Active noise control with adaptive algorithms applied to a finite element model

The simulation of an active noise control (ANC) system in a design stage typically occurs after the estimation of the impulse responses characterizing the environment and exploiting the recordings of the disturbing and perceived noise signals as inputs to an analytical or a numerical model. However, this way of proceeding can be a limiting factor in some circumstances, since the impulse responses are typ-ically fixed in time after the estimation and do not consider potential environmental changes. This article is aimed at investigating the application of an ANC system with an adaptive algorithm to an acoustic model realized with the finite element method (FEM). The FEM model was realized with COMSOL Multiphysics® and numerically solved in the time domain. At each time instant, the adaptive algorithm implemented in MATLAB® interacts with the FEM model through LiveLinkTM with MATLAB® to adaptively modify the boundary condition associated with the control source, generating the cancelling noise signal. In this preliminary study, the secondary path inside the FEM model was estimated with a procedure analogue to the classical offline estimation carried out in a real application. The algorithm chosen for the validation of the method is the classical feedforward version of the filtered-X least mean squares (FXLMS). The analyses were carried out at different temperatures and with a different value of the wall impedance, observing the influence on the ANC performance in the presence of different pure tone signals. The results show the potentiality of this method as a support during the design stage, with the possibility of interaction with other physics.