Local overheating self-adjusting method and system of amorphous alloy transformer

Abstract

The invention discloses a local overheating self-adjusting method and system of an amorphous alloy transformer, and belongs to the technical field of transformers, and the method comprises the steps: collecting temperature data in real time through a multi-point distributed temperature sensing network arranged on the amorphous alloy transformer; thermal sensitive characteristic parameters of the amorphous alloy material are called to carry out local thermal anomaly identification on the temperature data set, a dynamic thermal field difference analysis module is introduced, and local temperature rise trend is predicted in combination with local thermal anomaly data; performing partition detection on the amorphous alloy transformer according to the thermal trend prediction result, generating a plurality of region thermal detection results, performing overheating clustering analysis, and determining a multi-region overheating data class; and carrying out temperature change balance analysis based on the multi-zone overheating data class, formulating a zone cooling strategy, and carrying out intelligent deployment. The technical problems that in the prior art, local overheating of the amorphous alloy transformer is difficult to accurately recognize, thermal trend pre-judgment is not accurate, and resource mismatching and high operation risk are caused by extensive cooling strategies are solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of transformers, in particular to a local overheating self-regulating method and system of amorphous alloy transformers. BACKGROUND

[0002] Amorphous alloy transformers are widely used in intelligent power distribution systems and other scenarios due to their low no-load loss characteristics, and are the core equipment for ensuring efficient power transmission and supporting stable system operation. With the expansion of the scale of intelligent power distribution systems and the increase in load complexity, local overheating of transformers, such as core joints and local areas of windings, can accelerate insulation aging, seriously threatening system reliability.

[0003] Existing control methods mostly rely on single-point temperature detection, which cannot cover the entire thermal field and is prone to miss local hot spots. Moreover, they mostly use fixed threshold alarms, which are prone to misjudgment due to transient load fluctuations. Cooling strategies are mostly whole-on or whole-off, which cannot allocate resources according to local overheating zones, resulting in control lag and high energy consumption, making it difficult to adapt to fine operation and maintenance requirements. SUMMARY

[0004] The present application provides a local overheating self-regulating method and system of amorphous alloy transformers, aiming to solve the technical problems of inaccurate identification of local overheating of amorphous alloy transformers, inaccurate prediction of thermal trends, and resource mismatch and high operation risk caused by extensive cooling strategies.

[0005] In view of the above problems, the present application provides a local overheating self-regulating method and system of amorphous alloy transformers.

[0006] The first aspect of the present application provides a local overheating self-regulating method of amorphous alloy transformers, which comprises collecting temperature data in real time through a multi-point distributed temperature sensing network arranged in the amorphous alloy transformer to obtain a temperature data set; retrieving thermal sensitivity characteristic parameters of amorphous alloy materials to identify local thermal abnormalities of the temperature data set, introducing a dynamic thermal field difference analysis module, combining local thermal abnormal data to predict local temperature rising trends, and generating a thermal trend prediction result; performing zoned detection on the amorphous alloy transformer according to the thermal trend prediction result, generating multiple regional thermal detection results for overheating clustering analysis, and determining multiple-zone overheating data classes; performing temperature change balance analysis based on the multiple-zone overheating data classes, formulating zoned cooling strategies, and executing the zoned cooling strategies for intelligent allocation.

[0007] In another aspect of the present application, a local overheating self-regulating system of an amorphous alloy transformer is provided, which comprises: a temperature data set obtaining module for obtaining temperature data sets by collecting temperature data in real time through a multi-point distributed temperature sensing network arranged in the amorphous alloy transformer; a thermal trend prediction result generating module for performing local thermal anomaly identification on the temperature data sets by calling thermal sensitivity characteristic parameters of the amorphous alloy material, introducing a dynamic thermal field difference analysis module, combining local thermal anomaly data to predict a local temperature rise trend, and generating thermal trend prediction results; a multi-zone overheating data class determining module for performing partition detection on the amorphous alloy transformer according to the thermal trend prediction results, generating a plurality of regional thermal detection results for overheating clustering analysis, and determining a multi-zone overheating data class; and an intelligent adjustment module for performing temperature change balance analysis based on the multi-zone overheating data class, formulating a partition cooling strategy, and executing the partition cooling strategy for intelligent adjustment.

[0008] One or more technical solutions provided in the present application have at least the following technical effects or advantages: Since the multi-point distributed temperature sensing network is arranged to collect temperature data of the amorphous alloy transformer, the thermal sensitivity characteristic parameters of the material are called to identify local thermal anomalies, the dynamic thermal field difference analysis module is introduced to predict the temperature rise trend, and the partition detection clustering and partition cooling intelligent adjustment technical solution are combined, the technical problems of inaccurate identification of local overheating of the amorphous alloy transformer, inaccurate prediction of thermal trend, and resource mismatch caused by extensive cooling strategy and high operation risk in the prior art are solved, and the technical effects of accurate identification of thermal anomalies, accurate prediction of temperature change trend, efficient regulation and control of local overheating, and stable operation of the transformer are achieved.

[0009] The above description is only a summary of the technical solutions of the present application. In order to more clearly understand the technical means of the present application, the present application can be implemented according to the content of the specification, and in order to make the above and other purposes, characteristics and advantages of the present application more obvious and easy to understand, the following specific embodiments of the present application are described. BRIEF DESCRIPTION OF DRAWINGS

[0010] Figure 1 A flowchart of a local overheating self-regulating method of an amorphous alloy transformer is provided for the embodiments of the present application; Figure 2 A flowchart of a process of calling thermal sensitivity characteristic parameters of an amorphous alloy material in a local overheating self-regulating method of an amorphous alloy transformer is provided for the embodiments of the present application; Figure 3 A structural schematic diagram of a local overheating self-regulating system of an amorphous alloy transformer is provided for the embodiments of the present application.

[0011] Figure labeling: Temperature dataset acquisition module 11, thermal trend prediction result generation module 12, multi-zone overheating data class determination module 13, intelligent allocation module 14. Detailed Implementation

[0012] The overall concept of the technical solution provided in this application is as follows: This application provides a method and system for self-regulating local overheating of amorphous alloy transformers. Based on the collection of temperature data of amorphous alloy transformers, the method retrieves the thermal sensitivity parameters of the material to identify local thermal anomalies, introduces a dynamic thermal field difference analysis module to predict the temperature rise trend, and then, through zonal detection and overheating clustering, formulates zonal cooling strategies and intelligently allocates them to specifically solve the problem of local overheating control.

[0013] After introducing the basic principles of this application, various non-limiting embodiments of this application will be described in detail below with reference to the accompanying drawings. Example 1

[0014] like Figure 1 As shown in the embodiment of this application, a method for self-regulating local overheating of an amorphous alloy transformer is provided. The method includes: Step S100: Temperature data is collected in real time through a multi-point distributed temperature sensing network deployed on the amorphous alloy transformer to obtain a temperature dataset.

[0015] Specifically, amorphous alloy transformers are transformers with amorphous alloy cores and are widely used in intelligent power distribution systems, offering significant energy-saving advantages. Multi-point distributed temperature sensing networks are networked detection systems formed by arranging multiple temperature sensors according to specific rules in different areas of the transformer and connecting them via communication links.

[0016] Specifically, firstly, based on the structural characteristics of amorphous alloy transformers, such as key areas like the core, high and low voltage windings, and tank, suitable temperature sensors are selected. For example, fiber optic grating sensors are used in winding areas with high temperatures and strong electromagnetic interference because they are resistant to electromagnetic interference and have a wide measurement range; platinum resistance sensors are used in more stable environments such as the tank surface because they are less expensive. Then, sensors are arranged according to the principle of increasing density in key areas and simplifying in non-critical areas. Next, the sensors are connected into a network through wireless transmission modules or wired buses, and data acquisition terminals, such as edge computing gateways, receive data from each sensor in real time. Finally, these discrete data containing location, time, and temperature are integrated into a structured temperature dataset.

[0017] This step achieves full coverage monitoring of the transformer's thermal field through a multi-point distributed network. Combined with real-time acquisition characteristics, it ensures that temperature data can dynamically reflect the instantaneous thermal state under load fluctuations, providing high-density and high-timeliness raw data support for subsequent identification of local thermal anomalies and prediction of thermal trends.

[0018] Step S200: Retrieve the thermosensitive characteristic parameters of the amorphous alloy material to identify local thermal anomalies in the temperature dataset, introduce the dynamic thermal field difference analysis module, and combine the local thermal anomaly data to predict the local heating trend and generate thermal trend prediction results.

[0019] Specifically, thermosensitive characteristic parameters refer to the characteristic parameters of the thermophysical properties of amorphous alloy materials that change with temperature, mainly including thermal conductivity and specific heat capacity. These parameters directly affect the heat generation and dissipation laws of the material. The dynamic thermal field difference analysis module is used to analyze the changes in heat distribution within the three-dimensional space of a transformer over time in real time. It can calculate the differences in temperature gradient, heat flow direction, etc., of the thermal field at different times, reflecting the heat diffusion trend. For the dynamic thermal field difference analysis module, the input temperature dataset and transformer structural parameters are used to divide the thermal field into snapshots according to the time series; the temperature distribution characteristics of each time period are extracted using a sliding window, and the temperature gradient vector is calculated in combination with the three-dimensional thermal field; the changes in heat flow parameters are analyzed through a spatiotemporal comparison algorithm to calculate the thermal field difference values ​​at different times; the output is the difference analysis results including changes in heat flow direction and intensity, supporting the prediction of temperature rise trends.

[0020] Specifically, the thermosensitive characteristic parameters of the amorphous alloy material are retrieved, and the temperature dataset is compared with the normal temperature threshold corresponding to the parameters. A machine learning-based anomaly detection algorithm is used to identify local thermal anomalies. Subsequently, a dynamic thermal field difference analysis module is introduced to map the temperature data onto the three-dimensional model of the transformer and calculate the thermal field difference at different times. Combining the local thermal anomaly data and the thermal field difference results, a time-series prediction model is used to predict the temperature rise trend of the abnormal area, generating a thermal trend prediction result that includes the temperature peak and the temperature rise rate.

[0021] This step introduces the thermosensitive characteristic parameters of amorphous alloy materials, making thermal anomaly identification more in line with the material's nature; dynamic thermal field difference analysis realizes the spatiotemporal correlation analysis of heat distribution, significantly improving the prediction accuracy of local heating trends.

[0022] Step S300: Perform zoned testing on the amorphous alloy transformer according to the thermal trend prediction results, generate multiple regional thermal test results, perform overheating cluster analysis, and determine the overheating data class of multiple zones.

[0023] Specifically, zone testing is a process of dividing the transformer into multiple independent testing units based on the transformer structure and thermal trend prediction results, such as the core area, high-voltage winding area, low-voltage winding area, and oil tank area, and specifically monitoring the thermal performance of each area.

[0024] Specifically, based on the structural and thermal trend prediction results of the amorphous alloy transformer, the transformer is divided into multiple detection zones, including the core zone, high-voltage winding zone, low-voltage winding zone, front tank zone, and rear tank zone. Then, an infrared thermal imager is used to detect the temperature distribution in each zone, and distributed fiber optic sensors record the temperature time sequence to perform thermal testing. The regional thermal testing results are generated using a LabVIEW data acquisition system. Furthermore, the DBSCAN clustering algorithm is invoked, using heating rate, thermal stability, and heat dissipation rate as feature parameters to cluster the detection results for each zone, thus determining the overheating data classes for multiple zones.

[0025] This step uses zonal detection to focus thermal analysis on specific components, avoiding the obscuring of local problems by macroscopic data; cluster analysis categorizes regions with similar overheating characteristics, reducing the complexity of subsequent cooling strategy formulation and mitigating overcooling or undercooling problems caused by misjudgment of regional characteristics.

[0026] Step S400: Perform temperature change balance analysis based on the multi-zone overheating data, formulate a zoned cooling strategy, and execute the zoned cooling strategy for intelligent allocation.

[0027] Specifically, the zoned cooling strategy refers to a differentiated cooling scheme formulated for the thermal characteristics of different multi-zone overheated data types, including parameters such as cooling method, intensity, and duration.

[0028] Specifically, temperature change balance analysis is performed on the thermal trend prediction results of multi-zone overheating data to calculate the cooling demand difference of each zone and assess the regional thermal interaction impact; combined with the cooling equipment parameters, the optimization toolbox of MATLAB is used to formulate a zoned cooling strategy; and intelligent allocation is completed through a PLC control system.

[0029] This step avoids misallocation of cooling resources through temperature balance analysis; the zoning strategy improves cooling efficiency, enabling precise, efficient, and low-consumption regulation of local overheating in amorphous alloy transformers, and extending equipment insulation.

[0030] Furthermore, such as Figure 2 As shown, the process of retrieving the thermosensitive characteristic parameters of an amorphous alloy material includes: setting multiple temperature ranges and multiple magnetic flux densities; performing thermal analysis on the amorphous alloy material according to the multiple temperature ranges to obtain a first thermophysical parameter; performing thermal analysis on the amorphous alloy material according to the multiple magnetic flux densities to obtain a second thermophysical parameter; performing thermosensitive detection based on the first and second thermophysical parameters to construct a thermosensitive parameter curve; and dynamically selecting based on the magnetic flux density distribution information of the amorphous alloy transformer combined with the thermosensitive parameter curve to determine the thermosensitive characteristic parameters of the amorphous alloy material.

[0031] Specifically, multiple temperature ranges refer to several continuous temperature segments divided according to the actual operating temperature range of the amorphous alloy transformer, covering possible normal and abnormal temperature scenarios. Multiple magnetic flux densities refer to the multiple magnetic flux density values ​​that may exist in the transformer core, covering the range of magnetic flux changes under different operating conditions such as no-load, full-load, and overload. The first thermophysical parameter refers to the thermal characteristic parameters of the amorphous alloy material measured under multiple temperature ranges, mainly including parameters directly related to temperature such as thermal conductivity and specific heat capacity. The second thermophysical parameter refers to the thermal characteristic parameters of the amorphous alloy material measured under multiple magnetic flux densities, mainly including heating parameters related to electromagnetic induction such as hysteresis loss power and eddy current loss coefficient. Magnetic flux density distribution information refers to the magnetic flux density distribution data of various regions of the core during the actual operation of the amorphous alloy transformer.

[0032] Specifically, based on the design operating range of the amorphous alloy transformer, multiple temperature ranges and multiple magnetic flux densities are set. A differential scanning calorimeter is used to test the thermal conductivity and specific heat capacity of the material at each temperature range to obtain the first thermophysical parameter. Simultaneously, a vibrating sample magnetometer combined with a thermal imager is used to test hysteresis loss and eddy current loss at each magnetic flux density to obtain the second thermophysical parameter. Furthermore, the first and second thermophysical parameters are thermally sensitively detected using MATLAB data fitting tools, and outliers are removed to construct thermal sensitivity parameter curves. Finally, real-time magnetic flux density distribution information is obtained through the transformer online monitoring system, and the applicable thermal sensitivity parameters for each region are determined by matching the parameters at the corresponding temperatures based on this distribution in the thermal sensitivity parameter curves.

[0033] This step, through multi-dimensional testing and dynamic matching, makes the obtained thermal characteristic parameters more consistent with the actual operating conditions of the transformer, providing a precise material characteristic basis for the entire self-adjustment method.

[0034] Furthermore, the temperature dataset is analyzed by retrieving the thermosensitive characteristic parameters of the amorphous alloy material to identify local thermal anomalies. A dynamic thermal field difference analysis module is introduced, and the local heating trend is predicted based on the local thermal anomaly data to generate a thermal trend prediction result. The method includes: introducing the structural parameters of the amorphous alloy transformer to perform mesh division and determine multiple regions to be analyzed; performing heat source temperature rise analysis on the multiple regions to be analyzed based on the thermosensitive characteristic parameters of the amorphous alloy material and setting multiple expected temperature values; comparing the temperature dataset with the multiple regions to be analyzed according to the differences between the multiple expected temperature values ​​to identify local thermal anomaly data; constructing a three-dimensional thermal field of the amorphous alloy transformer, mapping the temperature dataset to the three-dimensional thermal field to calculate the temperature gradient vector, and determining the transformer heat flow parameters, which include heat flow direction parameters and heat flow intensity parameters; performing spatiotemporal variation difference analysis on the multiple regions to be analyzed based on the heat flow direction parameters and the heat flow intensity parameters to generate a thermal field difference analysis result; and predicting the local heating trend by combining the local thermal anomaly data and the thermal field difference analysis results according to a time series to generate the thermal trend prediction result.

[0035] Specifically, the structural parameters of amorphous alloy transformers are key parameters describing the transformer's physical structure, including core dimensions, number of winding layers and turns, tank volume, and heat sink spacing, reflecting the spatial position and connection relationships of each component. The expected temperature value refers to the temperature range under normal operation for each analyzed region, determined based on heat source temperature rise analysis; exceeding this range may indicate anomalies. The temperature gradient vector represents the rate of temperature change in the three-dimensional thermal field, pointing in the direction of the fastest temperature increase, and its magnitude reflects the rate of temperature change. Transformer heat flux parameters describe the characteristics of heat transfer, where the heat flux direction parameter indicates the spatial direction of heat transfer, such as from the windings to the tank, and the heat flux intensity parameter represents the amount of heat passing through a unit area per unit time.

[0036] Specifically, the structural parameters of the amorphous alloy transformer are obtained, and finite element software is used to generate a mesh, determining multiple regions to be analyzed. Then, based on the previously obtained thermosensitive characteristic parameters, COMSOL Multiphysics is used to perform heat source temperature rise simulation analysis, setting expected temperature values ​​for each region. The temperature dataset is grouped according to the regions to be analyzed, and the deviation between the measured temperature and the expected value is calculated to identify the region as a local thermal anomaly. Next, a three-dimensional thermal field of the transformer is constructed using a 3D modeling tool, mapping the temperature data from each sensor to the corresponding mesh. The finite difference method is used to calculate the temperature gradient vector and determine the heat flux parameters. Then, a dynamic thermal field difference analysis module is introduced to compare historical heat flux parameters with current heat flux parameters, generating thermal field difference analysis results. Finally, the local thermal anomaly data and thermal field difference results are input into an LSTM neural network model to predict future temperature changes, generating thermal trend prediction results. First, the local thermal anomaly data and thermal field difference data are normalized to construct the "time step-feature" input sample; then, the network is built: the input layer is followed by a 2LSTM, and the output layer uses a fully connected layer to output the temperature prediction value; Adam is used as the optimizer and MSE is used as the loss function, and the training and validation sets are iteratively trained using an 8:2 ratio. After parameter tuning, the LSTM neural network model is completed.

[0037] This step, through structured region division and multi-dimensional thermal field analysis, allows the local overheating analysis to focus on specific components, avoiding blind spots in macroscopic analysis; the temperature expectation value set by combining the thermosensitive characteristic parameters is more in line with the actual heat generation law of the material.

[0038] Furthermore, the temperature dataset is compared with the multiple expected temperature values ​​according to the multiple regions to be analyzed to identify local thermal anomaly data. The method includes: grouping the temperature dataset according to the multiple regions to be analyzed to determine multiple temperature data subsets; calculating the difference between the multiple temperature data subsets and the multiple expected temperature values ​​to obtain multiple absolute deviation values; performing deviation extreme value analysis on the multiple regions to be analyzed based on the multiple expected temperature values ​​to set a region allowable deviation threshold; comparing the absolute deviation values ​​with the region allowable deviation threshold to identify multiple out-of-limit data points; continuously verifying the multiple out-of-limit data points; and performing spatiotemporal clustering analysis on the verified out-of-limit data points to generate the local thermal anomaly data.

[0039] Specifically, a temperature data subset refers to the dataset obtained by splitting the original temperature dataset according to the region to be analyzed. Each subset contains temperature data from all sensors within the corresponding region. The absolute deviation value refers to the absolute difference between the measured temperature in the temperature data subset and the expected temperature value for the corresponding region. The regional allowable deviation threshold refers to the maximum permissible temperature deviation set based on deviation extreme value analysis. Exceeding-limit data points refer to temperature data points whose absolute deviation values ​​exceed the regional allowable deviation threshold.

[0040] Specifically, the temperature dataset is grouped according to the region to be analyzed, resulting in multiple temperature data subsets. The NumPy library is used to calculate the absolute deviation between the measured temperature of each subset and the expected temperature value of the corresponding region. SPSS statistical tools are used to perform deviation extreme value analysis on the historical normal data of each region. Then, the absolute deviation value is compared with the threshold to identify the data points that exceed the limit. The time series analysis tool is then used to continuously monitor the points that exceed the limit to confirm that they continue to exceed the limit. Finally, the DBSCAN clustering algorithm is used to perform spatiotemporal clustering on the verified points that exceed the limit, clustering them into a continuous abnormal region to generate local thermal anomaly data.

[0041] This step, through multi-layered screening including partitioning analysis, dynamic thresholding, continuous verification, and spatiotemporal clustering, significantly improves the accuracy of local thermal anomaly identification, providing a highly reliable foundation of anomaly data for subsequent thermal trend prediction.

[0042] Furthermore, the local thermal anomaly data and the thermal field difference analysis results are combined according to a time series to predict the local temperature rise trend and generate the thermal trend prediction result. The method includes: extracting the abnormal temporal features of the local thermal anomaly data according to the time series; detecting the spatial distribution of temperature based on the thermal field difference analysis results to obtain spatial features; integrating the abnormal temporal features and the spatial features to construct a three-dimensional spatiotemporal feature matrix; segmenting the three-dimensional spatiotemporal feature matrix to obtain multiple segmented features, the multiple segmented features including trend features and periodic features; performing deep learning based on the trend features and the periodic features, and performing multivariate time series prediction based on the learning results to generate a local area temperature change prediction result; and performing thermal change assessment based on the local area temperature change prediction result to construct the thermal trend prediction result.

[0043] Specifically, anomaly time-series features refer to the time-varying characteristics extracted from the time series of local thermal anomaly data, including heating rate, fluctuation frequency, and duration, reflecting the temporal evolution of anomalous temperatures. Spatial features are the spatial distribution characteristics of temperature extracted based on the results of thermal field difference analysis, including the geometry of the anomalous region, its spatial location, and its temperature gradient with surrounding areas, reflecting the spatial distribution pattern of anomalous temperatures. The three-dimensional spatiotemporal feature matrix is ​​a three-dimensional data structure integrating anomaly time-series features and spatial features, with dimensions of time step, spatial coordinates, and eigenvalues, containing both temporal dynamics and spatial distribution information. Segmented features refer to the features obtained by splitting the three-dimensional spatiotemporal feature matrix according to the time dimension, where trend features reflect the long-term direction of temperature change, and periodic features reflect short-term periodic fluctuations. Thermal change assessment refers to assessing the risk level of temperature changes based on temperature change prediction results, providing a basis for judgment of thermal trend prediction results.

[0044] Specifically, time-series analysis is performed on local thermal anomaly data to extract anomaly time-series features. Simultaneously, 3D visualization tools are used to extract spatial features from the thermal field difference analysis results, such as temperature distribution maps at different times, to determine the geometry of the region and the temperature gradient of adjacent regions. Then, a NumPy array is used to integrate the anomaly time-series features and spatial features to construct a 3D spatiotemporal feature matrix. Pandas' time-series piecewise function is used to split the matrix into trend features and periodic features over time. These piecewise features are then input into a multivariate LSTM model built on TensorFlow, using historical operating data as the training set to predict future temperature changes. Finally, a risk assessment module is used to assess thermal changes, generating a thermal trend prediction result containing the predicted curve and risk level. Specifically, the input includes the local area temperature change prediction result, the equipment critical temperature, heat flux intensity parameters, and regional importance weights. Risk level thresholds are set: Low: predicted temperature less than the critical value; Medium: critical value reached but temperature rise is slow; High: above the critical value and temperature rise is rapid. The comprehensive risk value is calculated using the analytic hierarchy process (AHP) and implemented using Python's Scikit-learn algorithm. Historical failure data is used to validate and adjust the weights, and the final output is an assessment result with risk level.

[0045] This step captures the dynamic changes in temperature through abnormal time series features, and spatial features reflect the spatial constraints of heat diffusion. The combination of the two makes the three-dimensional spatiotemporal feature matrix closer to the actual physical process; segmented processing separates trend and periodic features to avoid short-term fluctuations interfering with the judgment of long-term trends.

[0046] Furthermore, the amorphous alloy transformer is subjected to zonal testing according to the thermal trend prediction results to generate multiple regional thermal test results. The method includes: traversing the multiple regions to be analyzed for thermal performance analysis and setting multiple thermal performance test indicators; using a sliding time window to perform thermal behavior analysis on the multiple regions to be analyzed according to the multiple thermal performance test indicators and the thermal trend prediction results to generate thermal behavior characteristics; calculating the temperature distribution uniformity of the multiple regions to be analyzed based on the thermal behavior characteristics and identifying multiple local zone hotspots; and performing zonal testing based on the multiple local zone hotspots to determine the multiple regional thermal test results.

[0047] Specifically, thermal performance analysis refers to the quantitative analysis of the heat generation and heat dissipation characteristics of each area under analysis, including heat generation rate and heat dissipation efficiency, to assess the thermal stability of the area. Thermal behavior analysis, based on thermal trend prediction results and sliding window data, analyzes the temperature change patterns of an area over time, such as continuous temperature rise, periodic fluctuations, and patterns of rapid rise followed by stabilization. Temperature distribution uniformity describes the degree of uniformity of temperature distribution within an area, usually expressed as the difference between the highest and lowest temperatures within the area or the temperature standard deviation. Local hotspots refer to small, localized areas with low temperature distribution uniformity and significantly higher temperature values ​​than the surrounding areas.

[0048] The process involves iterating through the region to be analyzed and setting thermal performance testing indicators based on the characteristics of amorphous alloy transformers. A sliding time window is then set, and based on thermal trend prediction results, the temperature dataset for each region is truncated. The data within the window is analyzed using SciPy statistical tools to generate thermal behavior characteristics. Next, Matlab's spatial analysis module is used to calculate the temperature distribution uniformity of each region based on these thermal behavior characteristics. Finally, the above information is integrated to generate regional thermal performance testing results for each region to be analyzed, including indicator values, behavioral characteristics, and hotspot locations.

[0049] This step captures instantaneous changes in thermal behavior through a sliding time window, avoiding short-term anomalies masked by long-term average data; the multi-indicator system comprehensively reflects regional thermal performance; the combination of temperature distribution uniformity calculation and hotspot identification improves the overheating location accuracy from "regional level" to "grid level", providing high-resolution basic data for subsequent overheating cluster analysis and effectively supporting the formulation of differentiated cooling strategies.

[0050] Furthermore, multiple regional thermal detection results are generated for overheating cluster analysis to determine the overheating data classes in multiple regions. The method includes: performing thermal similarity calculation based on the multiple local hot spots to obtain thermal similarity scores for multiple regions; initially grouping the multiple regional thermal detection results according to the thermal similarity scores to determine the number of clusters; verifying the clustering stability based on the number of clusters; performing hierarchical clustering on the multiple regional thermal detection results based on the verification results and the number of clusters to generate initial overheating data; evaluating the clustering effect based on the initial overheating data to obtain a clustering score; filtering the initial overheating data according to the clustering score to construct an overheating distribution map, wherein the overheating distribution map contains the multiple regions of overheating data classes.

[0051] Specifically, the zonal thermal similarity represents a quantitative value indicating the degree of similarity in thermal characteristics between two regions to be analyzed; a higher value indicates that the thermal behavior is more similar. The overheating distribution map presents a spatial distribution map of multi-zone overheating data classes in a visual manner, marking the regional location and common characteristics of each data class. Multi-zone overheating data classes refer to the set of regions with stable common thermal characteristics determined after screening, and are the core content of the overheating distribution map.

[0052] First, core features are extracted from the regional thermal monitoring results: heating rate, number of hot spots, standard deviation of temperature fluctuation, heat accumulation, and heat dissipation rate. The cosine_similarity function is called to calculate the thermal similarity between regions. Preliminary grouping is performed based on the similarity matrix, and the elbow method is used to analyze the data distribution to determine the number of clusters. The Bootstrap sampling method is used for stability verification, followed by hierarchical clustering to generate initial overheating data. The silhouette coefficient is used to evaluate the initial data and obtain the cluster score. Finally, an overheating distribution map is drawn, and the spatial location and common characteristics of each data class are marked to determine the overheating data classes in multiple regions.

[0053] This step uses thermal similarity calculation based on multi-dimensional features to make the similarity more closely match the actual thermal characteristics; the dual screening of cluster stability verification and effect evaluation reduces clustering error.

[0054] Furthermore, based on the multi-zone overheating data, temperature balance analysis is performed to formulate a zoned cooling strategy and execute the zoned cooling strategy for intelligent allocation. The method includes: evaluating the cooling performance of the amorphous alloy transformer to obtain extreme cooling performance values; performing heat dissipation analysis on the amorphous alloy transformer based on the multi-zone overheating data and setting heat dissipation demand priorities; allocating cooling resources according to the heat dissipation demand priorities and the extreme cooling performance values ​​to formulate a cooling parameter set; distributing the cooling parameter set according to the multiple regions to be analyzed to generate a zoned cooling strategy, the zoned cooling strategy including a cooling command sequence; executing the cooling command sequence to monitor the cooling effect in real time, and dynamically adjusting the feedback based on the cooling effect to adaptively adjust the local overheating of the amorphous alloy transformer.

[0055] Specifically, cooling assessment refers to the process of detecting and quantifying the maximum cooling capacity and operating thresholds of a transformer cooling system, such as maximum fan speed and maximum coolant flow rate, to determine the system's physical limits. Cooling performance extremes refer to the maximum cooling effect parameters that the cooling system can achieve, such as maximum fan speed and maximum liquid cooling system flow rate, reflecting the upper limit of cooling resources. Heat dissipation demand priority refers to the cooling sequence set according to regional importance and overheating risk level.

[0056] Specifically, firstly, an infrared thermal imager and a power meter are used to evaluate the cooling system of the transformer, measuring the maximum heat dissipation power of the air-cooled system and the maximum flow rate of the liquid-cooled system to determine the extreme values ​​of cooling performance. Next, based on the multi-zone overheating data, the hierarchical analysis method is used to set the priority of heat dissipation demand, and combined with the extreme values ​​of cooling performance, cooling resources are allocated to formulate a set of cooling parameters. Then, the parameter set is decomposed according to the area to be analyzed to generate a cooling instruction sequence containing equipment number, action, parameter, and duration. Finally, the instructions are executed through the PLC control system, while distributed fiber optic sensors monitor the temperature in real time, triggering dynamic adjustment feedback and continuously adaptively adjusting until the temperature of each area stabilizes within a safe range.

[0057] This step enables precise and dynamic utilization of cooling resources: extreme value assessment of cooling performance avoids overloading of the cooling system; priority allocation ensures that high-risk areas receive resources first, significantly reducing the risk of insulation aging.

[0058] Furthermore, based on the multi-zone overheating data class, a heat dissipation analysis is performed on the amorphous alloy transformer, and a priority for heat dissipation requirements is set. The method includes: traversing the overheating distribution map according to the multi-zone overheating data class to determine the overheating region distribution pattern; extracting multiple overheating regions based on the multi-zone overheating data class, calculating the temperature rise of the multiple overheating regions based on the overheating region distribution pattern to obtain a regional temperature rise dataset, wherein the regional temperature rise dataset includes heat accumulation rate data and heat diffusion trend data; conducting an operational safety impact assessment on the amorphous alloy transformer according to the heat accumulation rate data and the heat diffusion trend data to generate an operational risk level; and performing a heat dissipation requirement analysis based on the overheating region distribution pattern and the operational risk level to set the priority for heat dissipation requirements.

[0059] Specifically, the overheating area distribution pattern refers to the spatial distribution characteristics of overheated areas within the transformer, such as "concentrated," where overheating is densely concentrated in a certain layer of the winding; "diffused," where it gradually spreads from the core to the winding; and "discrete," where multiple independent small areas are randomly distributed, reflecting the spatial diffusion law of overheating. Heat accumulation rate data is a parameter used to describe how quickly heat accumulates in the overheated area; a higher value indicates faster heat accumulation and a more rapid increase in overheating risk. Heat diffusion trend data reflects the dynamic characteristics of heat transfer from the overheated area to the surrounding area, including diffusion speed, diffusion direction, and diffusion range.

[0060] Specifically, the process involves traversing the overheating distribution map, analyzing the spatial distribution of multiple multi-zone overheating data classes, and determining the distribution pattern of overheated areas. Then, the corresponding overheated areas are extracted, and the NumPy library is used to calculate the regional temperature rise dataset. Subsequently, the Analytic Hierarchy Process (AHP) combined with the FMEA model is used to assess the operational safety impact. For example, Class 1, located near the high-voltage winding insulation layer, experiences rapid heat accumulation and is prone to insulation aging, thus assessed as "high" risk. Class 2, although diffused, is far from critical insulation components, thus assessed as "medium" risk. Class 3, with low temperature and located on the edge, is assessed as "low" risk. Finally, combining the distribution pattern and operational risk level, TOPSIS is used to analyze heat dissipation requirements and set priorities.

[0061] This step avoids overcooling of diffuse areas by identifying overheated zone distribution patterns; quantifies dynamic changes in risk based on heat accumulation and diffusion data; and enables accurate assessment of safety impacts through operational risk level evaluation. This improves resource utilization while preventing wasted cooling in low-risk areas.

[0062] In summary, the local overheating self-adjustment method for amorphous alloy transformers provided in this application has the following technical effects: 1. This method constructs a closed-loop process of "temperature measurement-identification-prediction-clustering-cooling," breaking through the limitations of traditional intelligent power distribution systems' overall transformer temperature measurement and crude cooling. It achieves accurate identification, trend prediction, and zoned control of localized overheating. This effectively solves problems such as difficulty in detecting overheating in hidden areas and misallocation of cooling resources, improves the equipment's response to localized thermal anomalies, ensures long-term stable transformer operation, and reduces the risk of failures due to overheating.

[0063] 2. By conducting parameter tests across multiple temperature ranges and magnetic flux densities, and dynamically selecting thermistor parameters based on the actual magnetic flux distribution of the transformer, the deviation caused by the disconnect between traditional fixed parameters and operating conditions is avoided. This ensures that the acquired parameters more closely match the true thermophysical properties of amorphous alloy materials, providing accurate underlying data support for subsequent thermal anomaly identification and improving the analytical reliability of the entire regulation method from the source.

[0064] 3. By combining overheating zone distribution patterns, temperature rise data, and operational risk assessments to set priorities, it breaks through the limitations of traditional prioritization based solely on temperature values. It can accurately determine the impact of different overheating zones on equipment safety, ensuring that cooling resources are prioritized for high-risk, high-impact areas, avoiding excessive resource consumption in low-risk areas, and significantly improving the efficiency and targeting of cooling resource utilization. Example 2

[0065] Based on the same inventive concept as the local overheat self-regulation method for amorphous alloy transformers in the foregoing embodiments, such as Figure 3 As shown in the embodiment of this application, a local overheat self-regulation system for an amorphous alloy transformer is provided. The system includes: The temperature dataset acquisition module 11 is used to collect temperature data in real time through a multi-point distributed temperature sensing network deployed on the amorphous alloy transformer to obtain a temperature dataset; the thermal trend prediction result generation module 12 is used to retrieve the thermosensitive characteristic parameters of the amorphous alloy material to identify local thermal anomalies in the temperature dataset, introduce a dynamic thermal field difference analysis module, and combine the local thermal anomaly data to predict the local temperature rise trend and generate a thermal trend prediction result; the multi-zone overheating data class determination module 13 is used to perform zone detection on the amorphous alloy transformer according to the thermal trend prediction result, generate multiple zone thermal detection results for overheating cluster analysis, and determine the multi-zone overheating data class; the intelligent allocation module 14 is used to perform temperature change balance analysis based on the multi-zone overheating data class, formulate a zoned cooling strategy, and execute the zoned cooling strategy for intelligent allocation.

[0066] Furthermore, the thermal trend prediction result generation module 12 is also used to perform the following steps: setting multiple temperature ranges and multiple magnetic flux densities; performing thermal analysis on the amorphous alloy material according to the multiple temperature ranges to obtain a first thermophysical parameter; performing thermal analysis on the amorphous alloy material according to the multiple magnetic flux densities to obtain a second thermophysical parameter; performing thermosensitive detection based on the first thermophysical parameter and the second thermophysical parameter to construct a thermosensitive parameter curve; and dynamically selecting based on the magnetic flux density distribution information of the amorphous alloy transformer combined with the thermosensitive parameter curve to determine the thermosensitive characteristic parameters of the amorphous alloy material.

[0067] Furthermore, the thermal trend prediction result generation module 12 is also used to perform the following steps: introducing the structural parameters of the amorphous alloy transformer for mesh division to determine multiple regions to be analyzed; performing heat source temperature rise analysis on the multiple regions to be analyzed based on the thermosensitive characteristic parameters of the amorphous alloy material, and setting multiple expected temperature values; comparing the temperature dataset with the multiple regions to be analyzed according to the differences between the multiple expected temperature values, and identifying local thermal anomaly data; constructing a three-dimensional thermal field of the amorphous alloy transformer, mapping the temperature dataset to the three-dimensional thermal field to calculate the temperature gradient vector, and determining the transformer heat flow parameters, which include heat flow direction parameters and heat flow intensity parameters; performing spatiotemporal variation difference analysis on the multiple regions to be analyzed according to the heat flow direction parameters and the heat flow intensity parameters, and generating thermal field difference analysis results; predicting the local temperature rise trend by comparing the local thermal anomaly data and the thermal field difference analysis results according to the time series, and generating the thermal trend prediction result.

[0068] Furthermore, the thermal trend prediction result generation module 12 is also used to perform the following steps: grouping the temperature dataset according to the multiple regions to be analyzed to determine multiple temperature data subsets; calculating the difference between the multiple temperature data subsets and the multiple expected temperature values ​​to obtain multiple absolute deviation values; performing deviation extreme value analysis on the multiple regions to be analyzed based on the multiple expected temperature values ​​to set a region allowable deviation threshold; comparing the absolute deviation values ​​with the region allowable deviation threshold to identify multiple out-of-limit data points; continuously verifying the multiple out-of-limit data points, and performing spatiotemporal clustering analysis on the verified out-of-limit data points to generate the local thermal anomaly data.

[0069] Furthermore, the thermal trend prediction result generation module 12 is also used to perform the following steps: extracting abnormal temporal features of local thermal anomaly data according to the time series; detecting spatial temperature distribution based on the thermal field difference analysis results to obtain spatial features; integrating the abnormal temporal features and the spatial features to construct a three-dimensional spatiotemporal feature matrix; segmenting the three-dimensional spatiotemporal feature matrix to obtain multiple segmented features, the multiple segmented features including trend features and periodic features; performing deep learning based on the trend features and the periodic features, performing multivariate temporal prediction based on the learning results, and generating local area temperature change prediction results; and performing thermal change assessment based on the local area temperature change prediction results to construct the thermal trend prediction results.

[0070] Furthermore, the multi-zone overheating data classification module 13 is also used to perform the following steps: traversing the multiple regions to be analyzed for thermal performance analysis, and setting multiple thermal performance detection indicators; using a sliding time window to perform thermal behavior analysis on the multiple regions to be analyzed based on the thermal trend prediction results according to the multiple thermal performance detection indicators, and generating thermal behavior features; calculating the temperature distribution uniformity of the multiple regions to be analyzed based on the thermal behavior features, and identifying multiple local hot spots; performing zone detection based on the multiple local hot spots, and determining the thermal performance detection results of the multiple regions.

[0071] Furthermore, the multi-zone overheating data class determination module 13 is also used to perform the following steps: perform thermal similarity calculation based on the multiple local partition hotspots to obtain multiple partition thermal similarity; perform preliminary grouping of the multiple regional thermal detection results according to the multiple partition thermal similarity to determine the number of clusters; perform clustering stability verification based on the number of clusters, and perform hierarchical clustering of the multiple regional thermal detection results according to the verification results and the number of clusters to generate initial overheating data; evaluate the clustering effect based on the initial overheating data to obtain a clustering score, and filter the initial overheating data according to the clustering score to construct an overheating distribution map, wherein the overheating distribution map contains the multi-zone overheating data class.

[0072] Furthermore, the intelligent allocation module 14 is also used to perform the following steps: perform cooling evaluation on the amorphous alloy transformer to obtain the extreme value of cooling performance; perform heat dissipation analysis on the amorphous alloy transformer based on the multi-zone overheating data class and set the priority of heat dissipation demand; allocate cooling resources according to the priority of heat dissipation demand combined with the extreme value of cooling performance and formulate a cooling parameter set; allocate the cooling parameter set according to the multiple regions to be analyzed to generate a partitioned cooling strategy, the partitioned cooling strategy including a cooling command sequence; execute the cooling command sequence to monitor the cooling effect in real time, and perform dynamic adjustment feedback according to the cooling effect to adaptively adjust the local overheating of the amorphous alloy transformer.

[0073] Furthermore, the intelligent allocation module 14 is also used to perform the following steps: traversing the overheating distribution map according to the multi-zone overheating data class to analyze and determine the overheating area distribution pattern; extracting multiple overheating areas based on the multi-zone overheating data class, and calculating the temperature rise of the multiple overheating areas based on the overheating area distribution pattern to obtain a regional temperature rise dataset, the regional temperature rise dataset including heat accumulation rate data and heat diffusion trend data; conducting an operational safety impact assessment on the amorphous alloy transformer according to the heat accumulation rate data and the heat diffusion trend data to generate an operational risk level; and performing a heat dissipation demand analysis based on the overheating area distribution pattern and the operational risk level, and setting the priority of the heat dissipation demand.

[0074] In summary, any step of the method described above can be stored as a computer instruction or program in an unrestricted computer memory, and can be called and identified by an unrestricted computer processor to implement any method in the embodiments of this application, without any additional restrictions.

[0075] Furthermore, the "first" or "second" mentioned above may not only represent a sequential relationship, but may also represent a specific concept, and / or refer to the individual or collective selection of multiple elements. Clearly, those skilled in the art can make various modifications and variations to this application without departing from its scope. Therefore, if such modifications and variations fall within the scope of this application and its equivalents, this application intends to include such modifications and variations.

Claims

1. A method for self-regulating local overheating in an amorphous alloy transformer, characterized in that, The method includes: Temperature data is collected in real time by deploying a multi-point distributed temperature sensing network on an amorphous alloy transformer to obtain a temperature dataset. The temperature dataset is used to identify local thermal anomalies by retrieving the thermosensitive characteristic parameters of the amorphous alloy material. A dynamic thermal field difference analysis module is introduced to predict the local heating trend by combining the local thermal anomaly data and generate thermal trend prediction results. Based on the thermal trend prediction results, the amorphous alloy transformer is subjected to zonal testing, and the thermal test results of multiple regions are generated for overheating cluster analysis to determine the overheating data class of multiple regions. Based on the multi-zone overheating data, temperature balance analysis is performed to formulate a zoned cooling strategy, and the zoned cooling strategy is executed for intelligent allocation.

2. The method for local overheating self-regulation of an amorphous alloy transformer as described in claim 1, characterized in that, The process of obtaining the thermosensitive characteristic parameters of amorphous alloy materials includes the following methods: Multiple temperature ranges and multiple magnetic flux densities are set, and thermal analysis is performed on the amorphous alloy material according to the multiple temperature ranges to obtain the first thermophysical parameter. Thermal analysis is then performed on the amorphous alloy material according to the multiple magnetic flux densities to obtain the second thermophysical parameter. Thermistor detection is performed based on the first thermophysical parameter and the second thermophysical parameter, and a thermistor parameter curve is constructed. The thermal characteristic parameters of the amorphous alloy material are determined by dynamically selecting parameters based on the magnetic flux density distribution information of the amorphous alloy transformer and the thermal parameter curve.

3. The method for local overheating self-adjustment of an amorphous alloy transformer as described in claim 1, characterized in that, The method involves retrieving the thermosensitive characteristic parameters of the amorphous alloy material to identify local thermal anomalies in the temperature dataset, introducing a dynamic thermal field difference analysis module, and combining the local thermal anomaly data to predict the local temperature rise trend, thereby generating a thermal trend prediction result. The method includes: The structural parameters of the amorphous alloy transformer are introduced to generate a mesh, and multiple regions to be analyzed are determined. Based on the thermosensitive characteristic parameters of the amorphous alloy material, the heat source temperature rise analysis is performed on the multiple regions to be analyzed, and multiple expected temperature values ​​are set. The temperature dataset is compared with the multiple regions to be analyzed and the multiple expected temperature values ​​to identify local thermal anomaly data; A three-dimensional thermal field of an amorphous alloy transformer is constructed. The temperature dataset is mapped to the three-dimensional thermal field to calculate the temperature gradient vector and determine the transformer heat flow parameters, which include heat flow direction parameters and heat flow intensity parameters. Based on the heat flow direction parameter and the heat flow intensity parameter, spatiotemporal variation difference analysis is performed on the multiple regions to be analyzed to generate thermal field difference analysis results; The local thermal anomaly data and the thermal field difference analysis results are combined in a time series to predict the local temperature rise trend, generating the thermal trend prediction result.

4. The method for local overheating self-adjustment of an amorphous alloy transformer as described in claim 3, characterized in that, The method involves comparing the temperature dataset with the multiple regions to be analyzed and the multiple expected temperature values ​​to identify local thermal anomalies. The temperature dataset is grouped according to the multiple regions to be analyzed to determine multiple temperature data subsets; The differences between the multiple temperature data subsets and the multiple expected temperature values ​​are calculated to obtain multiple absolute deviation values; Based on the multiple expected temperature values, deviation extreme value analysis is performed on the multiple regions to be analyzed, and a region allowable deviation threshold is set. The absolute deviation value is compared with the allowable deviation threshold for the region to identify multiple data points that exceed the limit. The multiple out-of-limit data points are continuously verified, and the verified out-of-limit data points are subjected to spatiotemporal clustering analysis to generate the local thermal anomaly data.

5. The method for local overheating self-regulation of an amorphous alloy transformer as described in claim 3, characterized in that, The method involves predicting the local temperature rise trend by combining the local thermal anomaly data and the thermal field difference analysis results according to a time series, thereby generating the thermal trend prediction result. Extract the anomalous temporal features of local thermal anomaly data according to the time series; Based on the thermal field difference analysis results, the spatial distribution of temperature is detected to obtain spatial characteristics; The abnormal temporal features and spatial features are integrated to construct a three-dimensional spatiotemporal feature matrix. The three-dimensional spatiotemporal feature matrix is ​​segmented to obtain multiple segmented features, which include trend features and periodic features. Deep learning is performed based on the trend characteristics and the periodic characteristics, and multivariate time series prediction is performed based on the learning results to generate local regional temperature change prediction results. Based on the predicted temperature changes in the local area, a thermal change assessment is performed to construct the predicted thermal trend results.

6. The method for local overheating self-adjustment of an amorphous alloy transformer as described in claim 3, characterized in that, Based on the aforementioned thermal trend prediction results, amorphous alloy transformers are subjected to zoned testing to generate thermal test results for multiple regions. The method includes: The thermal performance analysis is performed by traversing the multiple regions to be analyzed, and multiple thermal performance test indicators are set. Using a sliding time window, thermal behavior analysis is performed on multiple regions to be analyzed based on the thermal trend prediction results according to the multiple thermal performance detection indicators, and thermal behavior characteristics are generated. Based on the thermal behavior characteristics, the temperature distribution uniformity of the multiple regions to be analyzed is calculated, and multiple local hot spots are identified. Based on the multiple local hot spots, partition detection is performed to determine the thermal detection results of the multiple regions.

7. The method for local overheating self-adjustment of an amorphous alloy transformer as described in claim 6, characterized in that, Multiple regional thermal detection results are generated for overheating cluster analysis to determine the overheating data classes in multiple regions. The methods include: Based on the multiple local partition hotspots, thermal similarity calculation is performed to obtain the thermal similarity of multiple partitions; The thermal detection results of the multiple regions are initially grouped according to the thermal similarity of the multiple partitions to determine the number of clusters; Cluster stability is verified based on the number of clusters. Based on the verification results and the number of clusters, hierarchical clustering is performed on the thermal detection results of the multiple regions to generate initial overheating data. Clustering effect is evaluated based on the initial overheating data to obtain a clustering score. The initial overheating data is then filtered according to the clustering score to construct an overheating distribution map, which includes the multi-region overheating data classes.

8. The method for local overheating self-regulation of an amorphous alloy transformer as described in claim 7, characterized in that, Based on the aforementioned multi-zone overheating data, temperature balance analysis is performed to formulate zoned cooling strategies. These strategies are then executed for intelligent allocation. The method includes: Cooling performance of amorphous alloy transformers was evaluated to obtain extreme values ​​of cooling performance. Based on the multi-zone overheating data, heat dissipation analysis is performed on the amorphous alloy transformer, and heat dissipation requirements are prioritized. Cooling resources are allocated according to the priority of heat dissipation requirements and the extreme values ​​of cooling performance, and a set of cooling parameters is formulated. The cooling parameter set is allocated according to the multiple regions to be analyzed to generate a partitioned cooling strategy, which includes a cooling instruction sequence. The cooling command sequence is executed to monitor the cooling effect in real time, and dynamic adjustment feedback is performed based on the cooling effect to adaptively adjust the local overheating of the amorphous alloy transformer.

9. The method for local overheating self-adjustment of an amorphous alloy transformer as described in claim 8, characterized in that, Based on the aforementioned multi-zone overheating data, a heat dissipation analysis is performed on the amorphous alloy transformer, and heat dissipation requirements are prioritized. The method includes: The overheating distribution map is analyzed by traversing the overheating data of the multi-zone overheating region to determine the distribution pattern of the overheating region. Based on the multi-zone overheating data, multiple overheating regions are extracted. Based on the distribution pattern of the overheating regions, the temperature rise of the multiple overheating regions is calculated to obtain a regional temperature rise dataset. The regional temperature rise dataset includes heat accumulation rate data and heat diffusion trend data. Based on the heat accumulation rate data and the heat diffusion trend data, an operational safety impact assessment of the amorphous alloy transformer is conducted to generate an operational risk level. Based on the distribution pattern of the overheated area and the operational risk level, a heat dissipation demand analysis is conducted, and the priority of the heat dissipation demand is set.

10. A local overheat self-regulating system for an amorphous alloy transformer, characterized in that, The system is used to perform the local overheat self-regulation method for amorphous alloy transformers according to any one of claims 1 to 9, the system comprising: The temperature dataset acquisition module is used to collect temperature data in real time through a multi-point distributed temperature sensing network deployed on the amorphous alloy transformer to obtain a temperature dataset. The thermal trend prediction result generation module is used to retrieve the thermosensitive characteristic parameters of the amorphous alloy material to identify local thermal anomalies in the temperature dataset, introduce the dynamic thermal field difference analysis module, and combine the local thermal anomaly data to predict the local temperature rise trend and generate thermal trend prediction results. The multi-zone overheating data class determination module is used to perform zone detection on the amorphous alloy transformer according to the thermal trend prediction results, generate multiple zone thermal detection results for overheating cluster analysis, and determine the multi-zone overheating data class. The intelligent allocation module is used to perform temperature change balance analysis based on the multi-zone overheating data, formulate zoned cooling strategies, and execute the zoned cooling strategies for intelligent allocation.

Images