Errors in the measurement of structure-borne power flow using two-accelerometer techniques

This paper presents a theoretical study of three errors in the measurement of structure-borne power flow in one-dimensional structures using two-accelerometer techniques. It is assumed that the physical and material properties of the test structure are known. For the measurement of bending wave power flow, the errors due to the neglect of shear force and rotating inertia and the presence of longitudinal waves are examined individually. It is shown that the omission of shear force and rotating inertia results in a large bias error at high frequencies. The presence of incoherent longitudinal waves results in no bias error for the biaxial accelerometer technique or usually a negligible bias error for the two-accelerometer array technique. However, if longitudinal waves are coherent with bending waves the bias error increases with increasing the coherence and the longitudinal to bending wave energy ratio and becomes large when the bending wave energy is smaller than the highly coherent longitudinal wave energy. For the measurement of longitudinal wave power the effect of bending waves is important. As long as the bending wave power is not much smaller than the longitudinal wave power, the bias error is large even if bending and longitudinal waves are incoherent. Measures should be taken during measurements to eliminate or reduce the contribution from bending waves to a certain extent, otherwise, the longitudinal wave power cannot be measured using the two-accelerometer array technique. © 1997 Academic Press Limited.