摘要
为了解决无法直接得到小型发电机组的物理声源及其灵敏度的问题,采用经验模态分解(EMD)和支持向量机(SVM)协同分析的方法,开展噪声信号盲源分离,提取独立分量(IC)并识别主要物理声源,以测点声压级、机组功率和独立分量的声压级作为样本数据建立支持向量机回归模型,推导灵敏度计算函数,计算主要物理声源的灵敏度,得到了小型发电机组主要物理声源对辐射噪声的影响率.研究结果表明:影响该小型发电机组辐射噪声的主要因素有功率、配气机构噪声和驱动平衡轴的齿轮噪声,应用EMD-SVM协同分析可得到物理声源灵敏度,对于噪声控制具有重要指导意义.
To solve the problem that the physical noise sources of small generator and its sensitivity can't be obtained directly,the empirical mode decomposition(EMD)and support vector machine(SVM)were combined to separate the blind sources of noise signals,obtain the independent components(IC)and identify the physical noise sources.The SVM regression model was established on the samples of sound pressure level of measuring point,generator power and sound pressure level of the IC.The sensitivity calculation function was derived,the sensitivity of main physical noise source was calculated,and the influence rate of main physical noise source on small generator radiated noise was obtained.The research concludes that the main factors affecting the radiated noise of the small generator are power,valve train noise and balancing shaft gear noise,and the sensitivity of physical noise source can be obtained by using EMD-SVM,which has important guiding significance for noise control.
出处
《天津大学学报(自然科学与工程技术版)》
EI
CSCD
北大核心
2017年第10期1077-1083,共7页
Journal of Tianjin University:Science and Technology
基金
国家科技支撑计划资助项目(2015BAF07B04)~~
关键词
小型发电机组
物理声源
灵敏度
经验模态分解
支持向量机
small generator
physical noise source
sensitivity
empirical mode decomposition(EMD)
support vector machine(SVM)