Abstract
This work examines the acoustic modelling and design of Helmholtz resonators (HRs) for the noise in a shop-floor cabin housing a friction stir processing (FSP) machine, whose spectrum has a dominant tone at 470 Hz. The main contribution lies in the modelling and design of the resonators and in the identification of the methodological requirements for an in-cabin validation; the experimental part is exploratory. Four HR layouts (single, dual, in-line and in-line dual) were selected by direct finite element (FE) simulation, 3D-printed in PLA and tested in the cabin with and without a pyramid acoustic panel. A lumped-element model with thermoviscous losses, combined with a transfer-matrix model, reproduces all FE resonance frequencies within 10% and shows that the bandwidth is set by the neck conductance and the peak by the neck resistance. A response surface fitted to a central composite design was not significant (p = 0.60) and was not used, so the geometries are FE-selected designs, not statistically validated optima. In the cabin, the 470 Hz level changed by 0.1–2.1 dB without the panel and 1.2–6.3 dB with it, but repeat recordings differ by up to 5.3 dB, and a room energy balance predicts reductions below 1 dB. Although the influence of various resonator models and acoustic panels in controlling noise is clearly established, the reduction in noise level remains on the lower side. The machine radiates the tone, and a resonator near the source could change its sound power by at most 0.5 dB. No reduction of the overall, A-weighted or occupational exposure level is claimed. The study provides a design method and a measurement protocol for a conclusive in-cabin validation.