Abstract
According to the production site, a three-level network monitoring system, which consists of a host computer, master, and slave nodes, is built to transmit data from 10 automatic assembly lines. This system achieves field data collection, site control, and on-site operation simulation. The monitoring software is developed based on configuration software and an Access database to realize functions such as real-time operation monitoring, alarm recording, maintenance, and production statistics. Historical production and fault information can be queried via the database. Finally, using Web publishing technology, production site data is published to the Internet to achieve remote monitoring.
The Challenge: Fragmented Production Management
As the demand for $\text{LiMnO}_2$ button cells—the ubiquitous "coin" batteries found in small electronics—has grown, manufacturers have deployed dozens of automated assembly lines. However, managing these lines in isolation is inefficient. Without a centralized system, it is difficult to track equipment efficiency, identify the root causes of downtime across multiple machines, or monitor production status from a remote location. The challenge lies in bridging the gap between the "shop floor" (physical assembly hardware) and the "top floor" (management and remote monitoring) using a robust industrial network.
System Architecture and Methods
The researchers implemented a hierarchical, three-layer monitoring architecture designed for high reliability and low latency:
- Management Level (Host PC): A central computer runs custom monitoring software developed using industrial configuration tools and a Microsoft Access database. This layer handles data storage, historical queries, and Web publishing.
- Master Level (Master PLC): One OMRON CJ-series PLC is designated as the "Master." It communicates with the Host PC via an Ethernet module, serving as the gateway for the entire production floor.
- Field Level (Slave PLCs): The remaining nine assembly lines are controlled by "Slave" PLCs. These are networked to the Master PLC using Controller Link technology—a high-speed industrial protocol specifically designed for PLC-to-PLC communication.
Key Insights: Modular Logic and Hybrid Networking
- Modular Assembly Logic: The paper breaks down the complex battery assembly into distinct modules: lithium-chip cutting, pressure-forming, paper cutting, electrolyte injection, manganese-chip addition, and final sealing. This modularity allows the monitoring system to pinpoint exactly which stage of the process is failing.
- Hybrid Networking: Rather than putting all 10 PLCs on a standard Ethernet network—which can be prone to collisions in an industrial setting—the authors used a Master/Slave hierarchy. The Controller Link handles the heavy lifting of machine-to-machine data, while Ethernet is reserved for high-level data transfer to the PC.
- Web Integration: By using Web publishing technology, the system moves beyond the local factory floor, allowing stakeholders to monitor production via the Internet, a precursor to modern Industrial IoT (IIoT) standards.
Key Results
The authors successfully deployed the system across 10 active lines. Key functional results include:
- Real-time Visualization: A graphical interface provides a "dashboard" view of all 10 lines simultaneously, using color-coded indicators for status.
- Fault Analysis: The system captures the exact time and location of shutdowns, significantly reducing the time required for maintenance and repair.
- Reporting: The system automatically generates yield summary reports and downtime curves, providing a data-driven basis for improving production efficiency.
Limitations and Future Work
While effective, the system is heavily reliant on OMRON-specific hardware and protocols (Controller Link). Expanding this to a factory with mixed-vendor hardware would require a more universal protocol like OPC UA. Additionally, while Microsoft Access is sufficient for 10 lines, a more robust SQL-based relational database would be necessary for larger-scale operations. Future work may focus on moving from reactive fault logging to proactive predictive maintenance.
Summary updated