基于激光与红外的复合式贯流风叶缺陷在线监测系统设计

发布时间：2018-04-18 17:42

  本文选题：贯流风叶 + 激光　； 参考：《江苏大学》2017年硕士论文

【摘要】：贯流风叶作为空调等风力输送设备的重要部件,其结构完整性对风力输送过程中的噪声水平控制和气流输送平稳性有直接影响。贯流风叶通常采用注塑工艺生产,由于模具磨损或注塑过程控制参数变化等原因,会造成贯流风叶叶片边角缺料缺陷,如果不能及时检测并剔除这些不合格产品,将导致产品不合格率上升,严重影响产品质量。目前,贯流风叶缺料缺陷主要是通过人工目视检测,由于生产批量大,生产线对产品检测速度要求高,人员在长时间目视检测会产生视觉疲劳,导致误判时有发生。另外,人工目视检测还存在易受人为主观因素影响的弊端,进一步导致误判率上升。为解决上述问题,本课题组设计了第一代贯流风叶注塑缺料缺陷在线监测系统,由于采用单一类型的红外传感器对缺料缺陷进行检测,还存在检测准确率不稳定和无法有效检出小缺料缺陷的不足。为此,本论文在第一代红外监测系统的基础上,改进了检测方案,采用激光与红外复合阵列对产品进行监测,并提取了新的信号特征,采用了特征参数融合算法和基于支持向量机的缺陷识别算法,并增加了产品自动放置识别功能,完善了人机交互功能。同时,为了提高检测系统的检测稳定性和抗干扰能力,改善了电路的设计。并且完成了第二代样机的制作和调试工作,通过对实际产品的检测,其性能指标优于第一代设备。其检测准确率、稳定性和对小缺陷(缺陷面积≤1mm~2)的检出率均得到了提高。在实验室条件下,第二代样机对缺料缺陷的检出率超过99.9%,其中对小缺陷的检出率超过99%,而第一代样机对大缺陷的检出率约为97%,且无法有效检出小缺陷。本文的主要研究内容如下:(1)在分析第一代监测系统性能特点的基础上,针对其存在的不足,提出了新的检测原理方案,采用了基于激光与红外复合式检测方法,并进行了第二代监测系统总体方案设计;(2)设计了激光与红外复合式检测电路,主要包括信号激励电路、接收电路、模拟信号处理电路、上位机通讯电路;并且设计了激光与红外复合阵列布设方案与机械结构,进一步完善了第一代样机的电路设计,提高了电路抗干扰能力和长时间检测稳定性;并增加了产品自动放置识别功能;(3)设计了基于C8051F020单片机的主控电路;(4)研究和分析了贯流风叶缺料缺陷的光学信号特征,提取了缺陷的特征参数;提出了缺陷特征参数融合方法,并研究了基于支持向量机的缺陷识别算法;(5)编写监测系统软件,主要包括初始化程序、产品放置识别程序、数据采集程序、特征参数提取与融合程序、基于支持向量机的缺陷识别程序、检测结果输出程序、以及人机交互程序等;(6)制作监测系统样机,进行调试和试验,验证监测系统的性能。
[Abstract]:As an important part of air conditioning and other wind transport equipment, the structural integrity of tubular blade has a direct impact on the noise level control and airflow smoothness in the process of wind transport.The tubular vane is usually produced by injection molding process. Because of the wear of mould or the change of injection process parameters, it will cause defects in the edge and corner of the blade. If these unqualified products can not be detected and eliminated in time,Will cause the product nonconformity rate to rise, seriously affects the product quality.At present, the defects of material shortage in the tubular blade are mainly detected by manual visual inspection. Because of the large production lot, the production line requires high testing speed of the product, and the personnel will produce visual fatigue in the long time visual inspection, which leads to the misjudgment occurring from time to time.In addition, artificial visual detection is vulnerable to human subjective factors, resulting in a further increase in the rate of miscarriage of justice.In order to solve the above problems, the first generation of on-line monitoring system for defects in injection molding of tubular vane was designed. Because of the use of a single type of infrared sensor to detect the defects of missing materials,There are also some shortcomings such as unstable detection accuracy and inability to effectively detect small defects.Therefore, based on the first generation infrared monitoring system, this paper improves the detection scheme, uses the laser and infrared composite array to monitor the product, and extracts the new signal features.The feature parameter fusion algorithm and the defect recognition algorithm based on support vector machine are adopted, and the function of automatic product placement and identification is added to improve the human-computer interaction function.At the same time, in order to improve the detection stability and anti-interference ability of the detection system, the circuit design is improved.The second generation prototype is made and debugged, and its performance index is better than the first generation equipment by testing the actual product.The detection accuracy, stability and detection rate of small defects (defect area 鈮，

本文编号：1769391