Abstract
This paper presents numerical verification of error detection method of the frequency response function (FRF) between an actual acoustic field and its finite element (FE) model. The model updating approach that uses a sensitivity of FRF and optimization algorithm has been conventionally applied for greater accuracy. Usually, a parametric design variable such as the acoustic impedance is adopted as an updating variable, which allows the calculation to be more easily generated. However, the actual error factors are included in a non-parametric case. For this reason, its application is limited. The accessible concept of utilizing the driving point FRF has often been practiced in engineering. With this technique, a driving point FRF of each point in the acoustic field is compared with an actual acoustic field and the FE model to detect the error factor. Nevertheless, it is difficult to recognize such a factor precisely. Because, the FRF is calculated from the time history, which contains reflection waves from the entire system, and the driving point FRF involves not only the participation around the excitation point but also the entire system. We therefore propose a new approach using a driving point transient response. When a certain point is excited, the sound wave propagates from the excitation point toward the whole system. A driving point transient response, which is immediately measured after excitation, includes only the participation around the excitation point. By comparing the driving point transient response with an actual acoustic field and the FE model, it is possible to clarify whether the excited segment is an error factor. In this paper, as a verification, numerical result of error detection method is shown.
