Smart Manufacturing Technology for 2035

Source: Chinese Journal of Mechanical Engineering

In recent years, my country has carried out a lot of research work on intelligent manufacturing technology and its application. Facing the emerging new technologies, new ideas and new models of intelligent manufacturing, in order to better promote the orderly development of follow-up intelligent manufacturing technology related research, intelligent The six directions of intelligent products in manufacturing, discrete manufacturing, process-based manufacturing, new models and new formats, industrial intelligent networking, and intelligent manufacturing cloud, respectively, through technical system construction and technical situation scanning, technical list formulation, questionnaire surveys and expert discussions and other methods and processes, draw a roadmap for intelligent manufacturing technology for 2035 from the three levels of target layer, implementation layer, and guarantee layer, and point out the list of key technologies that smart manufacturing needs to concentrate on at present, with a view to laying a foundation for the future development of intelligent manufacturing technology. It provides references for policy formulation, value evaluation, competitiveness evaluation, strategic management and scientific research.

foreword

As a technology management method that explores the dynamic relationship between organizational goals, technical resources and changing environments, the technology roadmap reflects the development trend of future technology with its unique framework and organizational form and through the mechanism of exchange and integration of ideas. In the late 1970s, the technology roadmap was first applied to product technology planning. Driven by the formulation and practice of relevant industrial technology roadmaps, the technology roadmap has been widely used at the enterprise and industry levels, and gradually expanded to the national science and technology strategy planning level. In the complex and ever-changing technology competition environment, many countries use technology roadmaps to plan the development path of technology. For example, South Korea proposes strategic products and key technologies for the concept of technological development, and Japan has compiled strategic technologies in several key areas. route map. China's scientific research institutions have carried out a series of studies on technology roadmaps, representative ones are the national technology roadmap research carried out by the Ministry of Science and Technology in 2007, the Chinese Academy of Sciences' strategic research on China's scientific and technological development roadmaps in important areas until 2050, and The Chinese Academy of Engineering and the National Natural Science Foundation of China jointly conducted a research on the 2035 development strategy of China's engineering science and technology.

At present, the intelligent manufacturing technology formed by the deep integration of the new generation of information technology and advanced manufacturing technology, especially the new generation of artificial intelligence technology and advanced manufacturing technology, has become the core of the fourth industrial revolution. technology and core drivers. Intelligent manufacturing is leading and promoting the fourth industrial revolution, triggering major and profound changes in manufacturing development concepts and manufacturing models, reshaping manufacturing technology systems, production models, development elements and value chains, and promoting China's manufacturing industry to gain competition New advantages will push the development of the global manufacturing industry into a new stage and achieve an overall leap in social productivity. In recent years, my country has carried out a lot of research work on the relevant theories and application technologies of intelligent manufacturing. Facing the emerging new technologies, new ideas, and new models of intelligent manufacturing, in order to carry out orderly research on intelligent manufacturing technology, it is urgent to study The development trend of intelligent manufacturing in 2035 and the key technologies that need to be concentrated to overcome, and prepare a technology development roadmap.

1. Research method and process of intelligent manufacturing technology roadmap

The research team of the Chinese Academy of Engineering continued to promote the construction of the theoretical system of intelligent manufacturing in the "Research on the Strategy of Manufacturing Power", proposed the development strategy of China's intelligent manufacturing, and launched the research work of "Foresight and Roadmap of Intelligent Manufacturing Technology for 2035" in 2019. According to the previous research results of this research team, the intelligent manufacturing system is composed of three functional systems of intelligent products, intelligent production and intelligent services, and two supporting systems of industrial intelligent networking and intelligent manufacturing cloud. Among them, intelligent products are the main body, and intelligent production is the main line, the transformation of the industrial model centered on intelligent services is the theme, and the intelligent manufacturing cloud and industrial intelligent networking are the supports, as shown in Figure 1. Therefore, the research on the technology forecast and technology roadmap of intelligent manufacturing also closely revolves around these aspects, but they are not exactly the same. According to the distribution of expert research fields, the roadmap work is divided into six sub-directions: intelligent products, discrete manufacturing, process-based manufacturing, new models and new formats, industrial intelligent networking, and intelligent manufacturing cloud.

Figure 1. Intelligent manufacturing system integration

In the process of research and preparation of the intelligent manufacturing technology roadmap, the research team combined general methods and specific practical experience, adopted a "qualitative and quantitative" roadmap research and drawing method, with experts as the core, process as the norm, data as the support, Use interaction as a means to improve the forward-looking, scientific and normative nature of the roadmap. The roadmap research work is led by more than ten academicians including Li Peigen, Tan Jianrong, Chai Tianyou, Lu Bingheng, and Li Bohu. More than 1,600 experts, scholars, and engineers, including experts from the United States, Germany, and Japan, are organized to conduct research on the latest developments in domestic and foreign technologies and Analysis of trends, as well as research work on forecasts and roadmaps for the development of smart manufacturing-related technologies. Researchers in the six directions of intelligent manufacturing rely on the intelligent decision-making system platform (Intelligent Supporting System, referred to as iSS, website: iss.ckcest.cn) of engineering technology strategy consulting, according to the technical system and situation analysis, technical list, questionnaire survey and expert discussion, The technology roadmap draws four steps to carry out research work, as shown in Figure 2.

Figure 2. Smart manufacturing technology roadmap drawing process

In the process of drawing the intelligent manufacturing technology roadmap, the data analysis results interact with experts for multiple rounds. On the one hand, the data is used to support the expert's research and judgment on the problem, and on the other hand, the experts are guided to carry out the intelligent manufacturing technology roadmap drawing work according to the standardized process. . At the same time, the data analysis results were revised according to expert opinions.

1.1 The first step: technical system and technical situation analysis

(1) Technology system construction: Firstly, build a multi-level technology structure in 6 directions to form a technology system for each direction, which is used to describe the relationship between technologies in various fields of intelligent manufacturing, sort out the technical context, and divide the research boundaries. As a visual form reflecting expert knowledge and consensus, the technical system can guide the technical situation scanning of objective data.

Figure 3 is an example of the industrial Internet technology system, in which the industrial Internet is regarded as the first-level technology, network intelligence technology, network connection technology, network security technology, network identification analysis technology, etc. , identification coding technology, industrial network intelligent management system, etc. as the third-level technology.

Figure 3. Example of industrial Internet technology system

(2) Technical situation scanning: According to the various technical contents in the technical system, determine the search keywords, as shown in Table 1, which is the keyword table determined for the process intelligent manufacturing factory. The first layer corresponds to the second-level technology of the technology system, the second layer corresponds to the third-level technology of the technology system, and the third layer is keywords for searching data such as papers and patents. After the keywords are determined, each direction obtains objective data in the field of intelligent manufacturing such as papers, patents, and research reports through the database, and analyzes the field of intelligent manufacturing from multiple dimensions such as the world, country, researchers, and research topics.

Table 1. Search keywords for process manufacturing smart factory

In the process of building the entire technical system and scanning the technical situation, experts determine the technical system, researchers determine the search keywords (or data search formula) according to the technical system, and obtain data; after analyzing the data using the iSS platform, the experts propose key Word modification opinions, multiple rounds of interaction between the analysis results and experts, and correction and iteration of the data analysis results, and finally complete the scanning of the intelligent manufacturing technology situation in six directions. Figure 4 is an example of the situation analysis of smart manufacturing technology.

Figure 4. Example of smart manufacturing technology situation analysis

The results of the smart manufacturing technology situation scan can support experts to clarify the past and current macro development trends in the field of smart manufacturing from a quantitative perspective. The introduction of objective data helps to reduce the preference in the analysis of the development direction of intelligent manufacturing technology, and helps experts form a more consistent understanding of the research background. At the same time, integrating expert opinions into the process of data analysis in an iterative and interactive manner can improve the accuracy of smart manufacturing technology situation scanning.

1.2 Step 2: Technical Checklist

Based on the results of intelligent manufacturing technology situation scanning, cluster analysis and natural language processing methods were used in the six directions to excavate core research topics in the field of intelligent manufacturing. After manual sorting, several key technical items were summarized to form an initial technology list. Then search and screen the future-oriented key technology items in the field of intelligent manufacturing carried out in various countries and regions to supplement the initial technology list and obtain the candidate technology list. Finally, three rounds of expert seminars were held. In the first round of research, experts supplemented the list of candidate technologies and added technical items missed in the analysis; in the second round of seminars, technical items that were inappropriate or too small in granularity were deleted , Merge technical items with similar content; in the third round of seminars, experts adjusted the granularity of technical items in the list to keep the granularity of all technical items in the list basically the same, and wrote the scope and connotation of each technical item, and finally formed 6 A list of key technologies for smart manufacturing oriented towards 2035. Table 2 is an example of a list of key technologies for intelligent manufacturing. The first column is the technology number, the second column is the selected key technology, and the third column is the technical category and connotation, which mainly describes the main content and boundaries of the technology, and the main functions realized and application scenarios, etc.

Table 2. Example of a list of key technologies for smart manufacturing

1.3 The third step: questionnaire survey and expert discussion

Based on the selected list of key technologies in the field of intelligent manufacturing, questionnaires or seminars are widely distributed to universities, scientific research institutes, and enterprise experts in the field of intelligent manufacturing around the world, focusing on the importance, core, driving, subversive, and maturity of technologies Solicit expert opinions in terms of degree, leading countries, and technology realization time, summarize and analyze expert opinions, and sort out and determine important milestones in technology development.

Based on the results of two rounds of questionnaire surveys and expert discussions, the single-factor indicator score is calculated based on the experts' familiarity with the technical field and the number of votes. Table 3 is an example of calculating the indicators of this questionnaire. Sorted according to the scores, 6 directions propose key technologies in this field, forming a pool of 86 key technologies. Focusing on the key technology pool, two rounds of 6 research group expert collective discussions were carried out, and the list of key technologies was re-selected and merged, finally forming a list of 27 smart manufacturing technologies for 2035.

Table 3. Calculation example of questionnaire indicators for intelligent manufacturing technology roadmap

1.4 Step 4: Draw a technology roadmap

The intelligent manufacturing technology roadmap mainly includes the target layer, implementation layer, guarantee layer and other layers. The target layer is composed of experts in 6 directions according to the intelligence

Discussion and formulation of manufacturing development vision, future economic and social needs, and strategic goals and tasks in the field. The implementation layer is mainly drawn based on the selected key technology list and important time nodes collected through expert opinions. The security layer is discussed by experts based on the needs of policies, talents, and financial support.

On the basis of drawing the first draft of the roadmap, academicians from six directions will lead the organization to adjust the items of the target layer, implementation layer, and guarantee layer in the roadmap, such as adding, deleting, and time adjustment, with expert opinions as the core. After several rounds of expert discussions, the experts' opinions were basically converged, and after a consensus was reached, the final technical roadmap was determined. Figure 5 shows the basic framework of this roadmap. The roadmap has four basic elements: based on time series; layered presentation; clear milestone nodes; and connections between layers.

Figure 5. Schematic diagram of the basic framework of the smart manufacturing technology roadmap

2. List of smart manufacturing technologies

As described in Section 1.2 and Section 1.3, after technical situation scanning, manual collation, questionnaire survey and three rounds of expert discussions, etc.

Finally, a list of key technologies for intelligent manufacturing facing 2035 in six directions was formed, as shown in Table 4.

Table 4 List of smart manufacturing technologies for 2035 (2020 version)

2.1 Knowledge base for product design and process

The knowledge base for product design and process involves all aspects of product life cycle knowledge, including data and knowledge on materials, structure, design, manufacturing, service (sales, maintenance) and scrapping. The main content and core sub-techniques include knowledge acquisition technology, knowledge management technology, knowledge application technology and so on. The knowledge base oriented to product design and process can realize full digital design of products, simulation and simulation optimization of structure, performance and function in parallel and collaboratively in a virtual digital environment.

2.2 Data acquisition, processing and analysis technology

Data collection, processing and analysis technology is a technology for real-time and efficient collection of user data in scenarios of various access terminals and various network types. It mainly solves the integration and access of heterogeneous communication protocol data sources, the unification of real-time data interfaces, the fusion of multi-source heterogeneous data, the massive storage of real-time data, the efficient query of real-time data read and write operations and historical data, data quality evaluation and cleaning, Real-time computing and analytical processing, organization of real-time data, and access rights management.

2.3 Distributed intelligent control technology

Distributed intelligent control technology is a combination of artificial intelligence and distributed computing. It is mainly used in large-scale areas, multi-heterogeneous platform collaborative operations, and multiple intelligent machines (high-reliability intelligent machines) working together. Study behavior coordination and work task coordination between different agents, and each agent has its own goals and wishes. Through distributed artificial intelligence, the multi-objective solution problem of a complex system is divided into sub-problems with relatively low complexity layer by layer, and then completed by different agents through communication, collaboration and independent decision-making, which can overcome the lack of resources and capabilities of a single intelligent machine and the Single function and other limitations. It is necessary to focus on breakthroughs in the distributed control architecture of cluster machines in the cloud computing environment. On this basis, research on the integration of real-time resource scheduling and control methods for edge controllers, task-oriented semantic programming and automatic generation mechanisms, and fast and high-precision collaboration Observational models of multi-intelligent machine systems, and task assignment, coordination mechanisms, and distributed control of multi-intelligent machines.

2.4 Human-machine fusion robot

A human-machine fusion robot is an intelligent robot that combines the advantages of a living system with those of an electromechanical system. Through the research on the technology and method of deep integration of life system and electromechanical system, the regulation mechanism of bio-electromechanical system integration and related performance optimization models, the formation of intelligent biological functions based on the integration of life system and electromechanical system for new perception, drive and energy supply Device unit, and realize a new generation of human-machine fusion robot through system integration. The development of many disciplines such as micro-electromechanical systems, micro-nano processing technology, and life sciences are promoting research in the field of life-like robots.

2.5 Smart Sensor Technology

The development direction of smart sensor technology includes multi-source sensor fusion technology and bionic sensor technology. Multi-source sensor fusion technology refers to the use of multi-sensor information resources in different time and space to automatically analyze, synthesize, dominate and use the observation information obtained in time series under certain criteria, and obtain consistent interpretation and accuracy of the measured object. Described to accomplish the required decisions and tasks so that the system achieves superior performance over its individual components. Its main research contents include data association, multi-sensor ID/trajectory estimation, collection management, etc. Biomimetic sensors are new sensors composed of immobilized cells, enzymes or other biologically active substances and transducers. A new type of perception technology developed by the mutual penetration of technology and engineering.

2.6 Enterprise intelligent decision-making system

Enterprise intelligent decision-making system includes enterprise strategy intelligent decision-making system, product map intelligent decision-making system, supply chain management intelligent decision-making system and process selection intelligent decision-making system to achieve integrated optimization decision-making of enterprise goals, plan scheduling, operation indicators, production instructions and control instructions . The enterprise strategic intelligent decision-making system analyzes and makes decisions on the enterprise's competitive advantages, technological innovation system, innovation performance, environmental uncertainty, industry and technological development trends. The product map intelligent decision-making system conducts a full-process and multi-factor analysis of product life cycle and competitive advantages, conducts product grouping, development map, and realizes path planning to maximize the value of products. The supply chain management intelligent decision-making system analyzes all elements of the supply chain to realize intelligent supply chain management with high efficiency and zero inventory. The process selection intelligent decision-making system carries out product design and process flow intelligent planning, and makes strategic selection of product manufacturing mode.

2.7 Intelligent CNC machining technology and equipment

Intelligent NC machining technology includes human, computer, and machine integration theory and technology; multi-source information perception theory and technology; thermal deformation traceability, temperature field theory, sensor layout and compensation technology; geometric error modeling and compensation technology; vibration construction Model and suppression technology tool processing model and processing state perception technology; on-machine quality inspection method technology; workpiece processing progress extraction technology based on CNC system; fault online identification theory and technology; energy flow model and energy efficiency detection technology in processing process; intelligent decision theory and technology; intelligent execution theory and technology; intelligent maintenance theory and technology; intelligent machine tool comprehensive capability evaluation theory and technology, etc. Intelligent CNC machining equipment, such as intelligent CNC machining centers and intelligent machine tools, enhances the perception of processing status on the basis of digital control technology, realizes the interconnection between equipment through network technology, and applies big data and artificial intelligence technology. Capabilities of self-perception, self-analysis, self-adaptation, self-maintenance, and learning can realize functions such as processing optimization, real-time compensation, intelligent measurement, remote monitoring, and diagnosis.

2.8 Additive manufacturing technology and equipment

Additive manufacturing technology and equipment include metal additive manufacturing technology and equipment, functionally graded materials and structural additive manufacturing technology and equipment,

Bio-additive manufacturing technology and equipment, as well as process directions such as additive, subtractive, and other material-integrated intelligent hybrid manufacturing technologies. Metal Additive Manufacturing Technology and Equipment Through modeling and simulation of energy beam generating devices and multiple energy fields, the precise control of the coupling effects of multiple energy fields on the performance of component preparation, the prevention of deformation and cracking, and the precise control of shape during component manufacturing At the same time, it can realize the shape-controlled manufacturing of large and complex components including amorphous alloys, high-entropy alloys and other special performance materials. Functionally graded materials and structural additive manufacturing technology and equipment can carry out 3D printing of mixed materials such as metals, non-metals, composite materials, ceramics, etc., realize the combination of multi-category additive manufacturing processes, and the combination of additive manufacturing and other manufacturing processes, and can Different materials are used in different parts of the components, and the gradual transition of different materials is used to achieve complementary advantages or unique combinations of material properties, so that the components have extraordinary performance. Bio-additive manufacturing technology and equipment can realize the matching design of inactive devices and bioactive tissues and organs; through the intelligent control of multi-scale, multi-material, multi-cell, and multi-tissue fluid channels, the inactive devices and bioactive tissues can be realized. Precise printing of organs; combined with micro-nano biosensing and neuron regeneration to achieve integration with host tissue and nervous system.

2.9 Discrete Smart Factory

Discrete smart factories can carry out intelligent selection of product design and manufacturing according to product performance requirements, and realize personalized customization and flexible manufacturing mixed-flow production. Through the design and simulation software, the bionic, generative and topology optimization design of the product is realized. Through intelligent manufacturing equipment, the whole process simulation and process parameter decision-making of the manufacturing process can be carried out to realize the selection of manufacturing mode and multi-category and multi-mode hybrid processing and manufacturing, and realize the comprehensive improvement of product performance and manufacturing efficiency. Based on the collaborative optimization technology of the whole process information, it realizes the real-time control and collaborative optimization of the whole process of the factory including R&D design, process and equipment, logistics, quality, warehousing, sales and so on. Discrete smart factories focus on the production needs of 3C products in small batches, multiple varieties, and rapid iteration; for aviation, aerospace, and ship parts with super-large, complex structures, lightweight, and high-quality production needs; for large gas turbines and electric propulsion engines Manufacturing requirements for high-performance engines such as automotive products; multi-system, multi-component and personalized customization requirements for automotive products.

2.10 Intelligent modeling and simulation technology

Intelligent modeling and simulation automatically analyzes and synthesizes information and data from multi-sensors and multi-scales under certain criteria, integrates heterogeneous data and structural data, and combines mechanism models and data models to realize The integrated modeling of multi-level, multi-scale and multi-field coupling in the whole process assembles simulation model software in different fields into a comprehensive simulation software system with multiple functions through unified interface, software bus, data sharing or network technologies. When simulating a large-scale complex system, it is possible to use a coordinated structure, standard and protocol, and use network equipment to interconnect simulation devices scattered in various places to form a comprehensive simulation environment.

2.11 Intelligent manufacturing standard system

The object of intelligent manufacturing standards is the standards required for the integration of information technology and manufacturing technology. The research on the standard system includes the research on the standard system architecture, and clarifies the scope and description dimensions of the standard system. According to the framework, it is expanded into a standard system structure, and the standard classification, hierarchical structure and main development direction of standards in the standard system are clarified. The standard system construction of smart factories, including basic common standards such as foundation, safety, management, testing and evaluation and reliability, and key technical standards such as smart equipment, smart factories, smart services, industrial software and big data, and industrial Internet, such as manufacturing process standards , data standards, communication protocols and standards, technical application standards.

2.12 Intelligent optimization decision-making technology and system

Intelligent optimization decision-making technology and system is a technology and system that quickly selects the optimal solution that can achieve the goal from all feasible solutions. Intelligent optimization decision-making techniques include optimality conditions, convex optimization, linear optimization, unconstrained optimization solving methods, constrained optimization solving methods, dynamic programming, intelligent algorithms for solving optimization problems, decision theory, game theory, graph and network analysis, Queuing theory, storage theory, etc. The intelligent decision-making system adopts intelligent analysis technology of working condition protocol, multi-source heterogeneous data fusion technology, high-dimensional nonlinear strong coupling process statistical learning theory characterized by information depth perception, multi-quality index inverse mapping modeling method, and data-based knowledge learning With the rule extraction method, self-healing control and self-optimization are realized.

2.13 Process Smart Factory

The process intelligent factory aims to optimize the operation index, adaptively decides the set value of the control system, and realizes the optimal control and independent control of the operation index. It can predict and diagnose abnormal working conditions in time. When abnormal working conditions occur, through self-healing control, abnormal working conditions can be eliminated to achieve safe and optimized operation; the mechanism model and data model can be deeply integrated to establish an effective dynamic intelligent model to realize production equipment. Dynamic autonomous learning and data-driven autonomous control. Realize the whole process quality management and free flow of data. Focus on meeting the technical and system requirements of steel, petrochemical, mineral processing, nonferrous metals and other process smart factories.

2.14 Digital twin technology

In a digital environment, digital twins map physical entities such as humans, machines, and objects to form information virtual bodies. It "understands" the changes of the corresponding physical entities and responds to the changes through the real-time data from the physical entities. With the help of information space's ability to comprehensively analyze and process data, it can respond to changes in the external complex environment, make effective decisions, and act on physical entities. In this process, physical entities and information virtual bodies are interactively linked, virtual and real are mapped, and through data fusion analysis, decision-making iterative optimization and other means, the continuous optimization of manufacturing activities is realized, and a new space-time dimension is provided for manufacturing activities.

2.15 Equipment Health Assessment and Fault Prediction Technology

Equipment health assessment and failure prediction Through failure mechanism analysis, damage evolution modeling, and decline analysis and prediction technologies, a theoretical model for life-cycle design and predictive maintenance based on failure mechanisms is established. Based on technologies such as equipment operation data, data mining, feature learning, information sharing, security and privacy protection, and integrating equipment principles, expert knowledge and data models, strong correlation is made for early weak failures or extremely weak abnormal information of basic components of equipment Effective separation of fault features, enhancement and extraction of early weak fault features, multi-dimensional space feature mapping and extraction, so as to effectively identify early weak faults and composite faults, promote technological progress in the fields of remote monitoring, diagnosis, health management and predictive maintenance, and provide equipment maintenance predictive recommendations.

2.16 Shared manufacturing

Shared manufacturing can be specifically divided into: ① manufacturing capacity sharing. Focus on the shared innovation of processing and manufacturing capabilities, focus on the development of a shared platform that gathers manufacturing resources such as production equipment, special tools, and production lines, develop shared manufacturing services for multi-factory collaboration, develop shared factories that gather the common manufacturing needs of small and medium-sized enterprises, and develop rent-based sales , On-demand equipment sharing services; ② Sharing of innovative capabilities. Focusing on the flexible, diverse and low-cost innovation needs of small and medium-sized enterprises and start-ups, develop product design and development capability sharing that gathers diverse social intellectual resources, and expand scientific research equipment and testing capability sharing; ③ service capability sharing. Focusing on the common service needs of enterprises such as logistics warehousing, product testing, equipment maintenance, inspection and factory inspection, supply chain management, data storage and analysis, integrate massive social service resources, and explore and develop intensive, intelligent and personalized services Capability sharing.

2.17 Personalized scale customization

Personalized scale customization combines artificial intelligence and decision support systems. Through expert systems, decision support systems can more fully apply human knowledge, including descriptive knowledge of problems, procedural knowledge of decision-making processes, and reasoning for solving problems. Sexual knowledge, etc., help solve complex decision-making problems in personalized customization problems through logical reasoning.

2.18 Knowledge engineering and industrial knowledge software

The softwareization of industrial technology is the process of making explicit, modeling, digitizing, systematizing and intelligentizing experience and knowledge in industrial technology. Technologies and methods that promote the automatic use of knowledge by machines and the efficient use of knowledge by humans in various fields of industry. The softwareization of industrial technology includes platform technology and various related industrial APPs; its maturity reflects the depth and level of a country's integration of industrialization and informationization. Industrial technology objects that can be softwareized include: industrial products, the process of forming industrial products, abstract results of experience, various independent algorithm tools and knowledge contained in the process, and even a combination of multiple industrial APPs. The software of industrial technology is essentially knowledge engineering method and technology.

2.19 Intelligent Industrial Network

Intelligent industrial network is a network management system with intelligent functions such as network status self-awareness, network data visualization, fault self-location and self-recovery, network self-optimization, etc. It has the ability to detect network status, accurate diagnosis, dynamic optimization and remedy.

2.20 New Generation Mobile and Data Communication Technology

The new generation of mobile communication has a flexible and configurable new network architecture, which integrates evolution technologies such as large-scale antennas, new duplexing, and advanced coding, as well as innovative technologies such as new carriers and new dimensions. It has larger capacity, higher speed, and more diverse needs. Application scenarios and business needs and other characteristics. The network nodes of this network can realize functions such as high-performance routing and forwarding, virtualized management and efficient scheduling of network resources, dynamic binding and efficient resolution between various types of identifiers, and transmit IP-based Represents packet data and service information represented by content. For the new generation of mobile and data communication systems, 5G enhancement technology and 6G will be the focus of subsequent research. On the basis of continuously promoting the technological evolution of enhanced mobile broadband scenarios, 5G enhanced technology will focus on the standardization of key technologies and system designs for low-latency, high-reliability, and massive machine-type communication scenarios. Research a new generation of optical transmission technology for higher speed, large capacity and low delay, and actively promote new spectrum resource utilization technologies such as terahertz and visible light communication; second, promote fusion technologies with big data and artificial intelligence; third, It is necessary to explore the fusion architecture and key technologies of non-cellular networks such as satellite communications. At present, there are many technical directions for the future data network, showing a fragmented development trend.

2.21 Edge Intelligence Technology

Edge intelligence is a higher stage in the development of edge computing. Through the combination of edge computing and artificial intelligence, each edge computing node has the ability of deep computing and intelligent decision-making, and is deeply integrated with industrial applications. Edge intelligence is an open platform set up close to the source of objects or data. It integrates network communication, high-performance computing, large-capacity data storage and application core capabilities, and provides the nearest end service, resulting in faster network service response. And it can meet the requirements of real-time, security and privacy protection. The platform can also make full use of the massive field data and terminal computing capabilities on the edge side, cooperate with the large-scale simulation and calculation capabilities of the industrial cloud center, and efficiently realize the fusion analysis of simulation data, virtual-real interactive feedback and iterative decision-making optimization in industrial applications. Using the combination of artificial intelligence technology and edge computing to provide edge intelligence for intelligent data perception, semantic marking, and real-time processing based on industrial operating mechanisms is the key direction of future development. It mainly focuses on the supporting technologies related to edge computing and edge intelligence, including two categories, one is new information technology at the software, platform, and system levels applied to edge computing and edge intelligence; the other is network-level technologies that support mobile communications. New communication technology.

2.22 Identification analysis and management technology

Identification analysis and management technology is an important technology of the Industrial Internet. Logo resolution refers to the technical service that can query the network location or related information of the target object according to the logo code; the main technical paths are, one is the logo resolution technology that can operate independently from the Internet domain name system, and the other is the logo resolution that needs to rely on the operation of the Internet domain name system technology. Logo management refers to the supporting technical means necessary for logo-related registration, distribution, verification, and inspection.

2.23 Network Security Technology

Network security technology is an important technology of the industrial Internet, including the technology to prevent messages from being tampered with, deleted, replayed and forged, to enable the sent messages to be verified, and to enable the recipient or a third party to identify and confirm the authenticity of the message , and the technology of disguising information so that illegal accessors cannot understand the true meaning of information. By processing the collected data to judge the security status of the network, reflect the security change trend of the network and information system, do a good job in network security protection in advance, and reduce the potential losses caused by network security incidents. According to the system security situation and possible attacks, the network elements are dynamically reconfigured and changed, and the offense is the defense. By actively detecting the network security situation and attack situation, predicting the attack pattern, and improving the target system's security in the process of continuous self-learning defense level.

2.24 Semantic-based intelligent recognition technology

The key technologies of semantic-based intelligent recognition are natural language processing and semantic understanding. Natural language processing (NLP) research involves computer science, artificial intelligence, and linguistics, a field concerned with the interaction between computers and human (natural) language. Through the establishment of high-quality databases, understanding of common domains, scalable algorithm frameworks, and data-driven closed-loop processes, effective communication between humans and computers using natural language can be achieved to achieve semantic understanding, logically correct inferences, and specific The application of knowledge promotes understanding in application scenarios.

2.25 Hybrid Enhancement Technology

Hybrid augmentation technology is to introduce human functions or human cognitive models into artificial intelligence systems to form a form of "hybrid augmented intelligence". This form is a feasible and important growth model for artificial intelligence. Intelligence enhancement technology can enhance human-computer interaction, human-biological neural network interaction, and enhance the ability of the human brain to process, store and extract information. Deep reinforcement learning combines the representational ability of deep learning and the learning model of reinforcement learning, and uses reinforcement learning to drive agents to quickly explore various architectures, node types, connections, hyperparameter settings, and deep learning, machine learning and other AI models.

2.26 Human-cyber-physical systems (HCPS)

The intelligent manufacturing system is an intelligent system composed of people, information systems and physical systems, that is, the human-cyber-physical system (HCPS), in which the physical system is the executor and finisher of manufacturing activities; the information system is the link between the information flow of manufacturing activities. The core is to help human beings perceive, recognize, analyze, decide and control the physical system, so that the physical system can run as optimally as possible. Human is always the master of HCPS. Human is the creator of physical system and information system, and also the user and manager of physical system and information system. HCPS highlights the central position of human beings in intelligent systems, and puts more emphasis on the integration and synergy of the respective advantages of human intelligence and machine intelligence in intelligent systems.

2.27 Industrial E-commerce

Industrial e-commerce promotes the transformation and upgrading of traditional capabilities of enterprises in R&D and innovation, production control, supply chain management, operation control, financial control, and user services through the networking, collaboration, and intelligence of industrial enterprise transaction methods and business models. Help industrial enterprises to accelerate the cultivation of new capabilities based on precise identification and definition of needs, dynamic integration of resources, rapid delivery of products or services, and dynamic services throughout the life cycle, and form new models and new formats such as personalized customization, service transformation, and network collaboration .

3. Development roadmap of intelligent manufacturing technology

The technology roadmap in this study takes time as the main axis, and is oriented towards the medium and long term of 2035. It presents intelligent manufacturing technologies and products, discrete intelligent factories, process-based intelligent factories, new models of intelligent manufacturing, intelligent manufacturing clouds, industrial Development goals, demand trends, key technologies, key tasks, auxiliary support resources and other five aspects of the future development direction of the six major technical fields including Internet technology, as well as the main upgrade paths and key time nodes (Figure 6-11). Specifically, in terms of development goals, the vision for domain or cross-domain major projects is analyzed. In terms of demand trends, the major engineering technology needs for economic, social and industrial development are sorted out. In terms of key technologies, it pointed out the priority development topics in the field and the development direction across fields, identified the core technologies required by the field, and discovered the technology groups that need breakthroughs. In terms of key tasks, it clarified the major scientific and technological research projects, key scientific issues and important directions of basic research in the future. In terms of auxiliary support resources, the required policies, scientific research environment and guarantee conditions, as well as policy tools and management measures are proposed. Figure 6 shows the technology roadmap for smart products to 2035. Among them, in terms of demand, before 2035, it is mainly reflected in the following aspects.

(1) Intensified competition in the international information field requires high-performance computing technology, network security technology and sensory perception technology.

(2) The production of complex, high-performance, and high-precision components in key areas of the national economy requires efficient and intelligent processing technology and equipment.

(3) The transformation and development of the future society towards intelligent manufacturing and intelligent life will require a new generation of intelligent robots.

(4) Building an active control traffic system with intelligent vehicles, intelligent transportation facilities, and collaborative management services.

(5) The medical service model is gradually changing to personalization and intelligence, and there is a demand for building a digital, networked, and intelligent medical service information system. The overall goal of smart product development is divided into two steps.

(1) By 2025, the new generation of artificial intelligence system technology will be successfully applied in typical products, and the digitalization, networking, and intelligence of products will make significant progress. Key breakthroughs have been made in the application and manufacture of typical intelligent products such as intelligent electromechanical and intelligent medical products, as well as intelligent machine tools, intelligent robots, and intelligent forming equipment in the manufacturing field, which will play a demonstration role in the realization of the development of intelligent products.

(2) By 2035, realize the deep integration of the new generation of artificial intelligence technology and products, overcome a number of basic common technologies and key cutting-edge technologies related to intelligent products, master a number of internationally leading key core technologies, and develop in intelligent robots, intelligent machine tools, Intelligent forming equipment, intelligent construction machinery, unmanned aerial vehicles, intelligent networked vehicles, intelligent ships, intelligent rail transit and intelligent household appliances have formed typical advantageous products, forming a global competitive advantage in the relevant intelligent product industry chain, and the overall competitiveness Reach the level of a world power.

Figure 6. Smart product technology roadmap

The goals of the smart product focus area include the following. (1) By around 2025, a number of basic common technologies for intelligent manufacturing will be overcome, and some technologies will achieve original innovation breakthroughs. By around 2035, intelligent products and technologies will be fully applied in key industries and key enterprises.

(2) By about 2025, more than 60% of large enterprises or specialized enterprises in key industries will implement digital and intelligent manufacturing, and more than 70% of equipment will be digitalized and intelligent. By about 2035, 90% of key components and intelligent manufacturing equipment will be realized autonomous.

(3) By around 2025, a complete rail transit equipment key technology and major equipment system, supporting systems and equipment, key parts and basic parts manufacturing capacity will be significantly improved and widely promoted, and by around 2035, it will become the world's advanced robot The innovation center has become the world's largest robot application market, formed several intelligent robot industry clusters with international influence, and produced a number of original technologies in the industry.

(4) By about 2025, an independent, controllable and complete intelligent networked automobile industry chain and intelligent transportation system will be basically established, and it will enter the ranks of automobile powerhouses. By about 2035, the automobile manufacturing industry will be upgraded to an intelligent manufacturing system. The large-scale customized production of interconnection and cooperation, the integration of manufacturing service providers and service manufacturers, intelligent networked vehicles are more safe, reliable, energy-saving and environmentally friendly, comfortable and convenient.

(5) By around 2025, the green and intelligent level of mainstream ships will be internationally advanced, and the independent design and construction capabilities of high-tech ships will be fully mastered. By around 2035, a complete ship design, assembly and construction, equipment supply, and technical service industry will be formed. system and standard specification system.

(6) By around 2025, major breakthroughs will be made in new mobile medical care, new diagnosis and treatment, interventional therapy and wearable smart devices, digital medicine and human-machine interface technology, new biomaterials and nanobiotechnology, and a comprehensive Intelligent and integrated health care and health service system; comprehensively establish a nationwide medical information technology system, and form a world-leading modern health industry system around the health care and health information network.

(7) By around 2025, the intelligent manufacturing equipment industry will become a leading industry with international competitiveness, and 70% of the key components and manufacturing equipment required for intelligent manufacturing will be self-independent. By around 2035, the labor productivity of enterprises applying intelligent manufacturing will increase significantly. Improvement, material and energy consumption are significantly reduced, and product quality and consistency have reached the international leading level. The key tasks of intelligent products include knowledge base for product design and process, data acquisition and processing analysis technology, distributed intelligent control technology, human-machine fusion robot, intelligent sensor technology, etc., and its development roadmap is shown in the corresponding position in Figure 6 Each key task corresponds to several subtasks. Facing 2035, the strategic support and guarantee for the development of smart products includes the following contents.

(1) Integrate innovation system resources, build a national manufacturing intelligent innovation center, and promote the in-depth integration of industry, education and research.

(2) Improve the mode of funding input and give full play to the matching, unity and continuity of the implementation of national multi-sectoral policies.

(3) Strengthen the construction of a multi-level talent team, improve the treatment of innovative talents, and prevent brain drain. Figure 7 shows the discrete smart factory technology roadmap for 2035. In terms of demand, it will be mainly reflected in the following aspects before 2035.

(1) With the rapid development of ICT technology, digital technology, Internet, Internet of Things technology, artificial intelligence technology, big data technology, and virtual reality technology are developing rapidly.

(2) Consumers in the fields of national defense, industry, and civil use have diverse, personalized, high-quality, time-effective, low-cost, and service-oriented demands for products.

(3) Demands for the development of new manufacturing technology revolutions such as 3D printing, laser processing, micro-nano manufacturing, bio-manufacturing, robotics, and intelligent manufacturing.

(4) The needs of manufacturing enterprises and social transformation and upgrading, the needs of the development of green economy and service-oriented economy, the needs of a new round of technological revolution to promote industrial transformation and upgrading, and the development of new formats and new kinetic energy.

(5) The international environment in the new era requires competition from enterprises, and the new international and domestic dual-cycle development pattern requires manufacturing transformation. The development goals of discrete smart factories include the following.

(1) By around 2025, digital and networked manufacturing will be popularized and deeply applied across the country, and key technologies for smart manufacturing such as typical smart manufacturing equipment, industrial Internet and big data technology, and smart factory enabling technology will make breakthroughs and be successfully applied. By 2032 Around 2020, a new generation of intelligent manufacturing technology and intelligent factories will be widely applied in the manufacturing industry, realizing the transformation and upgrading of China's manufacturing industry.

(2) By around 2027, the new generation of smart manufacturing will be piloted and demonstrated in key areas and achieved remarkable results, and 10 iconic smart factories will be built, and they will be promoted and applied in some enterprises. By around 2035, discrete smart factories will help the manufacturing industry The overall level has reached the world's advanced level, and some fields are at the world's leading level. Facing 2035, the key tasks for the development of discrete intelligent factories include enterprise intelligent decision-making systems, intelligent numerical control processing technology and equipment, additive manufacturing technology and equipment, intelligent modeling and simulation technology, discrete intelligent factories, intelligent manufacturing standard systems, etc. , the corresponding subtasks and their roadmaps are shown in Figure 7. Facing 2035, the strategic support and guarantee for the development of discrete smart factories include the following.

(1) Combined with the actual situation of China's manufacturing industry to support the transformation and development of enterprises, quickly make up for the missing content of Industry 2.0 and 3.0, and actively explore the cutting-edge technologies of intelligent manufacturing in Industry 4.0. The two are organically connected and adhere to the principle of "parallel advancement and integrated development" technical route.

Figure 7. Discrete Smart Factory Technology Roadmap

(2) Support the continuous implementation of the 04 project. Intelligent manufacturing equipment is the basic support of the smart factory. The 04 project has achieved important results. It has achieved independent control in major national tasks, aerospace and other major fields, and is in the climbing stage. In desperate need of a smart upgrade.

(3) Take the lead in supporting typical enterprises with foundations in key fields to build demonstration smart factories, break through typical key technical problems, strive to be replicable and popularizable, and play a wide range of reference, radiation and leading role.

(4) Vigorously support and guide the establishment of a group of intelligent manufacturing professional technical service enterprises to form a regional manufacturing chain and intelligent manufacturing ecology.

(5) Encourage and support key engineering colleges and universities to set up undergraduate majors related to the direction of intelligent manufacturing, cultivate specialized technical and management talents in the field of intelligent manufacturing in a planned and batch manner, and accumulate reserve forces in the field of intelligent manufacturing in the future. New knowledge and skill requirements, planned training and training for on-the-job engineering and technical personnel.

(6) In the development of smart factories, it reflects the combination of "major key manufacturing equipment, industrial software, and standard specifications", highlights the "Made in China" of the above three, and guards against and prevents China's smart manufacturing from becoming a developed country's high-end equipment and industry in the future The dumping place of software prevents problems such as the "hollowing" of high-end equipment and core technologies and the manipulation of Chinese manufacturing by others. Figure 8 shows the technology roadmap for process manufacturing smart factories in 2035. In terms of demand, it will be mainly reflected in the following aspects before 2035.

Figure 8. Process Manufacturing Smart Factory Technology Roadmap

(1) The new generation of information technology represented by cloud computing, the Internet of Things, and big data is deeply integrated with modern manufacturing and producer services to promote industrial transformation and upgrading.

(2) Promote the intelligent optimization of the production process and the manufacturing mode characterized by the overall intelligent optimization of the entire production process.

(3) Realize the efficiency and greening of the overall situation of the enterprise and the whole process of production and operation. The goals of process manufacturing smart factories include the following.

(1) National goal: By 2025, the Made in China 2025 software technology standard and ecological system will be basically formed. By 2035, my country will be transformed from a large manufacturing country in the process industry into a manufacturing power. World leading level.

(2) Industry goals: By around 2025, establish an intelligent autonomous control system to realize intelligent perception of changes in production conditions, and by around 2035, establish an intelligent collaborative optimization control system for the entire manufacturing process, and build an intelligent optimization decision-making system. The roadmap of key products (key projects) includes: setting up a special topic of "intelligent optimized manufacturing in process industry" in the major project of "intelligent manufacturing and robotics", promoting the pilot demonstration of intelligent manufacturing in process industry, promoting the intelligent upgrading of major process industrial processes,) data and Information fusion and promotion to form an intelligent innovation system for the process industry. Key common technologies include intelligent optimization decision-making technology and systems, digital twin technology, and process smart factories. The corresponding subtasks and their roadmap are shown in Figure 8. The strategic support and guarantee for the development of smart factories in process manufacturing includes the following contents.

(1) Policy: It is recommended that relevant government departments organize a strategic research group composed of academics, R&D and enterprises to jointly carry out strategic planning and top-level design of intelligent optimized manufacturing in the process industry; Intelligent and optimized manufacturing; highlight the strategic position of intelligent manufacturing in the process industry, enhance the innovation capabilities of process industry enterprises; strengthen infrastructure construction, strengthen the dominant position of enterprise innovation, optimize the innovation environment of process-based smart factories, integrate basic and cutting-edge research, key research and development plans, and the Ministry of Industry and Information Technology The in-depth integration of informatization and industrialization promotes the overall deployment.

(2) Funds: Focus on supporting the investment in intelligent perception and remote operation and maintenance technology of major equipment in process manufacturing; increase capital investment in knowledge-based automation-based process industry intelligent optimization decision-making and collaborative control integration technology; Capital investment in product quality monitoring, traceability, diagnosis, forecasting and optimization technologies throughout the life cycle; strengthen financial support for CPS-based smart energy integrated management systems.

(3) Talents: Introduce talents with key technical professional qualities, focus on cultivating and creating practical engineering and technical talents oriented to industrial innovation needs, and cultivate application-oriented R&D talents with solid literacy. Figure 9 shows the technology foresight and roadmap of new formats and models of manufacturing industry for 2035. By 2035, the demand will come from the following aspects.

(1) The demand for the development of the manufacturing industry towards collaboration, customization, platformization and serviceization.

(2) Based on data, networked and intelligentized to support the demand for service-oriented manufacturing.

(3) Enterprises carry out personalized customization and mass production around the needs of customers.

(4) Collect existing resources, give full play to professional advantages, optimize resource allocation, flexible matching, dynamic sharing, and realize the needs of new models and new formats of customer value-added.

(5) Improve equipment health assessment and fault prediction capabilities, and build a health analysis and fault prediction platform covering the entire industry chain and the entire field. Facing 2035, the specific goals of the new business model and new model of the manufacturing industry include the following two aspects.

(1) National goal: By 2025, a pattern of leading special projects, demonstration by typical enterprises, and active exploration by enterprises will be formed. By 2030, the overall competitiveness of the manufacturing industry will be greatly improved, its position in the global industrial division of labor and value chain will be significantly improved, and new models and new formats will continue to grow. By 2035, the industrial base will be advanced, the industrial chain will be modernized, and the overall competitiveness of the manufacturing industry will reach the upper-middle level of the world's manufacturing powerhouses.

(2) Industry goals: Enterprises actively explore new formats and models, and form a development pattern led by special projects and demonstrated by typical enterprises (around 2024). The ratio of income from new business forms and new models to the main business income of enterprises will continue to increase, and the overall competitiveness will increase (around 2029). Become a new competitive advantage and an important source of profit (around 2032). The income from new models and new formats will account for 45% of the company's main business income (around 2035). The key tasks for the development of new formats and new models of manufacturing include equipment health assessment and failure prediction, shared manufacturing (collaboration and sharing), personalized scale customization, industrial e-commerce, knowledge engineering, and softwareization of industrial knowledge. The corresponding subtasks and their roadmap are shown in Figure 9. By 2035, the strategic support and guarantee for the development of new formats and new models of manufacturing industry will include the following aspects.

(1) Grasp the changes in trends, plan forward-looking layouts, and organically connect new projects with current special projects.

(2) Overall coordination, resource sharing, layered guidance, and collaborative advancement.

(3) Strengthen the guidance of industry classification, promote development in stages, with key points, and for a long time.

(4) Strengthen application demonstration, implement dynamic evaluation, and establish a normal mechanism for follow-up research, evaluation, and demonstration.

(5) Establish an innovation exchange platform and build a long-term cooperation mechanism.

(6) Optimizing the policy environment, building and improving a multi-channel and multi-level policy system and mechanism.

Figure 9. Roadmap for New Models and New Formats

Figure 10 shows the technology forecast and roadmap of intelligent manufacturing cloud for 2035. By 2035, the demand will mainly come from: through virtualization, service, container technology, cloud computing, big data analysis and mining technology and other technologies, the Massive multi-source heterogeneous data is integrated into the cloud platform, and based on the big data engine service, artificial intelligence engine service, and simulation engine service provided by the cloud platform, it realizes modeling and optimization based on big data, thereby supporting cloud application service functions

Figure 10. Smart manufacturing cloud technology roadmap

Facing 2035, the specific goals of the intelligent manufacturing cloud include two aspects.

(1) 智能制造过程横向集成：到 2025 年，实现论证／设计／仿真／生产／试验／管理／销售／运营／维修／报废等制造全生命周期活动的集成；到2035 年，消除企业各环节或者产业链上各企业之间的交互的冗余和非增值过程，实现企业与企业、企业与产品之间的生态化协作，在社会范围内实现人、技术流、管理流、数据流、物流、资金流的共享集成和优化应用。

(2) 智能制造过程纵向集成：到 2025 年左右，实现智能机床、CAE 软件等智能集成制造装备的互联、集成和管控的装备／产品集成装备与生产线、仓储与物流、制造执行管理、车间生产决策等环节的车间集成；到 2035 年左右，实现智能设计、智能试验、智能生产、智能保障、智能管理等跨部门企业/工厂集成及同行业内各企业的设计研发、生产制造、服务保障等资源的集成和优化配置的行业集成。智能制造云的相关重点产品包括：重要硬件设备、关键引擎、HCPS 集成系统。关键技术包括基于语义的智能识别技术、边缘智能技术、混合增强技术等。 图10 中展示了对应的重点产品和关键技术的实施路线图。同时，图 10 还包括了示范项目的实施路线图。

图11·工业互联网技术路线图

智能制造云的战略支撑与保障如下所述。

(1) 支持行业产业联盟体系：到 2025 年，倡导各典型行业龙头企业牵头成立行业级智能集成制造系统产业联盟；到 2030 年，行业联盟建立行业级标准体系。

(2) 支持跨学科领域产业联盟：到 2030 年，倡导以行业龙头企业为主要参与单位的跨行业智能集成制造系统联盟，开展跨行业领域数据标准、数据接口标准体系研究与建立。

(3) 打造自主行业／通用标准体系，参与国际标准体系构建。到 2035 年，支持各行业智能集成制造系统产业联盟与跨行业智能集成制造系统联盟，建立自主行业／通用标准体系，参与国际自主行业/通用标准体系构建。图 11 展示了面向 2035 年的工业互联网的技术预见及路线图，预计到 2022 年左右，工业互联网的产业经济规模将达 3.1 万亿元，到 2035 年左右，工业互联网的产业经济规模全球比重将显著提高。面向 2035 年，工业互联网的具体目标包括以下两方面内容。

(1) 中国目标：到 2025 年，网络实现基础设施基本完善；到 2030 年，全面建成低时延高可靠广覆盖网络，到 2035 年，建成领先网络基础设施和平台。

(2) 全球目标：到 2035 年左右，全球经济价值10～15 万亿美元。工业互联网的关键技术包括：新一代移动和数据通信技术、智能工业网络、标识解析与管理技术、网络安全技术等。重点产品包括新一代移动和数据通信技术、智能工业网络、标识解析与管理技术、基于自治愈的网等。图 11 中展示了对应的关键技术和重点产品的重要时间点和实施路线图。工业互联网的战略支撑与保障如下所述。

(1) 政策：完善协同推进体系，成立工业互联网专项工作组；壮大工业互联网产业联盟等产业组织，联合产业各方开展交流活动；开展工业智联网网络安全相关法规政策的研究；开展工业互联网相关法律、行政法规和规章立法、完善工作；推动非金融企业债务融资工具应用，支持保险公司开发 产品。

(2) 资金：推动银行业金融机构探索数据资产质押、知识产权质押、绿色信贷。

(3) 人才：建设智库，建立高端人才引进绿色通道，完善配套政策；完善技术入股、股权期权激励、科技成果转化收益分配等机制。

四、结论

智能制造技术预见和路线图是智能制造的发展蓝图，是关于智能制造发展的技术方向、关键技术和技术路线的共识性框架，可为智能制造技术有序发展提供参考，在支撑科学决策和明确发展路径中具有重要的应用价值，主要包括三个方面。第一，助力制造业实现跨越式发展。目前信息技术开始大量涌入智能产品、智能生产、智能服务等各个领域，在这种情况下，提出适合我国国情的智能制造发展技术路线图、明确智能制造技术发展的方针和优先行动，对我国广大企业制定智能制造升级路径具有借鉴意义。第二，为“十四五”制定智能制造规划提供了参考。本研究通过科学严谨的方法，提出了面向2035 的智能制造发展的关键技术清单，具有科学性、前瞻性和战略性。研究对关键技术的技术内容作了清晰的描述，并且明确了发展方向，已作为制定“十四五”期间我国智能制造发展规划的参考依据之一。第三，可为今后的技术开发指明发展方向。本研究所制定的路线图具有行动导向作用，可以使不同主体的科研和投资活动在一个系统框架下形成协调的、长期的、稳定的合作，减少技术发展的盲目性和重复性，进而提高科研工作的有效性。未来，中国工程院智能制造技术路线图项目组，将以本次智能制造技术路线图为基础，2 年一个周期对智能制造技术路线图进行迭代更新；并深入智能制造的应用和支撑行业，推动制定我国细分领域的智能制造技术路线图，如智能机床技术路线图等，以期为我国智能制造发展提供理论建议，为我国制造强国建设提供战略支撑。