Study On Moving Noise Source Identification Based On Acoustic Holography Of Neural Network

Noise source identification is a primary and an important task in noise control, because it can provide an effective means to test the quality and improve the noise vibration system. At present the main methods for identifing the noise source include the elimination method, sound intensity measurement method, acoustic holography method etc. Acoustic holography method belongs to the non-contact measurements, and has no limitation of the noise type, so this method is widely used now. But in China, using acoustic holography method for moving noise source identification is still only in its infancy stage.Acoustic holography field transformation technology measures the pressure of holographic surface which is surrounded by the holographic measurements plane, then use the relation of the space between the surface acoustic pressure and transform hologram to get the real pressure of the surface acoustic source. This research developed a set of moving noise source identification and analysis system based on acoustic holography theory. Considering disadvantages of common arithmetic, a new holography reconstruction method based on Neural Network was proposed. In this paper, the main research work completed as follows:1.This paper made a detailed analysis on Doppler effect and Doppler correction. Discussed the characteristic of movement source signal and analysed the Doppler effect which is caused by two main factors: the measurement mikes accept time interval finer signals from the sound source which isn't radiating in interval of time, that leads to the signal frequency shift; Because of the distance between the mikes and moving vehicle noise sources is changing by time, which results in the signal amplitudes changes follow the distances between the sources and mikes. According to the reasons that cause the Doppler effects, some methods are used to get original amplitude message and frequency message of moving noise source.2. Introduced the basic theory of acoustic holography and analysed the sound field numerical transformation process. Considering disadvantages of common sound holography reconstruction algorithm, this article proposed a new acoustical holography method which is based on neural network, to realize the reconstruction algorithm. The whole process of the arithmetic was described in the thesis. At first, Neural Networks has the possibility for reconstrution, then K-L transformation is applied to extract the input, at last, confirm the structure of Neural Networks with an improved BP algorithm. After that, the result of simulation experiment shows that the arithmetic has the feasibility.3.The system can be used to achieve kinds of functions, such as sound source identification, 1/3 octave analysis, the acoustic intensity computation, the data file memory and the read, the system adjustment and so on. The virtual instrument technology is used in system's software design, and the part of software is composed of data acquisition module, data processing module, system calibration module, document preservation and read module. Using virtual instrument can save the system development cost and simplify the hardware structure.4.With experiments we verify the acoustic holography method is correct and superiority. The test was divided into two parts: Firstly, sound boxes were used as sound source, then analysed the sound field of them in order to prove the accuracy and universal of the new methodolog. Secondly, moving vehicle was used as moving noise source. Two single-frequency speakers were put in the vehicle. When the vehicle passed by, two speakers also issued a voice frequency that can be regarded as the main noise source from the moving vehicle, then used acoustic holography method to reconstruct the noise source position, and the results proved that the acoustic holography method can identify the location of the moving noise source effectively and accuratly.The oretical analysis, numerical simulation and practical application all verify that acoustic holography method can identify moving noise source availability and reliability. Previous work supplies great technical support for the development of measurement instrument, and accumulates experience for widespread application of acoustic holography method.