Book Description

Active noise control (ANC) has drawn widespread research interest in both academia and industry during the last two decades, especially in the automotive industry where better vehicle noise, vibration and harshness (NVH) performance is one of the important concerns for vehicle manufacturers. However, most of the current state-of-the-art ANC systems are mainly focused on stationary noises by following Gaussian distribution, and there is limited research work on ANC of impulsive and/or impact noise. In practice, there are many unwanted impact responses at industrial sites and in motor vehicles, such as noise generated by punching machines, combustion engines and pile drivers as well as impact road noise due to road bumps. This type of noise may pose a difficult challenge for the prevalent adaptive control algorithms, namely the filtered-x least mean square (FXLMS) algorithm. Hence, more in-depth understanding is required to design an ANC system with robust and efficient adaptive algorithms for treating impact acoustic noise. The first part of the dissertation concentrates on the development of more robust adaptive algorithms for generic impulsive noise. First, an enhanced modified filtered-x least mean M-estimate algorithm (MFXLMM) is proposed to tackle the impulsive-type noise, where the reference signal is fed through the Hampel's three-part M-estimator by setting thresholds for impulses. Second, a family of enhanced M-estimator based algorithms is proposed, where different types of robust criteria based on robust statistic theory are applied for non-Gaussian impulsive noise. This family of algorithms generalizes all the existing studies by previous researchers. Third, more specific observations are made for the convergence performance of the modified FXLMS algorithm for impact noise, where thresholds in the reference and error signal paths are added. This analysis is to address the underlying mechanism of the enhanced algorithm for impact noise due to the white noise and impulses with different durations. The second part of the dissertation is on impact noise with certain repetitiveness. First, a specific analytical convergence analysis of the FXLMS algorithm for repetitive impulse-induced noise is developed, where pure delay secondary path models are assumed. A step size bound is obtained which is in agreement with the previous studies. Moreover, a general secondary model is considered in the numerical simulations to further demonstrate the feasibility of applying the FXLMS algorithm for repetitive impact noise. Second, comparative experimental and simulation work are conducted to further validate the performance of FXLMS algorithm for repetitive impact noise. The feedback system with the iterative learning control (ILC) algorithm is also designed to compare with the FXLMS algorithm. Finally, a concluding work is performed to apply the enhanced FXLMS algorithm to impact road noise control. A frequency response function (FRF)-based sub-structuring technique is implemented to develop a coupled vehicle system model, where the passenger compartment is simplified as a 3-dimensional flexible-panel backed cavity model and the tire-wheel system is modeled as a flexible ring element and rigid wheel. Numerical simulations are conducted to validate the effectiveness of the proposed ANC system for interior impact road noise.