Industrial Reference Architecture Model 4.0 (4.0 RAMI) to display a three-dimensional model of all the key elements involved in the manufacturing, the hierarchical levels of the model dimension (the right horizontal axis) describes the automation of a hierarchy of seven layers, shown in Figure 1 from bottom to top are: product, field devices , control equipment, workstations, operations center, enterprise, connected world.
Figure 1. 4.0 industrial automation hierarchy
predetermined level consistent classification criteria for this function, and IEC 62264 ( “Enterprise – Control System Integration”) and IEC 61512 ( “Batch Control”). In an actual plant environment, in order to more clearly analyzed at different levels for the different requirements of the communications network, we level between the field device to the workstation made further subdivision, as shown in FIG.
Figure 2. Drag Automation Model
The data size, cycle, transmission distance, wherein the number of index nodes, etc., typical values between the different levels of the network communication graph as follows:
Table 1. Drag automation requirements and traffic characteristics
because different levels of requirements for the communication network, the current network technology at the conditions, usually only the cloud storage data operation, using artificial intelligence algorithms depth learning, online identification and implementation of the system modeling, optimizing operating efficiency and maintenance of equipment or operating state of the device, and does not directly control the driver to achieve control of the controlled object. 3D cloud print, for example, transmitted from the platform to the 3D cloud G-code printer to control the printing apparatus, shown in Figure 3:
FIG. 3. A main problem cloud 3D printing system architecture
The architecture of the present that once finalized purchase equipment, controller hardware can not be upgraded (providing scalable computing performance). In order to improve the efficiency and accuracy of 3D printing, a team developed a software algorithm called “FBS vibration compensation”, can effectively improve the 3D printing speed doubled, however, because some 3D printer controller “of computing power and memory is very low, can not support user algorithms. As another example, more than one robot coordinated motion (RoboTeam) scenario, the current is limited by the computing power of the robot controller, generally only support four synchronized motion of the robot, can not be extended to moreNumber of robots cooperative movement. Further, if the field controller device needs to be upgraded or replaced, or the entire device requires the whole production line downtime , and thus put the new device controller, which would lead to the production suspension; some occasions to meet operational requirements of high reliability, the need for a redundant control platform (hot standby PLC), this approach is too costly and complicated work. With the network technology (5G Control Engineering Copyright , Wi-Fi 6) and network control theory (predictive control, data-driven control) continues to develop, there are so-called cloud control system (Cloud control system) concept, which combines the advantages of cloud computing and network control, the system topology shown in Figure 4.
Figure 4. Drive control system network topology
In this architecture, because the cloud is an elastic service, a user need not know the physical location of the service provider, only to have the network connection, you can then be configured according to the virtualized resources required (computing, software, data access and storage) on the virtualized resource various data (real-time data, historical data) and the analysis processing, thereby generating a control system signal. Finally, left at the scene of the terminal may be simplified to a simple input and output devices or simply have limited computing / storage capacity, the above-mentioned problems can be resolved in this architecture, such as the even field device controller run not support “FBS vibration compensation algorithm”, by extending the cloud computing power controller, operating efficiency can be improved 3D printer , as shown in FIG.
Figure 5. Cloud 3D printing control system architecture
From the motion control, the robot and the 3D printer architectures were similar, as computing power and memory controller are low Control Engineering Copyright (C) , and thus can not support the issue of the new algorithm. There is the same kind of problem in the robotics industry, this cloud control system provides a new idea. While cloud control system has many advantages, but at this stage still faces many challenges, such as: challenges of the information transmission and processing, how to protect the large delay underQuality control certificate stability and closed-loop system; control system security challenge, not only to resist random disturbances and uncertainties of the physical layer, but also to protect against network layer strategy of targeted attacks and so on. So the development of this control architecture does not completely rewrite the reference architecture model for the industry is well known, but in the near future, with the continuous development of network technology and network control, cloud control system will include a variety of devices, including robot and a variety of practical applications play a positive role in promoting.