Industrial big data refers to all kinds of data and related technologies and applications generated in the industrial field around the typical intelligent manufacturing model, from customer demand to sales, orders, planning, research and development, design, process, manufacturing, procurement, supply, inventory, shipment and delivery, after-sales service, operation and maintenance, scrap or recycling and remanufacturing of the entire product lifecycle of each link. Industrial big data is the core of intelligent manufacturing, based on "big data + industrial Internet", cloud computing, big data, Internet of Things, artificial intelligence and other technologies to lead the change of industrial production methods, and drive the innovative development of the industrial economy.
Detailed description of industrial big data technology and application
I. the definition of industrial big data
Industrial big data refers to the industrial field, around the typical intelligent manufacturing model, from customer demand to sales, orders, planning, research and development, design, process, manufacturing, procurement, supply, inventory, shipment and delivery, after-sales service, operation and maintenance, scrap or recycling and remanufacturing of the entire product lifecycle of various types of data generated by various aspects of the general term, and the related technology and applications. It is centered on product data, which greatly extends the scope of traditional industrial data, and also includes industrial big data-related technologies and applications. The main sources of industrial big data are the following three categories.
1. Business data related to production and operation
Production and operation-related business data mainly come from the scope of traditional enterprise informatization and are stored inside enterprise information systems, including traditional industrial design and manufacturing software, enterprise resource planning (ERP), product lifecycle management (PLM), supply chain management (S CM), customer relationship management (CRM), and environmental management system (EMS), etc. These enterprise information systems have accumulated a large amount of data. These enterprise information systems have accumulated a large amount of product development data, production data, operational data, customer information data, logistics and supply data, and environmental data. This kind of data is the traditional data asset in the industrial field, and is gradually expanding its scope under the environment of new technology applications such as mobile Internet.
2. Equipment IoT data
Equipment IoT data mainly refers to industrial production equipment and target products in the IoT operation mode, real-time generation and collection of data covering operation and operation, working conditions, environmental parameters and other data reflecting the operation status of equipment and products. Such data is the new and fastest growing source of industrial big data. Narrow industrial big data refers to this type of data, that is, a large amount of data generated quickly by industrial equipment and products and the existence of time-series differences.
3. External data
External data refers to the production activities of industrial enterprises and products related to the enterprise's external Internet sources of data, for example, the evaluation of the enterprise's environmental performance of environmental regulations, predicting the product market macro-socio-economic data and so on. Industrial big data technology is a series of technologies and methods that enable the value contained in industrial big data to be mined and displayed, including data planning, acquisition, pre-processing, storage, analysis and mining, visualization and intelligent control. Industrial big data application is the process of integrating and applying a series of industrial big data technologies and methods to a specific industrial big data set to obtain valuable information. The research and breakthrough of industrial big data technology is essentially aimed at discovering new patterns and knowledge from complex data sets, and mining valuable new information, so as to promote the product innovation of manufacturing enterprises, improve the level of operation and production efficiency and expand new business models.
II. Industrial big data characteristics
In addition to the characteristics of general big data (large data volume, variety, rapidity and low value density), industrial big data also has the characteristics of temporal sequence, strong correlation, accuracy and closed loop.
Large data volume: the size of the data determines the value and potential information of the data under consideration. Industrial data volume is relatively large, a large number of high-frequency data from machines and equipment and Internet data continue to pour in, and the data sets of large industrial enterprises will reach the PB or even EB level.
Variety (variety): refers to the diversity of data types and a wide range of sources. Industrial data is widely distributed in various aspects such as machines and equipment, industrial products, management systems, and the Internet, and the structure is complex, with structured and semi-structured sensing data as well as unstructured data.
Fast (velocity): refers to the speed of obtaining and processing data. Industrial data processing speed needs are diverse, production site-level requirements for analysis time frame to the millisecond level, management and decision-making applications need to support interactive or batch data analysis.
Low value density (value): industrial big data puts more emphasis on user value-driven and the usability of the data itself, including: improving the innovation ability and production and operation efficiency and promoting personalized customization, service transformation and other new modes of intelligent manufacturing change.
Sequence: Industrial Big Data has a strong temporal sequence, such as orders, equipment status data.
Strong-relevance: On the one hand, the data at the same stage of the product life cycle has strong relevance, such as the composition of product parts, working conditions, equipment status, maintenance, and supplemental procurement of parts; on the other hand, the data at different stages of the product life cycle, such as R&D and design, production, and service, need to be associated.
Accuracy: mainly refers to the authenticity, completeness and reliability of the data, and pays more attention to the quality of the data as well as the reliability of the processing and analyzing techniques and methods. Higher confidence requirements for data analysis, relying only on statistical correlation analysis is not enough to support fault diagnosis, prediction and early warning and other industrial applications, the need to combine the physical model with the data model, mining cause and effect relationship.
Closed-loop: including the closure and association of the data chain in the horizontal process of the whole life cycle of the product, as well as the vertical data acquisition and processing process of intelligent manufacturing, which needs to support the dynamic and continuous adjustment and optimization under the closed-loop scenarios of state sensing, analysis, feedback, and control.
Due to the above characteristics, industrial big data, as an application industry of big data, has a broad application prospect while posing a great challenge to the traditional data management technology and data analysis technology.
III. Industrial Big Data Architecture
Industrial big data architecture contains three dimensions: life cycle and value stream, enterprise vertical layer and IT value chain.
In the life cycle and value stream layer, according to the application areas of industrial big data, it can be divided into three areas: product development and design before the start of the product production phase, production and supply chain management before product delivery, and operation and maintenance and service management after product delivery.
In the vertical layer of the enterprise, according to the data collection method and application level, it can be divided into information physical system layer, enterprise management information system layer and platform interconnection system layer.
In the IT value chain layer, it can be divided into business architecture, information system architecture and IT technology architecture at three levels, of which information system architecture can be divided into application architecture and information architecture.
1. Life cycle and value stream dimension
The life cycle and value stream dimension in the industrial big data architecture covers the stages of the entire product life cycle, i.e., R&D and design, production, logistics, sales, operation and maintenance, and service. Among them, production, logistics and sales can be further categorized into production and supply chain, and the life cycle and value stream dimension includes three fields: R&D and design, production and supply chain, and operation and maintenance and service. The application scenarios of each domain are shown in Figure 2.
01. R&D and Design
R&D data is accumulated by R&D personnel in the process of R&D and design, which comes from all aspects of the product life cycle, including: user demand big data, R&D knowledge big data, product reuse big data, R&D collaborative big data, etc., with cross-products and cross-industry, and a wide variety of characteristics.
Personalized product customization design enterprises can collect users' personalized product demand, product customer interaction and transaction data through the Internet platform. Mining and analyzing these customer dynamic data can help customers participate in product demand analysis and product design activities to achieve customized design, and then relying on the flexible production process, you can produce tailor-made products for users.
Realize the simulation design based on big data traditional production enterprises in the testing and verification of the link needs to be produced in kind to evaluate its performance and other indicators, the cost with the increase in the number of tests and rising. The use of virtual simulation technology can achieve the original R & D design process of simulation, analysis, evaluation, verification and optimization, thereby reducing the amount of engineering changes, optimize the production process, reduce costs and energy consumption.
Realize personalized custom design automation based on big data traditional enterprise product types, styles are not many, can be used to manually design product models, production samples, and then mass production production production mode, but in the face of personalized, small batch production requirements, the traditional mode will lead to product production cycle is too long, the cost is too high. By accumulating a large amount of product design model data, analyzing the correlation between the design data, and with the help of big data technology and other auxiliary design tools, the automation of personalized customized design and model generation can be realized.
Promote the integration and sharing of R&D resources and innovation and collaborative design enterprises through the construction and improvement of R&D and design knowledge base, to promote the digital drawings, standard parts library and other design data within the enterprise as well as upstream and downstream enterprises in the supply chain of resource sharing and innovation and collaboration, to enhance the enterprise's cross-regional R & D resources integrated management and industry chain collaborative design capabilities. Enhance the ability of enterprises to manage and utilize global R&D resources, optimize and reorganize R&D processes, and improve R&D efficiency.
Cultivate new modes of R&D based on the socialized sharing and participation of design resources, and enable enterprises to carry out new modes of R&D such as crowdsourcing and crowdsourcing based on their own R&D needs, so as to enhance the ability of enterprises to utilize socialized innovation and capital resources.
02.Production and Supply Chain
Production big data not only includes product production information, order information, equipment information, control information, material information, personnel work scheduling, but also includes internal management information flow, capital flow, product production of upstream and downstream suppliers and customer management and other related auxiliary production management information, the collection of production data relies on the enterprise's existing resource management, manufacturing execution, industrial control management, Supply chain management, supplier management, customer management, business management and other information systems.
It realizes real-time monitoring and management of the production process and predictive maintenance of production equipment, improves the management level of the production process and equipment, optimizes the production process and improves product quality. Modern industrial manufacturing production lines are installed with thousands of small sensors to detect the working status of production equipment, such as temperature, pressure, heat, vibration and noise, etc. The use of these data can realize real-time monitoring of the production process, equipment failure diagnosis and prediction, energy consumption analysis, quality accident analysis. In addition, it can also integrate and aggregate data from all aspects of manufacturing, establish virtual models of the production process, simulate and optimize the production process.
Realize personalized customized scale production and promote the establishment of modern production system. Through the automation of data flow in the whole life cycle of the product and the automated and intelligent control of the whole manufacturing process, it will promote information sharing, system integration and business collaboration, improve the ability of precision manufacturing, high-end manufacturing and agile manufacturing, realize personalized customized scale production, accelerate the establishment of modernized production systems such as smart workshops and smart factories, and realize intelligent production.
Realize networked collaborative manufacturing and manufacturing sharing economy. Through the "Internet +", the integration and optimization of production resources within or among enterprises, and the realization of vertical collaborative manufacturing within enterprises or horizontal collaborative manufacturing among enterprises. Through the Internet + sharing economy, the sharing of innovation resources, production capacity, inventory and other production resources to realize the manufacturing sharing economy.
Optimize the industrial supply chain. Electronic identification technology such as radio frequency identification (RFID), Internet of Things (IoT) technology, and mobile Internet technology can help industrial enterprises obtain big data on the complete product supply chain, and the use of such data for analysis will bring about a significant increase in the efficiency of warehousing, distribution, and sales as well as a significant reduction in costs.
Realize demand forecasting to better arrange incoming goods and production, and when demand drops, trace the cause of the problem and solve it.
Realize customer profiling and precision marketing and customer behavior analysis, which can expand customer sources, improve the success rate of marketing and original customer satisfaction and loyalty.
03.Operation and Maintenance and Service Field
There are many sources of data in the field of operation and maintenance and service, mainly including: real-time running status data and surrounding environment data collected by the sensors embedded in the products with the permission of the customers; product sales data obtained through the business platform, customer data and the corresponding product evaluation or use feedback; customer complaints and the corresponding processing records; product returns/returns and the corresponding maintenance records.
By monitoring and analyzing the real-time operation status data of products collected remotely, online value-added services such as remote monitoring and management, fault diagnosis and predictive maintenance can be realized, which can reduce maintenance costs and improve product utilization.
By analyzing the customer usage data and surrounding environment data of the equipment, it can also provide extended services for the users, expand the value space of the products, and realize the transformation of the product-centered business model to the model of "manufacturing + service".
By analyzing customer product evaluation or usage feedback, customer complaints, incorporating useful comments into product design and product improvement, and categorizing customer complaints, we can improve product quality and after-sales service quality, reduce the complaint rate, and increase customer satisfaction and loyalty.
By analyzing the reasons for product return or repair, and taking timely and effective measures, we can improve product quality and reduce the return rate and repair rate.
2. Enterprise vertical layer
The enterprise vertical layer of industrial big data architecture is divided into five layers from the perspective of physical domain from bottom up, which are equipment layer, control layer, workshop layer, enterprise layer and collaboration layer. In the equipment layer, control layer, and workshop layer, the Internet of Things can be used to realize smart factories based on the information physical system; in the enterprise layer, the enterprise integrates various internal informatization applications, carries out the integration and transformation of internal business processes, and improves the operational efficiency of the enterprise; and in the collaboration layer, the industrial cloud and other platform technologies are used to realize the external collaborative manufacturing of the enterprise and innovative business models such as the manufacturing service-oriented model. The vertical dimension of the enterprise can be divided into three subsystems: information physical system, enterprise management information system and interconnection platform system.
01.Information Physical System
Information physical system collects and aggregates machine operation data and production site data through sensors and various information systems to realize ubiquitous sensing, and applies data integration and processing technology to collect and exchange industrial data, manufacturing feedback and control, to achieve control and interaction with equipment and workshops, to realize interconnection and interoperability of internal and external physical systems of factories, and to provide a data basis for industrial modeling/simulation and analysis, and then provide support for decision-making optimization of workshop/factory operations. It also provides data basis for industrial modeling/simulation and analysis, which in turn provides support services for workshop/factory operation decision optimization. In the vertical layer of the industrial big data architecture enterprise, the information physical system used for information collection on the industrial equipment layer is based on big data, network and mass computation, and through the core intelligent perception, analysis, mining, evaluation, prediction, optimization, collaboration and other technical means, the computation, communication, and control can realize the organic fusion and in-depth cooperation, so as to achieve the deep fusion of the cyberspace and physical space of industrial equipments, environments, and groups. Deep integration. The essence of information physical system lies in connecting physical equipment to the Internet, so that physical equipment has five major functions: computation, communication, precise control, remote coordination and autonomy.
02.Enterprise Management Information System
Enterprise informatization is the process of applying information technology and products in enterprises. Enterprise informatization is the process of comprehensive penetration of information technology from local to global, from tactical level to strategic level to enterprises, applying it to process management and supporting enterprise operation and management. Enterprise informatization mainly involves production process control, enterprise management, product life cycle management, supply chain optimization and management processes. Production process control informatization focuses on product development and design, production process flow, workshop management, quality inspection and other design and production links. Enterprise management informatization is the largest proportion of enterprise information construction, the most difficult, the most widely used in a field, involving the business of enterprise management and all levels. Informatization construction of enterprise management is to collect, process, organize and integrate information resources effectively through the information integration application system on the basis of standardizing management basic work and optimizing business processes, improving management efficiency, and providing real-time dynamic management information and decision-making information. The informatization of enterprise supply chain management makes the production and management activities of the enterprise extend forward and backward. Enterprises from the procurement of raw materials, components, transportation, storage, processing and manufacturing, sales, until the final delivery and service to customers, forming a chain structure consisting of upstream suppliers, intermediate producers and third-party service providers, downstream sales customers, and the production activities of enterprises, management processes are subject to the constraints of this supply chain and influence.
03.Connected Platform System
At present, the industrial development of China and most countries are facing great difficulties and challenges, including: serious overcapacity, the scarcity of personalized products, products are becoming more and more complex, the means of production can not be effectively configured, and the market for large-scale equipment is becoming increasingly saturated, etc., and urgently need to seek a strategic solution for the return of the industry and the transformation and upgrading of the industry. "Internet +" very dynamic thinking and innovative business models, for the plight of the manufacturing industry in urgent need of transformation and upgrading provides a new direction of transformation, through the reform of production methods and business models and improve production technology, to achieve customer personalization of customized products, small batch, large-scale production, to solve the problem of large-scale production capacity of industrial products. By reforming production methods and business models and improving production technology, we can realize small-lot, large-scale production of personalized products to solve the problem of large-scale overcapacity of industrial products and the inability to meet the customers' personalized demand for products, so as to satisfy the demand for customers' respect and affirmation as well as self-realization.
Try to solve the problem of weak core technology and low manufacturing capacity of high-end products through networked collaborative manufacturing, i.e., with the help of the Internet or industrial cloud platform, develop new modes of collaborative research and development, crowdsourcing design, supply chain collaboration and so on among enterprises, in order to effectively reduce the cost of acquiring resources, significantly extend the scope of resource utilization, break the closed boundaries and accelerate the transformation from "fighting alone" to "industrial collaboration". "to industrial synergy, and promote the overall competitiveness of the industry.
Try to improve the problem of overcapacity without effective allocation of resources and weak independent innovation ability through innovation and entrepreneurship, manufacturing sharing economy.
With the core product as the axis, through the collection and analysis of product customer use data and the surrounding environment data, to provide users with extended services, expanding the value space of the product, expanding new markets, and realizing the transformation of the product-centered business model to the "manufacturing + service" model.
3. IT Value Chain
The value of big data is realized through activities such as data collection, pre-processing, analysis, visualization and access.
In the IT value chain dimension, the value of big data is realized through the provision of networks, infrastructure, platforms, application tools and other services that house big data for big data applications, thereby improving operational efficiency and supporting business innovation. The enterprise architecture supported by big data technology can be divided into three levels: business architecture, information system architecture and IT technology architecture with reference to TOGAF division method.
01. Business Architecture
Business architecture defines business strategy, management, organization and key business processes, and is the foundation of the enterprise's comprehensive information technology strategy and information system architecture, as well as the determinant of data, application and technology architecture. Business architecture is the channel that translates an organization's business strategy into daily operations, and business strategy determines business architecture. Business architecture converts high-level business strategies and goals into operational business models. Business architecture is an expression of the key business strategy of the enterprise and of the business functions and processes, usually a business design implemented on the basis of a business model, which describes the business modules and the relationships between them, i.e., the main processes of the business, from different perspectives. Business architecture is a proper delineation of the main and shared processes of the business, and the articulation and analysis of the lifecycle of business elements. The value of industrial big data can be obtained by strategically planning industrial big data business and building enterprise architecture.
02.Information System Architecture
In order to give full play to the value of industrial big data and avoid the formation of "information islands", it is necessary to build a unified information system architecture to realize user access and interoperability of various application systems and data. The information system architecture based on the industrial big data business strategy is an architecture that reflects the relationship between the various components of a manufacturing enterprise's information system, as well as the relationship between the information system and the related business, and between the information system and the related technology. Information system architecture includes application architecture and data architecture. Among them, application architecture describes the blueprint of the application system required to support enterprise operation, including application levels, functions, realization methods and construction standards, etc. It mainly studies the interaction between application systems and the correspondence between applications and core business, and is the focus of the research on the overall framework of the enterprise, which can be said to be the bridge between the business architecture and the technical architecture. Industrial big data application architecture contains both application systems corresponding to the various levels of the vertical layer of the enterprise in industrial big data architecture and application systems based on big data technology. The data architecture, on the other hand, is a description of the structure and interaction of the main data types and sources, logical data assets, physical data assets, and data management resources of a complex organizational body.
03.Information Technology Architecture
With the rise of the wave of Industry 4.0, information technologies such as Internet of Things (IoT), Cloud Computing, Big Data, Artificial Intelligence, Augmented Reality/Virtual Reality and other information technologies are continuously integrating and penetrating into the industrial field, which lays a solid technological foundation for the implementation of industrial big data applications. Among them, IoT technology makes ubiquitous end equipment and facilities, which can be connected to the Internet according to agreed protocols through information sensing devices such as radio frequency identification, infrared sensors, global positioning systems, etc., for information exchange and communication, making the items and their status visible, so as to realize intelligent identification, localization, tracking, monitoring, and management; cloud computing technology provides a kind of dynamic and scalable computing service that can be realized through the network on-demand. Cloud computing technology provides an on-demand, dynamically scalable and inexpensive computing services through the network; big data technology and AI technology makes it possible to analyze and mine the potential value of massive data in an acceptable amount of time, as well as to realize the trend prediction and group intelligence mode; AR/VR technology can realize the simulation and enhancement of the experience of the factory environment, industrial equipment, and so on. Manufacturing enterprises are generating a large amount of data every day or even every moment, with a wide variety, covering the whole life cycle of industrial products, including design data, production data, value chain data and related external data. These data either come from sensors, or from the data acquisition and monitoring control system of intelligent equipment, or from the design model and information system of the enterprise.
The realization of industrial big data application requires data collection and pre-processing, storage, analysis and mining, application for specific business and finally display the results, accordingly, the industrial big data information technology architecture is divided into six layers: data layer, data collection layer, storage layer, computing layer, application layer and display layer. If enterprises build each layer alone, the construction threshold will be relatively high. A series of open-source technology components related to industrial cloud services and (industrial) big data processing have been developed and completed at home and abroad, and artificial intelligence technology has further made great progress, which provides optional solutions for the intelligent transformation of the industrial field, and also reduces the threshold for the construction and implementation of industrial big data applications. Enterprises do not need to realize each component of the technical architecture on their own if not necessary, and can choose to use the corresponding open source components to build industrial big data applications according to their needs, as well as different types and levels of industrial cloud services according to their needs, and implement industrial big data applications on this basis, so as to focus more on the business areas and technical fields that the enterprises are good at.
Industrial big data analysis technology, as one of the core technologies of industrial big data, can enable industrial big data products to have the ability to mine massive data, integrate multi-source data, model multi-type knowledge, analyze multi-business scenarios, and discover multi-domain knowledge, etc., which plays a significant role in driving the business innovation and transformation and upgrading of enterprises.