Introduction (text/talk about things Song Huazhen)

Speaking in Kan Kan is not pragmatic, and solving problems is king!

Description: This article is selected from the official account of WeChat, Marketing Manager of B&R Industrial Automation (China), and is published with authorization!

The upsurge of industrial internet also raises the question of whether its surname is "work" or "net". But in my opinion, the entanglement of this issue is somewhat "self-centered", although I do observe that the IT industry lacks understanding of the industry and some are self-righteous.

However, we have to admit that ICT technology does have great potential to create some space for manufacturing. In fact, if we all look at the problem from the user's point of view-we must say that the industrial Internet must serve the customer's value, and the judgment and choice are actually determined by the user's experience. On the other hand, Kan Kan's talk is not pragmatic, and solving problems is king, probably because people at the ot side have been busy debugging equipment on the spot and have no time to write PPT.

0 1, the misunderstanding of automation in chronology development

Of course, as an automatic person, I also have some views on the present situation, because it is a very common phenomenon that people classify the development process of manufacturing industry as mechanization, automation, digitalization and intelligence in the way of "dating history".

Of course, all kinds of lists clearly list automation as a process. It seems that intelligence is a higher process and automation is only a lower process. Obviously, the real industry is not like this, because there is no pure automation on the spot, and there will be no machinery, digitalization and intelligence, and there will be no intelligent era.

The process without automation is logically untenable, because machinery, automatic control, information system and intelligent application are often regarded as a vertical level, rather than a historical process. Therefore, it is not appropriate to define the function and position of automation in the way of "dating history". Moreover, this definition of dating does not conform to the actual integration and infiltration of automation technology and intelligent technology, nor can it reflect the fact that Tian Kai Automation Company is basically a complex.

Automation is not the dynastic history of manufacturing industry. In the future, it is always changing, and it is always looking for digital technology, intelligent technology and the integration of various technologies to solve production problems. Should the name of the automation industry be changed?

02. Automation is far from the automation of the past.

In fact, people always understand automation as PLC, servo drive, HMI, motor ... In fact, the role of automation is constantly expanding, and the automation industry is close to the scene and constantly solves problems for the scene, making the problems in automation more comprehensive and holistic, rather than a partial product concept.

Several notable features have become the explanation for the development of automation:

Is automation a software industry?

More and more automation enterprises realize the importance of software, and you will find software everywhere in the automation industry, including several aspects:

① basic RTOS

Whether RTOS(VxWorks, QNX, μC/OS-II, etc. ) or based on RT-Linux and Windows Embedded, at least there are few pure hardware logic PLCs lacking operating system at present, because they have to run various development environments and program scheduling, including multi-threaded task processing and different task levels.

② The importance of software engineering

Traditional logic-based programs can be written by a programmer in one cycle, but today, we have to write complex algorithms, such as robots, CNC and positioning synchronization control in PLCopen PartIV, professional libraries such as filtering, basic axes and CNC, hydraulics and so on. And advanced to the industry-level application algorithms, all software.

③ Open software interface:

Like Siemens, B&; Both R and RA have modeling and simulation interfaces for MATLAB/Simulink, which can be used for electromechanical and soft integration design, including the progress of digital twin technology. These are all software.

④ Architecture of digital platform:

Automation vendors such as ABBAbility and SIEMENS Mindshpere have extended to the stage of edge, private cloud and background application development and intelligent algorithm integration.

If you look at automation from the perspective of the above software, will you understand automation as an industry that only sells PLC and motor drives?

03. Digitalization and automation have been progressing.

Of course, many people understand digitalization as CAD, CAE and CAPP, and call it digitalization. However, this is the digital part of the design, and the other part is the digital part of field work.

The most intuitive industrial digitalization is the field bus, which itself digitizes the signals on the spot, first for the convenience of transmission (reducing interference), and also for the later calculation and processing to be more direct, from the original Modbus, CAN, Profibus, DeviceNet, CANopen, Then developed to Profinet, POWERLINK, EtherCAT, Ethernet/IP, CC-Link IE, and today's OPC UA over TSN technology, various automation manufacturers have been advancing the process of digitalization.

This digitization process is closely related to the industrial application in the field. At first, it was only to solve the problems of signal control, transmission interference suppression, complex wiring and remote acquisition control. Later, with the improvement of machine automation and the demand of motion control, robotics and other applications, real-time Ethernet technology was put forward, which is the demand for more general low-cost networks, realized by it technology, and then developed. With the interaction with the management system and the interconnection of machines, there is a demand for "information" and interoperability, which leads to the unified transmission of OPC UA and TSN for collaborative real-time and non-real-time data.

The value of OPC UA is far from being an interoperable specification. It is the norm and standard of integrating information model, security, transmission, preprocessing and industry in digitalization.

Some people reflect why our asset management, including coding, has not been unified, and how to carry out workshop management. I told him that AutoID is a companion protocol of OPC UA, which can solve this problem, and vertical industry information models-PACKML and Euromap can solve it. VDMA has also developed a collaborative information model of robot and vision, including the management shell of Industry 4.0, which is also based on OPC UA. Recently, I talked to a friend of OPC UA Foundation about this topic. In fact, the biggest problem of MES promotion is the lack of unified specifications such as OPC UA. Therefore, MES projects often involve a lot of engineering investment, and the user cost is expensive, but the implementer has extremely low profits because of the large engineering investment. Everyone is very hard. If OPC UA is adopted, the project cost will be greatly reduced, and digitalization also needs good methods and tools!

Digitization-isn't that what automation people have been doing?

Intelligent technology was made many years ago.

This is also very interesting. Although the AI scene is hot today, we will think about how important AI will be. In fact, intelligent algorithms in the industry have been carried out for a long time, including strategy analysis, path optimization, logistics and system engineering at the edge layer. It has been several decades, including the realization of machine learning in PLC, which is completely feasible, because today's PLC is no longer the controller of hardware logic instructions. Including B&R RTOS+ Windows-based PLC (which can be developed by C/C++) or APROL platform (which can be programmed by Python), so we can no longer look at automation according to the traditional vision.

As far as cybernetics itself is concerned, there have been various models, methods and practices about intelligence decades ago, and the key lies in "economy". Just like last year, I talked to a few friends about a company that produced Dr. Xu Qiang's intelligent algorithm about driving more than ten years ago, because the cost of processing chip needed to realize this algorithm is very high, and the cost of FPGA today is very low.

Therefore, it is economical. Similarly, in the past, multi-loop PID coupling and decoupling calculation under complex dynamic conditions required very strong processing power, but traditionally, ordinary controllers and PCs did not have such computing power, so it is expensive to build high-speed dynamic computing networks for many industrial scenarios. Nowadays, ICT technology makes it economical to test and verify servers, high real-time response networks and software algorithms. Only in this way can many intelligent algorithms in the original automation field be really put into industrial application. The application can have two modes. One is that the automation system integrates algorithms locally with AI chips and software, or puts a large amount of data with large time particles in the cloud for calculation, which depends on which is more economical and applicable, not whose technology is better.

Therefore, any innovation is based on economy, and technology has long existed. Even the problems we are discussing today will have a historical origin of more than twenty or thirty years. However, with the increasing economic growth, these problems can be really applied.

05, we all serve the value creation of manufacturing users.

Therefore, in the era of industrial internet, IT doesn't matter what OT and IT surnames are. On the one hand, we should give full play to our own specialties, integrate with each other with an open mind, work hard on the interface construction of norms and standards, and integrate with each other to solve problems. However, automation is not the traditional products we understand, such as PLC, servo drive and motor. It has long been blurred with digitalization and intelligence, and is becoming the core force of manufacturing upgrading. Become the landing place of OICT industrial Internet integration.