Research On Several Key Problems Of Adaptive Vibration Noise Active Control

The reciprocating movement of the mechanical system around its equilibrium positioncalled vibration. It is a common phenomenon in nature and engineering.Vibration is an important part of basic physics and classical mechanics research. Vibration of mechanical structure is transmitted to people in air by the form of sound. On the interaction of human and mechanical, the vibration in airis usually classified as noise. It can be seen that noise is a special kind of vibration, related research of vibrationand noise problem is often analogical and uniform.Vibration and noise are very common in our living and nutrual environment, such as industrial production and manufacturing, construction process, and all kinds of traffic tools, household appliances work. The wavelength is longer of low-frequency vibration noise below 500 Hz. The traditional passive control method for low frequency vibration noise control is poor, not easy to implement.In order to solve such problem, active vibration noise control gradually began to be widely research and applicationwith the development computer, controller, actuator and sensor.The main principle of active vibration noise control technology is getting the responsedata of mechanical system by real-time acquisition as input parameters, the control signals are calculated by certain active control algorithm and intervenedwith primary acoustic field and structure vibration, so as to achieve the aim of damping noise reduction.For some controlled system structureparameter serious uncertainty condition, using adaptive active vibration noise control can avoid the calculation of the structural parameters, improve the efficiency of control.Classical adaptive active control model is feedforward control system model based on Filtered- x LMS algorithm.However, the algorithm is influenced by series of key problems such as secondary channel modeling, output time-delay and channel coupling in the application process, the actual application is not yet mature.So under the background, this paper studies the influenceofsecondary pathmodeling error on the filtered-x LMS algorithmand the accurate and efficient online secondary path modeling algorithm, and puts forward an online output delay estimation and compensation method, then, applies these research results to multi-channel active control system, and studies centralized multi-channel control system coupling problem,puts forward decoupling scheme. The full paper is divided into seven chapters. The main contents can be summarized as:1, Theoretical basis of adaptive vibration noise active control is introduced. At first, the physical mechanism of active vibration noise control is discussed base on finite beam model and the free sound field model respectively.Starting from the equation of motion, the physical mechanism of active control of vibration noise is derived.Secondly, comparing the advantage of adaptive feedforward algorithm, the derivation process of the filter- X LMS algorithm is given, and analyzed the stability and convergence of the algorithm, making theoretical foil for the key problems of follow-up study.2, Influence of secondary path on active control algorithm is analyzed, the effects of secondary path modeling error on the stability, convergence and control effect of active control system is studied. In order to improve the online secondary path modeling algorithm based on the three aspects, three kinds of online secondary path modeling of algorithms are put forward. Under the principle of trying to reduce the number ofadaptive filters and reduce the complexity of the algorithm, the influence of active control link and secondary channel identification link is eliminated, speeding up the convergence speed and precision of online modeling of the online secondary path modeling system.3, Analysis of the causes of the output delay and its influence to the active control system is analyzed. The output delay estimation and compensation are studied.Firstly, the output delay offline and online estimation method is proposed. Secondly, in order to eliminate the influence of the output delay, filtered time-delay compensation algorithm is proposed. Finally, in order to further improve the convergence of the time-delay compensation method, the output delay link is regarded as a low pass filter. Adaptive compensator was designed to make up for the phase frequency. A time-delay prediction based on LMS algorithm is proposed, realizing the online projections for output time-delay compensation, and eliminating the influence of output delay on control system.4, Because of the slow convergence speed characteristicof LMS algorithm for active broadband vibration noise control, RLS algorithm with fast convergence ability is usedfor broadband vibration noise control.Transfer function of infinite impulse response(IIR) filter is one that has the characteristics of the structure of the zero and pole. Conbining with RLS with IIR structure, adaptive FURLS algorithm can solve the unstable problem of the control system lead from vibration noise secondary feedback, which improves the convergence of the adaptive active control algorithm.Expand single channel online secondary path modelingalgorithm to multichannel adaptive active control system, multi-channel online secondary channel modeling algorithm is derived, for each sub-channel online modeling has it own convergence step length adjustment strategy, online secondary path modeling problem of multi-channel has been solved.5, Coupling between channels weakens the stability ofcentralized multi-channel active control algorithm.Failure of any single actuator or error sensor will lead to the failure of the whole system, and with the increase of the number of channels, multi-channel coupling algorithm will inevitably make a massive increase in the amount of calculation of the control system, increasing the difficulty of the control algorithm.In order to solve this problem, a distributed multi-channel adaptive active control algorithm is proposed. The stability and convergence of the algorithm are analyzed.Matrix decoupling method are put forward inserting decoupling compensator, can effectively completely decoupled centralized multi-channel active controlalgorithm.6, In order to verify the research onadaptive active vibration noise control systems.Active vibration noise control experiment platform is structured based on the CompactRIO of National Instruments. Online secondary path modeling algirithm, output time delay filter compensation algorithm and multi-channel adaptive active vibration control are operated to verify the engineering applicability of these algorithms.