Method and device for active early warning of abnormal transformer temperature

Abstract

A method and device for active early warning of abnormal transformer temperature, the method comprising: obtaining the temperature of each temperature measuring point outside the transformer at each temperature measuring moment and the environmental state parameters of the transformer during the temperature measurement process; based on the temperature of each temperature measuring point at each temperature measuring moment, using the transformer internal hot spot temperature inversion model to obtain the internal hot spot position and temperature of the transformer at each temperature measuring moment as inverted temperature field spatiotemporal distribution data, and based on the environmental state parameters, using the power transformer multi-scale reduced order model to obtain simulated temperature field spatiotemporal distribution data; performing a significant difference analysis on the inverted temperature field spatiotemporal distribution data and the simulated temperature field spatiotemporal distribution data, and issuing an early warning of temperature anomaly if there is a significant difference. The method and device provided by the embodiment of the present invention can achieve differentiated active early warning of abnormal heating of transformers without the need for power outage modification of the operating transformers.

Description

Technical Field

[0001] The present invention relates to the technical field of power equipment status monitoring, and in particular to a method and device for active early warning of abnormal transformer temperature. Background Art

[0002] Transformer thermal defects are one of the important reasons for transformer failure. Currently, the methods for monitoring transformer thermal defects include oil chromatography, oil temperature meter and core grounding current, which are relatively mature.

[0003] Oil chromatography online monitoring can only reflect more serious transformer thermal defects. As oil and solid insulation gradually age and deteriorate, they decompose into very small amounts of gases (mainly including hydrogen H2, methane CH4, ethane C2H6, ethylene C2H4, acetylene C2H2, carbon monoxide CO, carbon dioxide CO2, and other gases). When an overheating fault or discharge fault occurs inside the transformer, or when the internal insulation becomes damp, the content of these gases will gradually increase, requiring a long period of mixing before being detected by the oil chromatography online monitoring device. In particular, it is impossible to identify early defects or abnormalities such as abnormal local heating of the transformer that do not generate new gas components.

[0004] It is difficult to effectively and accurately cover the temperature measurement and early warning needs of all operating transformers through methods such as oil temperature gauges, direct temperature measurement with optical fibers, and indirect calculation of winding temperature. First, the temperature probes of the oil temperature gauge are generally arranged at several fixed locations such as the top oil temperature and the bottom oil temperature, which cannot effectively and comprehensively reflect the overall distribution of the oil temperature; second, direct temperature measurement with optical fibers and gratings requires the transformer to be disassembled to lay optical fibers and gratings, which are generally arranged at key heat source locations such as windings. There are problems such as process reliability, and it cannot effectively cover the temperature measurement needs of operating transformers; third, the accuracy of indirect calculation of winding temperature is limited. Indirect calculation is performed through limited oil temperature gauge temperature measurement points, and the calculation accuracy is limited by empirical formulas, and differentiated early warnings cannot be performed.

[0005] It is difficult to respond to different types of thermal defects through online monitoring of the core grounding current. It can only identify the heat caused by defects such as multi-point grounding of the core and clamps, and cannot effectively identify other types of thermal defects. Summary of the Invention

[0006] In view of this, the present invention proposes a method and device for active early warning of abnormal transformer temperature, aiming to solve the problems of existing transformer temperature monitoring that cannot cover the entire transformer interior, is inconvenient to operate, has poor process reliability, and cannot provide differentiated early warning.

[0007] In the first aspect, an embodiment of the present invention provides a method for active early warning of abnormal temperature of a transformer, comprising: obtaining the temperature of each temperature measuring point outside the transformer at each temperature measuring moment and the environmental state parameters of the transformer during the temperature measurement process; according to the temperature of each temperature measuring point at each temperature measuring moment, using the internal hot spot temperature inversion model of the transformer, obtaining the internal hot spot position and temperature of the transformer at each temperature measuring moment as the inverted temperature field spatiotemporal distribution data, and according to the environmental state parameters, using the multi-scale reduced order model of the power transformer to obtain the simulated temperature field spatiotemporal distribution data; performing a significance difference analysis on the inverted temperature field spatiotemporal distribution data and the simulated temperature field spatiotemporal distribution data, and if there is a significant difference, issuing an early warning for the temperature anomaly.

[0008] Furthermore, the multi-scale reduced-order model of the power transformer is obtained in the following manner: based on the transformer properties, a full calculation model of the electromagnetic fluid temperature field of the power transformer is established; based on the full calculation model of the electromagnetic fluid temperature field of the power transformer, a multi-scale reduced-order model of the power transformer is constructed by using POD reduction.

[0009] Furthermore, after establishing a complete model for calculating the electromagnetic fluid temperature field of a power transformer based on transformer properties, the method further includes: obtaining historical temperature data of the outside of the transformer under different environmental state parameters, and correcting the complete model for calculating the electromagnetic fluid temperature field of the power transformer according to the historical temperature data, wherein the historical temperature data of the outside of the transformer under different environmental state parameters is collected by a temperature sensor array on the transformer casing at historical temperature measurement times.

[0010] Furthermore, based on the full model of the electromagnetic fluid temperature field calculation of the power transformer, a multi-scale reduced-order model of the power transformer is constructed by using the POD reduction method, which includes: performing simulation based on the multi-scale reduced-order model of the power transformer to obtain the spatiotemporal distribution data of temperature under different working conditions to construct a spatiotemporal distribution database of the power transformer temperature field.

[0011] Furthermore, simulation is performed based on the multi-scale reduced-order model of the power transformer to obtain the spatiotemporal distribution data of temperature under different working conditions, including: based on the multi-scale reduced-order model of the power transformer, the spatiotemporal distribution data of temperature under different working conditions between sequence parameters are obtained by discrete empirical interpolation method, and the spatiotemporal distribution data of temperature under different working conditions outside the sequence parameters are quickly obtained by adaptive snapshot.

[0012] Furthermore, after simulation is performed based on the multi-scale reduced-order model of the power transformer to obtain the temperature spatiotemporal distribution data under different working conditions, it includes: optimizing the deployment position of the temperature sensor array on the transformer casing based on the sensitivity of the temperature spatiotemporal distribution data under different working conditions.

[0013] Furthermore, the transformer internal hotspot temperature inversion model is obtained in the following manner: obtaining optimized historical temperature data of the outside of the transformer, wherein the optimized historical temperature data of the outside of the transformer is collected by deploying a temperature sensor array with optimized position on the transformer casing at the historical temperature measurement time; based on the optimized historical temperature data of the outside of the transformer, searching the temporal and spatial distribution database of the power transformer temperature field to obtain the hotspot position and temperature data inside the transformer at the historical temperature measurement time; using the optimized historical temperature data of the outside of the transformer and the searched hotspot position and temperature data inside the transformer, training the initialized transformer internal hotspot temperature inversion model, and obtaining the final transformer internal hotspot temperature inversion model.

[0014] Furthermore, based on the optimized historical temperature data outside the transformer, the hotspot position and temperature data inside the transformer at the historical temperature measurement moment are searched in the spatiotemporal distribution database of the power transformer temperature field, including: based on the optimized historical temperature data outside the transformer, the hotspot position and temperature data inside the transformer at the historical temperature measurement moment are obtained by local sensitive hash search in the spatiotemporal distribution database of the power transformer temperature field.

[0015] Furthermore, the transformer internal hot spot temperature inversion model adopts a Kriging model.

[0016] Furthermore, a significance difference analysis is performed on the inverted temperature field spatiotemporal distribution data and the simulated temperature field spatiotemporal distribution data. If there is a significant difference, an early warning is issued for temperature anomaly, including: using a t-test to perform a significance difference analysis on the inverted temperature field spatiotemporal distribution data and the simulated temperature field spatiotemporal distribution data. If the p-value is less than the significance level, it is determined that the internal temperature of the transformer is abnormal, and an early warning is issued for the hot spot location inside the transformer with the abnormal temperature.

[0017] In the second aspect, an embodiment of the present invention also provides a device for active early warning of abnormal temperature of a transformer, comprising: an acquisition unit for obtaining the temperature of each temperature measuring point outside the transformer at each temperature measurement moment and the environmental state parameters of the transformer during the temperature measurement process; a processing unit for obtaining the internal hot spot position and temperature of the transformer at each temperature measurement moment based on the temperature of each temperature measuring point at each temperature measurement moment, using the internal hot spot temperature inversion model of the transformer as the inverted temperature field spatiotemporal distribution data, and obtaining the simulated temperature field spatiotemporal distribution data based on the environmental state parameters using the multi-scale reduced order model of the power transformer; an early warning unit for performing a significance difference analysis on the inverted temperature field spatiotemporal distribution data and the simulated temperature field spatiotemporal distribution data, and issuing an early warning for the temperature anomaly if there is a significant difference.

[0018] Furthermore, the multi-scale reduced-order model of the power transformer is obtained in the following manner: based on the transformer properties, a full calculation model of the electromagnetic fluid temperature field of the power transformer is established; based on the full calculation model of the electromagnetic fluid temperature field of the power transformer, a multi-scale reduced-order model of the power transformer is constructed by using POD reduction.

[0019] Furthermore, after establishing a complete model for calculating the electromagnetic fluid temperature field of a power transformer based on transformer properties, the method further includes: obtaining historical temperature data of the outside of the transformer under different environmental state parameters, and correcting the complete model for calculating the electromagnetic fluid temperature field of the power transformer according to the historical temperature data, wherein the historical temperature data of the outside of the transformer under different environmental state parameters is collected by a temperature sensor array on the transformer casing at historical temperature measurement times.

[0020] Furthermore, based on the full model of the electromagnetic fluid temperature field calculation of the power transformer, a multi-scale reduced-order model of the power transformer is constructed by using the POD reduction method, which includes: performing simulation based on the multi-scale reduced-order model of the power transformer to obtain the spatiotemporal distribution data of temperature under different working conditions to construct a spatiotemporal distribution database of the power transformer temperature field.

[0021] Furthermore, simulation is performed based on the multi-scale reduced-order model of the power transformer to obtain the spatiotemporal distribution data of temperature under different working conditions, including: based on the multi-scale reduced-order model of the power transformer, the spatiotemporal distribution data of temperature under different working conditions between sequence parameters are obtained by discrete empirical interpolation method, and the spatiotemporal distribution data of temperature under different working conditions outside the sequence parameters are quickly obtained by adaptive snapshot.

[0022] Furthermore, after simulation is performed based on the multi-scale reduced-order model of the power transformer to obtain the temperature spatiotemporal distribution data under different working conditions, it includes: optimizing the deployment position of the temperature sensor array on the transformer casing based on the sensitivity of the temperature spatiotemporal distribution data under different working conditions.

[0023] Furthermore, the transformer internal hotspot temperature inversion model is obtained in the following manner: obtaining optimized historical temperature data of the outside of the transformer, wherein the optimized historical temperature data of the outside of the transformer is collected by deploying a temperature sensor array with optimized position on the transformer casing at the historical temperature measurement time; based on the optimized historical temperature data of the outside of the transformer, searching the temporal and spatial distribution database of the power transformer temperature field to obtain the hotspot position and temperature data inside the transformer at the historical temperature measurement time; using the optimized historical temperature data of the outside of the transformer and the searched hotspot position and temperature data inside the transformer, training the initialized transformer internal hotspot temperature inversion model, and obtaining the final transformer internal hotspot temperature inversion model.

[0024] Furthermore, based on the optimized historical temperature data outside the transformer, the hotspot position and temperature data inside the transformer at the historical temperature measurement moment are searched in the spatiotemporal distribution database of the power transformer temperature field, including: based on the optimized historical temperature data outside the transformer, the hotspot position and temperature data inside the transformer at the historical temperature measurement moment are obtained by local sensitive hash search in the spatiotemporal distribution database of the power transformer temperature field.

[0025] Furthermore, the transformer internal hot spot temperature inversion model adopts a Kriging model.

[0026] Furthermore, the early warning unit is also used to: use a t-test to perform a significant difference analysis on the inverted temperature field spatiotemporal distribution data and the simulated temperature field spatiotemporal distribution data; if the p-value is less than the significance level, it is determined that the internal temperature of the transformer is abnormal, and an early warning is issued for the hot spot location inside the transformer with abnormal temperature.

[0027] In a third aspect, an embodiment of the present invention further provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the methods provided in the above embodiments are implemented.

[0028] In a fourth aspect, an embodiment of the present invention further provides an electronic device comprising: a processor; a memory for storing executable instructions of the processor; the processor for reading the executable instructions from the memory and executing the executable instructions to implement the methods provided in the above embodiments.

[0029] The method and device for active early warning of abnormal transformer temperature provided by the embodiment of the present invention obtain the internal hot spot position and temperature of the transformer at each temperature measurement moment as the inverted temperature field spatiotemporal distribution data by using the transformer internal hot spot temperature inversion model according to the temperature of each temperature measurement point at each temperature measurement moment, and obtain the simulated temperature field spatiotemporal distribution data by using the power transformer multi-scale reduction model according to the environmental state parameters, and perform a significance difference analysis on the inverted temperature field spatiotemporal distribution data and the simulated temperature field spatiotemporal distribution data. If there is a significant difference, an early warning of the temperature anomaly is issued. There is no need to perform power outage modification on the operating transformer. Only distributed temperature measurement points need to be arranged on the transformer casing. At the same time, differentiated active early warning of abnormal heating of the transformer can be achieved, avoiding the further development of latent thermal defects and causing major equipment shutdown accidents, and effectively improving the level of intelligent operation and maintenance of equipment, applying equipment-level rapid simulation to equipment operation and maintenance, and realizing active perception of equipment status. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 An exemplary flow chart of a method for active early warning of abnormal transformer temperature according to an embodiment of the present invention is shown;

[0031] Figure 2 A schematic diagram of a multi-scale reduced-order model of a power transformer according to an embodiment of the present invention is shown;

[0032] Figure 3 A schematic diagram illustrating an optimized deployment of a non-intrusive temperature sensor array for a transformer housing according to an embodiment of the present invention is shown;

[0033] Figure 4 A schematic diagram of constructing a transformer internal hotspot temperature inversion model according to an embodiment of the present invention is shown;

[0034] Figure 5 A schematic diagram showing comparison of inverted temperature field spatiotemporal distribution data and simulated temperature field spatiotemporal distribution data according to an embodiment of the present invention is shown;

[0035] Figure 6 A data schematic diagram of a significant difference analysis according to an embodiment of the present invention is shown;

[0036] Figure 7 A schematic structural diagram of a device for active early warning of abnormal transformer temperature according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0037] Exemplary embodiments of the present invention will now be described with reference to the accompanying drawings. However, the present invention may be embodied in many different forms and is not limited to the embodiments described herein. These embodiments are provided to provide a thorough and complete disclosure of the present invention and to fully convey the scope of the present invention to those skilled in the art. The terminology used in the exemplary embodiments shown in the accompanying drawings is not intended to limit the present invention. In the accompanying drawings, identical elements are denoted by the same reference numerals.

[0038] Unless otherwise specified, the terms used herein (including technical terms) have the meanings commonly understood by those skilled in the art. In addition, it is understood that terms defined in commonly used dictionaries should be understood to have the same meanings as those in the context of the relevant fields, and should not be understood as idealized or overly formal meanings.

[0039] Figure 1 An exemplary flow chart of a method for active early warning of abnormal transformer temperature according to an embodiment of the present invention is shown.

[0040] like Figure 1 As shown, the method includes:

[0041] Step S101: obtaining the temperature of each temperature measurement point outside the transformer at each temperature measurement moment and the environmental state parameters of the transformer during the temperature measurement process.

[0042] Specifically, the temperature at each measurement point on the transformer's exterior is collected using an array of non-invasive temperature sensors placed on the transformer's housing. Distributed temperature measurement on the transformer's exterior can utilize platinum resistance sensors, with at least 24 sets distributed across key external components of the transformer, including the oil tank top, four side surfaces, riser, cooler, and three-phase bushings. The platinum resistance sensors have an accuracy of 0.1°C / ±1% of maximum temperature and a measurement range of -200°C to +850°C. These sensors are connected to a front-end acquisition device via a signal cable. The front-end acquisition device includes a signal amplification module, a synchronization clock module, and a signal transmission module.

[0043] Furthermore, the deployment position of the non-invasive temperature sensor array is pre-optimized based on the sensitivity of the temperature spatiotemporal distribution data under different working conditions.

[0044] Specifically, environmental state parameters include load, sunshine, rainfall, wind speed, and fan and oil pump state parameters. Specifically, load parameters refer to voltage and current, including voltage amplitude, frequency, and phase, and current amplitude, frequency, and phase. Sunshine parameters refer to sunshine angle, light radiation power, and duration of sunshine. Rainfall parameters refer to rainfall rate, rainfall intensity, rainfall duration, and rainwater temperature. Wind speed parameters refer to wind direction, wind speed, and wind flow duration. Fan parameters refer to fan speed, number of fan groups on, and fan operating time. Oil pump state parameters include oil pump flow rate and oil pump operating time.

[0045] Step S102: Based on the temperature of each temperature measurement point at each temperature measurement moment, the transformer internal hotspot temperature inversion model is used to obtain the transformer internal hotspot position and temperature at each temperature measurement moment as the inverted temperature field spatiotemporal distribution data, and based on the environmental state parameters, the power transformer multi-scale reduced order model is used to obtain the simulated temperature field spatiotemporal distribution data.

[0046] Furthermore, the multi-scale reduced-order model of the power transformer is obtained in the following way:

[0047] Based on the transformer properties, a complete calculation model of the electromagnetic fluid temperature field of the power transformer is established;

[0048] Based on the full calculation model of the electromagnetic fluid temperature field of the power transformer, a multi-scale reduced-order model of the power transformer is constructed using the POD reduction method.

[0049] Furthermore, based on the transformer properties, a complete calculation model of the electromagnetic fluid temperature field of the power transformer is established, including:

[0050] A complete calculation model of the electromagnetic fluid temperature field of the power transformer based on the finite element and finite volume hybrid method is established using the full size and full material properties of the power transformer to calculate the temporal and spatial distribution of the temperature field under typical operating conditions of the power transformer.

[0051] Specifically, the full size and full material properties of the power transformer include the size, installation position, and material properties of key components such as the power transformer oil tank, core, winding, oil storage cabinet, cooler, fan, oil pump, and bushing.

[0052] Specifically, the hybrid finite element and finite volume method involves the combined application of finite element and finite volume methods to solve the electromagnetic fluid temperature field of a power transformer. The power transformer is excited by the voltage and current of the power grid, and the finite element method is used to solve the electromagnetic field. Separately, the finite volume method is used to solve the fluid and temperature fields. The calculation of the electromagnetic fluid temperature field of the power transformer is achieved by transferring intermediate results between the different fields through weak coupling. Weak coupling refers to parameter transfer, and the electromagnetic and fluid thermal field results are obtained through iterative solution convergence.

[0053] Specifically, the temporal and spatial distribution of the power transformer temperature field refers to the temperature field distribution of all components of the power transformer, such as the iron core, winding, and oil tank, from the application of voltage and current excitation to the steady state.

[0054] Furthermore, based on the transformer properties, a full model for calculating the electromagnetic fluid temperature field of the power transformer is established, including:

[0055] The historical temperature data of the transformer exterior under different environmental state parameters are obtained, and the full calculation model of the electromagnetic fluid temperature field of the power transformer is corrected based on the historical temperature data. The historical temperature data of the transformer exterior under different environmental state parameters are collected by the temperature sensor array on the transformer casing at the historical temperature measurement time.

[0056] Specifically, load, sunshine, rainfall, wind speed, fan, and oil pump state parameters are collected as simulation conditions. Temperatures from distributed temperature measurement points outside the power transformer are simultaneously collected to revise the full model for calculating the electromagnetic fluid temperature field of the power transformer under typical operating conditions. The full model for calculating the electromagnetic fluid temperature field of the power transformer is then corrected using the temperatures from these distributed temperature measurement points by aligning the spatial coordinates of the distributed temperature measurement points with the spatial position of the power transformer's three-dimensional model. Adjustable parameters of the full model are optimized using a gradient descent method.

[0057] Furthermore, based on the full calculation model of the electromagnetic fluid temperature field of the power transformer, a multi-scale reduced-order model of the power transformer is constructed using the POD reduction method, including:

[0058] A multi-scale reduced-order model of the power transformer is constructed based on POD reduction, and the full model is calculated using the verified electromagnetic fluid temperature field of the power transformer to verify the multi-scale reduced-order model of the power transformer.

[0059] Specifically, it is difficult to effectively simulate the temperature field distribution of the power transformer under all working conditions using the full model for calculating the electromagnetic fluid temperature field of the power transformer. By using the POD reduction method to construct a multi-scale reduced-order model of the power transformer, the temperature field simulation calculation time can be effectively shortened by 95%, but there is a loss of accuracy, which is less than 1%.

[0060] Specifically, multi-scale refers to decomposing the entire calculation model of the electromagnetic fluid temperature field of the power transformer into different weakly coupled component models for iterative coupling simulation. Figure 2 FIG. 1 shows a schematic diagram of a multi-scale reduced-order model of a power transformer according to an embodiment of the present invention. Figure 2 As shown in the figure, the transformer can be divided into several parts, such as the transformer body, bushing, oil pillow, and cooler system. The weak coupling method is to transfer the simulation results of the component model to other component models and perform iterative simulation.

[0061] Specifically, POD order reduction refers to: intrinsic orthogonal decomposition order reduction. (1) Data matrix construction: The simulation results of the temperature field are organized into a data matrix according to the time step. Each column represents a time step, and each row represents a position or node of a physical field. (2) Singular value decomposition: The data matrix is ​​decomposed into three matrices. (3) Principal component selection: The principal components are selected from the decomposition. The principal components are a set of synthetic vectors, each of which represents the joint change pattern of multiple physical fields. (4) Dimensionality reduction: The high-dimensional data of multiple physical fields are reduced to a lower dimension using the selected POD principal components. This involves multiplying the data of multiple physical fields with the principal components to obtain a reduced-dimensional representation. (5) Reconstruction: The original multi-physical field data can be reconstructed using the principal components and weights after dimensionality reduction.

[0062] Specifically, the verification method uses the mean square error (MSE) and mean absolute error (MAE) indicators to quantify the difference between the model prediction and the actual observation; the time series rolling forecast verification method is used to ensure the prediction accuracy of the model at future time points; and the ANOVA test is used to check whether the model residuals obey the normal distribution.

[0063] Furthermore, based on the full calculation model of the electromagnetic fluid temperature field of the power transformer, a multi-scale reduced-order model of the power transformer is constructed using the POD reduction method, including:

[0064] Based on the multi-scale reduced-order model of the power transformer, simulation is performed to obtain the spatiotemporal distribution data of temperature under different working conditions to construct a spatiotemporal distribution database of the power transformer temperature field.

[0065] Specifically, the multi-scale reduced-order model for power transformers can simulate temperature field distributions at different spatial and temporal scales. It uses sequence parameters to simulate the spatiotemporal distribution of the power transformer temperature field under different operating conditions and construct a database of the spatiotemporal distribution of the temperature field. By serializing the input parameters, for example, simulating the ambient temperature from -20°C to 50°C in 0.5°C steps, the reduced-order model can obtain the spatiotemporal distribution of the temperature field under different operating conditions.

[0066] Furthermore, simulations were performed based on the multi-scale reduced-order model of the power transformer to obtain the spatiotemporal temperature distribution data under different operating conditions, including:

[0067] Based on the multi-scale reduced-order model of power transformer, the temporal and spatial distribution data of temperature under different working conditions between sequence parameters are obtained by discrete empirical interpolation method, and the temporal and spatial distribution data of temperature under different working conditions outside sequence parameters are quickly obtained by adaptive snapshot.

[0068] Specifically, for the temperature field distribution between sequence parameters, the discrete empirical difference method is used to fit the approximate solution; for the temperature field distribution outside the sequence parameters, the adaptive snapshot method is used to quickly solve the problem using the parameters, grids, and intermediate results saved at the key nodes.

[0069] Furthermore, simulations were performed based on the multi-scale reduced-order model of the power transformer to obtain the spatiotemporal distribution data of temperature under different operating conditions, including:

[0070] Based on the sensitivity of the temperature spatiotemporal distribution data under different working conditions, the deployment position of the temperature sensor array on the transformer casing is optimized.

[0071] Specifically, the initial deployment location of the distributed temperature measurement sensors of the power transformer was selected based on manual experience (the deployment location in the step of correcting the full model of the electromagnetic fluid temperature field calculation of the power transformer) and is not the optimal deployment solution. Since the subsequent use of distributed temperature measurement to invert the internal hotspot temperature is involved, the distributed temperature measurement sensors of the power transformer can be optimized before constructing the internal hotspot temperature inversion model of the transformer. Based on the rapid simulation results under different working conditions, the temperature gradient sensitivity {S1, S2, S3, ..., Sk} and the temperature change rate sensitivity {δ1, δ2, δ3, ..., δk} of each temperature point {P1, P2, P3, ..., Pk} on the outside of the transformer are used as the evaluation basis, where k is the number of temperature measurement points (temperature sensors).

[0072] Temperature gradient sensitivity = |θx / θT| + |θy / θT|;

[0073] Where T is the temperature field temperature, and x and y are the coordinates of the temperature sensor location.

[0074] Temperature change rate sensitivity = |Δt / ΔT|;

[0075] Where ΔT is the change in temperature of the temperature field, and Δt is the time interval.

[0076] Figure 3 FIG. 1 shows a schematic diagram of optimizing the deployment of a non-invasive temperature sensor array for a transformer housing according to an embodiment of the present invention. Figure 3 As shown in the figure, the square average of the temperature gradient sensitivity and the temperature change rate sensitivity is the largest under the optimized sensor position, which can effectively reflect the different working conditions of the power transformer.

[0077] Furthermore, the inversion model of the hot spot temperature inside the transformer is obtained in the following way:

[0078] Obtain optimized historical temperature data of the transformer exterior, wherein the optimized historical temperature data of the transformer exterior is collected by deploying a temperature sensor array with optimized positions on the transformer housing at historical temperature measurement times;

[0079] Based on the optimized historical temperature data of the transformer's exterior, the hotspot location and temperature data inside the transformer at the historical temperature measurement time are searched in the temporal and spatial distribution database of the power transformer's temperature field;

[0080] The optimized historical temperature data outside the transformer and the searched hotspot position and temperature data inside the transformer are used to train the initialized transformer internal hotspot temperature inversion model to obtain the final transformer internal hotspot temperature inversion model.

[0081] Furthermore, based on the optimized historical temperature data of the transformer's exterior, the hotspot location and temperature data inside the transformer at the historical temperature measurement time are searched in the temporal and spatial distribution database of the power transformer's temperature field, including:

[0082] According to the optimized historical temperature data of the transformer outside, the hotspot location and temperature data inside the transformer at the historical temperature measurement moment are obtained through local sensitive hash search in the spatiotemporal distribution database of the power transformer temperature field.

[0083] Furthermore, the Kriging model is used as the inversion model for the hot spot temperature inside the transformer.

[0084] Specifically, based on the spatiotemporal distribution database of the power transformer temperature field under all working conditions established in the previous steps, the hotspot position and hotspot temperature inside the transformer are quickly searched by deploying a distributed temperature sensor array with optimized position.

[0085] Specifically, the similarity between the temperature of the distributed temperature sensor array and the temperature of the same point in the spatiotemporal distribution database of the temperature field is evaluated based on Local Sensitivity Hashing (LSH). (1) Hash function family: The hash function maps the input vector to a hash bucket. (2) Local sensitivity: That is, similar vectors have a higher probability of falling together in certain hash buckets, while dissimilar vectors have a lower probability. (3) Multiple hash functions: In order to increase local sensitivity, multiple hash functions are usually used to generate multiple hash buckets. When querying, it is necessary to search in multiple hash buckets to obtain potential similar vectors. Query processing: When executing a query, LSH will map the query vector to multiple hash buckets using the same hash function family. Then, LSH will search in these hash buckets to find potential similar vectors. These potential similar vectors may require further verification to ensure that they are true nearest neighbors.

[0086] Specifically, Figure 4 FIG. 1 shows a schematic diagram of constructing a transformer internal hotspot temperature inversion model according to an embodiment of the present invention. Figure 4 As shown in FIG, the temperature distribution in the spatiotemporal distribution database of the power transformer temperature field is searched by local sensitive hashing, and the hotspot position and hotspot temperature under the distribution are output.

[0087] Specifically, the inversion model is trained based on the temperature, hotspot location, and hotspot temperature of the distributed temperature sensor array under all operating conditions. The inversion model uses the Kriging method, (1) Semivariogram modeling: Determine the semivariogram to describe the temperature correlation between different locations inside the transformer. The selection and parameterization of the semivariogram need to consider the temperature variation characteristics inside the transformer. Generally, the semivariogram can be spherical, exponential, Gaussian, etc., depending on the spatial correlation of the data. (2) Semivariogram parameter estimation: Use known observation data to estimate the parameters of the semivariogram. This can be done through statistical methods such as maximum likelihood estimation. (3) Kriging model establishment: Based on the estimated semivariogram and parameters, a Kriging model is constructed, which can be used to estimate the temperature value at an unknown location inside the transformer. (4) Three-dimensional temperature field interpolation: Use the Kriging model to interpolate the three-dimensional temperature field at an unknown location inside the transformer. This can be achieved by applying the Kriging model on a three-dimensional coordinate grid. (5) Uncertainty estimation: Kriging can also provide uncertainty estimates for the interpolation results, including the variance or confidence interval of the temperature value. This can help determine the reliability of the estimate. (6) Model validation: The performance of the kriging model needs to be verified, usually by comparing with independent measurement data or using cross-validation to assess the accuracy of the interpolation model.

[0088] Step S103: performing a significance difference analysis on the inverted temperature field spatiotemporal distribution data and the simulated temperature field spatiotemporal distribution data. If there is a significant difference, an early warning of temperature anomaly is issued.

[0089] Furthermore, the inverted temperature field spatiotemporal distribution data and the simulated temperature field spatiotemporal distribution data are analyzed for significant differences. If there is a significant difference, an early warning of temperature anomaly is issued, including:

[0090] The t-test is used to analyze the significant differences between the inverted temperature field spatiotemporal distribution data and the simulated temperature field spatiotemporal distribution data. If the p-value is less than the significance level, the internal temperature of the transformer is judged to be abnormal, and an early warning is issued for the hot spot location inside the transformer with abnormal temperature.

[0091] Specifically, the inverted and simulated temperature field temporal and spatial distribution data for each transformer temperature measurement point are used to generate an inverted temperature time distribution curve and a simulated temperature time distribution curve. For the same transformer temperature measurement point, the inverted and simulated temperature time distribution curves form a set of curves, which are then analyzed for significant differences to provide a temperature anomaly warning. Figure 5 A schematic diagram of comparing inverted temperature field spatiotemporal distribution data with simulated temperature field spatiotemporal distribution data according to an embodiment of the present invention is shown. Figure 6 Schematic diagram of data of significant difference analysis according to an embodiment of the present invention is shown. Figure 5 and Figure 6 As shown in the figure, by independently testing the similarity between the inverted temperature-time distribution curve and the simulated temperature-time distribution curve of the same temperature measurement point over a period of time, it is determined whether the temperature has shifted abnormally. The method adopted is to use the independent sample t-test to test whether there is a significant difference in the means of the two groups of temperature series. If the p-value is less than the significance level (usually 0.05), it means that there is a significant difference.

[0092] Specifically, a boxplot is used to compare the data distribution of different temperature series. The boxplot can display the median, upper and lower quartiles, and outliers of the data, allowing you to quickly locate the spatial location of temperature differences.

[0093] Specifically, the diagnosed transformer temperature anomaly can be comprehensively compared with other non-homologous thermal defect anomaly diagnoses, including but not limited to: oil chromatography analysis, core grounding current and other online monitoring.

[0094] The method for proactively warning of abnormal transformer temperatures provided by this embodiment can play a significant role in improving the operational reliability of large-capacity transformers. Taking six 500kV transformers as an example, applying the method provided by this embodiment reduces power outages and repairs on each 500kV transmission line by five hours per year, with a transmission capacity of 5000MVA. The resulting annual losses are: (5 x 500) x 6 x 0.6 = 90 million yuan.

[0095] In the above embodiment, the hot spot position and temperature of the transformer at each temperature measurement moment are obtained by using the transformer internal hot spot temperature inversion model according to the temperature of each temperature measurement point at each temperature measurement moment as the inverted temperature field spatiotemporal distribution data, and the simulated temperature field spatiotemporal distribution data are obtained by using the power transformer multi-scale reduction model according to the environmental state parameters, and the inverted temperature field spatiotemporal distribution data and the simulated temperature field spatiotemporal distribution data are analyzed for significance differences. If there is a significant difference, an early warning of temperature anomaly is issued, and there is no need to perform power outage modification on the operating transformer. Only distributed temperature measurement points need to be arranged on the transformer casing. At the same time, differentiated active warning of early abnormal heating of the transformer can be realized, thereby avoiding the further development of latent thermal defects and causing major equipment shutdown accidents, and effectively improving the level of intelligent operation and maintenance of equipment, applying equipment-level rapid simulation to equipment operation and maintenance, and realizing active perception of equipment status.

[0096] Figure 7 A schematic structural diagram of a device for active early warning of abnormal transformer temperature according to an embodiment of the present invention is shown.

[0097] like Figure 7 As shown, the device comprises:

[0098] An obtaining unit 701 is configured to obtain the temperature of each temperature measuring point outside the transformer at each temperature measuring moment and the environmental state parameters of the transformer during the temperature measuring process;

[0099] Processing unit 702 is configured to obtain the location and temperature of the hot spot inside the transformer at each temperature measurement moment using the transformer internal hot spot temperature inversion model based on the temperature of each temperature measurement point at each temperature measurement moment as inverted temperature field spatiotemporal distribution data, and to obtain simulated temperature field spatiotemporal distribution data using the power transformer multi-scale reduced order model based on the environmental state parameters;

[0100] The early warning unit 703 is used to perform a significant difference analysis on the inverted temperature field spatiotemporal distribution data and the simulated temperature field spatiotemporal distribution data, and issue an early warning of temperature anomaly if there is a significant difference.

[0101] Furthermore, the multi-scale reduced-order model of the power transformer is obtained in the following way:

[0102] Based on the transformer properties, a complete calculation model of the electromagnetic fluid temperature field of the power transformer is established;

[0103] Based on the full calculation model of the electromagnetic fluid temperature field of the power transformer, a multi-scale reduced-order model of the power transformer is constructed using the POD reduction method.

[0104] Furthermore, based on the transformer properties, a full model for calculating the electromagnetic fluid temperature field of the power transformer is established, including:

[0105] The historical temperature data of the transformer exterior under different environmental state parameters are obtained, and the full calculation model of the electromagnetic fluid temperature field of the power transformer is corrected based on the historical temperature data. The historical temperature data of the transformer exterior under different environmental state parameters are collected by the temperature sensor array on the transformer casing at the historical temperature measurement time.

[0106] Furthermore, based on the full calculation model of the electromagnetic fluid temperature field of the power transformer, a multi-scale reduced-order model of the power transformer is constructed using the POD reduction method, including:

[0107] Based on the multi-scale reduced-order model of the power transformer, simulation is performed to obtain the spatiotemporal distribution data of temperature under different working conditions to construct a spatiotemporal distribution database of the power transformer temperature field.

[0108] Furthermore, simulations were performed based on the multi-scale reduced-order model of the power transformer to obtain the spatiotemporal temperature distribution data under different operating conditions, including:

[0109] Based on the multi-scale reduced-order model of power transformer, the temporal and spatial distribution data of temperature under different working conditions between sequence parameters are obtained by discrete empirical interpolation method, and the temporal and spatial distribution data of temperature under different working conditions outside sequence parameters are quickly obtained by adaptive snapshot.

[0110] Furthermore, simulations were performed based on the multi-scale reduced-order model of the power transformer to obtain the spatiotemporal distribution data of temperature under different operating conditions, including:

[0111] Based on the sensitivity of the temperature spatiotemporal distribution data under different working conditions, the deployment position of the temperature sensor array on the transformer casing is optimized.

[0112] Furthermore, the inversion model of the hot spot temperature inside the transformer is obtained in the following way:

[0113] Obtain optimized historical temperature data of the transformer exterior, wherein the optimized historical temperature data of the transformer exterior is collected by deploying a temperature sensor array with optimized positions on the transformer housing at historical temperature measurement times;

[0114] Based on the optimized historical temperature data of the transformer's exterior, the hotspot location and temperature data inside the transformer at the historical temperature measurement time are searched in the temporal and spatial distribution database of the power transformer's temperature field;

[0115] The optimized historical temperature data outside the transformer and the searched hotspot position and temperature data inside the transformer are used to train the initialized transformer internal hotspot temperature inversion model to obtain the final transformer internal hotspot temperature inversion model.

[0116] Furthermore, based on the optimized historical temperature data of the transformer's exterior, the hotspot location and temperature data inside the transformer at the historical temperature measurement time are searched in the temporal and spatial distribution database of the power transformer's temperature field, including:

[0117] According to the optimized historical temperature data of the transformer outside, the hotspot location and temperature data inside the transformer at the historical temperature measurement time are obtained through local sensitive hash search in the spatiotemporal distribution database of the power transformer temperature field.

[0118] Furthermore, the Kriging model is used as the inversion model for the hot spot temperature inside the transformer.

[0119] Furthermore, the early warning unit 703 is further configured to:

[0120] The t-test is used to analyze the significant differences between the inverted temperature field spatiotemporal distribution data and the simulated temperature field spatiotemporal distribution data. If the p-value is less than the significance level, the internal temperature of the transformer is judged to be abnormal, and an early warning is issued for the hot spot location inside the transformer with abnormal temperature.

[0121] In the above embodiment, the hot spot position and temperature of the transformer at each temperature measurement moment are obtained by using the transformer internal hot spot temperature inversion model according to the temperature of each temperature measurement point at each temperature measurement moment as the inverted temperature field spatiotemporal distribution data, and the simulated temperature field spatiotemporal distribution data are obtained by using the power transformer multi-scale reduction model according to the environmental state parameters, and the inverted temperature field spatiotemporal distribution data and the simulated temperature field spatiotemporal distribution data are analyzed for significance differences. If there is a significant difference, an early warning of temperature anomaly is issued, and there is no need to perform power outage modification on the operating transformer. Only distributed temperature measurement points need to be arranged on the transformer casing. At the same time, differentiated active warning of early abnormal heating of the transformer can be realized, thereby avoiding the further development of latent thermal defects and causing major equipment shutdown accidents, and effectively improving the level of intelligent operation and maintenance of equipment, applying equipment-level rapid simulation to equipment operation and maintenance, and realizing active perception of equipment status.

[0122] It should be noted that the apparatus provided in the above embodiments is merely illustrated by the division of the above functional modules when implementing its functions. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. In addition, the apparatus and method embodiments provided in the above embodiments are based on the same concept. The specific implementation process is detailed in the method embodiment and will not be repeated here.

[0123] An embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the method for active early warning of abnormal transformer temperature provided in the above embodiments is implemented.

[0124] An embodiment of the present invention also provides an electronic device, comprising: a processor; a memory for storing processor executable instructions; the processor is configured to read the executable instructions from the memory and execute the instructions to implement the method for active warning of abnormal transformer temperature provided in the above embodiments.

[0125] The invention has been described above with reference to a few embodiments. However, it is readily apparent to a person skilled in the art that other embodiments than the ones disclosed above are equally within the scope of the invention, as defined by the appended patent claims.

[0126] Generally, all terms used in the claims are to be interpreted according to their ordinary meaning in the technical field, unless explicitly defined otherwise herein. All references to "a / the [means, component, etc.]" are to be interpreted openly as referring to at least one instance of the means, component, etc., unless explicitly stated otherwise. The steps of any method disclosed herein do not necessarily need to be performed in the exact order disclosed, unless explicitly stated otherwise.

[0127] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0128] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0129] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0130] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0131] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered by the scope of protection of the claims of the present invention.

Claims

1. A method for active early warning of abnormal transformer temperature, characterized in that: include: Obtain the temperature of each temperature measurement point outside the transformer at each temperature measurement moment and the environmental state parameters of the transformer during the temperature measurement process; According to the temperature of each temperature measurement point at each temperature measurement moment, the transformer internal hot spot temperature inversion model is used to obtain the transformer internal hot spot position and temperature at each temperature measurement moment as the inverted temperature field spatiotemporal distribution data, and according to the environmental state parameters, the power transformer multi-scale reduced order model is used to obtain the simulated temperature field spatiotemporal distribution data; wherein, the power transformer multi-scale reduced order model is obtained by the following method: Based on the transformer properties, a complete calculation model of the electromagnetic fluid temperature field of the power transformer is established; Based on the full calculation model of the electromagnetic fluid temperature field of the power transformer, a multi-scale reduced-order model of the power transformer is constructed by adopting the POD reduction method; Simulation is performed based on the multi-scale reduced-order model of the power transformer to obtain the spatiotemporal distribution data of temperature under different working conditions, so as to construct a spatiotemporal distribution database of the power transformer temperature field; The inverted temperature field spatiotemporal distribution data and the simulated temperature field spatiotemporal distribution data are analyzed for significant differences. If there is a significant difference, an early warning of temperature anomaly is issued.

2. The method according to claim 1, characterized in that After establishing a full model for calculating the electromagnetic fluid temperature field of a power transformer based on transformer properties, the following steps are included: Obtain historical temperature data of the transformer exterior under different environmental state parameters, and correct the full calculation model of the electromagnetic fluid temperature field of the power transformer based on the historical temperature data, wherein the historical temperature data of the transformer exterior under different environmental state parameters is collected by a temperature sensor array on the transformer housing at historical temperature measurement moments.

3. The method according to claim 1, characterized in that Based on the multi-scale reduced-order model of the power transformer, simulation is performed to obtain the spatiotemporal distribution data of temperature under different working conditions, including: Based on the multi-scale reduced-order model of the power transformer, the temperature spatiotemporal distribution data under different working conditions between the sequence parameters are obtained by discrete empirical interpolation method, and the temperature spatiotemporal distribution data under different working conditions outside the sequence parameters are quickly obtained by adaptive snapshot.

4. The method according to claim 1, wherein After performing simulation based on the multi-scale reduced-order model of the power transformer and obtaining the temporal and spatial distribution data of temperature under different working conditions, the following steps are performed: Based on the sensitivity of the temperature spatiotemporal distribution data under different working conditions, the deployment position of the temperature sensor array on the transformer casing is optimized.

5. The method according to claim 4, characterized in that The transformer internal hot spot temperature inversion model is obtained in the following way: Obtaining optimized historical temperature data of the transformer exterior, wherein the optimized historical temperature data of the transformer exterior is collected by deploying a temperature sensor array with optimized positions on the transformer housing at historical temperature measurement times; According to the optimized historical temperature data of the transformer exterior, searching the temporal and spatial distribution database of the power transformer temperature field to obtain the hotspot position and temperature data inside the transformer at the historical temperature measurement time; The optimized historical temperature data outside the transformer and the searched hotspot position and temperature data inside the transformer are used to train an initialized transformer internal hotspot temperature inversion model to obtain a final transformer internal hotspot temperature inversion model.

6. The method according to claim 5, characterized in that According to the optimized historical temperature data of the transformer exterior, the hotspot location and temperature data of the transformer interior at the historical temperature measurement time are searched in the temporal and spatial distribution database of the power transformer temperature field, including: According to the optimized historical temperature data of the transformer exterior, the hotspot position and temperature data of the transformer interior at the historical temperature measurement moment are obtained by local sensitive hash search in the temporal and spatial distribution database of the power transformer temperature field.

7. The method according to claim 5, characterized in that The transformer internal hot spot temperature inversion model adopts the Kriging model.

8. The method according to claim 1, characterized in that Performing a significant difference analysis on the inverted temperature field spatiotemporal distribution data and the simulated temperature field spatiotemporal distribution data, and issuing an early warning of temperature anomaly if there is a significant difference, including: The t-test is used to perform a significant difference analysis on the inverted temperature field spatiotemporal distribution data and the simulated temperature field spatiotemporal distribution data. If the p-value is less than the significance level, it is determined that the internal temperature of the transformer is abnormal, and an early warning is issued for the hot spot location inside the transformer with abnormal temperature.

9. A device for active early warning of abnormal transformer temperature, characterized in that: include: An acquisition unit is used to obtain the temperature of each temperature measurement point outside the transformer at each temperature measurement moment and the environmental state parameters of the transformer during the temperature measurement process; The processing unit is configured to obtain the position and temperature of the hot spot inside the transformer at each temperature measurement moment using the transformer internal hot spot temperature inversion model according to the temperature of each temperature measurement point at each temperature measurement moment as the inverted temperature field spatiotemporal distribution data, and to obtain the simulated temperature field spatiotemporal distribution data using the power transformer multi-scale reduced order model according to the environmental state parameters; wherein the power transformer multi-scale reduced order model is obtained in the following manner: Based on the transformer properties, a complete calculation model of the electromagnetic fluid temperature field of the power transformer is established; Based on the full calculation model of the electromagnetic fluid temperature field of the power transformer, a multi-scale reduced-order model of the power transformer is constructed by adopting the POD reduction method; Simulation is performed based on the multi-scale reduced-order model of the power transformer to obtain the spatiotemporal distribution data of temperature under different working conditions, so as to construct a spatiotemporal distribution database of the power transformer temperature field; The early warning unit is used to perform a significant difference analysis on the inverted temperature field spatiotemporal distribution data and the simulated temperature field spatiotemporal distribution data, and issue an early warning of temperature anomaly if there is a significant difference.

10. The device according to claim 9, characterized in that After establishing a full model for calculating the electromagnetic fluid temperature field of a power transformer based on transformer properties, the following steps are included: Obtain historical temperature data of the transformer exterior under different environmental state parameters, and correct the full calculation model of the electromagnetic fluid temperature field of the power transformer based on the historical temperature data, wherein the historical temperature data of the transformer exterior under different environmental state parameters is collected by a temperature sensor array on the transformer housing at historical temperature measurement moments.

11. The device according to claim 9, characterized in that Based on the multi-scale reduced-order model of the power transformer, simulation is performed to obtain the spatiotemporal distribution data of temperature under different working conditions, including: Based on the multi-scale reduced-order model of the power transformer, the temperature spatiotemporal distribution data under different working conditions between the sequence parameters are obtained by discrete empirical interpolation method, and the temperature spatiotemporal distribution data under different working conditions outside the sequence parameters are quickly obtained by adaptive snapshot.

12. The device according to claim 9, characterized in that After performing simulation based on the multi-scale reduced-order model of the power transformer and obtaining the temporal and spatial distribution data of temperature under different working conditions, the following steps are performed: Based on the sensitivity of the temperature spatiotemporal distribution data under different working conditions, the deployment position of the temperature sensor array on the transformer casing is optimized.

13. The device according to claim 12, characterized in that The transformer internal hot spot temperature inversion model is obtained in the following way: Obtaining optimized historical temperature data of the transformer exterior, wherein the optimized historical temperature data of the transformer exterior is collected by deploying a temperature sensor array with optimized positions on the transformer housing at historical temperature measurement times; According to the optimized historical temperature data of the transformer exterior, searching the temporal and spatial distribution database of the power transformer temperature field to obtain the hotspot position and temperature data inside the transformer at the historical temperature measurement time; The optimized historical temperature data outside the transformer and the searched hotspot position and temperature data inside the transformer are used to train an initialized transformer internal hotspot temperature inversion model to obtain a final transformer internal hotspot temperature inversion model.

14. The device according to claim 13, characterized in that According to the optimized historical temperature data of the transformer exterior, the hotspot location and temperature data of the transformer interior at the historical temperature measurement time are searched in the temporal and spatial distribution database of the power transformer temperature field, including: According to the optimized historical temperature data of the transformer exterior, the hotspot position and temperature data of the transformer interior at the historical temperature measurement moment are obtained by local sensitive hash search in the temporal and spatial distribution database of the power transformer temperature field.

15. The device according to claim 13, characterized in that The transformer internal hot spot temperature inversion model adopts the Kriging model.

16. The device according to claim 9, characterized in that The early warning unit is further used to: The t-test is used to perform a significant difference analysis on the inverted temperature field spatiotemporal distribution data and the simulated temperature field spatiotemporal distribution data. If the p-value is less than the significance level, it is determined that the internal temperature of the transformer is abnormal, and an early warning is issued for the hot spot location inside the transformer with abnormal temperature.

17. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 8 is implemented.

18. An electronic device comprising: processor; a memory for storing instructions executable by the processor; The processor is configured to read the executable instructions from the memory and execute the executable instructions to implement the method according to any one of claims 1 to 8.

Images