Analysis of active noise control plants using identification and signal processing methodology

Abstract

In active noise control (ANC), to support the control systems design and controllers parameterisation, it is important to become familiar with plant features and possess its good parametric model. Due to significant complexity of coupled acoustic and electric phenomena, it is not feasible to build a sufficiently precise phenomenological model, although a physical analysis is obviously very useful. The experimental methodology developed in this paper provides information on passive features of ANC plants as well as offers some guidelines, supported by theoretical explanations, for the choice of sampling rate, system configuration and set-up. It is also proposed how this methodology can be applied to examine non-linear effects and evaluate the associated performance limitations. Plant non-stationarity is also checked. Next, the procedure for determining parametric model of the plant of proper structure is given. Within this stage, non-minimum phase phenomena of two kinds are widely discussed. Although the analysis presented in this paper is general, as the example for experiments, an active personal hearing protection device (APHPD) applied to the artificial head is used. It is a representation of compact acoustic plants (CAP). Such plants are characterised by the fact that due to the geometrical arrangement of loudspeakers and microphones, in discrete control, transmission times of acoustic waves are shorter than the transmission times of corresponding electric signals.