基于大数据的卷烟工业网络诊断探索  

Exploration on Network Diagnosis of Cigarette Industry Based on Big Data

作　　者：王德吉[1] 汪翠兰 张旭[1] 范九歌 房广顺 

机构地区：[1]中国烟草总公司职工进修学院 [2]河南轻工职业学院 [3]许昌卷烟厂

出　　处：《自动化博览》2022年第1期50-54,共5页Automation Panorama1

摘　　要：本文针对卷烟工业网络中出现的故障现象,采用大数据分析手段,基于Spark streaming框架的实时检测,通过消息订阅读取上层Kafka组件中的实时数据流,并加载离线构建的检测模型,对经过预处理后的规范数据流进行实时检测,具有高效的计算能力和并发处理能力。通过某卷烟厂实验证明了方法的有效性。In view of the fault phenomenon in the cigarette industrial network,this paper adopts big data analysis methods,which consists of the real-time detection based on Spark streaming framework,reads the real-time data stream in the upper-layer Kafka component through message subscription,and loads the detection model constructed offline.The preprocessed normative data stream is detected in real time.The method has effcient computing power and concurrent processing capability and its effectiveness is proved by experiments in a cigarette factory.

关 键 词：卷烟工厂 工业网络 故障诊断 大数据 实时检测 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论] TP393.08[自动化与计算机技术—计算机科学与技术] TS48[农业科学—烟草工业]