LMS算法.docVIP

学 号 密 级 哈尔滨工程大学论文 院系名 称动力与能源工程学院专 业 名 称： 轮机工程学 生 姓 名： 黎文科 指 导 教 师： 杨铁军 教授2014年6月 摘要振动主动控制技术传统的被动控制技术的不足可以用它来弥补，主动控制技术可以有效地控制低频噪声和振动并且自动跟踪声振频率的变化。在实际的控制系统中，误差通道是影响减振降噪效果和系统稳定性的主要因素之一，它主要包括D/A、功率放大器、A/D等。考虑到误差通道传递函数S(z)的影响，FXLMS算法作为LMS的延伸在主动控制中得到广泛的应用。在实际系统中，S(z)是时变或者非线性的。因此，保证FXLMS算法在声振主动控制系统中的收敛性，对误差通道的辨识有着重要的实际意义。 虽然传统的前馈FXLMS算法以其良好的控制效果和自适应性而得到广泛应用，但其有一个严重的缺点：需要参考信号这在很多情况下是很难保证的。因此，需要采用反馈结构的控制算法也称为内模算法，它通过误差信号来估计原始的声振信号，并用估计值来作为参考输入信号。 采用的自适应内模算法来实现声振的主动控制在MATLAB环境下，对主动控制系统进行了仿真研究。仿真结果，在不同形式激励条件下，采用具有功能的自适应内模算法来实现的主动控制取得了比较满意的控制效果鲁棒性。 Abstract Active vibration control (AVC) has been widely applied in industry to reduce environmental vibration because of its more efficient and economical than the traditional passive methods for low-frequency noise and vibration suppression and its ability of tracking the disturbance under the time varying phenomena．In practical control systems, the secondary path comprising the D/A converter, smoothing filter, power amplifier and A/D converter actuator, physical path, error sensor, and other components is one of the key affecting factors for noise and vibration reduction and the stability of the system .The FXLMS algorithm is an extension of LMS algorithm for active noise and vibration control systems, which takes into account of the influence of secondary path transfer function S(z) .In some practical cases, S(z) can be time varying or non-linear. For these cases, online modeling of S(z) is required to ensure the convergence of the FXLMS algorithm for the active noise and vibration control system .So the modeling of secondary path is important and practical. Though typical Feed Forward Filtered-x Least Mean Square algorithm (FXLMS) has the advantage of high control correction rate and strong adaptive capacity for non-stationary response, it has a critical defect that the reference signal of the external excitation should be obtained which is very difficult for some situation. So a fe