Practical improvements to real and imaginary spectral based modal parameter measurements of SDOF systems

Determining material damping type by comparing modal frequency estimators

“…Unfortunately, this method is sensitive to phase errors. 5 It has been assumed here that one damping type is dominant; in some applications both types may be active. Finally, it is seen to be valuable to use statistical analysis to allow conclusions regarding the damping type to be drawn, due to variations in repeated experimental measurements.…”

“…From Eq. (4) it is easily seen that Re = n independent of both the response parameter and the damping mechanism, 5 where the trivial solution = 0 is ignored. The use of acceleration ͑͗p͘ =a͒ is not considered further here.…”

“…This is due to the fortuitous alignment of the discrete points on the frequency axis with the true position of the characteristics of the spectrum. 5 However, here the focus is on non-lightly damped systems. It is seen that ␣ Re : X determines the damping correctly independent of damping type.…”

Improving the accuracy of the n-dB method for determining damping of non-lightly damped systems

“…The authors have previously shown that using the real spectral part the system natural frequency can be determined using any response parameter, and that the damping can be determined accurately for non-lightly damped systems with either viscous or hysteretic damping [3]. Such measurements are sensitive to phase offset errors, although techniques exist to correct for them [4]. Here, damping measurements are considered using the spectral magnitude, which is still a popular technique.…”

Accuracy and robustness of four basic single degree of freedom methods for determining the modal parameters of non-lightly damped systems

“…Here the effects of measurement noise or quantisation of the frequency axis are not considered, the former could be reduced by averaging multiple measurements and the latter has been studied by the first author [9] for modal parameter measurement using quadrature spectral parts. Whilst both these affect accuracy, even if they are minimised, an exact measurement strategy that has no bias error should still be employed.…”