节点文献
天然气管道缺陷检测系统设计研究
Research and Design on Defects Detection System for Natural Gas Pipeline
【作者】 张伟;
【导师】 刘胜;
【作者基本信息】 哈尔滨工程大学 , 控制理论与控制工程, 2013, 硕士
【摘要】 这些年来,由天然气管道泄漏引起的爆炸事故逐年增加,这是由于管道使用时间长、维护保养不到位、管道内壁缺陷发现不及时等原因引起的,因此需要加大对天然气管道内壁缺陷的检测研究。随着超声波检测理论的成熟、数字信号处理技术以及自动化与智能化技术等学科和技术的迅猛发展,利用超声波检测技术进行管道检测有着检测速度快、效率高、结果准确、没有环境污染和成本小等优势,因此应用范围很广泛。本文以天然气管道为研究对象,利用脉冲反射法检测管道缺陷,以DSP芯片为核心的数字信号处理技术应用到超声波检测系统中,并通过SD卡存储数据。本文是针对天然气管道内壁进行缺陷识别,首先对机器人本体结构部分进行了设计,包括驱动调速结构、超声波探头的排列方式以及安全性的设计;然后通过对超声波检测理论与技术的分析,提出基于数字信号处理器(DSP)的超声检测系统的总体设计方案;系统的硬件设计主要包括超声波发射电路以及接收电路中的信号限幅保护、放大、滤波设计,DSP应用系统设计包括电源设计、时钟和复位电路、电池的选择以及存储器SD卡接口电路设计。系统的软件设计包含有DSP系统初始化、A/D信号转换的软件实现以及SD存储卡的在SPI模式下实现SD存储卡的底层读写,并建立文件系统实现对SD存储卡的文件操作。超声回波信号是一种非平稳时变信号,不仅包含很多有用信息,同时还夹杂着大量的噪声信号,最常使用小波阈值去噪法进行降噪,但不同的阈值函数对信号去噪能力是有非常大的差异,通过对软硬阈值函数的分析,构造一个新的阈值函数,它能够避免软阈值处理后某一些有用信息丢失的缺点,也能够避免硬阈值函数在阈值处产生附加振荡的缺点。通过对几种常用的母小波函数进行仿真实验,并结合超声回波信号的特点,确定合适的母小波函数,对实际的超声回波信号使用软硬阈值函数和改进的阈值函数进行去噪实验,通过实验结果验证了:改进的阈值函数能够更好地实现去噪,并且在低信噪比下也具有良好的去噪能力。
【Abstract】 Over the years, the explosion caused by the natural gas pipeline leak has increased yearby year, which is due to the following reasons:prolonged use of the pipeline, imperfectpipeline maintenance,delay pipe wall defects found. Therefore, it is meaningful to doresearch on gas pipeline wall defect detection.With the rapid development of the ultrasonic detection theory,the disciplines of digitalsignal processing technology and automated with intelligent technology, it is proved thatusing ultrasonic testing technology for pipeline detection has a fast speed, high efficiency andaccurate results, no environmental pollution and small cost. Therefore the application range isvery wide. Based on gas pipeline as the research object, this paper applies the pulse reflectionmethod to detect pipeline defects, uses DSP chip as the core of the digital signal processingtechnology applied to ultrasonic testing system and the SD card to store data.This article is aimed at the inner wall of the gas pipeline defect recognition. Firstly, thepaper completes the design of robot body structure, including the drive speed controlstructure, the arrangement of the ultrasonic probe as well as the safety of the design; Secondly,through the analysis of the theory and technology of ultrasonic detection, the overall designof the ultrasonic testing system is proposed based on a digital signal processor (DSP), Thehardware design of the system includes limiting the protection of the signal in the ultrasonictransmitting circuit and a receiving circuit, amplification, filter design; the design of DSPapplication system includes power supply design, clock and reset circuit, battery selectionand memory SD card interface circuit design. The software design of the system includes theinitialization of DSP system, the software implementations of A/D signal conversion, andthe SD memory card reads and writes in SPI mode, and creates a file system on the SDmemory card to manage document operation.Ultrasonic echo signal is a kind of time-varying signal, not only contains many usefulinformation, but also mingled with a lot of noise,. It is common to use wavelet thresholddenoising method to reduce noise, but denoising capacity of different threshold functionshave different effects. After analysis of hard and soft threshold function, a new thresholdfunction is constructed, which can avoid loss of useful information after treated by the softthreshold function, as well as avoid the additional oscillation after treated by the hard threshold function.A suitable mother wavelet function is determined by simulation of several commonmother wavelet function, as well as combined with the characteristics of the ultrasonic echosignal. In denoising experiments, the paper applies the actual use of soft and hard thresholdfunction and the improved threshold function to ultrasonic echo signal. Validated byexperimental results, improved threshold function has the very good denoising ability, andalso has a good denoising effect in a low signal-to-noise ratio.
【Key words】 Natural Gas Pipeline; Ultrasonic Test; SD Card; Signal Denoise; WaveletAnalysis; Threshold Value;