Six thousand words let you understand what is a digital twin?

1. Background

Recently, the company is considering to gradually digitize the campus to build a smart campus based on digital twin technology ecology when serving B-side customers. Now in the stage of providing the construction plan, we need to do some technical research and make the following summary.

2. Digital twin basics

2.1 Concept

Digital twin, also known as "digital twin", is to map the structure, state, behavior, function and performance of industrial products, manufacturing systems, cities and other complex physical systems to the digital virtual world, through real-time sensing, connection mapping, precise analysis and Immersive interaction is used to describe, predict and control physical systems, realize the integration of virtual and real complex systems, and maximize the closed-loop optimization of all elements, processes, and value chains of the system.

Digital twin integrates technologies such as artificial intelligence (AI) and machine learning (ML), combines data, algorithms, and decision-making analysis to establish a simulation, that is, a virtual mapping of physical objects, to detect problems before they occur, and to monitor physical objects Changes in the virtual model, diagnosis based on artificial intelligence-based multi-dimensional data complex processing and abnormal analysis, and predict potential risks, rationally and effectively plan or maintain related equipment.

2.2 Value

It reflects the current asset condition and includes historical data related to the asset. Digital twins can be used to assess the latest condition of an asset and more importantly predict future behavior, refine control or optimize operations.

• Simulate, monitor, diagnose, predict and control product formation and behavior in real-world environments.
• Fundamentally promote efficient collaboration throughout the product life cycle and drive continuous innovation.
• The digital product life cycle archives have laid a data foundation for the whole process traceability and continuous improvement of research and development.

3. Technology ecology

3.1 Technical system

• Data Assurance Layer
• modeling computing layer
• functional layer
• immersive experience layer

3.2 Core Technology

3.2.1 Multi-domain, multi-scale fusion modeling

Multi-domain modeling refers to the cross-domain design understanding and modeling of physical systems from different domains and in-depth mechanism levels, starting from the initial conceptual design stage under normal and abnormal conditions.

3.2.2 State assessment based on data-driven and physical model fusion

The data-driven method can use the historical and real-time operating data of the system to update, correct, connect and supplement the physical model, fully integrate the system mechanism characteristics and operating data characteristics, and better combine the real-time operating status of the system to obtain An evaluation system that dynamically follows the state of the target system in real time.

A data-driven and physical model fusion approach:

1. The analytical physical model is mainly used, and the parameters of the analytical physical model are corrected by using the data-driven method.
2. The analytic physical model and the data-driven method will be used in parallel, and finally weighted according to the reliability of the output of the two to obtain the final evaluation result.

3.2.3 Data collection and transmission

The collection and rapid transmission of high-precision sensor data is the basis of the entire digital twin system. The performance of various types of sensors, including temperature, pressure, vibration, etc., must reach the optimal state to reproduce the operating state of the physical target system.

3.2.4 Full lifecycle data management

The full lifecycle data storage and management of complex systems is an important support for the digital twin system. The cloud server is used to carry out distributed management of the massive operating data of the system, realize high-speed reading of data and secure redundant backup, provide sufficient and reliable data sources for data intelligent analysis algorithms, and play an important role in maintaining the operation of the entire digital twin system .

3.2.5 Virtual Reality Presentation

Virtual reality (VR) technology can present the manufacturing, operation, and maintenance status of the system in a surreal form, conduct multi-field and multi-scale status monitoring and evaluation of various subsystems of complex systems, and attach intelligent monitoring and analysis results to In the various subsystems and components of the system, while perfectly reproducing the physical system, the digital analysis results are superimposed on the created twin system in the form of virtual mapping, providing immersive virtual reality in terms of vision, sound, touch, etc. Realistic experience to realize real-time and continuous human-computer interaction.

The rendering layer can rely on the solutions of Tencent Cloud vendors:
Cloud Application Rendering relies on Tencent's rich edge computing nodes, flexible GPU virtualization technology, and stable and low-latency audio and video streaming capabilities to help you apply customer Terminal cloudization enables users to operate cloud applications only through lightweight terminals without downloading application packages. At the same time, Applied Cloud Rendering provides cloud API + full-end SDK to meet the needs of your users in various terminals and business scenarios.

3.2.6 High Performance Computing

The realization of the complex functions of the digital twin system depends largely on the computing platform behind it, and real-time performance is an important indicator to measure the performance of the digital twin system. Therefore, the cloud server platform based on distributed computing is an important guarantee for the system, and optimizing the data structure and algorithm structure to improve the task execution speed of the system is an important means to ensure the real-time performance of the system.

Overall solution architecture (Tencent Cloud):

3.3 Construction

3.3.1 Key points

3.3.1.1 Digital twin process design and information requirements

Use standard process design techniques to show
how business processes, process managers, business applications, information and physical assets interact, create related diagrams, connect production processes and applications, data requirements and sensors required to create digital twins information type. Process design will be enhanced with several features to improve cost, time and asset efficiency, which form the basis of the digital twin, and the
enhanced effectiveness of the digital twin begins here.

3.3.1.2 Conceptual architecture of digital twin

By creating enabling technologies that integrate real assets and their digital twins, sensor data flows in real time with operational and transactional information in enterprise core systems. The digital twin concept architecture can be divided into six easy-to-understand steps:

Detailed description:
creation
The creation step includes equipping the physical process with a large number of sensors to detect and obtain key data of the physical process and its environment. The data detected by the sensor is converted into protected digital information by the encoder and transmitted to the digital twin system. Signals from sensors can be enhanced with process-oriented information from manufacturing execution systems, enterprise resource planning systems, CAD models, and supply chain systems, providing a large amount of continuously updated data for digital twin systems for analysis.

Transmission
Network transmission is one of the major changes that make digital twins a reality, facilitating seamless, real-time, two-way integration/interconnection between real processes and digital platforms. Transmission includes the following three components.

Aggregation
The Aggregation step supports storing the acquired data in a data repository, processing it and preparing it for analysis. Data aggregation and processing can be done on-site or in the cloud.

Analyze
In the analyze step, the data is analyzed and visualized. Data scientists and analysts leverage advanced data analytics platforms and techniques to develop iterative models to uncover insights, make recommendations, and guide decision-making.

Insights
In the Insights step, insights unearthed by analytics tools are presented as visualizations in dashboards, highlighting unacceptable differences in performance between the digital twin and physical world analogs in one or more dimensions, indicating possible needs Area to investigate or replace.

Action
Action steps refer to the actionable insights formed in the previous steps that are fed back to physical assets and digital processes to realize the role of digital twins. Insights decoded into actuators responsible for movement or control mechanisms on the physical asset process, or updated in back-end systems governing supply chain and order behavior, can be human-intervened to complete the physical world and digital twin The last link of the closed-loop connection between.

3.3.2 Deployment Key Points

Envision Possibilities
Envision and select a range of scenarios where digital twins can yield benefits. Although different companies or different environments have different applicable solutions, they usually have the following two important characteristics.

Scenario Evaluation
Evaluate each scenario after it has been selected to identify processes that can leverage digital twins for quick wins. A centralized ideation session is recommended, with members of operations, business, and technology leadership working together to facilitate the evaluation process.

Identify process
Identify digital twin pilot models with the highest potential value and highest probability of success. Comprehensively consider operational, commercial, and organizational change management factors to create the best pilot plan. At the same time, focus on areas where there is potential to scale up equipment, siting, or technology.

Pilot Projects
Through agile iterative cycles, rapidly pilot projects to accelerate the learning process and maximize return on investment through effective risk management. In the process of advancing the pilot project, the implementation team should always emphasize adaptability and open thinking to create an unknown open ecosystem
, which can integrate new data and accept new technologies and partners in a timely manner.

Industrialize the process
Immediately after the success of the pilot project, existing tools, techniques and scripts can be used to industrialize the digital twin development and deployment process. This process includes integrating the various fragmented implementation processes of the enterprise, implementing data lakes, improving performance and productivity, improving governance and standardizing
data, and promoting organizational changes to support digital twins.

Scale up the digital twin
After successful industrialization, the focus should be on opportunities to scale up the digital twin. Targets should be focused on similar processes and processes related to the pilot project. Learn from the trial run experience of the project, adopt the tools, technologies and scripts used during the trial run, and quickly expand the scale.

Monitoring and detection
Monitor the solution and objectively detect the value created by the digital twin; determine whether tangible benefits can be generated within the cycle, improve productivity, quality, utilization, and reduce incidents and costs; repeatedly debug the digital twin process and observe the results , to determine the best configuration.

4. Digital twin cities

Digital twin city is the wide application of digital twin technology at the city level. By constructing a complex giant system with one-to-one correspondence, mutual mapping, and collaborative interaction with the physical world of the city and the network virtual space, a matching and corresponding city is recreated in the network space. Twin cities realize
digitalization and virtualization of all elements of the city, real-time and visualization of the full state of the city, and collaborative and intelligent urban management decision-making.

4.1 Features

4.1.1 Accurate Mapping

The digital twin city realizes comprehensive digital modeling of urban roads, bridges, manhole covers, light poles, buildings and other infrastructure through the deployment of sensors at various levels such as air, ground, underground, and river channels, as well as full perception and dynamics of the city's operating status. Monitoring forms
the accurate information expression and mapping of the virtual city to the physical city in the information dimension.

4.1.2 Virtual-real interaction

There are traces of urban infrastructure and construction of various components, and urban residents and visitors can get information when they contact the Internet. In the future digital twin city, all kinds of traces can be observed in the physical space of the city, and various information can be searched in the virtual space of the city. Urban planning, construction and various activities of
the people It has been greatly expanded, and the fusion of virtual and real, and the collaboration of virtual and real will define a new model for the future development of the city.

4.1.3 Software Definition

The digital twin city establishes a corresponding virtual model for the physical city, and uses software to simulate the behavior of people, things, and things in the city in a real environment. Through cloud and edge computing, it softly guides and controls the city's traffic signal control, electric heating Energy scheduling, major project cycle management, infrastructure site selection and construction.

4.1.4 Intelligent Intervention

Through planning and design, simulation, etc. on the "digital twin city", intelligent early warning of possible adverse effects, conflicts, and potential dangers that may occur in the city, and provide reasonable and feasible countermeasures and suggestions, intelligently intervene in the original development of the city from a future perspective Track and operation, and then guide and optimize the planning and management of physical cities, improve the supply of citizen services, and endow urban life with "smartness"

4.2 Overall Architecture

4.2.1 End side

Crowd perception, visible and controllable. City perception terminals form group intelligence perception capabilities in "groups". Perception facilities will develop rapidly from single RFID and sensor nodes to intelligent hardware with stronger perception, communication, and computing capabilities, intelligent poles, and intelligent unmanned vehicles. At the same time, the smart phones and smart terminals held by individuals will integrate more and more sophisticated sensing capabilities, with increasingly powerful perception, computing, storage and communication capabilities, and become "strong" nodes for sensing the surrounding environment and residents of the city, forming Crowd perception for large-scale, large-scale, and collaborative pervasive computing.

Based on signs and perception systems, the intelligence level of traditional infrastructure is comprehensively improved. Through the establishment of an urban comprehensive utility corridor based on intelligent signs and monitoring, the coordinated planning of the utility corridor, the visualization of construction and operation, and the retention of all process data are realized. By establishing an intelligent road network, the intelligent monitoring, maintenance and two-way control management of road network, fences, bridges and other facilities can be realized. Multifunctional information poles and other new intelligent facilities are deployed in the whole area to realize intelligent capabilities such as intelligent lighting, information interaction, wireless service, motor vehicle charging, emergency call, and environmental monitoring.

4.2.2 Grid side

Ubiquitous high-speed, heaven and earth integration. Provide ubiquitous high-speed, multi-network synergy access services. Comprehensively promote multi-network coordinated deployment such as 4G/5G/WLAN/NB-IoT/eMTC, realize three-dimensional seamless coverage based on virtualization and cloud technologies, provide wireless perception, mobile broadband and Internet of Everything access services, and support the new generation Converged applications of mobile communication networks in vertical industries.

Form an integrated information network of "heaven and earth" to support cloud services. Comprehensively utilize new information network technology, give full play to the respective advantages of "air, space, and ground" information technology, and effectively acquire, coordinate, transmit and gather multi-dimensional information such as "air, space, ground, and sea", as well as coordinate the processing of resources , task distribution, action organization and management, realize the integrated comprehensive processing and maximum effective utilization of time-space complex networks, and provide real-time, reliable and on-demand services for various users with ubiquitous, mobile, efficient, intelligent, Collaborative information infrastructure and decision support systems.

4.2.3 Cloud side

On-demand scheduling, inclusive and convenient.

High-speed information processing capabilities are provided by edge computing and quantum computing facilities. In urban factories, roads, transfer boxes and other places, build edge computing nodes with surrounding environment sensing, on-demand distribution and intelligent feedback response. Deploy various quantum computing facilities based on atoms, ions, superconducting circuits, and photons
to provide massive information processing capabilities for ultra-large-scale data retrieval, precise urban weather forecasts, computationally optimized traffic command, and artificial intelligence research and exploration. support.

Artificial intelligence and blockchain facilities for smart contract execution. Construct artificial intelligence unified computing platforms and facilities that support knowledge reasoning, probability statistics, deep learning, etc., as well as intelligent support capabilities for knowledge computing, cognitive reasoning, motion execution, and human-computer interaction capabilities; establish highly customized and personalized deployment blocks Chain service facilities, supporting the automatic execution of business smart contracts such as identity verification for various applications, electronic evidence preservation, supply chain management, and product traceability.

Deploy cloud computing and big data facilities. Establish a virtual integrated cloud computing service platform and big data analysis center, based on SDN technology to realize the scheduling capabilities of cross-regional servers, networks, and storage resources, to meet the needs of smart government affairs office and public services, comprehensive governance, and industrial development. Compute needs.

5. Summary

Digital twins have not yet been applied on a large scale, and basically require targeted construction. The cost of building small businesses is relatively high. Currently, it can be realized with the help of Tencent Cloud's overall ecological solution.

reference:

"Digital Twin" by Postdoctoral Chen Gen

IoT Device Insight IoT Insight

Tencent Smart Building Operation System

Digital twin real-time rendering solution