping ratio for the same mode reaches 5%.</p> Conclusion <p>In comparison with conventional techniques—which typically exhibit reduced accuracy in higher modes—the proposed PSDT-CF method consistently achieves natural frequency estimation errors below 2.5% across all identified modes, demonstrating superior robustness and reliability in modal parameter identification.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A Novel Complex Frequency-Domain Approach Using Power Spectral Density Transmissibility to Extract Dynamic Characteristics

  • 期刊论文

  • 发表日期：

    2026年7月6日

查看全文
  • Reza Tarinejad,
  • Farhad Amanzad

摘要

Purpose

Conventional techniques often yield substantial errors in pole identification, particularly when dealing with closely spaced modes. Furthermore, the presence of non-periodic frequency components in short-duration earthquake records increases spectral dispersion, particularly in higher modes. Additionally, violations of FFT assumptions for discrete, non-periodic signals result in spectral leakage across all structural modes.

Methods

To address these shortcomings, a novel method termed Power Spectral Density Transmissibility–Complex Frequency (PSDT-CF) is introduced. The method utilizes forward and backward frequency shifts via the z-transform to generate modified transmissibility functions. This procedure suppresses spectral leakage in the vicinity of the poles and enhances modal discrimination in systems with closely spaced modes. The proposed approach is validated through the dynamic analysis of the Koyna Dam subjected to three distinct excitation records.

Results

Based on noise-contaminated structural response data, the results indicate that for frequency shifts exceeding 120, the maximum error in estimating the natural frequency of the fourth mode is limited to 2%, while the corresponding maximum error in the estimation of the damping ratio for the same mode reaches 5%.

Conclusion

In comparison with conventional techniques—which typically exhibit reduced accuracy in higher modes—the proposed PSDT-CF method consistently achieves natural frequency estimation errors below 2.5% across all identified modes, demonstrating superior robustness and reliability in modal parameter identification.