Adaptive membrane-type acoustic metamaterial with autonomous tuning to incident sound fields
Final Report Abstract
Membrane-type acoustic metamaterials (MAM) consist of a thin pre-stressed membrane with masses attached to the membrane. Even though these metamaterials are very lightweight and thin, they exhibit narrow frequency bands (so-called anti-resonances) in the low frequency range with very high sound insulation values - much higher than a homogeneous wall with the same mass. This makes MAM a very promising solution for noise control applications with strict constraints on the mass and size of noise control treatments. The narrow bandwidth of the anti-resonances, however, is a considerably inhibiting factor for the application of MAM to noise sources with changing frequencies of tones (e.g. generated by propellers). The project “Adaptive membrane-type acoustic metamaterial with autonomous tuning to incident sound fields” aimed at developing and demonstrating an active MAM which is capable of autonomously changing its properties (e.g. the anti-resonances) to the characteristics of an incident sound field with a spectrum containing both broadband and tonal components. The main objectives of the project were to demonstrate a much larger noise insulation performance of the adaptive MAM, compared to the non-adaptive MAM, to show that the adaptive MAM is robust with respect to environmental influences, and to have a real-time capable adaptive MAM, which can respond rapidly to changes in the incident sound field. During the project, different actuation mechanisms have been evaluated which, based on an electrical input signal, apply a force to the MAM that can be used to adapt the properties of the MAM. An electrodynamic exciter, consisting of a voice coil driving an inertial mass, was identified as a suitable actuator, because these exciters are readily available and can be very small and lightweight. Numerical simulation models of MAM with these actuators were created to study the dynamics of the MAM-exciter system and investigate possibilities for reducing the complexity of large-scale set-ups containing multiple MAM. Apart from the actuation, suitable control algorithm were studied which use one or more input signals from sensors and convert these into electrical signals provided to the actuator. These algorithms were implemented in simulation models which were coupled to the MAM-exciter models to evaluate their performance for different incident noise fields with varying tonal frequencies. The following control algorithms proved to be most promising for the adaptive MAM: (1) A frequency tracking algorithm which identifies the frequency of the loudest tone in the incident noise and adjusts a gain value in the controller according to a pre-computed look-up table, so that the anti-resonance frequency of the MAM is shifted to match the tonal frequency; (2) an adaptive filter based on the filtered-reference least mean squares (FxLMS) algorithm, which is widely applied in active noise control applications. The simulations performed in this project indicated that the frequency tracking algorithm is capable of very rapidly (less than 100 ms) responding to large changes in the tonal frequency. However, the pre-computed look-up table was not very robust when the properties of the MAM changed, e.g. due to changing ambient temperature leading to changes in the membrane tension. The FxLMS algorithm, on the other hand, albeit somewhat slower in response time, was also capable of tracking tonal frequency changes and proved to be very robust. Also, using the FxLMS algorithm it was possible to achieve improved noise insulation properties over a broad frequency band. At the end of the project, test samples of the adaptive MAM with a miniature electrodynamic exciter have been built to demonstrate their performance in an experimental impedance tube setup. Altogether, the experimental results obtained in the project confirm the behaviour of the adaptive MAM expected from the simulations. Furthermore, the stability limits of the different control algorithms could be identified. In the final step, extending studies have been performed on the experimental setup using an accelerometer on the MAM aiming to replace the error microphones and making the integrated combination of sensors, actuators, and controllers within the added mass of the MAM possible, paving the way for miniaturization and optimization of adaptive MAM designs in follow-up research.
Publications
-
A Preliminary Investigation of an Active Membrane-type Acoustic Metamaterial. In: Proceedings of Inter-Noise 2022. Glasgow, 2022
F. Langfeldt & J. Cheer
-
Evaluation of Active Control Concepts for a Self-adjusting Membrane-type Acoustic Metamaterial. In: Proceedings of DAGA 2022. Stuttgart, 2022, pp. 240–243
F. Langfeldt & J. Cheer
