Proceedings of the XIV International Symposium on Dynamic Problems of Mechanics (DINAME 2011),
Si., Kurka, P. R. G. (Editors), ABCM, São Sebastião, SP , Brazil, March 13th - March 18th, 2011
Performance evaluation of control strategies with different
feedforward error sensors in active noise control
Siviero, D. A. , Goldenstein, A. L. , and Arruda, J. R. F.
1Universidade Estadual de Campinas, Departamento de Mecânica Computacional, Rua Mendeleyev, 200, Campinas,
Abstract: One of the problems faced today in the implementation of active noise control (ANC) applications is the
choice of error sensors to provide the best control performance for a given target noise abatement, for example, in an
airplane, where to place error sensors and which typeof sensors to use to improve aircraft panel transmission loss at
given frequencies Following these lines, this work shows the performance comparison of feedfo rward controllers in
both time and frequency domain with different error sensors, namely a pressure sensor (microphone), widely used in
this type of application and a particle velocity sensor (Microflown®, mod. Standard PU), which isbelieved to be less
subject to ambient noise not correlated with the noise to be abated (e.g., boundary layer noise). Tests have been
conducted in a plane wave tube using as control actuator a smart-foam under development, which uses a piezoelectric
actuator. The controller performances were evaluated by the gain in sound transmission loss when the active noise
control was on.
Keywords: AcousticNoise Control, Transmission Loss, Adaptive Filtering, Smart Foam
Feedforward control algorithms based on adaptive filtering are widely used and tested in different areas due to the
simplicity of design and ease of application. (Kuo and Morgan, 1996; Widrow and Stearns, 1985 ). In ANC problems,
many studies have used the Filtered-X LMS, FX-LMS, (Siviero et al, 2010, Donadon etal, 2006; Kim, Kim, Roh, 2002;
Guigou and Fuller, 1998; Gentry, Guigou and Fuller, 1997), where the control signal which is sent to the acoustic
actuator is obtained through an adaptive filtering of a reference signal that must to be highly correlated with the primary
disturbance (noise to be controlled) and preferably not affected by the control signal (otherwise this influence must be
filteredout). The adaptation of this filter is performed based on the instantaneous error signal, which is the sum of the
plant response when it is excited by an exogenous signal and the plant response to the control signal
The well known equation that adjusts the FX -LMS controller filter weights which works with a reference signal, a
control signal and an error signal can be written as:
whereis the reference signal, filtered by the estimated secondary plant obtained from an offline (or online)
is the instantaneous error signal,
is a vector with the gain values of the FIR filters,
is a vector with the future values of FIR filter, before the adaptation step and μ represents the size of adaptation step,
which must be between 0 and 1 to ensure the normalizedalgorithm stability (Widrow and Stearns, 1985).
As the control strategy tested is adaptive and attempts to minimize the instantaneous error, the more the error signal
is correlated with the sum of plant responses to excitation and the control signal, the better is the performance of the
controller, or, in other words, if the sensors capture less noise uncorrelated with the signals of interest thecontroller will
have better performance. This fact motivates the attempt to use of particle velocity sensors. The sound pressure and
particle velocity have a different physical nature. While the sound pressure has omnidirectional sensitivity, the particle
velocity is a vector, and, therefore, is directional. Thus, the particle velocity sensor will only capture the particle
velocity in the...
Ler documento completo
Por favor, assinar para o acesso.