Stress analysis method and system based on multi-physical field coupling GIS bus

Abstract

The application discloses a stress analysis method and system based on a multi-physical field coupling GIS bus, and relates to the technical field of electric power, comprising: collecting engineering structure parameters, and determining a historical typical thermal-electric-power working condition operation parameter set; identifying a parameter missing area, and determining a parameter missing area set; respectively performing stress field completion on the parameter missing area set, and obtaining a completed stress field set; constructing a driving projection completion model based on the historical typical thermal-electric-power working condition operation parameter set, the parameter missing area set and the completed stress field set; and performing stress field completion analysis on a target GIS bus under thermal-electric-power multi-physical field coupling through the driving projection completion model, and determining a target completed stress field. The application solves the technical problem that the prior art cannot accurately identify and complete the stress missing area under the multi-physical field coupling of the GIS bus, and achieves the technical effect of improving the completeness and accuracy of stress analysis.

Description

Technical Field

[0001] This invention relates to the field of power technology, and more specifically to a stress analysis method and system based on multi-physics field coupled GIS busbars. Background Technology

[0002] During the operation of a GIS busbar, thermal, electric, and structural force fields interact and interact, resulting in complex and variable engineering structural parameters and operating conditions. This often leads to missing or abnormal stress field data. Traditional simulation methods often rely on modeling a single physical field, making it difficult to reflect the true stress response characteristics under multi-field interaction. This is especially true in critical structural components such as supports, interfaces, and corrugated joints, where timely and accurate stress data is hard to obtain. In such cases, missing areas cannot be effectively identified, thus affecting the accuracy and reliability of the overall stress analysis results. Summary of the Invention

[0003] This application provides a stress analysis method and system based on multi-physics field coupled GIS busbars, which is used to address the technical problem that existing technologies cannot accurately identify and complete the stress missing areas under multi-physics field coupling of GIS busbars.

[0004] In view of the above problems, this application provides a stress analysis method and system based on multi-physics coupled GIS busbars.

[0005] A first aspect of this application provides a stress analysis method based on a multi-physics coupled GIS busbar, the method comprising:

[0006] The engineering structural parameters of the target GIS bus are collected, and a set of historical typical thermal-electrical-power operating parameters is determined based on the matching of these parameters. Missing parameter regions are identified by combining the historical typical thermal-electrical-power operating parameter set and the engineering structural parameters, thus determining a set of missing parameter regions. Stress field completion is performed on each of these missing parameter regions to obtain a complete stress field set. A driving projection completion model is constructed based on the historical typical thermal-electrical-power operating parameter set, the set of missing parameter regions, and the complete stress field set. The driving projection completion model is used to perform stress field completion analysis on the target GIS bus under thermal-electrical-power multiphysics coupling to determine the target complete stress field.

[0007] A second aspect of this application provides a stress analysis system based on a multiphysics coupled GIS busbar, the system comprising:

[0008] The system comprises the following modules: a parameter acquisition module for acquiring engineering structural parameters of the target GIS bus, and determining a set of historical typical thermal-electrical-power operating parameters based on these parameters; a region identification module for identifying missing parameter regions by combining the historical typical thermal-electrical-power operating parameter set and the engineering structural parameters; a stress field completion module for completing the stress field of each of the missing parameter regions to obtain a completed stress field set; a model construction module for constructing a driven projection completion model based on the historical typical thermal-electrical-power operating parameter set, the missing parameter region set, and the completed stress field set; and a completion analysis module for performing stress field completion analysis on the target GIS bus under thermal-electrical-power multiphysics coupling using the driven projection completion model to determine the target completed stress field.

[0009] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0010] This application collects the engineering structural parameters of the target GIS bus, and determines a set of historical typical thermal-electrical-power operating parameters based on the matching of these parameters. It then identifies missing parameter regions by combining the historical typical thermal-electrical-power operating parameter set with the engineering structural parameters, thus determining a set of missing parameter regions. Stress field completion is performed on each of these missing parameter regions to obtain a complete stress field set. A driving projection completion model is constructed based on the historical typical thermal-electrical-power operating parameter set, the set of missing parameter regions, and the complete stress field set. The driving projection completion model is used to perform stress field completion analysis on the target GIS bus under thermal-electrical-power multi-physics coupling to determine the target complete stress field. This invention solves the technical problem of existing technologies being unable to accurately identify and complete stress missing regions under multi-physics coupling of GIS buses. By constructing a driving projection completion model for stress field completion analysis, it achieves the technical effect of improving the completeness and accuracy of stress analysis. Attached Figure Description

[0011] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 A schematic diagram of the stress analysis method based on multi-physics coupled GIS bus provided in the embodiments of this application;

[0013] Figure 2 A schematic diagram of the stress analysis system based on a multi-physics field coupled GIS bus provided in this application embodiment.

[0014] Figure labeling: Parameter acquisition module 11, Region identification module 12, Stress field completion module 13, Model construction module 14, Completion analysis module 15. Detailed Implementation

[0015] This application provides a stress analysis method and system based on multi-physics field coupled GIS busbars. It addresses the technical problem that existing technologies cannot accurately identify and complete the stress missing areas under multi-physics field coupling of GIS busbars. By constructing a driving projection completion model to perform stress field completion analysis, it achieves the technical effect of improving the completeness and accuracy of stress analysis.

[0016] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0017] It should be noted that any variation of the terms "comprising" and "having" is intended to cover non-exclusive inclusion, for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or modules that are not explicitly listed or that are inherent to such processes, methods, products, or devices.

[0018] Example 1, as Figure 1 As shown, this application provides a stress analysis method based on multi-physics coupled GIS busbars, the method comprising:

[0019] Step S100: Collect the engineering structural parameters of the target GIS bus, and determine the set of historical typical thermal-electric-power operating parameters based on the engineering structural parameters.

[0020] In this embodiment, the engineering structural parameters of the target GIS busbar are first obtained by extracting key parameters such as the geometric dimensions, structural layout, connection interface form and material properties of the busbar based on the pre-stored design drawings or 3D modeling data.

[0021] Subsequently, a set of historical typical thermal-electrical-power operating parameters was determined based on matching engineering structural parameters. In this process, similar operating conditions were first retrieved from the historical GIS bus operation database using engineering structural parameters to obtain a preliminary set of thermal-electrical-power operating condition data. This set was then mapped onto a three-dimensional coordinate system with thermal, electrical, and force parameters as dimensions, constructing a multi-dimensional competition space. Within this space, clustering or filtering was performed based on data distribution characteristics to extract representative typical operating condition data, forming a set of historical typical thermal-electrical-power operating parameters.

[0022] Furthermore, in the method provided in the application embodiments, collecting the engineering structural parameters of the target GIS bus and determining the set of historical typical thermal-electrical-power operating parameters based on the matching of the engineering structural parameters further includes:

[0023] Using the engineering structure parameters as an index, the thermal-electric-power operating parameters are matched in the historical GIS bus operation database to determine the historical thermal-electric-power operating parameter set; the historical thermal-electric-power operating parameter set is input into a three-dimensional coordinate system to construct a three-dimensional competition space, wherein the three-dimensional competition space includes multiple historical competition coordinate points; typical thermal-electric-power operating parameters are filtered in the three-dimensional competition space to determine the historical typical thermal-electric-power operating parameter set.

[0024] In this embodiment, the engineering structural parameters of the target GIS busbar are first used as an index. These parameters include the busbar's external dimensions, material properties, and interface structure. By utilizing a structural feature similarity matching algorithm, such as cosine similarity calculation, the structural parameters of existing data samples in the historical GIS busbar operation database are compared. Historical samples with structural similarity exceeding a set threshold are selected to preliminarily determine a set of historical thermal-electrical-power operating parameters similar to the target structure. For example, for a section of aluminum alloy straight pipe busbar with corrugations, the matching process will focus on comparing parameters such as pipe diameter, length, and corrugation depth.

[0025] Subsequently, the historical thermal-electrical-mechanical operating parameters were mapped onto a three-dimensional coordinate system consisting of thermal, electrical, and mechanical parameters, constructing a three-dimensional competition space. In this process, a standardized normalization mapping method was employed, that is, Z-score standardization was performed on the numerical data of each type of physical field to eliminate the influence of different physical dimensions on the coordinate scale. This ensures that each sample data point becomes a historical competition coordinate point with a uniform scale, and the constructed three-dimensional competition space includes multiple historical competition coordinate points.

[0026] Finally, typical thermal-electrical-power operating parameters are screened in the three-dimensional competition space. In this step, firstly, based on a preset three-dimensional discrete approximation threshold and a preset coordinate point interval threshold, a density extraction algorithm is used to screen multiple historical competition coordinate points, extracting a set of historical typical competition coordinate points. Then, all historical competition coordinate points are grouped into the corresponding aggregation neighborhood of the historical typical competition point according to the principle of maximum similarity, forming a set of historical typical competition coordinate point neighborhoods. Finally, a centralized screening is performed within the neighborhoods of the historical typical competition coordinate point neighborhoods to determine the set of historical typical thermal-electrical-power operating parameters.

[0027] Furthermore, in the method provided in the application embodiments, the process of filtering typical thermo-electric-power operating parameters in the three-dimensional competition space to determine the set of historical typical thermo-electric-power operating parameters further includes:

[0028] Based on a preset three-dimensional discrete approximation threshold and a preset coordinate point interval threshold, historical typical competitive coordinate points are extracted from the multiple historical competitive coordinate points to determine a set of historical typical competitive coordinate points; the multiple historical competitive coordinate points are added to the neighborhood of the historical typical competitive coordinate point corresponding to the maximum value of three-dimensional aggregated similarity to determine a set of neighborhood of historical typical competitive coordinate points; and the neighborhood of the set of neighborhood of historical typical competitive coordinate points is centrally filtered to determine a set of historical typical thermal-electrical-power operating parameters.

[0029] In this embodiment, when extracting typical historical competitive coordinate points from multiple historical competitive coordinate points, a coordinate point is first randomly selected from all historical competitive coordinate points as the initial typical historical competitive coordinate point and added to the set of typical historical competitive coordinate points as the starting reference point for subsequent screening. Next, all other unselected historical competitive coordinate points are traversed, and each candidate point is compared with all selected points in the current set of typical historical competitive coordinate points, using a dual judgment based on a preset three-dimensional discrete approximation threshold and a preset coordinate point interval threshold. The cosine similarity calculation method is used to evaluate whether the candidate point and the selected typical points have similar operating condition combination trends in the three dimensions of thermal parameters, electrical parameters, and force parameters. If the similarity is lower than the set discrete approximation threshold, it indicates that their operating condition modes are independent. Simultaneously, the Euclidean distance calculation method is used to determine whether the spatial distance between candidate points in the three-dimensional parameter space is greater than the preset coordinate point interval threshold, ensuring that the distribution of the selected coordinate points has spatial differences. When both conditions are met, the candidate point is included in the set of typical historical competitive coordinate points. Through continuous iterative comparison and screening, a set of historically typical competing coordinate points that are representative in both thermal-electrical-mechanical operating condition change patterns and parameter spatial distribution was finally obtained. The preset three-dimensional discrete approximation threshold and the preset coordinate point interval threshold were pre-set by technical experts.

[0030] After constructing the set of historical typical competing coordinate points, the remaining unselected historical competing coordinate points are processed and classified using a three-dimensional aggregation similarity maximum assignment method. This method uses cosine similarity as a metric to calculate the similarity value between each unassigned point and all historical typical competing coordinate points, judging their similarity in the trends of thermal-electrical-mechanical parameter changes. Each unassigned point is assigned to the neighborhood corresponding to the historical typical competing coordinate point with the highest similarity, thereby establishing multiple assignment groups centered on typical points, forming a neighborhood set of historical typical competing coordinate points.

[0031] Finally, a neighborhood-based centralized screening is performed on the neighborhood sets of historical typical competing coordinate points. In this process, each coordinate point in the historical typical competing coordinate point set serves as a starting reference point. Following a preset centralized screening step size, the process iteratively expands within its corresponding historical typical competing coordinate point neighborhood set, gradually constructing a historical iterative competing coordinate point set. In each iteration, the centralized screening quantity of the current result is evaluated and compared with the original centralized screening quantity of that historical typical competing coordinate point. If the current centralized screening quantity reaches or exceeds the reference value, the process continues to expand to a more distant neighborhood region according to the centralized screening step size. This process terminates when the preset number of iterations is met or a stable screening state is reached. Ultimately, based on the samples obtained from the iterative screening in each neighborhood, the set of historical typical thermal-electrical-power operating parameters is determined.

[0032] Furthermore, in the method provided in the application embodiments, the method further includes performing neighborhood-based centralized screening on the neighborhood set of the historical typical competing coordinate points to determine the set of historical typical thermal-electrical-power operating parameters, and also includes:

[0033] Starting from the set of historical typical competitive coordinate points, each point is iterated within its neighborhood according to a preset set of concentrated screening steps to determine the set of historical iterative competitive coordinate points. It is then determined whether the concentrated screening amount of each historical iterative competitive coordinate point in the set of historical iterative competitive coordinate points is greater than or equal to the concentrated screening amount of the corresponding historical typical competitive coordinate point in the set of historical typical competitive coordinate points. If so, the set of historical iterative competitive coordinate points is iterated according to the preset set of concentrated screening steps until the preset number of iterations is met, thus determining the set of historical typical thermal-electrical-power operating parameters.

[0034] In this embodiment, during the process of performing neighborhood-based centralized screening on the neighborhood set of historical typical competitive coordinate points, each historical typical competitive coordinate point in the neighborhood set is taken as the starting point. Within its corresponding neighborhood set, iterative processing is performed according to a preset centralized screening step size. Using the Euclidean distance calculation method, all historical competitive coordinate points in the current neighborhood set whose distance from the historical typical competitive coordinate point is less than or equal to the preset centralized screening step size are selected. From these, the historical competitive coordinate point farthest from the historical typical competitive coordinate point is selected as the historical iterative competitive coordinate point for this round of iteration.

[0035] Subsequently, based on Euclidean distance, within the same neighborhood set of typical historical competitive coordinate points, taking both the historical iterative competitive coordinate point and the original typical historical competitive coordinate point as centers, the number of historical competitive coordinate points whose distance is less than or equal to one-third of the preset set-wide selection step size is counted, serving as their respective set-wide selection values. When the set-wide selection value of a historical iterative competitive coordinate point is greater than or equal to the set-wide selection value of its corresponding typical historical competitive coordinate point, it is considered to have better clustering characteristics and working condition representativeness in the neighborhood, and is thus used to replace the current reference point, continuing the next iteration. Each iteration is repeated within the same neighborhood set, including selecting the farthest point, calculating the set-wide selection value, comparing and determining replacement, until the preset number of iterations is met. When the set-wide selection value of a historical iterative competitive coordinate point is less than the set-wide selection value of its corresponding typical historical competitive coordinate point, it indicates that the iteration no longer improves the clustering effect, and this is considered a convergence state, immediately ending the iteration process of the current neighborhood set.

[0036] Finally, the historical iteration competition coordinate point at the end of each iteration in each neighborhood is taken as the optimal representative point of that neighborhood, and the corresponding thermal, electrical, and force parameter operation data are extracted. The set of three-dimensional operation parameter data corresponding to all final historical iteration competition coordinate points constitutes the set of historical typical thermal-electrical-force operating parameters.

[0037] Step S200: Identify missing parameter regions by combining the historical typical thermal-electric-power operating parameter set and engineering structural parameters, and determine the set of missing parameter regions.

[0038] In this embodiment, when identifying missing parameter regions, the historical typical thermo-electric-mechanical operating parameter set and engineering structural parameters are first input into the simulation software to establish a multi-physics coupling model of thermal, electric, and force fields, generating a simulated stress field set. This simulated stress field set is then traversed and analyzed. By identifying numerical anomalies, discontinuous distributions, or missing response regions, the location set of anomaly regions is extracted. The type of missing region is determined based on structural characteristics, generating a corresponding missing type label set. Finally, each anomaly region location is bound one-to-one with its corresponding missing type label to determine the complete set of missing parameter regions.

[0039] Furthermore, the method provided in the application embodiment, which combines the historical typical thermal-electric-power operating parameter set and engineering structural parameters to identify parameter missing regions and determine the parameter missing region set, also includes:

[0040] The historical typical thermal-electric-power operating parameters set and the engineering structure parameters are input into the simulation software for multiphysics coupling to determine the simulated stress field set; the simulated stress field set is traversed to identify abnormal regions, and the abnormal region location set and the missing type label set are determined; the abnormal region location set and the missing type label set are bound one-to-one to determine the parameter missing region set.

[0041] In this embodiment, historical typical thermal-electric-mechanical operating parameters and engineering structural parameters are first input into multiphysics simulation software, such as ANSYS Multiphysics, to construct a three-dimensional finite element model coupling thermal, electric, and force fields. During this modeling process, the historical typical thermal-electric-mechanical operating parameters are set as boundary conditions and load inputs, including thermal parameters (such as ambient temperature and conductor heat generation), electrical parameters (such as operating voltage and breakdown voltage), and force parameters (such as internal pressure and thermal stress). The engineering structural parameters are used to define the solid geometry, material properties, and key structural features of the busbar, including specific modeling requirements for busbar dimensions, flange connection areas, weld locations, corrugated joint shapes, and supporting base locations. A high-precision mesh is generated using the finite element modeling tool, and after applying coupled loads, the solution is executed, outputting nodal stress response data under the combined action of multiple physics fields, forming a simulated stress field set.

[0042] Subsequently, the simulated stress field set is traversed to identify abnormal regions, determining the set of abnormal region locations and the set of missing type labels. Specifically, in identifying abnormal regions in the simulated stress field, a training dataset is first constructed, obtaining multiple sample sets of simulated stress fields and their corresponding sets of abnormal region locations and missing type labels as positive sample sets. Then, the abnormal locations and labels in these samples are randomly perturbed to generate inconsistent pairings, forming a negative sample set. Based on the positive and negative sample sets, a Support Vector Machine (SVM) algorithm is used for supervised learning to train an abnormal region identifyer with discriminative capabilities. Finally, this identifyer is applied to the current simulated stress field set to analyze the stress distribution and identify the set of abnormal region locations and their corresponding sets of missing type labels.

[0043] Finally, the set of abnormal region locations is bound one-to-one with the set of missing type labels to form a structured dataset, and the output is a set of parameter missing regions.

[0044] Furthermore, in the method provided in the application embodiments, the process of traversing the simulated stress field set to identify abnormal regions and determining the set of abnormal region locations and the set of missing type labels also includes:

[0045] A positive sample set is obtained by acquiring multiple sets of simulated stress fields, multiple sets of abnormal region locations, and multiple sets of missing type labels. The sets of abnormal region locations and missing type labels are then randomly modified, and the results of these random modifications, along with the sets of simulated stress fields, are used as a negative sample set. A support vector machine is trained under supervision using the positive and negative sample sets to construct an abnormal region identifier. The abnormal region identifier is then used to identify abnormal regions in the simulated stress field set, determining the sets of abnormal region locations and missing type labels.

[0046] In this embodiment, multiple sets of simulated stress fields, multiple sets of abnormal region locations, and multiple sets of missing type labels are first obtained from a historical database as the positive sample set required for training. The multiple sets of simulated stress fields originate from stress response data generated by multiphysics simulation software under different GIS busbar structures and thermal-electrical-mechanical conditions, including node numbers, node coordinates, thermal stress, electrical stress, and structural stress. The multiple sets of abnormal region locations record the locations of stress data anomalies identified in the simulated stress fields, including anomaly node or unit numbers, spatial coordinates, structural component (e.g., flange, weld, corrugated joint), boundary condition information, and anomaly region extent. The multiple sets of missing type labels provide a labeling of the missing cause for each anomaly region, using a standardized classification system. Labels include, but are not limited to, geometric occlusion, mesh degradation, and measurement point blind spots, and are accompanied by label numbers.

[0047] Subsequently, multiple sets of sample anomaly location locations and multiple sets of sample missing type labels are randomly modified. In this process, firstly, the sets of sample anomaly location locations are perturbed by adding random spatial offsets (such as uniform or Gaussian perturbations) to the original anomaly points, causing the perturbed locations to deviate from the true anomaly regions but remain within the effective area of ​​the simulation model. This operation simulates non-real anomaly regions, constructing pseudo-anomaly spatial locations. Secondly, the sets of multiple samples missing type labels are perturbed by scrambling the labels through label mismatch or label swapping operations, randomly combining the originally correctly paired anomaly locations with the missing type labels, thus forming incorrect anomaly cause labels. The perturbed sets of anomaly location locations and perturbed sets of missing type labels are then recombined with the original sets of multiple sample simulated stress fields to generate logically invalid anomaly data pairs, which serve as the negative sample set.

[0048] Subsequently, a support vector classification model is trained using positive and negative sample sets. During training, multiple feature parameters are extracted from each node or unit in the simulated stress field, including thermal stress, electrical stress, structural stress, stress gradient, boundary state, and spatial proximity, etc., to construct an input feature set. Based on the labeled locations of anomalous regions and their corresponding missing type labels, the model is trained to learn to identify anomalous regions in the stress field and output the corresponding anomalous cause type. Through this process, the anomalous region identification model is completed.

[0049] Finally, an anomaly region identifier is used to identify anomalies in the simulated stress field set. The anomaly region identifier performs feature classification on each simulation unit, automatically outputting whether it is an anomaly region, and assigns a corresponding missing type label to the anomaly region based on the model learning results. Ultimately, an identification result containing a set of anomaly region locations and a set of missing type labels is generated, constituting a parameter missing region set.

[0050] Step S300: Perform stress field completion on the set of regions with missing parameters to obtain a complete stress field set.

[0051] In this embodiment, when performing stress field completion on a set of missing parameter regions, firstly, based on the missing type label corresponding to each missing region, an applicable completion method model is matched to form a set of matching completion method models. Then, multiple completion indicators related to completion are extracted from the simulated stress field set where the missing region is located, including adjacent stress values, gradient changes, boundary conditions, etc., as completion input features, and respectively input into the corresponding completion method model to predict the stress value results for each missing region. Finally, the obtained stress prediction results are filled into the missing positions of the original simulated stress field to complete the overall stress completion and form a complete completed stress field set.

[0052] Furthermore, in the method provided in the application embodiments, stress field completion is performed on the set of missing parameter regions to obtain a completed stress field set, which further includes:

[0053] Extract the missing type labels from the set of missing parameter regions and perform model matching for completion methods to determine the set of matching completion methods; according to the preset completion index, extract the set of completion indexes for the simulation stress field set corresponding to the set of missing parameter regions and input them into the corresponding set of matching completion methods to obtain the set of stress prediction results for the missing parameter regions; write the set of stress prediction results for the missing parameter regions into the corresponding position in the simulation stress field set to obtain the completed stress field set.

[0054] Furthermore, the method provided in the application embodiments also includes:

[0055] The preset completion indicators are the location of the abnormal region, the local topological information of the busbar structure, and the stress gradient of the neighborhood.

[0056] In this embodiment, during stress field completion of a set of missing parameter regions, the corresponding missing type label is first extracted for each missing region in the set. Based on the matching rules between this label and a preset completion model, the applicable completion model set is determined. Specifically, when the missing type label indicates geometric occlusion or measurement point blind spots caused by structural features such as flange slots, bellows cavities, weld blind spots, or support bottom surface obstruction, a spatial interpolation model is used. When the missing type label indicates that the region suffers from mesh degradation simulation failure due to mesh distortion, element degradation, or discontinuous boundary connections, a graph structure reconstruction model is used. If the missing type label indicates that the missing region is located at a thermo-electrical-mechanical interface with significant boundary disturbance, belonging to a multi-physics coupled boundary interference type of missing region, a multi-condition boundary regression model is used. The spatial interpolation model, graph structure reconstruction model, and multi-condition boundary regression model are all trained based on missing region samples labeled in a historical simulation database. Among them, the training data for the spatial interpolation model comes from stress distribution data without missing nodes in a large number of complete simulation samples. The input is three-dimensional spatial coordinates, and the output is the nodal stress value. The training data for the graph structure reconstruction model includes areas in the busbar finite element model where there is simulation distortion. By constructing a local topology graph (node ​​connection relationship, element attribute, material information) and using known stress values ​​as input, the propagation law of stress on the graph structure is learned. The training samples for the multi-condition boundary regression model come from simulation data under multiple combinations of heat source, charge, and mechanical load conditions. The model input is boundary condition features and local structural attributes, and the output is the stress response value of the target area.

[0057] After completing the model matching for the missing parameter region set, the corresponding region in the simulated stress field set is extracted according to the preset missing parameter feature structure, forming a set of missing parameter indices. Specifically, based on the location identifier of each missing node or element in the missing parameter region set, the location of its abnormal region is retrieved from the finite element model data, including spatial coordinates and mesh number, for precise location of the missing parameter target. Then, the structural range of the missing region is analyzed, and the local topological information of the busbar structure in that region is extracted from the mesh file, including the connection relationships between nodes, element types (e.g., 3D solid elements, shell elements), material properties (e.g., elastic modulus, Poisson's ratio), boundary conditions (e.g., fixed constraints, load action), etc., to characterize the structural characteristics and boundary environment of the region. Finally, based on the surrounding adjacent elements or neighboring nodes of the missing node or element, its neighborhood stress gradient is calculated, i.e., the magnitude and direction of stress value changes among adjacent effective nodes. This operation extracts the stress values ​​of surrounding nodes by setting a gradient radius range, and then calculates the ratio between the relative position and the stress value difference to characterize the stress change trend. After performing the above steps, the location of the abnormal region, local topological information and neighborhood stress gradient extracted from each missing region are integrated into a structured input to form a set of completion indicators for that region.

[0058] The set of completion indicators is then input into the set of corresponding completion method models obtained from the previous matching, and the model is called to make predictions, thereby obtaining the set of parameter missing region stress prediction results for each missing region.

[0059] Finally, based on the node number or spatial location index, the stress values ​​obtained from the stress prediction result set for the region with missing parameters are written to the corresponding missing positions in the simulation stress field set, completing the completion process for that region. The entire set of regions with missing parameters is traversed, and predictions and completions are performed one by one, ultimately forming a complete stress field set.

[0060] Step S400: Construct a driving projection completion model based on the set of historical typical thermo-electric-mechanical operating parameters, the set of parameter missing regions, and the set of completed stress fields.

[0061] In this embodiment, when constructing the driving projection completion model, the set of historical typical thermo-electric-mechanical operating parameters and the set of parameter missing regions are used as input features, and the corresponding set of completed stress fields is used as the supervised output. A feedforward neural network is used as the modeling framework for end-to-end supervised training. By continuously optimizing the network weights, the model can accurately capture the mapping relationship between the operating condition driving and stress completion processes, ultimately obtaining a driving projection completion model with generalization ability under multiphysics conditions.

[0062] Furthermore, the method provided in the application embodiments also includes:

[0063] Using the set of historical typical thermo-electric-mechanical operating parameters and the set of parameter missing regions as input, and the completed stress field set as the mapping output, the framework constructed based on the feedforward neural network is trained under supervision until the training converges, thereby obtaining the driving projection completion model.

[0064] In this embodiment, firstly, each set of operating parameters in the historical typical thermo-electric-force operating parameter set is mapped one-to-one with its corresponding set of missing parameter regions. Each set of data is then jointly encoded to form a high-dimensional input vector containing features such as thermo-electric-force boundary conditions, three-dimensional coordinates of the missing region, and local structural connection information. Simultaneously, the stress values ​​of the target region are extracted from the completed stress field set generated during the completion process under this operating condition, serving as a monitoring label.

[0065] Subsequently, a feedforward neural network structure with multiple hidden layers was constructed. The input layer of this neural network receives high-dimensional joint features, the hidden layers learn the relationship between complex working conditions and stress response through nonlinear activation functions (such as ReLU), and the output layer corresponds to the stress completion result for missing regions. During training, mean squared error is used as the loss function, and the network weights are continuously adjusted through backpropagation to iteratively optimize model performance.

[0066] When the network loss value stabilizes after multiple rounds of training, and the output results are highly consistent with the historical stress distribution of the completion process, it indicates that the model has achieved a stable fit. At this point, the training converges, and a driving projection completion model with generalization ability is finally obtained.

[0067] Step S500: Perform stress field completion analysis on the target GIS bus under thermal-electrical-mechanical multiphysics coupling using the driving projection completion model to determine the target completion stress field.

[0068] In this embodiment, the engineering structural parameters of the target GIS bus and the thermal-electrical-mechanical operating parameters under the current working conditions are first collected. Combined with the set of missing parameter regions identified during the modeling process, the required input information is jointly encoded to form an input feature vector containing thermal, mechanical, and electrical boundary conditions and structural missing information.

[0069] The feature vector is then input into the trained driving projection completion model for processing, obtaining the predicted stress values ​​for each missing parameter region in the target GIS bus under the current thermo-electric-mechanical coupling condition. Based on the location indices of these predictions, they are then written into the corresponding missing regions in the original simulated stress field, completing the reconstruction of the overall stress distribution of the target GIS bus. This process completes the stress completion of the target GIS bus and determines the target completed stress field.

[0070] In summary, the embodiments of this application have at least the following technical effects:

[0071] This application collects the engineering structural parameters of the target GIS bus, and determines a set of historical typical thermal-electrical-power operating parameters based on the matching of these parameters. It then identifies missing parameter regions by combining the historical typical thermal-electrical-power operating parameter set with the engineering structural parameters, thus determining a set of missing parameter regions. Stress field completion is performed on each of these missing parameter regions to obtain a complete stress field set. A driving projection completion model is constructed based on the historical typical thermal-electrical-power operating parameter set, the set of missing parameter regions, and the complete stress field set. The driving projection completion model is used to perform stress field completion analysis on the target GIS bus under thermal-electrical-power multi-physics coupling to determine the target complete stress field. This invention solves the technical problem of existing technologies being unable to accurately identify and complete stress missing regions under multi-physics coupling of GIS buses. By constructing a driving projection completion model for stress field completion analysis, it achieves the technical effect of improving the completeness and accuracy of stress analysis.

[0072] Example 2, based on the same inventive concept as the stress analysis method for multi-physics coupled GIS busbars in the previous examples, such as... Figure 2 As shown, this application provides a stress analysis system based on multi-physics coupled GIS busbars. The system and method embodiments in this application are based on the same inventive concept. The system includes:

[0073] The parameter acquisition module 11 is used to acquire the engineering structural parameters of the target GIS bus, and determine the set of historical typical thermal-electric-power operating parameters based on the matching of the engineering structural parameters; the region identification module 12 is used to identify the parameter missing regions by combining the set of historical typical thermal-electric-power operating parameters and the engineering structural parameters, and determine the set of parameter missing regions; the stress field completion module 13 is used to complete the stress field of the set of parameter missing regions respectively, and obtain the completed stress field set; the model construction module 14 is used to construct a driven projection completion model based on the set of historical typical thermal-electric-power operating parameters, the set of parameter missing regions and the completed stress field set; the completion analysis module 15 is used to perform stress field completion analysis under thermal-electric-power multi-physics coupling on the target GIS bus through the driven projection completion model, and determine the target completed stress field.

[0074] Furthermore, the system is also used to implement the following functions:

[0075] Using the engineering structure parameters as an index, the thermal-electric-power operating parameters are matched in the historical GIS bus operation database to determine the historical thermal-electric-power operating parameter set; the historical thermal-electric-power operating parameter set is input into a three-dimensional coordinate system to construct a three-dimensional competition space, wherein the three-dimensional competition space includes multiple historical competition coordinate points; typical thermal-electric-power operating parameters are filtered in the three-dimensional competition space to determine the historical typical thermal-electric-power operating parameter set.

[0076] Furthermore, the system is also used to implement the following functions:

[0077] Based on a preset three-dimensional discrete approximation threshold and a preset coordinate point interval threshold, historical typical competitive coordinate points are extracted from the multiple historical competitive coordinate points to determine a set of historical typical competitive coordinate points; the multiple historical competitive coordinate points are added to the neighborhood of the historical typical competitive coordinate point corresponding to the maximum value of three-dimensional aggregated similarity to determine a set of neighborhood of historical typical competitive coordinate points; and the neighborhood of the set of neighborhood of historical typical competitive coordinate points is centrally filtered to determine a set of historical typical thermal-electrical-power operating parameters.

[0078] Furthermore, the system is also used to implement the following functions:

[0079] Starting from the set of historical typical competitive coordinate points, each point is iterated within its neighborhood according to a preset set of concentrated screening steps to determine the set of historical iterative competitive coordinate points. It is then determined whether the concentrated screening amount of each historical iterative competitive coordinate point in the set of historical iterative competitive coordinate points is greater than or equal to the concentrated screening amount of the corresponding historical typical competitive coordinate point in the set of historical typical competitive coordinate points. If so, the set of historical iterative competitive coordinate points is iterated according to the preset set of concentrated screening steps until the preset number of iterations is met, thus determining the set of historical typical thermal-electrical-power operating parameters.

[0080] Furthermore, the system is also used to implement the following functions:

[0081] The historical typical thermal-electric-power operating parameters set and the engineering structure parameters are input into the simulation software for multiphysics coupling to determine the simulated stress field set; the simulated stress field set is traversed to identify abnormal regions, and the abnormal region location set and the missing type label set are determined; the abnormal region location set and the missing type label set are bound one-to-one to determine the parameter missing region set.

[0082] Furthermore, the system is also used to implement the following functions:

[0083] A positive sample set is obtained by acquiring multiple sets of simulated stress fields, multiple sets of abnormal region locations, and multiple sets of missing type labels. The sets of abnormal region locations and missing type labels are then randomly modified, and the results of these random modifications, along with the sets of simulated stress fields, are used as a negative sample set. A support vector machine is trained under supervision using the positive and negative sample sets to construct an abnormal region identifier. The abnormal region identifier is then used to identify abnormal regions in the simulated stress field set, determining the sets of abnormal region locations and missing type labels.

[0084] Furthermore, the system is also used to implement the following functions:

[0085] Extract the missing type labels from the set of missing parameter regions and perform model matching for completion methods to determine the set of matching completion methods; according to the preset completion index, extract the set of completion indexes for the simulation stress field set corresponding to the set of missing parameter regions and input them into the corresponding set of matching completion methods to obtain the set of stress prediction results for the missing parameter regions; write the set of stress prediction results for the missing parameter regions into the corresponding position in the simulation stress field set to obtain the completed stress field set.

[0086] Furthermore, the system is also used to implement the following functions:

[0087] The preset completion indicators are the location of the abnormal region, the local topological information of the busbar structure, and the stress gradient of the neighborhood.

[0088] Furthermore, the system is also used to implement the following functions:

[0089] Using the set of historical typical thermo-electric-mechanical operating parameters and the set of parameter missing regions as input, and the completed stress field set as the mapping output, the framework constructed based on the feedforward neural network is trained under supervision until the training converges, thereby obtaining the driving projection completion model.

[0090] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0091] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0092] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and modifications fall within the scope of this application and its equivalents, this application intends to include such modifications and modifications.

Claims

1. A stress analysis method based on multiphysics coupled GIS busbars, characterized in that, The method includes: Collect the engineering structural parameters of the target GIS busbar, and determine the set of historical typical thermal-electric-power operating parameters based on the matching of the engineering structural parameters; By combining the historical typical thermal-electric-power operating parameter set and engineering structural parameters, parameter missing regions are identified, and the parameter missing region set is determined. Stress fields are completed for the set of regions with missing parameters to obtain a complete set of stress fields. A driving projection completion model is constructed based on the set of historical typical thermo-electric-mechanical operating parameters, the set of parameter missing regions, and the set of completed stress fields. The target GIS busbar is subjected to stress field completion analysis under thermo-electric-mechanical multiphysics coupling through the driving projection completion model to determine the target completion stress field. By combining the historical typical thermal-electric-power operating parameter set and engineering structural parameters, parameter missing regions are identified, and the set of parameter missing regions is determined, including: The set of historical typical thermo-electric-power operating parameters and the engineering structure parameters are input into the simulation software to perform multiphysics coupling and determine the set of simulated stress fields. The simulated stress field set is traversed to identify abnormal regions, and the set of abnormal region locations and the set of missing type labels are determined. The abnormal region location set and the missing type label set are bound one-to-one to determine the parameter missing region set; The simulated stress field set is traversed to identify abnormal regions, determining the set of abnormal region locations and the set of missing type labels, including: Multiple sets of simulated stress fields, multiple sets of abnormal region locations, and multiple sets of missing type labels for samples are obtained as positive sample sets. The set of multiple abnormal region locations and the set of multiple missing type labels of samples are randomly varied, and the result of the random variation and the set of multiple simulated stress fields of samples are used as the negative sample set. The support vector machine is trained under supervision using the positive and negative sample sets to construct an anomaly region identifier. The abnormal region identifier is used to identify abnormal regions in the simulated stress field set, and to determine the set of abnormal region locations and the set of missing type labels.

2. The stress analysis method based on multiphysics coupled GIS busbar as described in claim 1, characterized in that, Collect the engineering structural parameters of the target GIS busbar, and determine the set of historical typical thermal-electrical-power operating parameters based on the matching of the engineering structural parameters, including: Using the aforementioned engineering structural parameters as an index, the thermal-electric-power operating parameters are matched in the historical GIS bus operation database to determine the set of historical thermal-electric-power operating parameters; The historical thermal-electric-power operating parameters are input into a three-dimensional coordinate system to construct a three-dimensional competition space, wherein the three-dimensional competition space includes multiple historical competition coordinate points; Typical thermal-electric-power operating parameters are screened in the three-dimensional competitive space to determine the set of historical typical thermal-electric-power operating parameters.

3. The stress analysis method based on multi-physics coupled GIS busbar as described in claim 2, characterized in that, Typical thermal-electric-power operating parameters are screened in the three-dimensional competitive space to determine the set of historical typical thermal-electric-power operating parameters, including: Based on a preset three-dimensional discrete approximation threshold and a preset coordinate point interval threshold, the multiple historical competition coordinate points are extracted to determine a set of historical typical competition coordinate points. Each of the multiple historical competing coordinate points is added to the neighborhood of the historical typical competing coordinate point corresponding to the maximum value of the three-dimensional aggregated similarity, and the neighborhood set of the historical typical competing coordinate point is determined. Each of the historical typical competitive coordinate point neighborhood sets is subjected to centralized screening within the neighborhood to determine the historical typical thermal-electrical-power operating parameter set.

4. The stress analysis method based on multiphysics coupled GIS busbar as described in claim 3, characterized in that, Each of the historical typical competing coordinate point neighborhood sets is subjected to centralized screening within the neighborhood to determine the historical typical thermal-electrical-power operating parameter set, including: Starting from the set of historical typical competitive coordinate points, the system iterates within the neighborhood set of the corresponding historical typical competitive coordinate points according to a preset set of screening steps to determine the set of historical iterative competitive coordinate points. Each historical iteration competition coordinate point in the historical iteration competition coordinate point set is determined to have a set of historical typical competition coordinate points. If so, the historical iteration competition coordinate point set is iterated according to a preset set of centralized screening steps until the preset number of iterations is met, thereby determining the set of historical typical thermal-electrical-power operating parameters.

5. The stress analysis method based on multi-physics coupled GIS busbar as described in claim 1, characterized in that, Stress field completion is performed on the set of regions with missing parameters to obtain a complete stress field set, including: Extract the missing type labels from the set of missing parameter regions and perform model matching for completion methods to determine the set of matching completion method models; According to the preset completion index, the completion index set of the simulation stress field set corresponding to the parameter missing region set is extracted and input into the corresponding matching completion method model set to obtain the stress prediction result set of the parameter missing region. Write the set of stress prediction results for the region with missing parameters into the corresponding position in the simulated stress field set to obtain the complete stress field set.

6. The stress analysis method based on multiphysics coupled GIS busbar as described in claim 5, characterized in that, The preset completion indicators are the location of the abnormal region, the local topological information of the busbar structure, and the stress gradient of the neighborhood.

7. The stress analysis method based on multiphysics coupled GIS busbar as described in claim 1, characterized in that, Using the set of historical typical thermo-electric-mechanical operating parameters and the set of parameter missing regions as input, and the completed stress field set as the mapping output, the framework constructed based on the feedforward neural network is trained under supervision until the training converges, thereby obtaining the driving projection completion model.

8. A stress analysis system based on multi-physics coupled GIS busbars, characterized in that, The system is used to execute the stress analysis method based on multi-physics coupled GIS bus as described in any one of claims 1-7, and the system includes: The parameter acquisition module is used to collect the engineering structural parameters of the target GIS busbar and determine the set of historical typical thermal-electric-power operating parameters based on the engineering structural parameters. The region identification module is used to identify the parameter missing regions by combining the set of historical typical thermal-electric-power operating parameters and engineering structural parameters, and to determine the set of parameter missing regions. The stress field completion module is used to complete the stress field for the set of missing parameter regions to obtain a completed stress field set. The model building module is used to build a driven projection completion model based on the set of historical typical thermal-electric-mechanical operating parameters, the set of parameter missing regions, and the set of completed stress fields. The completion analysis module is used to perform stress field completion analysis on the target GIS bus under the coupling of thermal, electrical and mechanical multiphysics fields through the driving projection completion model, and to determine the target completion stress field.

Images