摘要
针对离散制造企业生产环境和工艺的复杂性容易引起产品标识由于磨损、污染等因素而不可识读的问题,提出一种基于制造历史数据的产品标识失效预测与补救方法。给出了基于直接标刻技术的产品制造历史信息模型,分析了标识失效的影响因素并对历史数据进行Z-score标准化,通过主成分分析方法优化提取的特征,建立了神经网络标识失效预测模型,结合神经网络预测结果和标识转移与继承方法进行了失效标识的恢复补救。实例结果表明,该方法能够较好地预测产品标识失效并进行补救。
Aiming at the problem that product identification was not read due to wear, pollution and other factors caused by production environment and process complexity of discrete manufacturing enterprise, a prediction and re- mediation method of failed product identification was presented based on manufacturing history data. The product manufacturing history data model based on Direct Part Marking（DPM）was given. The factors of failed product iden- tification were analyzed and the history data was standardized by Z-score, and the extracted feature was optimized through Principle Component Analysis （PCA） method. The neural network model for prediction failed product iden- tification was established, and product identification was remedied by using neural network prediction results with i- dentification transfer and inheritance method. The experimental results showed that the proposed method could bet- ter predict or remedy the failed product identification.
出处
《计算机集成制造系统》
EI
CSCD
北大核心
2015年第9期2494-2503,共10页
Computer Integrated Manufacturing Systems
基金
国家自然科学基金资助项目(51275419)
国防基础科研资助项目(A2720110011)~~
关键词
直接标刻技术
制造历史数据
预测与补救
主成分分析
神经网络
direct part marking
manufacturing history data
prediction and remediation principal component analy-sis
neural networks
