A device digital twin model updating method, system, terminal and medium

Abstract

This application discloses a method, system, terminal, and medium for updating a digital twin model of a device. The method includes the following steps: collecting data from several locations of the device; dynamically adjusting the monitoring frequency for any data collection point on the device based on the sensitivity index of the monitoring data change; preprocessing the collected data; updating the digital twin model of the device based on the processed data; when updating the model: dynamically adjusting the allocation of computing resources based on the changing trend of the data; and using an incremental learning algorithm to update the parts of the model affected by changes in physical entities. The advantage of this application is that it can improve the update efficiency of the digital twin model of the device, significantly improve the real-time performance and reliability of the digital twin system, and provide strong support for the safe operation, maintenance, and management of large-scale equipment.

Description

Technical Field

[0001] The present application relates to the field of digital twin technology, and in particular to a method, system, terminal and medium for updating a digital twin model of a device. Background Art

[0002] A key issue in the application of digital twin technology is real-time model updates. As physical entities operate, their states constantly change, such as due to wear and tear on equipment and structural deformation. Digital twin models need to be updated accordingly to accurately reflect these changes. However, existing technologies have numerous limitations. For example, the limited frequency and coverage of sensors prevents them from fully and timely capturing changes in physical entities. Furthermore, the algorithms and computing resources required to update the models are complex, restricting the real-time performance of digital twin systems. This is particularly true for large equipment, such as large cranes. Structural changes under wind loads and loads are difficult to reflect in real time in digital twin models, significantly impacting the effectiveness and application value of digital twin technology. Summary of the Invention

[0003] This application provides a device digital twin model update method, system, terminal and medium, which has the advantage of improving the update efficiency of the device digital twin model, significantly improving the real-time performance and reliability of the digital twin system, and providing strong support for the safe operation, maintenance and management of large-scale equipment.

[0004] The technical solution of this application is as follows:

[0005] In one aspect, the present application provides a method for updating a device digital twin model, comprising the following steps:

[0006] S1: collects data from several locations of the device;

[0007] For any data collection point on the device, the monitoring frequency is dynamically adjusted according to the sensitivity index of the monitoring data change;

[0008] S2: Preprocessing of collected data, including data cleaning, outlier detection and noise reduction;

[0009] S3: Update the digital twin model of the device based on the processed data. When the model is updated:

[0010] Select an update algorithm based on the data type. For structural deformation strain data, an update algorithm based on finite element analysis is used. For equipment wear data, an algorithm based on historical wear data and machine learning is used to predict wear trends and update the component wear status in the model.

[0011] Dynamically adjust the allocation of computing resources based on data trends, and select a model update algorithm based on the allocation ratio of computing resources;

[0012] An incremental learning algorithm is used to update the parts of the model that are affected by changes in physical entities.

[0013] Furthermore, in step S1, the collected equipment data includes strain data, displacement data and vibration data of different parts of the equipment.

[0014] Furthermore, in step S1, a distributed sensor network is used to collect data from several locations of the device, and the collection frequency of the sensor is obtained by the following formula:

[0015]

[0016] in, is the initial monitoring frequency of the sensor, k is the adjustment coefficient, and S is the sensitivity index of the change of the monitoring data;

[0017] The change sensitivity index S of the monitoring data is obtained by any of the following methods:

[0018] (1) For structural strain data monitoring:

[0019]

[0020] : The variance of the data in the current time window; Time window length T: set according to the response time of the physical entity;

[0021] : Historical maximum variance, historical variance update cycle: daily or weekly update;

[0022] (2) For displacement data and vibration data:

[0023]

[0024] : Time gradient of data; gradient calculation cycle : consistent with the sensor sampling frequency;

[0025] : historical maximum gradient;

[0026] When the time gradient of the data exceeds the preset threshold multiple times continuously, it is judged as abnormal and an abnormal warning is issued;

[0027] (3) For complex working conditions of equipment:

[0028]

[0029] : Weight coefficient; through historical data regression analysis, select the model that minimizes the prediction error , value.

[0030] Furthermore, in step S1 , a distributed sensor network is used to collect data from several locations of the device, and sensors in the sensor network communicate with each other.

[0031] Furthermore, in step S2, the statistical analysis method is used to identify and remove abnormal data points. The 3σ principle is used. If the data x satisfies , is the mean, σ is the standard deviation, then the data is judged to be an outlier and excluded;

[0032] Moving average filtering is used to reduce the noise of the data. The formula is:

[0033]

[0034] in is the filtered value of the nth data point, are the original data points and N is the size of the moving average window.

[0035] Furthermore, in step S3, the allocation of computing resources is dynamically adjusted according to the changing trend of the data as follows:

[0036] When a set of data is used to update the device digital twin model, the computing resource allocation ratio R of the set of data is:

[0037]

[0038] Where a and b are coefficients determined based on system performance and equipment characteristics;

[0039] Entropy H can be calculated using the Shannon entropy formula:

[0040]

[0041] is the probability distribution of the data.

[0042] Furthermore, in step S3, an incremental learning algorithm is used to update the part of the model affected by the change of the physical entity;

[0043] The incremental update of the finite element model is based on the variational principle and the virtual work principle; assuming that the structural deformation , which satisfies the following variational equation:

[0044]

[0045] in: is the domain of the structure, σ is the stress tensor, is the strain tensor, It's physical strength. is the surface force on the boundary, is the virtual displacement; by solving this equation, the deformation state of the structure is updated, and then the structural information in the digital twin model is updated.

[0046] In another aspect, the present application provides a device digital twin model update system, comprising:

[0047] A data acquisition unit, comprising a plurality of data acquisition devices distributedly deployed at a plurality of locations on the equipment to collect data from the equipment; the data acquisition device comprises a sensor module, a data processing module, and a communication module; the sensor module comprises one or more of a strain sensor, a displacement sensor, and a vibration sensor; and the data processing module is used to pre-process the data collected by the sensor module;

[0048] A data center is communicatively connected to a data acquisition device and configured to receive data from the data acquisition device. The data center updates the digital twin model of the equipment based on the received data. When updating the model, an update algorithm is selected based on the data type. For strain data of structural deformation, an update algorithm based on finite element analysis is used. For equipment wear data, an algorithm based on historical wear data and machine learning is used to predict wear trends and update the wear status of components in the model. The allocation of computing resources is dynamically adjusted based on the changing trends of the data. An incremental learning algorithm is used to update the parts of the model affected by changes in physical entities.

[0049] and a storage and display unit for storing and displaying digital twin models.

[0050] On the other hand, the present application provides a digital twin model terminal, including a memory, a processor and a display, wherein the display is used to display the digital twin model, and the memory stores a computer program. When the computer program is called and executed by the processor, the device digital twin model update method as described above is implemented.

[0051] On the other hand, the present application provides a computer-readable medium, characterized in that the computer-readable medium memory stores a computer program, and when the computer program is called and executed by a computer, it implements the device digital twin model updating method as described above.

[0052] In summary, the beneficial effects of this application are:

[0053] 1. Dynamically adjust the monitoring frequency based on the operating status and environment of the physical entity, achieving optimal resource allocation. When the data volume is small, less data is collected, reducing the amount of data processing. The monitoring frequency is calculated using the sensitivity index of the monitoring data changes, allowing the sensor to intelligently adapt to the data collection needs under different working conditions.

[0054] 2. Use the 3σ principle and moving average filtering to accurately detect outliers and reduce data noise, providing high-quality data for subsequent model updates;

[0055] 3. Flexibly selecting update algorithms based on different data types and dynamically allocating computing resources based on data changes avoids the limitations of fixed algorithms and resource allocations in traditional methods, improving computing resource utilization efficiency. Dynamic allocation of computing resources based on data entropy enables adaptive resource adjustment. The introduction of an incremental learning algorithm updates only the affected model parts, significantly reducing the computational effort and time required for updates and making model updates more real-time. Incremental updates of finite element models are achieved using the variational principle and the principle of virtual work, ensuring the physical rationality and accuracy of model updates. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 It is a schematic diagram of the device digital twin model update system in this application. DETAILED DESCRIPTION

[0057] The specific implementation of the present application is described in detail below with reference to the accompanying drawings.

[0058] A specific embodiment of the present application provides a method for updating a device digital twin model, comprising the following steps:

[0059] S1: collects data from several locations of the device;

[0060] For any data collection point on the device, the monitoring frequency is dynamically adjusted according to the sensitivity index of the monitoring data change;

[0061] The collected equipment data includes strain data, displacement data and vibration data of different parts of the equipment.

[0062] A distributed sensor network is used to collect data from several locations of the device. The sensor collection frequency is obtained by the following formula:

[0063]

[0064] in, is the initial monitoring frequency of the sensor, k is the adjustment coefficient, and S is the sensitivity index of the change of the monitoring data;

[0065] The change sensitivity index S of the monitoring data is obtained by any of the following methods:

[0066] (1) Based on data variance calculation, it is applicable to scenarios where the data changes significantly but slowly, typically for structural strain data monitoring:

[0067]

[0068] : The variance of the data in the current time window; the time window length T: is set according to the response time of the physical entity (for example, crane strain monitoring, T = 10s);

[0069] : Historical maximum variance (used for normalization), historical variance update cycle: daily or weekly updates to prevent long-term drift.

[0070] (2) Based on data gradient calculation, it is applicable to scenarios where the data change rate is sensitive, typically for vibration monitoring and displacement data detection:

[0071]

[0072] : Time gradient of data; gradient calculation cycle : consistent with the sensor sampling frequency (for example = 0.1s);

[0073] : historical maximum gradient;

[0074] When the time gradient of the data exceeds the preset threshold multiple times continuously, it is judged as abnormal and an abnormal warning is issued.

[0075] (3) Based on the calculation of data mixed indicators (variance + gradient), it is applicable to complex working conditions with comprehensive change amplitude and rate, typically for crane swing monitoring under wind load:

[0076]

[0077] : Weight coefficient, which needs to be optimized according to the scenario (for example, =0.6, =0.4); through historical data regression analysis, select the one that minimizes the model prediction error , value.

[0078] Distributed sensor networks consist of multiple smart sensors deployed at key locations and components of physical entities. Smart sensors not only possess high precision and sensitivity but also possess intelligent processing capabilities, enabling preliminary processing of collected data. For example, large cranes utilize various types of sensors, including strain sensors, displacement sensors, and vibration sensors, installed in key load-bearing structures, key connecting components, and transmission mechanisms.

[0079] The sensors used are shown in Table 1:

[0080] Table 1: Sensor types in distributed sensor networks

[0081] Serial number type brand model applicability 1 strain sensors HottingerBaldwinMesstechnikGmbH C10 High precision, high stability, and excellent linearity enable accurate measurement of minute strain changes in structures. Accuracy can reach ±0.1% FS (full scale), meeting the precision requirements for structural stress and strain monitoring in equipment such as large cranes. 2 Displacement Sensor Micro-Epsilon ILD1402 Using the principle of laser triangulation, high-precision displacement measurement can be achieved, with a resolution of up to sub-micron level. The measurement range can be selected from a few millimeters to several meters depending on the model. It is very suitable for measuring the displacement changes of crane hoisting mechanisms, booms and other parts. 3 Vibration Sensor Bruel&Kjaer 4507 With a wide frequency response range (from a few Hz to tens of kHz), it can accurately measure vibration during crane operation, playing a vital role in detecting abnormal vibration and providing fault warnings. Its measurement accuracy remains high across different frequency bands, effectively capturing subtle vibration changes.

[0082] Sensors are deployed using a distributed network structure, automatically adjusting their monitoring frequency and scope based on the complexity and importance of the physical entity. For example, if an abnormal change (such as stress concentration) is detected in a critical part of the equipment, the sensor in that area will automatically increase its monitoring frequency to obtain more detailed data. For relatively stable components, the monitoring frequency can be appropriately reduced to save energy and data transmission resources.

[0083] In other embodiments, the communication links between sensors implement self-organizing communication to form an intelligent network. When some sensors fail or data transmission is blocked, adjacent sensors can adjust their monitoring strategies through the self-organizing communication mechanism to ensure data continuity and integrity.

[0084] In this embodiment, in the sensor network, the parameters are as follows:

[0085] (1) Sensor nodes:

[0086] Core chip:

[0087] Microprocessor: ARM Cortex-M7 (300MHz main frequency, supporting floating-point operations, meeting local filtering and decision-making calculations);

[0088] Communication module: Semtech SX1276 LoRa chip (frequency band 868 MHz, transmission power 14 dBm, communication distance 1 km).

[0089] Energy Management:

[0090] Power supply: 3.7V lithium thionyl chloride battery (capacity 19Ah, low self-discharge rate);

[0091] Dynamic power consumption adjustment: sleep mode current ≤ 1μA, active mode current ≤ 20mA.

[0092] (2) Edge computing unit:

[0093] Hardware configuration:

[0094] Processor: NVIDIA Jetson Nano (quad-core ARM A57, 128-core GPU, supports TensorFlow Lite);

[0095] Storage: 64GB eMMC (read and write speed 100MB / s).

[0096] (3) Communication network:

[0097] 5G module: Quectel RM500Q-GL (supports Sub-6 GHz, peak rate 2.5 Gbps);

[0098] Protocol configuration: CoAP over UDP (small header overhead, suitable for low-bandwidth transmission).

[0099] Example of implementing a sensor network that self-organizes communication and dynamically adjusts monitoring strategies (taking a large crane scenario as an example):

[0100] Failure scenario: A strain sensor on a crane boom fails due to mechanical shock.

[0101] Dynamic adjustment process:

[0102] 1. Neighboring sensors detect failed nodes through the ant colony routing protocol and replan the data transmission path;

[0103] 2. Displacement sensors near the failed node trigger a reinforcement learning model, increasing the monitoring frequency from 5 Hz to 10 Hz to compensate for the missing strain data.

[0104] 3. The data preprocessing module recovers the historical data of the failed node from other nodes through redundant coding;

[0105] 4. The adaptive algorithm module infers the local stress changes of the boom based on the fused displacement and vibration data and updates the digital twin model.

[0106] The above technologies can achieve the following technical effects:

[0107] Data continuity: The data loss rate of faulty nodes is reduced to less than 5%;

[0108] Real-time: Network self-recovery time is controlled within 200ms;

[0109] Energy efficiency: Dynamic energy management extends the life of sensor networks by 30%.

[0110] Through distributed intelligence and adaptive mechanisms, the sensor network achieves high robustness, low latency and high energy efficiency, providing reliable guarantees for real-time updates of digital twin models.

[0111] S2: Preprocessing of collected data, including data cleaning, outlier detection and noise reduction;

[0112] In step S2, the statistical analysis method is used to identify and remove abnormal data points. The 3σ principle is used. If the data x meets , If is the mean and σ is the standard deviation, the data is considered to be an outlier and will be excluded;

[0113] Moving average filtering is used to reduce the noise of the data. The formula is:

[0114]

[0115] in is the filtered value of the nth data point, are the original data points and N is the size of the moving average window.

[0116] S3: Update the digital twin model of the device based on the processed data. When the model is updated:

[0117] The update algorithm is selected according to the data type. For strain data of structural deformation, an update algorithm based on finite element analysis is adopted. For equipment wear data, an algorithm based on historical wear data and machine learning is used to predict wear trends and update the wear status of components in the model.

[0118] According to the changing trend of data, the allocation of computing resources is dynamically adjusted, and the model update algorithm is selected according to the allocation ratio of computing resources. When the data changes relatively smoothly, fewer computing resources are allocated; when sudden changes or anomalies occur, more computing resources are automatically called for accurate calculations and model updates. Specifically:

[0119] When a set of data is used to update the device digital twin model, the computing resource allocation ratio R of the set of data is:

[0120]

[0121] Where a and b are coefficients determined based on system performance and equipment characteristics;

[0122] Entropy H can be calculated using the Shannon entropy formula:

[0123]

[0124] is the probability distribution of the data.

[0125] A high rate of data change indicates significant physical changes in the corresponding device components, requiring more computing resources to process the data and update the model. When the data change rate reaches a set threshold, a high-precision finite element model is used for update calculations. During normal operation, the data change rate is low, requiring fewer resources to process the data, and a simplified model update algorithm is used to update the model.

[0126] The simplified model updating algorithm is: a fast response method based on dynamic order reduction.

[0127] The Dynamic Reduced-Order Model (DROM) is a simplified algorithm designed specifically for low-resource scenarios. Through data-driven selection of primary modes and construction of a reduced-order space, it significantly reduces computational complexity while ensuring usable accuracy for model updates. Its core innovation lies in its dynamic adjustment mechanism and resource-aware capabilities, filling a gap in simplified update algorithms for digital twin models.

[0128] When the computing resource allocation ratio R is low, the system uses the Dynamic Reduced Order Model (DROM) as a simplified algorithm to achieve efficient and real-time digital twin model updates. The following are the specific implementation steps and technical details:

[0129] While maintaining a certain level of accuracy, this approach significantly reduces computational complexity and is suitable for low-resource scenarios (e.g., R < 0.3). Driven by real-time data, it dynamically selects key modes (e.g., principal strains and principal displacement directions), constructs a reduced-order space, and updates the model only within this subspace.

[0130] The specific implementation steps are:

[0131] 1. Data-driven modality selection

[0132] Input: Current sensor data (strain, displacement, etc.) ---- Principal component analysis (PCA) results of historical data ---- Standardize the real-time data to eliminate the dimension effect ---- Calculate the projection coefficient of the real-time data and the historical principal component:

[0133]

[0134] U: principal component matrix of historical PCA;

[0135] : The mean of historical data.

[0136] The first k modes with the largest absolute values of projection coefficients are selected (for example, k = 3, experiments show that the first three principal components can explain more than 90% of the data variance, taking into account both efficiency and accuracy) to form the reduced-order subspace.

[0137] 2. Reduced-order model construction

[0138] In the reduced-order subspace, the original finite element model is simplified to:

[0139]

[0140] : Reduced stiffness matrix (U k are the first k principal components);

[0141] : Reduced-order load vector.

[0142] 3. Incremental Updates and Refactoring

[0143] Directly solve the reduced-order equation to obtain the reduced-order displacement increment , reconstruct the global displacement field:

[0144]

[0145] Update model: Superimpose it on the current displacement field and update the stress and strain synchronously.

[0146] 4. Dynamic mode adjustment.

[0147] Adaptive mechanism: If the residual error of the reduced-order model exceeds a threshold (e.g. , determined through A / B testing, a value higher than this will result in a model error > 5%, requiring modal expansion), automatically expanding the modal number k (e.g. );

[0148] When resources are sufficient, the principal component matrix U is periodically recalculated. The principal component update cycle is 24 hours, which is based on the device operation cycle to avoid frequent principal component calculations that consume resources.

[0149] An incremental learning algorithm is used to update the parts of the model affected by changes in physical entities. This only updates the parts of the model affected by physical changes, avoiding large-scale recalculation of the entire model and greatly improving update efficiency.

[0150] The incremental update of the finite element model is based on the variational principle and the virtual work principle; assuming that the structural deformation , which satisfies the following variational equation:

[0151]

[0152] in: is the domain of the structure, σ is the stress tensor, is the strain tensor, It's physical strength. is the surface force on the boundary, is the virtual displacement, Γ is the boundary of the domain Ω; by solving this equation, the deformation state of the structure is updated, and then the structural information in the digital twin model is updated.

[0153] The above method is further explained below using the digital twin application of a large crane as an example.

[0154] 1. Deployment of distributed intelligent sensor networks

[0155] For digital twin applications of large cranes, smart sensors are deployed in key locations such as the boom, tower, hoisting mechanism, slewing mechanism, and traveling mechanism. For example, strain sensors are installed at different sections and key connection points of the boom, and displacement and vibration sensors are installed on the ropes and transmission components of the hoisting mechanism.

[0156] Each intelligent sensor is equipped with a microprocessor and a communication module. The microprocessor is responsible for preliminary processing of the collected data, and the communication module enables self-organizing communication between sensors. The sensor determines the initial monitoring frequency based on its location and monitoring data through a built-in decision-making algorithm. For example, in the initial state, for the root of the boom, which is subject to greater stress, the strain sensor's monitoring frequency is set to 10 times per second, while for the walking mechanism, the displacement sensor's monitoring frequency is set to 5 times per second. Specifically, assuming that the strain sensor's adjustment coefficient, the sensitivity index to changes in monitoring data calculated based on the initial strain data variance, is 0.2, then its adjusted monitoring frequency is:

[0157]

[0158] (2) Data transmission and preprocessing

[0159] The data collected by the sensors is transmitted to the data center via low-latency wireless communication protocols (such as 5G and LoRa). During the transmission process, the data is encrypted and integrity verified to ensure data security and accuracy.

[0160] In the data center, sliding window technology is used to clean data in real time, removing noise and abnormal data caused by environmental interference or sensor anomalies. For abnormal data, comparison with adjacent data and analysis of historical data are used to determine whether it is a true anomaly. For example, if the strain data at a certain moment differs from the data at the previous and subsequent moments by more than a certain threshold and does not conform to the historical data trend, it is determined to be abnormal data and excluded. When using a moving average filter, if the window size N=5, the strain data of the nth data point is filtered to obtain:

[0161]

[0162] (3) Adaptive model update algorithm

[0163] The system automatically determines the update algorithm based on the incoming data type. When receiving strain data on structural deformation, an update algorithm based on finite element analysis is used to update the structural mechanical properties in the digital twin model based on the strain data. For equipment wear data, an algorithm based on historical wear data and machine learning is used to predict wear trends and update the component wear status in the model.

[0164] The system dynamically allocates computing resources based on the magnitude and rate of data change. For example, when the rate of change of strain data exceeds a set threshold, more computing resources are called upon to enable high-precision finite element model update calculations; in normal operation, a simplified model update algorithm is used. Assuming that the probability distribution of strain data is known, its entropy is calculated.

[0165]

[0166] According to the coefficient , calculate the computing resource allocation ratio:

[0167]

[0168] For model updating, an incremental learning algorithm is used. Assuming that when a local deformation occurs in the crane boom, only the finite element mesh and structural mechanics parameters related to the local deformation are updated instead of recalculating the finite element model of the entire boom. According to the variational equation

[0169]

[0170] The updated deformation information is substituted into the equation and solved by the finite element solver to obtain the updated structural deformation field and stress field, and then the corresponding parts in the digital twin model are updated.

[0171] (IV) Digital Twin Model Storage and Display

[0172] The updated digital twin model is stored in a distributed storage system, using a distributed file system or distributed database to ensure data redundancy and scalability.

[0173] The digital twin model is displayed to users through a 3D visualization engine. Users can view the status of the crane under different working conditions through the interactive interface, such as observing the deformation of the boom under wind load, and rotate and zoom the model with the mouse to view detailed information and parameters of different components.

[0174] Another specific embodiment of the present application provides a device digital twin model update system, including:

[0175] A data acquisition unit, comprising a plurality of data acquisition devices distributedly deployed at a plurality of locations on the equipment to collect data from the equipment; the data acquisition device comprises a sensor module, a data processing module, and a communication module; the sensor module comprises one or more of a strain sensor, a displacement sensor, and a vibration sensor; and the data processing module is used to pre-process the data collected by the sensor module;

[0176] The data center is connected to the data acquisition device and is used to receive data from the data acquisition device. A low-latency, high-bandwidth data transmission protocol is used to ensure that the data can reach the data center quickly and accurately.

[0177] The data center updates the digital twin model of the equipment based on the received data. When updating the model: an update algorithm is selected based on the data type. For strain data of structural deformation, an update algorithm based on finite element analysis is adopted. For equipment wear data, an algorithm based on historical wear data and machine learning is used to predict wear trends and update the wear status of components in the model. The allocation of computing resources is dynamically adjusted based on the changing trends of the data. An incremental learning algorithm is used to update the parts of the model affected by changes in physical entities.

[0178] and a storage and display unit for storing and displaying digital twin models.

[0179] Another specific embodiment of the present application provides a digital twin model terminal, including a memory, a processor and a display, wherein the display is used to display the digital twin model, and the memory stores a computer program. When the computer program is called and executed by the processor, the device digital twin model update method as described above is implemented.

[0180] Another specific embodiment of the present application provides a computer-readable medium, characterized in that the computer-readable medium memory stores a computer program, and when the computer program is called and executed by a computer, it implements the device digital twin model updating method as described above.

[0181] The above is only a preferred embodiment of the present application. It should be pointed out that for ordinary technicians in this field, several variations and improvements can be made without departing from the creative concept of the present application, and these all fall within the scope of protection of the present application.

Claims

1. A method for updating a device digital twin model, characterized in that: The following steps are involved: S1: Collecting data from several locations of the equipment; in step S1, the collected equipment data includes wear data, strain data, displacement data, and vibration data from different parts of the equipment; For any data collection point on the device, the monitoring frequency is dynamically adjusted according to the sensitivity index of the monitoring data change; A distributed sensor network is used to collect data from several locations of the device. The sensor collection frequency is obtained by the following formula: ； in, is the initial monitoring frequency of the sensor, k is the adjustment coefficient, and S is the sensitivity index of the change of the monitoring data; S2: Preprocessing of collected data, including data cleaning, outlier detection and noise reduction; S3: Update the digital twin model of the device based on the processed data. When the model is updated: Select an update algorithm based on the data type. For structural deformation strain data, an update algorithm based on finite element analysis is used. For equipment wear data, an algorithm based on historical wear data and machine learning is used to predict wear trends and update the component wear status in the model. Dynamically adjust the allocation of computing resources based on data trends, and select a model update algorithm based on the allocation ratio of computing resources; An incremental learning algorithm is used to update the parts of the model that are affected by changes in physical entities.

2. The device digital twin model updating method according to claim 1, characterized in that: In step S1, The change sensitivity index S of the monitoring data is obtained by any of the following methods: (1) For structural strain data monitoring: ； : The variance of the data in the current time window; Time window length T: set according to the response time of the physical entity; : Historical maximum variance, historical variance update cycle: daily or weekly update; (2) For displacement data and vibration data: ； : Time gradient of data; gradient calculation cycle : consistent with the sensor sampling frequency; : historical maximum gradient; When the time gradient of the data exceeds the preset threshold multiple times continuously, it is judged as abnormal and an abnormal warning is issued; (3) For complex working conditions of equipment: ； : Weight coefficient; through historical data regression analysis, select the model that minimizes the prediction error , value.

3. The device digital twin model updating method according to claim 1, characterized in that: In step S2, the statistical analysis method is used to identify and remove abnormal data points. The 3σ principle is used. If the data x satisfy , If is the mean and σ is the standard deviation, the data is considered to be an outlier and will be excluded; Moving average filtering is used to reduce the noise of the data. The formula is: ； in is the filtered value of the nth data point, are the original data points and N is the size of the moving average window.

4. The device digital twin model updating method according to claim 1, characterized in that: In step S3, the allocation of computing resources is dynamically adjusted according to the changing trend of the data as follows: When a set of data is used to update the device digital twin model, the computing resource allocation ratio R of the set of data is: ； Where a and b are coefficients determined based on system performance and equipment characteristics; Entropy H can be calculated using the Shannon entropy formula: ； is the probability distribution of the data.

5. The device digital twin model updating method according to claim 1, characterized in that: In step S3, an incremental learning algorithm is used to update the part of the model affected by the change of the physical entity; The incremental update of the finite element model is based on the variational principle and the virtual work principle; assuming that the structural deformation , which satisfies the following variational equation: ； in: is the domain of the structure, σ is the stress tensor, is the strain tensor, It's physical strength. is the surface force on the boundary, is the virtual displacement, Γ is the boundary of the domain Ω; by solving this equation, the deformation state of the structure is updated, and then the structural information in the digital twin model is updated.

6. A device digital twin model update system, characterized in that: include: A data acquisition unit, comprising a plurality of data acquisition devices, which are distributed and deployed at a plurality of locations of the device to collect data of the device; The data acquisition device includes a sensor module, a data processing module, and a communication module. The sensor module includes one or more of a strain sensor, a displacement sensor, and a vibration sensor. The data processing module is used to pre-process the data collected by the sensor module. The acquisition frequency of the sensor is obtained by the following formula: ； in, is the initial monitoring frequency of the sensor, k is the adjustment coefficient, and S is the sensitivity index of the change of the monitoring data; A data center is communicatively connected to a data acquisition device and configured to receive data from the data acquisition device. The data center updates the digital twin model of the equipment based on the received data. When updating the model, an update algorithm is selected based on the data type. For strain data of structural deformation, an update algorithm based on finite element analysis is used. For equipment wear data, an algorithm based on historical wear data and machine learning is used to predict wear trends and update the wear status of components in the model. The allocation of computing resources is dynamically adjusted based on the changing trends of the data. An incremental learning algorithm is used to update the parts of the model affected by changes in physical entities. and a storage and display unit for storing and displaying digital twin models.

7. A digital twin model terminal, characterized in that: It includes a memory, a processor and a display, the display is used to display the digital twin model, the memory stores a computer program, and when the computer program is called and executed by the processor, it implements the device digital twin model updating method as described in any one of claims 1 to 5.

8. A computer-readable medium, characterized in that The computer-readable medium memory stores a computer program, and when the computer program is called and executed by a computer, it implements the device digital twin model updating method according to any one of claims 1 to 5.

Images