Three-dimensional point location binding method and device, and nuclear power station simulation operation method and system

Abstract

The invention discloses a three-dimensional point position binding method and device and a nuclear power station simulation operation method and system, and relates to the technical field of three-dimensional visualization. The three-dimensional point location binding method comprises the following steps: in response to a first input operation of a user based on a first interface, obtaining at least one to-be-bound data source and at least one to-be-bound three-dimensional point location; in response to a first binding determination operation of a user based on a second interface, for each to-be-bound data source in the at least one to-be-bound data source, determining a to-be-bound three-dimensional point location corresponding to the to-be-bound data source from the at least one to-be-bound three-dimensional point location by adopting a target binding model, and binding the to-be-bound data source with the corresponding to-be-bound three-dimensional point location to obtain a first binding relationship. According to the embodiment of the invention, the method can achieve the automatic binding of the three-dimensional point location and the data source, and can improve the binding efficiency of the three-dimensional point location.

Description

Technical Field

[0001] This application belongs to the technical field of 3D visualization, and specifically relates to a 3D point binding method and device, a nuclear power plant simulation operation method and system. Background Art

[0002] 3D visualization technology refers to a technical system that, based on physical entities, uses 3D modeling technology to digitally model them and form corresponding virtual objects. 3D visualization technology can be associated with various fields. Currently, 3D visualization technology has been widely applied in various fields such as mechanical manufacturing, architectural design, and the nuclear industry. 3D visualization technology has broad application prospects in the future digital and intelligent fields.

[0003] With the continuous development of 3D scenes, during the construction of 3D scenes, limited by the characteristics of 3D models themselves and the computing power of computers, the system requires professional developers to operate, with a relatively high threshold and inconvenient operation and maintenance. As a result, operators need to spend a lot of energy on the binding work between 3D points and data sources, leading to low efficiency in 3D point binding. Summary of the Invention

[0004] The technical problem to be solved by this application is to address the above-mentioned deficiencies in the prior art and provide a 3D point binding method and device, a nuclear power plant simulation operation method and system. Using this 3D point binding method, it is possible to automatically bind the 3D points to be bound and their corresponding data sources to be bound, thereby improving the efficiency of 3D point binding.

[0005] In a first aspect, an embodiment of this application provides a 3D point binding method, including:

[0006] In response to a first input operation by the user based on a first interface, obtain at least one data source to be bound and at least one 3D point to be bound;

[0007] In response to a first confirm binding operation by the user based on a second interface, for each data source to be bound among at least one data source to be bound, use a target binding model to determine the 3D point to be bound corresponding to the data source to be bound from at least one 3D point to be bound, and bind the data source to be bound with its corresponding 3D point to be bound to obtain a first binding relationship;

[0008] Wherein, the target binding model is obtained by training a large language model with sample data sources and their corresponding sample 3D points.

[0009] In some embodiments of the first aspect, before responding to the first confirm binding operation by the user based on the second interface, the method further includes:

[0010] In response to a second input operation of the user based on the third interface, obtain a target matching rule;

[0011] Using the target binding model to determine the corresponding three-dimensional point to be bound from at least one three-dimensional point to be bound includes:

[0012] Based on the target matching rule, using the target binding model to determine the corresponding three-dimensional point to be bound from at least one three-dimensional point to be bound.

[0013] In some embodiments of the first aspect, before using the target binding model to determine the corresponding three-dimensional point to be bound from at least one three-dimensional point to be bound based on the target matching rule, the method further includes:

[0014] Extract a three-dimensional point screening rule from the target matching rule;

[0015] According to the three-dimensional point screening rule, screen at least one three-dimensional point to be bound to obtain at least one target three-dimensional point;

[0016] Based on the target matching rule, using the target binding model to determine the corresponding three-dimensional point to be bound of the data source to be bound from at least one three-dimensional point to be bound includes:

[0017] Based on the target matching rule, using the target binding model to determine the corresponding target three-dimensional point of the data source to be bound from at least one target three-dimensional point.

[0018] In some embodiments of the first aspect, before using the target binding model to determine the corresponding three-dimensional point to be bound of the data source to be bound from at least one three-dimensional point to be bound based on the target matching rule, the method further includes:

[0019] Extract a data source screening rule from the target matching rule;

[0020] According to the data source screening rule, screen at least one data source to be bound to obtain at least one target data source;

[0021] Based on the target matching rule, using the target binding model to determine the corresponding three-dimensional point to be bound of the data source to be bound from at least one three-dimensional point to be bound includes:

[0022] Based on the target matching rule, using the target binding model to determine the corresponding three-dimensional point to be bound of the target data source from at least one three-dimensional point to be bound.

[0023] In some embodiments of the first aspect, after binding the data source to be bound with its corresponding three-dimensional point to be bound to obtain a first binding relationship, the method further includes:

[0024] Invoke the agent to perform a logical check on the configuration file to which the first binding relationship belongs, and obtain the first check result;

[0025] When the first check result is passed, invoke the agent to execute the first binding relationship.

[0026] In some embodiments of the first aspect, after invoking the agent to execute the first binding relationship, the method further includes:

[0027] Invoke the 3D platform for inspection to obtain a second check result, wherein the 3D platform inspects the first binding relationship based on the target matching rule;

[0028] When the second check result is passed, visually display the first binding relationship on the fourth interface.

[0029] In some embodiments of the first aspect, after visually displaying the first binding relationship, the method further includes:

[0030] Respond to the user's first modification operation based on the fourth interface, and modify the first binding relationship;

[0031] Or,

[0032] Respond to the user's second modification operation based on the fourth interface, modify the target matching rule, and obtain the modified target matching rule;

[0033] Based on the modified target matching rule, use the target binding model to determine the corresponding to-be-bound 3D point from at least one to-be-bound 3D point, and bind the to-be-bound data source to its corresponding to-be-bound 3D point to obtain the updated first binding relationship.

[0034] In some embodiments of the first aspect, before using the target binding model to determine the to-be-bound 3D point corresponding to the to-be-bound data source from at least one to-be-bound 3D point, the method further includes:

[0035] Respond to the user's third input operation based on the fifth interface, and determine the user's task type;

[0036] When the task type is the target task type, display the first interface.

[0037] In some embodiments of the first aspect, before using the target binding model to determine the to-be-bound 3D point corresponding to the to-be-bound data source from at least one to-be-bound 3D point, the method further includes:

[0038] Create a target binding model;

[0039] Creating a target binding model specifically includes:

[0040] Obtain the sample data source and its corresponding sample 3D point;

[0041] Training a large language model with a sample data source and its corresponding sample three-dimensional point positions to obtain a target binding model.

[0042] Based on the same inventive concept, in a second aspect, an embodiment of the present application further provides a method for simulating the operation of a nuclear power plant, the method including:

[0043] Determining a first binding relationship according to the three-dimensional point position binding method of any item in the first aspect;

[0044] Simulating the operation of a nuclear power plant according to the first binding relationship.

[0045] Based on the same inventive concept, in a third aspect, an embodiment of the present application further provides a three-dimensional point position binding device, including:

[0046] A first acquisition module, configured to acquire at least one data source to be bound and at least one three-dimensional point position to be bound in response to a first input operation of a user based on a first interface;

[0047] A first determination module, connected to the first acquisition module, configured to, in response to a first determination and binding operation of a user based on a second interface, for each data source to be bound in at least one data source to be bound, use the target binding model to determine the three-dimensional point position to be bound corresponding to the data source to be bound from at least one three-dimensional point position to be bound, and bind the data source to be bound and its corresponding three-dimensional point position to be bound to obtain a first binding relationship;

[0048] Wherein, the target binding model is obtained by training a large language model with a sample data source and its corresponding sample three-dimensional point positions.

[0049] In some embodiments of the third aspect, the device further includes:

[0050] A second acquisition module, connected to the first determination module, configured to acquire a target matching rule in response to a second input operation of a user based on a third interface;

[0051] The first determination module is specifically configured to:

[0052] Based on the target matching rule, use the target binding model to determine the three-dimensional point position to be bound corresponding to it from at least one three-dimensional point position to be bound.

[0053] In some embodiments of the third aspect, the device further includes:

[0054] A first extraction module, configured to screen at least one three-dimensional point position to be bound according to a three-dimensional point position screening rule to obtain at least one target three-dimensional point position;

[0055] The first screening module, connected to the first extraction module, is used to screen at least one three-dimensional point to be bound according to the three-dimensional point screening rule, and obtain at least one target three-dimensional point;

[0056] The first determination module, also connected to the first screening module, is specifically used for:

[0057] Based on the target matching rule, the target binding model is used to determine the target three-dimensional point corresponding to the data source to be bound from at least one target three-dimensional point.

[0058] In some embodiments of the third aspect, the device further includes:

[0059] The second extraction module is used to extract the data source screening rule from the target matching rule;

[0060] The second screening module, connected to the second extraction module, is used to screen at least one data source to be bound according to the data source screening rule, and obtain at least one target data source;

[0061] The first determination module, also connected to the second screening module, is specifically used for:

[0062] Based on the target matching rule, the target binding model is used to determine the three-dimensional point to be bound corresponding to the target data source from at least one three-dimensional point to be bound.

[0063] Based on the same inventive concept, in the fourth aspect, the embodiment of the present application further provides a nuclear power plant simulation operation system, and the system includes:

[0064] The three-dimensional point binding device according to any one of the first aspects, which is used to determine the first binding relationship;

[0065] The simulation operation module, connected to the three-dimensional point binding device, is used to simulate the operation of the nuclear power plant according to the first binding relationship.

[0066] According to the three-dimensional point binding method and device, nuclear power plant simulation operation method and system provided by the embodiments of the present application, first, in response to the user's first input operation based on the first interface, at least one data source to be bound and at least one three-dimensional point to be bound are obtained; then, in response to the user's first determination and binding operation based on the second interface, for each data source to be bound in at least one data source to be bound, the target binding model is used to determine the three-dimensional point to be bound corresponding to the data source to be bound from at least one three-dimensional point to be bound, and the data source to be bound is bound to its corresponding three-dimensional point to be bound to obtain the first binding relationship, so as to automatically bind the three-dimensional point to be bound and its corresponding data source to be bound, thereby improving the efficiency of three-dimensional point binding. Description of the Drawings

[0067] Figure 1Shows a schematic flowchart of a three-dimensional point position binding method provided by an embodiment of the present application;

[0068] Figure 2 Shows a schematic diagram of a basic platform functional architecture provided by an embodiment of the present application;

[0069] Figure 3 Shows another schematic flowchart of a three-dimensional point position binding method provided by an embodiment of the present application;

[0070] Figure 4 Shows a schematic diagram of a technical framework provided by an embodiment of the present application;

[0071] Figure 5 Shows a schematic structural diagram of a three-dimensional point position binding device provided by an embodiment of the present application. Detailed implementation manners

[0072] To enable those skilled in the art to better understand the technical solutions of the present application, the present application will be further described in detail below with reference to the accompanying drawings and embodiments.

[0073] The features and exemplary embodiments of various aspects of the present application will be described in detail below. To make the objectives, technical solutions, and advantages of the present application clearer and more understandable, the present application will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only configured to explain the present application and are not configured to limit the present application. For those skilled in the art, the present application can be implemented without some of these specific details. The following description of the embodiments is only provided to provide a better understanding of the present application by showing examples of the present application.

[0074] It should be noted that in this document, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements, but also includes other elements not expressly listed, or further includes elements inherent to such process, method, article or device. Without further limitation, elements defined by the statement "including..." do not exclude the presence of additional identical elements in the process, method, article or device including the said elements.

[0075] As described in the background art, the operator needs to spend a lot of energy on three-dimensional point position binding work such as model processing and import, model relationship setting, point position placement, data collection, and model-data binding, resulting in low efficiency of three-dimensional point position binding.

[0076] Example 1

[0077] The three-dimensional point position binding method provided by the embodiment of the present application is applicable to the simulation operation process of a nuclear power plant. The three-dimensional point position binding method can be executed by a three-dimensional point position binding device, an electronic device, etc. Hereinafter, an example will be given in which the three-dimensional point position binding method is executed by an electronic device.

[0078] As Figure 1 shown, the three-dimensional point position binding method provided by the embodiment of the present application may include steps S110 to S120.

[0079] S110. In response to a first input operation of the user based on the first interface, obtain at least one data source to be bound and at least one three-dimensional point position to be bound.

[0080] S120. In response to a first confirmation binding operation of the user based on the second interface, for each data source to be bound in at least one data source to be bound, use the target binding model to determine the three-dimensional point position to be bound corresponding to the data source to be bound from at least one three-dimensional point position to be bound, and bind the data source to be bound with its corresponding three-dimensional point position to be bound to obtain a first binding relationship.

[0081] Among them, the target binding model is obtained by training a large language model (LLM) with a sample data source and its corresponding sample three-dimensional point position.

[0082] According to the three-dimensional point position binding method provided by the embodiment of the present application, first, in response to a first input operation of the user based on the first interface, obtain at least one data source to be bound and at least one three-dimensional point position to be bound; then, in response to a first confirmation binding operation of the user based on the second interface, for each data source to be bound in at least one data source to be bound, use the target binding model to determine the three-dimensional point position to be bound corresponding to the data source to be bound from at least one three-dimensional point position to be bound, and bind the data source to be bound with its corresponding three-dimensional point position to be bound to obtain a first binding relationship, so as to automatically bind the three-dimensional point position to be bound and its corresponding data source to be bound, thereby improving the efficiency of three-dimensional point position binding.

[0083] In step S110, exemplarily, the first interface may be an interface capable of inputting at least one data source to be bound and at least one three-dimensional point position to be bound.

[0084] Exemplarily, the first input operation may be an operation capable of inputting at least one data source to be bound and at least one three-dimensional point position to be bound.

[0085] Exemplarily, the data source to be bound can be a data source that is needed but not bound to three-dimensional point positions. Among them, the data source can be the identification and parameters of an object entity, etc. For example, the data source can be the temperature value detected by a temperature sensor.

[0086] Exemplarily, the three-dimensional point position to be bound can be a three-dimensional point position that is needed but not bound to a data source. Among them, the three-dimensional point position can be a position point clearly defined by a coordinate system in three-dimensional space, and its core function is to accurately describe the position and shape of an object or phenomenon in three-dimensional space.

[0087] As an example, an electronic device can be communicatively connected to an Internet of Things (IOT) platform system and a three-dimensional platform. The electronic device can call the IOT platform interface to obtain multiple data sources from the IOT platform system and display the multiple data sources in the form of options on a first interface; the electronic device can also call the three-dimensional platform interface to obtain multiple three-dimensional point positions from the three-dimensional platform and display the multiple three-dimensional point positions in the form of options on the first interface. At this time, the first input operation can be an operation in which the user can batch-select the options corresponding to the data source and the options corresponding to the three-dimensional point position based on the first interface.

[0088] As another example, a data source input box to be bound and a three-dimensional point position input box to be bound are displayed on the first interface. The first input operation can be an operation in which the user inputs the data source to be bound in the data source input box to be bound and an operation in which the user inputs the three-dimensional point position to be bound in the three-dimensional point position input box to be bound.

[0089] In some embodiments, before responding to the user's first input operation based on the first interface, the method further includes:

[0090] Responding to the user's third input operation based on the fifth interface, determining the user's task type;

[0091] When the task type is the target task type, display the first interface.

[0092] In this embodiment, the first interface is displayed only when the task type is the target task type, so that while the user performs the first input operation based on the first interface, requests that are not of the target task type can be intercepted, thereby reducing the execution of meaningless tasks and saving computing resources.

[0093] Exemplarily, the fifth interface can be an interface capable of inputting the user's task type.

[0094] Exemplarily, the third input operation may be an operation for the user to input a task type. For example, the third input operation may be an operation of inputting a task type in a task type input box displayed on the fifth interface, or the third input operation may also be an operation of selecting a preset task type from multiple preset task types displayed on the fifth interface as the task type.

[0095] Exemplarily, the target task type may be a task type in which three-dimensional points are bound to a data source.

[0096] Exemplarily, when the task type is not the target task type, a first error prompt message is displayed on the fifth interface, and the first error prompt message is used to indicate that the task type is incorrect.

[0097] In step S120, after the electronic device obtains at least one data source to be bound and at least one three-dimensional point to be bound in response to the user's first input operation based on the first interface, it may also respond to the user's first confirmation binding operation based on the second interface, and for each data source to be bound in the at least one data source to be bound, determine the three-dimensional point to be bound corresponding to the data source to be bound from the at least one three-dimensional point to be bound by using the target binding model, and bind the data source to be bound with its corresponding three-dimensional point to be bound to obtain a first binding relationship.

[0098] It should be noted that the core purpose of binding the data source to be bound with its corresponding three-dimensional point to be bound is to realize the dynamic mapping between the physical entity and the virtual three-dimensional model.

[0099] Exemplarily, the second interface may be an interface capable of performing a confirmation binding operation.

[0100] Exemplarily, the first confirmation binding operation may be an operation of confirming to bind at least one data source to be bound with at least one three-dimensional point to be bound.

[0101] Exemplarily, the first binding relationship may be a binding relationship between at least one data source to be bound and its corresponding three-dimensional point to be bound.

[0102] For example, a confirmation binding option may be displayed on the second interface, and the first confirmation binding operation may be an operation for the user to click the confirmation binding option.

[0103] It can be understood that the first interface and the second interface may be the same interface or different interfaces, which is not limited herein.

[0104] It should be noted that the number of data sources to be bound and the number of three-dimensional points to be bound may be the same or different, which is not limited herein.

[0105] Here, the electronic device invokes a large language model to output code. The code content mainly calls the three-dimensional platform interface for data source entry docking. The output configuration file includes a first binding relationship, and this configuration file can call the three-dimensional platform interface to write the first binding relationship.

[0106] Exemplarily, the large language model may include the deepseek model, etc.

[0107] In some embodiments, before using the target binding model to determine the three-dimensional point to be bound corresponding to the data source to be bound from at least one three-dimensional point to be bound, the method further includes:

[0108] Create a target binding model;

[0109] Creating a target binding model specifically includes:

[0110] Obtain the sample data source and its corresponding sample three-dimensional point;

[0111] Use the sample data source and its corresponding sample three-dimensional point to train the large language model to obtain the target binding model.

[0112] In this embodiment, by using the sample data source and its corresponding sample three-dimensional point to train the large language model to obtain the target binding model, it provides a basis for automatically binding the three-dimensional points using this target binding model later.

[0113] Exemplarily, the previously bound data source and its corresponding three-dimensional point can be used as the sample data source and its corresponding sample three-dimensional point respectively.

[0114] Exemplarily, based on the fine-tuning technology of the large language model, use the sample data source and its corresponding sample three-dimensional point to train the large language model to obtain the target binding model.

[0115] Exemplarily, the embodiments of the present application use a large language model as the main body, and the fine-tuning of the large language model is the key technology of the project. The samples used in the fine-tuning need to be customized and organized according to actual needs, so the joint cooperation of business experts and technical experts is required. The embodiments of the present application have innovation in data organization (i.e., the sample data source and its corresponding sample three-dimensional point). Fine-tuning itself, as a general technology, does not have innovative optimization.

[0116] It should be noted that in actual work, the binding work of three-dimensional points and data sources is very cumbersome and there is no possibility of simple regularization. By training with a large language model to obtain the target binding model and automatically binding the data source and three-dimensional points through the target binding model, the work of the operator can be simplified, and thus the efficiency of automatically binding the data source and three-dimensional points can be improved.

[0117] In some embodiments, before the first determination and binding operation by the user based on the second interface, the method further includes:

[0118] In response to a second input operation by the user based on the third interface, obtain a target matching rule;

[0119] Determine the corresponding three-dimensional point to be bound from at least one three-dimensional point to be bound by using a target binding model, including:

[0120] Based on the target matching rule, use the target binding model to determine the corresponding three-dimensional point to be bound from at least one three-dimensional point to be bound.

[0121] In this embodiment, by interacting with the user, obtain the target matching rule, and then based on the target matching rule, use the target binding model to determine the corresponding three-dimensional point to be bound from at least one three-dimensional point to be bound, which can achieve precise services for users and improve the user experience. In addition, based on the target matching rule input by the user, the three-dimensional point to be bound and the data source to be bound can be automatically bound, that is, the matching rule can be dynamically adjusted, which can improve the flexibility of three-dimensional point binding.

[0122] Exemplarily, the third interface can be an interface capable of inputting a target matching rule. It should be noted that the third interface can be the same as at least one of the first interface and the second interface, or can be different from both the first interface and the second interface, which is not limited herein.

[0123] Exemplarily, the second input operation can be an operation capable of inputting a target matching rule. For example, the second input operation can be an operation of inputting a target matching rule in the matching rule input box on the third interface.

[0124] Exemplarily, the target matching rule can be a matching rule between the three-dimensional point to be bound and the data source to be bound. For example, the target matching rule can include: the identifier of the data source to be bound is Ti, and the identifier of the three-dimensional point to be bound corresponding to the data source to be bound is CAMi, where i is a positive integer.

[0125] It can be understood that the target binding model determines the three-dimensional point to be bound corresponding to the data source to be bound from at least one three-dimensional point to be bound according to the target matching rule.

[0126] In some embodiments, before using the target binding model to determine the corresponding three-dimensional point to be bound from at least one three-dimensional point to be bound based on the target matching rule, the method further includes:

[0127] Extract a three-dimensional point screening rule from the target matching rule;

[0128] According to the three-dimensional point position screening rule, at least one to-be-bound three-dimensional point position is screened to obtain at least one target three-dimensional point position;

[0129] Based on the target matching rule, using the target binding model to determine the to-be-bound three-dimensional point position corresponding to the to-be-bound data source from at least one to-be-bound three-dimensional point position, including:

[0130] Based on the target matching rule, using the target binding model to determine the target three-dimensional point position corresponding to the to-be-bound data source from at least one target three-dimensional point position.

[0131] In this embodiment, by screening at least one to-be-bound three-dimensional point position through the three-dimensional point position screening rule to obtain at least one target three-dimensional point position, the to-be-bound three-dimensional point positions that do not conform to the three-dimensional point position screening rule can be removed, thereby reducing the data processing volume and saving computing resources.

[0132] Exemplarily, the three-dimensional point position screening rule can be a rule capable of screening at least one target three-dimensional point position from at least one to-be-bound three-dimensional point position.

[0133] Exemplarily, the to-be-bound three-dimensional point positions that conform to the three-dimensional point position screening rule among at least one to-be-bound three-dimensional point position are determined as target three-dimensional point positions to obtain at least one target three-dimensional point position. For example, the identifiers of the to-be-bound three-dimensional point positions include CAM1, CAM2, CAM4, CAM5, M1, M3, W5, and the three-dimensional point position screening rule includes that the identifier corresponding to the to-be-bound three-dimensional point position is CAMi, then M1, M3, and W5 are removed, and CAM1, CAM2, CAM4, and CAM5 are determined as target three-dimensional point positions.

[0134] Exemplarily, the target binding model determines the to-be-bound three-dimensional point position corresponding to the target data source from at least one to-be-bound three-dimensional point position based on the target matching rule.

[0135] In some embodiments, before using the target binding model to determine the to-be-bound three-dimensional point position corresponding to the to-be-bound data source from at least one to-be-bound three-dimensional point position based on the target matching rule, the method further includes:

[0136] Extract the data source screening rule from the target matching rule;

[0137] According to the data source screening rule, screen at least one to-be-bound data source to obtain at least one target data source;

[0138] Based on the target matching rule, using the target binding model to determine the to-be-bound three-dimensional point position corresponding to the to-be-bound data source from at least one to-be-bound three-dimensional point position, including:

[0139] Based on the target matching rule, a target binding model is used to determine the three-dimensional point to be bound corresponding to the target data source from at least one three-dimensional point to be bound.

[0140] In this embodiment, at least one target data source is obtained by screening at least one data source to be bound through a data source screening rule, which can remove the data sources to be bound that do not conform to the data source screening rule, and thus can reduce the amount of data processing and save computing resources.

[0141] Exemplarily, the data source screening rule can be a rule that can screen at least one target data source from at least one data source to be bound.

[0142] Exemplarily, the data source to be bound that conforms to the data source screening rule among at least one data source to be bound is determined as the target data source to obtain at least one target data source. For example, the identifiers of the data sources to be bound include T1, T2, T5, T7, K1, K3, L1, and the three-dimensional point screening rule includes that the identifier corresponding to the data source to be bound is Ti, then K1, K3, and L1 are removed, and T1, T2, T5, and T7 are determined as the target three-dimensional points.

[0143] Exemplarily, the target binding model determines the three-dimensional point to be bound corresponding to the data source to be bound from at least one target three-dimensional point based on the target matching rule.

[0144] In some embodiments, after binding the data source to be bound with the three-dimensional point to be bound corresponding thereto to obtain a first binding relationship, the method further includes:

[0145] Invoking an Agent to perform a logical check on the first binding relationship to obtain a first check result;

[0146] In the case where the first check result is passed, invoking the Agent to execute the first binding relationship.

[0147] In this embodiment, by invoking the Agent to perform a logical check on the configuration file to which the first binding relationship belongs, and in the case where the first check result is passed, invoking the Agent to execute the first binding relationship, the logical accuracy of the first binding relationship executed by the Agent can be ensured.

[0148] Specifically, call the Agent to perform Python code checking and logical checking on the configuration file to which the first binding relationship belongs, and check whether the code requirements and logical requirements are met. When the code requirements and logical requirements are met, the first check result is passed; when at least one of the code requirements and logical requirements is not met, the first check result is not passed. When the first check result is passed, call the Agent to execute the Python code on the configuration file to which the first binding relationship belongs. When the first check result is not passed, output a second error prompt message, which is used to indicate that the logical check fails. Alternatively, when the first check result is not passed, modify the configuration file to which the first binding relationship belongs until the first check result corresponding to the configuration file to which the first binding relationship belongs is passed.

[0149] Among them, the code requirements and logical requirements can be preset in the electronic device according to the actual situation, or can be obtained according to the user's operations on the input code requirements and logical requirements based on the electronic device. For example, the code requirements may include whether the three-dimensional platform interface call parameters conform to the OpenAPI specification, and the three-dimensional platform interface call parameters may include authentication headers and data format declarations, etc.; the logical requirements may include whether the interface call order meets the requirements of the three-dimensional platform, etc.

[0150] In some embodiments, after calling the intelligent agent to execute the first binding relationship, the method further includes:

[0151] Call the three-dimensional platform for checking to obtain a second check result, where the three-dimensional platform checks the first binding relationship based on the target matching rule;

[0152] When the second check result is passed, visually display the first binding relationship on the fourth interface.

[0153] In this embodiment, by calling the three-dimensional platform for checking to obtain a second check result, and when the second check result is passed, visually displaying the first binding relationship on the fourth interface, the accuracy of the first binding relationship can be improved, and the user can intuitively know the first binding relationship.

[0154] Exemplarily, the fourth interface can be an interface capable of displaying the first binding relationship.

[0155] Exemplarily, when the first binding relationship conforms to the target matching rule and the execution result is correct, it is determined that the second check result is passed; when the first binding relationship does not conform to the target matching rule, is not executed or the execution result is incorrect, it is determined that the second check result is not passed.

[0156] Exemplarily, in the case where the second inspection result fails, steps S110 to S120 can be re-executed to update the first binding relationship.

[0157] In some embodiments, after visually presenting the first binding relationship, the method further includes:

[0158] responding to a first modification operation of the user based on the fourth interface to modify the first binding relationship;

[0159] Or,

[0160] responding to a second modification operation of the user based on the fourth interface to modify the target matching rule to obtain a modified target matching rule;

[0161] Based on the modified target matching rule, using the target binding model to determine the corresponding three-dimensional point to be bound from at least one three-dimensional point to be bound, and binding the data source to be bound to its corresponding three-dimensional point to be bound to obtain an updated first binding relationship.

[0162] In this embodiment, by responding to the first modification operation of the user based on the fourth interface to modify the first binding relationship, that is, manually correcting the first binding relationship, the accuracy of the first binding relationship can be further improved. Or, responding to the second modification operation of the user based on the fourth interface to modify the target matching rule to obtain a modified target matching rule, and based on the modified target matching rule, using the target binding model to determine the corresponding three-dimensional point to be bound from at least one three-dimensional point to be bound, that is, by manually adjusting the target matching rule, so that the updated first binding relationship determined according to the modified target matching rule better meets the user's needs.

[0163] Exemplarily, the first modification operation can be an operation capable of modifying the first binding relationship. For example, the first modification operation can be an operation of modifying the first binding relationship in the first binding relationship editing box on the fourth interface.

[0164] Exemplarily, the second modification operation can be an operation capable of modifying the target matching rule. For example, the second modification operation can be an operation of modifying the target matching rule in the matching rule editing box on the fourth interface.

[0165] To better understand the three-dimensional point binding method provided by the embodiments of the present application, a specific application embodiment will be described below.

[0166] A three-dimensional visualization point automatic matching platform for the nuclear industry uses an artificial intelligence (AI) large language model for core logic processing and builds an Agent framework. This automatic matching platform includes the following functions: data source (i.e., the data source to be bound) selection, model (i.e., the three-dimensional points to be bound) selection, matching rule (i.e., the target matching rule) description, automatic matching, and matching adjustment.

[0167] Data source selection: Conduct IOT platform system interaction on the interface (i.e., the first interface) to batch-select the data sources maintained by the IOT platform.

[0168] Model selection: Select models in the three-dimensional scene, and batch selection, group selection, rule selection, etc. can be used.

[0169] Matching rule description: Communicate with the LLM in the interaction interface (i.e., the third interface) to describe the matching rules.

[0170] Automatic matching: Click the button (i.e., the first confirm binding operation) to trigger the automatic matching work, write an automatic matching script based on the data source, model source (i.e., the three-dimensional points to be bound), and rule (i.e., the target matching rule), and execute it.

[0171] Matching adjustment: View the matching situation on the interaction interface (i.e., the fifth interface), adjust the matching rules or directly adjust the matching results (i.e., the first binding relationship).

[0172] As Figure 2 shown, the tool solution includes four parts: interactive Web, LLM, Agent, and task planning. Other parts such as external systems like the IOT platform and three-dimensional scene rendering platform (i.e., the three-dimensional platform), as well as the backend service, are replaceable components, and only the interfaces need to be kept available in different projects.

[0173] 1.1 Interactive Web

[0174] Provide a multi-round dialogue interface for user language communication in the interactive Web. In the interface, users communicate about matching rules, model source selection, data source selection, and adjust matching results.

[0175] Develop using Vue technology, connect to the backend interface of the backend service and display data.

[0176] Support unauthorized access and single sign-on, and are convenient to be embedded in other systems.

[0177] 1.2 Task Planning

[0178] Generate rules based on the prompts of the large language model. The system has a set of default and optimized prompts for interacting with the large language model, and performs matching workflow orchestration based on the data source input by the user (i.e., the data source to be bound). The orchestrated process is as follows Figure 3 As shown, the process is as follows: 1) Requirement understanding: Understand the content input by the user (i.e., the user's task type), and reject and return non-matching tasks (i.e., tasks that are not the target task type). When the user's requirement is a match, proceed to the following steps. 2) Model source parsing: Call the three-dimensional platform interface to obtain a list of model names (i.e., at least one three-dimensional point), and understand the model list (i.e., according to the three-dimensional point screening rules, screen at least one three-dimensional point to be bound to obtain at least one target three-dimensional point). 3) Data source parsing: Call the IOT platform interface to obtain a list of data sources, and understand the data source name and configuration (i.e., according to the data source screening rules, screen at least one data source to be bound to obtain at least one target data source). 4) Code writing: Call the LLM to output code. The code content (i.e., the configuration file to which the first binding relationship belongs) is mainly to call the three-dimensional platform interface for data source entry docking. 5) Code checking: Call the Agent (i.e., the intelligent agent) to perform Python code checking and logical checking to see if there is any content that does not meet the requirements. 6) Code execution: Call the Agent to execute the Python code. 7) Result checking: Call the three-dimensional platform to check the point-source matching result to see if it is executed and whether the execution result is correct.

[0179] 1.3 Agent

[0180] Run code and call interfaces based on Python. For different types of interfaces, call the LLM to generate call code, and then execute the code through the Python executor.

[0181] 1.4 LLM

[0182] Based on the dynamic quantization version of the large language model of DeepSeek, a 671B volume model using 1.58-bit quantization is deployed, and the model operation deployment is adjusted so that the system runs the large language model by mixing Dynamic Random-Access Memory (DRAM) and Video Random-Access Memory (VRAM).

[0183] 1.5 Back-end service

[0184] The back-end service built based on Python Flask provides capabilities such as user login, permission management, conversation data storage, and call log storage.

[0185] 1.6 External System

[0186] The external system mainly includes an IOT platform and a 3D platform. The IOT platform provides a data source interface, through which an IOT data source list, including the list item name, attributes, field types, refresh frequency, etc., can be obtained. The 3D platform provides a model list acquisition interface and a model-data source matching interface. The model acquisition interface provides the scene model name, location, and grouping (model hierarchical relationship) information. The model-data source matching interface provides the binding operation between the model and the data source, enabling the 3D system to obtain data refresh by itself through data source settings.

[0187] Exemplarily, in the embodiments of the present application, the user can associate the relationships among the model, devices (i.e., data sources), and points through the dialogue LLM. In the fine-tuning key technologies, supervised fine-tuning (SFT) and reinforcement learning fine-tuning are the two main methods, which are introduced below:

[0188] 2.1 SFT Fine-Tuning Management

[0189] Build SFT data before fine-tuning to prepare for fine-tuning. The data sources are as follows: 1) Historical actual matching data. 2) Rule construction data. 3) Other LLM-generated data. 4) Manually designed data. Among them, the historical data needs to be sorted, cleaned, and structured, and the historical matching results are labeled as positive and negative categories. The rule-generated data needs to implement various possible rules and then generate matching data through the rules. The LLM-generated data uses prompts for prompting words, based on the historical data and rule construction data as the basis, and the LLM generates possible artificial conversations. The manually designed data adjusts the background, requirements, rule descriptions, etc. based on the data generated by the LLM for manual construction.

[0190] 2.2 Model Data Management

[0191] Use the Unsloth framework for fine-tuning. After fine-tuning, store the weights as gguf (GPT-Generated Unified Format), and split the file to ensure that no file exceptions occur. At the same time, perform access permission management on the system to control the access (copying, moving, etc.) of the model and protect the intellectual property rights of the model data. Among them, gguf is an efficient file storage format designed specifically for large language models, the GPT-generated unified file format.

[0192] 2.3 Runtime Data Management

[0193] Operation data management records the customer conversations, matching results, and correction situations during actual operation and stores them in the database. In the next cycle, targeted fine-tuning is performed for cases of matching failure, and the matching is incorporated into the test database to provide data support for subsequent iterations.

[0194] 2.4 Strengthen Fine-Tuning Data Management

[0195] The strengthened fine-tuning data is composed of the actual user usage data stored in the operation data management after cleaning, screening, and structuring. Among them, items with matching failures are extracted first, the content in the user conversation that does not describe the matching rules is deleted, then the corresponding error descriptions are written, and then combined into a fixed data structure for use as strengthened fine-tuning data in fine-tuning.

[0196] Exemplarily, the technical framework diagram provided by the embodiments of this application is as Figure 4 shown. The system is divided into several parts: front-end, back-end, Agent, task planning, and LLM. The front-end includes Vue (Vue.js), HTML (HyperText Markup Language), and Js (JavaScript); the back-end includes REST API (Representational State Transfer Application Programming Interface), Python Flask, and MySQL; Agent includes Dify, and task planning includes Workflow; LLM includes model data management, fine-tuning data management, and Unsloth (Unsloth Efficient Fine-Tuning Framework) model fine-tuning and deployment. Among them, Vue is a progressive JavaScript framework for building user interfaces; Python Flask is a lightweight web framework; Dify is an AI (Artificial Intelligence) intelligent agent development platform. In the front-end part, the front-end page is developed using Vue, HTML, and JS technologies to provide a user interaction interface; in the back-end part, Python-Flask is used to provide RESTAPI interfaces to implement functions such as front-end interface login, LLM access forwarding, Agent scheduling, and task planning scheduling, and MySQL is used for data persistence; in Agent, the Dify framework is used for development, and the external interfaces of the IOT platform and the three-dimensional scene platform (i.e., the three-dimensional platform) are called to implement the aforementioned functions; in task planning, Workflow is used for development, and task plans are made based on the plans returned by LLM; in the LLM part, file management is used for model data management and fine-tuning data management, and at the same time, Unsloth is used for model fine-tuning and deployment.

[0197] like Figure 3 As shown, task scheduling management is executed in the order of requirements understanding, data source parsing, model source parsing, code writing, code checking, and code execution functions; this order is provided to LLM as a prompt, and LLM returns the specific task plan for this conversation, such as understanding the requirements to match points and configure the point type, the data source is XXX file, and it needs to be opened in XX mode, and each field attribute is X. After the task scheduling is generated, the system will mark the task scheduling every time it is executed to indicate whether the task is completed or not. The task scheduling will also manage the current situation and call back the back-end interface to return the current progress and assign the next execution task content (code generation LLM or code checking, etc.). After the task scheduling receives the task callback for the last time and completes all tasks, the system will call back and return the back-end matching task completion flag.

[0198] The embodiments of the present application have at least the following innovative features:

[0199] 1. Use LLM to match the model (i.e. the 3D point to be bound) and the data source (i.e. the data source to be bound)

[0200] There will be many points in industrial scenes. The points need to be managed in three-dimensional scenes and associated with relevant data panels. This system matches the model and data source through LLM dialogue communication. In the environment where LLM is still widely used for dialogue applications, LLM is applied to industrial point matching work.

[0201] 2. System advancement analysis

[0202] While LLM is still widely used in dialogue generation, article writing and other tasks, we have attempted to use LLM in industry, which is advanced in the current environment.

[0203] 3. Application Effect

[0204] The embodiment of the present application provides a nuclear industry three-dimensional visualization development platform. By adopting the above technical solution, the platform has the following beneficial effects:

[0205] There are many points in industrial scenes, which need to be managed in three-dimensional scenes and associated with relevant data panels. This system matches models and data sources through LLM dialogue communication, and can complete the matching with almost no human intervention after the data source and rules are given, thereby reducing personnel costs and subsequent maintenance costs.

[0206] Shorten development cycle: Through virtual verification and rapid prototyping, the development cycle is greatly shortened. Designers and engineers can quickly iterate and optimize in a digital environment, reducing the time for design rework and correction.

[0207] Enhance market competitiveness: By quickly responding to market demands and providing rapid deployment capabilities, the enterprise has enhanced its market competitiveness.

[0208] Embodiment 2

[0209] The nuclear power plant simulation operation method provided by the embodiments of the present application is applicable to the process of nuclear power plant simulation operation. This nuclear power plant simulation operation method can be executed by a nuclear power plant simulation operation system, an electronic device, etc. Hereinafter, an example will be given where this nuclear power plant simulation operation method is executed by an electronic device for illustration.

[0210] The nuclear power plant simulation operation method provided by the embodiments of the present application may include steps S210 to S220.

[0211] S210. Determine the first binding relationship according to the three-dimensional point position binding method of Embodiment 1.

[0212] S220. Simulate the operation of the nuclear power plant according to the first binding relationship.

[0213] According to the nuclear power plant simulation operation method provided by the embodiments of the present application, through the three-dimensional point position binding method of Embodiment 1, the first binding relationship can be automatically determined, thereby improving the efficiency of three-dimensional point position binding and enabling the timely simulation operation of the nuclear power plant.

[0214] For the specific implementation manner of step S210, please refer to Embodiment 1, and details will not be elaborated here.

[0215] In step S220, exemplarily, when the temperature of a certain heat pipe is abnormal, the fuel cladding temperature can be simulated and predicted whether it exceeds the limit according to the first binding relationship to achieve the simulation operation of the nuclear power plant.

[0216] Embodiment 3

[0217] As Figure 5 shown, the three-dimensional point position binding device provided by the embodiments of the present application may include:

[0218] The first acquisition module 510 is configured to acquire at least one data source to be bound and at least one three-dimensional point position to be bound in response to a first input operation of the user based on the first interface;

[0219] The first determination module 520 is connected to the first acquisition module 510 and is configured to, in response to a first determination and binding operation of the user based on the second interface, for each data source to be bound in at least one data source to be bound, determine the three-dimensional point position to be bound corresponding to the data source to be bound from at least one three-dimensional point position to be bound by using a target binding model, and bind the data source to be bound with its corresponding three-dimensional point position to be bound to obtain a first binding relationship;

[0220] Among them, the target binding model is obtained by training a large language model with a sample data source and its corresponding sample three-dimensional point positions.

[0221] In some embodiments, the device further includes:

[0222] A second acquisition module, connected to the first determination module, for acquiring a target matching rule in response to a second input operation of the user based on a third interface;

[0223] The first determination module 520 is specifically configured to:

[0224] Based on the target matching rule, use the target binding model to determine the corresponding three-dimensional point position to be bound from at least one three-dimensional point position to be bound.

[0225] In some embodiments, the device further includes:

[0226] A first extraction module, for screening at least one three-dimensional point position to be bound according to a three-dimensional point position screening rule to obtain at least one target three-dimensional point position;

[0227] A first screening module, connected to the first extraction module, for screening at least one three-dimensional point position to be bound according to a three-dimensional point position screening rule to obtain at least one target three-dimensional point position;

[0228] The first determination module 520 is further connected to the first screening module, and is specifically configured to:

[0229] Based on the target matching rule, use the target binding model to determine the target three-dimensional point position corresponding to the data source to be bound from at least one target three-dimensional point position.

[0230] In some embodiments, the device further includes:

[0231] A second extraction module, for extracting a data source screening rule from the target matching rule;

[0232] A second screening module, connected to the second extraction module, for screening at least one data source to be bound according to the data source screening rule to obtain at least one target data source;

[0233] The first determination module 520 is further connected to the second screening module, and is specifically configured to:

[0234] Based on the target matching rule, use the target binding model to determine the three-dimensional point position to be bound corresponding to the target data source from at least one three-dimensional point position to be bound.

[0235] In some embodiments, the device further includes:

[0236] The first inspection module is used to call an intelligent agent to perform a logical inspection on the configuration file to which the first binding relationship belongs, and obtain a first inspection result;

[0237] The first execution module is connected to the first inspection module and is used to call an intelligent agent to execute the first binding relationship when the first inspection result is passed.

[0238] In some embodiments, the device further includes:

[0239] The second inspection module is used to call a three-dimensional platform for inspection to obtain a second inspection result, wherein the three-dimensional platform inspects the first binding relationship based on a target matching rule;

[0240] The first display module is connected to the second inspection module and is used to visually display the first binding relationship on a fourth interface when the second inspection result is passed.

[0241] In some embodiments, the device further includes:

[0242] The first modification module is used to modify the first binding relationship in response to a first modification operation of the user based on the fourth interface;

[0243] Or,

[0244] The second modification module is used to modify the target matching rule in response to a second modification operation of the user based on the fourth interface, and obtain a modified target matching rule;

[0245] The second determination module is connected to the second modification module and is used to determine the corresponding to-be-bound three-dimensional point from at least one to-be-bound three-dimensional point based on the modified target matching rule by using a target binding model, and bind the to-be-bound data source to its corresponding to-be-bound three-dimensional point to obtain an updated first binding relationship.

[0246] In some embodiments, the device further includes:

[0247] The third determination module is used to determine the task type of the user in response to a third input operation of the user based on a fifth interface;

[0248] The first display module is used to display a first interface when the task type is a target task type.

[0249] In some embodiments, the device further includes:

[0250] The creation module is used to create a target binding model;

[0251] Specifically, the creation module is used for:

[0252] Obtain a sample data source and its corresponding sample three-dimensional point;

[0253] Training a large language model using a sample data source and its corresponding sample three-dimensional point positions to obtain a target binding model.

[0254] The three-dimensional point position binding device provided by the embodiments of the present application has the beneficial effects and implementation manners of the three-dimensional point position binding method provided by Embodiment 1 of the present application. Specifically, reference can be made to the specific description of the three-dimensional point position binding method in the above Embodiment 1, and details are not described herein again.

[0255] Embodiment 4

[0256] The embodiments of the present application further provide a nuclear power plant simulation operation system, which may include:

[0257] The three-dimensional point position binding device of Embodiment 3, configured to determine a first binding relationship;

[0258] A simulation operation module, connected to the three-dimensional point position binding device, configured to simulate the operation of a nuclear power plant according to the first binding relationship.

[0259] The nuclear power plant simulation operation system provided by the embodiments of the present application has the beneficial effects and implementation manners of the nuclear power plant simulation operation method provided by Embodiment 2 of the present application. Specifically, reference can be made to the specific description of the nuclear power plant simulation operation method in the above Embodiment 2, and details are not described herein again.

[0260] It can be understood that the above embodiments are merely exemplary embodiments adopted to illustrate the principle of the present application. However, the present application is not limited thereto. For those of ordinary skill in the art, various modifications and improvements can be made without departing from the spirit and essence of the present application, and these modifications and improvements are also regarded as the protection scope of the present application.

Claims

1. A three-dimensional point position binding method, characterized in that, Including: In response to a first input operation of the user based on a first interface, obtain at least one data source to be bound and at least one three-dimensional point to be bound; In response to a first determination and binding operation of the user based on a second interface, for each data source to be bound in the at least one data source to be bound, use a target binding model to determine a three-dimensional point to be bound corresponding to the data source to be bound from the at least one three-dimensional point to be bound, and bind the data source to be bound with its corresponding three-dimensional point to be bound to obtain a first binding relationship; Wherein, the target binding model is obtained by training a large language model with a sample data source and its corresponding sample three-dimensional point.

2. The method according to claim 1, characterized in that Before the step of responding to the first determination and binding operation of the user based on the second interface, the method further includes: In response to a second input operation of the user based on a third interface, obtain a target matching rule; The step of using the target binding model to determine the three-dimensional point to be bound corresponding to it from the at least one three-dimensional point to be bound includes: Based on the target matching rule, use the target binding model to determine the three-dimensional point to be bound corresponding to it from the at least one three-dimensional point to be bound.

3. The method according to claim 2, characterized in that, Before the step of using the target binding model to determine the three-dimensional point to be bound corresponding to the data source to be bound from the at least one three-dimensional point to be bound based on the target matching rule, the method further includes: Extract a three-dimensional point screening rule from the target matching rule; According to the three-dimensional point screening rule, screen the at least one three-dimensional point to be bound to obtain at least one target three-dimensional point; The step of using the target binding model to determine the three-dimensional point to be bound corresponding to the data source to be bound from the at least one three-dimensional point to be bound based on the target matching rule includes: Based on the target matching rule, use the target binding model to determine the target three-dimensional point corresponding to the data source to be bound from the at least one target three-dimensional point.

4. The method according to claim 2, wherein Before the step of using the target binding model to determine the three-dimensional point to be bound corresponding to the data source to be bound from the at least one three-dimensional point to be bound based on the target matching rule, the method further includes: Extract a data source screening rule from the target matching rule; According to the data source screening rule, screen the at least one data source to be bound to obtain at least one target data source; The step of using the target binding model to determine the three-dimensional point to be bound corresponding to the data source to be bound from the at least one three-dimensional point to be bound based on the target matching rule includes: Based on the target matching rule, use the target binding model to determine the three-dimensional point to be bound corresponding to the target data source from the at least one three-dimensional point to be bound.

5. The method according to claim 2, wherein After the step of binding the data source to be bound with its corresponding three-dimensional point to be bound to obtain a first binding relationship, the method further includes: Call an intelligent agent to perform a logical check on the configuration file to which the first binding relationship belongs to obtain a first check result; In the case where the first check result is passed, call the intelligent agent to execute the first binding relationship.

6. The method according to claim 5, wherein After the step of calling the intelligent agent to execute the first binding relationship, the method further includes: Invoke a 3D platform for inspection to obtain a second inspection result, where the 3D platform inspects the first binding relationship based on the target matching rule; When the second inspection result is passed, visually display the first binding relationship on the fourth interface.

7. The method according to claim 6, wherein After the first binding relationship is visually displayed, the method further includes: In response to a first modification operation by the user based on the fourth interface, modify the first binding relationship; Or, In response to a second modification operation by the user based on the fourth interface, modify the target matching rule to obtain a modified target matching rule; Based on the modified target matching rule, use the target binding model to determine the corresponding to-be-bound 3D points from the at least one to-be-bound 3D points, and bind the to-be-bound data source to its corresponding to-be-bound 3D points to obtain an updated first binding relationship.

8. The method according to claim 1, characterized in that, Before the response to the first input operation by the user based on the first interface, the method further includes: In response to a third input operation by the user based on the fifth interface, determine the user's task type; When the task type is the target task type, display the first interface.

9. The method according to claim 1, characterized in that, Before using the target binding model to determine the to-be-bound 3D points corresponding to the to-be-bound data source from the at least one to-be-bound 3D points, the method further includes: Create a target binding model; The creation of the target binding model specifically includes: Obtain sample data sources and their corresponding sample 3D points; Use the sample data sources and their corresponding sample 3D points to train a large language model to obtain the target binding model.

10. A method for simulating the operation of a nuclear power plant, characterized in that, Include: According to the 3D point binding method according to any one of claims 1 to 9, determine the first binding relationship; Simulate the operation of a nuclear power plant according to the first binding relationship.

11. A three-dimensional point position binding device, characterized in that, Include: A first acquisition module, configured to acquire at least one to-be-bound data source and at least one to-be-bound 3D point in response to a first input operation by the user based on the first interface; A first determination module, connected to the first acquisition module, configured to, in response to a first determination and binding operation by the user based on the second interface, for each to-be-bound data source in the at least one to-be-bound data source, use the target binding model to determine the to-be-bound 3D points corresponding to the to-be-bound data source from the at least one to-be-bound 3D points, and bind the to-be-bound data source to its corresponding to-be-bound 3D points to obtain a first binding relationship; Wherein, the target binding model is obtained by training a large language model using sample data sources and their corresponding sample 3D points.

12. The device according to claim 11, characterized in that, The device further includes: A second acquisition module, connected to the first determination module, configured to acquire a target matching rule in response to a second input operation by the user based on the third interface; The first determination module is specifically configured to: Based on the target matching rule, use the target binding model to determine the corresponding to-be-bound 3D points from the at least one to-be-bound 3D points.

13. The apparatus according to claim 12, wherein The device further includes: A first extraction module, configured to screen the at least one to-be-bound 3D points according to the 3D point screening rule to obtain at least one target 3D point; The first screening module, connected to the first extraction module, is configured to screen the at least one three-dimensional point to be bound according to the three-dimensional point screening rule, so as to obtain at least one target three-dimensional point; The first determination module is further connected to the first screening module, and specifically configured to: Based on the target matching rule, use the target binding model to determine the target three-dimensional point corresponding to the data source to be bound from the at least one target three-dimensional point.

14. The device according to claim 12, characterized in that, The apparatus further includes: A second extraction module, configured to extract a data source screening rule from the target matching rule; A second screening module, connected to the second extraction module, is configured to screen the at least one data source to be bound according to the data source screening rule, so as to obtain at least one target data source; The first determination module is further connected to the second screening module, and specifically configured to: Based on the target matching rule, use the target binding model to determine the three-dimensional point to be bound corresponding to the target data source from the at least one three-dimensional point to be bound.

15. A nuclear power plant simulation operation system, characterized in that, Comprising: The three-dimensional point binding apparatus according to any one of claims 11 to 14, configured to determine a first binding relationship; A simulation operation module, connected to the three-dimensional point binding apparatus, is configured to simulate the operation of a nuclear power plant according to the first binding relationship.

Images