714 Accuracy evaluation index for running transfer path analysis

Abstract

In this study, an accuracy evaluation index for running transfer path analysis (TPA) employing principal component regression method was considered and the reliability was verified. To evaluate the TPA accuracy, the consistency of response signals in which the calculated response signal from the TPA model was compared with the measured response signal was used frequently until now. Also, the condition number that indicates the inverse matrix condition used in the calculation procedure of acceleration transfer function was employed in some cases. In addition to these accuracy evaluation methods, correlated principal component number (CPCN), that indicates the number of principal component correlating with the response signal generated at the running test, was proposed. The reliabilities of these accuracy evaluation methods were verified through simulation. As a result, the conventional two methods (consistency method and condition number) could not evaluate the accuracy well. However, CPCN index could indicate the frequency bands where the TPA accuracy was high or low. Consequently, the new index was found to be a suitable index for accuracy evaluation of the running TPA.