The active noise-reducing casing developed and promoted by the authors in recent publications have multiple advantages over other active noise control methods. When compared to classical solutions, it allows for obtaining global reduction of noise generated by a device enclosed in the casing. Moreover, the system does not require loudspeakers, and much smaller actuators attached to the casing walls are used instead. In turn, when compared to passive casings, the walls can be made thinner, lighter and with much better thermal transfer than sound-absorbing materials. For active noise control a feedforward structure is usually used. However, it requires an in-advance reference signal, which can be difficult to be acquired for some applications. Fortunately, usually the dominant noise components are due to rotational operations of the enclosed device parts, and thus they are tonal and multitonal. Therefore, it can be adequately predicted and the Internal Model Control structure can be used to benefit from algorithms well developed for feedforward systems. The authors have already tested that approach for a rigid casing, where interaction of the walls was significantly reduced. In this paper the idea is further explored and applied for a light-weight casing, more frequently met in practice, where each vibrating wall of the casing influences all the other walls. The system is verified in laboratory experiments.

It is possible to enhance acoustic isolation of the device from the environment by appropriately controlling vibration of a device casing. Sound insulation efficiency of this technique for a rigid casing was confirmed by the authors in previous publications. In this paper, a light-weight casing is investigated, where vibrational couplings between walls are much greater due to lack of a rigid frame. A laboratory setup is described in details. The influence of the cross-paths on successful global noise reduction is considered. Multiple vibration actuators are installed on each of the casing walls. An adaptive control strategy based on the Least Mean Square (LMS) algorithm is used to update control filter parameters. Obtained results are reported, discussed, and conclusions for future research are drawn.