Data interface and service development method for nuclear power industry heterogeneous mechanism model

Abstract

The invention discloses a data interface and service development method for a nuclear power industry heterogeneous mechanism model, and the method comprises the following steps: S1, creating a data set for the training, testing and verification of a model, and creating a plurality of service components as the development basis of different types of models; s2, creating a project for project management, creating a model under the project, calling a corresponding service component according to the classification of the model to perform model development and debugging, and uploading the debugged model to a mechanism model library; and S3, creating a model service, selecting a model from the mechanism model library to be bound with the model service, and calling a corresponding service component according to the classification of the selected model to operate and publish the model service. According to the invention, an integrated development mode is provided for a nuclear power industrial mechanism model.

Description

Technical Field

[0001] The present invention relates to the technical field of nuclear power industry Internet platform optimization, and in particular to a data interface and service development method for heterogeneous mechanism models of the nuclear power industry. Background Art

[0002] At present, there are many nuclear power plants in China. Each nuclear power plant has a large number of operating units and involves a wide range of models. In addition, many power plants have been in operation for many years. The nuclear power business involves many types and a large number of mechanism models. How to well manage and utilize these nuclear power mechanism models, how to meet the arrangement and debugging requirements of various different mechanism models, and improve user experience have become urgent problems to be solved. Summary of the Invention

[0003] In view of this, it is necessary to provide a data interface and service development method for heterogeneous mechanism models in the nuclear power industry to effectively solve the technical problem of the lack of integrated development and management of mechanism models in the nuclear power industry.

[0004] The present invention provides a data interface and service development method for a heterogeneous mechanism model of the nuclear power industry, comprising the following steps:

[0005] Step S1: Create a dataset for model training, testing, and verification, and create multiple service components as the basis for developing different types of models;

[0006] Step S2: Create a project for project management, create a model under the project, call the corresponding service component according to the model classification to develop and debug the model, and upload the debugged model to the mechanism model library;

[0007] Step S3: Create a model service, select a model from the mechanism model library and bind it to the model service, and call the corresponding service component according to the classification of the selected model to run and publish the model service.

[0008] Preferably, the step S1 creates a data set for model training, testing and verification, specifically:

[0009] The datasets include public datasets and user-created custom datasets; users use the data in the public datasets by adding them to their favorites; users enter basic dataset information and upload data to create custom datasets, which can be set to public or private.

[0010] Preferably, the data set in step S1 also includes an external data set;

[0011] When accessing the external data set, if the accessed data is asset data, first apply for the asset package in the data asset platform, and then query the data through the http interface; if the accessed data is sample data shared in the sample tool, query and download the data through the sample tool http interface.

[0012] Preferably, the service components in step S1 specifically include real-time computing service components, AI service components, cloud simulation service components and industrial software service components.

[0013] Preferably, the project is created in step S2 for project management, specifically:

[0014] Enter the basic information of the project to create the project. The data between projects is isolated and called through open interfaces. The project includes general projects and application projects. The general project provides general model services. The application calls the general model service through the industrial service open platform. The application project is used to provide corresponding model services for specific applications. The application directly calls the corresponding model service.

[0015] Preferably, in step S2, a model is created under the project, and corresponding service components are called according to the classification of the model to perform model development and debugging, specifically:

[0016] Input the basic information of the model. Models are categorized into real-time computing models, AI models, cloud-based simulation models, and industrial software models. Based on the model classification, call the corresponding service component interface to complete model creation.

[0017] After the model is created, the corresponding service component provides development functions according to the model type. If the corresponding service component provides algorithm configuration functions, jump to the corresponding algorithm configuration page to develop the model. If the service component does not provide algorithm configuration functions, upload the model file to develop the model.

[0018] After the model development is completed, the corresponding service component provides debugging functions according to the model type. If the corresponding service component provides algorithm configuration functions, jump to the corresponding algorithm configuration page to debug the model. If the service component does not provide algorithm configuration functions, configure the model parameters through the platform, select the debugging data set, and then call the model running interface of the corresponding service component to submit the debugging related data to the corresponding service component to run the debugging model.

[0019] Preferably, in step S3, the corresponding service component is called according to the classification of the selected model to run the model service, specifically:

[0020] If the service component corresponding to the model classification provides a configuration page, jump to the configuration page to set the operating parameters. If the service component corresponding to the model classification does not provide a configuration page, enter the operating parameters or upload the operating parameter file; select the operating environment according to the model classification; configure the model operation mode according to the model classification;

[0021] The model structure file, the operating parameters, and the operating mode are packaged and downloaded to the operating environment, and an operating task is created based on the model calculation engine of the operating environment to run the model bound to the model service.

[0022] Preferably, in step S3, the corresponding service component is called according to the classification of the selected model to publish the model service, specifically:

[0023] After the model service is completed, the corresponding service component is called according to the model classification to create a model service API interface for publication. After publication, the model service provides a model service interface with permission authentication to the outside world to provide the computing power of the model to the outside world;

[0024] If the project to which the model service belongs is a general project, the model service interface is published to the industrial service open platform, and the model service is called through the industrial service open platform; if the project to which the model service belongs is not a general project, access permissions are set for the model service interface.

[0025] Preferably, the step S3 further includes:

[0026] After the model service is running, the model operation status is monitored and online interactive instructions are sent to the running model tasks; the model service operation results are saved as a new dataset for subsequent training, testing and verification of the model.

[0027] Compared with the existing technology, the present invention has the following beneficial effects: the present invention creates data sets and service components as the basis for model integration development, and calls different service components according to the classification of different models to develop and debug the model, meeting the orchestration and debugging requirements of various types of different mechanism models; different service components are called separately to manage the operation and release of model services. After the model service is released, the user can call the required model service very conveniently, thereby improving the user experience; at the same time, a project is created to manage different models, and the models are divided from a functional perspective. Users can also select model services according to the project. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0029] FIG1 is a flow chart of an embodiment of a data interface and service development method for a heterogeneous mechanism model of the nuclear power industry provided by the present invention;

[0030] Figure 2 yes Figure 1 A schematic diagram of integrated development and operation of an embodiment of a data interface and service development method for a heterogeneous mechanism model of the nuclear power industry in the illustrated embodiment;

[0031] Figure 3a yes Figure 1 A schematic diagram of a model addition, deletion, and modification in an embodiment of a data interface and service development method for a heterogeneous mechanism model of the nuclear power industry;

[0032] Figure 3b yes Figure 1 A schematic diagram of model development for a data interface and service development method for a heterogeneous mechanism model of the nuclear power industry in the embodiment shown;

[0033] Figure 3c yes Figure 1 A schematic diagram of model debugging of an embodiment of a data interface and service development method for a heterogeneous mechanism model of the nuclear power industry in the illustrated embodiment;

[0034] Figure 4a yes Figure 1 A schematic diagram of model service operation of an embodiment of a data interface and service development method for a heterogeneous mechanism model of the nuclear power industry in the illustrated embodiment;

[0035] Figure 4b yes Figure 1 The illustrated embodiment is a schematic diagram of model service publishing for a data interface and service development method for heterogeneous mechanism models in the nuclear power industry. DETAILED DESCRIPTION

[0036] The preferred embodiments of the present invention will be described in detail below in conjunction with the accompanying drawings, wherein the accompanying drawings constitute a part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, and are not used to limit the scope of the present invention.

[0037] Example 1

[0038] See also Figure 1 The data interface and service development method for the heterogeneous mechanism model of the nuclear power industry in this embodiment includes the following steps:

[0039] Step S1: Create a dataset for model training, testing, and verification, and create multiple service components as the basis for developing different types of models;

[0040] Step S2: Create a project for project management, create a model under the project, call the corresponding service component according to the model classification to develop and debug the model, and upload the debugged model to the mechanism model library;

[0041] Step S3: Create a model service, select a model from the mechanism model library and bind it to the model service, and call the corresponding service component according to the classification of the selected model to run and publish the model service.

[0042] like Figure 2 As shown, this embodiment creates a data set and service components as the basis for model integration development. The data set provides a data set for model training, testing, and verification, which can be of various types such as text, image, audio, or multimodal. Users call different service components based on the classification of different models to develop and debug models, meeting the orchestration and debugging requirements of various different mechanism models; call different service components to manage the operation and release of model services. After the model service is released, users can easily call the required model service, improving the user experience; at the same time, create projects to manage different models and divide the models from a functional perspective. Users can also select model services based on the project.

[0043] Specifically, in step S1, a data set is created for model training, testing, and verification, specifically:

[0044] The datasets include public datasets and user-created custom datasets; users use the data in the public datasets by adding them to their favorites; users enter basic dataset information and upload data to create custom datasets, which can be set to public or private.

[0045] The platform provides public datasets, which users can collect and use later. Users can also create their own datasets and upload data, which can then be made public to the platform's public datasets.

[0046] The platform's public datasets are derived from user-created and publicly available data. Display fields include the dataset's Chinese name, English name, dataset description, tags, creation time, and number of favorites. Datasets can be searched using keywords such as dataset name and tag name. Click the Save button to add a dataset to your favorites, and click again to remove it from the favorites list. Users can add datasets to their favorites or make them public. Clicking a dataset opens the dataset details page. Dataset details include basic information, a dataset description, dataset files, and a data preview. Basic information includes the dataset's Chinese name, English name, tags, creation time, and number of favorites. The Dataset Introduction tab displays a brief introduction to the dataset, explaining its background, content, source, uses, and instructions. The Dataset Files tab displays a list of the dataset's files, including file name, file size, and update time. Folders can be displayed. The Data Preview tab is used to preview dataset files. Select a previewable dataset file to preview. Only CSV format files are supported.

[0047] In dataset maintenance, users can search for created and favorite datasets using keywords such as dataset name and tag name. The Create tab displays user-created datasets in pages, displaying fields such as the Chinese name, English name, description, tags, creation time, public status, and buttons for editing and deleting. The Favorites tab displays user-favorited datasets in pages, displaying fields such as the Chinese name, English name, description, tags, creation time, and a favorites button.

[0048] Click the Create Dataset button to pop up the Create Dataset page. The dataset fields include the Chinese name of the dataset, the English name of the dataset, labels, a brief introduction to the dataset, whether it is public, etc. You can upload one or more dataset files, which can be text, images, audio, or multimodal. Multiple files can be organized in folders or compressed packages. The basic information of the dataset created by the user is stored in the database, and the uploaded dataset files are stored in the object storage service, such as MinIO; users can modify the datasets they created, including modifying basic information, adding or deleting dataset files; users can delete the datasets they created, and all files in the dataset will be completely deleted from the object storage service. Dataset Sharing User-created datasets can be shared to public datasets by setting whether to be public.

[0049] Specifically, such as Figure 2 As shown, the data set in step S1 also includes an external data set;

[0050] When accessing the external data set, if the accessed data is asset data, first apply for the asset package in the data asset platform, and then query the data through the http interface; if the accessed data is sample data shared in the sample tool, query and download the data through the sample tool http interface.

[0051] In addition to the created datasets, external datasets can also be accessed. To access data assets, you need to apply for an asset package on the Data Asset Platform. Once approved, you can query the data through the HTTP interface. Samples shared in the sample tool can be queried and downloaded through the sample tool's HTTP interface. Sample files are in CSV format, and multiple files are stored in zip format. First, call the sample tool's Get Token interface, then call the sample tool's Query interface to query the samples shared by the sample tool. Select the sample data you want to use, obtain the sample data ID, and call the Sample Tool's Download Sample interface to download the sample file.

[0052] Specifically, such as Figure 2 As shown, the service components in step S1 specifically include real-time computing service components, AI service components, cloud simulation service components and industrial software service components.

[0053] Specifically, such as Figure 2 As shown, the project is created in step S2 for project management, specifically:

[0054] Enter the basic information of the project to create the project. The data between projects is isolated and called through open interfaces. The project includes general projects and application projects. The general project provides general model services. The application calls the general model service through the industrial service open platform. The application project is used to provide corresponding model services for specific applications. The application directly calls the corresponding model service.

[0055] The platform allows you to create projects to manage them. Data between projects is isolated and can be accessed through open interfaces. Basic project settings include adding, modifying, deleting, and querying. Basic project authorization management includes adding and revoking authorizations, as well as authorizing and opening data interfaces based on permissions. Model management manages all platform models within a project, including real-time computing models, cloud-based simulation models, industrial software models, and AI models.

[0056] Users can search for model information they created or downloaded by model category, model name, or tag name, and view detailed model information, including basic model information and model introduction. Model information sources include user-created models and models downloaded from the mechanism model library (IML).

[0057] Self-created models can be modified, deleted, developed, debugged, and uploaded. If the model classification does not support configuration development, the Develop button will not appear. Models downloaded from the mechanism model library can be modified, deleted, and debugged; users can create, modify, and delete models. Based on the model classification, the add, delete, and modify model interfaces of the four component services are called. After the interface call is successful, the platform model's add, delete, and modify operations can be performed.

[0058] like Figure 3a As shown, create a model: Click the Create Model button to open the Create Model page. Users can create their own models. Basic model information includes the model's Chinese name, English name, description, version number, tags, and model category. If the model category does not support configuration development, you can directly upload the model file for debugging and running. After filling in the basic model information, call the add model interface of the four component services based on the model category. If the interface call is successful, the component returns the model ID created. Then, add the platform model data and bind the model ID returned by the interface.

[0059] Modify Model: Select the platform model data to be modified and obtain the corresponding model IDs of the four component services. Then, based on the model classification, call the Modify Model API of the four component services and submit the model IDs and the modified model basic information to the API. After the API call is successful, the basic information of the platform model is modified.

[0060] Delete Model: Select the platform model data to be deleted and obtain the corresponding model IDs of the four component services. Then, based on the model classification, call the delete model API of the four component services and submit the model IDs of the four component services to the delete model API of the four component services. After the API call is successful, the platform model data is deleted.

[0061] After the basic information of the model is created, different component services can be called to develop the model according to different model categories.

[0062] Specifically, in step S2, a model is created under the project, and corresponding service components are called according to the classification of the model to perform model development and debugging, specifically:

[0063] Input the basic information of the model. Models are categorized into real-time computing models, AI models, cloud-based simulation models, and industrial software models. Based on the model classification, call the corresponding service component interface to complete model creation.

[0064] After the model is created, the corresponding service component provides development functions according to the model type. If the corresponding service component provides algorithm configuration functions, jump to the corresponding algorithm configuration page to develop the model. If the service component does not provide algorithm configuration functions, upload the model file to develop the model.

[0065] After the model development is completed, the corresponding service component provides debugging functions according to the model type. If the corresponding service component provides algorithm configuration functions, jump to the corresponding algorithm configuration page to debug the model. If the service component does not provide algorithm configuration functions, configure the model parameters through the platform, select the debugging data set, and then call the model running interface of the corresponding service component to submit the debugging related data to the corresponding service component to run the debugging model.

[0066] like Figure 3b As shown in the figure, the corresponding component service provides development functions for models of different categories. If the component service provides algorithm configuration, it will jump to the corresponding algorithm configuration page according to the model classification to develop the model. After the model development is completed, the model file and other information will be saved. If the component service does not provide algorithm configuration, you can upload the model file. The platform will call the component service's upload model file interface based on the model classification to upload the model file.

[0067] like Figure 3c As shown, for model debugging, the corresponding component service provides debugging functions for models of different categories. If the component service provides algorithm configuration functions, it will jump to the corresponding algorithm configuration page according to the model classification to debug the model. You can configure the parameters of each node of the model, select the debugging dataset through the platform dataset interface, and finally submit the model and parameters to the component service to run the model. If the component service does not provide algorithm configuration functions, you can configure the model parameters through the platform, select the debugging dataset, and then call the model running interface of the component service to submit the model ID, parameters, etc. to the component service to run the model. The component service provides an interface for querying model running results. The running result formats include forms, JSON files, visual interfaces, logs, etc.

[0068] Model upload: users can upload their own models to the mechanism model library IML. The uploaded information includes basic model information and model files. For example, real-time algorithm model files are XML and RTT files, and AI model files are JSON files. Users select the model to be uploaded, and the platform calls the upload IML interface of the four component services according to the model classification, and sends the model ID to be uploaded to the component service. The upload IML interface of the component service finds the corresponding model basic information and model file based on the passed model ID, calls the upload interface of the mechanism model library, uploads the model basic information and model file to the mechanism model library, and queries or downloads the model by calling the mechanism model library interface.

[0069] To download the model, click the Download Model button on the Model Management page to open the Mechanism Model List page. Users can query the model information that has been applied for and approved on the Mechanism Model Library and perform download operations. The platform calls the interface provided by the Mechanism Model Library IML to query the model information that the user has applied for and approved on the Mechanism Model Library in pages. The list fields include the model's Chinese name, model's English name, model category, version, introduction, and status. For downloaded models, the status is displayed as downloaded, and for models that have not been downloaded, the status is displayed as not downloaded and can be downloaded. When the user clicks the Download button after the model information, the Download IML of the four component services is called according to the model category, and the model interface is created. The model ID of the Mechanism Model Library model to be downloaded is sent to the component service. The Download IML of the component service and the Model Interface are created. First, the Download Model Interface of the Mechanism Model Library is called to download the model, and then the model information is created and the model ID is returned. After the platform successfully calls the Download IML of the component service and the Model Interface is created, the platform model information is created synchronously. After the entire download operation is completed, the user can debug and run the model.

[0070] Specifically, in step S3, the corresponding service component is called according to the classification of the selected model to run the model service, specifically:

[0071] If the service component corresponding to the model classification provides a configuration page, jump to the configuration page to set the operating parameters. If the service component corresponding to the model classification does not provide a configuration page, enter the operating parameters or upload the operating parameter file; select the operating environment according to the model classification; configure the model operation mode according to the model classification;

[0072] The model structure file, the operating parameters, and the operating mode are packaged and downloaded to the operating environment, and an operating task is created based on the model calculation engine of the operating environment to run the model bound to the model service.

[0073] Model service management is where all model services under a project are managed, supporting the addition, deletion, modification, query, operation, release, and monitoring of model services.

[0074] Model service maintenance, model service query: users can query the model services under the project, and support the creation, modification, deletion, operation, release and monitoring of model services. Supports conditional filtering according to model service name, model category, status, etc. Display fields include model service name, model service introduction, model name, model category, model version, creator, creation time, and status. Model service creation: Click the Create Model Service button to pop up the Create Model Service page. Fill in the basic information of the model service. The basic information includes the Chinese name of the model service, the English name, the service introduction, etc. Then select the model from the platform model list of the project, and bind the model to the model service to support the subsequent operation and release of the model service; Model service modification: supports the modification of the basic information of the model service; Model service deletion: If the model service is already running or published, call the interfaces of the four component services, take the model running task offline, and then synchronously delete the model service information of the platform.

[0075] To run the model service, click the Run button after the model service, and the Run Model Service page will pop up. Running the model service means running the model bound to the model service. You need to fill in or upload the model running parameters, running environment, running mode and other information, and then download it to the running environment for running. Figure 4a As shown in the figure, if the component service corresponding to the model category provides a configuration page for parameter setting, the model will be redirected to this page to set the parameters. After completing the configuration, the model runtime parameters will be obtained. If the component service corresponding to the model category does not provide a configuration page for parameter setting, the model runtime parameters can be entered through a form on the platform or a model runtime parameter file can be directly uploaded. After setting the model runtime parameters, the runtime environment is selected based on the model category. Different model categories have different runtime environments, including the model calculation engine, runtime scheduler, and dependency libraries. After selecting the runtime environment, the model runtime mode is configured based on the model category. The runtime mode is divided into periodic and non-periodic. Periodic runtime requires configuring a scheduling policy, and the model calculation engine will schedule the model according to the configured scheduling policy. The platform packages the model structure file, model runtime parameters, and runtime mode of the model to be run, and then downloads them to the selected runtime environment. The model calculation engine in the runtime environment creates a runtime task and runs the model, while also recording the runtime log and results. The platform can view the model's runtime status, runtime log, and runtime results.

[0076] Specifically, in step S3, the corresponding service component is called according to the classification of the selected model to publish the model service, specifically:

[0077] After the model service is completed, the corresponding service component is called according to the model classification to create a model service API interface for publication. After publication, the model service provides a model service interface with permission authentication to the outside world to provide the computing power of the model to the outside world;

[0078] If the project to which the model service belongs is a general project, the model service interface is published to the industrial service open platform, and the model service is called through the industrial service open platform; if the project to which the model service belongs is not a general project, access permissions are set for the model service interface.

[0079] Model service publishing: Model services support service interfaces with permission authentication, providing external access to the model's computing capabilities, accepting input parameters and returning calculation results. Already running model services can be published, and after publishing, the model service interface can be provided externally.

[0080] like Figure 4b As shown, after the model service is running, click the publish button to call the create model service API interface of the four component services according to the model classification to obtain information such as the interface address, request method, and request parameters. Then the publish model service page pops up and is displayed. If the project to which the model service belongs is a general project, the model service interface is published to the industrial service open platform, and the application calls the model service through the industrial service open platform; if the project to which the model service belongs is not a general project, you can set permissions for API access, including setting access tokens (access_token), request rate limits, request number limits, robot detection, and IP geo-fencing. After setting access permissions, it can be opened to the outside world, and external applications can directly call the model service.

[0081] Specifically, such as Figure 2 As shown, the step S3 further includes:

[0082] After the model service is running, the model operation status is monitored and online interactive instructions are sent to the running model tasks; the model service operation results are saved as a new dataset for subsequent training, testing and verification of the model.

[0083] Model service monitoring: Once the model service is running, you can monitor the model's operation status, including task information, status, logs, historical execution records, and results. The platform can send online interactive commands to running model tasks, including common scheduling commands such as start, stop, and restart. The platform supports viewing and downloading operation results in text and file formats, and supports visual viewing of operation results. Model operation results can be added to the user's dataset. The component services corresponding to the model classification provide model operation monitoring capabilities, and the platform calls the component service interfaces to monitor model operation.

[0084] The above description is only a preferred specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by any technician familiar with this technical field within the technical scope disclosed in the present invention should be covered by the present invention.

Claims

1. A data interface and service development method for heterogeneous mechanism models in the nuclear power industry, characterized by: The following steps are involved: Step S1: Create a dataset for model training, testing, and verification, and create multiple service components as the basis for developing different types of models; Step S2: Create a project for project management, create a model under the project, call the corresponding service component according to the model classification to develop and debug the model, and upload the debugged model to the mechanism model library; Step S3: Create a model service, select a model from the mechanism model library and bind it to the model service, and call the corresponding service component according to the classification of the selected model to run and publish the model service.

2. The data interface and service development method for heterogeneous mechanism models in the nuclear power industry according to claim 1 is characterized in that: In step S1, a data set is created for model training, testing, and verification, specifically: The datasets include public datasets and user-created custom datasets; The user uses the data of the public dataset by adding it to the collection; the user enters basic information of the dataset and uploads data to create the custom dataset, which can be set to a public state or a private state.

3. The data interface and service development method for heterogeneous mechanism models in the nuclear power industry according to claim 1 is characterized in that: The data set in step S1 also includes an external data set; When accessing the external data set, if the accessed data is asset data, first apply for the asset package in the data asset platform, and then query the data through the http interface; if the accessed data is sample data shared in the sample tool, query and download the data through the sample tool http interface.

4. The data interface and service development method for heterogeneous mechanism models in the nuclear power industry according to claim 1 is characterized in that: The service components in step S1 specifically include real-time computing service components, AI service components, cloud simulation service components and industrial software service components.

5. The data interface and service development method for heterogeneous mechanism models of the nuclear power industry according to claim 1 is characterized in that: The step S2 is to create a project for project management, specifically: Enter the basic information of the project to create the project. The data between projects is isolated and called through open interfaces. The project includes general projects and application projects. The general project provides general model services. The application calls the general model service through the industrial service open platform. The application project is used to provide corresponding model services for specific applications. The application directly calls the corresponding model service.

6. The data interface and service development method for heterogeneous mechanism models in the nuclear power industry according to claim 1 is characterized in that: In step S2, a model is created under the project, and corresponding service components are called according to the classification of the model to perform model development and debugging, specifically: Input the basic information of the model. Models are categorized into real-time computing models, AI models, cloud-based simulation models, and industrial software models. Based on the model classification, call the corresponding service component interface to complete model creation. After the model is created, the corresponding service component provides development functions according to the model type. If the corresponding service component provides algorithm configuration functions, jump to the corresponding algorithm configuration page to develop the model. If the service component does not provide algorithm configuration functions, upload the model file to develop the model. After the model development is completed, the corresponding service component provides debugging functions according to the model type. If the corresponding service component provides algorithm configuration functions, jump to the corresponding algorithm configuration page to debug the model. If the service component does not provide algorithm configuration functions, configure the model parameters through the platform, select the debugging data set, and then call the model running interface of the corresponding service component to submit the debugging related data to the corresponding service component to run the debugging model.

7. The data interface and service development method for heterogeneous mechanism models in the nuclear power industry according to claim 1 is characterized in that: In step S3, the corresponding service component is called according to the classification of the selected model to run the model service, specifically: If the service component corresponding to the model classification provides a configuration page, jump to the configuration page to set the operating parameters. If the service component corresponding to the model classification does not provide a configuration page, enter the operating parameters or upload an operating parameter file; Select the operating environment according to the model classification; configure the model operation mode according to the model classification; The model structure file, the operating parameters, and the operating mode are packaged and downloaded to the operating environment, and an operating task is created based on the model calculation engine of the operating environment to run the model bound to the model service.

8. The data interface and service development method for heterogeneous mechanism models in the nuclear power industry according to claim 1 is characterized in that: In step S3, the corresponding service component is called according to the classification of the selected model to publish the model service, specifically: After the model service is completed, the corresponding service component is called according to the model classification to create a model service API interface for publication. After publication, the model service provides a model service interface with permission authentication to the outside world to provide the computing power of the model to the outside world; If the project to which the model service belongs is a general project, the model service interface is published to the industrial service open platform, and the model service is called through the industrial service open platform; If the project to which the model service belongs is not a general project, access permission is set for the model service interface.

9. The data interface and service development method for heterogeneous mechanism models in the nuclear power industry according to claim 1 is characterized in that: The step S3 also includes, after the model service is running, monitoring the model operation status and sending online interactive instructions to the running model tasks; saving the model service operation results as a new data set for subsequent training, testing and verification of the model.

Images