The next generation of smart factory comes with three core technology decryption

The central control is the installation of the robot arm, the glue is the automatic operation of the robot, the front windshield is the installation of the robot, the tire is the automatic assembly, the tool spare parts are unmanned management, the frame is intelligent sorting, the digital twin technology is efficiently used, the digital flow of the equipment and the system is interconnected. …

This is not the imagination of the future smart factory , this is the real scene of FAW's smart factory.

In recent years, lighthouse factories, unmanned chemical plants, and industrial robot intelligent factories, representing the highest level of digitalization in the current manufacturing industry, have emerged in large numbers in China. A number of high-tech companies, including Xiaomi, have set up their own smart factory demonstration bases in Beijing.

Recently, Lin Fusheng, vice-chairman of the Beijing Municipal Committee of the Chinese People's Political Consultative Conference (CPPCC), gave a speech. Smart factories build the future” The theme of the 2024 Zhongguancun Forum said that since 2021, a total of 103 smart factories and digital workshops have been identified in Beijing, and 47 enterprises have been selected into the list of intelligent manufacturing demonstration factories of the Ministry of Industry and Information Technology, that is, the list of excellent scenes.

At a press conference held by The State Council Information Office on January 19, Tao Qing, spokesman for the Ministry of Industry and Information Technology and director of the Operation Monitoring and Coordination Bureau, said that by the end of December 2023, the Ministry of Industry and Information Technology had cultivated 421 national demonstration factories and more than 10,000 provincial digital workshops and smart factories.

Three core technologies of smart factory

Li Peigen, academician of the Chinese Academy of Engineering, said at the forum that the construction of smart factories is an important starting point for new industrialization enabled by AI, and its core lies in the deep penetration and application of key technologies such as intelligent robots, digital twins, and large models in industrial scenes.

Regarding intelligent robots, Li Peigen pointed out that robots in the future manufacturing industry can go out of the storage and manufacturing links and enter the open environment, which is an important direction in the future, and the realization of this scenario mainly relies on autonomous mobile technology (AMR). — Based on map positioning technology, it scans the working environment and updates the map independently, without the need for auxiliary fixed beacons, and has almost no need to transform the work site, which is very suitable for deployment in complex and dynamic production scenarios.

In the workshop of Makino Makino in Japan, robots using AMR technology can take care of multiple machine tools at the same time, loading, unloading, cutting tools and other operations are not a problem.

The scope of application of robots should not be limited to the enterprise, but should be able to cope with a variety of operating conditions, densely populated Spaces, and even other automation equipment. “ Out of the past controlled warehousing and manufacturing links, robots will bring development opportunities to all walks of life. ” Lee Bacon said. For example, humanoid robots have a high degree of automation and are suitable for a variety of complex environments, which is an important exploration direction for the construction of smart factories. Tesla plans to be the first to introduce humanoid robots in the assembly process of cars. Mercedes is working with robot developer Apptranik to bring humanoid robots to factories to assemble parts.

“ With human-like capabilities, robots can cope with complex human environments without modifying the environment, and use a wider range of scenarios, which can not only replace simple, repetitive and dangerous labor, but also meet emotional needs and interactions while adapting to non-standard service scenarios. ” Li Peigen said. In addition, the robot also has a learning function, the more it is used, the smarter it is, and the factory can also realize the skills inheritance of new and old employees through robots.

Talk about digital twins. Li Peigen stressed that smart factories should be factories that fully apply digital twin technology. The digital twin is not only the digital twin of products and equipment, but also the digital twin of workshops, factories, and supply chains. For example, Huazhong Numerical Control uses digital twin technology in CNC machine tools to control machine equipment in real time. Through the analysis and simulation of the twin data collected from the machine tool, the user can understand the behavior of the machine tool more clearly, so as to carry out more accurate control and operation.

“ The digital twin factory should not only be some simple digital record and digital presentation of the physical factory, but should be able to collect real-time data through the digital twin system, and in turn guide the operation of the workshop or the factory, and even cope with supply chain competition, so that the factory can run in the best state. ” Lee Bacon said.

In fact, what the workshop needs is a digital twin factory platform. The digital system records the operation of the real system in real time, analyzes which parameters need to be adjusted, and ensures the quality and effect through real-time adjustment. The supply chain digital twin is reflected in the enterprise, the enterprise should not only have a good supply chain system, but also have a digital supply chain. And unlike traditional, offline, independent simulations that have long been used in supply chains, digital twin-based simulations are real-time, dynamic, and can interact with physical systems.

The third important aspect of smart factories is the AI grand model. Li Peigen believes that the smart factory of the future requires insight into some complex and high-level relationships. The understanding of the high-level relevance of AI models in the world has far exceeded that of humans, and its application in the construction process of smart factories should be paid attention to.

A large model is a machine learning model with a large number of parameters and a complex structure. Its Transformer architecture has two important features: one is vectorization of words, like ChatGPT, which can extract 300 to 500 attributes at a time to describe “ Apple ” This one word, different types, different regions of the apple it is very clear, this is difficult for human beings to do; The other is self-attention, emphasizing the search for relationships. After reading a book of nearly 2,000 pages, ChatGPT5 can grasp all the relationships between words from beginning to end, which is difficult for human cognition to shoulder.

Back to the industrial manufacturing scenario, the electronically controlled big data inside the CNC system is the main data source. The big data inside the CNC system includes the state of the moving axis, the state of the spindle, the running state of the machine tool, the operating state of the airport, the program state, etc. If these data are collected and analyzed, it will be found that there is a complex correlation between the quality, accuracy and processing efficiency of the parts, and the corresponding measures can significantly improve the production quality.

“ If we trace the machine tool accuracy, assembly error, thermal and dynamic error, machine tool life cycle error control and other issues, there may be countless complex relationships that we need to consider, such as machine tool parts, parts, accuracy, motion parameters, speed, etc., these relationships are difficult to consider in the manufacturing process. ” Li Peigen admitted. However, the emergence of large models may help smart factories gain insight into more complex higher-order relationships.

At the forum, Yu Xiaohui, president of the China Academy of Information and Communications Technology, released the "Thinking of the Next Generation of smart factories" and introduced the eight reform models leading the construction of the next generation of smart factories. It includes intelligent research and development mode based on digital twin, large-scale and efficient independent manufacturing mode, adaptive and reconfigurable flexible manufacturing mode, net zero emission sustainable manufacturing mode, and ultra-conventional limit manufacturing mode.

In the context of new industrialization, in China's intelligent manufacturing strategy, the exploration of the next generation smart factory model will become the focus of attention and occupation by the outside world.

How intelligent applications enable equipment manufacturing

Joe, Vice President and Head of Industrial Automation China End-User Business at Schneider Electric, explained how Schneider Electric empowers industrial customers for the full life cycle of industrial applications.

As an automation and energy management company, Schneider Electric has refined a complete smart factory construction path to help users establish a dual base of data: static data and dynamic data at the lowest level of the enterprise digital factory. On top of the two data centers, help users deploy advanced applications to solve a variety of different business problems, which can also be understood as apps on the phone. Such as energy management module, manufacturing execution system, production operation system, equipment management platform, personnel management platform, can also provide users with more visualization, convenient access to applications, such as automatic alarm, automatic generation of management reports, automatic distribution of work orders, and enterprise ERP docking, to achieve seamless connection between production and operation and enterprise management.

Liu Wenrui, vice president of BOE Group and director of LCD Manufacturing Center of Display device and Internet of Things innovation business, introduced his BOE smart factory construction experience.

As a semiconductor display enterprise, BOE has two intelligent four-level smart factories assessed by the Institute of Electronic Technology Standards, as well as the lighthouse factory recognized by the World Davos Economic Forum, from product design, production and manufacturing, process monitoring, logistics, warehousing the whole chain to achieve intelligent production.

First dynamic automatic scheduling system. Semiconductor display process is extremely complex, involving dozens of production processes, as well as hundreds of equipment production at the same time, based on scheduling logic and algorithm, through genetic algorithm every 15 minutes iteration, combined with real-time production on the production line and equipment conditions for dynamic scheduling.

Secondly, quality control. Through the coverage of all models and processes, intelligent monitoring of actions, combined with historical experience to build a digital model, combined with real-time production data, to predict possible deviations in advance, dynamic adjustment, to achieve an improved closed loop. Finally, energy management. The semiconductor shows that the annual power consumption is very large, taking the 8.5 line as an example, the annual power consumption is about 800 million degrees. Boe arranges tens of thousands of energy acquisition sensors on thousands of devices, dynamically analyzes energy consumption problems on the production line through data upload and model establishment, accurately confirms the equipment through data analysis, and sends it to the field engineers for intervention. It reduces energy consumption per unit of production by 39%, and also enables automated models to automatically look for improvement opportunities and continuously reduce the plant's energy use.

At the same time, based on years of independent research and development capabilities and production and intelligent transformation experience, BOE has built an industrial Internet platform covering the entire process of production and manufacturing, which has certain promotion and application and peer reference value, and helps to promote the intelligent upgrade of China's high-end display module manufacturing industry.

AI exploration in the field of digital twin space

Beijing Satellite Factory, as the main research and development unit of China's artificial satellite, is the birthplace of China's first artificial satellite, the first communication satellite, the first manned spacecraft, the first lunar probe, the first Mars probe and the first space station. Han Jianchao, deputy general manager of Beijing Satellite Manufacturing Co., LTD., introduced the achievements of spacecraft intelligent factory construction at the forum.

Beijing Satellite Factory has established a set of homologous data model, which runs through the whole cycle of the development of the space station, realizes the digital twin and the integration of heaven and earth, and constructs a model covering four major specialties, nine analysis systems and more than 700 high-execution equipment.