水轮机空化的小波奇异性检测研究

空化现象是水轮机中难以避免的主要危害。空化噪声的谱分析一般采用功率谱分析，而功率谱分析在研究空化噪声这种具有很强突变性的信号具有明显的缺陷。小波分析具有很好的局部性，在奇异性检测和奇异性定位研究中得到了广泛的应用，文中在小波变换模极大值奇异性检测的理论基础上提出了水轮机空化的小波奇异性检测方法，并提出了小波基选择方案和突变点检测条件。较之以往的基于模极大值线搜索的方法，该方法具有运算量小的优点。分析了实际检测数据，结果表明该方法的有效性，并很好的检测出空化初生和空化形态转变。

RESEARCH ON TURBINE CAVITATION TESTING BASED ON WAVELET SINGULARITY DETECTION

Cavitation is the main harm in turbine, which is difficult to avoid. Power Spectrum analysis is always used in spectrum analysis of cavitation noise. However, it is obviously unfit for research on the transient signal such as cavitation noise. Wavelet transform is just localized, which is popularly used in detection and localization of singularity. Methods of turbine cavitation testing based on wavelet singularity detection are presented in this paper. At the same time, mother wavelet option methods and determined conditions are presented. The computation of the methods is simpler than ever methods based on detection of modulus maxima of wavelet transform. The methods of turbine cavitation testing are used in real data analysis and it is proved that the methods are effective and incipient cavitation and change of cavitation pattern can both be detected effectively.