Active control of noise through open windows

Active noise control (ANC) is a noise abatement technique that preserves an abode’s natural ventilation when applied on domestic windows. ANC systems for open windows mitigates noise at the receivers’ end, which augments solutions that are less effective for urban noise in high-rise cities. The target of the ANC system is to reduce noise that propagates through an open aperture, such as a window, into the interior of a room. This thesis presents a systematic approach to the open window ANC system design to address the complex practical implementation challenges. Firstly, the fundamental limits of the ANC system in controlling noise through an open aperture are determined by numerical simulations. The investigation into the performance of different physical arrangements reveals that the upper limit frequency limit of control is determined by the separation distance between the sources of the planar array, distributed evenly in the aperture. As a result, the minimum number of sources needed for good control is generalized for a rectangular aperture and the dominant noise incidence angles. Based on the design guidelines for the open aperture, a further numerical study was conducted to investigate the influence of passive elements (i.e. glass panel) in the aperture on the active control performance. Moreover, it was shown that the arrangement of control sources in a partially glazed aperture (~75%) based on the established design guidelines could provide more attenuation than a fully-glazed aperture without active control. Next, an experimental setup with a full-sized domestic window is designed and installed on a room model placed in a recording studio. The active control units were custom designed and secured within the aperture via the security grille. A total of 24 control sources of the multiple-input-multiple-output (MIMO) ANC system are implemented on a fully-opened two-panel sliding window. The active control attenuation performance was predicted to be 10 dB at the error microphones through offline simulations with measured transfer functions. Due to the high computational complexity, the real-time active control system was implemented on an FPGA platform, which operated at 25 kHz sampling rate. The active control performance was determined by an array of 18 microphones distributed within the interior of the test chamber, of which 12 microphones monitored the noise directly in front of the aperture in a plane array, and the rest were distributed according to the ISO standard to measure the energy-averaged sound pressure level of the room interior. The active control performance was compared to the passive insulation provided by a fully-closed window. In both tonal and broadband noise scenarios, the active control system was comparable to the passive insulation of a fully-glazed window in the mid-frequencies and at oblique angles of incidence for the entire frequency range of interest. Lastly, considerations to increase the practicability of the active noise control system for domestic windows are discussed. Namely, the omission of error microphones but implementing fixed filters and the reduction of the number of control sources that are distributed in the aperture in an effort to reduce the physical and visual obstructions.Doctor of Philosoph