SIA OpenIR  > 工业控制网络与系统研究室
基于噪声分析的传感器故障诊断技术研究与开发
其他题名Sensor Failure Diagnostics Research and Development Based on Noise Analysis
刘元锋1,2
导师刘明哲
分类号TP212
关键词故障诊断 快速傅里叶变换 温度传感器 压力传感器 功率谱估计
索取号TP212/L76/2015
页数60页
学位专业控制工程
学位名称硕士
2015-05-26
学位授予单位中国科学院沈阳自动化研究所
学位授予地点沈阳
作者部门工业控制网络与系统研究室
摘要我国目前的工业生产形势非常的严峻,在生产过程中发生任何事故都会对工厂或者社会带来严重的影响,威胁着经济的快速增长和社会进步的步伐,因此工业系统的运行安全性变得极为重要。现在自动化技术已成为确保安全生产的有效技术手段。安全仪表系统能够保证设备的安全运行,可有效避免危险事故的发生。仪表传感器作为安全控制系统的重要组成部分,它能够检测生产环境的关键性输入,在确保生产过程的安全可靠运行过程中发挥了非常重要的作用。 传统的仪表传感器故障诊断的方法多采用事后、离线、周期性的诊断方式。在两次周期性检查期间,传感器有可能就已经发生了故障,因此会带来严重的安全隐患。另外,在检查的过程中,工厂需要停止生产,这将会对工厂的经济效益产生严重影响。论文针对传统的传感器故障,采用一种求取系统响应时间特征值的基于快速傅里叶变换的噪声分析法,通过比较响应时间的大小来实时检测传感器是否发生故障。这对于提升生产效率和保障生产安全有重要意义。 首先,针对现在工业上使用率较高的热电阻、热电偶温度传感器,说明了其系统运行的响应时间,就热电阻温度传感器套管脱落故障,搭建了系统响应时间与套管脱落时的数学模型。针对电容差压式压力变送器传感器,阐述了其运行原理,根据其运行原理将电容式压力变送器等效成一个弹簧二阶系统,由此可以得到差压式压力变送器的系统模型。文中选取了在测压过程中常见的引压管阻塞故障,搭建了引压管阻塞与相应时间之间的数学模型。 然后研究实现了适用于仪表传感器故障数据的数据处理方法。本文将基于快速傅里叶变换的噪声分析故障诊断技术应用到嵌入式仪表传感器故障诊断当中去,详细介绍了应用基于快速傅里叶变换的噪声分析法诊断嵌入式传感器故障的实现步骤。对采集到的实验数据首先要进行提取噪声数据,对获取的噪声数据进行低通滤波器。滤波选取的巴特沃斯低通滤波器。针对基于快速傅里叶变换的求取功率谱估计的噪声分析法,比较了常用的功率谱估计法图周期法、平均周期法、welch法、多窗口法来求取功率谱密度的优缺点。最终选择了welch法用来求取功率谱密度曲线。并给出了welch详细的实现过程。 最后,分别就温度传感器的套管脱落故障、压力传感器的引压管阻塞故障搭建了物理测试平台。首选就温度传感器的套管脱落故障采集了故障模式和正常模式两种状态下的实验数据,对采集到的数据进行基于快速傅里叶变换的噪声分析处理,并借用canvas对数据处理结果实时的显示到web界面上。同样对于电容式压力传感器,选取了两根粗细不同的引压管,用细引压管模拟粗引压管阻塞的情况。在相同的压力源下,分别用粗细两根引压管采集实验数据,对采集到的实验数据进行基于快速傅里叶变换的噪声分析。并借用canvas将数据处理结果进行实时显示。结果表明将基于快速傅里叶变换的噪声分析法可以用来检测嵌入式温度变送器的套管脱落故障以及压力变送器传感器引压管阻塞故障。
其他摘要The current grim situation of our country's industrial production and the great economic losses of the production accident has caused the development of economy and the progress of the society seriously. As a result the security of the system becomes very important. Nowadays the automation technology has become an effective technical approach to ensure the safety in production. Safety instrument system can ensure the safe operation of equipment, which can effectively avoid dangerous accidents. As an important part of safety control system, instrument sensor can detect production key input. In order to ensure the safe and reliable operation of the production process, instrument sensor has become more and more important. The traditional diagnosis of sensor failure is based on the measures of after event diagnosis, offline diagnosis and periodic diagnosis. However, the sensor may be broken down between two consecutive diagnosis, this will lead to potential safety hazard. Furthermore, the low maintenance efficiency of the sensor can also lead to the waste of time and money. In this paper, a failure diagnosis method based on fast Fourier transform noise analysis is proposed, the noise signal in the sensor output signal is used to achieve fault detection, which is significance to improve production efficiency and ensure production safety. Firstly, in this paper, the noise analysis failure diagnosis technology based on fast Fourier transform is applied to embedded instrument sensor failure diagnosis. Besides,the realization of the noise analysis failure diagnosis technology based on fast Fourier transform to diagnose embedded sensor failure is described in detail. In terms of the high utilization rate of thermal resistance temperature sensor and thermocouple temperature sensor, the response time of the system operation is introduced. When the thermowell of the temperature sensor is missing, a mathematical model between the system response time and the missed thermowell is set up. For capacitive pressure sensor, the operation principle is described. According to the operation principle, the capacitive pressure sensor system can be to equivalent a second order spring system .So when the pressure tube is blocked, a mathematical relationship between the system response time and blocked tube is set. Secondly, the instrument sensor data processing method is studied and realized. First of all, extract the noise data out from collected experimental data, then, through a low-pass filter. Butterworth low-pass filter is selected. And then the advantages and disadvantages of the commonly used power spectrum estimation method to calculate of power spectral density is compared, the method such as figure period, Bartlett, Welch and multitaper method. At last, Welch method is selected. And detailed implementation process of Welch method is presented. At last, when the thermowell of the temperature sensor is missing and when the pressure tube is blocked, the physical test platform is established. When the thermowell of the temperature sensor is normal and when thermowell of the temperature sensor is missing, the experimental data is collected, respectively. The collected data is processed by noise analysis method based on fast Fourier transform. And the real-time data processing result is displayed to the web interface through the canvas at last. For capacitive pressure sensor, two different degree of pressure tube are selected and the thin tube is used to simulate rough tube obstruction. Under the same pressure source, we use the thin tube and the rough tube to select the experimental data separately. The collected data is processed by noise analysis method based on fast Fourier transform. And the real-time data processing result is displayed to the web interface through the canvas method at last. The results show that noise analysis method based on fast Fourier transform can be used to detect the missed thermowell fault of the embedded temperature transmitter and the blocked tube fault of pressure sensor.
语种中文
产权排序1
文献类型学位论文
条目标识符http://ir.sia.cn/handle/173321/16745
专题工业控制网络与系统研究室
作者单位1.中国科学院沈阳自动化研究所
2.中国科学院大学
推荐引用方式
GB/T 7714
刘元锋. 基于噪声分析的传感器故障诊断技术研究与开发[D]. 沈阳. 中国科学院沈阳自动化研究所,2015.
条目包含的文件
文件名称/大小 文献类型 版本类型 开放类型 使用许可
基于噪声分析的传感器故障诊断技术研究与开(3086KB) 开放获取CC BY请求全文
个性服务
推荐该条目
保存到收藏夹
查看访问统计
导出为Endnote文件
谷歌学术
谷歌学术中相似的文章
[刘元锋]的文章
百度学术
百度学术中相似的文章
[刘元锋]的文章
必应学术
必应学术中相似的文章
[刘元锋]的文章
相关权益政策
暂无数据
收藏/分享
所有评论 (0)
暂无评论
 

除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。