基于谐波失真的扬声器异音故障检测方法研究
发布时间：2019-01-05 15:52
【摘要】：在扬声器的生产过程中,异音故障检测主要还是通过人工检听的形式来完成,标准难以统一和量化,并且长时间的检听,对人耳造成很大的伤害。随着人们对电子产品的需求日益增长,扬声器的产量也与日俱增,扬声器的检听工作更加艰巨,这是电声行业亟待解决的问题,因此开发出一套由机器代替人耳的自动化检测异音故障的系统是企业迫切的需求。本文分析了扬声器异音故障产生的原因,以及响应信号的特性,研究了一种在线检测扬声器异音故障的方法。对扬声器响应信号进行短时傅里叶变换得到时频图。通过谐波失真的理论,在时频图上进行各阶谐波划分,对谐波能量进行统计。然后设计特征提取器,提取出各阶谐波的特征,找到合格扬声器与异音故障扬声器的区别。最后设计分类器,从而来对合格和故障扬声器进行分类,达到自动检测的目的。本课题的研究工作主要包括:1.分析和研究扬声器异音故障及其产生机理;2.针对扬声器响应信号进行时频分析,根据谐波失真机理,提取特征,分类学习,提出基于谐波失真的扬声器异音故障检测方法;3.对于不同种类的扬声器的阈值设定;4.对扬声器异音故障在线检测系统进行软硬件系统的集成。
[Abstract]:In the production process of loudspeaker, the abnormal sound fault detection is mainly completed by manual inspection and listening, the standard is difficult to be unified and quantified, and the long period of time detection and hearing, causing great harm to the human ear. With the increasing demand for electronic products, the output of loudspeakers is increasing, and the detection and hearing of loudspeakers is more difficult, which is an urgent problem to be solved in the electro-acoustic industry. Therefore, it is urgent for enterprises to develop an automatic system for detecting abnormal sounds by replacing human ears with machines. In this paper, the causes of loudspeaker abnormal sound fault and the characteristics of response signal are analyzed, and a method of on-line detecting loudspeaker abnormal sound fault is studied. The short-time Fourier transform is applied to the loudspeaker response signal to obtain the time-frequency diagram. Based on the theory of harmonic distortion, the harmonic energy of each order is divided on the time-frequency diagram and the harmonic energy is counted. Then a feature extractor is designed to extract the features of each order harmonic and find out the difference between the qualified loudspeaker and the abnormal loudspeaker. Finally, a classifier is designed to classify qualified and faulty loudspeakers to achieve the purpose of automatic detection. The research work of this topic mainly includes: 1. Analysis and study of loudspeaker abnormal sound fault and its generation mechanism; 2. Based on the time-frequency analysis of loudspeaker response signal, according to the mechanism of harmonic distortion, extracting features, classifying and learning, a method of loudspeaker abnormal sound fault detection based on harmonic distortion is proposed. 3. Set the threshold for different types of loudspeakers; 4. The software and hardware integration of the on-line detection system of loudspeaker abnormal sound fault is carried out.
【学位授予单位】：天津科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN643
本文编号：2401977
[Abstract]:In the production process of loudspeaker, the abnormal sound fault detection is mainly completed by manual inspection and listening, the standard is difficult to be unified and quantified, and the long period of time detection and hearing, causing great harm to the human ear. With the increasing demand for electronic products, the output of loudspeakers is increasing, and the detection and hearing of loudspeakers is more difficult, which is an urgent problem to be solved in the electro-acoustic industry. Therefore, it is urgent for enterprises to develop an automatic system for detecting abnormal sounds by replacing human ears with machines. In this paper, the causes of loudspeaker abnormal sound fault and the characteristics of response signal are analyzed, and a method of on-line detecting loudspeaker abnormal sound fault is studied. The short-time Fourier transform is applied to the loudspeaker response signal to obtain the time-frequency diagram. Based on the theory of harmonic distortion, the harmonic energy of each order is divided on the time-frequency diagram and the harmonic energy is counted. Then a feature extractor is designed to extract the features of each order harmonic and find out the difference between the qualified loudspeaker and the abnormal loudspeaker. Finally, a classifier is designed to classify qualified and faulty loudspeakers to achieve the purpose of automatic detection. The research work of this topic mainly includes: 1. Analysis and study of loudspeaker abnormal sound fault and its generation mechanism; 2. Based on the time-frequency analysis of loudspeaker response signal, according to the mechanism of harmonic distortion, extracting features, classifying and learning, a method of loudspeaker abnormal sound fault detection based on harmonic distortion is proposed. 3. Set the threshold for different types of loudspeakers; 4. The software and hardware integration of the on-line detection system of loudspeaker abnormal sound fault is carried out.
【学位授予单位】：天津科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN643
【参考文献】
相关期刊论文 前10条
1 祝仰宽;许增朴;穆瑞林;王永强;周聪玲;;YDT318-9AU扬声器异音故障检测方法的研究[J];电声技术;2016年07期
2 王以真;胡秉奇;;扬声器的失真特性[J];电声技术;2015年10期
3 李云红;李小英;周静雷;潘杨;;基于HHT算法的扬声器异常音检测技术[J];计算机工程与应用;2015年11期
4 李绩科;尚新春;金明昱;;微型扬声器谐波失真分析[J];电声技术;2014年11期
5 崔璨;袁英才;;窗函数在信号处理中的应用[J];北京印刷学院学报;2014年04期
6 韩捷;张坡;尤国雷;;国家标准《扬声器主要性能测试方法》解读[J];电声技术;2012年11期
7 桑松;柴玉华;孙影;;基于小波变换和快速傅里叶变换的谐波检测[J];电测与仪表;2012年07期
8 王以真;;世界的扬声器,扬声器的世界[J];家庭影院技术;2012年03期
9 韦峻峰;杨益;温周斌;冯海泓;;一种扬声器异常音的时域特征检测方法[J];振动与冲击;2011年10期
10 张东;吴晓琳;;短时傅立叶变换在振动信号处理中的应用[J];计算机与数字工程;2011年08期
相关硕士学位论文 前2条
1 王帆;扬声器异常声信号检测中的门限与类型判定方法的研究[D];西安工程大学;2013年
2 丁翔;基于虚拟仪器的手机扬声器测试系统的设计与实现[D];苏州大学;2006年
，本文编号：2401977
本文链接：https://www.wllwen.com/kejilunwen/dianzigongchenglunwen/2401977.html
