46 J. C. Akers and Joel Sills Fig. 5.13 CMIF overlaid with pole estimates Fig. 5.14 Test vs FEM cross orthogonality polyreference techniques extracted the majority of the test modes, with a Single Degree Of Freedom (SDOF) polynomial curve fitting tool used a few times. Advanced modal extraction techniques, such as mode enhancement and spatial filtering, were not used. Figure 5.13 shows a CMIF overlaid with pole estimates forming well behaved race track patterns indicating the test modes are straight forward to extract. Self and cross-orthogonality criteria were used to determine how well the “test” primary target modes could be identified. If the magnitudes of the off-diagonal of the self orthogonality of the “test” primary target modes and the cross orthogonality of the “test” and FEM primary target modes being 5% (10%) or less, and the magnitudes of the diagonals of the cross orthogonality of the “test” and FEM primary target modes being 95% (90%) or greater, then the “test” primary target modes were well (adequately) identified. Figure 5.14 shows the cross orthogonality between all identified “test” modes and FEM modes in the frequency range of interest with the ML having 1% modal damping and sensor and ambient background noise included. Figure 5.15 shows the corresponding cross orthogonality between just the identified “test” primary target modes
RkJQdWJsaXNoZXIy MTMzNzEzMQ==