基于小波包分析的声强计算方法

传统的基于FFT的声强计算方法是分析平稳噪声信号的有效工具，但是它却无法对机电设备发生故障时所辐射出的非平稳噪声信号进行有效的分析。由于小波包分析可以实现非平稳噪声信号在不同频带和不同时刻的合理分离，因此可以利用小波包分析对声强进行计算。文中应用自行研制的噪声自动分析系统对声强的计算方法进行了研究，提出了一种基于小波包信号分析技术的声强计算方法：并通过实验验证了该方法的正确性。该方法不同于传统的基于FFT分析的声强计算方法，可以实现对非平稳故障噪声信号的分析，为机电设备的噪声监测和故障诊断提供了一条研究途径。

Sound Intensity Computing Method Based on Wavelet Packets

The traditional sound intensity computing method based on FFT is an effective tool to analyze stationary noise signals, but it isn't suitable for nonstationary noise signals emitted from faulty mechatronic equipments. The advent of wavelet packets gives a chance to solve the above problem. Wavelet packets analysis can decompose nonstationary noise signals reasonably into different frequency bands at different times. So wavelet packets can be applied in sound intensity analysis. A new type of automatic noise analysis system, which is self-developed, is used to investigate the application of wavelet packets to sound intensity. The theory of wavelet packets is introduced, and a sound intensity computing method based on wavelet packets is presented. The validity of the method is confirmed by experiments, and the characteristics of the method and its application to fault diagnosis are discussed. The method differs from traditional FFT transform in its capability to analyze nonstationary noise signals of faulty mechatronics equipments. It offers a way for noise monitoring and fault diagnosis.