基于质量数据融合及规则挖掘的离散制造过程监控方法

发布时间：2018-08-06 20:15

【摘要】：为有效地从企业积累的制造过程质量数据中挖掘加工工序间的关联规则、实现制造过程的智能化监控,建立了基于质量数据融合和规则挖掘的离散制造过程监控模型。借鉴证据理论思想,提出一种基于模糊证据理论的多传感器质量数据融合算法,首先将所有测量值构成的集合视为辨识框架,应用模糊数学原理引入隶属度函数,设计了多传感器之间互支持度计算的新方法,并借鉴信任分配思想将测量值转换为相应的证据,最终通过基于冲突分配的证据组合规则得到融合结果,实现制造过程质量数据的精确采集。设计了基于质量规则挖掘的离散制造过程分析方法,针对传统关联规则挖掘结果繁杂且有不符合实际加工顺序规则生成的情况,借鉴有监督学习思想,以实际工序顺序为约束,挖掘导致产品质量问题的关键因素,结合复杂事件处理技术对制造过程进行实时监控。通过实验结果验证了该监控方法的有效性和实用性。
[Abstract]:A discrete manufacturing process monitoring model based on quality data fusion and rule mining is established in order to effectively mine association rules between manufacturing processes from the quality data accumulated by enterprises and to realize intelligent monitoring of manufacturing processes. Based on the theory of evidence, a multi-sensor quality data fusion algorithm based on fuzzy evidence theory is proposed. Firstly, the set of all the measured values is regarded as the identification framework, and the membership function is introduced by applying the fuzzy mathematics principle. A new method for calculating the mutual support between sensors is designed, and the measurement value is converted into the corresponding evidence by using the idea of trust assignment. Finally, the fusion result is obtained by the rule of evidence combination based on conflict assignment. Accurate acquisition of manufacturing process quality data is realized. A discrete manufacturing process analysis method based on quality rule mining is designed. In view of the complicated results of traditional association rules mining and the generation of rules that do not conform to the actual processing order, the supervised learning idea is used for reference and the actual process order is taken as the constraint. The key factors that lead to product quality problems are excavated, and real-time monitoring of manufacturing process is carried out with complex event processing technology. The experimental results show the effectiveness and practicability of the method.
【作者单位】：中国科学院成都计算机应用研究所;贵州大学现代制造技术教育部重点实验室;贵州大学计算机科学与技术学院;
【基金】：国家自然科学基金资助项目(51475097) 工信部智能制造示范项目(工信部联装[2016]213号) 贵州省重大基础研究资助项目(黔科合JZ字[2014]2001)~~
【分类号】：F273.2;TP311.13

【相似文献】