A Study On The Clustered Active Noise Control System

With the development of engineering technology,active noise control(ANC)technology is widely used in open space or large size enclosed space.However,in order to reduce noise in such environments,ANC system usually requires a large number of secondary sources and error sensors.The traditional centralized control system leads to an increasing cost in computing capacity,hardware complexity of controller implementation with the increase of the system size.To reduce the complexity of large-scale systems,a decentralized control system is proposed.The system consists of a number of independent feedback control units,each of which includes one error sensor and one secondary source.In such an unit,only the secondary-path transfer function of its own needs to be considered,regardless of the influence of other control units.However,comparing to the centralized system,the decentralized control system may produce global control performance reduction and system instability,and the secondary source and error sensor must be corresponded.The clustered control system consists of a number of localized multi-channel sub-control units,which are small centralized control systems.Compared with the decentralized system,the number of secondary sources and error sensors for each control unit of clustered system can be flexibly selected as necessary.The control unit includes more interactions between error sensors and secondary sources,thus better global performance and stability can be expected.This dissertation systematically studies the adaptive algorithm,control unit partition methods,system layout,steady-state performance and stability of the clustered ANC system in various acoustic environments.The main contents include(1)The performance analysis of the clustered and centralized system using adaptive algorithm is conducted,a stable condition of clustered system is obtained.(2)The equivalent condition of the clustered system and centralized system is studied in free space.The condition is that the number of the system error sensors is less than or equal to the number of the secondary sources,and the secondary-path impedance matrix of the system and secondary-path impedance matrix of the control unit are full rank.The rank of secondary-path impedance matrix of the control unit is positively correlated with the control performance.A guideline for configuring the control units of clustered system is proposed.The influence on rank of numbers and positions of secondary sources and error sensors of control unit is also studied.(3)In a rectangular enclosed space,the influence of positions of the secondary sources and the error sensors of the clustered system on performance is studied.It is found that the system has better control effect when the secondary source and the error sensor avoid the zero line of acoustic modal,and the impedance matrix of the clustered system is always full rank in the rectangular enclosed space.The clustered system using adaptive algorithm is researched,and it is found that the clustered system is usually unstable in the enclosed space.The clustered system can be stabilized by regularization processing of the cost function.A simulation study is conducted for an aircraft cabin.The result of the study which is in a rectangular enclosed space is also appropriate for complex enclosed space.(4)In a full-sized helicopter cabin mock-up,performance analysis of the centralized,decentralized and clustered system is carried out,respectively.The results show that the control effect of three kinds of systems is consistent when the equivalent conditions are satisfied.The stable frequency range of the decentralized system and the clustered system are compared.The result implies the advantage of the clustered system than the decentralized system in stability.(5)Through cabin mock-up experiment and the real aircraft flight experiment,the clustered system in propeller aircraft cabin is studied.The results show that the noise reduction effect at the monitoring points of the clustered system is similar to the centralized system.It also preliminarily proves the feasibility of the clustered system in prctical projects.