基于MES的变速箱装配质量动态控制与预测研究

发布时间：2018-03-09 18:53

本文选题：变速箱装配　切入点：Kalman滤波　出处：《合肥工业大学》2017年硕士论文　论文类型：学位论文

【摘要】：在日趋激烈的市场竞争环境下,产品质量已成为提升企业竞争力的关键要素。变速箱作为一种典型的复杂机械产品,其整机性能由设计、制造和装配三个环节共同保证。其中,装配是变速箱生产过程中最为重要的环节,装配环节中各工序的装配结果对整机的质量有着极大的影响,因此,需要对变速箱装配过程质量进行控制,并建立实时、高效的装配质量管控平台。本文主要对变速箱装配质量的动态控制与预测进行研究。本文以变速箱装配过程为研究对象,在对变速箱装配工艺和装配特点分析的基础上,提出了变速箱装配质量控制系统的功能需求,构建了基于MES的变速箱装配质量动态控制与预测体系,并对关键技术进行了深入分析和研究。首先,针对变速装配质量控制问题,构建了小批量生产模式下的质量控制模型。针对短期微小偏移的问题,提出了基于Kalman滤波的EWMA控制图模型。对于长期较大偏移的监测,引用了基于相似性理论和公差带数据转换方法的质量特性数据标准化方法,改善了小批量生产质量特性数据不足的问题。其次,针对短期的装配质量预测问题,通过研究粒子群算法和BP神经网络,建立了基于PSO-BP神经网络的装配质量预测模型,实现了对变速箱装配质量的短期预测。最后,基于以上理论和方法,本文以某汽车股份公司双离合变速箱生产过程为应用背景,设计并开发了原型系统,并将该系统实际运用到变速箱的装配活动中,收获了较好的预期效果,为提高变速箱的装配质量提供了理论和技术基础。
[Abstract]:In the increasingly fierce market competition environment, product quality has become a key element to enhance the competitiveness of enterprises. As a typical complex mechanical product, the performance of the gearbox is guaranteed by three links: design, manufacture and assembly. The assembly is the most important link in the gearbox production process. The assembly result of each procedure in the assembly link has a great influence on the quality of the whole machine. Therefore, it is necessary to control the quality of the gearbox assembly process and establish the real time. This paper mainly studies the dynamic control and prediction of gearbox assembly quality. This paper takes the gearbox assembly process as the research object, based on the analysis of the gearbox assembly process and assembly characteristics. The functional requirement of gearbox assembly quality control system is put forward, the dynamic control and prediction system of gearbox assembly quality based on MES is constructed, and the key technologies are deeply analyzed and studied. The quality control model in small batch production mode is constructed. Aiming at the problem of short term small migration, a EWMA control chart model based on Kalman filter is proposed. The quality characteristic data standardization method based on similarity theory and tolerance band data conversion method is introduced to improve the problem of shortage of quality characteristic data in small batch production. Secondly, the short-term assembly quality prediction problem is discussed. By studying particle swarm optimization algorithm and BP neural network, the assembly quality prediction model based on PSO-BP neural network is established, and the short-term prediction of gearbox assembly quality is realized. In this paper, a prototype system is designed and developed based on the production process of double clutch gearbox in a certain automobile joint-stock company. The prototype system is applied to the assembly of the gearbox, and the expected results are obtained. It provides a theoretical and technical basis for improving the assembly quality of the gearbox.
【学位授予单位】：合肥工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TH132.46;TG95

【相似文献】