基于磨损识别的齿轮故障诊断研究

发布时间：2018-07-09 16:27

本文选题：齿轮故障 + 数字图像处理　；参考：《武汉工程大学》2012年硕士论文

【摘要】：齿轮是机械设备中最常用的一种传动零件，在动力的传递过程中起着重要的作用，其正常的工作状态和可靠性是确保传动系统效率的有力保障。目前，对齿轮进行故障诊断的常用技术包括传统诊断技术和智能诊断技术，如振动监测技术、噪声监测技术、油液分析、红外热成像、无损探伤、声发射技术和神经网络、模糊逻辑、专家系统、遗传算法等，它们在齿轮运行状态监测和故障诊断中均有着重要的应用。随着图像处理技术的不断发展，在机械工业中的应用也越来越广泛。为此，本文提出了一种基于数字图像处理技术的故障诊断研究思路，将其应用到齿轮磨损故障诊断中进行了研究，为齿轮磨损检测提供了一种全新的手段，实现了通过磨损图像识别的方法进行齿轮磨损故障的判别。本文的研究工作以建立可供实际应用的齿轮磨损故障判别系统为目标，分为五大部分：首先对齿轮图像进行了预处理研究，通过对图像中存在的各种失真和噪声进行了处理，获得了一幅干净且目标特征清晰的图像；其次是对齿轮图像的边缘检测和提取算法进行了研究，通过比较分析选择最优算子，提取出了能够真实反映实际轮齿边缘的齿廓曲线；再对图像进行离散化处理，将通过解析式获得的标准齿轮齿廓与通过图像识别获得的实际齿廓进行了对比研究，得出了齿轮磨损量；然后为了获得齿轮磨损量的阈值，进行了齿轮磨损实验；最后对故障诊断系统进行了开发，利用相关软件建立了一套可视化的齿轮磨损故障诊断系统。本文研究内容涉及到多种数字图像处理技术和系统开发技术，研究思路与方法比较新颖，目前国内进行的相关研究较少，因此具有一定的理论意义和创新价值。该方法丰富了齿轮故障诊断技术，，为齿轮零件的磨损故障诊断提供了一个新的方向，具有极大的现实意义。
[Abstract]:Gear is one of the most commonly used transmission parts in mechanical equipment, which plays an important role in the transmission of power. Its normal working state and reliability are the effective guarantee to ensure the efficiency of transmission system. At present, the commonly used technologies for gear fault diagnosis include traditional diagnosis technology and intelligent diagnosis technology, such as vibration monitoring technology, noise monitoring technology, oil analysis, infrared thermal imaging, non-destructive flaw detection, acoustic emission technology and neural network. Fuzzy logic, expert system, genetic algorithm and so on, they have important applications in gear running state monitoring and fault diagnosis. With the development of image processing technology, it is widely used in machinery industry. Therefore, this paper puts forward a research idea of fault diagnosis based on digital image processing technology, and applies it to gear wear fault diagnosis, which provides a new method for gear wear detection. The fault identification of gear wear is realized by the method of wear image recognition. The aim of this paper is to set up a gear wear fault discrimination system for practical application, which is divided into five parts: firstly, the preprocessing of gear image is carried out, and the distortion and noise in the image are processed. A clean image with clear target features is obtained. Secondly, the edge detection and extraction algorithm of gear image is studied. By comparing and selecting the optimal operator, the tooth profile curve which can truly reflect the actual tooth edge is extracted. Then the image is discretized, the standard gear tooth profile obtained by analytic formula is compared with the actual gear profile obtained by image recognition, the gear wear quantity is obtained, and then in order to obtain the gear wear threshold, Finally, the fault diagnosis system is developed, and a set of visual gear wear fault diagnosis system is established by using related software. The research content of this paper involves a variety of digital image processing technology and system development technology, the research ideas and methods are relatively new, at present, there are few related studies in China, so it has certain theoretical significance and innovative value. This method enriches gear fault diagnosis technology and provides a new direction for gear wear fault diagnosis, which has great practical significance.
【学位授予单位】：武汉工程大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TH132.41;TP391.41

【参考文献】