Oil immersed transformer bushing temperature rise simulation data processing method and system

Abstract

The invention relates to the technical field of electric data processing, in particular to an oil immersed transformer bushing temperature rise simulation data processing method and system, and the method comprises the steps: obtaining a plurality of standardized operation parameters in a transformer simulation model, carrying out the segmentation processing of any operation parameter, obtaining a plurality of representative points, and calculating the local density of the representative points; selecting a first representative point and a second representative point, taking a negative correlation value of the distance between the target point and the first representative point and the distance between the target point and the second representative point as a weight, and performing weighted summation on the local density to obtain the target density of the target point; and determining a local data range, obtaining a preliminary neighborhood data set of the target point according to the local data range, and executing a DBSCAN clustering algorithm based on the distance between the target point and the preliminary neighborhood data set. According to the method, global density is converted into local representative point density calculation, and the calculation amount of a traditional DBSCAN algorithm is effectively reduced.

Description

Technical Field

[0001] This invention relates to the field of electrical data processing technology, and in particular to a method and system for simulating temperature rise data of oil-immersed transformer bushings. Background Technology

[0002] As a core piece of equipment in power systems, oil-immersed transformers are evolving towards higher voltage, larger capacity, and intelligent assessment. During the operation and maintenance of oil-immersed transformers, the temperature rise distribution of the bushing and its riser area plays a crucial role as a key indicator for assessing insulation condition and operational safety. Simulation models, obtained through finite element analysis, capture multi-dimensional spatial node information such as temperature, stress, and pressure in the bushing and its surrounding area. This extensive simulation data is used to characterize local hotspots and evaluate structural reliability.

[0003] However, processing simulation data on temperature rise of oil-immersed transformer bushings remains a challenge in the industry. Considering the spatial heterogeneity and multi-field coupling characteristics of the simulation data, especially in geometrically complex areas such as the welded joints of the riser, where the temperature rise gradient varies significantly, traditional processing methods struggle to obtain characteristic values ​​reflecting physical properties. Furthermore, accurate data processing is crucial for improving heat dissipation design and enhancing the accuracy of fault prediction. Therefore, researching a method for processing simulation data on temperature rise of oil-immersed transformer bushings is of great significance for improving the operation and maintenance efficiency of power equipment.

[0004] In related technologies, the DBSCAN algorithm is commonly used to process simulation data of oil-immersed transformers. This algorithm identifies clusters by calculating the Euclidean distance between simulation nodes, thereby dividing the temperature rise hotspot regions. However, due to the large node size and high dimensionality of bushing temperature rise simulation data, the DBSCAN algorithm needs to calculate the Euclidean distance between every two simulation nodes, resulting in an increase in computational load as the data size increases, increasing the consumption of computing resources and making it difficult to cope with the rapid assessment and dynamic adjustment of temperature rise status. At the same time, the calculation of local density by existing algorithms is affected by global parameters, making it difficult to distinguish between physical hotspots and fluctuations in simulation calculations, thus limiting its application in refined transformer simulation scenarios. Summary of the Invention

[0005] To address the technical problems mentioned above regarding insufficient identification accuracy in processing simulation data of temperature rise in oil-immersed transformer bushings, where the full calculation of Euclidean distance leads to high computational resource consumption and the local density calculation makes it difficult to distinguish between actual temperature rise hotspots and simulation fluctuations, this invention provides solutions in the following aspects.

[0006] In a first aspect, the present invention provides a method for processing simulation data of temperature rise in the bushing of an oil-immersed transformer, the method comprising the steps of:

[0007] Simulation data at several moments in the transformer simulation process is acquired and preprocessed to obtain standardized simulation data. The simulation data at each moment includes multiple operating parameters of the transformer simulation. Any one of the multiple operating parameters is designated as the current operating parameter. The current operating parameter is segmented, and each segment point and the two endpoints of the current operating parameter are designated as representative points. The local density of the representative points is obtained based on the distribution of their neighborhood data. Taking the current moment of the current operating parameter as the target point, the largest representative point whose value is less than the target point is selected as the first representative point, and the largest representative point whose value is greater than the target point is selected as the second representative point. The smallest representative point of the target point is used as the second representative point. The negative correlation between the distances of the target point and the first and second representative points is used as weights to sum the local densities of the representative points corresponding to the first and second representative points to obtain the target density of the target point. The local data range of the target point is obtained according to the ratio of the neighborhood radius to the target density of the target point. The neighborhood data points of the target point are determined according to the local data range. The neighborhood data points of the target point for multiple operating parameters are calculated and merged to form the preliminary neighborhood dataset of the target point. Based on the distance between the target point and its preliminary neighborhood data points, the DBSCAN clustering algorithm is executed.

[0008] This invention lays a data foundation for subsequent clustering analysis by acquiring multiple operating parameters of the transformer simulation process and performing standardized preprocessing. To cope with the computational pressure brought by large-scale simulation data, this invention extracts representative points through segmented processing and calculates their local density. The target density of the target point is evaluated by weighted summation of the local densities of the representative points, and then the local data range of the target point is dynamically adjusted by the ratio of the neighborhood radius to the target density. This mechanism enables the local data range to be shrunk in densely distributed areas of simulation data to reduce redundant calculations, and the range to be appropriately expanded in sparsely distributed areas to ensure the integrity of the search, thereby selecting an initial neighborhood dataset. By transforming the global search into density-based local filtering, the distance calculation amount when executing the clustering algorithm is reduced. While maintaining the accuracy of identifying hot spots in bushing temperature rise, the efficiency of simulation data processing is improved, and the computational resource occupation pressure brought by high-dimensional transformer operating parameters is alleviated.

[0009] Preferably, segmenting the current item's operating parameters includes: recording the ratio of the absolute difference between the maximum and minimum values ​​of the current item's operating parameters to a preset neighborhood radius as the segment number; and segmenting the current item's operating parameters into equal intervals according to the segment number.

[0010] This invention uses the ratio of the range of the current running parameter to the preset neighborhood radius as the basis for segmentation, thereby achieving dynamic matching between the number of representative points and the distribution characteristics of simulation data. This ensures that the representative points after equidistant segmentation can uniformly and completely cover the physical value space of the running parameter, improving the characterization of subsequent density assessment.

[0011] Preferably, the local density of the representative point satisfies the following relationship: ; in, It is the first The first item in the running parameters Local density of a representative point; It is the first The running parameters of item number 1 and 2 are related to the first item. The number of parameters whose distance between representative points is less than the preset neighborhood radius; It is the first The first item in the running parameters Local standard deviation of data values ​​within the neighborhood of a representative point; It is a preset smoothing constant; It is the preset scaling factor for the preset neighborhood radius; It is a natural exponential function.

[0012] This invention introduces local standard deviation and natural exponential function into local density calculation, and uses the fluctuation characteristics of simulation values ​​as stability weights. This allows the local density of representative points to not only reflect the degree of data aggregation, but also to reduce the weight of areas with large fluctuations in local simulation values, thereby reducing the impact of random interference in simulation calculations on density evaluation results.

[0013] Preferably, when the representative point is a segmentation point of the current item's running parameters, the preset scaling factor is equal to twice the preset neighborhood radius.

[0014] Preferably, when the representative points are the two endpoints of the current item's running parameters, the preset scaling factor is equal to the preset neighborhood radius.

[0015] Preferably, the step of using the negative correlation value of the distance between the target point and the first representative point and the second representative point as weights includes: calculating the sum of the negative exponential function of the absolute difference between the target point and the second representative point and the negative exponential function of the absolute difference between the target point and the first representative point, denoted as the bipolar distance exponential sum; the weight of the local density of the representative point corresponding to the first representative point is the ratio of the negative exponential function of the absolute difference between the target point and the first representative point to the bipolar distance exponential sum; the weight of the local density of the representative point corresponding to the second representative point is the ratio of the negative exponential function of the absolute difference between the target point and the second representative point to the bipolar distance exponential sum.

[0016] This invention uses a negative exponential function to calculate the weight of the distance between the target point and each representative point, and performs normalization based on the bipolar distance exponent, so that the target density of the target point can smoothly transition according to the physical distance, ensuring that the target density can truly reflect the local density distribution trend of the target point between the two representative points.

[0017] Preferably, the step of acquiring simulation data at several moments in the transformer simulation process and preprocessing it to obtain standardized simulation data includes: acquiring simulation data at several moments in the transformer simulation process; normalizing multiple operating parameters in the simulation data by linear normalization to obtain normalized multiple operating parameters; obtaining the weight of each operating parameter by analytic hierarchy process; and standardizing it according to the weight of each operating parameter to obtain standardized simulation data.

[0018] Preferably, the standardized simulation data includes standardized data for multiple operating parameters, and the standardized data value of each operating parameter satisfies the following relationship: ; in, It is the first The standardized data values ​​of the project's operating parameters It is the total number of running parameters included in the simulation data. It is the first The weight of the item's running parameters, It is the normalized first Item running parameters.

[0019] This invention calculates standardized data values ​​by introducing adjustment terms into the weights, which stabilizes the standardized data range within a specific interval. At the same time, this proportional relationship preserves the differences in the importance of each operating parameter in the transformer simulation model, thereby improving the robustness of data feature extraction.

[0020] Preferably, the plurality of operating parameters include at least casing temperature, equivalent stress, and fluid pressure.

[0021] In a second aspect, the present invention provides a simulation data processing system for temperature rise of bushings in oil-immersed transformers. The system includes a memory and a processor. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, the simulation data processing method for temperature rise of bushings in oil-immersed transformers according to the first aspect of the present invention is implemented.

[0022] By adopting the above technical solution, a computer program is generated from the simulation data processing method for temperature rise of oil-immersed transformer bushings according to the first aspect of the present invention, and stored in a memory so that it can be loaded and executed by a processor. A terminal device can then be made based on the memory and the processor for convenient use.

[0023] The beneficial effects of this invention are as follows: By acquiring multiple standardized operating parameters from the transformer simulation process and introducing a representative point evaluation mechanism, the global density calculation of high-dimensional operating parameters is transformed into a density-weighted calculation of local representative points. The target density is obtained by utilizing the physical correlation between the local density of representative points and the location of the target point, and then the local data range is dynamically adjusted to filter the initial neighborhood dataset. This processing logic effectively addresses the characteristics of large data volume and obvious nonlinear distribution in the simulation of bushing temperature rise of oil-immersed transformers. It achieves reasonable compression of the retrieval space before executing the clustering algorithm, reduces the number of invalid Euclidean distance calculations between data points, and reduces the occupation of computing resources. At the same time, this invention incorporates local standard deviation as a stability weight in the local density calculation, which can adjust the numerical fluctuations generated by the simulation calculation, further reducing the probability of misjudgment caused by random interference, and achieving efficient and accurate extraction of key temperature rise features of transformers, providing reliable data support for the evaluation of transformer operating status. Attached Figure Description

[0024] Figure 1 A flowchart of a method for processing simulation data of temperature rise in bushings of an oil-immersed transformer, provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of clustering results provided in an embodiment of the present invention; Figure 3 This is a structural block diagram of a simulation data processing system for temperature rise of oil-immersed transformer bushings, provided in an embodiment of the present invention. Detailed Implementation

[0025] The first aspect of this invention provides a method for processing simulation data of temperature rise in the bushings of an oil-immersed transformer, such as... Figure 1 As shown, the method includes steps S100-S500: Step S100: Obtain simulation data at several moments in the transformer simulation process, preprocess the data to obtain standardized simulation data, and the simulation data at each moment includes multiple operating parameters of the transformer simulation.

[0026] It should be noted that, considering the operational safety and structural reliability of the transformer simulation model, the acquisition of multiple operating parameters serves as the foundation for simulation analysis and plays a crucial role in characterizing the equipment's condition. Specifically, excessively high bushing temperatures accelerate insulation aging, and excessive stress in the riser seat can lead to cracking at the welded locations. Therefore, the coordinated matching of bushing temperature, equivalent stress, fluid pressure, and displacement is used to assess the reliability of transformer operation and is key to ensuring safe operation and structural integrity. Meanwhile, high-voltage riser seat stress and other parameters are core indicators for assessing structural strength; whether their values ​​are within the standard range directly relates to whether the simulation object meets the technical evaluation requirements. Parameters such as tank bottom deformation and tank edge stress are closely related to mechanical properties and directly affect the overall service life of the transformer. These operating parameters are of great significance for the reliable operation of the transformer simulation model; therefore, acquiring these data is the first step.

[0027] Specifically, the operating parameters include: bushing temperature, equivalent stress, fluid pressure, displacement, high-voltage riser stress, tank bottom deformation, and tank edge stress, etc. These operating parameters are usually stored in a simulation database. Therefore, this invention obtains these multiple operating parameters during the transformer simulation process through a simulation platform.

[0028] It should be noted that, considering the differences in the numerical ranges of various operating parameters during simulation data processing, directly using the acquired data for calculations would interfere with the accuracy of the results. Therefore, it is necessary to standardize each operating parameter to reduce the interference caused by differences in numerical ranges. Furthermore, different operating parameters have varying degrees of influence in the transformer simulation model. This invention uses the analytic hierarchy process (AHP) to assign weights to each operating parameter, thereby characterizing their actual influence during the standardization process and ensuring that subsequent analysis results accurately reflect the operating characteristics.

[0029] Specifically, firstly, each operating parameter is processed through linear normalization to obtain normalized data, and then the 1st... The first time to obtain The normalized value of the running parameter is denoted as The important weights of each operating parameter are obtained through the analytic hierarchy process. According to weight The data is standardized to obtain standardized simulation data. The standardized data values ​​of each operating parameter satisfy the following relationship: ; in, It is the first The standardized data values ​​of the project's operating parameters It is the total number of running parameters included in the simulation data. It is the first The weight of the item's running parameters, It is the normalized first Item running parameters.

[0030] This relationship stabilizes the maximum value of all data at [value]. Nearby, considering that the average weight of each parameter obtained by the analytic hierarchy process is... By adding to each weight Adjustment item, to adjust the weight benchmark to This weight is then used to adjust the normalized data, thereby controlling the range of the standardized data. Furthermore, linear normalization and the analytic hierarchy process (AHP) are related techniques, which will not be elaborated upon here.

[0031] It should be noted that, considering the complexity of the distribution of transformer simulation data in the multidimensional parameter space, this invention preferably employs the density-based spatial clustering (DBSCAN) algorithm to perform pattern recognition. DBSCAN is a classic density-based clustering method that forms clusters by identifying node regions within the sample space that meet a preset density, and marks nodes in low-density distributions as outliers. This invention chooses DBSCAN precisely because of its characteristics of not requiring a preset number of clusters and being able to identify distribution features of arbitrary shapes. It can effectively handle the nonlinear distribution features caused by drastic fluctuations in temperature rise gradients in the transformer simulation field, and can effectively eliminate noise fluctuations generated by simulation calculations based on density detection logic, thereby ensuring the accurate extraction and identification of the core hotspot regions of the transformer bushing.

[0032] Specifically, when using the DBSCAN algorithm, two main parameters need to be set: the preset neighborhood radius and the minimum number of points. This invention sets the preset neighborhood radius to 0.1. For the minimum number of points, 100 data points can be randomly selected from the standardized simulation data, and the average number of other data points within their neighborhood can be calculated. This average value is then used as the minimum number of points. The preset neighborhood radius and the minimum number of points can be set according to the actual implementation requirements.

[0033] Thus, we obtained a number of standardized operating parameters.

[0034] Step S200: Record any one of the multiple operating parameters as the current operating parameter, divide the current operating parameter into segments, and record each segment point and the two endpoints of the current operating parameter as representative points; obtain the local density of the representative points based on the distribution of the neighborhood data of the representative points.

[0035] It should be noted that, to avoid the excessive computational resource consumption caused by the traditional DBSCAN algorithm calculating the Euclidean distance of all data points, this invention proposes a targeted improvement: by first determining the local data range of each data point, filtering out local data points within that range, and calculating the Euclidean distance between the data point and the local data points, the number of distances to be calculated is reduced, thereby lowering the overall computational load. The following section will focus on how to obtain the local data range of the data points.

[0036] Considering that directly calculating the Euclidean distance between all data points would result in high computational complexity, this invention adopts a representative point density evaluation strategy: by selecting representative feature points from the running parameters, subsequent analysis is based on these representative points, which reduces the amount of computation while preserving the distribution characteristics of the data.

[0037] Specifically, the present invention records the ratio of the absolute difference between the maximum and minimum values ​​of the current item's operating parameter to the preset neighborhood radius as the segment number, and divides the current item's operating parameter into equal segments according to the segment number.

[0038] Based on the above logic, the number of segments satisfies the following relation: ; in, It is the first The number of segments for the item's running parameters. It is the first The standardized data values ​​of the project's operating parameters , They are the first The maximum and minimum values ​​of the item's running parameters. It is the neighborhood radius.

[0039] This relationship is based on the range of values ​​for the running parameters. The system dynamically calculates the number of segments, adapting to variations in data distribution across different operating parameters. When the range of operating parameters is large, the number of segments increases accordingly. The number of segments will increase, making the segmentation more refined; when the values ​​of the running parameters are concentrated, the number of segments will be less, avoiding redundant subdivisions and ensuring that the segments fit the actual characteristics of the data.

[0040] After calculating the number of segments using the piecewise relation, the number of segments is used to determine the segment number of the first segment. The running parameters are divided into equal segments to generate The segmentation points are then combined with the segmentation points of the nth segmentation point. If the two endpoints of the running parameters are used together as the representative point, then the first... There are a total of 100 running parameters. One representative point.

[0041] It should be noted that when obtaining the first... After identifying representative points for the operational parameters, the local density of each representative point can be further calculated. This local density calculation method not only reflects the density of data within the local neighborhood of a representative point but also eliminates the interference of simulation numerical fluctuations through local stability operators, achieving a fusion of physical characteristics and computational confidence. Its calculation logic is based on the distribution of data points within the neighborhood radius of the representative point. By integrating and calculating factors such as the number of data points, the neighborhood range, and the local numerical dispersion, the local density is characterized, laying a solid foundation for subsequent density-based clustering and other data processing steps.

[0042] Specifically, based on the above logic, the local density of the representative point satisfies the following relationship: ; in, It is the first The first item in the running parameters Local density of a representative point; It is the first The running parameters of item number 1 and 2 are related to the first item. The number of parameters whose distance between representative points is less than the preset neighborhood radius; It is the first The first item in the running parameters The local standard deviation of data values ​​within the neighborhood of a representative point is used to characterize the fluctuation characteristics of local data. It is a preset smoothing constant; It is a preset scaling factor for the preset neighborhood radius. When the representative point is a segmentation point of the current item's running parameters, When the representative points are the two endpoints of the current item's running parameters, , It is the neighborhood radius; It is a natural exponential function.

[0043] In this relation, Used to characterize the The first item of the running parameters The density of data points within the neighborhood of a representative point. As a stability weighting term, it is used to evaluate the reliability of local data. If the local standard deviation... A large local standard deviation indicates computational instability in the temperature rise values ​​of the region. This index term will significantly reduce the local density value of the representative point, effectively suppressing the interference of simulation fluctuations on the hotspot identification results. A smaller density indicates that the data distribution in the region is stable, and the local density is mainly determined by the degree of node clustering.

[0044] Thus, the first [item] was obtained. The local density of representative points of the running parameters.

[0045] Step S300: Taking the current time of the current item's running parameters as the target point, select the largest representative point in the current item's running parameters whose value is less than the target point as the first representative point, and the smallest representative point whose value is greater than the target point as the second representative point. Use the negative correlation value of the distance between the target point and the first and second representative points as weights, and perform a weighted summation of the local densities of the representative points corresponding to the first and second representative points to obtain the target density of the target point.

[0046] It should be noted that after obtaining the local density of each representative point, it is also necessary to calculate the density of each data point. This process can be completed based on the representative points and their densities. It should also be noted that when calculating the density of data points, the following logic should be followed: the farther away from the representative point, the smaller the weight; the closer the distance, the larger the weight. The target density of the target point is calculated based on the weights obtained from this logic. This approach can depict the density distribution around the data point while focusing on the local area and relating it to global representative point information.

[0047] Specifically, regarding the first The first time to obtain The current data point of the running parameters is taken as the target point, and the first... The largest representative point whose value in the running parameters is less than the target point is denoted as the first representative point. Select the first The smallest representative point whose value in the running parameters is greater than the target point is denoted as the second representative point. .

[0048] Based on the above logic, the first The first time to obtain The target density of the target points for the running parameters satisfies the following relationship: ; ; ; in, It is the first The first time to obtain The target density of the target points for the item's running parameters. , They are the first The first time to obtain The first representative point, the second representative point, and the third running parameter The first time to obtain The negative exponential function of the absolute difference of the standardized data values ​​of the project's operating parameters. , They are the first The first time to obtain The first and second representative points of the item's operating parameters , They are the first The first time to obtain The local density of the first representative point and the local density of the second representative point for each running parameter. It is the first The first time to obtain The target point data values ​​after standardization of the project's operating parameters. It is the absolute value symbol. It is a natural exponential function.

[0049] In this relation, and Through the first The first time to obtain The negative correlation between the distance between each running parameter and its corresponding representative point is used to weight the local density of the representative points, so that the representative points that are farther away from the target point have smaller weights. For any target point, the greater the target density of each running parameter, the smaller the target point can be selected within a local range.

[0050] Thus, the first [item] was obtained. The first time to obtain The target density of the running parameters.

[0051] Step S400: Obtain the local data range of the target point based on the ratio of the neighborhood radius to the target density of the target point, and determine the neighborhood data points of the target point based on the local data range.

[0052] It should be noted that the first The first time to obtain The target density of the target points for each running parameter should be inversely related to the local data range. The higher the target density, the denser the distribution around the target points, meaning that a smaller area is needed to cover enough similar data. Therefore, the corresponding local data range should be smaller, and fewer local data points need to be acquired.

[0053] Specifically, based on the above logic, the first The first time to obtain The local data range of the target point of the running parameters satisfies the following relationship: ; in, It is the first The first time to obtain Local data range of running parameters It is a preset neighborhood radius. It is the first The first time to obtain The target density of the operating parameters, It is a preset microvalue used to prevent the denominator from being zero. It can be set to 0.001; It is a preset scaling factor for the preset neighborhood radius. When the calculation position corresponding to the target point is in the non-boundary region of the current item's running parameters, the scaling factor corresponding to the segment point is selected. When the calculation location corresponding to the target point is within the boundary region of the current item's running parameters, the scaling factor corresponding to the endpoint is selected. .

[0054] In this relationship, the neighborhood radius is used in conjunction with the first... The first time to obtain The ratio of the target density to the operating parameters is used to calculate the local data range, enabling dynamic control of density over the range by introducing a preset scaling factor. Multiplying by the target density transforms the denominator into a dimensionless form, ensuring homogeneity of both sides of the formula in terms of physical dimensions. A higher target density results in a smaller local data range, incorporating fewer local data points into the calculation, allowing for focused and detailed analysis of high-density regions. Conversely, a lower target density results in a larger local data range, covering more local data points, adapting to retrieval in low-density regions, and ensuring accurate matching between data range filtering and density distribution characteristics.

[0055] Thus, the first [item] was obtained. The first time to obtain Local neighborhood data points of the target point in the running parameters: Data points in the data.

[0056] Step S500: Calculate the neighborhood data points of the target point for multiple operating parameters, merge them and record them as the preliminary neighborhood dataset of the target point, and execute the DBSCAN clustering algorithm based on the distance between the target point and its preliminary neighborhood data points.

[0057] It should be noted that, considering the high coupling of oil-immersed transformer simulation data in a multi-dimensional parameter space, a direct full search would lead to an increased proportion of invalid calculations. Single-parameter filtering can pre-exclude nodes that do not physically or logically meet the neighborhood conditions, thereby reducing resource consumption while ensuring that the subsequently fused dataset can centrally represent the local features of hotspot areas, improving the reliability of clustering identification. Based on this, to accurately calculate the Euclidean distance, this invention employs a strategy of single-parameter filtering and global-local data point fusion.

[0058] First, perform single-parameter filtering, including: for any data point in the transformer simulation data used as the target point. Regarding its first The numerical range for each running parameter is determined based on the aforementioned local data range formula: Based on this numerical range, select the first... All data points whose operating parameters fall within this interval are denoted as... .

[0059] Then, global and local data point fusion is performed, including: acquiring data points. The numerical range of all parameters is obtained by merging all data points to get the local data points of all running parameters. This local data point set is denoted as the preliminary neighborhood dataset.

[0060] It should be noted that DBSCAN clustering is performed based on the Euclidean distance between each data point and the data points in its initial neighborhood dataset.

[0061] Specifically, the core sample points of the DBSCAN algorithm are selected based on the Euclidean distance of the preliminary neighborhood dataset. By filtering data within a local data range, calculations and filtering are performed only on a single running parameter, reducing the number of times the Euclidean distance needs to be calculated for each data point. This can reduce resource consumption and improve processing efficiency when dealing with large-scale simulation data.

[0062] like Figure 2 As shown in the figure, this is a schematic diagram of the clustering results. The horizontal axis represents the standardized parameter 1, and the vertical axis represents the standardized parameter 2. In the figure, circles represent cluster 0, crosses represent cluster 1, and squares represent noise points. Observing the figure, it can be seen that this invention identifies the dense region with parameter 1 around 0.86 as cluster 1 and the region around 0.94 as cluster 0, achieving separation of the two dense regions. This indicates that this invention achieves adaptive scaling of the neighborhood radius by calculating the target density of data points and dynamically adjusting the local data range. This mechanism ensures accurate segmentation of high-density hotspot regions while identifying subtle outlier anomalies, reducing the possibility of insufficient clustering or misjudgment.

[0063] This completes the processing of the transformer simulation data.

[0064] The second aspect of this embodiment provides a simulation data processing system for the temperature rise of bushings in oil-immersed transformers, such as... Figure 3 As shown, the oil-immersed transformer bushing temperature rise simulation data processing system includes a memory and a processor. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, the oil-immersed transformer bushing temperature rise simulation data processing method of the first aspect of the present invention is implemented.

[0065] The oil-immersed transformer bushing temperature rise simulation data processing system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces. Their settings and functions are known in the art and will not be described in detail here.

[0066] In this invention, the aforementioned memory can be any tangible medium containing or storing a program that can be used or combined with an instruction execution system, apparatus, or device. For example, a computer-readable storage medium can be any suitable magnetic or magneto-optical storage medium, such as resistive random access memory (DRAM), dynamic random access memory (DRAM), static random access memory (SRAM), enhanced dynamic random access memory (DRAM), high-bandwidth memory, hybrid memory cube, etc., or any other medium that can be used to store desired information and can be accessed by an application, module, or both. Any such computer storage medium can be part of a device or accessible to or connected to a device.

[0067] The above are all preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Therefore, all equivalent changes made in accordance with the structure, shape and principle of the present invention should be covered within the scope of protection of the present invention.

Claims

1. An oil-immersed transformer bushing temperature rise simulation data processing method, characterized in that, The application relates to a transformer simulation data preprocessing method and device. The method comprises the following steps: acquiring simulation data of a transformer simulation process at several time points, and preprocessing the simulation data to obtain standardized simulation data, wherein the simulation data at each time point comprises multiple operation parameters of the transformer simulation; taking any one of the multiple operation parameters as a current operation parameter, segmenting the current operation parameter, taking each segment point and two end points of the current operation parameter as representative points, and acquiring local density of the representative points according to the distribution of neighborhood data of the representative points; taking the current time point of the current operation parameter as a target point, selecting a maximum representative point with a value smaller than the target point in the current operation parameter as a first representative point, and selecting a minimum representative point with a value larger than the target point in the current operation parameter as a second representative point, taking a negative correlation value of distances between the target point and the first representative point and the second representative point as a weight, and performing weighted summation on the local density of the representative points corresponding to the first representative point and the second representative point to obtain target density of the target point; acquiring a local data range of the target point according to a ratio of a neighborhood radius to the target density of the target point, and determining neighborhood data points of the target point according to the local data range; 2. The oil-immersed transformer bushing temperature rise simulation data processing method according to claim 1, characterized in that, calculating the neighborhood data points of the multiple operation parameters of the target point, merging the neighborhood data points to obtain a preliminary neighborhood data set of the target point, and executing a DBSCAN clustering algorithm based on distances between the target point and the preliminary neighborhood data points of the target point. The segmentation of the current operation parameter comprises the following steps: taking a ratio of an absolute difference value between a maximum value and a minimum value of the current operation parameter to a preset neighborhood radius as a segmentation number; 3. The oil-immersed transformer bushing temperature rise simulation data processing method according to claim 1, characterized in that, segmenting the current operation parameter at equal intervals according to the segmentation number. ； wherein, is the first item running parameter; local density of the first item running parameter; is the first item running parameter; is the first item running parameter; local standard deviation of data values within the neighborhood of the first is a preset smoothing constant; is a preset scaling factor of the preset neighborhood radius; is a natural exponential function.

4. The oil-immersed transformer bushing temperature rise simulation data processing method according to claim 3, characterized in that, The local density of the representative point satisfies a relationship formula.

5. The oil-immersed transformer bushing temperature rise simulation data processing method according to claim 3, characterized in that, When the representative point is a segment point of the current operation parameter, the preset scaling factor is equal to 2 times of the preset neighborhood radius.

6. The oil-immersed transformer bushing temperature rise simulation data processing method according to claim 1, characterized in that, When the representative point is an end point of the current operation parameter, the preset scaling factor is equal to the preset neighborhood radius. The weight of the distance between the target point and the first representative point and the second representative point comprises the following steps: calculating a sum of a negative exponential function of an absolute difference value between the target point and the second representative point and a negative exponential function of an absolute difference value between the target point and the first representative point, and taking the sum as a bipolar distance exponential sum; the weight of the local density of the representative point corresponding to the first representative point is a ratio of the negative exponential function of the absolute difference value between the target point and the first representative point to the bipolar distance exponential sum; 7. The oil-immersed transformer bushing temperature rise simulation data processing method according to claim 1, characterized in that, the weight of the local density of the representative point corresponding to the second representative point is a ratio of the negative exponential function of the absolute difference value between the target point and the second representative point to the bipolar distance exponential sum. The acquisition of the simulation data of the transformer simulation process at the several time points and the preprocessing of the simulation data to obtain the standardized simulation data comprises the following steps: acquiring simulation data of a transformer simulation process at several time points; normalizing the multiple operation parameters in the simulation data through linear normalization to obtain normalized multiple operation parameters; obtaining weights of each operation parameter through an analytic hierarchy process; 8. The oil-immersed transformer bushing temperature rise simulation data processing method according to claim 1, characterized in that, standardizing each operation parameter according to the weight to obtain standardized simulation data. The standardized simulation data comprises standardized data of the multiple operation parameters, and the standardized data value of each operation parameter satisfies a relationship formula. ； in, It is the first The standardized data values ​​of the project's operating parameters It is the total number of running parameters included in the simulation data. It is the first The weight of the item's running parameters, It is the normalized first Item running parameters.

9. The oil-immersed transformer bushing temperature rise simulation data processing method according to claim 1, characterized in that, The multiple operating parameters at least include a bushing temperature, an equivalent stress, and a fluid pressure.

10. An oil-immersed transformer bushing temperature rise simulation data processing system characterized by comprising: The oil-immersed transformer bushing temperature rise simulation data processing system comprises a processor and a memory, and the memory stores computer program instructions which, when executed by the processor, implement the oil-immersed transformer bushing temperature rise simulation data processing method according to any one of claims 1-9.

Images