Method for calculating insulation field intensity of transformer under influence of moisture and temperature distribution

Abstract

The invention discloses a method for calculating the insulation field intensity of a transformer under the influence of moisture and temperature distribution, and relates to the technical field of electric power engineering, and the method comprises the steps: constructing a multi-physical field coupling simulation platform for monitoring and analyzing the temperature, humidity and electric field distribution in the transformer; setting a humidity sensitive element array, collecting humidity data, and transmitting the humidity data to a central processing unit for processing; constructing an insulation prediction model through a deep neural network algorithm, inputting the insulation prediction model based on historical humidity, historical temperature and insulation state data to complete training, and predicting a change trend of insulation field intensity under a specific working condition; processing the real-time data, inputting the processed real-time data into the insulation prediction model to obtain a real-time insulation state prediction value, and dynamically adjusting parameters of the insulation prediction model; an abnormity early warning mechanism is established based on a real-time insulation state prediction value, and once an insulation state abnormity is detected, the system immediately starts early warning to inform operation and maintenance personnel to check and maintain.

Description

Technical Field

[0001] The invention relates to the technical field of electric power engineering, in particular to a method for calculating transformer insulation field strength under the influence of moisture and temperature distribution. Background Art

[0002] In the field of power engineering, transformers are key components of power systems, and their safe and stable operation is crucial to the reliability of the power grid. In recent years, with the rapid development of smart grids, the demand for transformer status monitoring and health assessment has been growing. Traditional transformer monitoring technology mainly relies on regular manual inspections and offline tests. This method is not only inefficient, but also difficult to capture early fault signs in real time. Especially in the face of complex and changeable operating environments, the impact of humidity and temperature fluctuations on the insulation state of the transformer is often ignored, which leads to misjudgment and missed faults.

[0003] Existing transformer condition monitoring technology has shortcomings in dealing with the complexity of the impact of environmental factors such as humidity and temperature on insulation status. On the one hand, the lack of multi-physical field coupling effect makes it difficult for the monitoring system to comprehensively evaluate the dynamic changes in the insulation aging process, especially the performance degradation of insulation materials caused by humidity and temperature changes, which directly affects the safe operation of the transformer. On the other hand, the lack of effective data processing and analysis tools and the application of deep learning models limit the monitoring system from mining valuable information from massive data, especially in predicting the evolution trend of insulation status, making it difficult to warn of potential failure risks in advance. Summary of the invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides a method for calculating the insulation field strength of transformers under the influence of moisture and temperature distribution to solve the problem that the lack of multi-physical field coupling effect makes it difficult for the monitoring system to comprehensively evaluate the dynamic changes in the insulation aging process, especially the performance degradation of insulation materials caused by humidity and temperature changes, which directly affects the safe operation of the transformer. On the other hand, the lack of effective data processing and analysis tools and the application of deep learning models limit the monitoring system from mining valuable information from massive data, especially in terms of the limited ability to predict the evolution trend of insulation status, making it difficult to provide early warning of potential failure risks.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In a first aspect, an embodiment of the present invention provides a method for calculating transformer insulation field strength under the influence of moisture and temperature distribution, which comprises:

[0008] Build a multi-physics coupling simulation platform to monitor and analyze the temperature, humidity, and electric field distribution inside the transformer;

[0009] Setting up a humidity sensitive element array, collecting humidity data, and transmitting the humidity data to a central processor for processing;

[0010] The insulation prediction model is built through a deep neural network algorithm. The historical humidity, historical temperature and insulation status data are input into the insulation prediction model to complete the training, and the change trend of the insulation field strength under specific working conditions is predicted;

[0011] The real-time data is processed and input into the insulation prediction model to obtain the real-time insulation state prediction value, and the parameters of the insulation prediction model are dynamically adjusted;

[0012] An abnormal warning mechanism is established based on the real-time insulation status prediction value. Once an abnormal insulation status is detected, the system immediately activates the warning and notifies the operation and maintenance personnel to conduct inspection and maintenance.

[0013] As a preferred solution of the method for calculating transformer insulation field strength under the influence of moisture and temperature distribution of the present invention, wherein: a multi-physical field coupling simulation platform is constructed to monitor and analyze the temperature, humidity and electric field distribution inside the transformer, and the specific steps are:

[0014] A multi-physics coupling model is constructed based on partial differential equations (PDEs) to describe the dynamic changes of temperature T, humidity and electric field. The core equations of the multi-physics coupling model include the temperature diffusion equation, the humidity diffusion equation and the electric field distribution equation.

[0015] The particle swarm optimization algorithm is used to dynamically update the PSO thermal conductivity, humidity diffusion coefficient and charge density;

[0016] The central processing unit, sensor network, wireless communication module, data preprocessing algorithm and multi-physics coupling model are integrated to obtain a multi-physics coupling simulation platform, wherein the central processing unit is used to coordinate the processing of data flow and the operation of the multi-physics coupling model.

[0017] As a preferred solution of the transformer insulation field strength calculation method under the influence of moisture and temperature distribution of the present invention, wherein: the humidity sensitive element array is set to collect humidity data, and the humidity data is transmitted to the central processor for processing, and the specific steps are:

[0018] Inside the transformer, selected locations including windings, core and tank wall are integrated with temperature sensors, humidity sensors and electric field sensors;

[0019] The area of ​​the selected part is divided into a uniform grid, with each grid cell representing a potential monitoring point;

[0020] The distance between the elements is calculated, and a humidity sensitive element is placed at the center of each grid cell to form a two-dimensional array;

[0021] Each humidity sensor has a built-in microcontroller with a sampling frequency f s =10Hz continuous collection of humidity data;

[0022] Manchester encoding is used to encode humidity data;

[0023] After completing Manchester encoding of humidity data, encapsulate it into data packets;

[0024] After receiving the data, the central processor first decodes the data packet and then pre-processes the humidity data using a Kalman filter to eliminate noise and improve data quality;

[0025] Distributed Kalman filter algorithm is used to fuse multiple component data from the same area;

[0026] Based on the real-time humidity data collected by the humidity sensitive element array, an abnormal value detection algorithm is used to detect abnormal values. When it is detected that the humidity in a certain area exceeds the normal range, an early warning is triggered immediately;

[0027] The humidity diffusion coefficient function in the multi-physics field coupling model is dynamically adjusted according to real-time humidity data to ensure that the model parameters are consistent with the actual working conditions.

[0028] As a preferred solution of the method for calculating transformer insulation field strength under the influence of moisture and temperature distribution of the present invention, the insulation prediction model is constructed by a deep neural network algorithm, and the specific steps are as follows:

[0029] Select the long short-term memory network LSTM as the basis of the insulation prediction model;

[0030] The insulation prediction model includes an input layer, several LSTM layers, a fully connected layer, and an output layer.

[0031] As a preferred solution of the transformer insulation field strength calculation method under the influence of moisture and temperature distribution of the present invention, wherein: the historical humidity, historical temperature and insulation state data are combined to input the insulation prediction model to complete the training and predict the change trend of the insulation field strength, and the specific steps are:

[0032] Extract humidity, temperature and insulation condition data related to transformer insulation condition from historical data sets;

[0033] Interpolation method was used to fill missing data, and Z-score standardization was used to normalize the data;

[0034] Based on the original humidity and temperature, principal component analysis (PCA) is used to extract the most influential cross-features for insulation status prediction, which is used to capture the interactive effects of humidity and temperature on insulation status.

[0035] The cross-features are input into the insulation prediction model to output the predicted value of the insulation status.

[0036] As a preferred solution of the method for calculating transformer insulation field strength under the influence of moisture and temperature distribution of the present invention, wherein: the real-time data is processed and input into the insulation prediction model to obtain the real-time insulation state prediction value, and the parameters of the insulation prediction model are dynamically adjusted, and the specific steps are:

[0037] Collect humidity, temperature and electric field data inside the transformer in real time through a sensor network;

[0038] Before data is transmitted to the central processor, Manchester coding and Kalman filter are used to process real-time data to improve data quality and transmission robustness;

[0039] The processed real-time data is passed as input to the insulation prediction model and the model parameters are updated;

[0040] The PSO algorithm is used to dynamically update thermal conductivity, humidity diffusion coefficient and charge density;

[0041] After the real-time data is input into the insulation prediction model, the real-time insulation status prediction value is output;

[0042] The mean square error (MSE) is selected as the loss function to measure the difference between the predicted value of the insulation prediction model and the actual insulation status;

[0043] Set the threshold LO;

[0044] When the MSE value exceeds the preset threshold, it indicates that there is a large deviation between the model prediction and the actual state. It is necessary to adjust the model parameters through the back propagation algorithm and optimize the model to reduce the MSE and improve the prediction accuracy.

[0045] As a preferred solution of the transformer insulation field strength calculation method under the influence of moisture and temperature distribution of the present invention, wherein: the abnormal warning mechanism is established based on the real-time insulation state prediction value. Once the insulation state is abnormal, the system immediately starts the warning and notifies the operation and maintenance personnel to conduct inspection and maintenance. The specific steps are:

[0046] Use anomaly detection algorithms to identify abnormal fluctuations in real-time insulation status prediction values;

[0047] Real-time monitoring of insulation status prediction values. Once the prediction values ​​deviate from the normal range, the early warning mechanism is triggered immediately.

[0048] After detecting an abnormality, the central processor immediately sends an alarm to the operation and maintenance personnel and starts the self-diagnosis program to find the potential cause of the fault and take timely measures to prevent the insulation condition from deteriorating.

[0049] As a preferred solution of the transformer insulation field strength calculation method under the influence of moisture and temperature distribution of the present invention, wherein: the insulation state prediction value is monitored in real time, and once the prediction value is found to deviate from the normal range, an early warning mechanism is immediately triggered, specifically:

[0050] Compare the dynamic threshold with the real-time insulation status prediction value, and then divide it into normal interval, warning interval, slight abnormal interval and serious abnormal interval;

[0051] Set a certain safety margin as △;

[0052] when When the system is in normal working condition, basic maintenance such as regular cleaning, lubrication, tightening and functional testing should be performed according to the equipment's usage frequency and historical maintenance records;

[0053] when When the system enters the early warning state, in addition to routine maintenance, it is necessary to increase the frequency of online monitoring, conduct remote diagnosis, and arrange professional personnel to conduct on-site inspections to ensure that any potential problems can be discovered and handled in a timely manner;

[0054] when When the system determines that it is a minor abnormality, the operation and maintenance personnel will be notified immediately to conduct on-site inspections, and the self-diagnosis program will be started to find possible fault sources and perform preliminary repair and adjustment measures;

[0055] when When the system faces serious abnormalities and a major failure is about to occur, it will be shut down for inspection immediately to avoid greater damage. At the same time, a high-level troubleshooting process will be initiated and damaged parts will be replaced if necessary.

[0056] In a second aspect, an embodiment of the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, any step of the method for calculating the insulation field strength of a transformer under the influence of moisture and temperature distribution as described in the first aspect of the present invention is implemented.

[0057] In a third aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, any step of the method for calculating the insulation field strength of a transformer under the influence of moisture and temperature distribution as described in the first aspect of the present invention is implemented.

[0058] The beneficial effects of the present invention are as follows: the present invention realizes comprehensive monitoring and analysis of the temperature, humidity and electric field distribution inside the transformer by constructing a multi-physical field coupling simulation platform. In particular, the present invention utilizes deep neural network technology to construct an insulation prediction model, and dynamically adjusts the model parameters in combination with real-time data. It can not only accurately predict the changing trend of the insulation state, but also quickly activate the early warning mechanism when an abnormal situation occurs, and provide timely fault diagnosis information for operation and maintenance personnel. The present invention particularly focuses on the intelligent analysis and early warning of the impact of humidity and temperature on the insulation state, providing a new solution for the health management of transformers, which helps to improve the operating safety and economy of the power system. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0060] Figure 1 This is a flow chart of the method for calculating transformer insulation field strength under the influence of moisture and temperature distribution in Example 1.

[0061] Figure 2 This is a flow chart of the method for calculating transformer insulation field strength under the influence of moisture and temperature distribution in Example 1. DETAILED DESCRIPTION

[0062] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0063] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0064] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0065] Example 1, reference Figure 1 and Figure 2 , which is the first embodiment of the present invention, provides a method for calculating transformer insulation field strength under the influence of moisture and temperature distribution, comprising the following steps:

[0066] S1 builds a multi-physics coupling simulation platform to monitor and analyze the temperature, humidity and electric field distribution inside the transformer;

[0067] A multi-physics coupling model is constructed based on partial differential equations (PDEs) to describe the dynamic changes of temperature T, humidity and electric field. The core equations of the multi-physics coupling model include the temperature diffusion equation, the humidity diffusion equation and the electric field distribution equation.

[0068] The temperature diffusion equation is expressed as:

[0069]

[0070] in, represents the rate of change of temperature field with time, is a gradient operator used to calculate the spatial gradient of the temperature field. k represents the temperature- and humidity-dependent thermal conductivity, T represents temperature, H represents humidity, T(x, y, z, t) represents the temperature field, and Q loss represents the heat loss function;

[0071] The humidity diffusion equation is expressed as:

[0072]

[0073] in, represents the rate of change of humidity field over time, D(H) represents the humidity diffusion coefficient function, H(x, y, z, t) represents the humidity field, S m Indicates humidity changes;

[0074] The electric field distribution equation is expressed as:

[0075]

[0076] Among them, E(x, y, z, t) represents the electric field distribution, and p is the charge density function, which represents the charge distribution under the influence of temperature field, humidity field and electric field distribution.

[0077] The particle swarm optimization algorithm is used to dynamically update the PSO thermal conductivity, humidity diffusion coefficient and charge density. The PSO algorithm can adjust these parameters according to real-time data to make the model more suitable for actual working conditions.

[0078] The central processing unit, sensor network, wireless communication module, data preprocessing algorithm and multi-physics coupling model are integrated to obtain a multi-physics coupling simulation platform, wherein the central processing unit is used to coordinate the processing of data flow and the operation of the multi-physics coupling model.

[0079] Heat loss function. The heat loss mainly comes from resistance loss. The resistance loss expression is:

[0080] Q loss (T,E)=σ(T)|E| 2 ;

[0081] Where σ represents the conductivity;

[0082] The humidity diffusion coefficient function represents the diffusion rate of water molecules in the medium. The humidity diffusion coefficient usually changes with the increase of humidity. The expression is:

[0083]

[0084] Where D0 represents the diffusion coefficient of humidity, E a represents the activation energy, which indicates the difficulty of water diffusion and is related to the chemical properties of the medium, and R represents the universal gas constant;

[0085] Charge density function. In transformer insulation materials, charge density is mainly generated by polarization charge and ion migration, and the expression is:

[0086] ρ(T,H,E)=ρ p (T,H)+ρ m (E);

[0087] Among them, p p (T,H) represents the temperature and humidity dependent polarization charge density, reflecting the polarization degree of the material at different temperatures and humidities. m (E) represents the electric field-dependent migration charge density, which describes the extent to which the electric field induces charge migration.

[0088] S2 sets the humidity sensitive element array, collects humidity data, and transmits the humidity data to the central processor for processing;

[0089] Select humidity sensitive elements based on the dielectric constant change principle. This element reflects humidity changes by measuring the dielectric constant change of the medium, and has the characteristics of high accuracy and fast response;

[0090] The gold electrode structure on the polyimide substrate is used because polyimide has stable dielectric properties in a wide humidity range, and the gold electrode is sensitive to humidity changes and can be effectively converted into electrical signals;

[0091] The LoRa technology is used for wireless data transmission. Due to its long-distance and low-power characteristics, it is very suitable for use in a closed environment such as inside a transformer.

[0092] Inside the transformer, selected locations including windings, core and tank wall are integrated with temperature sensors, humidity sensors and electric field sensors;

[0093] Divide the area of ​​the selected site into a uniform grid, with each grid cell representing a potential monitoring point. The size of the grid should take into account the expected gradient of humidity change and the required data resolution.

[0094] The distance between elements is calculated, and a humidity sensitive element is placed at the center of each grid unit to form a two-dimensional array so that each area is covered by at least one element, avoiding excessive overlap between elements and reducing data redundancy;

[0095] Each humidity sensor has a built-in microcontroller with a sampling frequency f s =10Hz continuous collection of humidity data;

[0096] Manchester coding is used to encode humidity data to improve the robustness and anti-interference ability of wireless transmission;

[0097] After completing Manchester encoding of humidity data, it is encapsulated into data packets, each of which includes component ID, encoded humidity data, timestamp and CRC check code;

[0098] After receiving the data, the central processor first decodes the data packet and then pre-processes the humidity data using a Kalman filter to eliminate noise and improve data quality;

[0099] A distributed Kalman filter algorithm is used to fuse data from multiple components in the same area to obtain a more accurate humidity estimate;

[0100] Based on the real-time humidity data collected by the humidity sensitive element array, anomaly detection algorithm is used to detect outliers. When the humidity in a certain area is detected to be beyond the normal range, an early warning is triggered immediately. Once an abnormality is detected, the central processor immediately sends an alarm to the operation and maintenance personnel and starts the self-diagnosis program to find the potential cause of the fault;

[0101] Dynamically adjust the humidity diffusion coefficient function in the multi-physics field coupling model based on real-time humidity data to ensure that the model parameters are consistent with the actual working conditions;

[0102] Calculate the distance between components, specifically:

[0103] Set the target coverage to C, which means the ratio of the area that a single element can effectively monitor to the total area. It can be set according to the needs. The expression of the maximum distance between elements is:

[0104]

[0105] Where d represents the maximum distance between components, and r represents the monitoring radius of each humidity sensitive component;

[0106] Manchester encoding is used to encode humidity data. The specific steps of encoding are:

[0107] Convert the humidity data into binary format, set the humidity data as a 16-bit number, and convert it into a binary string after quantization and encoding;

[0108] Manchester encoding is used to convert each bit 0 and 1 into two pulse signals. The duration of each pulse signal is half a bit period. Bit 0 is represented by a pulse from low to high, and bit 1 is represented by a pulse from high to low.

[0109] When the bit is 0, the signal remains low in the first half of the bit period and jumps to high in the second half;

[0110] When the bit is 1, the signal remains high in the first half of the bit period and drops to low in the second half;

[0111] Based on the sampling frequency, calculate the bit period, which is equal to 1 divided by twice the sampling frequency.

[0112] S3 builds an insulation prediction model through a deep neural network algorithm. The insulation prediction model is trained based on historical humidity, historical temperature and insulation status data, and the changing trend of insulation field strength under specific working conditions is predicted.

[0113] Collect insulation condition data, including insulation resistance and dielectric spectrum, through regular inspections of transformers;

[0114] We choose the long short-term memory network (LSTM) as the basis of the insulation prediction model because it is good at processing time series data and can capture the long-term dependencies of data.

[0115] The insulation prediction model includes an input layer, several LSTM layers, a fully connected layer, and an output layer;

[0116] The LSTM layer is used to capture time series features, the fully connected layer is used for feature fusion, and the output layer gives the insulation state prediction;

[0117] Insulation prediction model, the expression is:

[0118]

[0119] Among them, △I(t) represents the change of insulation state at time point t, α is a positive coefficient used to adjust the influence of temperature on insulation state change, e -λ(t) is an exponential decay function based on temperature, and β is another positive coefficient used to adjust the degree of influence of humidity on the change of insulation state. It means that as the humidity increases, the function value gradually increases from -0.5 to close to 0.5, indicating that the negative impact of humidity on the insulation state increases. μ is a positive number that controls the steepness of the humidity effect. γ is a positive coefficient that is used to adjust the intensity of the combined impact of other physical parameters. w k represents the weight factor of k(T,H), w D represents the weight factor of D(H), w p represents the weight factor of P(T,H,E);

[0120] The output layer of the insulation prediction model is expressed as:

[0121]

[0122] in, It represents the predicted value of the insulation state at time point t, and I(t-1) represents the actual insulation state value at the previous moment;

[0123] Extract humidity, temperature and insulation condition data related to transformer insulation condition from historical data sets;

[0124] Interpolation methods are used to fill missing data to ensure the integrity and consistency of the data set, and Z-score standardization is used to normalize the data so that features of different scales are at the same order of magnitude, improving the efficiency of model training;

[0125] Based on the original humidity and temperature, principal component analysis (PCA) is used to extract the most influential cross-features for insulation status prediction, which is used to capture the interactive effects of humidity and temperature on insulation status.

[0126] The cross-features are input into the insulation prediction model to output the predicted value of the insulation state, that is, the variation trend of the insulation field strength;

[0127] The specific steps for extracting cross features are:

[0128] Set the standardized data matrix to X;

[0129] Calculate the covariance matrix C of X, the expression is:

[0130]

[0131] Where N represents the number of samples;

[0132] Calculate the eigenvalues ​​and eigenvectors of the covariance matrix C, the expression is:

[0133] Cv=λv;

[0134] Among them, V represents the eigenvector, λ represents the eigenvector;

[0135] Arrange the eigenvectors in descending order of eigenvalues. The larger the eigenvalue, the more important the principal component represented by the corresponding eigenvector.

[0136] Accumulate the proportion of each eigenvalue to the total eigenvalue, and select the principal components whose cumulative contribution rate of the first k eigenvectors reaches 80% as the final feature set;

[0137] Project the eigenvector matrix into the matrix consisting of the first k selected eigenvectors;

[0138] The projection formula is:

[0139] Y=XV k ;

[0140] Among them, V k represents the matrix composed of the first k eigenvectors, and Y represents the projected data matrix.

[0141] S4 processes the real-time data and inputs it into the insulation prediction model to obtain the real-time insulation state prediction value, and dynamically adjusts the parameters of the insulation prediction model;

[0142] Collect humidity, temperature and electric field data inside the transformer in real time through a sensor network;

[0143] Before data is transmitted to the central processor, Manchester coding and Kalman filter are used to process real-time data to improve data quality and transmission robustness;

[0144] The processed real-time data is passed as input to the insulation prediction model, and the model parameters are updated, especially those that depend on temperature, humidity and electric field, humidity diffusion coefficient and charge density, which are dynamically adjusted according to the real-time data to reflect the current working conditions;

[0145] Use the PSO algorithm to dynamically update thermal conductivity, humidity diffusion coefficient and charge density to ensure that the model parameters are consistent with the actual working conditions and improve prediction accuracy;

[0146] After the real-time data is input into the insulation prediction model, the real-time insulation status prediction value is output;

[0147] The mean square error (MSE) is selected as the loss function to measure the difference between the predicted value of the insulation prediction model and the actual insulation state. The expression is:

[0148]

[0149] Where N represents the number of samples, y i represents the true insulation state value of the i-th sample, represents the insulation state value predicted by the i-th sample;

[0150] Set the threshold LO;

[0151] When the MSE value exceeds the preset threshold, it indicates that there is a large deviation between the model prediction and the actual state. It is necessary to adjust the model parameters through the back propagation algorithm and optimize the model to reduce the MSE and improve the prediction accuracy.

[0152] The continuous input of real-time data and the dynamic adjustment of model parameters form a closed loop, which continuously optimizes the prediction ability of the model and ensures the real-time and accuracy of insulation status assessment.

[0153] S5 establishes an abnormal warning mechanism based on the real-time insulation status prediction value. Once an abnormal insulation status is detected, the system immediately starts the warning and notifies the operation and maintenance personnel to conduct inspection and maintenance;

[0154] The anomaly detection algorithm is used to identify abnormal fluctuations in the real-time insulation status prediction value. The expression is:

[0155]

[0156] Among them, Θ(t) is a dynamic threshold, which is used to determine whether the real-time insulation status fluctuates abnormally. When the real-time insulation status value exceeds this threshold, the system will determine it as an abnormal situation. Indicates the average insulation state value, σ t represents the standard deviation, Φ -1 represents the inverse cumulative distribution function, and p represents the probability threshold of abnormal occurrence, which can be set by yourself;

[0157] Real-time monitoring of insulation status prediction values. Once the prediction values ​​deviate from the normal range, the early warning mechanism is triggered immediately.

[0158] When the central processor detects an abnormality, it immediately sends an alarm to the operation and maintenance personnel and starts the self-diagnosis program to find the potential cause of the fault and take timely measures to prevent the insulation condition from deteriorating;

[0159] Compare the dynamic threshold with the real-time insulation status prediction value, and then divide it into normal interval, warning interval, slight abnormal interval and serious abnormal interval;

[0160] Set a certain safety margin as △;

[0161] when When the system is in normal working condition, basic maintenance such as regular cleaning, lubrication, tightening and functional testing should be performed according to the equipment's usage frequency and historical maintenance records;

[0162] when When the system enters the early warning state, in addition to routine maintenance, it is necessary to increase the frequency of online monitoring, conduct remote diagnosis, and arrange professional personnel to conduct on-site inspections to ensure that any potential problems can be discovered and handled in a timely manner;

[0163] when When the system determines that it is a minor abnormality, the operation and maintenance personnel will be notified immediately to conduct on-site inspections, and the self-diagnosis program will be started to find possible fault sources and perform preliminary repair and adjustment measures;

[0164] when When the system faces serious abnormalities and a major failure is about to occur, it will be shut down for inspection immediately to avoid greater damage. At the same time, a high-level troubleshooting process will be initiated and damaged parts will be replaced if necessary.

[0165] This embodiment also provides a computer device, which is suitable for the case of a method for calculating the insulation field strength of a transformer under the influence of moisture and temperature distribution, including: a memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions to implement the method for calculating the insulation field strength of a transformer under the influence of moisture and temperature distribution proposed in the above embodiment.

[0166] The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0167] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the method for calculating the insulation field strength of a transformer under the influence of moisture and temperature distribution proposed in the above embodiment is implemented; the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (Static Random Access Memory, referred to as SRAM), electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory, referred to as EEPROM), erasable programmable read-only memory (Erasable Programmable Read Only Memory, referred to as EPROM), programmable read-only memory (Programmable Red-Only Memory, referred to as PROM), read-only memory (Read-Only Memory, referred to as ROM), magnetic storage, flash memory, magnetic disk or optical disk.

[0168] In summary, the present invention achieves comprehensive monitoring and analysis of the temperature, humidity and electric field distribution inside the transformer by: constructing a multi-physical field coupling simulation platform. In particular, the present invention uses deep neural network technology to construct an insulation prediction model, and dynamically adjusts the model parameters in combination with real-time data. It can not only accurately predict the changing trend of the insulation state, but also quickly activate the early warning mechanism when an abnormal situation occurs, and provide timely fault diagnosis information for operation and maintenance personnel. The present invention particularly focuses on the intelligent analysis and early warning of the impact of humidity and temperature on the insulation state, providing a new solution for the health management of transformers, which helps to improve the operating safety and economy of the power system.

[0169] Example 2

[0170] Referring to Table 1, which is the second embodiment of the present invention, in order to further verify the advancement of the present invention, experimental simulation data of the transformer insulation field strength calculation method under the influence of moisture and temperature distribution are provided.

[0171] In order to verify the effectiveness and innovativeness of the invention, a series of experiments were carried out to simulate the influence of humidity, temperature and electric field distribution on the insulation state of the transformer under different working conditions. The experiments were carried out under laboratory conditions and used a multi-physics field coupling simulation platform, which integrated temperature, humidity and electric field sensors, as well as an insulation prediction model based on a deep neural network. In the experiment, an array of 100 humidity-sensitive elements was first constructed, and the maximum distance between the elements was set to 0.5 meters to ensure effective coverage of the entire internal space of the transformer. These elements collected data at a frequency of 10 Hz and transmitted it to the central processor for processing through LoRa wireless communication technology. In addition, a variety of working conditions were set, including different temperatures (20°C to 80°C), humidity (30% RH to 90% RH) and electric field strength (0V / m to 1000V / m) conditions to comprehensively evaluate the changing trend of the insulation state.

[0172] During the experiment, the insulation prediction model constructed using the long short-term memory network (LSTM) was trained based on a historical data set. The data set included humidity, temperature and insulation status data of the transformer under various working conditions in the past year, totaling 10,000 records. After the model training was completed, real-time data was used for testing to verify its prediction performance. In order to evaluate the prediction accuracy of the model, the mean square error (MSE) was set as the loss function, with the goal of controlling the MSE within 0.01. During the experiment, a set of abnormal warning mechanisms was also designed based on the real-time insulation status prediction value. When the deviation between the predicted value and the average value exceeds 3 standard deviations, the system will automatically start the warning and notify the operation and maintenance personnel to check.

[0173] The details are shown in Table 1:

[0174] Table 1 Experimental record table

[0175]

[0176]

[0177] Through the analysis of experimental data, it can be observed that with the increase of temperature, humidity and electric field strength, the insulation state of the transformer shows a clear downward trend. For example, in working conditions 001 to 100, the insulation state drops from 0.98 to 0.65, which is consistent with the expected theoretical analysis. More importantly, the insulation prediction model based on deep neural network shows excellent prediction ability. In all test conditions, the mean square error (MSE) between the insulation state predicted by the model and the actual value is less than 0.01, which shows that the model can accurately capture the dynamic trend of the insulation state as the working conditions change.

[0178] It is particularly noteworthy that under extreme conditions of high temperature, high humidity and high electric field strength, the model can still maintain a low MSE, which reflects the robustness and stability of the model in complex environments. For example, in operating condition 100, although the actual insulation state has dropped to 0.65, the model predicted value is 0.66 and the MSE is only 0.001, which fully demonstrates the prediction accuracy of the model. In addition, the abnormal warning mechanism also played a key role in the experiment, successfully identifying all cases where the predicted values ​​deviated from the mean value by more than 3 standard deviations, ensuring the safe operation of the system.

[0179] In summary, the multi-physics field coupling simulation platform and the insulation prediction model based on deep neural network proposed in the present invention can not only accurately monitor and predict the insulation state changes of the transformer under different working conditions, but also timely warn of potential insulation failures, significantly improving the operation safety and reliability of the power system. This achievement has important innovative significance and broad application prospects in the field of power engineering.

[0180] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for calculating transformer insulation field strength under the influence of moisture and temperature distribution, Features: include, Build a multi-physics coupling simulation platform to monitor and analyze the temperature, humidity, and electric field distribution inside the transformer; Setting up a humidity sensitive element array, collecting humidity data, and transmitting the humidity data to a central processor for processing; The insulation prediction model is built through a deep neural network algorithm. The historical humidity, historical temperature and insulation status data are input into the insulation prediction model to complete the training, and the change trend of the insulation field strength under specific working conditions is predicted; The real-time data is processed and input into the insulation prediction model to obtain the real-time insulation state prediction value, and the parameters of the insulation prediction model are dynamically adjusted; An abnormal warning mechanism is established based on the real-time insulation status prediction value. Once an abnormal insulation status is detected, the system immediately activates the warning and notifies the operation and maintenance personnel to conduct inspection and maintenance.

2. The method for calculating transformer insulation field strength under the influence of moisture and temperature distribution as claimed in claim 1, Features: The multi-physics field coupling simulation platform is constructed to monitor and analyze the temperature, humidity and electric field distribution inside the transformer. The specific steps are as follows: A multi-physics coupling model is constructed based on partial differential equations (PDEs) to describe the dynamic changes of temperature, humidity and electric field. The core equations of the multi-physics coupling model include the temperature diffusion equation, the humidity diffusion equation and the electric field distribution equation. The particle swarm optimization algorithm is used to dynamically update the PSO thermal conductivity, humidity diffusion coefficient and charge density; The central processing unit, sensor network, wireless communication module, data preprocessing algorithm and multi-physics coupling model are integrated to obtain a multi-physics coupling simulation platform, wherein the central processing unit is used to coordinate the processing of data flow and the operation of the multi-physics coupling model.

3. The method for calculating transformer insulation field strength under the influence of moisture and temperature distribution as claimed in claim 2, Features: The humidity sensitive element array is set to collect humidity data, and the humidity data is transmitted to the central processor for processing, and the specific steps are as follows: Inside the transformer, selected locations including windings, core and tank wall are integrated with temperature sensors, humidity sensors and electric field sensors; Divide the area of ​​the selected site into a uniform grid, with each grid cell representing a potential monitoring point. The size of the grid should take into account the expected gradient of humidity change and the required data resolution. The distance between the elements is calculated, and a humidity sensitive element is placed at the center of each grid cell to form a two-dimensional array; Each humidity sensor has a built-in microcontroller with a sampling frequency f s =10Hz continuous collection of humidity data; Manchester encoding is used to encode humidity data; After completing Manchester encoding of humidity data, encapsulate it into data packets; After receiving the data, the central processor first decodes the data packet and then pre-processes the humidity data using a Kalman filter to eliminate noise and improve data quality; Distributed Kalman filter algorithm is used to fuse multiple component data from the same area; Based on the real-time humidity data collected by the humidity sensitive element array, an abnormal value detection algorithm is used to detect abnormal values. When it is detected that the humidity in a certain area exceeds the normal range, an early warning is triggered immediately; The humidity diffusion coefficient function in the multi-physics field coupling model is dynamically adjusted according to real-time humidity data to ensure that the model parameters are consistent with the actual working conditions.

4. The method for calculating transformer insulation field strength under the influence of moisture and temperature distribution as claimed in claim 3, Features: The insulation prediction model is constructed by using a deep neural network algorithm, and the specific steps are as follows: Select the long short-term memory network LSTM as the basis of the insulation prediction model; The insulation prediction model includes an input layer, several LSTM layers, a fully connected layer, and an output layer.

5. The method for calculating transformer insulation field strength under the influence of moisture and temperature distribution as claimed in claim 4, Features: The insulation prediction model is trained based on historical humidity, historical temperature and insulation status data to predict the change trend of insulation field strength. The specific steps are as follows: Extract humidity, temperature and insulation condition data related to transformer insulation condition from historical data sets; Interpolation method was used to fill missing data, and Z-score standardization was used to normalize the data; Based on the original humidity and temperature, principal component analysis (PCA) is used to extract the most influential cross-features for insulation status prediction, which is used to capture the interactive effects of humidity and temperature on insulation status. The cross-features are input into the insulation prediction model to output the predicted value of the insulation status.

6. The method for calculating transformer insulation field strength under the influence of moisture and temperature distribution as claimed in claim 5, Features: The real-time data is processed and input into the insulation prediction model to obtain the real-time insulation state prediction value, and the parameters of the insulation prediction model are dynamically adjusted. The specific steps are: Collect humidity, temperature and electric field data inside the transformer in real time through a sensor network; Before data is transmitted to the central processor, Manchester coding and Kalman filter are used to process real-time data to improve data quality and transmission robustness; The processed real-time data is passed as input to the insulation prediction model and the model parameters are updated; The PSO algorithm is used to dynamically update thermal conductivity, humidity diffusion coefficient and charge density; After the real-time data is input into the insulation prediction model, the real-time insulation status prediction value is output; The mean square error (MSE) is selected as the loss function to measure the difference between the predicted value of the insulation prediction model and the actual insulation status; Set the threshold LO; When the MSE value exceeds the preset threshold, it indicates that there is a large deviation between the model prediction and the actual state. It is necessary to adjust the model parameters through the back propagation algorithm and optimize the model to reduce the MSE and improve the prediction accuracy.

7. The method for calculating transformer insulation field strength under the influence of moisture and temperature distribution as claimed in claim 6, Features: The abnormal warning mechanism is established based on the real-time insulation state value prediction value. Once the insulation state is detected to be abnormal, the system immediately starts the warning and notifies the operation and maintenance personnel to perform inspection and maintenance. The specific steps are as follows: Use anomaly detection algorithms to identify abnormal fluctuations in real-time insulation status prediction values; Real-time monitoring of insulation status prediction values. Once the prediction values ​​deviate from the normal range, the early warning mechanism is triggered immediately. After detecting an abnormality, the central processor immediately sends an alarm to the operation and maintenance personnel and starts the self-diagnosis program to find the potential cause of the fault and take timely measures to prevent the insulation condition from deteriorating.

8. The method for calculating transformer insulation field strength under the influence of moisture and temperature distribution as claimed in claim 7, Features: The real-time monitoring of the insulation status prediction value, once the prediction value is found to deviate from the normal range, immediately triggers the early warning mechanism, specifically: Compare the dynamic threshold with the real-time insulation status prediction value, and then divide it into normal interval, warning interval, slight abnormal interval and serious abnormal interval; Set a certain safety margin as △; when When the system is in normal working condition, basic maintenance such as regular cleaning, lubrication, tightening and functional testing should be performed according to the equipment's usage frequency and historical maintenance records; when When the system enters the early warning state, in addition to routine maintenance, it is necessary to increase the frequency of online monitoring, conduct remote diagnosis, and arrange professional personnel to conduct on-site inspections to ensure that any potential problems can be discovered and handled in a timely manner; when When the system determines that it is a minor abnormality, the operation and maintenance personnel will be notified immediately to conduct on-site inspections, and the self-diagnosis program will be started to find possible fault sources and perform preliminary repair and adjustment measures; when When the system faces serious abnormalities and a major failure is about to occur, it will be shut down for inspection immediately to avoid greater damage. At the same time, a high-level troubleshooting process will be initiated and damaged parts will be replaced if necessary.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program. Features: When the processor executes the computer program, the steps of the method for calculating the insulation field strength of a transformer under the influence of moisture and temperature distribution as described in any one of claims 1 to 8 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, Features: When the computer program is executed by a processor, the steps of the method for calculating the insulation field strength of a transformer under the influence of moisture and temperature distribution as described in any one of claims 1 to 8 are implemented.

Images