A dynamic analysis method for short-circuit resistance of an in-service transformer considering health state

Abstract

The present application belongs to the technical field of power equipment monitoring, and relates to a dynamic analysis method for short-circuit resistance of an operating transformer considering health state, comprising: obtaining health state sensing data of the transformer at a historical time; calculating an equivalent loss coefficient of short-circuit resistance according to a dynamic equivalent short-circuit current; iteratively training a preset neural network with the health state sensing data as input and the equivalent loss coefficient of short-circuit resistance as output to obtain a dynamic analysis model for short-circuit resistance; inputting real-time health state sensing data into the dynamic analysis model for short-circuit resistance to obtain a real-time equivalent loss coefficient of short-circuit resistance through the dynamic analysis model for short-circuit resistance, i.e. obtaining an evaluation result for short-circuit resistance of the transformer. The present application maps the health state and the equivalent loss coefficient of short-circuit resistance through dynamic equivalent short-circuit current and neural network, solving the defect that the traditional method cannot quantize real-time influence of insulation aging.

Description

Technical Field

[0001] The present invention belongs to the technical field of power equipment monitoring, and in particular relates to a dynamic analysis method for the short-circuit resistance of an in-service transformer taking into account its health status. Background Art

[0002] As core equipment in power systems, the reliability and safety of transformers are directly linked to the stable operation of the power grid. In actual operation, transformers are often exposed to the risk of short-circuit shocks, and their short-circuit resistance is a key indicator of the equipment's safe operation. However, with long-term operation, the transformer's short-circuit resistance gradually declines due to insulation aging, mechanical structure degradation, and the influence of external environmental factors. This change in health status makes it difficult for traditional static assessment methods to accurately reflect the transformer's actual short-circuit resistance. Chinese Patent Application No. 202210057776.1, "A Dynamic Assessment Method for Power Transformer Short-Circuit Resistance Based on Multi-Parameters," currently assesses the short-circuit resistance of transformers based on design parameters and initial conditions. However, this method relies on segmented historical data and cannot dynamically track the loss of transformer capacity due to deterioration in health during operation. It also suffers from poor real-time performance. Furthermore, this method only uses normalized weighting and lacks the ability to fit nonlinear data from multiple sources and heterogeneous data, resulting in significant assessment errors and potential safety hazards. In addition, although existing monitoring methods can collect multi-source data such as transformer vibration, noise, and dissolved gas in oil, they lack an analytical model that dynamically associates this data with the short-circuit resistance capability, resulting in the inability to achieve real-time assessment of the short-circuit resistance capability. Summary of the Invention

[0003] In order to solve the problem that it is impossible to dynamically calculate in real time the loss of short-circuit resistance of operating transformers due to insulation aging and health deterioration after operation, the present invention proposes a dynamic analysis method for the short-circuit resistance of operating transformers taking into account the health status. By combining historical health status perception data, calculation of dynamic equivalent short-circuit current and static equivalent short-circuit current, training of neural network model and real-time evaluation, accurate dynamic analysis of the short-circuit resistance of operating transformers is achieved.

[0004] The present invention provides a method for dynamically analyzing the short-circuit capability of an in-service transformer taking into account its health status, comprising:

[0005] Step 1: Obtain the health status perception data of the transformer at the historical moment;

[0006] Step 2: Under a preset vibration, iteratively inversely calculate the static equivalent short-circuit current of the transformer's short-circuit resistance according to a preset coupling model;

[0007] Step 3: Obtain the current peak value in the actual short-circuit impact event and correct it to obtain the dynamic equivalent short-circuit current, and calculate the equivalent short-circuit protection capability loss factor based on the dynamic equivalent short-circuit current;

[0008] Step 4: Correlate the health status perception data of the transformer at the moment of short-circuit impact with the equivalent impairment coefficient of the short-circuit resistance capability; use the health status perception data as input and the equivalent impairment coefficient of the short-circuit resistance capability as output, iteratively train a preset neural network to obtain a dynamic analysis model of the short-circuit resistance capability;

[0009] Step 5. Input the real-time health status perception data into the dynamic analysis model of the short-circuit resistance capability. The real-time short-circuit resistance capability equivalent loss coefficient is obtained through the output of the dynamic analysis model of the short-circuit resistance capability. The actual short-circuit current that the operating transformer can currently withstand is calculated based on the real-time short-circuit resistance capability equivalent loss coefficient and the static equivalent short-circuit current, thus obtaining the transformer short-circuit resistance capability evaluation result.

[0010] Further preferably, in step 1, the health status sensing data includes: short-circuit impact cracking conditions, anti-short-circuit sensitive monitoring data, winding maintenance test data, and product design data;

[0011] Short-circuit impact cracking process data: historical short-circuit current peak , short circuit impact duration And cumulative number of impacts ;

[0012] Short-circuit sensitive monitoring data: Winding vibration acceleration spectrum , noise sound pressure level , concentration of dissolved gas (H2, C2H2) in oil ;

[0013] Winding maintenance test data: winding deformation , insulation resistance R ins ;

[0014] Product design data: short-circuit impedance , winding elastic modulus E and insulation spacer spacing d.

[0015] More preferably, the step 2 comprises:

[0016] Using the winding elastic modulus E and the insulation spacer spacing d in the product design data, the winding displacement equation is established:

[0017] ,

[0018] Where, is the winding displacement, is the winding material density, is the core magnetic permeability, For time, is the static equivalent characteristic current, It is the unit distance of the insulating spacer in the direction perpendicular to the force;

[0019] In the event of historical short-circuit impulse current ratio exceeding 50%, extract the vibration acceleration peak value A peak ;

[0020] Tongdang The iteration is terminated when is the critical winding displacement, and the output static equivalent characteristic current is used as the static equivalent short-circuit current.

[0021] Further preferably, in step 3, the current peak value in the actual short circuit impact event is obtained And correct it to the dynamic equivalent short-circuit current according to the following rules :

[0022] Historical short-circuit current peak Limit: If ,but ;

[0023] Short circuit shock duration Attenuation: If More than 0.5 seconds, press attenuation;

[0024] Cumulative number of impacts Correction: If >10, additional correction factor ,Right now .

[0025] Further preferably, in step 3, the calculation process of the equivalent loss coefficient of the short-circuit resistance is as follows:

[0026] Calculate dynamic equivalent short-circuit current based on finite element simulation Theoretical vibration under and theoretical leakage flux ; During short circuit shock events, the actual vibration is recorded simultaneously and actual leakage flux ;

[0027] Equivalent loss factor of short-circuit resistance Expressed as:

[0028] ,

[0029] In the formula 、 Respectively represent the weights of vibration and leakage flux against short-circuit attenuation, is the vibration frequency.

[0030] Further preferably, in the step 4, the short-circuit resistance dynamic analysis model is modeled by mapping the health state with the short-circuit resistance equivalent impairment coefficient, specifically including:

[0031] Step S41, input feature construction: the health status data at the time of short circuit impact and the equivalent loss coefficient of short circuit resistance Correlation, input features include: vibration deviation , magnetic flux leakage deviation , winding deformation , dissolved gas concentration in oil , insulation resistance R ins , short-circuit impedance , winding elastic modulus E and insulation spacer spacing d;

[0032] Step S42: Neural network model training:

[0033] The neural network structure includes the input layer: receiving feature vector ;

[0034] Hidden layer: 2 fully connected layers;

[0035] Output layer: Sigmoid function constraint ∈(0,1], directly mapping the equivalent loss coefficient of short-circuit resistance;

[0036] The training optimization dataset contains The historical short-circuit event data of the label is divided into training set, validation set, and test set in a ratio of 7:2:1; and optimized using the Adam algorithm.

[0037] Further preferably, in step 5, the real-time health status perception data is preprocessed to construct a real-time feature vector X real ; X real To the trained neural network model, output the current real-time short-circuit resistance equivalent loss coefficient of the operating transformer .

[0038] Further preferably, the actual short-circuit current that the transformer can withstand is calculated based on the real-time short-circuit resistance equivalent loss factor. :

[0039] ,in, is the static equivalent characteristic current, It is the equivalent loss coefficient of real-time short-circuit resistance capability.

[0040] The beneficial effects of the present invention are: by combining historical health status perception data, the calculation of dynamic and static equivalent short-circuit currents, and the training and real-time evaluation of a neural network model, it overcomes the inability of traditional methods to dynamically calculate the transformer's short-circuit resistance in real time. This enables real-time dynamic evaluation and provides reliable technical support for the safe operation of the transformer. By using dynamic equivalent short-circuit current correction rules and a neural network to map health status and impairment coefficients, the present invention overcomes the limitation of traditional methods in being unable to quantify the real-time impact of insulation aging. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 Flowchart of the present invention. DETAILED DESCRIPTION

[0042] The present invention will be further described in detail below with reference to the accompanying drawings.

[0043] Reference Figure 1 A dynamic analysis method for the short-circuit capability of an in-service transformer considering its health status is proposed, including:

[0044] Step 1: Obtain health status perception data of the transformer at a historical moment; the health status perception data includes: short-circuit impact cracking conditions, anti-short-circuit sensitive monitoring data, winding maintenance test data, and product design data;

[0045] Short-circuit impact cracking process data: historical short-circuit current peak , short circuit impact duration And cumulative number of impacts ;

[0046] Short-circuit sensitive monitoring data: Winding vibration acceleration spectrum , noise sound pressure level , concentration of dissolved gas (H2, C2H2) in oil ;

[0047] Winding maintenance test data: winding deformation , insulation resistance R ins ;

[0048] Product design data: short-circuit impedance , winding elastic modulus E and insulation spacer spacing d.

[0049] After obtaining the health status perception data of the transformer at a historical moment, an adaptive sliding window based on a multimodal joint distribution is constructed, and abnormal values ​​in the health status perception data are collaboratively detected through a dynamic probability graph model to obtain preprocessed health status perception data.

[0050] Step 2: Under a preset vibration, iteratively inversely calculate the static equivalent short-circuit current of the transformer's short-circuit resistance according to a preset coupling model;

[0051] Using the winding elastic modulus E and the insulation spacer spacing d in the product design data, the winding displacement equation is established:

[0052] ,

[0053] Where, is the winding displacement, is the winding material density, is the core magnetic permeability, For time, is the static equivalent characteristic current, It is the unit distance of the insulating spacer in the direction perpendicular to the force;

[0054] In the event of historical short-circuit impulse current ratio exceeding 50%, extract the vibration acceleration peak value A peak ;

[0055] By the winding displacement equation, when The iteration is terminated when =2mm), is the critical winding displacement, and the output static equivalent characteristic current is used as the static equivalent short-circuit current.

[0056] Step 3: Obtain the current peak value in the actual short-circuit impact event and correct it to obtain the dynamic equivalent short-circuit current, and calculate the equivalent short-circuit protection capability loss factor based on the dynamic equivalent short-circuit current;

[0057] Step 4: Correlate the health status perception data of the transformer at the moment of short-circuit impact with the equivalent impairment coefficient of the short-circuit resistance capability; use the health status perception data as input and the equivalent impairment coefficient of the short-circuit resistance capability as output, iteratively train a preset neural network to obtain a dynamic analysis model of the short-circuit resistance capability;

[0058] Step 5. Input the real-time health status perception data into the dynamic analysis model of the short-circuit resistance capability. The real-time short-circuit resistance capability equivalent loss coefficient is obtained through the output of the dynamic analysis model of the short-circuit resistance capability. The actual short-circuit current that the operating transformer can currently withstand is calculated based on the real-time short-circuit resistance capability equivalent loss coefficient and the static equivalent short-circuit current, thus obtaining the transformer short-circuit resistance capability evaluation result.

[0059] In step 3, obtain the current peak value in the actual short circuit impact event And correct it to the dynamic equivalent short-circuit current according to the following rules :

[0060] Historical short-circuit current peak Limit: If ,but ;

[0061] Short circuit shock duration Attenuation: If More than 0.5 seconds, press attenuation;

[0062] Cumulative number of impacts Correction: If >10, additional correction factor ,Right now .

[0063] The calculation process of the equivalent loss coefficient of short-circuit resistance is as follows:

[0064] Calculate dynamic equivalent short-circuit current based on finite element simulation Theoretical vibration under and theoretical leakage flux ; During short circuit shock events, the actual vibration is recorded simultaneously and actual leakage flux ;

[0065] Equivalent loss factor of short-circuit resistance Expressed as:

[0066] ,

[0067] In the formula 、 Respectively represent the weights of vibration and leakage flux against short-circuit attenuation, is the vibration frequency.

[0068] In the fourth step, the short-circuit resistance dynamic analysis model is modeled by mapping the health state with the short-circuit resistance equivalent impairment coefficient, specifically including:

[0069] Step S41: Input feature construction:

[0070] The health status data at the time of short circuit impact and the equivalent loss coefficient of short circuit resistance Association,input features include:

[0071] (1) Mechanical state: vibration deviation , magnetic flux leakage deviation , winding deformation ;

[0072] (2) Insulation state: concentration of dissolved gas in oil , insulation resistance R ins ;

[0073] (3) Product design parameters: short-circuit impedance , winding elastic modulus E and insulation spacer spacing d;

[0074] Step S42: Neural network model training:

[0075] The network structure includes the input layer: receiving feature vector Hidden layer: 2 fully connected layers (64 neurons, ReLU activation);

[0076] Output layer: Sigmoid function constraint ∈(0,1], directly mapping the equivalent loss coefficient of short-circuit resistance;

[0077] The training optimization dataset contains The historical short-circuit event data of the labels is divided into training set, validation set, and test set in a ratio of 7:2:1. The data is optimized using the Adam algorithm and the early stopping method is used to prevent overfitting.

[0078] In step 5, the real-time health status perception data is preprocessed to construct a real-time feature vector X real ; X real To the trained neural network model, output the current real-time short-circuit resistance equivalent loss coefficient of the operating transformer . Short-circuit current The calculation is as follows:

[0079] ,in, is the static equivalent characteristic current, It is the equivalent loss coefficient of real-time short-circuit resistance capability.

[0080] The technical solution of the present invention solves the problem that traditional methods cannot dynamically calculate the short-circuit resistance of transformers in real time by combining historical health status perception data, calculation of static and dynamic equivalent short-circuit current values, training of neural network models and real-time evaluation. In actual application scenarios, for example, transformers operating in a large substation will experience short-circuit shock events at regular intervals. Through the method of the present invention, the health status of the transformer can be monitored in real time, its short-circuit resistance can be dynamically evaluated, and maintenance plans can be formulated or operating strategies can be adjusted based on the evaluation results. Compared with traditional methods, the present invention not only improves the evaluation accuracy, but also significantly reduces the risk of failures caused by insufficient short-circuit resistance of the transformer, providing reliable protection for the safe and stable operation of the power system.

Claims

1. A dynamic analysis method for the short-circuit capability of an in-service transformer considering its health status, characterized in that: include: Step 1: Obtain the health status perception data of the transformer at the historical moment; Step 2: Under a preset vibration, iteratively inversely calculate the static equivalent short-circuit current of the transformer's short-circuit resistance according to a preset coupling model; Step 3: Obtain the current peak value in the actual short-circuit impact event and correct it to obtain the dynamic equivalent short-circuit current, and calculate the equivalent short-circuit protection capability loss factor based on the dynamic equivalent short-circuit current; Step 4: Correlate the transformer health status perception data at the moment of short-circuit impact with the equivalent loss factor of the short-circuit resistance capability; Taking the health status sensing data as input and the short-circuit resistance equivalent impairment coefficient as output, a preset neural network is iteratively trained to obtain a short-circuit resistance dynamic analysis model; Step 5. Input the real-time health status perception data into the dynamic analysis model of the short-circuit resistance capability. The real-time short-circuit resistance capability equivalent loss coefficient is obtained through the output of the dynamic analysis model. The actual short-circuit current that the operating transformer can currently withstand is calculated based on the real-time short-circuit resistance capability equivalent loss coefficient and the static equivalent short-circuit current, thus obtaining the transformer short-circuit resistance capability evaluation result.

2. The method for dynamic analysis of short-circuit capability of an in-service transformer considering health status according to claim 1, characterized in that: In step 1, the health status perception data includes: Short-circuit impact cracking process data: historical short-circuit current peak , short circuit impact duration And cumulative number of impacts ; Short-circuit sensitive monitoring data: Winding vibration acceleration spectrum , noise sound pressure level , dissolved gas concentration in oil ; Winding maintenance test data: winding deformation , insulation resistance R ins ; Product design data: short-circuit impedance , winding elastic modulus E and insulation spacer spacing d.

3. The method for dynamic analysis of short-circuit capability of an in-service transformer considering health status according to claim 1, characterized in that: The second step includes: Using the winding elastic modulus E and the insulation spacer spacing d in the product design data, the winding displacement equation is established: ， Where, is the winding displacement, is the winding material density, is the core magnetic permeability, For time, is the static equivalent characteristic current, It is the unit distance of the insulating spacer in the direction perpendicular to the force; In the event of historical short-circuit impulse current ratio exceeding 50%, extract the vibration acceleration peak value A peak ; Tongdang The iteration is terminated when is the critical winding displacement, and the output static equivalent characteristic current is used as the static equivalent short-circuit current.

4. The method for dynamic analysis of short-circuit capability of an in-service transformer considering health status according to claim 1, characterized in that: In the step 3, the peak current in the actual short circuit impact event is obtained. And correct it to the dynamic equivalent short-circuit current according to the following rules : Historical short-circuit current peak Limit: If ,but ; Short circuit shock duration Attenuation: If More than 0.5 seconds, press attenuation; Cumulative number of impacts Correction: If >10, additional correction factor ,Right now .

5. The method for dynamic analysis of short-circuit capability of an in-service transformer considering health status according to claim 1, characterized in that: In step 3, the calculation process of the equivalent loss coefficient of the short-circuit resistance is as follows: Calculate dynamic equivalent short-circuit current based on finite element simulation Theoretical vibration under and theoretical leakage flux ; During short circuit shock events, the actual vibration is recorded simultaneously and actual leakage flux ; Equivalent loss factor of short-circuit resistance Expressed as: ， In the formula 、 Respectively represent the weights of vibration and leakage flux against short-circuit attenuation, is the vibration frequency.

6. The method for dynamic analysis of short-circuit capability of an in-service transformer considering health status according to claim 1, characterized in that: In the fourth step, the short-circuit resistance dynamic analysis model is modeled by mapping the health state with the short-circuit resistance equivalent impairment coefficient, specifically including: Step S41, input feature construction: the health status data at the time of short circuit impact and the equivalent loss coefficient of short circuit resistance Correlation, input features include: vibration deviation , magnetic flux leakage deviation , winding deformation , dissolved gas concentration in oil , insulation resistance R ins , short-circuit impedance , winding elastic modulus E and insulation spacer spacing d; Step S42: Neural network model training: The neural network structure includes the input layer: receiving feature vector Hidden layer: 2 fully connected layers; Output layer: Sigmoid function constraint ∈(0,1], directly mapping the equivalent loss coefficient of short-circuit resistance; The training optimization dataset contains The historical short-circuit event data of the label is divided into training set, validation set, and test set in a ratio of 7:2:1; and optimized using the Adam algorithm.

7. The method for dynamic analysis of short-circuit capability of an in-service transformer considering health status according to claim 1, characterized in that: In step 5, the real-time health status perception data is preprocessed to construct a real-time feature vector X real ; X real To the trained neural network model, output the current real-time short-circuit resistance equivalent loss coefficient of the operating transformer .

8. The method for dynamic analysis of short-circuit capability of an in-service transformer considering health status according to claim 1, characterized in that: Short-circuit current Calculate according to the following formula; ,in, is the static equivalent characteristic current, It is the equivalent loss coefficient of real-time short-circuit resistance capability.

Images