摘要
为实现复杂的板带生产质量相关故障精准识别与诊断,以板带质量相关故障为研究对象。首先,针对贝叶斯网络中节点过多导致网络参数过多和计算速度慢的问题,提出了一种基于协方差和蚁群优化BN结构的质量相关故障诊断模型。通过蚂蚁优化算法对贝叶斯网络进行优化,简化推理过程。其次,将节点间的互信息作为概率推理的重要因素,并将其集成到贝叶斯推理过程中。最后,利用热轧板带生产实测数据展开验证。结果表明,该模型能有效识别质量相关故障,模型的准确率达80%以上,满足实际生产需求,验证了优化诊断模型的有效性。
To realize accurate identification and diagnosis of production quality-related faults of complex strip,the strip quality-related faults were taken as the research objects.Firstly,aiming at the problem of excessive network parameters and slow calculation speed caused by too many nodes in bayesian network,a quality-related fault diagnosis model based on covariance and ant colony(Cov-ACO)optimization of BN structure was proposed.The bayesian network was optimized by ant optimization algorithm to simplify the reasoning process.Secondly,the mutual information between nodes was taken as an important factor of probabilistic reasoning,and it was integrated into the bayesian reasoning process.Finally,the verification was carried out by the measured data of hot rolling strip production.The results show that the model can effectively identify quality-related faults,and the accuracy of the model is more than 80%,which meets the needs of actual production,and verifies the effectiveness of the optimized diagnosis model.
作者
郭贺松
孙建亮
杨振
彭艳
GUO He-song;SUN Jian-liang;YANG Zhen;PENG Yan(School of Mechanical Engineering,Yanshan University,Qinhuangdao 066004,China;National Engineering Research Center for Equipment and Technology of Cold Rolled Strip,Yanshan University,Qinhuangdao 066004,China)
出处
《塑性工程学报》
CAS
CSCD
北大核心
2022年第10期126-134,共9页
Journal of Plasticity Engineering
基金
河北省自然科学基金资助项目(E2020203029)
河北省创新研究群体项目(E2021203011)
中央引导地方科技发展资金资助项目(206Z1601G
216Z1802G)
教育部产学合作协同育人项目(202101004003)
河北省教育厅在读研究生创新能力培养资助项目(CXZZBS2021129)。
关键词
热轧
贝叶斯网络
质量相关故障
蚁群优化
质量诊断
hot rolling
bayesian network
quality-related fault
ant optimization
quality diagnosis
