SPC在变速器装配线质量管理系统中的应用研究

CORC > 沈阳自动化研究所 > 中国科学院沈阳自动化研究所 > 智能检测与装备研究室

题名	SPC在变速器装配线质量管理系统中的应用研究
作者	肖忠保
学位类别	硕士
答辩日期	2016-05-25
授予单位	中国科学院沈阳自动化研究所
导师	陈书宏
关键词	统计过程控制支持向量机控制图模式质量管理系统
其他题名	Application Research of SPC in TheQuality Management System of Transmission Assembly Line
学位专业	控制工程
中文摘要	SPC(Statistical Process Control ，统计过程控制)控制图作为广泛应用在企业中的质量管理工具，对控制和改进企业质量具有重要的作用。本课题以沈阳华晨动力机械有限公司的某系列变速器装配线项目背景，根据课题组和企业签订的技术协议要求和更好的完成装配任务以及提高整条线体的自动化水平，开发了这套基于SPC的变速器装配质量管理系统。其中的管理系统除了包括常见的下生产计划、生产数据采集及监控和统计报表以外，在质量控制方面把SPC控制图应用到系统中加强了对质量的监控。在系统中不仅实现了国标规定的八条判异规则，而且针对生产中常见的控制图模式现象，本文使用SVM(support vector machine，支持向量机)对其进行识别。最后，通过Matlab环境下的仿真实验，发现改进的支持向量机算法取得了很高的识别精度。本文首先对华晨项目的变速器装配线情况做了简单介绍，包括装配线的各条线体，并且通过对装配线的工艺分析概括出了变速器装配的特点。结合装配特点对装配质量控制进行了分析，选取了SPC控制图作为质量控制手段。然后研究了SPC的基本理论和其在工业环境下的应用流程。接着讲述了关于支持向量机这种分类算法的原理，选取了基于PSO(particle swarm optimization，粒子群优化算法)的SVM对常见的6种控制图模式进行识别。本文在传统的粒子群优化SVM参数的基础上，给控制图的8个形状特征和4个统计特征赋予一个适应值，并把每一个适应值也都作为粒子搜索空间的一维，根据优化以后适应值的大小决定其对应的特征是否作为SVM的输入特征。然后在Matlab环境下做了网格搜索算法、标准PSO算法以及改进后的PSO算法的联合仿真实验，通过仿真发现改进后的PSO-SVM算法不仅在识别精度上较高而且收敛速度较快，为这种算法的在线实时识别应用提供了一个很好的支持。最后在对华晨项目的整体网络架构及设备分析以后，设计和开发了SPC质量管理系统并且对相关的生产管理系统功能做了详尽介绍。该系统成功完成对设备数据的实时分析，监控设备的状态，有效预防由设备不受控带来的装配问题，保证装配质量，同时减少企业成本。
英文摘要	SPC (Statistical Process Control)control chart is widely used as the tool of enterprise quality management, which is important for enterprise to improve and control the quality. This article is based on ShenYang HuaChen Power Machinery Co., Ltd’s transmission assembly line project. According to the technical protocol between research group and the enterprise, completing the assembly work better and improving the automation level of the whole line, we developed this quality management system based on SPC. This system included the common function like making production plan, data collection and monitor and statistical form besides in the quality control we usedthe SPC control chart in the system to reinforce the quality monitor. In the system, it achieved the national standard specified eight abnormal rules. Aimed at recognizing the control chart patterns, we used a improved support vector machine(SVM)using a particle swarm optimization(PSO) algorithm to optimize the parameter of the SVM kernel function. Then, we did our simulation in the Matlab environment, simulation results showed the propose algorithm achieves a high recognition accuracy. First, we introduced the transmission assembly line of HuaChen project including the three lines. We concluded the main feature of the transmission assembly line by the work flow analysis. So we decided to take advantage of the SPC to control our assembly quality. Then, we researched the basic theory of SPC and the workflow of application in industrial environment. In the second part we did research on the classification principle of SVM and chose the PSO as the SVM parameter optimization, and used it to recognize the common six control chart patterns. Here the PSO not only took the kernel function parameters as variables but also each adaptive value of the eight shape features and four statistical features ,each adaptive value decides if the feature is the input of the SVM. What’s more, we compared the grid search method and standard PSO method and improved PSO method in the Matlab simulation environment, the simulation results showed the improved algorithm achieves a high recognition accuracy and less iterations and support the goal of online using in factory. Finally we developed an SPC quality management system and introduced the function of the production management system which we would introduce in detail based on the analyzing the whole network architecture and device. The SPC system completed the real time data analyzing and monitoring the device state successfully. As a result, we could control the device in the normal state and ensure the quality of assembly at the same time we decrease the enterprise cost.
语种	中文
产权排序	1
页码	67页
内容类型	学位论文
源URL	[http://ir.sia.cn/handle/173321/19661]
专题	沈阳自动化研究所_智能检测与装备研究室
推荐引用方式 GB/T 7714	肖忠保. SPC在变速器装配线质量管理系统中的应用研究[D]. 中国科学院沈阳自动化研究所. 2016.