多品种小批量制造模式下的过程质量诊断技术研究

发布时间：2018-01-02 18:12

本文关键词：多品种小批量制造模式下的过程质量诊断技术研究　出处：《浙江工业大学》2011年硕士论文　论文类型：学位论文

【摘要】：以浙江省科技厅重大优先主题项目“面向服务架构的数字化设计与制造关键技术研究及其在离散制造企业中的应用”为依托,针对多品种小批量生产模式下数据样本少、质量诊断困难的问题,以优化该生产模式下质量诊断方法为目的,拟开展多品种小批量制造模式下的质量诊断技术研究。主要研究工作和成果如下： 1.控制图混合模式识别。针对多品种小批量生产模式下质量数据样本少的问题,同时考虑质量过程数据常会有多种异常现象混合的情况,提出了小波分析与SVM相结合的控制图混合模式识别方法,并将PSO算法引入到SVM中来提高控制图模式识别的精度,设计了三层控制图模式识别模型框架和基本流程。通过构造合理的仿真样本进行训练测试,验证了模型的有效性。 2.控制图模式参数估计。为给管理或技术人员提供质量过程调整的依据,在控制图模式识别的基础上,提出了基于PSO-SVM的控制图模式参数估计方法,设计了参数估计模型框架用于估计三种异常模式的四个参数,并采用仿真实例验证了模型的可行性。 3.质量异常原因诊断。对几种质量异因诊断方法进行了比较,通过借鉴专家系统的知识库和解释机制功能,构造相关数据库,设计了基于PSO-SVM的质量异因诊断模型,以及模型数据和用户可识别内容之间的转换规则。 4.实例的验证。在对SJ公司质量诊断控制现状分析的基础上,将质量控制图混合模式识别、参数估计、异常原因诊断模型应用于SJ公司的质量诊断控制中,证明了模型在实际应用中的可行性。
[Abstract]:It is based on the key technology research of digital design and manufacture of service-oriented architecture and its application in discrete manufacturing enterprises. Aiming at the problem of few data samples and difficult quality diagnosis in multi-variety and small-batch production mode, the aim of this paper is to optimize the quality diagnosis method in this production mode. It is planned to carry out the research on the quality diagnosis technology under the multi-variety and small-batch manufacturing mode. The main research work and results are as follows: 1. Mixed pattern recognition of control chart. Considering the problem of few samples of quality data in multi-variety and small-batch production mode, and considering that there are often a variety of abnormal phenomena mixing in the data of quality process. A hybrid pattern recognition method based on wavelet analysis and SVM is proposed, and the PSO algorithm is introduced into SVM to improve the accuracy of control chart pattern recognition. The model framework and basic flow of three-layer control chart pattern recognition are designed, and the validity of the model is verified by training and testing with reasonable simulation samples. 2. Control chart pattern parameter estimation. In order to provide management or technical personnel with the basis of quality process adjustment, on the basis of control chart pattern recognition. A control chart mode parameter estimation method based on PSO-SVM is proposed. A parameter estimation model framework is designed to estimate four parameters of three abnormal patterns. Simulation examples are used to verify the feasibility of the model. 3. Quality abnormal cause diagnosis. Several methods of quality heterogenetic diagnosis are compared, and the related database is constructed by using the functions of knowledge base and explanation mechanism of expert system for reference. A quality heterogeneity diagnosis model based on PSO-SVM is designed, and the conversion rules between model data and user-identifiable content are also presented. 4. Verification of examples. On the basis of analyzing the current situation of quality diagnosis and control in SJ Company, the mixed pattern recognition and parameter estimation of quality control chart are made. The abnormal cause diagnosis model is applied to the quality diagnosis control of SJ Company, which proves the feasibility of the model in practical application.
【学位授予单位】：浙江工业大学
【学位级别】：硕士
【学位授予年份】：2011
【分类号】：TH165.3

【相似文献】