Blogs

As Mr. Zhao Min said, “The big river and lake of the Industrial internet platform is too hot, and everyone in the hot rivers and lakes is about to boil and blister.” There are “world” or “global”, “China “, “XX Province” of various grades and specifications, the bosses invited are all well-known, millions of APPs and enterprises have moved to the cloud, it is really a prosperous scene of “the scenery is unique here”. This is a question – it’s not that people in the industry have a sour grape mentality, but “there are gods at the scene” – it seems that people in the industry are not so high-profile, it feels like a group of people are talking about your horses How to treat the disease, but you find that there is only one cow cultivating the land in your family!

Just like at a certain industrial Internet conference, there was a person who was doing “digital transformation”. I really don’t know what this company is for. It’s too abstract. Later, I asked what business it was and said that it can improve the ability of manufacturing through the Internet. – After listening to it, I said to him, “But the production site did not have the Internet in the past, nor does it have the Internet today, nor will there be Internet in the future.” – If you are talking about the office of the factory, there is, but there has never been Internet in the field production!

Whose business is the Industrial Internet?

What makes us envious is that the IT world is full of splendor. People who are involved in the industrial Internet are particularly good at packaging. There are many people who often play the role of “Internet thinking” success masters. There are really many people in the IT industry who can mix with ordinary people, but when it comes to the big guys in the OT circle, apart from their own circle and his relatives and friends, the general public obviously does not know it. This kind of innate The heroic charisma has also influenced the manufacturing industry, seemingly weaving a web full of overtaking shortcuts around corners.

I remember a time when the big guys of the Industrial Internet said that the power generation efficiency of wind turbines can be improved through big data. I questioned that what we need to do in the control system is to obtain the “maximum tip speed ratio” of the blades? I have adjusted to the maximum, how can you be bigger? What if the wind changes? – Then don’t I have a yaw system? The impact on the power grid can be avoided through big data, so doesn’t our control system have a “low voltage ride-through” function? Then I can do predictive maintenance – sorry, if your machine has so many fault signals, I think this company’s wind turbines have been eliminated by the market, simple logic – you have such a large amount of fault information let you study? But if you don’t have that many glitches – you’ll have problems with the accuracy of your learning!

Industrial Internet is like what Mr. Ren Zhengfei of Huawei said“Let’s not show off our hoes and forget about hoeing”– The real connection is not the physical Internet, but a network of “value”. What kind of value do you provide to your customers? What kind of synergy do you need for your production? How to improve the efficiency of information is based on the planning and design at the “operation management” level, and the Industrial Internet is just a tool.

Application first, or architecture first?

Mcrazy reminded me that the difference between IT thinking and OT thinking lies in the structure of the thinking mode. IT is a top-down architecture, emphasizing platform architecture design, while OT is a bottom-up execution, solving local problems first. , and then solve the global problem, because if the local problem is not solved, for control, you cannot even meet the basic production accuracy and production efficiency, and you don’t talk about upward data transmission at all – because you have not measured Exactly, what do you want that data for?

IT and OT are indeed integrating with each other, but OT manufacturers are also actively using IT technology to help themselves achieve “optimization”, and OT’s move to the Industrial Internet is based on the existing models that have been implemented by its own measurement and control. Level-by-level data integration and optimization, the advantage of OT over IT is that I understand the site, and IT does not understand what the site is planning, so its ideas cannot be implemented.

Which of the two paths is more suitable for industrial sites? It seems self-evident – architecting the big picture is good, but if you don’t understand objects, this won’t work.

Differences between industrial and commercial scenarios

People from IT are good at architecture and software, but their objects are often not physical entities, or are not “periodic” control tasks, and control tasks are based on “isochronous” periodic signal acquisition and execution. The objects are physical valves and actuators, hydraulic objects, servo motors, etc. These objects have their own mechanical and material properties.

People in the industrial scene dare not describe arbitrarily, and rhetoric lies in awe-Because the industry needs more stable, reliable and periodic transmission, and requires years of accumulation of know-how, there will be casualties and large cost losses.Therefore, it is not easy to make “promises” – how to improve efficiency and outline a bright future – if it is only on the PPT, it will not be able to convince people in the manufacturing industry.

Manufacturing is divided into discrete physical processing of materials. Most of the processes are chemical reactions to materials. In the process of finding economical production, there are too many constraints.Just like the injection molding machine needs to explore the process of a material, there may be thousands of types of materials, and the specifications of the products are also varied, so for each product, its process control is a process that needs to be accumulated very much. In the field of spinning leveling, USTER has accumulated 200 years of knowledge and experience in the field of metal heat treatment technology to determine the level of materials, and these have always been the inevitable path that the “strong foundation” of China’s manufacturing industry cannot overcome.

The industry is not the rapid iteration of “Internet” thinking, or the rapid growth of viral replication. The so-called “Internet thinking” success stories tell many people how to quickly grasp the capital “window”, get rich quickly, stocks, options, entrepreneurial teams , venture capital, this is the influence of Internet thinking that stimulates people’s desire to “get rich”. However, Internet thinking in the industrial field needs to be “medium” in terms of accumulation and development, which is not the right way. Therefore, for a period of time Many people have doubts about this path. What scenarios can be quickly replicated?

Digitization is beneficial to the manufacturing industry, as is data connection, and Internet thinking can also be used for reference—such as cross-border thinking and user thinking, which must be incorporated into the industrial field to help it upgrade.However, the “Internet thinking” that many people talk about today is not for the development of the manufacturing industry, but a desire for quick success and quick success.

Where is the industry so simple?

People talk about millions of APPs and hope to realize the clouding of millions of APPs through the industrial Internet. In fact, these have nothing to do with industry. Most of the applications are currently “displaying” such tasks.

In fact, the first step of the Industrial Internet itself is the connection problem, and a large number of companies currently existing in this scene are doing “manual work”—It is to provide data connection for different industries and connect complex fieldbuses. A friend asked how many fieldbuses there are. I found more than 50 kinds of data, and there are 18 kinds listed as standards by IEC. There are thousands of “protocols” generated on these bus interfaces, and these are what most “industrial Internet” companies are doing at present. In fact, they are “guide tables”, which compare the data in the machine’s data object dictionary with the data in the data object dictionary of the machine. The variables of the host software are mapped.

TSN came into being – Industrial Internet is different from internet

Connectivity is the first hurdle of the Industrial Internet, another hurdle involves the ownership of dataNow, you said that I sold a machine to you, so are the parameters of this machine yours? If I sell the control system to the machine maker, will the technology of this controller be sold to the machine manufacturer? It’s like buying an iPhone. Does the information in this phone belong to the company APPLE?

If it is only the parameters of the production state, it does not matter. What if it is related to the parameters of the process?This is the Know-How of machine manufacturers. These data cannot be opened. If they are not open, there will be incomplete data, and the conclusions drawn from incomplete data will be incomplete.

The difficulty of big data-based applications in industrial scenarios lies in “small data, big applications”. In fact, the amount of industrial data is not very large compared to graphics and image processing, financial transactions, and Mobike. Although periodic The data can be refreshed according to mS, but the fault data for vibration monitoring is not very large, because the premise is that there must be a large number of faults, which is not allowed.

Guided by application requirements, natural development path

Different from the industrial Internet, companies from the automation industry have also begun to move in this direction. However, it is obviously more cautious. Every industrial application in the automation industry comes from the actual needs of the industry. Here, IT is a tool , to solve practical problems in the industry through the integration of cross-border tools, which is the starting point of the industrial Internet of people in the automation industry.

For example: my client needs to solve the connection of production lines outside the control, to connect the injection molding machine and the robot, I need an interface, everyone will promote EUROMAP, and the packaging industry needs to connect filling and labeling to promote the PackML standard , to connect the machine tools, promote MTConnect, and then find that these machines may produce mixed production, then unify the specification of OPC UA, and in order to achieve interaction with IT, promote TSN technology.

OPC UA is constantly seeking to solve the problems of the Industrial Internet

Every technology comes from the needs of the manufacturing site. When we developed millions of industrial APPs, did we think about what these APPs do?

Every vertical industry has their own application Know-How, and these are people in the industry who have accumulated decades of experience. In fact, many Internet people who do not know the industrial scene think they are very powerful and can do anything. In fact, AI has been used in the manufacturing industry for many years. IT is a tool for the manufacturing industry. In the past, the computing power was insufficient, so many applications of AI It has not been promoted, and now with economy, many applications can be realized. Many technologies are not innovations of IT. IT only lowers the threshold of technology through commercial applications, making industrial applications feasible.

Know-How is still in the industry itself, IT is only a tool, not the essence of manufacturing.

In the Industrial Internet, we need to be in awe of the manufacturing site. Just like Kazuo Inamori said, “there are gods at the site” – if you don’t know the site, you must connect the industry. This is very scary, and it will not be able to land. To quote Mr. Ren Zhengfei “Let those who hear the fire make decisions” –Let people close to the scene guide the manufacturing industry, and we must be in awe of the scene.