汽车报交线缺陷预报警模型及控制方法研究

Research on the Alarm and Control Method of the Inspection Line in Automobile Factory

研究汽车报交线缺陷的预报警模型及基于大数据分析的最优控制方法。首先,在对业务流程重构以及基于在线看板实时信息化的基础上,建立了缺陷预报警的系统模型。其次,通过大数据分析方法以及蒙特卡洛模拟方法,以最小化漏报警概率和误报警概率为目标,比较分析休哈特控制图法和核密度估计法的优劣,并最终提出了最优的控制方法。最后,通过数值实验与实例验证,表明所提出的模型与方法的有效性与高效性。

The alarm models and control methods for the inspection production line of paint shop in automobile manufacturing is studied.The effective alarm model is built based on the reconstruction of business flow and the development of real time Kanban system.Through big data analysis and Monte Carlo method,the optimal control method is identified by comparing the methods of Shewhart control chart and Kernel density estimation methods in consideration of minimizing the alarm missing and error alarm.Finally,experiments and case studies prove the effectiveness of efficiency of proposed models and methods.