A Method For Accommodating Frequency Mismatch In Narrowband Active Noise Control System

In a narrowband feedforward active noise control(ANC) system, the reference signal is usually generated through a nonacoustic sensor. When the nonacoustic sensor or the cosine wave generator contains errors, the measured synchronous frequencies of the reference signal will be different from the true frequencies of the primary noise being targeted. This difference is referred as frequency mismatch(FM). The narrowband ANC system may suffer significant performance degradation when FM exists. In this thesis, the narrowband ANC system with adaptive linear combiner is considered for investigating FM problem.Based on AR model, a new FM compensation method is proposed in this thesis, through adding a delay module and a corresponding amplitude and phase adjuster. In essence, this method constructs signals containing implied true frequencies through delays and superposition so as to raise the phase matching speed and accuracy. Through lots of simulations, we investigate performances of the system based on delay model, as well as changes of the adaptive parameters when the system is running. We also analyse the influence of the adaptive parameters and retardation coefficient values on the performance of the system. Compared with the frequency compensation method based on AR model, the designed frequency compensation method can effectively deal with FM with faster convergence rate, which increases limited computational complexity.On the basis of simulation results, furtherly we design and build a one-dimensional duct ANC experimental system to implement the algorithm by using digital signal processor(DSP). The traditional narrowband ANC system and the proposed ANC system for mitigating FM are implemented in the experimental system. The experimental results verify the feasibility of the ANC experimental system and the effectiveness of the frequency compensation method based on delay model.