A cloud-edge collaborative-based intelligent elevator edge digital twin construction method and system

Abstract

This invention provides a method and system for constructing a smart elevator edge digital twin based on cloud-edge collaboration. The method includes: acquiring environmental information of the vertical elevator's operating scenario; generating a meta-model file based on the environmental information parameters, the meta-model file including an elevator shaft model, an elevator subsystem model, interface files, and a canvas file; acquiring current subsystem data of the vertical elevator; uploading and writing data to an edge module near the subsystem based on dynamic parameter data and static dimension data collected from control cabinets near the subsystem, thus obtaining an edge data center for the elevator subsystem; matching the model-driven data and state decision data processed by the edge computing module with the corresponding model component names under the general file package to form an edge digital twin model, and sending the processed data form to a cloud server; pulling the digital twin data and general file package of the vertical elevator from the cloud server to the digital twin front end, and displaying the elevator's operating status and early warning results in the digital twin model; realizing the coordinated implementation of vertical elevator equipment under cloud-edge collaborative computing and digital twin technology, improving the effective information extraction capability and real-time monitoring level of the monitored equipment, and better preventing potential safety threats to elevator equipment.

Description

Technical Field

[0001] This invention relates to the field of intelligent manufacturing technology, specifically to a method and system for constructing a smart elevator edge digital twin based on cloud-edge collaboration. Background Technology

[0002] As complex mechanical equipment, vertical elevators present challenges such as large data volumes and difficulties in centrally processing and displaying multi-dimensional information. Dividing the elevator system and configuring multiple edge computing modules addresses these issues to meet the monitoring needs for safe operation under different working scenarios. 3D visualization technology can realistically display elevator equipment, enabling operation management and monitoring of each layer of the elevator structure based on a 3D environment. Traditional elevator operation platforms mostly upload and write the characteristic parameters of the entire equipment, requiring loading and compiling all collected feature data. This increases the communication load of the operation platform and degrades its overall functionality. Furthermore, current elevator maintenance largely relies on manual, periodic maintenance, which may result in equipment operating with defects. Applying digital twin technology and cloud-edge collaboration technology enables collaborative information management between devices, achieving real-time monitoring and optimization of elevator operation.

[0003] Smart elevator edge digital twin refers to the process of creating and deploying models in elevator systems using digital twin technology, enabling elevators to operate and be maintained more intelligently. Digital twins are the digital modeling and simulation of real-world entities, processes, or systems, allowing for a better understanding, prediction, and optimization of their behavior. Building a smart elevator edge digital twin moves edge computing tasks closer to the data source, enabling faster data processing and lower latency. This meets the management and information integration needs of smart elevator equipment, providing a clear and intuitive understanding of the effective information in the operation of the modular system, and achieving visualized and real-time management.

[0004] Cloud-edge collaboration is a complementary and synergistic approach between cloud computing and edge computing. The emergence of edge computing models offers a new solution to the shortcomings of centralized cloud computing models and is a product of technological development. Cloud-edge collaboration tightly integrates cloud services with edge computing, achieving the decentralization of cloud computing by rationally allocating tasks between cloud services and edge computing, extending cloud services and cloud analytics to the edge. Compared to the traditional method of directly migrating platforms to the cloud, the cloud-edge collaboration method preprocesses the complex multi-information data of elevator equipment, compressing the equipment parameter data collected from the equipment control cabinet and reducing data redundancy. Summary of the Invention

[0005] This invention addresses the problem that, when building an elevator management platform, the usual practice is to upload all elevator equipment operation data and directly process the collected data in the cloud, leading to long project implementation cycles and increased cloud server load. Furthermore, different application scenarios for vertical elevators result in varying requirements for elevator system monitoring and management. To meet the needs of smart elevator construction methods that adapt to changing application scenarios, this invention proposes a method and system for constructing a smart elevator edge digital twin based on cloud-edge collaboration. This method allows for the configuration of subsystem edge modules to customize health management functions and addresses the real-time efficiency issue of a single platform system handling multi-party data processing.

[0006] To achieve the above objectives, a first aspect of the present invention provides a method for constructing a smart elevator edge digital twin based on cloud-edge collaboration, comprising:

[0007] S1. Obtain environmental information of the vertical elevator operation scenario, generate a meta-model file based on the environmental information parameters, the meta-model file includes an elevator shaft model, an elevator subsystem model, an interface file, a canvas file, and an operation map model, and store the meta-model file as a general file package on a cloud server;

[0008] S2. Obtain the current subsystem data of the vertical elevator. Based on the dynamic parameter data and static dimension data summarized by the control cabinet near the subsystem, upload the data and write it to the edge module near the subsystem to obtain the edge data center of the elevator subsystem. Match the model-driven data and state decision data processed by the edge module with the corresponding model component names under the general file package to form an edge digital twin model, and send the data to the cloud server.

[0009] S3. Pull the digital twin data and general file package of the vertical elevator from the cloud server to the digital twin front end, and display the elevator's operating status and early warning results in the digital twin model.

[0010] As can be seen from the above technical solution, the method of the present invention is mainly divided into three stages: building and storing the model using modeling equipment and cloud server; collecting information and assessing the safety status of the elevator group using edge module, and uploading the calculated data information to the cloud server for front-end access; and displaying the elevator operation status using digital twin front-end, thereby constructing a universal smart elevator management solution. Access to the cloud server is granted through a dedicated network, and the elevator operation position and calculation results are displayed in the digital twin model, realizing the universalization of the smart elevator model and the vivid display of edge computing results.

[0011] Furthermore, step S1, which involves obtaining environmental information about the vertical elevator's operating scenario and generating a meta-model file based on the environmental information parameters, includes:

[0012] S11. Use modeling equipment to model the simulation environment of the vertical elevator operation scene. The environment information includes model components: panoramic surround camera, lighting, elevator shaft model, car system model, door operator system model, traction machine system model, and canvas interface.

[0013] S12. The model components are programmed using C# language to obtain the trajectory scripts of each component; then the trajectory scripts are bound to the model components to complete the start and stop behavior of the elevator model between the specified floors.

[0014] S13. Generate a meta-model file based on the environmental information. The meta-model file includes an elevator shaft model, an elevator subsystem model, an interface file, and a canvas file.

[0015] Furthermore, step S13, based on the interface files included in the meta-model file, includes:

[0016] S131. Use the TCP / IP protocol to configure the environment for the metamodel-driven data source, match attribute names, find the corresponding component name in the file package, and add Update and Queue functions to build the script file for the corresponding model;

[0017] S132. Generate an interface file for the elevator model based on the created script file and data source address, which will be used in the twin service of the smart elevator.

[0018] Furthermore, step S2, which involves acquiring the current subsystem data of the vertical elevator, and uploading and writing the data to the edge module near the subsystem based on the dynamic parameter data and static dimension data collected from the control cabinet near the subsystem, to obtain the edge data center of the elevator subsystem, includes:

[0019] S21. Collect static and dynamic data of the elevator on site to the edge module using the bus protocol. Preprocess the collected data using the two-dimensional CNN-GRU state decision algorithm and TL algorithm deployed under the calculation module to accurately calculate the current elevator operation status result.

[0020] S22. Calculate the elevator's state information and static dimension information based on the edge module of the elevator subsystem to generate an elevator twin data form;

[0021] S23. Store the data form in the edge data center to store the visualized status information of the smart elevator.

[0022] Furthermore, the two-dimensional CNN-GRU algorithm and transfer learning method deployed in the computing module described in step S2-1 perform state discrimination on the preprocessed time series data, including:

[0023] S211. Perform continuous wavelet transform on the time series data stored in the computing module to generate a time-frequency diagram corresponding to the time series data;

[0024] S212. Define a feature extraction layer model based on the time-frequency information and image size specifications of the generated image to form a two-dimensional CNN-GRU state prediction model, train the state prediction model to obtain the corresponding feature learning parameters, and save it as a .h5 file in the current directory.

[0025] S213. By using feature transfer learning methods to freeze the low-level parameters of the feature extraction layer, the feature training time is reduced and the generalization ability of the state prediction model to various data collected from the edge module is enhanced.

[0026] Furthermore, step S2, which involves matching the model-driven data and state decision data obtained from the edge module with the corresponding model component names under the general file package to complete the mapping from entity to virtual and form an edge digital twin model, includes:

[0027] S23. Upload the form calculated by the edge module to the cloud server data storage center. According to the interface file under the general file package, match the form name with the model component name, call the update function and queue function to realize the elevator operation mapping process, and generate the edge digital twin model.

[0028] S24. The mapping process is implemented through the Update and Queue functions. The cloud server needs to receive the preprocessed result data uploaded by the edge device module and complete the interaction between the relational database and the front-end model based on the TCP / IP protocol and C# programming technology. The scripts containing the queue function and update function are mounted on the elevator model component to ensure the real-time synchronization of the elevator virtual-real mapping.

[0029] Furthermore, the generation of the edge digital twin model from the subsystem operating data, edge computing module, and meta-model file described in steps S2-3 includes:

[0030] S231. Let the aforementioned edge digital twin model be described as follows:

[0031] (1)

[0032] In the formula, The elevator's parameter data consists of static data on the geometric characteristics of each elevator subsystem and dynamic data on equipment operation.

[0033] (2)

[0034] As an interaction protocol, it enables data interconnection, wherein the aforementioned For subsystem serial number, for

[0035] Components in the current subsystem The interface file name for its component:

[0036] (3)

[0037] The equipment model file mainly includes simulation models of the main operating equipment of the elevator:

[0038] (4)

[0039] Furthermore, the update function and queue function required for the model mapping process described in steps S2-4 include:

[0040] S241. Assume the aforementioned model The position of direction The position of direction The positions of the directions are respectively

[0041] (5)

[0042] To ensure the data type remains unchanged, assign the data retrieved by the queue function to the model component.

[0043] but: (6)

[0044] S242. According to the frequency of model mapping, the data update time interval is... , Each time the call is executed, the time interval is incremented, and the return value is reset to zero. Greater than If so, the update function will be called once.

[0045] In addition, the present invention also provides a construction system for a smart elevator edge digital twin based on cloud-edge collaboration, including modeling equipment, edge devices, cloud servers, an elevator group consisting of multiple vertical elevators, and digital twin front-end software.

[0046] The modeling device includes a data acquisition and drawing module, used to acquire environmental information of the vertical elevator operation scenario and generate a meta-model file based on the environmental information parameters. The meta-model file includes an elevator shaft model, an elevator subsystem model, a canvas model, and interface files.

[0047] The edge module device includes a data storage center and an algorithm deployment module; wherein, the data storage center is used to acquire the current subsystem data of the vertical elevator, including dynamic parameter data and static dimension data summarized by the control cabinet near the subsystem and equipment safety status data processed by the algorithm; the algorithm deployment module is used to calculate the equipment safety status information from the dynamic data;

[0048] The cloud server includes a data storage module and a meta-model module; wherein, the data storage module is used to receive data information uploaded from the edge module; the meta-model module is used to construct a general file package based on the system model files generated by the drawing module.

[0049] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0050] This invention constructs a universal operation and management solution for elevator groups, allowing for model expansion and dynamic updates. By configuring edge modules near the monitored equipment, it avoids the traditional method of directly uploading elevator system data to the cloud. Instead, it utilizes computational methods deployed under the edge modules to preprocess the collected data, reducing cloud server computation time and ensuring the timeliness of the smart elevator's virtual-real interaction. This achieves a real-time mapping process between the real and virtual worlds, intuitively and vividly displaying the operational status on the digital twin front end. It also facilitates clear and observable observation and management by relevant personnel, enabling timely and appropriate rescue measures. Attached Figure Description

[0051] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below.

[0052] Figure 1 This is a flowchart illustrating a cloud-edge collaborative edge digital twin method for smart elevators.

[0053] Figure 2 This is a schematic diagram of the virtual-real interaction process in the method of the present invention;

[0054] Figure 3 This is a flowchart of the intelligent elevator operation solution based on cloud-edge collaboration technology of the present invention. Detailed Implementation

[0055] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments.

[0056] The invention will now be described in further detail with reference to the accompanying drawings.

[0057] Figure 1This is a flowchart of the method for constructing a smart elevator edge digital twin according to the present invention. This embodiment of the method can monitor the elevator operation in real time through cloud and edge devices and digital twin technology. It is mainly completed by using modeling equipment and cloud server for model creation; using edge module for data collection and status calculation; and using digital twin front end for smart elevator operation display. The details are as follows.

[0058] First, before the vertical elevator is in normal operation, a model is created using modeling equipment and a cloud server to obtain environmental information of the vertical elevator operation scenario. The environmental information parameters are then used to generate meta-model files, which are named. The meta-model files include elevator shaft models, elevator subsystem models, interface files, canvas files, and operation map models. The named meta-model files are then stored as a general file package on the cloud server.

[0059] Modeling equipment is used to simulate the operation scene of a vertical elevator. The model components obtained from the environmental information include a panoramic surround camera, lighting, elevator shaft model, car system model, door operator system model, traction machine system model, and canvas interface. These are the main model files. Related model files can be added or deleted as needed. The model component configuration script files are programmed using C# language to obtain the trajectory scripts of each component. The trajectory scripts are then bound to the model components to complete the start and stop behavior of the elevator model between specified floors.

[0060] The meta-model files generated based on the environmental information include elevator shaft models, elevator subsystem models, interface files, and canvas files. Since the interface files require data source configuration, the meta-model drives the data source through the TCP / IP protocol to perform environment configuration and attribute name matching, find the corresponding component names in the file package, and add Update and Queue functions to build the script files of the corresponding models.

[0061] The created script file and data source address are used to generate the interface file for the elevator model, which is then used in the twin service of the smart elevator.

[0062] Secondly, edge modules are used for data acquisition and status calculation to obtain the current subsystem data of the vertical elevator. Based on the dynamic parameter data and static dimensional data aggregated from the control cabinets near the subsystem, the data is uploaded and written to the edge modules near the subsystem, thus obtaining the edge data center of the elevator subsystem. The edge modules are deployed locally in the industrial field, supporting CRUD operations on the database. They are also equipped with computing modules to preprocess the acquired data. The edge modules connect to various IoT devices in the industrial field to acquire data, and connect to the cloud system via protocol interaction to upload the preprocessed data. They can also receive control commands from the cloud system.

[0063] The static and dynamic data of the elevator on site are collected to the edge module using a bus protocol. The collected data is preprocessed by the two-dimensional CNN-GRU state decision algorithm and TL algorithm deployed in the computing module to accurately calculate the current elevator operation status result.

[0064] The edge module of the subsystem calculates the elevator's status information and static dimension information to generate an elevator twin data form, which is then stored in the edge data center to store the visualized status information of the smart elevator.

[0065] The edge module's functionality is achieved through data preprocessing and algorithm model deployment. The time-series data stored in the computing module undergoes continuous wavelet transform to generate the corresponding time-frequency graph.

[0066] A two-dimensional CNN-GRU state prediction model is formed by defining a feature extraction layer model using the time-frequency information and image size specifications of the generated image. The state prediction model is then trained to obtain the corresponding feature learning parameters, which are saved as a .h5 file in the current directory.

[0067] By using feature transfer learning to freeze the low-level parameters of the feature extraction layer, the feature training time can be reduced, thereby enhancing the generalization ability of the state prediction model to various data collected from the edge module.

[0068] Then, the model-driven data and state decision data obtained from the edge module are matched with the corresponding model component names under the general file package to form an edge digital twin model.

[0069] The form calculated by the edge module is uploaded to the cloud server data storage center. According to the interface file under the general file package, the form name is matched with the model component name. The update function and queue function are called to realize the mapping process of elevator operation and generate the edge digital twin model.

[0070] The edge digital twin model includes subsystem operational data, edge computing modules, and meta-model files, as defined below:

[0071] Let the edge digital twin model be described as follows:

[0072] (1)

[0073] In the formula, The elevator's parameter data consists of static data on the geometric characteristics of each elevator subsystem and dynamic data on equipment operation.

[0074] (2)

[0075] As an interaction protocol, it enables data interconnection, wherein the aforementioned Subsystem number for

[0076] Components in the current subsystem The interface file name for its component:

[0077] (3)

[0078] The equipment model file mainly includes simulation models of the main operating equipment of the elevator:

[0079] (4)

[0080] The mapping process based on the virtual-real model is implemented through the Update and Queue functions. The cloud server needs to receive the preprocessed result data uploaded by the edge device module and complete the interaction between the relational database and the front-end model based on the TCP / IP protocol and C# programming technology. The scripts containing the queue function and update function are mounted on the elevator model component to ensure the real-time synchronization of the elevator virtual-real mapping.

[0081] The mapping process is implemented using the Update and Queue functions. First, it must be ensured that... The file code and The file codes correspond, and secondly, it ensures that the update function loops once every time the data source is called in the queue under the script file. The specific definition is as follows:

[0082] Let the model described above be... The position of direction The position of direction The positions of the directions are respectively

[0083] (5)

[0084] To ensure the data type remains unchanged, assign the data retrieved by the queue function to the model component.

[0085] but: (6)

[0086] Based on the frequency of model mapping, the data update interval is: , Each time the call is executed, the time interval is incremented, and the return value is reset to zero. Greater than If so, the update function will be called once.

[0087] Finally, the operation of the smart elevator is demonstrated using a digital twin front-end, such as... Figure 3As shown, the digital twin front end is the process from acquiring underlying data to pulling data from the cloud server for display.

[0088] The edge module acquires data from external sensors of various elevator subsystems. It aggregates the static and dynamic data integrated in the elevator control cabinet into the edge module data center using RS-485, Modbus, and serial port protocols. The data is then preprocessed using the edge module's feature extraction method, and the results are uploaded to the cloud database module.

[0089] A relational database is deployed on the cloud server to maintain data communication with each edge device. It is used to receive the data preprocessing results performed by the edge modules, which serve as a collection of twin data, geometric dimensions, and status results of the elevator equipment. Cloud computing resources are used to make decisions based on the data, ensuring the quality and availability of decision data. The relational database is connected to the client Unity3D application through a pre-configured port IP number, which facilitates the local digital twin front-end software to access the data status information corresponding to its components.

[0090] The twin layer is the application end of the smart elevator twin model. Figure 2 To illustrate the schematic diagram, the process mainly includes backend data reception and frontend model display. A request to retrieve data from the corresponding database list is sent to the cloud's feature data module. The client, established through a backend IP address, connects to the server. The cloud returns the required data and writes it to the platform system components. Data transformation is performed using queue and update functions in a C# script file. The Transform component information of the twin is traversed, and the canvas and component parameters under the model component unit are updated in a timely manner, thus mapping the real-time operation of the on-site vertical elevator system.

[0091] According to the above embodiments, this embodiment provides a construction system for a smart elevator edge digital twin based on cloud-edge collaboration, including modeling equipment, edge devices, cloud servers, an elevator group composed of multiple vertical elevators, and digital twin front-end software;

[0092] The modeling device includes a data acquisition and drawing module, used to acquire environmental information of the vertical elevator operation scenario, and generate a meta-model file based on the environmental information parameters. The meta-model file includes an elevator shaft model, an elevator subsystem model, a canvas model, and interface files.

[0093] The edge module device includes a data storage center and an algorithm deployment module; wherein, the data storage center is used to acquire the current subsystem data of the vertical elevator, including dynamic parameter data and static dimension data summarized by the control cabinet near the subsystem and equipment safety status data processed by the algorithm; the algorithm deployment module is used to calculate the equipment safety status information from the dynamic data;

[0094] The cloud server includes a data storage module and a meta-model module; wherein, the data storage module is used to receive data information uploaded from the edge module; the meta-model module is used to construct a general file package based on the system model files generated by the drawing module.

[0095] The system built above mainly divides its functions into elevator status monitoring and cloud-edge computing collaboration. This example is specifically implemented using software and hardware such as Unity3D, VS2019, cloud server, edge module, database and device sensors.

[0096] Elevator status monitoring visualizes on-site work scene data as a 3D scene model, and simultaneously displays the smart elevator operation parameter information corresponding to the 3D scene model;

[0097] Queue functions and update functions are used to update the information about the operation of the equipment on both sides of the interface in real time and record the status information of the elevator operation.

[0098] The model was converted to .FBX format using Blender model drawing software, imported into Assets in the Unity3D project management panel, the center coordinate position was adjusted, the elevator shaft model was rendered, code was written to capture the elevator car model, and the elevator car was set to light up the shaft corridor of the floor it was on when it arrived at each floor door, giving a dynamic simulation process of virtual and real mapping.

[0099] Utilizing layout features from Unity3D components, the two sides of the display interface are set as canvases, each with three canvases to display relevant information. Icons are controlled by a "Wave circle" function, and corresponding text is added. The layout includes bar charts, time-series charts, and memo charts, etc. Information is processed and stored in a cloud database. Within the Unity3D project files, `System.Data.SqlClient` is deployed as the protocol framework, using TCP / IP as the client to listen to the SQL relational database. Upon successful data listening, it accesses the database tables, and based on the data type information, uses a queue data structure and update functions to retrieve the list data from the database, displaying and updating the data in real-time on the corresponding canvases and components.

[0100] Cloud-edge computing collaboration involves placing edge devices near the data source of the elevator subsystem for data preprocessing, reducing the data load on the cloud, and enabling the fusion of multiple information data from the devices themselves for rapid decision-making.

[0101] The cloud server acts as the central hub, storing the geometric model data of the equipment in the industrial scenario. Edge devices near the subsystem collect data, and real-time operational data of the elevator traction drive system is collected based on the RS-485 communication protocol. Temperature control information of the elevator shaft and car is collected based on the Modbus protocol. Other elevator system parameter information is also collected using the above protocols. The data is preprocessed using a state discrimination algorithm, and the preprocessed data is stored and awaited uploading to the cloud for the interface file to call the cloud data source.

[0102] During the editing of the digital twin 3D model, corresponding data points are selected for the data interface of each digital twin unit model through data distribution. The data interface of each digital twin unit model is bound to the table name in the relational database, thereby transmitting the business data of the physical entity processed by the edge device components to the digital twin unit model. Driven by the data of the physical entity in the industrial field, the digital twin unit model accurately reflects the real-time operating status and fault information of the physical entity in the virtual environment.

[0103] On the interface published by the digital twin front end, input the running command to apply the generated edge digital twin smart elevator model in the maximized form of the current client. The monitoring screen is a monitor connected to the client. The client can view the running status of the digital twin model in real time, including the model's actions, real-time data, running status information, etc., and faithfully reflect the running situation of the actual scene. The client can view the real-time information and fault information of the corresponding physical entity through the supplementary information at the bottom of the screen.

[0104] This implementation case constructs elevator data interaction logic through a digital twin front-end and cloud-edge devices, realizing the configuration capability of edge digital twins for vertical elevators. Addressing the varying elevator safety management needs across different scenarios, the system divides into subsystems to configure edge modules locally, constructing an edge digital twin model. At the edge modules, a state discrimination method is used to process the operating parameter data of each subsystem. Finally, the results are written to the model component of the digital twin platform, realizing the mapping process of smart elevator operation. This reduces the development workload of elevator safety management status panels, alleviates the load pressure of cloud-based analysis and processing of complex elevator data, and better prevents potential safety threats throughout the elevator's entire lifecycle. It also provides a reference for the subsequent construction of customized multi-functional smart elevator modeling systems.

Claims

1. A method for constructing a smart elevator edge digital twin based on cloud-edge collaboration, characterized in that, include: S1. Obtain environmental information of the vertical elevator operation scenario, generate a meta-model file based on the environmental information parameters, the meta-model file includes an elevator shaft model, an elevator subsystem model, an interface file, and a canvas file, and store the meta-model file as a general file package on a cloud server. S2. Obtain the current subsystem data of the vertical elevator. Based on the dynamic parameter data and static dimension data summarized by the control cabinet near the subsystem, upload the data and write it to the edge module near the subsystem to obtain the edge data center of the elevator subsystem. Match the model-driven data and state decision data processed by the edge computing module with the corresponding model component names under the general file package to form an edge digital twin model. Send the processed data form to the cloud server. S3. Pull the digital twin data and general file package of the vertical elevator from the cloud server to the digital twin front end, and display the elevator's operating status and early warning results in the digital twin model.

2. The method for constructing a smart elevator edge digital twin based on cloud-edge collaboration according to claim 1, characterized in that, Step S1, which involves obtaining environmental information about the vertical elevator operation scenario and generating a meta-model file based on the environmental information parameters, includes: S1-1. Use modeling equipment to model the simulation environment of the vertical elevator operation scene. The environment information includes the following model components: panoramic surround camera, lighting, elevator shaft model, car system model, door operator system model, traction machine system model, and canvas interface. S1-2. Program the model components using C# language to obtain the trajectory scripts for each component; then bind the trajectory scripts to the model components to complete the start and stop behavior of the elevator model between the specified floors. S1-3. Generate a meta-model file based on the environmental information. The meta-model file includes an elevator shaft model, an elevator subsystem model, an interface file, and a canvas file.

3. The method for constructing a smart elevator edge digital twin based on cloud-edge collaboration according to claim 2, characterized in that, The interface files in the meta-model file mentioned in steps S1-3 include: S1-3-1. Configure the environment for the metamodel-driven data source using the TCP / IP protocol, match attribute names, find the corresponding component name in the file package, and add Update and Queue functions to build the script file for the corresponding model. S1-3-2. Generate the interface file for the elevator model based on the created script file and data source address, which is used in the twin service of the smart elevator.

4. The method for constructing a smart elevator edge digital twin based on cloud-edge collaboration according to claim 1, characterized in that, Step S2 involves acquiring the current subsystem data of the vertical elevator, uploading the data and writing it to the edge module near the subsystem based on the dynamic parameter data and static dimension data collected from the control cabinets near the subsystem, thus obtaining the edge data center of the elevator subsystem, including: S2-1. Collect and aggregate the static and dynamic data of the elevator on site to the edge module using the bus protocol. Preprocess the data according to the computing module under the edge module. Then, use the two-dimensional CNN-GRU algorithm and transfer learning method deployed under the computing module to perform state discrimination on the preprocessed time series data to obtain the current elevator operation status result. S2-2. Calculate the elevator's state information and static dimension information based on the edge module of the elevator subsystem to generate a data form for the elevator edge digital twin. S2-3. Store the data form in the edge data center to store the visualized status information of the smart elevator.

5. The method for constructing a smart elevator edge digital twin based on cloud-edge collaboration according to claim 4, characterized in that, The two-dimensional CNN-GRU algorithm and transfer learning method deployed in the computing module described in step S2-1 perform state discrimination on the preprocessed time series data, including: S2-1-1. Perform continuous wavelet transform on the time series data stored in the computing module to generate the time-frequency diagram corresponding to the time series data; S2-1-2. Define a feature extraction layer model based on the time-frequency information and image size specifications of the generated image to form a two-dimensional CNN-GRU state prediction model, train the state prediction model to obtain the corresponding feature learning parameters, and save it as a .h5 file in the current directory. S2-1-3. Use the feature transfer method of transfer learning to freeze the low-level parameters of the feature extraction layer, thereby reducing the feature training time and enhancing the generalization ability of the state prediction model to various data collected by the edge module.

6. The method for constructing a smart elevator edge digital twin based on cloud-edge collaboration according to claim 1, characterized in that, Step S2, which involves matching the model-driven data and state decision data obtained from the edge module with the corresponding model component names under the general file package to complete the mapping from entity to virtual and form an edge digital twin model, includes: S2-3. Upload the form calculated by the edge module to the cloud server data storage center. According to the interface file under the general file package, match the form name with the model component name, call the update function and queue function to realize the elevator operation mapping process, complete the mapping from entity to virtual, and the edge digital twin model is composed of subsystem operation data, edge computing module and meta-model file. S2-4. The mapping process is implemented through the Update and Queue functions. The cloud server needs to receive the preprocessed result data uploaded by the edge device module and complete the interaction between the relational database and the front-end model based on the TCP / IP protocol and C# programming technology. The script containing the update function and queue function is mounted on the elevator model component to ensure the real-time synchronization of the elevator virtual-real mapping.

7. The method for constructing a smart elevator edge digital twin based on cloud-edge collaboration according to claim 6, characterized in that, Step S2-3 involves using the subsystem operation data, edge computing module, and meta-model file to generate the edge digital twin model, including: S2-3-1. Let the edge digital twin model be described as follows: （1） In the formula, The elevator's parameter data consists of static data on the geometric characteristics of each elevator subsystem and dynamic data on equipment operation. （2） This is an interactive protocol that enables data interconnection, whereby... For subsystem serial number, for Components in the current subsystem The interface file name for its component: （3） The equipment model file mainly includes simulation models of the main operating equipment of the elevator: （4）。 8. The method for constructing a smart elevator edge digital twin based on cloud-edge collaboration according to claim 6, characterized in that, The update function and queue function required for the model mapping process described in steps S2-4 include: S2-4-1, Assume the aforementioned model The position of direction The position of direction The positions of the directions are respectively (5) To ensure the data type remains unchanged, the data retrieved by the queue function is assigned to the model component, then: (6) S2-4-2. According to the frequency of model mapping, the data update time interval is: , Each time the call is executed, the time interval is incremented, and the return value is reset to zero. Greater than If so, the update function will be called once.

9. A system for constructing a smart elevator edge digital twin based on cloud-edge collaboration, characterized in that, It includes modeling equipment, edge devices, cloud servers, an elevator group consisting of multiple vertical lifts, and digital twin front-end software; The modeling device includes a data acquisition and drawing module, used to acquire environmental information of the vertical elevator operation scenario, and generate a meta-model file based on the environmental information parameters. The meta-model file includes an elevator shaft model, an elevator subsystem model, a canvas model, and interface files. The edge module device includes a data storage center and an algorithm deployment module; wherein, the data storage center is used to acquire the current subsystem data of the vertical elevator, including dynamic parameter data and static dimension data summarized by the control cabinet near the subsystem and equipment safety status data processed by the algorithm; the algorithm deployment module is used to calculate the equipment safety status information from the dynamic data; The cloud server includes a data storage module and a meta-model module; wherein, the data storage module is used to receive data information uploaded from the edge module; the meta-model module is used to construct a general file package based on the system model files generated by the drawing module.

Images