Operational transfer path analysis with crosstalk cancellation based on least variance spectrum estimation

As the evaluation accuracy of operational transfer path analysis (OTPA) is seriously affected by crosstalk effect, OTPA with crosstalk cancellation based on least variance spectrum estimation with stronger anti-noise performance is proposed in this paper. Firstly, to overcome the insufficient variance performance of conventional methods, the least variance spectrum estimation method is proposed to obtain more stable power spectrum. Secondly, after Burg algorithm is used to determine the optimal iteration times, the prior transmissibility can be calculated from acquired power spectra. Thirdly, reference signals without crosstalk components can be obtained by solving the equations, which eliminates crosstalk effect of OTPA. Finally, according to numerical case studies on a cylindrical shell simulation system and experimental case studies on a shell testbed, the performance of the proposed method is comparatively studied with conventional methods. Generally, the proposed method can increase the reliability and accuracy of transfer path identification and contribution evaluation of OTPA, which can benefit vibration monitoring, reduction and control of mechanical systems.