1812results about "Transformers testing" patented technology

The invention relates to a transformer intelligent monitoring analysis method, system and device and a medium, and is applied to the technical field of transformer monitoring, and the method comprises the steps: obtaining historical fault information, monitoring data of a to-be-monitored transformer, and an operation state of the to-be-monitored transformer, the historical fault information comprises a historical fault transformer, historical abnormal data, a historical abnormal type, a historical operation state and a historical fault type; determining a threshold range based on the historical fault information and a threshold optimization model; determining abnormal information based on the monitoring data and the threshold range, wherein the abnormal information comprises abnormal monitoring data, an abnormal type and a predicted abnormal type; searching a fault type from a preset fault database according to the exception type and the predicted exception type to obtain fault information; and determining a fault processing strategy based on the fault information. The method and the device have the effect of improving the accuracy of transformer monitoring.

A transformer fault diagnosis method based on vibration mechanisms of an iron core and a winding. The method comprises: performing signal processing and feature extraction by means of calling simulation results of an iron core and a winding; and establishing a transformer fault diagnosis model. Thus, the accurate diagnosis of a transformer fault is realized. A conventional fault diagnosis method is improved, and chaotic mapping and an opposition-based learning strategy are introduced to optimize the diagnosis model, thereby improving the accuracy of diagnosis. Compared with the conventional method, a fault type and a fault degree of a transformer can be more accurately determined, so as to facilitate an engineering technician in maintenance, thereby improving the operation reliability of the transformer. In addition, a transformer fault diagnosis system based on vibration mechanisms of an iron core and a winding, and a computing device and a computer-readable storage medium are further provided.

A diagnosis system is provided for diagnosing the partial discharge faults in a power transformer. The system includes a feature extractor, a fault classifier and a domain discriminator and whose parameters are optimized by minimizing a combination of feature loss, classifier loss and domain discrepancy discriminator loss. The feature loss is a weighted sum of cosine similarity distance loss, batch-based instance separation loss, and batch-based feature decorrelation loss. The measured applied voltages and partial discharge voltages are pre-processing using sliding window method to represent the each partial discharge event as a series of statistics based moments for a set of overlapped time blocks within the duration of fault event, wherein statistics based moments include an average value of applied voltage for representing the applied voltage magnitude and fault event occurring moment for each time block, and the mean, standard deviation, Kurtosis, and Skewness of partial discharge voltages for representing the variations of partial discharging within each time block.

The invention discloses a distribution transformer fault diagnosis method based on full-band multi-modal data fusion. The method comprises the following steps: collecting different feature maps of a transformer; performing feature extraction and weighted summation on the collected different feature maps to obtain a multi-source fusion feature vector; performing data preprocessing on the multi-source fusion feature vector, and performing normalization processing after data preprocessing; training an improved probabilistic neural network by using the normalized multi-source fusion features; optimizing the polynomial order of the trained improved probabilistic neural network by adopting an improved parrot optimization algorithm; and carrying out transformer fault diagnosis by adopting the optimized improved probabilistic neural network. According to the method, the improved parrot-shaped optimization algorithm is adopted, the foraging behavior of the parrot-shaped algorithm is simulated, the polynomial order of the improved probabilistic neural network is optimized by effectively combining global search and local search strategies, and the accuracy of transformer data fusion diagnosis and the generalization ability of the probabilistic neural network are improved.

The invention belongs to the technical field of data exception identification, and particularly relates to a main transformer operation multi-mode data exception identification method, which comprises the following steps: step 1, data acquisition: collecting various data in operation of a main transformer, including current, voltage, temperature, vibration and noise; step 2, data preprocessing: cleaning and sorting the collected data, including noise removal, missing value filling and data smoothing operation; step 3, feature extraction: extracting key features, including statistical features and frequency domain features, from the preprocessed data, and converting the original data into a form for analysis; step 4, modeling and training: utilizing a machine learning or deep learning model to train the features, wherein the model comprises a decision tree and a support vector machine; according to the invention, through a multi-level and multi-dimensional anomaly detection and processing mechanism, the safety and reliability of the operation of the transformer are improved, and the efficiency and cost of operation and maintenance management are optimized.

The invention discloses a method and a device for calculating an electromagnetic-heat flow field under a turn-to-turn short circuit fault of a converter transformer. The method comprises the following steps: establishing a converter transformer multi-physical field numerical calculation theoretical model based on circuit-electromagnetic-heat flow coupling; based on the multi-physical field numerical calculation theoretical model, establishing a converter transformer electromagnetic-heat flow field model under the turn-to-turn short circuit fault; setting turn-to-turn short circuit faults at different positions in the electromagnetic-heat flow field model of the converter transformer; and performing winding current waveform, voltage waveform and electromagnetic field analysis and heat flow field analysis of the converter transformer under the set turn-to-turn short circuit fault.

The invention discloses a coal mine underground power supply intelligent monitoring system which comprises the following steps: after a corresponding set monitoring time interval is reached, acquiring operation parameters of a corresponding transformer in a set time period; the electrical parameters, the temperature parameters and the sound parameters of the transformers can be comprehensively monitored, and the limitation that only basic electrical parameters are concerned in a traditional monitoring mode is overcome. Therefore, the health condition of the transformer can be evaluated more accurately, a visual quantitative index is provided for maintenance personnel, and potential problems can be found in time.

The invention relates to the technical field of power equipment state monitoring, in particular to a transformer iron core alternating current and direct current mixed current on-line monitoring method and system based on a magnetoelectric coupling sensor, and the method comprises the steps: monitoring the change of magnetic flux generated by the vibration of an iron core in a transformer in real time through a magnetoelectric coupling sensor array; the magnetic flux change is converted into an original electric signal, the original electric signal is processed, an alternating current signal component and a direct current signal component in the mixed current are extracted, and historical operation data and real-time environment parameters of the alternating current signal component and the direct current signal component are input into a deep recurrent neural network to be trained; and the deep recurrent neural network outputs the enhanced stray current and the direct current signal after the environmental influence is eliminated after iterative training. According to the invention, high-precision prediction of the temperature of the iron core is realized by using a method of combining a deep recurrent neural network and adaptive decomposition; and the neural network is trained by using the dynamic data, so that the dynamic processing capability of the network is enhanced and the operation reliability of the transformer can be effectively improved.

The invention belongs to the field of power equipment monitoring, and discloses an electrical signal acquisition system based on intelligent dynamic sampling, which comprises a signal acquisition module, a signal conditioning module, an abrupt change detection module, an impedance calculation module and an upper computer, and is characterized in that the signal acquisition module comprises a first ADC module, a second ADC module, a first MCU module and a second MCU module. Short-circuit impedance real-time calculation and waveform abrupt change detection are adopted for cooperative judgment, and when equipment operates stably, the sampling rate of the system is kept low; when impedance changes or waveform changes suddenly, the sampling rate is improved automatically, important signal changes are recorded accurately, and waveform abrupt change detection is carried out by adopting an edge detection algorithm improved based on an FIR (Finite Impulse Response) filter. The system ensures the time sequence consistency of data under different sampling frequencies through multi-channel data synchronization and timestamp management, and meanwhile, smooth transition is performed by applying an interpolation algorithm during sampling switching, so that the continuity of a data sequence is ensured. The data redundancy is reduced; and the integrity of key information is ensured.

The invention relates to the technical field of magnetic bias monitoring, in particular to a real-time monitoring method and system for direct-current magnetic bias of a transformer. The method comprises the following steps: acquiring a primary side current signal and an iron core vibration signal of the transformer; analyzing the zero-flux closed-loop characteristic of the primary side current signal of the transformer and performing temperature drift elimination on the primary side current signal of the transformer to generate a primary side optimization signal of the transformer; performing magnetostrictive vibration noise separation on the iron core vibration signal to generate an iron core vibration separation signal; dynamically filtering a power frequency fundamental wave of the primary side optimization signal of the transformer to extract a pure direct current component; and performing wavelet packet decomposition on the iron core vibration separation signal, and performing magnetostriction characteristic spectrum extraction on the decomposed iron core vibration separation signal to obtain an iron core magnetostriction characteristic spectrum. According to the invention, through multi-source signal fusion and material characteristic modeling, the accuracy, real-time performance and graded protection response capability of transformer DC magnetic bias monitoring are improved.

The invention discloses a transformer fault detection device based on a fuzzy logic algorithm, and the device comprises a data collection module which obtains the operation original data of a transformer in real time through combining the dissolved gas in oil with the temperature, vibration, current and voltage; the data preprocessing module is used for carrying out missing value processing, noise removal and abnormal value detection and processing on the original data; the feature extraction module is used for realizing dynamic feature selection based on data analysis provided by the data preprocessing module; the fault identification module is used for carrying out abnormal waveform judgment on current, voltage, temperature and vibration parameters through a threshold calculation unit and carrying out threshold adjustment based on an optimization algorithm; and the fault detection module triggers the alarm unit or maintains a normal working state according to an identification result of the fault identification module. According to the invention, through monitoring analysis and timely alarm notification, accurate and efficient monitoring of the transformer fault is realized, and stable operation and long-term reliability of the transformer are ensured.

The invention relates to the technical field of transformers, and provides a transformer load capacity prediction method and device, equipment and a storage medium, and the method comprises the steps: obtaining current data, winding temperature data, oil temperature data, environment temperature data and historical load records of a target transformer; calculating a temperature rise correction value based on the winding temperature data and the oil temperature data; after the temperature rise change rate in unit time is calculated through the current data, the environment temperature data and the temperature rise correction value, time sequence matching analysis is carried out on the temperature rise change rate and historical load records, after the load capacity change trend is obtained, comparative analysis is carried out on the load capacity change trend and the historical load records, and a load capacity prediction result is generated. By utilizing collaborative analysis and heat conduction characteristic calculation of multi-source data and comprehensively constructing a time delay interval and a load capacity evaluation method, the precision and applicability of transformer load capacity prediction are improved, and the problem that the prediction precision is insufficient due to the fact that operation data and environment data of the transformer cannot be fully fused during multi-source data processing is solved.

The invention discloses a power transformer arc discharge multi-parameter detection simulation platform and a fault diagnosis method, and relates to the technical field of power system equipment state monitoring and fault diagnosis. The platform comprises a transformer body, a replaceable discharge module, a multi-parameter sensing unit and a signal processing and diagnosis module, and can truly reproduce various typical arc discharge faults of a needle plate, an air gap, a creeping surface, turn-to-turn and the like. The sensing unit is integrated with ultrahigh frequency and ultrasonic sensing probes, high-frequency current and voltage sensors, optical fiber temperature / pressure / strain sensors and the like, so that synchronous acquisition of multi-physical field signals is realized. According to the diagnosis method, through wavelet denoising and multi-dimensional feature extraction, a feature vector of multi-state parameter fusion in the process from partial discharge to arcing is constructed, and accurate classification of fault types is realized by using a support vector machine (SVM) model. The diagnosis method has high accuracy and early warning capability, effectively overcomes the limitation of single parameter diagnosis, and provides reliable technical support for transformer fault research and intelligent operation and maintenance.

This invention presents a method for detecting circuit anomalies in a current transformer. The method is executed by a module integrated into an intelligent electronic device. The process involves a multi-step analysis that starts with comparing the frequency of a detected current signal from the transformer's secondary winding with the frequency of a detected voltage signal from a primary conductor. If the frequency difference exceeds a predetermined threshold, the method advances to assess the magnitude of the detected current signal. If this magnitude falls below another predetermined threshold, the method proceeds to analyze the Total Harmonic Distortion (THD) of the current signal. Depending on whether the THD is above or below a set threshold, the method determines the presence of either an open or a short circuit within the current transformer. This approach provides a reliable and accurate means for identifying circuit anomalies, enhancing the safety of electrical systems.

Provided are a method for training a transformer fault detection model, a fault diagnosis method, and a related device. The method includes: obtaining an initial voiceprint signal of a transformer and a fault type corresponding to the initial voiceprint signal; preprocessing the initial voiceprint signal to obtain an input signal, and establishing an input signal dataset; performing feature extraction on a first input signal in the training dataset based on a preset feature extraction algorithm to obtain a first voiceprint feature; training an initial detection model based on the first voiceprint feature and a first fault type corresponding to the first input signal to obtain a first training result; determining a loss function based on the first training result and the first fault type; and iteratively adjusting a weight value of the initial detection model until the loss function converges to obtain a fault detection model.

The invention discloses a power transformer winding insulation state monitoring method, a power transformer winding insulation state monitoring device, power transformer winding insulation state monitoring equipment and a medium, and relates to the technical field of power system equipment state monitoring. And an improved particle swarm optimization algorithm is adopted to inversely calculate the insulation parameters of a plurality of windings of the transformer to the ground, so that the insulation state of each winding can be accurately evaluated. Compared with a traditional monitoring method, the method has the advantages that rich information contained in the common-mode leakage current can be fully utilized, the insulation degradation conditions of different windings can be accurately distinguished, more accurate insulation state evaluation is provided, and the safe operation level of the transformer is effectively improved. Moreover, multi-stage evaluation is carried out on the insulation state through a fuzzy comprehensive evaluation method, the insulation state can be comprehensively reflected, different severity degrees of the insulation state can be distinguished, the current state can be evaluated, the degradation trend can be predicted through dynamic indexes, and a scientific basis is provided for operation and maintenance decision making.

The invention relates to the technical field of voiceprint fault diagnosis, in particular to a multi-mode voiceprint fault diagnosis method for a converter transformer. The system comprises a multi-modal signal acquisition module, a multi-modal signal preprocessing module, a feature weighted fusion module, a depth feature extraction module and a fault identification and classification output module. A mechanical vibration signal and a voiceprint feature signal are synchronously collected through a vibration sensor and a voiceprint sensor, and a multi-modal feature vector is constructed after preprocessing; dynamic weighted fusion of vibration and voiceprint features is realized by adopting a channel attention mechanism, and a channel weight is generated through global average pooling and nonlinear mapping; and finally, voiceprint embedding vectors with time sequence distribution characteristics are extracted through attention statistical pooling, and accurate recognition of fault types is realized by adopting a Softmax classifier. According to the method, through collaborative optimization of physical signal coupling, algorithm feature fusion and deep representation learning, the detection capability of the early weak fault of the converter transformer is effectively improved.

The disclosure provides a transformer fault detection model training method, a fault diagnosis method, and a related device. The method comprises: obtaining an initial voiceprint signal of a transformer, and a fault type corresponding to the initial voiceprint signal; preprocessing the initial voiceprint signal by using a wavelet packet analysis method, so as to obtain an input signal, and according to the input signal and the fault type, establishing an input signal data set; according to a preset feature extraction algorithm, performing feature extraction on a first input signal in a training data set to obtain a first voiceprint feature; using the first voiceprint feature and a first fault type corresponding to the first input signal to train an initial detection model, so as to obtain a first training result; according to the first training result and the fault type, determining a loss function; on the basis of the loss function, using a back propagation algorithm to iteratively adjust weight values of the initial detection model, until the loss function converges, so as to obtain a fault detection model. In the present disclosure, accurate transformer fault detection is realized by utilizing the obtained fault detection model.

The invention discloses a method for detecting the short-circuit resistance of a power transformer based on a digital twinning technology, and the method comprises the steps: building a three-dimensional transformer vibration simulation model, and carrying out the coupling calculation of a circuit, a magnetic field, a structural force field and a flow field. Calculating current and voltage curves of the winding and magnetic field distribution of the transformer in electromagnetic field simulation; solving a deformation trend and a fault type of the winding in a structural force field; the vibration characteristics of the measuring points on the surface of the oil tank are calculated in the flow field, so that the relationship between transformer winding deformation and vibration and reactance characteristics is obtained, the short-circuit damage process of the transformer can be analyzed, the method is of great significance in mastering the damage mechanism of the transformer, the anti-short-circuit capability of the power transformer is detected from multiple aspects, the detection accuracy is improved, and the detection cost is reduced. Related tests can be carried out when transformer equipment factories or related scientific research institutions are combined with transformer short-circuit tests, transformer short-circuit laws are mastered, and the anti-short-circuit capacity of the transformer is optimized and improved.

The invention discloses a transformer on-line monitoring and fault early warning system, and particularly relates to the technical field of power equipment monitoring, and the system comprises a signal collection module, a signal filtering module, a broadband focusing module, a signal complexity analysis module, a threshold self-tuning module, a fault early warning module and a fault positioning module. The signal acquisition module acquires operation data of a transformer in real time through a sensor network, the filtering module removes noise, and the broadband focusing module converts broadband signals into narrow-band signals, so that the resolution is improved; the signal complexity analysis module evaluates the signal complexity, and the threshold value self-tuning module dynamically adjusts the threshold value; the fault early warning module judges whether to trigger early warning based on the adjusted threshold value, and the fault positioning module determines the position of the partial discharge source through a cross positioning algorithm; according to the invention, the problem that the accuracy of early warning is affected due to the fact that a traditional fixed-threshold monitoring method is easy to generate false alarm or missing alarm can be solved, and operation and maintenance personnel can be helped to quickly determine the position of a fault source through cross positioning.

The invention relates to the related technical field of distribution transformer insulation aging, in particular to a frequency domain diagnosis method, system and device for distribution transformer insulation aging and a medium, and the method comprises the steps: carrying out the frequency response test of a to-be-tested transformer, and obtaining a first frequency response curve; performing a frequency response simulation test on a transformer equivalent circuit model pre-constructed based on the to-be-tested transformer to obtain a second frequency response curve; constructing a target function based on the deviation of the first frequency response curve and the second frequency response curve; taking element parameters in the equivalent circuit model of the transformer as position vectors, taking the objective function as a fitness function, and adopting a multi-objective particle swarm algorithm to solve the element parameters to obtain optimal values of the element parameters; respectively comparing the optimal values of the element parameters with corresponding normal values to obtain insulation aging degrees; according to the method, the insulation aging degree can be obtained by comparing the element parameters obtained through optimization with the normal unaged element parameters.

The invention relates to the technical field of transformer detection, and discloses a nondestructive testing optimization method and system for an oil-immersed power transformer, and the method comprises the steps: building a three-dimensional dielectric response coordinate system, and generating a preliminary defect positioning map; aging-dominated and damp-dominated defects are detected and identified through spiral frequency sweep excitation; establishing a temperature gradient excitation scheme based on the defect type to generate a defect degree quantitative evaluation index; designing a sound wave modulation excitation scheme to generate a high-resolution defect characteristic spectrum; constructing a multi-mode intelligent sensor network, and combining a defect development trend prediction model to realize defect evolution prediction and generate a graded early warning signal; according to the method, the whole-process accurate detection of the insulation defect of the transformer from positioning, classification and quantitative evaluation to evolution prediction is realized, and a reliable basis is provided for operation and maintenance.

The embodiment of the invention discloses a transformer winding short circuit fault detection method. The method comprises the following steps: acquiring magnetic flux leakage data between transformer windings; determining a short-circuit fault area according to the magnetic flux leakage data; constructing a finite element model of the transformer; simulating the finite element model of the transformer based on the short-circuit fault area to obtain a short-circuit database; and acquiring magnetic flux leakage data between the transformer windings in real time, and determining the short-circuit fault position of the transformer windings according to the short-circuit database. According to the method, accurate detection of the short-circuit fault of the transformer winding is achieved by obtaining the magnetic flux leakage data, determining the short-circuit fault area, constructing the finite element model, obtaining the short-circuit database through simulation and locating the fault in real time, the accuracy and efficiency of fault detection are improved, and powerful guarantee is provided for safe operation of the transformer.

The invention relates to a fault detection method and device for a power transformer based on an infrared thermal imaging technology, and relates to the technical field of fault detection. The method comprises the following steps: identifying a target area where a to-be-detected power transformer in a power system is located; performing infrared detection on the target area based on an infrared thermal imaging technology to obtain a temperature distribution thermal image shown outside the to-be-detected power transformer; performing temperature analysis on the temperature distribution thermogram according to a pre-trained temperature analysis model to obtain a temperature analysis result; and determining fault information of the power transformer to be detected according to the temperature information of each component in the power transformer. By adopting the method, the detection accuracy of the power transformer in the power system can be improved.

The invention relates to the technical field of capacitance compensation matching, in particular to a dynamic tuning capacitance compensation matching method and system for a transformer simulation test. The method comprises the following steps: before a test is started, establishing a digital twin model based on parameters of a transformer to be tested, identifying a systematic measurement error source caused by inherent characteristics of a system through simulation, and generating error source correction parameters; in the test period, voltage and current original waveforms containing electrical disturbance in a test loop are obtained in real time through high-frequency synchronous sampling; a mixed disturbance separation model is used, the collected original waveform and the generated error source correction parameters are used as input, and random measurement disturbance components caused by load dynamic changes are calculated and separated in real time; and based on the separated disturbance component, driving a dynamic capacitance compensation module to generate a compensatory electric signal which is equal to the disturbance component in size and opposite to the disturbance component in phase, and injecting the compensatory electric signal into a test loop so as to realize active disturbance cancellation. According to the invention, the measurement precision of the transformer simulation test can be improved.

The invention discloses a transformer winding monitoring and diagnosis data acquisition method, system and device and a medium, and belongs to the technical field of power equipment state monitoring, and the method comprises the steps: constructing a multi-scale wavelet transform and improved adaptive Kalman filtering fusion architecture, the multi-scale wavelet transform and improved adaptive Kalman filtering fusion architecture is used for acquiring a monitoring signal of a transformer winding; performing frequency division processing on the monitoring signal by using the multi-scale wavelet transform, and determining each scale coefficient; adjusting a noise covariance matrix based on improved adaptive Kalman filtering, and carrying out filtering processing on each scale coefficient to determine a filtering result; and reconstructing the filtering result into a reconstructed signal, and performing signal quality evaluation on the reconstructed signal to determine the signal quality of the transformer winding. Through multi-scale decomposition and adaptive filtering, the limitation of a traditional method under non-stationary signal processing and strong noise interference is solved, and the retention degree and the signal-to-noise ratio of signal features are improved.

The invention belongs to the technical field of power equipment maintenance, and particularly relates to a transformer winding fault diagnosis method, a terminal and a storage medium, and the method comprises the steps: obtaining frequency response curves of a transformer winding in a normal state and an abnormal state through a frequency sweep test, the head end of the primary side winding is connected with a sweep frequency signal and the tail end is suspended; determining the number of units of an equivalent network according to the number of the frequency response curve characteristic frequencies of the windings in the normal state, and reconstructing a normal winding equivalent network composed of a plurality of RLC units; according to the method, the high-precision equivalent network is constructed, the frequency response residual error analysis is performed, and the Pearson's correlation coefficient is combined to screen the high-correlation element, so that the specific position of the winding deformation or fault can be accurately positioned, the fault type is distinguished, and the limitation that the traditional frequency response method can only judge the overall deformation degree is thoroughly solved.

The invention relates to the technical field of transformers, in particular to a transformer component fault diagnosis method based on machine learning, which comprises the following steps: analyzing medium data acquired by a transformer component, comparing a dielectric loss factor sequence with offset characteristics, training and generating an offset time sequence characteristic set, screening abnormal time periods and aggregating key discharge characteristics, and performing fault diagnosis. And obtaining an abnormal clustering signal group, concluding a trend evolution sequence, performing layered optimization on a path structure, and outputting fault link positioning data. According to the method, various parameters such as dielectric loss, discharge, temperature rise, gas and vibration are integrated, abnormal distribution is actively screened, key node trend change is captured in real time, historical and real-time characteristics in different operation scenes are fused, and efficient analysis of complex link evolution and implicit anomalies is realized. And the fault evolution relationship among the nodes is output in a structured manner, direct application of a diagnosis result in full-link traceability and multi-node trend research and judgment is supported, and the fault diagnosis pertinence and the hidden danger recognition capability are effectively improved.

The invention discloses a transformer fault ultrasonic detection system, which comprises an ultrasonic sensor module, a signal conditioning module, a data acquisition and processing module, a data fusion module, a learning module, a diagnosis decision module and a human-computer interaction interface, according to the ultrasonic sensor module, 6-8 ultrasonic sensors are arranged in a transformer shell in an array mode, piezoelectric ceramic sensors with the resonant frequency of 40 kHZ are adopted as the ultrasonic sensors, the sensitivity of the ultrasonic sensors is-65 dBV / uBar, and effective capture is weak. According to the ultrasonic detection system for the transformer fault, single ultrasonic detection is easily influenced by accidental interference, the system is synchronously connected into equipment such as an infrared thermal imager and a vibration accelerometer, and a multi-physical-quantity correlation analysis model is established. When high-frequency ultrasonic pulses are detected, the system automatically calls temperature data of the corresponding position, and if the temperature gradient exceeds 2 DEG C / cm, it is confirmed that discharging heating is conducted; and if the temperature rise does not exist and the low-frequency vibration of 10-200Hz is accompanied, the mechanical looseness is judged.

The invention provides a transformer winding state detection method based on temperature characteristics. The method comprises the following steps: (1) collecting operation data; (2) preprocessing the operation data; (3) taking the normalized transformer load current, the normalized active power, the normalized top oil temperature, the normalized ambient temperature and the normalized oil flow speed as input data, taking the normalized winding hot-spot temperature as output data, and constructing a data set; (4) establishing a VSN-TKAN-GRN deep learning network parallel computing model of the transformer winding hot-spot temperature considering the multi-head attention mechanism; (5) predicting the hot-spot temperature of the transformer winding based on a VSN-TKAN-GRN deep learning network parallel computing model; and (6) calculating the average relative error percentage of the transformer winding hot-spot temperature prediction result and the actual measurement result, and judging the winding state according to the change of the average relative error percentage. According to the method, the state of the transformer winding in operation can be efficiently and accurately judged, so that effective transformer operation and maintenance measures can be taken in time, and large faults are avoided.