The System for Active Control of Sound Transmission through a Window Panel – The Concept and Simulation Results

Abstract

High acoustic insulation windows are common mean to decrease sound transmission to closed spaces. Hence, the improvement of sound insulation of windows is very important research issue. In this paper partial results of research project aimed at synthesis of the window panel with actively controlled sound transmission are presented. Recent stage of the project is focused on the development of multichannel Active Structural Acoustic Control algorithm. High amplitude excitation produces nonlinear vibration effects. Using feedforward control strategy it is convenient to control not only the reference signal spectrum, but also additional frequencies generated as the cause of nonlinearity. Therefore neural network based algorithm is considered. To lower the computational burden of the algorithm, round robin based error backpropagation learning is employed. The results of numerical simulations are presented. Major conclusion of presented part of the research is that the algorithm is capable of controlling a wide-band noise efficiently. Reduction of the computational complexity leads to increased convergence time, not influencing the final value of mean square error.