Design of tobacco industry hierarchical early warning system based on big data

In view of the flaw found in the existing industrial control network system which cannot efficiently and stably receive and process the massive sensing data from the cigarette factory during the production process, a tobacco industrial data processing and storage system for sensor data of cigarette factory, based on the Hadoop framework, has thus been designed and implemented. This system is based on the existing Cyber Physical System (CPS) and uses Message Queuing Telemetry Transmission (MQTT) as the data access protocol, with Flink on Yarn-Cluster deployment mode as the flow computing platform, and with Kafka as the message queue buffer diversion. Flink is used to send different types of data to Kafka for temporary storage and processing. A grading pre-warning system for cigarette manufacturing department is designed to realize dynamic monitoring and real-time pre-warning of production equipment status. Finally, by combining production business with big data technology, a hierarchical early warning system for cigarette manufacturing department was designed to realize dynamic monitoring and real-time early warning of production equipment status.