A simplified method of predicting the output probability distribution based on the time series model for actual sound insulation systems with arbitrary random excitation by use of the statistical method of moment

Abstract

Most of the usual studies on a sound insulation system were introduced from the deterministic viewpoint in the frequency domain. This paper presents for the first time a fairly simplified identification method applicable to the actual situation of a sound insulation system under the introduction of a time series system model as seen from the probabilistic viewpoint. This identification method is based on the statistical data obtained from a subexperiment with use of a typical white noise input and the well-known moment method. Then, on the basis of these identified results, a simple method is proposed for predicting the output probability distribution of the insulation system with an arbitrary noise excitation. Finally, the proposed identification and prediction method is applied to several types of actual sound insulating systems and its effectiveness confirmed experimentally.