Current hot industry the Internet (IIoT), so a lot of companies flocking, but also to many companies around the detour. After GEPredix operate independently, we continue to adjust its strategy to deploy a new release of the privatization of the product is still a lack of real customer stories. German version Predix quite demeanor of industrial internet platform Axoom, in just the past July was unexpectedly sold its other stores Rich Dad, the industry set off a firestorm.
had crossed the mountains and the sea, but also through the sea of people of Siemens, to see the situation of fellow travelers, Zuoheganxiang? 2019 is already more than half gone it will reach a point in time when the “2020 Vision Company” program, Siemens is anxious not to worry? Earlier this year, Siemens and Harvard Business Review HBR jointly issued a research report “Internet industry to accelerate the schedule.” In this attempt to cover corporate C-suites global industry executives survey, 741 executives, 74% think things will help them build a competitive advantage in the market within two years. However, obstacles are obvious, 90% of executives believe the return on investment ROI projects are difficult to calculate things; there are many company executives said a lack of knowledgeable experts, the lack of confidence of things to deal with the complexities and challenges of the project. Many people expressed a dilemma: no hard ROI data, the company’s senior leadership in the moment temporarily unable to do things finalized investment projects. Conversely, however, if a competitor first to try things, a year or two it may cause poor can not catch up the gap. After
publish the report shortly, in April 2019, Siemens usher in a key node, Siemens Things Services division was formally established, the mission is to accelerate the IOT services. ● Why have Mindsphere, but also to set up things Services Division of Siemens after the cold thinking? ● IOT services business, Siemens and where to find growth opportunities exponentially? ● data sets, Siemens how to interpret? ● Siemens industrial layout quietly mapping knowledge, it will be the next super vent it? With these questions, and I’m president of Siemens China, General Manager of Siemens (China) Co., Ltd. things Zhu Xiao Xun Services Division conducted a dialogue challenge the limits of thinking, the war gaming experience for the things he has done service development feel the behavior of the logic behind his decision, choice and focus. Why to 01Things stand Services Division? Peng Zhao: What is the positioning of Things services division is? Zhu Xiao Xun: This is a business model innovation, the transition from selling products to selling services, with end to end services to help customers complete the digital transition. This picture shows the Siemens think about the seven types of things should be carried out in an integrated service platform CONTROL ENGINEERING China Copyright , pushing tightly integrated IT and OT, but also reflects the service Things to do .
Briefly, the procedure was the networking service is divided into seven steps: Step 1 ● advisory and prototyping ● Step 3, it is necessary to link the existing equipment ● Step 4, the conventional system may OT intelligent adaptation and upgrade ● step 5 need to link all these devices and systems, there is a need of things platform for management and integration, there may need to be other things existing platform access which ● 6 7 step, said after with the platform, you can develop a variety of digital applications according to customer needs. And long-term operations and services. This 7-step finish, Siemens really help customers walked the entire digital journey. Customers not only hardware and software, but also end to end services, networking services complement the existing digital capabilities, I think this is a complete process of establishing digital world. Peng Zhao: Internet of Things services should be how do? Zhu Xiao Xun: first, to assess the value of digitization can bring by way of consultation, clear the organization’s existing data situation. We have a modeling evaluation kits, to solve the problem of modeling a stock of any plant. If the new plant, will have a three-dimensional simulation is now widely flashy, company managers will see a simulation software to do it with dynamic simulation display. And although there is no stock of the factory this dynamic simulation display, but as there is data in the database to be used, as can timely reflect the value of the data, but we did not pass the corresponding tools and models to show it out. So modeling is a core step. Second, when the customer has a production line to be upgraded, in what should be what kind of sensors placed? From the perspective of Things services, we will provide recommendations for the deployment of IoT KPI (Key Performance Indicator) oriented. Because the factory director general to look at more than a dozen to two dozen KPI, he needs to know today FactoryHow well operations. Here we have the knowledge map as a core technology, it may in the end what point these KPI from the production line has been traced back to, what kind of data needed to solve the problem now. This is also the benefits of the knowledge map, we were able to accurately tell the director, what kind of data he was missing somewhere. How do IoT thus recommended factory deploy third, operational level optimization. It is also based on the knowledge map, it is to let the software do the algorithm of knowledge map. These three steps, the first step is to create knowledge maps, second and third step is to use knowledge map. Zhao Peng: this process from above, you can see knowledge map is very important. In the end what is the knowledge map? Zhu Xiao Xun: In the past we are talking about technology package is important, but the technology package are posters, after people read this technology package file, this process can know how the product is made out of. So how do you put inside the factory posters technology package, so that things become a machine can understand it? We are mapping knowledge through the technology package becomes something machines can understand. When the machine can understand, go write a variety of APP when you do not need to do too much customization section. The importance of knowledge map is that it provides a core support for the involvement of various algorithms. Without it, a lot of access to data is loose.
02 sets of data in just a cup, what is important is that bubble tea Peng Zhao: What knowledge map is a carrier? Zhu Xiao Xun: the data table. The data sets to run what is not important to do things for the service. It can run in the Siemens MindSphere above, you can also run in a customer’s existing cloud platform above, you can also run the factory at the customer’s server above, can be. Generally speaking, there are two very important. First, I am talking about in this data table is not necessarily MindSphere, the data table is not linked to industrial internet platform, which is platform-independent (Platform Agnostic). Second, the data in the table will not change the existing system architecture. Original original critical data will still exist several different systems, we only read data , not converging data. This can solve the customer to whom the data confusion. Even with data sets, data or stay in its original place, the data does not need to be handed whatsquare. Peng Zhao: So you can not MindSphere. Zhu Xiao Xun: It is not necessary. Data in the table itself is not dependent on MindSphere, but we see Siemens MindSphere is the industrial market from commercial maturity, APP layout view of the platform of choice. Zhao Peng: this data table, and BAT are talking about data in Taiwan, it is a means? Zhu Xiao Xun: I do not know what BAT’s data table in the end yes. I think that the data in the table, can draw an analogy, in fact, the data table is a cup. I want to drink tea cup, no cup, I can not make tea. So although I wanted to make tea, but my statements will change, I might say to you, you need a cup. In fact, who’s this cup is the cup look like, it does not matter to me. Data in the table is just a cup, the cup preparation is the key to pretend tea. No one can say that have data sets, but the data in the table is not capable of leading to real viable technology paths, by industry standards, to get knowledge map. Finally, the use of digital mapping knowledge as a basis for the plant twins, to build a sustainable use and sustainable development of digital platforms, this is the key. Zhao Peng: It is not to be understood as Siemens Things to final customer delivery service is the knowledge map and data bus, similar to this type of thing? Zhu Xiao Xun: so to speak. In the physical world is reflected in the data bus, what is missing is the customer data, we will give suggestions to add. In the virtual world, we offer customers the factory model to deliver knowledge map. This knowledge map is a living digital twins. There are many digital twins, we tend to think of the three-dimensional image after modeling with CAD digital twin is, in fact more than that. A plant, or even a room, can be based on demand modeling, creating knowledge maps. Although it is not flashy presentation of three-dimensional effect, but what you plant Riga new physical equipment, knowledge map will be reflected immediately. Moreover, when faced with a huge production line, when, how can we know in the end what kind of data needs? Generally there are tens of thousands of factories, hundreds of thousands of data points is normal. How to know what data is still missing? I do not have a ready-made software can tell. So when have the knowledge map, it can serve as the basis for the overall layout of things. In the business operations of state level, is not really used theThese data? It is not really used to analyze the data in real time for feedback, to verify? Knowledge maps can also be used as basis for real-time optimal state your perception is not in the production line, which part there is room for improvement. Zhao Peng: data sets, knowledge maps, digital twin … on the market a lot of people are doing, what Siemens unique place? Zhu Xiao Xun: I think is unique is that our technology path. Our unique place that is first by means of semantic model, the data from various sensors semantics, semantic tagging, become semantic data. This process is to heterogeneous data structured process, so that the data is associated.
Semantic data What role do? For example, a compressor, which is used in a power plant or chemical plant? What this compressor used in the scene? What this equipment and connected to the compressor? These are very important related information. This information helps machines to understand the machine. The second step, based on semantic model, increase people’s knowledge as input to generate knowledge maps. Encyclopedia of knowledge as a semantic map of the world, which is a set of encyclopedias machine can use. With this knowledge map, if a production line or a piece of equipment is a problem, the machine can easily be retrieved information, diagnostic guidelines given digitized, that is, let the machine to give effective advice. Beyond that is the next stage after the decision recommendation system, the machine can be based on knowledge of maps and data, automatic matching of similar models and decision-making mechanism, select the best from a variety of fault case base. The final step is application development based on knowledge and floor map. So I think there are two big industrial pillars of artificial intelligence, knowledge is the first map, followed by deep learning. At present, we should pay attention to most of the knowledge map.
Zhao Peng: knowledge map has not been mentioned a lot of people. Zhu Xiao Xun: In the absence of knowledge map, the pain points we face are very large. From the customer’s point of view, manufacturers will feel the lack of a universal digital software, all services are highly customized. The result must be the expensive software, expensive applications and services. From a developer’s point of view, very complex and industrial data scattered in various places, developers need to put a lot of acquisition costs, but also to manufacturing companies operating there are a lot of understanding and knowledge, with a very high threshold. Developer ecosystem shortcomings, in turn intensifiedThe problem manufacturers is difficult to achieve rapid return on investment through the digitization project. We propose manufacturing companies need data sets. Just like the metaphor, the data table is a cup, soak a good cup of tea this knowledge map, so that things can really energize the manufacturing sector, the need for such a unified data sets. From the ability, the data in the table first need a centralized data show. Although the show is not something to bring core values, but a necessary condition. Secondly, the data station also need to have the ability to integrate heterogeneous data, the ability to automatically connect, converting data in different formats. Third, the data sets need to be able to associate cross-cutting events CONTROL ENGINEERING China Copyright , association analysis can be made according to the context. Fourth, the deployment of the data in the table must be flexible. Data presented in the table does not mean that the world wants a unified platform, we put the data are placed here. Data in the table is the solution Siemens response to the current status quo and raised. China now has hundreds of industrial internet platform, there are a number of data sets, I do not want plus one thousand and one platform to solve data problems in manufacturing. 03 Where are the opportunities exponentially in? Peng Zhao: Internet of Things services, where China’s opportunities? Zhu Xiao Xun: It should be made clear that China and Germany are not comparable Industry 4.0. China has a lot of industry is not high level of automation, digital is a good means to help them improve efficiency. I do not advocate for these enterprises to automate a large number of shop equipment, but it is recommended to do the simulation and planning by digital means, and then decide how Lean. The German Industry 4.0 is developed under the premise of a high degree of automation, and therefore China’s market entry point for different nature. We will begin to start upgrading the stock from the factory, which is based on China’s current situation. China is a manufacturing country in the world, it is solved by manufacturing artificial in these most difficult technical problems. Especially to see some of China’s assembly plant types will find that a lot happens most automation can not solve the problem, most industrial robots can not solve the problem, they had a perfect realization of human use. Who is the world’s most intelligent machine, even the most powerful companies, in fact, there will be a very “industrial 2.0” production scenarios. Among the digital transformation process, if these technical problems can not wait for the newGeneration robot algorithm to solve the case, is not realistic. China and Germany of intelligent manufacturing industry 4.0 will be out of a completely different way. China’s stock market, from a purely manual, to semi-automatic, then automatic, not the same. Our concept stock reform is a hand-line does not need to go through the entire automation process. A handmade traditional manufacturing enterprises in China should be how to digitize it? Jokingly say, I often “deep-fried dough sticks,” the shop, for example within the team. I told the team that you do not have to first consider what automation is first assessed by digital means the efficiency of any form of production lines, allows you to discover evaluate the efficiency of these links are not really closely linked to asset utilization, staff utilization is not optimal? Can enhance efficiency, the benefits that digital technology. Without changing any business model, through digitization can enhance plant efficiency. Our main concept is the transformation of the stock of any style factory production line, through modeling, digital simulation, we can find room for improvement. Peng Chao: is not it also shows that from another perspective, the existing manufacturing enterprises is very confused, do not know what they want, or what to do how to do. Zhu Xiao Xun: most are not very clear. We are in the process of communication with customers, a lot of people to the digital still very confused. There are several puzzles: first, customers do not know what value digitized in the end, why invest so much to do such a thing? Done in the future, what kind of return? Second, customers do not know that they are now in the end how much data that is distributed inside the system in which the? Factories in the past 10–20 years, a different system in the end how much hidden treasure? Basically no companies really have the time to sort out this matter. An example of a real occurrence speak. He has a client for Apple and other companies to do OEM. The problems he faced was , he ran a production line to build two or three years after completion, this line would not have had, you have to replace the entire line according to the needs of the foundry. When he obtained OEM orders when there is no time to consider how to reuse some of the existing production facilities. In order to catch up with the progress of the foundry, he had to scratch to buy new equipment, build a new production line. The cause of his warehouse and a backlog of several hundred million of the original equipment is good, but these devices are not bad equipment. We are doing a consulting project with him, it is howBy modeling and asset management, to his existing equipment to master all the information clearly, and help him to reuse these facilities, to maximize cost savings and improve efficiency. Zhao Peng: too many people only value increment. I agree that the stock market is very big, there is no breakdown of some of the more customer-portrait? Zhu Xiao Xun: Our service is expected to grow things is very large, and did not make a strict division of various fields. Siemens has a vertical field of any traditional strengths, we are able to provide services. The transformation of the stock of the plant is a process of exploration, the market system of things and services currently do not exist, there is no similar program itself, to some extent, this is our first break and try. Like Siemens internal business, and create a new model. Zhao Peng: Siemens ready to invest much resources? Zhu Xiao Xun: We expect that by 2025, Internet of Things integration services market will grow by 10% to 15%. Siemens China is considered the United States, the European core markets side by side. Siemens’s traditional business is selling products, software sales model, and things Services Division is selling services rather than products linked with a business model. In China, we have more than 200 team dedicated to serving this market. Peng Zhao: What is the stock of the plant can successfully transform key point is? Zhu Xiao Xun: the leader or the entire business confidence and the will is very important. Truly successful transition, or after digitization pilot willing to invest, it is important that the concept of leadership. If the leaders do not think digital is the future survival of the core business, to do the pilot light that is of no use. I have not seen a case, after only a pilot, it changed the concept of leadership, not one. First, there often is a leader in the belief that digital is the only way my transition, and then to prove the concept of leadership is through pilot. Like Siemens White Paper “Things become a reality,” this figure painting. The outermost picture, is change management, reflects the confidence and the will of the leader is the most important prerequisite.
04 “from 0 to 1”, IIoT What’s missing link? Peng Chao: how do you see China is now the Internet industry development? Zhu Xiao Xun: China’s manufacturing intelligence, often referred to integration of the two, is the automation + information. This is what is the relationship with digital? I think the first step from a real digital are less than we are now throughProcess calendar from 0 to 1. So what is my opinion of digitization? First, as a traditional manufacturing enterprises, in the physical world is the real value generated by the processing of raw materials, put it into production product, converting the product into a value of finance. Second, the physical world of manufacturing companies, can be mapped to a digital world. Through the digital world of modeling, simulation, optimization, we can provide support in decision-making, to influence events in the physical world of the results, which reduce the possibility of mistakes in the physical world, to speed up all processes, so that the whole process becomes more smooth and efficient. Another point, from the industrial data path between the application of artificial intelligence have not setting up. AI special fire now, but artificial intelligence may not be successful in industrial applications. Now many industrial applications of artificial intelligence, on the whole is a single data source. Such as making parts of defect detection, in fact, all of the available data is image. If you do speech recognition, all the data is sound. Nobody really complicated project to talk about, and why? Because the data characteristics of the industry, a large amount of data industry, the kind of data is completely different from the structure of the data to unstructured data, from real-time data to a non-real-time data from a single data very little value to the accumulated long enough to be presented data aggregation certain trends … so many data processing enough headache, and finally the application of artificial intelligence will find that even the best industrial design, architecture, data will always be dispersed in a variety of subsystems in these They are unavoidable issue. If you want to make good use of artificial intelligence industry, can not bypass path is to establish a knowledge map. If you do not pre-semantic models and setting up the knowledge map, then AI is likely just a castle in the air. Model-based knowledge map has been established, ongoing acquisition and accumulation of data in order to later paved the way for digital and artificial intelligence applications. With the map in the future when Knowledge, artificial intelligence possible to achieve a final ultimate goal is that all the equipment completely independent and autonomous.
Peng Zhao: In your mind, there is no final scene of the Internet industry. Zhu Xiao Xun: I can not predict what will happen in the future, but I have some judgments and assumptions, which is the guiding assumptions of my own to do something. My hypothesis is that there will be a variety of non-stop industrial enterprises to launch their own Internet platform, homogenization of these platforms is inevitable. That’s why I now want to push some of the technical point of viewI think the right thing but not the same, because I think these homogeneous products can not solve customer problems. Secondly, I feel so much industry Internet platform will persist, they will become the future of things integral part of the ecosystem. I do not think IIoT will be like the consumer goods sector, as only a very few platforms dominate the political arena. Industrial ecological future of the Internet will be a rich, interesting ecology. Finally, it comes to the digital transition, I think we have to realize that this is not a purely technical matter that can be solved. Technology can only solve a small aspect, if the lack of good strategic and operational planning, will lead to various problems. In the digital process, because of the lack of strategic planning, a lot of examples of marking time. A lot of people do digital transformation project, you will observe that he always was doing a pilot, finished a pilot to do a pilot. Because these pilots can not together, not unified can zoom together to create value for the reproducibility of manufacturing companies, so I had to stop doing a pilot. Technical architecture needed to match the internal management system, or even the best software and systems, companies are not used up. A real case is that a well-known German car companies buy our full range of Siemens software, buy over discovery, finished, his staff had no desire to go with the digital stuff, enterprise software simply not used up. Finally, he came back to our specialized management training and corporate culture within the enterprise. So the core of the matter is that we develop with customers and target digitization, along with the development of route planning transformation and help customers to promote change throughout the organization. I mentioned that we are going through the process “from 0-1”, the physical layout of the ecological network is still very imperfect. Digital transformation is not a business can do, but there must be a group of people share the same faith to do this thing digitized, we can truly from “0” to cross “1” and from “1” to become infinite. Siemens service Things to do this thing, we are not alone on this team to achieve social impact. Finally, you will see that the success of things and services Siemens, Siemens is not a business success, but the success of an ecosystem. This Summary: 1. Siemens build things Services Division, is an innovative business model, the transition from selling products to selling services. 2. Data in the table is just a cup, the key is what this cup of tea ready to install. Data in the table is not capable of leading to real viable technology paths, create knowledge maps, and then use the knowledge map as a basis for the digital factory twins , to create a sustainable use, sustainable development digital platform, this is the key. 3. Industrial AI has two large pillars, the first is the knowledge map, followed by deep learning. At present, we should pay attention to most of the knowledge map. 4. semantic knowledge map of the world is like an encyclopedia, which is a set of encyclopedias machine can use, it is a living digital twins. The digital transformation is not a purely technical matter that can be solved. Technology can only solve a small aspect, if the lack of good strategic and operational planning, will lead to various problems.