Abstract: The nature of the production capacity and the equipment efficiency in the product-oriented manufacturing workshop provides great insights about the study of reducing the costs and increasing the competitiveness,which draws the focus of the researchers at home and abroad.Consequently,facing with the characteristic of the equipment intensive production line cluster in a general way,the Overall Equipment Effectiveness (OEE)accurate online monitoring was taken as the goal,the detailed feature factors and the characteristic spaces on the Surface Mounted Technology (SMT) production lines OEE calculation was critical reviewed and analyzed.Additionally,the shutdown type pattern recognition method based on the Synthetic Minority Oversampling Technique-Multilayer Perceptron (SMOTE-MLP) classification algorithm was constructed,which abandoned the traditional offline statistical calculation and provided a possibility for the online and real-time calculation of the availability rate,the performance rate,the quality rate and the OEE.Combined with the development trend of the intelligent manufacturing,arguing for the industrial big data-driven,equipment-intensive production line cluster OEE accurate online monitoring theory and method were proposed,which expected to contribute to the future research for the related work.

Key words: production line cluster, overall equipment effectiveness, accurate online monitoring, industrial big data

摘要： 随着国际竞争日益激烈,制造车间更多地关注产能与设备效率,以期通过降低停机时间、减少无效浪费,实现产能提升,从而摊薄成本,增加企业竞争力。有鉴于此,从工业大数据角度出发,以设备密集型产线集群设备综合效率(OEE)精准在线监测为目标,从时间维度和空间维度分析,以前后两个治具加工间隔、治具片数、检测时间、治具引脚数等为特征空间,构建基于合成少数类过采样技术和多层感知器神经网络分类模式的停机类型模式识别方法(SMOTE-MLP),实现不平衡数据处理与停机类型的精准划分,摒弃传统线下计算的方式,为时间稼动率、性能稼动率、良率以及OEE的在线、实时计算提供了可能,形成一套以工业大数据为驱动的、设备密集型产线集群OEE精准在线监测的理论和方法,为智能车间分析与优化提供一套可行方案。

关键词: 产线集群, 设备综合效率, 精准在线监测, 工业大数据