Industrial equipment digital twin system and intelligent optimization method

Abstract

The invention relates to the technical field of digital twinning, in particular to an industrial equipment digital twinning system and an intelligent optimization method, and integrates digital twinning modeling, edge computing, cloud computing and visualization technologies. The method comprises the following steps: firstly, constructing an equipment digital twinborn body through three-dimensional modeling and a CAD drawing, and simulating and acquiring data by using a virtual sensor; and the edge calculation module synchronizes physical equipment and digital model data to realize collaborative perception, reasoning, state evaluation and early warning. And the cloud computing module further analyzes the data and optimizes the operation parameters of the equipment. The visualization module integrates multivariate data in a three-dimensional interface, visually displays the equipment state and optimization information, responds to a user instruction, realizes comprehensive, real-time and interactive equipment monitoring and management, and improves the intelligent level and operation and maintenance efficiency of industrial equipment, and the systematic method not only improves the overall efficiency of industrial equipment management, but also improves the operation and maintenance efficiency of the industrial equipment. And reliable technical support is provided for digital transformation of enterprises.

Description

Technical Field

[0001] The present invention relates to the technical field of digital twins, and particularly to an industrial equipment digital twin system and an intelligent optimization method. Background Art

[0002] With the advent of the Industrial 4.0 era, digital transformation has become an inevitable trend in the manufacturing industry. Against this background, digital twin technology, as a bridge connecting the physical world and the digital world, has attracted wide attention in the industry. However, the application of current industrial equipment digital twin systems still faces many challenges.

[0003] Existing industrial equipment digital twin systems usually use simplified models to represent physical equipment, and these models are often difficult to comprehensively and accurately reflect the complex behavior of the equipment. For example, some systems only focus on the geometric features of the equipment, ignoring key factors such as material properties and internal structures, resulting in a large deviation between the simulation results and the actual situation. On the other hand, although some systems have established relatively complex models, due to the limitation of computing resources, they are unable to achieve real-time data synchronization and status update, making it difficult for digital twins to reflect the actual state of physical equipment in a timely manner.

[0004] In terms of data processing, existing systems generally have the problem of data islands. The massive data generated by physical equipment is often scattered in different systems, lacking effective integration and utilization. This not only leads to a waste of valuable data resources but also affects the accuracy and timeliness of decision-making. In addition, although some systems have achieved centralized data storage, when facing heterogeneous data sources, the data fusion and processing efficiency is still not satisfactory.

[0005] In terms of intelligent optimization, existing industrial equipment management systems mostly use rule-based methods for optimization and control. Although this method is simple and intuitive, it is difficult to cope with complex and changing industrial environments. Especially when facing multi-objective optimization problems, traditional methods often struggle to find the global optimal solution. At the same time, due to the lack of in-depth mining and learning of historical data, the optimization ability of the system cannot be continuously improved over time.

[0006] In terms of visualization and interaction, most current systems still remain at the level of two-dimensional charts and simple three-dimensional models, making it difficult to intuitively present complex industrial equipment and its operating status. This not only increases the difficulty for operators to understand and analyze data but also limits the decision-makers' grasp of the overall condition of the equipment.

[0007] In addition, the scalability of existing systems is generally poor. With the increase in the types and quantities of industrial equipment, as well as the continuous introduction of new sensors, the performance of the system often fails to meet the growing data processing requirements. This seriously restricts the application of digital twin technology in large-scale industrial environments. Summary of the Invention

[0008] In view of the above problems, the present invention proposes an industrial equipment digital twin system and an intelligent optimization method. By integrating advanced technologies such as digital twin modeling, edge computing, cloud computing, and visualization, this system achieves a high degree of integration between physical devices and digital models, providing a comprehensive and powerful solution for the intelligent management of industrial equipment.

[0009] The present invention proposes an industrial equipment digital twin system, including:

[0010] A digital twin modeling module, used for:

[0011] Constructing a digital twin model of industrial equipment based on 3D modeling software;

[0012] Establishing a 3D model of industrial equipment based on CAD drawings;

[0013] Simulating the digital twin model of industrial equipment using virtual sensors to obtain digital sensor data;

[0014] An edge computing module, data-connected to the digital twin modeling module, used for:

[0015] Receiving the digital sensor data sent by the digital twin modeling module;

[0016] Synchronizing the digital twin model of industrial equipment based on the physical device of the industrial equipment and the digital sensor data;

[0017] Based on the digital sensor data of the synchronized digital twin model of industrial equipment and the sensor data of the physical device of the industrial equipment obtained from the industrial Internet of Things module, synchronize the data of the two using the method of collaborative perception and reasoning;

[0018] Based on the differences between the two, the sensor data of the physical device of the industrial equipment, and the performance parameters of the industrial equipment, conduct equipment status assessment, trend prediction, and behavior warning to obtain sensor data and working condition data;

[0019] A cloud computing module, data-connected to the edge computing module, used for:

[0020] Receiving the sensor data and working condition data sent by the edge computing module;

[0021] Based on the statistical analysis, feature extraction, and intelligent learning of the sensor data and working condition data, determine the operating parameters of the equipment;

[0022] Based on the operating parameters of the equipment, optimize the parameters using an optimization algorithm;

[0023] A visualization module, which is data-connected to the cloud computing module and the edge computing module, is used for:

[0024] Visually presenting the digital twin model of industrial equipment;

[0025] Integrating the sensor data of the physical device, the digital sensor data of the digital twin model of industrial equipment, the working condition data of industrial equipment, and the equipment operation optimization information into the digital twin model, and displaying them on a three-dimensional visualization interface;

[0026] According to instructions from the user, present at least one of the sensor data of the physical device, the digital sensor data of the digital twin model of industrial equipment, the working condition data of industrial equipment, and the equipment operation optimization information on the three-dimensional visualization interface, and perform visualization.

[0027] Preferably, the digital twin modeling module is also used for:

[0028] Fusing the digital sensor data, the sensor data of the physical device, the working condition data of industrial equipment, and the equipment operation optimization information through a sensor configuration file;

[0029] Among them, the sensor configuration file is used to standardize the organization, structure, naming, and data interaction of data.

[0030] Preferably, the edge computing module is also used for:

[0031] Generating sensor data and working condition data based on the evaluation results of equipment status assessment, trend prediction, and behavior warning;

[0032] Among them, the equipment status assessment includes normal status assessment and fault status assessment;

[0033] The normal status assessment includes physical device position assessment, attitude assessment, and motion state assessment;

[0034] The fault status assessment includes physical device position collision detection and motion state collision detection;

[0035] The trend prediction includes physical device position trend and motion trend;

[0036] The behavior warning includes situations where the equipment performance parameter setting range is exceeded, the position and attitude of the digital twin model of industrial equipment are inconsistent with the physical model, and / or the position, attitude, and motion trajectory of the digital twin model of industrial equipment are inconsistent with the preset trajectory.

[0037] Preferably, the edge computing module is also used for:

[0038] Based on the evaluation results of device status assessment, trend prediction, and behavior warning, obtain the health indicators of the digital twin model and display them on the three-dimensional visualization interface.

[0039] Preferably, the optimization algorithms adopted by the cloud computing module include:

[0040] Based on a preset time span, based on the operating parameters of the device and the historical performance parameters during the device operation, optimize using an evolutionary algorithm; or

[0041] Based on a preset time span, based on the operating parameters of the device and the historical performance parameters during the device operation, combined with the historical operating parameters, optimize using a deep reinforcement learning algorithm.

[0042] Preferably, it further includes:

[0043] An industrial Internet of Things module, data-connected to the edge computing module, for:

[0044] Collect sensor data of the physical devices of industrial equipment;

[0045] Transmit the sensor data to the edge computing module.

[0046] Preferably, the visualization module is further used for:

[0047] Provide an interactive interface to allow users to view different data and information according to their needs;

[0048] Real-time update the data and information in the three-dimensional visualization interface.

[0049] Preferably, it further includes:

[0050] A data storage module, data-connected to the cloud computing module, for:

[0051] Store the sensor data, working condition data, device operating parameters, and optimized parameters;

[0052] Provide historical data support for the cloud computing module.

[0053] Preferably, the digital twin modeling module is further used for:

[0054] Construct a complete digital twin model based on a physical model, a behavior model, a rule constraint model, an operation model, and an environment model;

[0055] Among them, the physical model describes the physical characteristics such as the shape, size, and material of industrial equipment;

[0056] The behavior model describes the physical and chemical characteristics such as heat conduction and fluid mechanics of industrial equipment;

[0057] The rule constraint model includes various rules such as safe operation and minimum energy consumption;

[0058] The operation model records the control parameters and working process of industrial equipment;

[0059] The environment model considers external environmental factors such as temperature and humidity.

[0060] An intelligent optimization method for an industrial equipment digital twin system, using the described system, includes the following steps:

[0061] (1) Through the digital twin modeling module, based on 3D modeling software, construct an industrial equipment digital twin model, and establish a 3D model of industrial equipment based on CAD drawings; use virtual sensors to simulate the industrial equipment digital twin model to obtain digital sensor data;

[0062] (2) Through the edge computing module, based on the physical device of the industrial equipment and the digital sensor data obtained from the simulation, synchronize the industrial equipment digital twin model, and based on the digital sensor data of the synchronized industrial equipment digital twin model and the sensor data of the physical device of the industrial equipment obtained from the industrial Internet of Things module, use the method of collaborative perception and reasoning to synchronize the data of the two to make the data of the two synchronize in real time;

[0063] (3) After data synchronization, through the edge computing module, based on the differences between the two, the sensor data of the physical device of the industrial equipment, and the industrial equipment performance parameters, perform equipment status evaluation, trend prediction, and behavior warning to obtain sensor data and working condition data;

[0064] (4) Through the cloud computing module, based on the statistical analysis, feature extraction, and intelligent learning of the sensor data and working condition data, determine the operating parameters of the equipment; based on the operating parameters of the equipment, use an optimization algorithm to optimize the parameters;

[0065] (5) Through the visualization module, visually present the industrial equipment digital twin model, and integrate the sensor data of the physical device, the digital sensor data of the industrial equipment digital twin model, the working condition data of the industrial equipment, and the equipment operation optimization information into the digital twin model and display them on a 3D visualization interface;

[0066] (6) According to instructions from the user, present at least one of the sensor data of the physical device, the digital sensor data of the industrial equipment digital twin model, the working condition data of the industrial equipment, and the equipment operation optimization information on the 3D visualization interface and perform visualization.

[0067] The beneficial effects of the present invention are reflected at multiple levels:

[0068] First, at the macro level, the present invention constructs a complete digital twin ecosystem for industrial equipment. Through the organic combination of the digital twin modeling module, edge computing module, cloud computing module, and visualization module, full-process digital management from data collection, model construction, state synchronization to optimization control is achieved. This systematic approach not only improves the overall efficiency of industrial equipment management but also provides reliable technical support for the digital transformation of enterprises.

[0069] In terms of model construction, the present invention adopts a multi-dimensional modeling method, including physical models, behavior models, rule constraint models, operation models, and environment models. This comprehensive modeling method enables the digital twin to more accurately simulate and predict the behavior of actual industrial equipment. Especially in complex working conditions, such as equipment start-stop and load mutation, this system demonstrates superior simulation accuracy and prediction ability.

[0070] Data processing and synchronization are another innovation point of the present invention. The edge computing module adopts a method of collaborative perception and reasoning to achieve real-time synchronization of physical device and digital model data. This not only greatly reduces the data transmission volume but also significantly improves the system's response speed. For example, in the case of sudden changes in certain key parameters, the system can respond within milliseconds, much faster than traditional centralized processing methods.

[0071] In terms of intelligent optimization, the present invention combines advanced artificial intelligence algorithms and industrial domain knowledge. The cloud computing module can not only perform traditional statistical analysis and feature extraction but also continuously optimize the decision-making model through methods such as deep learning. This enables the system to adaptively handle various complex working conditions, such as changes in raw material composition and fluctuations in environmental conditions, thus achieving a more efficient and stable production process.

[0072] The innovative design of the visualization module greatly improves the usability of the system. The three-dimensional visualization interface not only intuitively displays the physical structure and operating status of the equipment but also flexibly presents various types of data and information according to user needs. This human-computer interaction method significantly reduces the cognitive load of operators and improves the accuracy and efficiency of decision-making.

[0073] At the micro level, the present invention has innovations and breakthroughs in multiple technical details. For example, in sensor data fusion, an innovative sensor profile method is adopted to effectively solve the integration problem of heterogeneous data sources. In terms of fault prediction, a combination of multiple advanced machine learning algorithms is used to greatly improve the accuracy and reliability of prediction.

[0074] It is worth mentioning that a high degree of coordination and complementarity has been achieved among the various modules of the present invention. For example, the real-time processing ability of the edge computing module complements the in-depth analysis ability of the cloud computing module, ensuring both the real-time nature of the system and the comprehensiveness and forward-looking nature of decision-making. Similarly, the high-precision simulation of the digital twin modeling module and the intuitive display of the visualization module promote each other, providing both an accurate data foundation and ensuring that this data can be effectively understood and utilized.

[0075] Generally speaking, the present invention not only solves the problems of the existing industrial equipment digital twin system in aspects such as model accuracy, data processing, and intelligent optimization, but also achieves a significant performance improvement through the synergistic effect among the modules. This innovation can not only greatly improve the operating efficiency and reliability of industrial equipment, but also provide strong technical support for the practice of Industry 4.0 and intelligent manufacturing. With the wide application of this technology, it is expected to promote the entire manufacturing industry to develop in a more efficient, intelligent, and sustainable direction. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] Figure 1 is the overall system architecture diagram of the present invention;

[0077] Figure 2 is the working flow chart of the digital twin modeling module of the present invention;

[0078] Figure 3 is the working flow chart of the edge computing module of the present invention;

[0079] Figure 4 is the working flow chart of the cloud computing module of the present invention;

[0080] Figure 5 is the working flow chart of the visualization module of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0081] As Figures 1-5 shown, the industrial equipment digital twin system of the present invention includes a digital twin modeling module 1, an edge computing module 2, a cloud computing module 3, and a visualization module 4. These modules work together to achieve a comprehensive mapping and intelligent optimization from the physical device to the digital twin.

[0082] The digital twin modeling module 1 is the foundation of this system, and its main function is to construct the digital twin model of industrial equipment. Specifically, this module first constructs the digital twin model of industrial equipment based on 3D modeling software (such as SolidWorks, AutoCAD, etc.). This process usually involves accurately modeling the geometric shape, material properties, internal structure, etc. of the equipment. For example, for an industrial machine tool, it is necessary to model its bed, spindle, workbench and other key components, and consider the relative positions and connection relationships among the components.

[0083] Next, the digital twin modeling module 1 establishes a three-dimensional model of the industrial equipment based on the CAD drawing. This step usually utilizes the import function of CAD software to convert the two-dimensional CAD drawing into a three-dimensional model. During this process, special attention needs to be paid to maintaining the accuracy and details of the model to ensure that the digital twin can accurately reflect the characteristics of the physical equipment.

[0084] Another important function of the digital twin modeling module 1 is to simulate the digital twin model of the industrial equipment using virtual sensors, thereby obtaining digital sensor data. Virtual sensors simulate the behavior of actual sensors through software and can "measure" various parameters of the model in a digital environment. For example, for an industrial machine tool, virtual temperature sensors, vibration sensors, position sensors, etc. can be set to simulate the performance of the equipment under different working conditions. The data generated by these virtual sensors provides an important basis for subsequent synchronization and optimization.

[0085] The edge computing module 2 is one of the core components of this system and is responsible for realizing data synchronization and real-time analysis between the physical equipment and the digital twin. This module first receives the digital sensor data sent by the digital twin modeling module 1. Then, based on the physical device of the industrial equipment and this digital sensor data, the digital twin model of the industrial equipment is synchronized. This process ensures that the digital twin can reflect the state of the physical equipment in real time.

[0086] A key innovation of the edge computing module 2 lies in the adoption of the method of collaborative perception and reasoning to synchronize data. Specifically, this module simultaneously processes the digital sensor data of the synchronized digital twin model of the industrial equipment and the sensor data of the physical device of the industrial equipment obtained from the industrial Internet of Things module. Through this two-way data flow, the system can more accurately capture the real-time state of the equipment and quickly identify potential anomalies.

[0087] Based on data synchronization, the edge computing module 2 further performs tasks such as equipment status assessment, trend prediction, and behavior warning. These tasks are based on three types of data: the data difference between the synchronized digital twin and the physical equipment, the sensor data of the physical device, and the predefined performance parameters of the industrial equipment. By comprehensively analyzing these data, the system can comprehensively evaluate the operating condition of the equipment, predict future trends, and issue warning signals in a timely manner.

[0088] For example, for an industrial machine tool, the system may monitor key parameters such as spindle temperature, vibration frequency, and machining accuracy. If it detects that the spindle temperature rises abnormally (e.g., exceeding the preset threshold of 80°C), the system will immediately issue a warning. At the same time, by analyzing the temperature change trend, the system can predict possible overheating failures and recommend preventive maintenance.

[0089] These analysis results are finally output in the form of sensor data and operating condition data, providing a basis for subsequent cloud analysis and optimization. Preferably, the edge computing module 2 can also generate a comprehensive health index to intuitively reflect the overall state of the device. This health index can be a score from 0 to 100, where 90 - 100 indicates excellent device state, 80 - 89 indicates good, 70 - 79 indicates average, 60 - 69 indicates attention is needed, and below 60 indicates possible serious problems.

[0090] The cloud computing module 3 is responsible for deeper data analysis and optimization. This module first receives the sensor data and operating condition data sent by the edge computing module 2. Then, based on these data, it conducts statistical analysis, feature extraction, and intelligent learning to finally determine the operating parameters of the device.

[0091] In this process, the cloud computing module 3 may apply a variety of advanced analysis techniques. For example, time series analysis can be used to identify periodic patterns in the data, clustering algorithms can be used to classify different operating states, or principal component analysis can be used to extract the most important features. These analysis results provide an important basis for subsequent parameter optimization.

[0092] Based on the determined device operating parameters, the cloud computing module 3 further adopts optimization algorithms to optimize these parameters. The present invention provides a choice of two optimization methods: evolutionary algorithms and deep reinforcement learning algorithms. These two methods have their own characteristics and can be flexibly selected according to specific application scenarios.

[0093] The optimization process of the evolutionary algorithm can be expressed as:

[0094] P t+1 =f(select(P t ),crossover(P t ),mutate(P t )),

[0095] Where P t represents the parameter population of the t-th generation, f represents the fitness function, and select, crossover, and mutate represent selection, crossover, and mutation operations respectively. The optimization process of the deep reinforcement learning algorithm can be represented by the Bellman equation:

[0096]

[0097] Where Q(s,a) represents the value of taking action a in state s, r is the immediate reward, γ is the discount factor, and s′ is the next state.

[0098] Both of these algorithms consider historical performance parameters within a preset time span, enabling long-term optimization. For example, for industrial machine tools, the system may optimize parameters such as cutting speed and feed rate to maximize production efficiency while ensuring machining quality.

[0099] The visualization module 4 is an important interface for the interaction between this system and users. This module first visually presents the digital twin model of industrial equipment, enabling users to intuitively understand the structure and status of the equipment. More importantly, this module integrates various data, such as sensor data of physical devices, digital sensor data of the digital twin model, operating condition data of industrial equipment, and equipment operation optimization information, into the digital twin model and displays them on a three-dimensional visualization interface.

[0100] For example, for industrial machine tools, users may see a realistic three-dimensional model where the colors of different components represent temperature distribution, the animation effect shows the real-time motion state, and various data indicators are presented in the form of floating text or charts. This intuitive display method greatly improves the readability and comprehensibility of the data.

[0101] In addition, the visualization module 4 also provides flexible interaction functions. Users can, according to their own needs, choose to present different types of data on the three-dimensional visualization interface, such as sensor data of physical devices, digital sensor data of the digital twin model, operating condition data, or equipment operation optimization information. This interactive design enables users to deeply explore the data they are interested in and thus make more informed decisions.

[0102] Generally speaking, the industrial equipment digital twin system of the present invention realizes seamless connection between physical devices and digital twins through the organic combination of modules such as digital twin modeling, edge computing, cloud computing, and visualization, providing powerful technical support for the real-time monitoring, status evaluation, and intelligent optimization of industrial equipment. This method not only improves the operation efficiency and reliability of the equipment but also provides a solid foundation for predictive maintenance and intelligent decision-making. In a preferred embodiment of the present invention, the digital twin modeling module 1 also has a more advanced data fusion function. Specifically, this module fuses digital sensor data, sensor data of physical devices, operating condition data of industrial equipment, and equipment operation optimization information through a sensor configuration file. The core advantage of this method lies in its ability to standardize the organization, structure, naming, and data interaction of data, thus greatly improving the efficiency and accuracy of data processing.

[0103] A sensor configuration file can be understood as a standardized data description format. In practical applications, this file is usually a document in XML or JSON format, which details information such as the attributes, units, and value ranges of each type of data. For example, for a temperature sensor, the configuration file may specify that its data type is a floating-point number, the unit is degrees Celsius, and the effective range is from -50°C to 150°C. This standardized description enables the system to automatically understand and process data from different sources.

[0104] Preferably, the sensor configuration file can also contain data conversion rules. For example, if the physical sensor outputs Fahrenheit temperature while the digital twin model uses Celsius, the configuration file can define a conversion function:

[0105]

[0106] where T C is the Celsius temperature and T F is the Fahrenheit temperature. This automatic conversion mechanism greatly simplifies the data processing flow and reduces the possibility of human errors.

[0107] In addition, the sensor configuration file can also define information such as the timestamp format and sampling frequency of the data, ensuring that data from different sources can be correctly aligned in the time dimension. This is crucial for achieving precise synchronization between physical devices and digital twins.

[0108] By using the sensor configuration file for data fusion, the system of the present invention can more flexibly process various types of data, whether these data come from physical sensors or virtual sensors. This method not only improves the scalability of the system but also facilitates the integration of new types of sensors or data sources in the future.

[0109] In another embodiment of the present invention, the functions of the edge computing module 2 are further extended and refined. This module not only performs simple data processing but also generates more comprehensive and in-depth sensor data and operating condition data based on the evaluation results of device status assessment, trend prediction, and behavior warning.

[0110] Specifically, device status assessment includes two main aspects: normal status assessment and fault status assessment. Normal status assessment mainly focuses on the position assessment, attitude assessment, and motion state assessment of the physical device. For example, for an industrial robot, the system will continuously monitor parameters such as the angles of its various joints, the spatial position of the end effector, the motion speed, and the acceleration. These parameters usually need to meet predetermined accuracy requirements, such as the position accuracy may need to reach ±0.1 mm and the attitude accuracy may need to reach ±0.1°.

[0111] Fault status assessment mainly includes physical device position collision detection and motion state collision detection. Here, "collision" not only refers to physical contact, but also includes situations where parameters exceed the safe range. For example, if it is detected that a certain joint angle of a robot exceeds its mechanical limit (usually specified by the manufacturer), the system will immediately trigger a collision alarm. Similarly, if it is detected that the robot's motion trajectory may collide with other devices or obstacles in the working environment, the system will also give a timely warning.

[0112] Trend prediction is another important innovation point of the present invention. This function mainly includes the prediction of physical device position trends and motion trends. The system analyzes historical data and uses advanced time series analysis techniques (such as ARIMA model, Prophet model, etc.) to predict the future state of the device. For example, for a numerically controlled machine tool, the system may predict the wear trend of the spindle bearing, so as to arrange maintenance in advance before a fault occurs.

[0113] The general form of the prediction model can be expressed as:

[0114] Y t = f(Y t-1 ,Y t-2 ,…,Y t-n ,X t ) + ∈ t ,

[0115] where, Y t is the predicted value at time t, Y t -1, Y t-2 ,…,Y t-n are historical observations, X t is other relevant factors, ∈ t is the error term. The specific form of the f function will depend on the prediction model used.

[0116] Behavior warning is the third key function of the edge computing module 2. The system of the present invention can identify and warn of various abnormal situations, including:

[0117] 1. The situation where device parameters exceed the performance setting range. For example, if the operating temperature of the motor exceeds 85°C (assuming the normal operating temperature range is -20°C to 80°C), the system will issue an overheat warning.

[0118] 2. The situation where the position and attitude of the digital twin model are inconsistent with the physical model. This may mean that the physical device has undergone unexpected changes, or the digital model needs to be recalibrated.

[0119] 3. The situation where the position, orientation, and motion trajectory of the digital twin model are inconsistent with the preset trajectory. This may indicate a problem with the device's control system or a change in the external environment.

[0120] Through these detailed evaluation and early warning mechanisms, the system of the present invention can comprehensively grasp the operating status of industrial equipment, promptly discover potential problems, thereby greatly improving the reliability and production efficiency of the equipment.

[0121] In another preferred embodiment of the present invention, the edge computing module 2 also has the function of generating health indicators. Specifically, based on the evaluation results of equipment status assessment, trend prediction, and behavior early warning, this module calculates a comprehensive health indicator and intuitively displays it on the three-dimensional visualization interface.

[0122] This health indicator can be understood as a quantitative representation of the overall state of the equipment, usually using a scoring system from 0 to 100. For example, the following calculation formula can be adopted:

[0123] H = w1S + w2T + w3A,

[0124] where H is the health indicator, S is the status assessment score, T is the trend prediction score, A is the behavior early warning score, w1, w2, and w3 are the corresponding weight coefficients, and w1 + w2 + w3 = 1.

[0125] Specific scoring criteria can be formulated based on industry experience and expert knowledge. For example:

[0126] 90 - 100 points: The equipment is in excellent condition, and all indicators are within the ideal range;

[0127] 80 - 89 points: The equipment is in good condition, and there may be slight deviations but it does not affect normal operation;

[0128] 70 - 79 points: The equipment is in general condition, and it is recommended to closely monitor and consider preventive maintenance;

[0129] 60 - 69 points: The equipment status requires attention, and there may be potential problems. It is recommended to check as soon as possible;

[0130] Below 60 points: The equipment is in an abnormal state, and there may be serious problems. It is recommended to immediately stop the machine for maintenance;

[0131] This intuitive health indicator can not only help operators quickly judge the equipment status but also provide decision-making support for managers. For example, when the health indicator is lower than 80 points, the system can automatically generate maintenance suggestions; when the indicator is lower than 60 points, it can trigger an emergency shutdown procedure.

[0132] In this way, the system of the present invention transforms complex multi-dimensional data into simple and understandable metrics, greatly improving the efficiency and accuracy of industrial equipment management. This not only helps to detect and solve problems in a timely manner, but also optimizes maintenance plans, extends equipment life, and ultimately enhances production efficiency and reduces costs. In a preferred embodiment of the present invention, the cloud computing module 3 employs two advanced optimization algorithms to optimize the operating parameters of the equipment. These two algorithms are the evolutionary algorithm and the deep reinforcement learning algorithm, respectively, and they are each applicable to different scenarios, providing flexible optimization strategy options for the system.

[0133] First, the application of the evolutionary algorithm is based on a preset time span, and it uses the operating parameters of the equipment and the historical performance parameters during the operation of the equipment for optimization. The core idea of the evolutionary algorithm is to simulate the biological evolution process, and continuously optimize the quality of the solution through operations such as selection, crossover, and mutation. In the present invention, the operating parameters of the equipment can be encoded as "chromosomes", and then the optimal parameter combination is found through iterative evolution.

[0134] For example, for an injection molding machine, its key parameters may include mold temperature, injection speed, holding pressure time, etc. The optimization process of the evolutionary algorithm can be expressed as:

[0135] P t+1 = f(select(P t ), crossover(P t ), mutate(P t ))

[0136] where P t represents the parameter population of the t-th generation, f is the fitness function used to evaluate the quality of the parameters. select, crossover, and mutate represent the selection, crossover, and mutation operations respectively. Through multiple generations of iteration, the algorithm can gradually find the parameter combination that can maximize production efficiency or minimize energy consumption. On the other hand, the deep reinforcement learning algorithm is based on a preset time span, and combines the operating parameters of the equipment, historical performance parameters, and historical operating parameters for optimization. This method is particularly suitable for complex and dynamically changing industrial environments. Deep reinforcement learning learns the optimal decision-making strategy through continuous interaction with the environment. In the present invention, the deep reinforcement learning algorithm can be expressed as:

[0137]

[0138] Here, Q(s,a) represents the value of taking action a in state s, r is the immediate reward, γ is the discount factor, and s′ is the next state. By continuously updating the Q value, the system can learn what actions should be taken (i.e., how to adjust the parameters) in different states to obtain the maximum long-term benefit.

[0139] Preferably, the system can dynamically select which optimization algorithm to use according to specific circumstances. For example, when the working environment is relatively stable, an evolutionary algorithm with higher computational efficiency can be selected; while in the face of complex and changeable situations, a more adaptable deep reinforcement learning algorithm can be adopted.

[0140] Another embodiment of the present invention introduces an industrial Internet of Things module, which further enhances the data acquisition ability of the system. This module is data-connected to the edge computing module 2 and is mainly responsible for collecting sensor data of the physical devices of industrial equipment and transmitting this data to the edge computing module 2.

[0141] The introduction of the industrial Internet of Things module enables this system to obtain more comprehensive real-time status information of the equipment. For example, for a production line, the industrial Internet of Things module may include a variety of sensors:

[0142] 1. Temperature sensors: Monitor the temperature of key components, such as motors, bearings, etc.

[0143] 2. Vibration sensors: Detect the vibration of the equipment and can detect mechanical failures at an early stage.

[0144] 3. Current sensors: Monitor the power consumption of the equipment and can judge the load status of the equipment.

[0145] 4. Pressure sensors: Used to monitor the working status of hydraulic or pneumatic systems.

[0146] 5. Position sensors: Track the position of moving parts to ensure motion accuracy.

[0147] These sensors usually use industrial-level communication protocols (such as Modbus, Profinet, OPCUA, etc.) to transmit data to the edge computing module 2. Preferably, encryption measures should be adopted during the data transmission process to ensure the security of the data.

[0148] In yet another preferred embodiment of the present invention, the function of the visualization module 4 is further enhanced. In addition to the basic data display function, this module also provides a rich interactive interface, allowing users to view different data and information according to their needs. At the same time, the visualization module 4 can update the data and information in the three-dimensional visualization interface in real time, providing users with a dynamic and intuitive display of the equipment status.

[0149] Specifically, the visualization module 4 can provide the following advanced functions:

[0150] 1. Multi-view switching: Users can freely switch between multiple views such as the overall view, local detail view, data trend chart, etc.

[0151] 2. Parameter customization: Users can customize the parameters that need to be focused on, and these parameters will be presented on the interface in a prominent manner.

[0152] 3. Alarm level setting: Users can set alarm thresholds for different parameters. When the parameters exceed the thresholds, the system will remind users by means of color change, flashing, etc.

[0153] 4. Historical data playback: The system supports the historical data playback function. Users can select a specific time period to view the operating status of the device during that period.

[0154] 5. Visualization of prediction results: Intuitively display the future trends predicted by the system in the form of charts to help users make preventive decisions.

[0155] The implementation of these functions depends on advanced front-end technologies such as WebGL and Three.js. Through these technologies, the visualization module 4 can present smooth 3D effects in the browser while ensuring real-time updates of a large amount of data.

[0156] In order to further improve the performance and reliability of the system, a data storage module is also included in an embodiment of the present invention. This module is connected to the cloud computing module 3 for data, and is mainly responsible for storing sensor data, working condition data, device operation parameters, and optimized parameters. At the same time, the data storage module also provides historical data support for the cloud computing module 3, which is crucial for long-term trend analysis and optimization.

[0157] The data storage module usually adopts a distributed storage architecture to meet the storage requirements of massive data in the industrial environment. For example, the Hadoop Distributed File System (HDFS) can be used to store raw data, and NoSQL databases such as HBase or Cassandra can be used to store structured and semi-structured data. To improve query efficiency, a search engine such as Elasticsearch can also be used to establish indexes.

[0158] Preferably, the data storage module should also implement the functions of hierarchical storage and automatic archiving of data. For example, the high-frequency data of the most recent month can be stored in high-speed storage devices, while the earlier data can be compressed and stored in low-speed large-capacity devices. This strategy not only ensures the response speed of the system but also takes into account the needs of long-term data storage.

[0159] Through the above optimizations and expansions, the industrial equipment digital twin system of the present invention has reached a very high level in aspects such as data processing, parameter optimization, information display, etc., providing strong technical support for the intelligent management of industrial equipment. In an important embodiment of the present invention, the function of the digital twin modeling module 1 has been further expanded and deepened. This module not only constructs a simple 3D model, but also constructs a comprehensive and complex digital twin model based on five key aspects: physical model, behavior model, rule constraint model, operation model, and environment model. This multi-dimensional modeling method enables the digital twin to more accurately simulate and predict the behavior of actual industrial equipment.

[0160] Specifically, the physical model describes the physical characteristics of industrial equipment such as shape, size, material, etc. For example, for an industrial robot, the physical model will include information such as the geometric dimensions of each joint, material properties (such as density, elastic modulus, etc.), and mass distribution. These parameters are crucial for accurately simulating the kinematic and dynamic characteristics of the equipment.

[0161] The behavior model focuses on describing various physical and chemical characteristics of industrial equipment, such as heat conduction, fluid mechanics, etc. Taking a heat treatment furnace as an example, its behavior model may include the heat conduction equation:

[0162]

[0163] where T is temperature, t is time, α is the thermal diffusivity, is the Laplace operator. By solving this equation, the system can simulate the change of temperature distribution during the heat treatment process.

[0164] The rule constraint model includes various operation rules such as safe operation and minimum energy consumption. These rules can be expressed as a series of constraint conditions, for example:

[0165]

[0166] where T represents temperature, P represents power, and E represents energy consumption. These constraint conditions ensure that the equipment operates in a safe and efficient state.

[0167] The operation model records the control parameters and working process of industrial equipment. This may include the startup sequence of the equipment, standard operating procedures, exception handling processes, etc. Preferably, the operation model can be represented by a state machine, where each state corresponds to a working stage of the equipment, and the transition between states is determined by specific trigger conditions.

[0168] The environment model takes into account external environmental factors such as temperature and humidity. These factors may significantly affect the performance and lifespan of the equipment. For example, for precision instruments, the environment model may include the following equation for the influence of humidity on measurement accuracy:

[0169] ΔL = αL(H - H0),

[0170] where ΔL is the length change caused by humidity change, L is the original length, α is the humidity expansion coefficient, and H and H0 are the current humidity and reference humidity respectively.

[0171] By integrating the models in these five aspects, the digital twin of the present invention can comprehensively and accurately simulate the behavior of actual industrial equipment. This not only improves the accuracy of the simulation but also lays a solid foundation for subsequent condition monitoring, fault diagnosis, and optimization control. The present invention can improve the model performance through a distributed training method and also includes a lightweight security module for protecting edge devices.

[0172] Finally, the present invention also proposes an intelligent optimization method for an industrial equipment digital twin system, which systematically integrates the functions of the foregoing various modules to form a complete workflow. The method includes the following steps:

[0173] First, through the digital twin modeling module 1, a digital twin model of the industrial equipment is constructed based on 3D modeling software, and a 3D model of the industrial equipment is established based on the CAD drawing. This step provides a basis for subsequent analysis and optimization. Preferably, various factors such as the geometric features, material properties, and internal structure of the equipment should be considered during the modeling process to ensure the accuracy of the model.

[0174] Next, the digital twin model of the industrial equipment is simulated using virtual sensors to obtain digital sensor data. These virtual sensors can simulate the measurement of various physical quantities, such as temperature, pressure, vibration, etc. The simulation process may involve solving complex physical equations, such as finite element analysis or computational fluid dynamics simulation.

[0175] Then, through the edge computing module 2, the digital twin model of the industrial equipment is synchronized based on the physical device of the industrial equipment and the digital sensor data obtained from the foregoing simulation. This step ensures that the digital twin can reflect the state of the physical device in real time. During the synchronization process, the system adopts a method of collaborative perception and reasoning, comprehensively using the digital sensor data of the digital twin model and the physical device sensor data obtained from the industrial Internet of Things module.

[0176] After data synchronization, the edge computing module 2 performs equipment status evaluation, trend prediction, and behavior warning based on the differences between the physical device and the digital twin, the sensor data of the physical device, and the preset performance parameters of the industrial equipment. The output of this step is the processed sensor data and working condition data, providing a basis for subsequent optimization.

[0177] Next, the cloud computing module 3 performs statistical analysis, feature extraction, and intelligent learning based on this sensor data and operating condition data to determine the operating parameters of the device. On this basis, the system uses an optimization algorithm (such as the aforementioned evolutionary algorithm or deep reinforcement learning algorithm) to optimize these parameters.

[0178] Finally, through the visualization module 4, the system visually presents the digital twin model of the industrial device, and integrates the sensor data of the physical device, the digital sensor data of the digital twin model, the operating condition data, and the operation optimization information into the model and displays them on the three-dimensional visualization interface. Users can select to present different types of data on the interface according to their needs and perform visual analysis.

[0179] The advantage of this method is that it realizes the full-process automation from data acquisition, model synchronization, state evaluation to parameter optimization. Through digital twin technology, the system can comprehensively analyze and optimize industrial devices in a virtual environment, greatly improving the accuracy and efficiency of decision-making. At the same time, the real-time visualization function makes complex data intuitive and easy to understand, helping operators quickly identify potential problems and respond.

[0180] Generally speaking, the industrial device digital twin system and its intelligent optimization method proposed in the present invention provide a comprehensive and powerful solution for the intelligent management of industrial devices by integrating advanced modeling technology, edge computing, cloud computing, artificial intelligence algorithms, and visualization technology. This system can not only improve the operating efficiency and reliability of devices, but also support advanced applications such as predictive maintenance and energy consumption optimization, and is expected to play an important role in the fields of Industry 4.0 and intelligent manufacturing.

[0181] To verify the effectiveness and superiority of the present invention, a typical industrial application scenario - the core device distillation column of a large petrochemical plant - was selected for simulation testing. The distillation column is a key device in the petrochemical industry, and its operating efficiency and stability directly affect the quality and economic benefits of the entire production process.

[0182] Example 1 adopted the complete system of the present invention, including a digital twin modeling module, an edge computing module, a cloud computing module, and a visualization module, and implemented the intelligent optimization method. Comparative Example 1 adopted a traditional rule-based control system without digital twin technology. Comparative Example 2 adopted partial digital twin technology but did not implement the intelligent optimization method.

[0183] During the test process, the following five indicators were focused on: energy efficiency, product quality stability, equipment fault prediction accuracy, decision response time, and system scalability. These indicators directly reflect the core innovation points and practical application value of the present invention.

[0184] Energy efficiency is evaluated by measuring the energy consumption per unit product, with the unit of kWh / ton. The product quality stability is measured by the standard deviation of the main product indicators, which is dimensionless. The accuracy of equipment fault prediction is evaluated using the F1 score, with a value range of 0 - 1. The decision response time refers to the time from the occurrence of an anomaly to the system's response, with the unit of seconds. The system scalability is evaluated by the system's ability to process newly added sensor data, with the unit of the number of data points that can be processed per second.

[0185] The test was continuously carried out for 6 months, during which various normal and abnormal working conditions were simulated. The test results are shown in the following table:

[0186]

[0187]

[0188] It can be seen from the test results that Example 1 is significantly superior to Comparative Example 1 and Comparative Example 2 in all indicators.

[0189] In terms of energy efficiency, Example 1 improved by 15.0% and 6.9% compared to Comparative Example 1 and Comparative Example 2 respectively. This is mainly due to the intelligent optimization method of the present invention, which can dynamically adjust the operating parameters of the distillation column, such as the reflux ratio and the top temperature, according to the real-time operating data and the prediction model, so as to minimize the energy consumption while ensuring the product quality.

[0190] The improvement in product quality stability is even more significant. The standard deviation of Example 1 decreased by 46.8% and 24.2% compared to Comparative Example 1 and Comparative Example 2 respectively. This reflects the strength of the digital twin technology combined with the intelligent optimization method. By synchronizing the physical device and the digital model in real time, the system can more precisely control the production process and reduce fluctuations.

[0191] In terms of fault prediction, the F1 score of Example 1 reached 0.92, which improved by 26.0% and 8.2% compared to Comparative Example 1 and Comparative Example 2 respectively. This benefits from the comprehensive data acquisition and analysis capabilities of the system of the present invention, as well as the advanced machine learning algorithms. The system can not only identify known fault patterns but also discover potential abnormal patterns, greatly improving the accuracy and comprehensiveness of the prediction.

[0192] The improvement in decision response time is particularly significant. Example 1 only needs 1.2 seconds to respond to abnormal situations, which is 79.3% and 55.6% faster than Comparative Example 1 and Comparative Example 2 respectively. This is mainly attributed to the edge computing module of the present invention, which can quickly process data and make preliminary decisions on the device side without the need for all data to be transmitted to the cloud for processing.

[0193] In terms of system scalability, Example 1 demonstrated excellent performance, capable of processing 100,000 data points per second, which is 10 times that of Comparative Example 1 and 2 times that of Comparative Example 2. This high scalability stems from the distributed architecture of the system of the present invention and the efficient data processing algorithm, enabling it to easily handle potential future system expansion requirements.

[0194] These test results fully demonstrate the superiority of the "Industrial Equipment Digital Twin System and Intelligent Optimization Method" of the present invention. It can not only significantly improve production efficiency and product quality, but also effectively reduce energy consumption and improve equipment reliability. More importantly, this system exhibits excellent real-time performance and scalability, which is particularly important for modern complex industrial production environments.

[0195] It is worth mentioning that during the testing process, Example 1 also demonstrated some additional advantages. For example, in the face of emergencies such as sudden changes in raw material composition or drastic changes in environmental temperature, Example 1 can quickly adjust the operating parameters to minimize the impact. This rapid adaptation ability is of great significance in actual production, which can effectively reduce unplanned downtime and product quality fluctuations.

[0196] Generally speaking, the present invention provides a comprehensive, efficient and reliable solution for the intelligent management of industrial equipment. It can not only improve the efficiency and stability of existing production lines, but also lay a technical foundation for the construction of future smart factories. With the in-depth promotion of Industry 4.0, this system based on digital twin and intelligent optimization will surely play an increasingly important role.

[0197] It should be noted that the above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the principles of the present invention shall be included within the protection scope of the present invention.

Claims

1. Industrial equipment digital twin system, characterized by: include: Digital twin modeling module for: Build digital twin models of industrial equipment based on 3D modeling software; Build 3D models of industrial equipment based on CAD drawings; Use virtual sensors to simulate the digital twin model of industrial equipment and obtain digital sensor data; The edge computing module is connected to the digital twin modeling module for: Receiving digitized sensor data sent by the digital twin modeling module; Synchronizing a digital twin model of the industrial equipment based on a physical device of the industrial equipment and the digitized sensor data; Based on the synchronized digital sensor data of the industrial equipment digital twin model and the sensor data of the physical device of the industrial equipment obtained from the industrial Internet of Things module, the collaborative perception and reasoning method is used to synchronize the data of the two. Based on the difference between the two, the sensor data of the physical device of the industrial equipment and the performance parameters of the industrial equipment, equipment status evaluation, trend prediction and behavior warning are performed to obtain sensor data and working condition data; The cloud computing module is data-connected to the edge computing module and is used to: Receiving sensor data and operating condition data sent by the edge computing module; Determine the operating parameters of the equipment based on statistical analysis, feature extraction and intelligent learning of the sensor data and operating condition data; Based on the operating parameters of the equipment, optimizing the parameters using an optimization algorithm; A visualization module, data-connected to the cloud computing module and the edge computing module, is used to: Visualize the digital twin model of industrial equipment; Integrate the sensor data of the physical device, the digitized sensor data of the digital twin model of the industrial equipment, the working condition data of the industrial equipment, and the equipment operation optimization information into the digital twin model, and display them on a three-dimensional visualization interface; According to instructions from the user, at least one of sensor data based on the physical device, digitized sensor data of the digital twin model of industrial equipment, operating condition data of the industrial equipment, and equipment operation optimization information is presented and visualized on a three-dimensional visualization interface.

2. The system according to claim 1, characterized in that The digital twin modeling module is also used to: fusing the digital sensor data, the sensor data of the physical device, the working condition data of the industrial equipment, and the equipment operation optimization information through the sensor configuration file; The sensor configuration file is used to standardize the organization, structure, naming and data interaction of data.

3. The system according to claim 1, characterized in that The edge computing module is also used for: Generate sensor data and operating condition data based on the evaluation results of equipment status evaluation, trend prediction and behavior warning; Wherein, the equipment status assessment includes normal status assessment and fault status assessment; Normal state assessment includes physical device position assessment, posture assessment, and motion state assessment; Fault status assessment includes physical device position collision detection and motion status collision detection; Trend prediction includes physical device location trends and movement trends; The behavior warning includes situations where the equipment performance parameters exceed the set range, where the position and posture of the digital twin model of the industrial equipment are inconsistent with the physical model, and / or where the position, posture and motion trajectory of the digital twin model of the industrial equipment are inconsistent with the preset trajectory.

4. The system according to claim 3, characterized in that The edge computing module is also used for: Based on the evaluation results of equipment status assessment, trend prediction and behavior warning, the health indicators of the digital twin model are obtained and displayed on a three-dimensional visualization interface.

5. The system according to claim 1, characterized in that The optimization algorithms used by the cloud computing module include: Based on a preset time span, based on the operating parameters of the device, and the historical performance parameters of the device during operation, an evolutionary algorithm is used for optimization; or Based on a preset time span, based on the operating parameters of the equipment and the historical performance parameters of the equipment during operation, combined with the historical operating parameters, a deep reinforcement learning algorithm is used for optimization.

6. The system according to claim 1, characterized in that Also includes: The industrial Internet of Things module is connected to the edge computing module for: Collect sensor data from physical devices of industrial equipment; The sensor data is transmitted to the edge computing module.

7. The system according to claim 1, characterized in that The visualization module is also used for: Provide an interactive interface that allows users to view different data and information according to their needs; Update data and information in the 3D visualization interface in real time.

8. The system according to claim 1, characterized in that Also includes: The data storage module is connected to the cloud computing module and is used to: Storing the sensor data, operating condition data, equipment operating parameters and optimized parameters; Provide historical data support for the cloud computing module.

9. The system according to claim 1, characterized in that The digital twin modeling module is also used to: Build a complete digital twin model based on physical model, behavior model, rule constraint model, operation model and environment model; Among them, the physical model describes the physical characteristics of industrial equipment, such as shape, size, and material; Behavioral models describe the physical and chemical properties of industrial equipment, such as heat transfer and fluid mechanics; The rule-constrained model contains various rules such as safe operation and minimum energy consumption; Operational models record the control parameters and working processes of industrial equipment; The environmental model takes into account external environmental factors such as temperature and humidity.

10. An intelligent optimization method for a digital twin system of industrial equipment, using the system according to any one of claims 1 to 9, characterized in that: The following steps are involved: (1) Using the digital twin modeling module, build a digital twin model of industrial equipment based on 3D modeling software, and build a 3D model of industrial equipment based on CAD drawings; use virtual sensors to simulate the digital twin model of industrial equipment to obtain digital sensor data; (2) synchronizing the digital twin model of the industrial equipment based on the physical device of the industrial equipment and the digital sensor data obtained by the simulation through the edge computing module, and synchronizing the data of the digital twin model of the industrial equipment based on the digital sensor data of the synchronized digital twin model of the industrial equipment and the sensor data of the physical device of the industrial equipment obtained from the industrial Internet of Things module by using a collaborative perception and reasoning method so that the data of the two are synchronized in real time; (3) After data synchronization, the edge computing module performs equipment status evaluation, trend prediction, and behavior warning based on the differences between the two, the sensor data of the physical device of the industrial equipment, and the performance parameters of the industrial equipment to obtain sensor data and working condition data; (4) determining the operating parameters of the equipment based on statistical analysis, feature extraction, and intelligent learning of the sensor data and the operating condition data through a cloud computing module; and optimizing the parameters using an optimization algorithm based on the operating parameters of the equipment; (5) Visualizing the digital twin model of industrial equipment through a visualization module, integrating sensor data of the physical device, digitized sensor data of the digital twin model of industrial equipment, operating condition data of the industrial equipment, and equipment operation optimization information into the digital twin model, and displaying them on a three-dimensional visualization interface; (6) Based on instructions from the user, at least one of the sensor data based on the physical device, the digitized sensor data of the digital twin model of the industrial equipment, the operating condition data of the industrial equipment, and the equipment operation optimization information is presented and visualized on a three-dimensional visualization interface.

Images