Abstract: This paper proposes a novel control network architecture—structured exchanged industrial Ethernet—through in-depth analysis of traditional industrial control networks. The characteristics of this system are analyzed, and key technologies for applying exchanged Ethernet control networks in industrial settings are explored. Keywords: Control network; Ethernet; Information network; Distributed intelligence [align=center][b]Research Of Exchanged Industrial Ethernet-Based Xi Yabin, Ma Yongguang, Lin Yongjun[/b] (Industry Process Simulation & Control Laboratory, North China Electric Power University, Baoding 071003, China)[/align] Keywords: control network; Ethernet; information network; Distributed Intelligence 0 Introduction The rapid development of information network technology is profoundly changing people's work and lifestyles, especially impacting the informatization and automation of enterprises. Enterprise informatization can effectively improve the quality and efficiency of production, operation, and management, thereby enhancing the enterprise's market competitiveness and sustainable development capabilities. In the field of enterprise informatization and automation, the combination of computer technology, control technology, and network and information technology has given rise to control network technology. The development of enterprise informatization urgently requires an open, unified, effective, and highly sustainable network communication platform for support. Ethernet has advantages such as high transmission speed, low power consumption, ease of installation, and good compatibility. Because it supports almost all popular network protocols, it is widely used in commercial systems. However, traditional Ethernet uses a bus topology and Carrier Sense Collision Detection (CSMA/CD) communication method. In situations with high real-time requirements, the transmission of important data can experience transmission delays, which is known as Ethernet's "uncertainty." Research shows that the transmission latency of commercial Ethernet in industrial applications is between 2-30ms, which is one of the important reasons why Ethernet has not been able to enter the process control field for a long time. This paper will conduct some research on this and other key technologies in industrial Ethernet. 1 Traditional Industrial Control Network Structure From the overall structure, traditional enterprise networks can be divided into three layers: management layer, monitoring layer, and field device layer. Its structural diagram is shown below: [align=center] Figure 1 Traditional Industrial Control Network[/align] The functions of each of the three layers, namely management layer, control layer, and device layer, are as follows: ● Management layer. Mainly office automation system, and also extracts relevant production data from the monitoring layer for making comprehensive management decisions. The management layer generally uses Ethernet for easy operation and can be connected to external networks. ● Monitoring layer. Obtains data from field devices, completes various control and operating parameter monitoring, alarm and trend analysis functions, and also includes the design and download of control configuration. The monitoring layer's functions are generally performed by the host computer. On one hand, it connects to the fieldbus via a network interface board in the expansion slot, coordinating data communication between network nodes; on the other hand, it connects the fieldbus network segment and the Ethernet segment through a dedicated fieldbus interface (converter). Its key technology is the interface between Ethernet and the underlying field device network, primarily responsible for converting fieldbus protocols to Ethernet protocols, ensuring correct interpretation and transmission of data packets. In addition to the above functions, the monitoring layer also provides a supporting environment for advanced control and remote operation optimization. ● Device Layer. Field devices are connected to the fieldbus network as network nodes. Following the fieldbus protocol standard, the devices adopt a functional block structure, completing various functions such as data acquisition, A/D conversion, digital filtering, temperature and pressure compensation, and PID control through configuration design. Furthermore, intelligent converters are used to digitally convert and compensate the current and voltage of traditional measuring instruments. 2. Differences between Control Networks and Information Networks and the Necessity of Integration Control network technology originates from computer network technology and shares many similarities with general information networks, but also has differences and unique aspects. Because industrial control systems place particular emphasis on reliability and real-time performance, data communication used in measurement and control differs from communication in general telecommunications networks and also from communication in general computer networks in information technology. Control network data communication aims to induce the movement of matter or energy. The main characteristics of data communication used for measurement and control are: allowing real-time event-driven communication, high data integrity, normal operation under electromagnetic interference and ground potential differences, and the use of dedicated communication networks. The specific differences between control networks and information networks are as follows: ① Timeliness of data transmission and real-time system response are the most basic requirements of control systems. Generally, the response time requirement for process control systems is 0.01–0.5 s, for manufacturing automation systems it is 0.5–2.0 s, and for information networks it is 2.0–6.0 s. In most applications of information networks, real-time performance is negligible. ② Control networks emphasize the integrity and reliability of data transmission in harsh environments. The control network should be capable of transmitting data continuously, reliably, and completely over long periods in industrial environments with high temperatures, humidity, vibration, corrosion, and especially electromagnetic interference, and should be able to withstand surges, drops, and spikes in industrial power grids. In flammable and explosive environments, the control network should also possess intrinsic safety performance. ③ In enterprise automation systems, because dispersed individual users need to access a system through the control network, communication methods often use broadcast or multicast: in information networks, one autonomous system typically establishes a one-to-one communication method with another. ④ The control network must address the compatibility, i.e., interoperability, of products and systems from multiple companies within the same network. Currently, enterprise networks generally consist of two parts: an information network that processes enterprise management and decision-making information, and a control network that processes real-time measurement and control information from the enterprise's field. The information network is generally located in the upper-middle layer of the enterprise, processing large volumes of changing and diverse information, characterized by high speed and comprehensiveness. The control network is mainly located in the lower-middle layer of the enterprise, processing real-time, field information, characterized by simple protocols, strong fault tolerance, security, reliability, and low cost. Therefore, the integration of control network and information network is essential and will create favorable conditions for enterprise computer integrated plant automation (CIPA) and informatization. This is mainly reflected in the following points: (1) Realize the information integration of control network and information network, establish a comprehensive real-time information database, and provide a basis for enterprise to optimize control, production scheduling, and planning decisions. (2) Establish distributed database management functions to ensure data consistency, integrity, and operability. (3) Realize remote monitoring, optimized scheduling, and remote diagnosis of control network operations. (4) Realize remote software maintenance and updates of control network. 3 Switched Industrial Ethernet The adoption of Ethernet-based distributed control system is precisely to better realize the integration of control network and information network and overcome the shortcomings of previous control systems. Compared with fieldbus, Ethernet's main advantage is that it can provide an open infrastructure, enabling enterprises to achieve comprehensive and seamless information integration from the field control layer to the management layer; and solving the problem of "automation islands" caused by the gap in network protocols. To meet the requirements of industrial control systems and drawing upon mature design methods for commercial networks, this paper designs a structured switched industrial Ethernet system. The entire system is based on an integrated Ethernet-TCP/IP design. Its system architecture diagram is shown in Figure 2. [align=center]Figure 2: Distributed Ethernet Architecture Diagram[/align] The entire network is centered around an Ethernet switch and includes a database server and a file server. The Ethernet switch has broadband ports of 10Mbps, 25Mbps, and 100Mbps. Typically, a 100Mbps network switch port is used to connect to the server to meet the network bandwidth requirements of industrial PCs, PLCs, embedded controllers, workstations, etc., when frequently accessing the server. The monitoring unit performs functions such as monitoring the entire control process and setting parameters. If multimedia functionality is required for the monitoring workstation, a 25Mbps port can be used. Field devices can be general industrial control computer systems, fieldbus control networks, PLCs, embedded control systems, etc. General industrial control computer systems are connected to network switches or switching hubs via Ethernet cards; fieldbus control networks are interconnected with Ethernet via data gateways; there are two ways to access PLCs: PLCs with Ethernet cards can be connected to switches via Ethernet cards, while ordinary PLCs without Ethernet cards need to be connected to network switches via 485/232 conversion and industrial control computers; embedded control systems can be connected to network switches via the Ethernet cards built into the embedded controllers. When the network scale is large, a segmented structure can be adopted to form a larger network, with each switch and control device constituting a relatively independent control subnet. Several control subnets are interconnected to form a large-scale control network. This control network system has the following characteristics: (1) Distributed peer-to-peer functional unit control system is divided into two layers according to function: management layer and control layer. The management layer is not significantly different from the management layer of traditional industrial networks. The control layer merges the original monitoring layer and equipment layer, and replaces them with several "units" divided according to function and department. The relationship between each unit is equal, and the unit is a relatively independent real-time control area. The control function is decentralized to the field intelligent instruments and equipment of each field control unit, achieving thorough decentralized control and improving the system's flexibility, autonomy and security reliability. (2) Layered Network Structure The entire network adopts a structured design, divided into two layers: the core layer and the distribution layer. Core Layer: This is the data exchange center. Communication between devices in the distribution layer is achieved through high-speed data exchange at this layer. The core switching device is a Layer 3 high-speed switch; it creates and maintains network routing tables to realize routing between different functional units or virtual LAN subnets; and it implements access control mechanisms. Distribution Layer: Layer 2 switches are used to realize data exchange within each functional unit. The switches are connected to field devices or downstream switches in a point-to-point/full-duplex manner, so that the main devices within the unit can enjoy dedicated bandwidth, thereby ensuring the real-time performance of system communication. Compared with flat networks, the layered network structure has significant advantages: (a) It can effectively divide the network into smaller broadcast and collision domains, providing greater bandwidth for terminal devices. (b) The layered structure improves the scalability of the network, and the performance of the network will not be reduced when new functional units are added. (c) It facilitates troubleshooting and limits the impact of faults. (d) It enables functional units to become peers, forming a distributed communication system. (3) Higher openness Based on Ethernet-TCP/IP industrial Ethernet, the management layer and control layer use the same communication protocol. In essence, the management unit can communicate freely with the field control unit and monitoring unit, so that the upper and lower layers can achieve “seamless” integration; if necessary, the decision-making unit of the management layer can directly obtain the data of the field control unit, which is conducive to improving the speed of integrated automation decision-making; at the same time, the management layer and control layer establish necessary access control mechanisms to protect the security of the field control unit and sensitive departments and the real-time communication. Due to the adoption of widely used open standards (protocols), as long as the equipment of various manufacturers adopts the same protocol or through intelligent converters, they can be easily integrated into the same system, improving interconnectivity and interoperability, thereby eliminating the “information islands” between different automation systems. (4) High transmission rate Generally speaking, the backbone link uses a transmission rate of 100Mbps, and the equipment and terminal access link uses a transmission rate of 10/100Mbps; in addition, the hierarchical network structure can limit the communication data of functional units to local as much as possible, saving backbone link bandwidth, reducing broadcast between units, and improving terminal bandwidth. 4 Key technologies for the application of switched Ethernet control network in industrial field ◆ Ensure real-time communication. For a long time, the uncertainty of Ethernet communication response has been one of the fatal weaknesses and major obstacles to its application in industrial field devices. The main reason for this uncertainty is the method of Ethernet access transmission media, which is prone to the "capture effect." Switched Ethernet control network technology not only greatly improves channel utilization, but more importantly, it can shield the latency uncertainty of traditional shared Ethernet, removing a major obstacle to its application in industrial control. However, as the scale of switched Ethernet control networks expands, loop and broadcast storm problems arise. ◆Solutions to loops and broadcast storms. The network structure of industrial control systems is becoming increasingly complex, making it difficult to guarantee that its topology does not contain closed loops. However, dynamic spanning tree algorithms can be used to calculate the routing of switches, thereby avoiding the impact of loops on data frame forwarding routes, while increasing the robustness and ease of networking of switched Ethernet control networks. ◆Support for network security. Currently, switched Ethernet control networks have integrated the traditional three-layer network system (i.e., information management layer, process monitoring layer, and field device layer), introducing a series of network security issues. In addition to introducing firewall mechanisms, the network body provides two aspects of security support: network device access control and data access control. ◆ Bus Power Supply. Bus power supply, or bus-fed power, refers to cables connecting to field devices that not only transmit data signals but also provide power to the devices. Using bus power supply reduces network cabling, lowers installation complexity and cost, and improves network and system maintainability. It is particularly important in harsh and hazardous environments. Since Ethernet was previously mainly used for commercial computer communication, general devices or workstations (such as computers) already have their own power supply and do not require bus power supply; therefore, the transmission medium is only used for information transmission. ◆ Interoperability. Interoperability refers to the ability of devices from different manufacturers connected to the same network to communicate and interoperate using the same application layer protocol. Devices with similar performance can be interchanged. As a characteristic of open systems, interoperability guarantees users that devices from different manufacturers can communicate with each other and work together in multi-vendor product integration environments. This improves system quality and provides users with greater market choices. Interoperability is crucial for the acceptance and widespread application of a communication technology by automation equipment manufacturers and users. ◆ Network Survivability. Network survivability refers to the strong network availability that Ethernet must possess when applied to industrial field control. The failure of any system component, regardless of whether it is hardware or not, will paralyze the operating system, network, controller, and applications, and even the entire system, indicating that the system's network survivability is weak. Therefore, to maximize network uptime, reliable technologies are needed to ensure uninterrupted system operation during network maintenance and improvements. ◆Intrinsic Safety and Explosion-Proof Technology. In production processes, many industrial sites inevitably contain flammable, explosive, and toxic environments. For intelligent equipment and communication devices used in these industrial sites, certain explosion-proof technologies must be implemented to ensure the safety of the industrial site. Explosion-proof technologies for field equipment include two categories: flameproof (such as increased safety, airtight, potting, etc.) and intrinsically safe. Compared with flameproof technologies, intrinsically safe technologies use the suppression of ignition source energy as an explosion-proof means, which can bring the following technical and economic advantages: simple structure, small size, light weight, and low cost; maintenance and replacement can be performed while the circuit is energized; high safety and reliability; and wide applicability. The key technologies for achieving intrinsic safety are low-power technology and intrinsically safe explosion-proof technology. The inherent security of Ethernet systems includes industrial field Ethernet switches, transmission media, and field devices such as Ethernet-based transmitters and actuators. Because current Ethernet transceivers generally consume a relatively large amount of power (60-70mA, 5V operating power), designing low-power Ethernet-based field devices and switches is relatively difficult. ◆Long-distance transmission. In industrial settings, production facilities are typically complex, and various measuring and control instruments are spatially dispersed, often at distances of several kilometers. In such cases, designing Ethernet using 10Base-T twisted-pair cabling is insufficient, while 10Base-2 or 10Base-5 coaxial cables cannot support full-duplex communication and are also costly. Similarly, while using fiber optic transmission media in the field might be costly, the cost of fiber optics will undoubtedly decrease significantly with the widespread adoption of the Internet and Ethernet technologies. 5 Conclusion The development of Ethernet and Internet technologies will completely transform the network architecture of traditional enterprises. Undoubtedly, Ethernet plays a crucial role in industrial automation networks. According to the latest research results of the three major automobile manufacturers, Ethernet can be applied to at least 70% of field networks, which means that industrial Ethernet has a very broad prospect. Control technology has evolved from centralized control system to distributed control system. It is now moving towards the third generation of control system - distributed intelligence. Due to the dual advantages of high bandwidth and delay determinism of switched Ethernet control network, it provides a communication network platform for realizing distributed intelligence, triggering a revolution in automation architecture. This shows that distributed intelligence is the inevitable future of automation, and switched Ethernet control network technology has made great contributions. References [1] Zhang Wenchao, Xu Baoguo, Industrial Computer Network. [J] Industrial Control Computer, 1991, 4: 23-26 [2] Xie Lingguang, Wu Naiyou, Huang Songjie. Ethernet technology for manufacturing automation [J]. Measurement and Control Technology, 2001. (6): 10 [3] Yang Changhun, Industrial Ethernet is entering the control field. [J] World Instrumentation and Automation, 2001, 1 [4] Walsh GC, Ye Hong. Scheduling of Networked Control System [J]. IEEE Control System Magazine, 2001, 2l(1): 57-65 [5] Austin. Fieldbus Foundation Releases High Speed ​​Ethernet (HSE) Final Specifications． Control Engineering． March 2000