基于神经网络的动反馈管道有源消声系统声学通道辨识  

IDENTIFICATION FOR ACOUSTICAL PATHS OF THE ACTIVE NOISE CONTROL SYSTEM IN DUCTS USING MOTIONAL FEEDBACK LOUDSPEAKER BASED ON NEURAL NETWORKS

作　　者：吴斌[1] 费仁元[1] 周大森[1] 杨立勇 

机构地区：[1]北京工业大学机电学院,北京100022

出　　处：《机械工程学报》2002年第2期100-103,共4页Journal of Mechanical Engineering

基　　金：北京市自然科学基金资助项目(3002004)。

摘　　要：介绍了一种采用动反馈式扬声器的管道有源自适应噪声控制系统，系统中动反馈扬声器既作为有源消声系统的执行元件，也作为检测控制效果的声传感器。分析了系统中存在的各声学通道，并进行了基于自适应线性神经网络的辨识。该方法实现简便，对工作环境无干扰，各声学通道传递函数估计精度高。An active adaptive noise control system in ducts is proposed，which adopts a motional feedback loudspeaker not only as an noise controlling actuator but also as a detecting sensor. By analysing each acoustical path, identification based adaptive linear neural networks is made. This kind of identifing method can be achieved expediently, and it would't bring disturbance to working environment. The estimated result of each acoustical path agrees well with its real acoustical character, respectively.

关 键 词：有源消声 动反馈扬声器 神经网络辨识 发动机 

分 类 号：TK403[动力工程及工程热物理—动力机械及工程] TB535.2[理学—物理]