Several Key Algorithms Of Active Noise Control System For Broadband And Narrowband Hybrid Noises In A Duct

Environmental noise has a serious impact on human’s work and life, conventional passive noise control systems are usually bulky, expensive, difficult to install, and inef-ficient at low-frequency noise. Active noise control (ANC) is particularly effective for low-frequency noise, which is based on the principle of destructive interference of sound waves. The secondary noise is generated by electroacoustic devices and is same amplitude but opposite phase position with the original noise. Since only plane waves propagate in narrow ducts, the sound field of a duct is simple for the application of ANC.Broadband ANC (BANC) and narrowband ANC (NANC) systems are two impor-tant directions in the active noise control fields, which divided the environmental noise into broadband and narrowband. Broadband refers to the energy being concentrated in a band, while narrowband refers to the energy being concentrated in a single or several frequencies. However, the real environmental noise shows the mixed characteristics of both broadband and narrowband. The new hybrid control system works on the control of the mixed noise of broadband and narrowband. The system consists of broadband ANC (BANC), narrowband ANC (NANC) and sinusoidal noise canceller (SNC), among which the SNC is the core of the system and requires a high estimation accuracy and tracking capabilities so that to coordinate the work of the subsystems effectively. Since the characteristics of the acoustic noise source and the environment are time-varying, the frequency content, amplitude, phase, and sound velocity of the undesired noise are non-stationary. An ANC system must be adaptive to cope with these variations. On the other hand, convergence speed, steady-state error are also the key characteristics that reflecting the system performance. Frequency Mismatch (FM), which is a frequency deviation of the reference signal measurement, may significantly deteriorate the steady-state perfor-mance. The traditional compensation algorithms ensured the steady-state performance at a sacrifice of the convergence rate. This paper dedicates to solving the above key technical problems, and providing a high-performance active noise control system.Firstly, an adaptive variable step-size LMS (VSS-LMS) analyzer is introduced to the SNC, and its corresponding analysis as well as various simulations is carried out for the guidance of the coefficients selecting. By introducing a step-size balance term to the cost function, the obtained algorithm can keep stable itself without the range limits setting. The analysis of the improved SNC system is made tractable by moving one step forward of the variable step size sequence. Difference equations governing the dynamic of the algorithm are developed. Extensive simulations are conducted to prove the effectiveness of the VSS-LMS analyzer, and to verify the validity of the analytical findings.Secondly, a variable step-size filtered-x LMS (VSS-FXLMS) algorithm and its mod-ified versions are proposed for the narrowband ANC system to cope with the system in the nonstationary noise environments. The different equations governing the dynamic of the algorithm are derived to provide the guidance of the coefficients selecting. Based on this system, an improved frequency mismatch (FM) compensation method is proposed to solve the problem of large nonstationary FM.Thirdly, the mean convergence of weight errors of the BANC system in the presence of multiple harmonic excitation is investigated. The obtained analytical results enrich our understanding of the BANC operated in an environment with multiple sinusoids. Exten-sive simulations are conducted to support the analytical findings. Then a modified VSS-FXLMS method is introduced to the main controllers of BANC subsystem in HANC. In this way, multi-VSS based hybrid ANC (MVSS-HANC) with three VSS based subsys-tems is composed to provide an excellent comprehensive performance.Finally, experiments based on tube ANC system are implemented for the VSS algo-rithm. Simulations based on real IIR secondary path and real factory noise are carried out to provide more convincing comparisons.