To address strong multi-source coupling, complex transmission paths, and reference-channel crosstalk in operational transfer path analysis (OTPA) of marine steam turbine generator sets, this study proposes a decoupled OTPA method based on vibration source coupling separation. Unlike OTPA, which directly uses measured mixed signals as inputs, the proposed method performs source decoupling at the input stage before solving the transfer matrix: a coupling separation model is established based on the mechanical coupling transmission relationships among sources, decomposing the measured mixed responses at the measurement points into mutually approximately independent pure source signals, thereby suppressing source crosstalk at its origin. Meanwhile, multiple coherence analysis based on partial coherence is employed to examine the correlations among source measurement points, sensor placement, and path completeness. When all 12 sources are included, the multiple coherence coefficients of all target points exceed 0.9, indicating that the selected reference channels capture the major vibration contributions. At 100 Hz for target point 1, the dominant path ranking obtained by decoupled OTPA agrees with conventional TPA on five of the top six paths, while classical OTPA shows one additional deviation. The decoupled response spectra are approximately 10–20 dB lower than conventional OTPA results, suggesting reduced overestimation caused by reference-channel coupling. Conventional OTPA exhibits obvious path contribution misallocation, whereas the decoupled OTPA can effectively reduce source crosstalk and improve the stability of transfer matrix solving and the accuracy of path contribution identification, providing a solution for rapid diagnosis of vibration transmission paths under operational conditions.
No takes yet. Share an insight, caveat, or question.
Li et al. (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: