文档介绍：分类号:TB52+6 密级: 公开
UDC: 621 学校代码: 10356
中国计量学院
本科毕业设计(论文)
基于MEMS声传感器的烟道气流
检测与识别研究
The Study on Detecting and Recognizing the Air of Flue based on MEMS Sonic Transducer
作者邓平学号 0700103235
申请学位工学学士指导教师赵子恺副教授
学科专业电气工程及其自动化培养单位中评阅人
2011年06月
基于MEMS声传感器的烟道气流检测与识别研究
摘要:任何声源发出的声音中均包含与其结构和状态等有关的信息。基于信号处理的声音特征提取和模式识别技术近些年发展较快, 声信号具有精确性和稳定性等特点, 借助计算机技术运用信号处理的方法对物体发出的声音进行分析, 提取其振动特征量, 可以为物体运转情况的预测提供精准的数据分析依据。声音模式识别在无人职守、故障检测、灾害预防等方面具有很广阔的应用前景。
油烟管道内电机旋转也会产生与其对应的声音信号,主要包括通风声、机械声和电磁声,其中通风声占据主导作用。我们可以通过检测其主频并结合统计学来判别管道各支路电机的运转情况。本文对模拟烟道进行了多组实验和数据分析,找到了合适的声音信号特征向量,基本上能正确识别烟道油烟机运行状况。文章最后,对本文工作进行总结和展望,提出解决问题的初步方法,谨供以后的研究者参考。
关键词:MEMS;声音信号;模式识别;烟道;快速傅里叶变换
中图分类号:TB52+6
The Study on Detecting and Recognizing the Air of Flue
based on MEMS Sonic Transducer
Abstract:The sound of any sound source contains information about its structure and processing-based voice feature extraction and pattern recognition technology developed rapidly in recent years,Acoustic signal characteristics with a precision and stability,so with the use puter technology,analyze the sound objects in signal processing methods,then extracte the Eigenvectors of the vibration,that can be a accurate data analysis for forecasting about the object's condition of pattern recognition has broad application prospects in the absence of duty, fault detection, disaster prevention.
A motor rotation in smoke pipe also make corresponding sound signals,mainly including ventilation noise, mechanical noise and ic noise,in which the wentilation and sound making the occupy a leading can detect the frequency bined with statistics to determine the functioning of motors of each paper conducted a multi-flue experiments and data analysis,and take the appropriate sound signal eigenvector,which been able to correctly identify the health of the flue , the paper summarizes the work and take a outlook of the initial approach to solve problems, it would