基于FBN的设备故障预测与健康评估仿真

Simulation Research on Equipment Fault Prediction and Health Evaluation Based on Fuzzy Bayesian Network

智能工厂设备故障分析与预测是精益生产的需求,有效地故障分析与预测对于合理地制定设备健康维护计划具有重要的指导意义。从智能工厂设备的历史故障数据出发,针对设备故障信息利用率较低,主要故障提取无代表性,各个故障对系统的交互作用体现不明显等问题,从分析某家电企业装配线单机设备的连接及工序入手,考虑连续生产工序上各单机设备故障之间的关系,提出连线生产组故障损失度作为相互关联设备故障分析指标。利用连线生产组设备故障发生率、故障损失度及产量数据、专家经验等信息,提出连线生产组设备的模糊贝叶斯网络故障预测与健康状态评估模型,进一步通过对某家电企业装配线有效数据的仿真,分析预测连线生产组设备故障,判断装配线的健康状态,达到了有效指导性维护目的。

Lean production requires failure analysis and prediction of intelligent factory equipment.Effective fail⁃ure analysis and prediction is of great significance for making reasonable equipment maintenance plan.Considering that there are many defects in the historical failure data of intelligent factory equipment,for example,the utilization rate for equipment failure information is low,the main fault extraction is not representative,and the interaction of va⁃rious failure on the system is unobvious.Taking the connection and process of single equipment in an assembly line of an electrical appliance enterprise as the research object,considering the relationship between the failures of single e⁃quipment under continuous production process,thus the failure loss degree of connecting production group was pro⁃posed as the failure analysis index of interrelation equipment.By utilizing the information of equipment failure occur⁃rence rate,fault loss degree,output data and expert experience,the fuzzy Bayesian network fault prediction and health status assessment model for the equipment in the connecting production group was proposed.Then,the equip⁃ment failure of the line production group was analyzed and predicted based on the simulation of the effective data of the assembly line of a household appliance enterprise,thus judging the health status of the assembly line.