工业是国民经济的命脉,工业发展不断推动着人类社会和经济的发展进步。在过去的两百多年里,世界工业已经历了三次重大变革与飞跃,从机械化、电气化发展到如今的自动化,极大提升了人类的物质文明。随着互联网、物联网、大数据、人工智能等新一代信息技术的快速发展和应用,未来工业发展正在迎来新的变革性契机,需要新的发展思路,而工业互联网正是实现这场变革的核心要素。从全球范围来看,工业互联网作为数字化转型的关键支撑力量,推动了传统产业加快转型升级、新兴产业加速发展壮大。对于我国而言,加快工业互联网创新发展步伐,快速构建我国制造业竞争新优势,抢占未来发展主动权具有重要的战略意义。
工业互联网通过互联网、物联网、大数据、人工智能等新一代信息技术在工业领域的深度融合和创新应用,建立了广泛连接人、机、物等各类生产要素的全球性网络,形成了贯穿全产业链的实体联网、数据联网、服务联网的开放平台,是重塑工业生产与服务体系,实现产业数字化、网络化、智能化发展的重要基础设施。本报告从信息技术的视角出发,将工业互联网的互联范围划分为工厂、企业、产业链三个层级,将工作流程归纳为智能感知(感)、网络互联(联)、数据分析(知)和控制协同(控)四个环节,提出“三层级四环节”的工业互联网逻辑架构。基于该架构,工业互联网连接了全生产系统、全产业链、全价值链,使原本割裂的工业数据在网络上流通,最终实现人、机、物的全面、深度、安全互联。
从知识体系来看,工业互联网具有多学科交叉、多应用驱动、多技术融合的特点。①多学科交叉:工业互联网涉及计算机、通信、控制、电子、机械、材料、制造等,几乎涵盖了信息、工程、材料领域的大部分学科。②多应用驱动:工业互联网应用场景丰富多样,面临的新需求层出不穷,通过解决应用问题可以驱动知识体系的完善,呈现应用带动学科发展的显著特征。③多技术融合:工业互联网融合了物联网、人工智能、大数据、云计算、机器人、移动通信、智能制造、柔性材料等多种前沿技术。
从层次结构来看,工业互联网的互联范围在工厂、企业和产业链三个层级不断延伸:①对于工厂级互联,覆盖“人、机、料、法、环”各方面,打通车间内部各系统间的“信息孤岛”;②对于企业级互联,实现企业业务全流程的联网,包含产品的设计、生产、售后等各个阶段;③对于产业链级互联,涉及各企业、产业和区域之间的价值链、企业链、供需链和空间链,涵盖产品生产或服务提供的全要素、全环节、全过程,借助工业互联网平台,实现产业链上下游企业(供应、制造、销售、金融)之间的互联,实现产业基础能力提升、运行模式优化、产业链控制力增强和治理能力提升。
就本质而言,工业互联网是通过信息网络使得原本割裂分散的工业大数据实现按需有序流通的,其关键科学问题体现在“感、联、知、控”四个环节上:首先,在“感”环节上,需要通过感知生产过程相关的人、机、物等各类生产要素及所在工业场景,刻画制造过程依赖关系和时空关联,以实现从物理世界到数字世界的映射;其次,在“联”环节上,需要通过海量多元工业实体泛在接入,构建安全、可靠、高效的工业互联网络,以实现异质离散工业实体(即人、机、料、法、环)数据的集成与汇聚;再次,在“知”环节上,需要通过研究有效的认知表达范式、认知实现方法与认知决策机制,以实现工业互联网与智能认知的充分融合;最后,在“控”环节上,需要通过对工业全流程的全局协同与控制,提高生产线的柔性反应能力和供应链的敏捷精准反应能力,以实现全流程柔性生产和智能制造。针对“感、联、知、控”四个环节,本报告将以上四大关键科学问题归结为:全模态信息表征、全要素互联组织、全场景智能认知、全流程柔性协同。
围绕“感、联、知、控”四大关键科学问题,工业互联网主要涉及六大核心技术领域:①工业智能感知,针对强干扰、大范围、多目标的复杂工业环境,实现全方位感知;②工业互联与信息集成,针对工业互联网异质工业实体与异构互联网络,设计高效的自适应互联技术,实现泛在工业互联网络数据实时传输和信息高度共享;③工业大数据与工业智能,针对多源异构工业数据,结合人工智能技术及工业领域知识,实现智能分析和决策;④工业互联网控制协同,针对物理世界和数字世界中的控制与协同难题,通过虚拟仿真、数据分析、认知决策等多个领域关键技术,实现全流程控制协同方法与机制;⑤工业互联网平台软件,面向工厂、企业到产业链的不同层级对资源管理和应用开发的不同需求,提供不同层级的共性软件平台,向下实现工业资源的有效管理,向上支撑工业场景的应用开发;⑥工业互联网安全,构建涵盖设备层、网络层、数据层、应用层、控制层的工业互联网安全防护体系。
展望未来,工业互联网将呈现三大发展趋势。一是泛在化。工业互联网将接入工业全场景,涵盖各环节的人、机、物,实现无处不在、无迹可寻的工业数字孪生世界。二是协同化。通过亚微秒级低时延、高可靠、广覆盖的基础设施,实现工业全场景、全链条的人、机、物高度协同。三是智能化。通过人工智能技术与工业场景、知识的深度结合,在工业设计、生产、管理、服务等各个环节实现模仿或超越人类的能力。
随着核心技术的突破和大规模部署应用,工业互联网必将催生工业生产制造新模式,重塑产品规划、设计、制造、销售环节,为工厂、企业乃至整个产业链带来全新发展机遇。若干发展可以预期。①工厂将实现生产柔性化:柔性化生产可以打通用户交互、产品创意产生、个性化订单下达、产品模块部件匹配、自动化生产等环节,此时,“量体裁衣”式的个性化生产将成为现实。②企业将实现生产服务化:在工业互联网时代,企业价值体系由以制造为中心向以服务为中心转变,即便产品已经交付使用,企业仍可以远程感知产品的运行数据,进而得以分析实时运行状况,为用户提供维护、预警、保养等附加工业服务,并从中产生新盈利点。③产业链将实现模式重塑:随着传统的工业化生产向数字化、网络化、智能化的方向发展,新一代信息技术将在制造、能源、交通、医疗等行业深度应用,并促进跨行业领域的融合,带来革命性的产业变革。
Abstract
Industry is the lifeblood of the national economy. Industrial development is constantly promoting the development and progress of human society and economy. In the past 200 years, the world industry has experienced three major changes and leaps, from mechanization and electrification to today’s automation, which has greatly improved the material civilization of mankind. With the rapid development and application of next-generation information technologies such as the Internet, Internet of Things, big data, and artificial intelligence, future industrial development is ushering in new transformative opportunities, which require new development ideas, and the Industrial Internet is the core element to realize this change. From a global perspective, the Industrial Internet, as a key supporting force for digital transformation, promotes the transformation and upgrading of traditional industries and the development and growth of emerging industries. For China, it is of great strategic significance to accelerate the pace of innovation and development of the Industrial Internet, quickly build new competitive advantages in China’s manufacturing industry, and seize the initiative in future development.
Through the deep integration and innovative application of nextgeneration information and communication technologies such as newnetworks, artifcial intelligence, and big data in the industry, the Industrial Internet establishes a global network that widely connects people, machines, things and other production factors. The open platform of the physical networking, data networking and service networking of the industry chain is an important infrastructure for reshaping the industrial production and service system, and realizing the digital, networked and intelligent development of the industry. From the perspective of information technology, this report divides the interconnection scope of the Industrial Internet into three levels of “factory,enterprise,industry chain”, summarizes the workflow into four phases, i.e., intelligent perception (sense), network interconnection (connection), data analysis (knowledge) and open service (control), and proposes a “three-level and four-phases” Industrial Internet logical architecture. Based on this architecture, the Industrial Internet connects the entire production system, the entire industrial chain, and the entire value chain, allowing the originally fragmented industrial data to circulate in the intelligent network, and finally realizes the comprehensive, deep, and secure interconnection of people, machines, and things.
From the perspective of knowledge hierarchy, the Industrial Internet has the interdisciplinary, multi-application driven, and multi-technology integration characteristics. (1) Interdisciplinary: The Industrial Internet involves computers, communications, control, electronics, machinery, materials, manufacturing, etc., covering almost all disciplines in the felds of information, engineering, and materials. (2) Multi-application driven: Application scenarios of the Industrial Internet are rich and diverse, in which new demands are emerging one after another. Solving these problems can drive the improvement of knowledge hierarchy, which shows the remarkable feature that application drives the development of disciplines. (3) Multi-technology integration: The Industrial Internet integrates various advanced technologies such as the Internet of Things, artificial intelligence, big data, cloud computing, robot, mobile communication, intelligent manufacturing, and fexible materials.
From the perspective of hierarchical structure, the interconnection scope of the Industrial Internet extends at the three levels of factories, enterprises and industrial chains. (1) For factory-level interconnection, it covers all aspects of “human, machine, material, method and environment”, and opens up the “information island” between various systems in the workshop. (2) For enterprise-level interconnection, it realizes the networking of the whole process of enterprise business, including product design, production, after-sales and other stages. (3) For the interconnection of the industrial chain, it involves the value chain, enterprise chain, supply and demand chain and space chain between enterprises, industries and regions, covering all elements, links and processes of product production and service provision. With the help of the Industrial Internet platform, the interconnection between upstream and downstream enterprises (supply, manufacturing, sales, and fnance) in the industrial chain can be realized, and the basic industrial capacity, operation mode, industrial chain control and governance capacity can be optimized or improved.
In essence, the Industrial Internet enables the originally fragmented industrial big data to flow orderly on demand through the information network. Its key scientifc issues are refected in the four links of “sense, connection, knowledge and control”. First of all, in the link of “sense”, it is necessary to depict the dependence and spatio-temporal relationship of the manufacturing process by perceiving various production factors related to the production process, such as people, machines, things and their industrial scenes, so as to realize the mapping from the physical world to the digital world. Secondly, in the link of “connection”, it is necessary to build a safe, reliable and efcient Industrial Internet through ubiquitous access of a large number of diversifed industrial entities, so as to realize the integration and aggregation of data of heterogeneous and discrete industrial entities (i.e., human, machine, material, law and environment). Thirdly, in the link of “knowledge”, it is necessary to establish the effective cognitive expression paradigm, cognitive realization method and cognitive decision-making mechanism, soas to realize the full integration of the Industrial Internet and intelligent cognition. Finally, in the link of “control”, it is necessary to improve the fexible response ability of the production line and the agile and accurate response ability of the supply chain through the overall coordination and control of the whole industrial process, so as to realize fexible production and intelligent manufacturing in the whole process. In this report, the above four key scientifc issues are summarized as full-mode information representation, all-element interconnection and organization, all-scene intelligent cognition, and whole-process fexible collaboration.
Focusing on the four key scientific issues of “sense, connection, knowledge and control”, the Industrial Internet mainly involves six core technical fields. (1) Industrial intelligent sense. Aiming at the complex industrial environment with strong interference, large range and multiple targets, it realizes all-round sensing. (2) Industrial interconnection and information integration. Aiming at heterogeneous industrial entities and heterogeneous interconnected networks in the Industrial Internet, it designs efcient adaptive interconnection technology to achieve real-time data transmission and high information sharing in ubiquitous industrial interconnection networks. (3) Industrial big data and industrial intelligence. Aiming at multi-source heterogeneous industrial data, it combines artificial intelligence technology and industrial domain knowledge to realize intelligent analysis and decision-making. (4) Industrial Internet control and collaboration. Aiming at the control and coordination problems in the physical world and the digital world, it realizes the method and mechanism of control and coordination in the whole process through virtual simulation, data analysis, cognitive decision-making and other key technologies in various felds. (5) Industrial Internet platform software. Facing the different needs of resource management and application development of factories, enterprises and industrial chains, it provides common software platforms at different levels, realizes efective management of industrial resources downwards, and supports application development of industrial scenarios upwards. (6) Industrial Internet security. It builds an Industrial Internet security protection system covering the device layer, network layer, data layer, application layer, and control layer.
Looking forward to the future, the Industrial Internet will show the following three development trends. The first is ubiquity. The entire industrial scene covering people, machines and things in all links will be connected to the Industrial Internet, realizing an invisible but ubiquitous industrial digital twin world. The second is collaboration. Through sub-microsecond-level low-latency, high-reliability and wide-coverage infrastructure, a high degree of collaboration between people, machines, and things in all industrial scenarios and chains will be realized. The third is intelligence. Through the in-depth integration of artifcial intelligence technology with industrial scenarios and knowledge, the ability to imitate or surpass human beings will be realized in various links such as industrial design, production, management and service.
With the breakthrough of core technologies and large-scale deployment and application, the Industrial Internet will surely give birth to a new model of industrial production and manufacturing, reshape links of product planning, design, manufacturing and sale, and bring new development opportunities for factories, enterprises and even the entire industrial chain. There are several expected developments. (1) Factories will realize flexible production. Flexible production can get through links such as user interaction, product idea generation, personalizedorder placement, product module component matching, and automated production. At this time, personalized production of “tailor-made” will become a reality. (2) Enterprises will realize the production service. In the era of the Industrial Internet, the enterprise value system has changed from manufacturing-centered to service-centered. Even if the product has been delivered for use, the enterprise can still remotely sense the product’s operating data, and then analyze the real-time operating status to provide users with additional industrial services such as maintenance, early warning and maintenance, and generate new profit points from them. (3) The industrial chain will be remodeled. With the development of traditional industrial production towards the direction of digitization, networking and intelligence, the next-generation information technology will be deeply applied in industries such as manufacturing, energy, transportation and medical, and promote the integration of cross-industry felds, bringing revolutionary industrial changes.