Research on embedded industrial Ethernet control scheme

I. Introduction

　　For the field of industrial automation, a large number of intelligent devices can be connected to the Internet through various channels, and information and data can be transmitted to each other through the network, so as to realize the functional autonomy of intelligent field equipment, the high decentralization of system structure and the integration of supervision and control.

　　Fieldbus is a new technology developed in response to this situation. The emergence of fieldbus marks the beginning of a new era in the field of industrial control technology. The development of this technology has played a huge role in promoting the realization of equipment-oriented automation systems. Compared with the traditional distributed control system (DCS), it has the advantages of full openness, full decentralization and interoperability, but it still has great limitations, which are mainly manifested in the following aspects:

　　1. The current computing power and information processing power of field instruments and equipment are relatively low, and the complex control functions are still concentrated on one control computer, which cannot realize full decentralized control and there is a phenomenon of risk concentration.

　　2. Fieldbus is only a component of the system and is located at the bottom of the system, which is not enough to realize the full open structure of the system. The system architecture is vertically combined, and there is a bottleneck in data communication.

　　3. The IEC61158 standard includes 8 types of fieldbuses, which are quite different from each other, cannot achieve interoperability, and have certain difficulties in connecting with each other.

　　4. All controllers in the system run independently, and each performs independent data processing, making it difficult
to share all information, resulting in unsatisfactory real-time performance of the system.

　　The above explanation shows that traditional classic PLC and fieldbus technologies are no longer suitable for this requirement. Even technologies such as industrial PC and OPC, as long as they are embedded in the traditional system structure, can only make marginal improvements to the system's functions.

　　Therefore, in order to reduce the heavy programming work and achieve system simplification, it is necessary to change the system structure. With the continuous leapfrog development of information technology, a control strategy that can make up for the defects of fieldbus and realize unified, efficient and real-time control of the whole system will inevitably emerge in the field of industrial control. Industrial Ethernet is a control technology that has developed rapidly to meet this need. Among all network technologies, Ethernet technology is the most ideal choice so far. It can meet all the following requirements:

　　1. Fully consider the needs of future development and have a high transmission rate, currently reaching 100 Mb/s.

　　2. High transmission security and reliability, and the certainty of hub technology.

　　3. The application of hubs does not need to consider the expansion of the network.

　　4. A standard is established: a new industrial control bus standard.

　　5. Connect with IT, the application of "world standard" TCP/IP technology.

　　6. Random network access technology in the entire network.

　　Ethernet is a technology for computers to access local area networks. Due to the substantial increase in Ethernet transmission rate, the industrialization of physical layer standards and the formation of Ethernet hub technology, the emergence of Gigabit Ethernet technology and collision-free full-duplex fiber technology, this advanced network technology has been promoted to industrial control networks that were previously considered unsuitable, forming industrial Ethernet technology. Compared with the current fieldbus-based control network, the control network based on industrial Ethernet technology is a low-cost (many commercial Ethernet chipsets and technologies can be borrowed) and high-performance control network solution.

II. Solution Analysis

　　(I) Embedded Industrial Ethernet Control System Solution Design

　　The control system network is divided into three layers: information layer, control layer and device layer (sensing/execution layer). Traditional control systems mostly use Ethernet in the information layer, while different fieldbuses or other dedicated networks are generally used in the control layer and device layer. At present, almost all PLC and remote I/O suppliers can provide products with Ethernet interfaces that support TCP/IP. After adopting the Ethernet architecture, the location of the controller can also break through the limitations of the traditional network architecture and can be located on-site or in the central control room. At present, the functions of controllers and even remote I/O to support Ethernet are becoming stronger and stronger. Web servers have been integrated in some controllers and remote I/O modules, allowing users of the information layer to directly obtain the current status values ​​in the controller and remote I/O module like users of the control layer.

　　In this solution, the network control system is divided into three parts:

　　1. Field device layer

　　It includes embedded node control modules and field working machines. The former mainly completes the collection of field data, the processing and storage of front-end data, and communicates with the upper layer through the Internet interface. The control module can realize the server function, and the information layer can be accessed through Web browsing (supporting point-to-multipoint communication). The latter is mainly responsible for some auxiliary and monitoring affairs, such as field data transmission, historical data processing, report output, etc.

　　2. Internal information layer

　　It is mainly composed of the internal Ethernet of the enterprise. It mainly completes the information collection and release of the entire system, that is, by accessing the Web server in the field node control module, the data of all monitoring nodes under monitoring are concentrated in the LAN server through the HUB hub, and are uniformly managed and saved, and released to the upper management department through Web browsing.

　　3. Internet network layer

　　This layer connects the various LANs of the enterprise through switches and routers to complete the global release of information. The department in the office can intuitively see the work situation on the site, the completion of the production plan, the working status of the equipment, etc. Even if it is thousands of miles away, it can grasp the operation of the enterprise (company) anytime and anywhere, making remote office a real reality. The industrial Ethernet control system solution is shown in Figure 1.

 

　(II) Implementation of embedded interface control module

　　In the industrial Ethernet architecture, Ethernet is used as the system bus to connect intelligent control modules. There is no difference between internal and external data communication. Hub technology is integrated in each controller, and internal communication is separated from external communication by allocating address space. The integration of hub technology and underlying protocols ensures the determinism, compatibility and integrity of Ethernet. At present, the protocols at the transport layer and network layer have basically been unified, and TCP/IP has become the standard network protocol, which is the "hub" for the normal operation of Ethernet. A key link in industrial Ethernet technology is to implement the TCP/IP network communication protocol in the field-level node control module (such as remote I/O module), that is, to establish a protocol stack. With the rapid development of electronic and information technology, it has become possible to embed the TCP/IP protocol into the node module through software or hardware. The software method embeds TCP/IP into the ROM of the microprocessor, and the hardware method is to design embedded processors and ASIC device chips and directly use them as network interfaces. [page]


　　The solution uses a single-chip microcomputer based on RISC structure with on-chip Flash program memory, which has in-system programming and debugging functions. Due to the use of CPU parallel pipeline mode and single clock cycle instructions, the instruction execution speed can reach 100 MI/s under 100 MHz crystal oscillator drive, and all I/O pins can be flexibly configured through programming. Based on the above characteristics, the virtual peripheral function can be realized: the CPU directly drives the ordinary I/O port to realize the hardware peripheral function (such as UART, I2C, SPI, CallerID, FSK, etc.) by executing the virtual software module. The most special thing is that this feature can be used to realize popular Internet protocol stacks, such as HTTP, SMTP, POP3, TCP, UDP, ICMP, IP, PPP.

　　The node module is implemented in a multi-tasking manner. The microcontroller can complete Internet protocol processing while performing data acquisition or completing I/O control tasks. At the application layer, any one of HTTP, SMTP, and POP3 can be selected as the communication protocol between the microcontroller system and the Internet remote management terminal; or other programs based on TCP and UDP protocols can be developed by yourself as application layer software. Using the Ethernet control chip, data packets can be sent to the Ethernet, and connected to the Internet through the Ethernet to realize a true embedded TCP/IP device. Figure 2 is a schematic diagram of the control module structure.

　　RJ45 is the interface between this system and the local area network. The data flow is as follows: the request information comes from the local area network, is sent to the network card control chip through RJ45, and the processed 05 data packet is sent to the single-chip microcomputer protocol stack, which parses the data packet to obtain the original request information. The request information is then processed by the single-chip microcomputer to generate a reply message. The process of replying information to the local area network is exactly the opposite of the above.

　　Features of the interface control module:

　　1. It does not rely on PCs or high-end single-chip microcomputers, and truly realizes the direct access of the 8b single-chip microcomputer system to the Internet. The entire system is completely self-sufficient.

　　2. It uses fewer peripheral devices and has a low system cost.

　　3. It supports IP, TCP, UDP, ICMP, HTTP, and SMTP protocols.

　　4. The system provides a 10/100 Base-T network interface and directly supports EthernetIEEE802.3 protocol.

　　5. Through the system's built-in RS232 serial interface and support for Web page download function, the monitoring point data can be displayed and controlled in real time and dynamically.

　　(III) Ethernet communication protocol development

　　Ethernet is designed according to the requirements of the local area network. The Ethernet standard (IEEE802.3) defines the physical layer and data link layer of the OSI reference model. The physical layer defines the cable type, connector, and signal level; the data link layer defines the frame format, error control method, channel allocation method, etc. However, Ethernet cannot complete the functions above the third layer of the OSI model. In this sense, it does not belong to a complete network protocol. How to implement the network layer, transport layer, and application layer of the OSI model based on the existing protocol is the key to solving the problem.

　　The implementation principle of the network protocol stack is shown in Figure 3.

　　The software protocol stacks in Figure 3 are all written in MCU language and stored in the Flash program memory of the MCU. Different manufacturers at home and abroad have launched different industrial Ethernet protocols for their Ethernet products. In general, protocol development is concentrated in the following three aspects:

　　1. Network layer protocol

　　The network layer mainly handles the activities of packets in the network, such as packet selection and route determination. Including: IP protocol (Internet Protocol), ICMP protocol (Internet Control Message Protocol) and IGMP protocol (Internet Group Management Protocol). IP protocol is the main application object. All data in the Internet is transmitted in the format of IP data packets. Its biggest feature is to provide unreliable and connectionless datagram transmission services.

　　In the development of the protocol stack in the embedded control module, the implementation of the IP layer is to package the message to be sent out with IP, that is, add an IP header to make it conform to the format of the IP data packet and send it to the physical layer; the received data packet from the physical layer is IP unpacked, that is, the header is removed and sent to the TCP layer.

　　2. Transport layer protocol

　　TCP provides a connection-oriented, reliable byte stream transmission service. Once the datagram is damaged or lost, it will be retransmitted by TCP. In the embedded module protocol stack, the TCP layer is implemented through the provided TCP API (application programming interface).

　    3. The application layer protocol

　　HTTP is a protocol on the WWW. When a user wants to browse a web page on the server, the HTTP request will be sent from the user's browser to the HTTP server. The server responds to this request and sends the specified web page back, and the user can see the web page. Because the TCP layer is below the HTTP layer, that is, HTTP uses TCP as its transmission protocol, so the above two TCP connections are also used by the Web Server. Naturally, the sending and receiving functions of the TCP API are also used to implement the Web Server's requests and corresponding commands to complete data transmission.

　　In the control module, the HTTP protocol can be used to construct a Web server, the SMTP protocol can be used to construct a client for mail services, and the PPP protocol can be used to construct a point-to-point system, and finally a C/S model can be established. All server resources are stored in the external memory (E2PROM), and its capacity determines the size of the resource file of the Web Server. HTTP uses a uniform resource locator to specify the network resources returned to the client (such as HTML, text documents, images, Java scripts, Java applet, PDF documents, etc.). Any type of Web server in the network can communicate interactively with it to achieve remote and real-time control.

　　(IV) Development of system application platform

　　The development of industrial control application software is mainly based on B/S network architecture. The control bottom layer realizes the collection of field data information and establishes a Web resource server by establishing an intelligent node module. The client only needs to read the data in real time through the browser and transmit the control command. The focus of this application system is on the development of the bottom server. At the same time, the system monitoring software adopts a network design, has good scalability and interconnectivity, with functions such as centralized parameter display and real-time data query, and special functions brought by the networking of all devices, such as automatic switching of control modules, network fault detection, resource sharing, etc.

III. Conclusion

　　The rapid development of network technology has a profound impact on the transformation of industrial automation technology. Embedded industrial Ethernet, a highly open, flexible and powerful new industrial control network, will connect the enterprise's field equipment layer, control layer and management layer with very high efficiency to form an enterprise information system based on network integrated automation. It will inevitably penetrate into all aspects of manufacturing industries such as machinery manufacturing, automobile manufacturing, semiconductor manufacturing, petrochemicals, etc., and will also be widely used in building automation, power system monitoring, robot control, textile packaging, printing and other fields that require digital information exchange and integration. Therefore, industrial Ethernet as a new "field bus" is the inevitable choice for future industrial control networks. It realizes the seamless connection between the field equipment layer and the enterprise's internal information network (Intranet), and while establishing a truly unified industrial control network, it applies the idea of ​​openness to industrial control networks to a higher degree.