Online monitoring and early warning method and system for gas in transformer oil
By monitoring the gas composition and load data in the transformer oil online, and establishing an early warning model using a multi-layer LSTM model and attention mechanism, solving the problems of monitoring time and complex operation in the existing technology, realizing timely warning and operation of transformer failures stability.
Patent Information
- Application Number
- CN202410839058.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-06-26
- Publication Date
- 2024-11-15
AI Technical Summary
The existing gas monitoring methods in transformer oil require on-site sampling and experimental analysis, which takes a long time and requires the operation level of the tester, so timely warning cannot be achieved.
The online monitoring method is adopted to detect gas composition data by transporting transformer oil to the oil-gas separation device, and adjust the oil pump speed based on the real-time detected gas content and transformer load data. A gas monitoring and early warning model is established using a multi-layer LSTM model and attention mechanism. Perform data dimensionless processing and real-time analysis to issue early warning signals.
It realizes timely early warning of transformer failures, improves monitoring reliability and stability, reduces the risk of failure, ensures the normal operation of the transformer and extends the service life.
Smart Images

Figure CN118962001A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of transformer oil gas monitoring, and in particular to an online monitoring and early warning method and system for gas in transformer oil. Background Art
[0002] Transformers have a wide range of applications and are one of the most important electrical products in power systems, especially in power transmission and transformation lines. In order to ensure the reliability and economy of power supply in power systems, online monitoring and fault diagnosis technology for transformers has emerged. Transformer oil is an insulating oil that plays the role of insulation, cooling and arc extinguishing in electrical equipment. Its quality is of great significance to ensure the safe operation of equipment. Transformer oil is a mineral oil obtained by distilling and refining natural petroleum. It is a mixed hydrocarbon composed of various hydrocarbons of different molecular weights. When faults such as overheating, partial discharge, electric sparks and continuous arcs occur during the operation of the transformer, the hydrogen and low molecular hydrocarbon gases formed in the oil will dissolve in the oil to form bubbles. These gases will affect the dielectric properties of the transformer oil and affect the normal operation of the transformer. The volume of these gases is closely related to the type and severity of the transformer fault. The detection of the volume of dissolved gas in transformer oil is of great significance for the early diagnosis and development of some potential faults of the transformer. Accurately measuring the volume of each component of the dissolved gas in the oil and improving the reliability of the experimental results are the premise and basis for accurately evaluating the operating status of the equipment.
[0003] Most of the existing transformer oil gas monitoring methods use gas chromatography, which requires on-site sampling and experimental analysis, resulting in a long time consumption and certain requirements on the operating level of the detection personnel, and cannot provide timely warnings. Summary of the invention
[0004] In view of the above-mentioned problems, the present invention is proposed.
[0005] Therefore, in order to solve the above technical problems, the present invention provides the following technical solutions: an online monitoring and early warning method for gas in transformer oil, comprising: delivering transformer oil to an oil-gas separation device and detecting separated gas composition data, and adjusting the oil pump speed according to the real-time detected gas content data and transformer load data;
[0006] The historical load data and gas content data of the transformer are dimensionless, and a multi-layer LSTM model is used to integrate the attention mechanism to establish a gas monitoring and early warning model. The gas monitoring and early warning model is trained and verified to obtain the predicted probability of transformer failure.
[0007] The detected gas component content and transformer oil quality are analyzed in real time, and the gas monitoring and early warning results are compensated and corrected to obtain compensation prediction results;
[0008] When the compensation prediction result exceeds the safety range, an alarm signal is issued according to the compensation prediction result to complete the early warning.
[0009] As a preferred solution of the online monitoring and early warning method for gas in transformer oil of the present invention, wherein: the gas composition data detected and separated includes:
[0010] Use gas sensors to monitor the gas composition in transformer oil in real time and collect gas concentration data;
[0011] The gas sensor comprises a plurality of gas detection sensors for detecting different types of gases, and the plurality of gas detection sensors for detecting different types of gases are arranged in an array;
[0012] The gas detection sensors include a hydrogen detection sensor, a carbon monoxide detection sensor, a carbon dioxide detection sensor, a methane detection sensor, an acetylene detection sensor, an ethylene detection sensor and an ethane detection sensor.
[0013] As a preferred solution of the online monitoring and early warning method for gas in transformer oil of the present invention, wherein: the oil pump speed adjustment includes:
[0014] According to the real-time detected gas content data and transformer load data, the real-time detected gas content data and transformer load data are calculated to obtain improved gas content data and improved transformer load data; the calculation formula of the improved gas content data is as follows:
[0015]
[0016] in, To improve the gas content data, S(τ) is the instantaneous gas concentration detected at time τ, t is the time, T g is the gas concentration detection time window, α is the gas attenuation coefficient;
[0017] The calculation formula for improving transformer load data is as follows:
[0018]
[0019] in, To improve transformer load data, β k is the weighting coefficient, N is the number of sampling points, Δt is the sampling interval, P(τ) is the load power of the transformer at time τ, and k is a discrete point in time;
[0020] Calculate the optimal oil pump speed according to the improved gas content data and the improved transformer load data, and send the adjustment signal to the oil pump control center;
[0021] The formula for calculating the optimum oil pump speed is as follows:
[0022]
[0023] Among them, Z is the normalization constant, γ is the feedback coefficient, δ n is the historical adjustment coefficient, M is the number of historical feedback points, V(t) is the oil pump speed at time t, and n is the index of historical feedback points;
[0024] The oil pump control center adjusts the speed of the oil pump according to the received regulation signal and accurately controls the operation of the oil pump through the frequency converter;
[0025] Regularly calibrate gas detection sensors to ensure the accuracy of detection data, record the results and time of each calibration, and generate a calibration log.
[0026] As a preferred solution of the online monitoring and early warning method for gas in transformer oil of the present invention, the establishment of a gas monitoring and early warning model includes:
[0027] The historical load data and gas content data of the transformer are collected and dimensionless processed. The dimensionless calculation formula is as follows:
[0028]
[0029] Among them, X t is the dimensionless data point, μ x and σ x are the mean and standard deviation of the original data respectively;
[0030] Use sliding window techniques and smoothing to eliminate noise and outliers in the data;
[0031] Divide the cleaned data into a training set and a validation set;
[0032] Train the model on the training set;
[0033] The model is tuned and verified on the validation set, which includes real-time detection data and is dynamically updated according to the working status of the transformer;
[0034] Construct a neural network model suitable for time series data, using a multi-layer LSTM model. The calculation formula of the multi-layer LSTM model is as follows:
[0035]
[0036] Among them, LSTM l represents the LSTM unit of the lth layer, L is the number of LSTM layers, and are the hidden state and cell state of the lth layer at time t respectively;
[0037] At the same time, the attention mechanism is introduced to strengthen the model's attention to key moments. The calculation formula for the attention output of the layer LSTM model is as follows:
[0038]
[0039] Among them, W q , W k , W v and b q , b k , b v are the weight matrix and bias term, A t is the output after the attention weights are applied, Q t is the query vector at time t, K t is the key vector at time t, V t is the value vector at time t, d k is the key vector K t Dimensions;
[0040] The final warning model output is calculated based on the attention output of the multi-layer LSTM model to obtain the predicted probability of transformer failure. The calculation formula is as follows:
[0041]
[0042] in, is the predicted probability of transformer failure, W o and b o are the weight matrix and bias term of the output layer respectively.
[0043] As a preferred solution of the online monitoring and early warning method for gas in transformer oil described in the present invention, the tuning and verification of the model on the verification set includes:
[0044] Use time series cross validation and optimize time series data to ensure the generalization ability and stability of the model;
[0045] Divide the original time series data X into N subsequences X i ;
[0046] For each subsequence X i , perform sliding window segmentation processing to obtain the segmented time series subset X i,j , where j = 1, 2, ..., M;
[0047] For each X i,j Use the gas monitoring and early warning model to make predictions and obtain prediction results;
[0048] Calculate the prediction error based on the prediction results;
[0049] The cross-validation error is obtained by weighted summation of all prediction errors. The specific calculation formula is as follows:
[0050]
[0051] Where CV(X) is the cross validation error, w i,j Corresponding to X i,j The weight, E i,j To calculate the prediction error, is the prediction result of the gas monitoring and early warning model, X i,j is the jth segment of the i-th subsequence;
[0052] Adjust the time series data, and when the cross validation error CV(X) reaches the minimum value, the optimized time series data is obtained.
[0053] As a preferred solution of the online monitoring and early warning method for gas in transformer oil of the present invention, the compensation and correction of the gas monitoring and early warning results include:
[0054] Collect real-time monitoring data of gas component content and transformer oil quality, and perform data cleaning and preprocessing;
[0055] Clean and preprocess the collected gas component content and transformer oil quality data to remove noise and outliers;
[0056] The cleaned and pre-processed data is input into the compensation and correction model to compensate and correct the gas monitoring warning results. The calculation formula is as follows:
[0057]
[0058] in, is the original gas monitoring warning result, x is the real-time analysis compensation data, m is the compensation weight factor, b is the compensation bias term, To compensate for the prediction results.
[0059] As a preferred solution of the online monitoring and early warning method for gas in transformer oil of the present invention, the step of issuing an alarm signal according to the compensation prediction result includes:
[0060] Compare the compensation prediction results with the set safety threshold. When the compensation prediction results exceed the safety range, identify potential failure risks and issue early warning signals.
[0061] The alarm signal includes sound and light alarm devices and remote alarm functions to ensure that relevant personnel can receive alarm information in a timely manner;
[0062] Record detailed information of all alarm events, including alarm time, alarm reason, gas content data and load data, and form an alarm log with detailed information of all alarm events;
[0063] Regularly analyze the frequency and causes of alarm events based on alarm logs, and optimize early warning models and system parameters;
[0064] Develop an emergency response plan. When operators receive an alarm signal, they will perform corresponding inspection and maintenance measures according to the emergency response plan;
[0065] The emergency response plan includes,
[0066] Level 1 alarm response includes when the operator receives a level 1 alarm signal, immediately stopping the operation of the transformer to avoid further damage, quickly checking the key parts of the transformer, including oil temperature, oil level and gas content, recording all data and uploading it to the central control center;
[0067] Level 2 alarm response includes when the level 2 alarm signal is triggered, the operator will conduct local inspection and maintenance according to the specific alarm cause to ensure that it does not affect the operation of the entire system, record all inspection data in detail, including gas content and load changes, and track the cause of the alarm;
[0068] The third-level alarm includes that when the third-level alarm signal is triggered, the operator continuously observes the operating status of the transformer through the remote monitoring system, records abnormal conditions, and regularly performs maintenance and service on the transformer.
[0069] Another object of the present invention is to provide an online monitoring and early warning system for gas in transformer oil. To solve the above technical problems, the present invention provides the following technical solutions: an online monitoring and early warning system for gas in transformer oil, comprising: a data acquisition module, an early warning modeling module, a data correction module, an alarm processing module, and an emergency response module;
[0070] The data acquisition module is used to transport transformer oil to the oil-gas separation device, detect gas composition data and monitor gas content data and transformer load data in real time, and adjust the oil pump speed;
[0071] The early warning modeling module is used to perform dimensionless processing on historical load data and gas content data, build a multi-layer LSTM model, and integrate the attention mechanism to establish a gas monitoring early warning model;
[0072] The data correction module is used to analyze the detected gas component content and transformer oil quality data in real time, and to compensate and correct the gas monitoring and early warning results;
[0073] The alarm processing module is used to compare the compensation prediction result with the set safety threshold, and when the compensation prediction result exceeds the safety range, an alarm signal is issued to complete the early warning;
[0074] The emergency response module is used to formulate an emergency response plan and execute corresponding inspection and maintenance measures according to the emergency response plan to ensure timely handling of abnormal situations.
[0075] A computer device comprises a memory and a processor, wherein the memory stores a computer program, and wherein the processor implements the steps of the above-mentioned method for online monitoring and early warning of gas in transformer oil when executing the computer program.
[0076] A computer-readable storage medium having a computer program stored thereon, characterized in that when the computer program is executed by a processor, the steps of the above-mentioned method for online monitoring and early warning of gas in transformer oil are implemented.
[0077] Beneficial effects of the present invention: The present invention proposes an online monitoring and early warning method for gas in transformer oil, which aims to timely discover possible failure risks of the transformer, so as to take preventive measures to ensure the normal operation of the transformer and extend its service life; by monitoring the gas composition data in the transformer oil, and combining the real-time transformer load data, the oil pump speed is adjusted to keep the gas content in the oil stable, effectively preventing transformer failures caused by gas anomalies, and improving reliability and stability; a gas monitoring and early warning model is established, and training and verification are carried out by combining the multi-layer LSTM model and the attention mechanism with the historical load data and the gas content data to realize the prediction of transformer failure; through real-time data analysis and correction, potential failure risks are discovered in time, and early warning signals are issued to ensure the safe operation of the transformer; according to the early warning signal, the emergency response plan is executed, and corresponding measures are taken to ensure the safe and stable operation of the power system, and the detailed information of the alarm event is recorded to provide data support for subsequent system optimization and maintenance; the steps are coordinated with each other to form a complete set of online monitoring and early warning system for gas in transformer oil, which can effectively prevent and respond to various possible failure situations of the transformer. BRIEF DESCRIPTION OF THE DRAWINGS
[0078] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.
[0079] Figure 1 This is a schematic diagram of the structure of a terminal in the first embodiment of the present invention;
[0080] Figure 2 It is a schematic diagram of a flow chart in the first embodiment of the present invention;
[0081] Figure 3 FIG. 4 is a schematic diagram of the structure of a computer device in a third embodiment of the present invention. DETAILED DESCRIPTION
[0082] In order to make the above-mentioned purposes, 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 drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.
[0083] 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.
[0084] Example 1
[0085] Reference Figure 1 , which is an embodiment of the present invention, provides an online monitoring and early warning method for gas in transformer oil.
[0086] First, the online monitoring and early warning method for gas in transformer oil provided by this application can be applied to Figure 1 In the terminal shown. Figure 1 As shown, the terminal may include one or two ( Figure 1 Only one is shown in the figure) processor and memory for storing data, wherein the processor may include but is not limited to a processing system such as a microprocessor MCU or a programmable logic device FPGA. The above terminal may also include a transmission device and an input and output device for communication functions. It can be understood by ordinary technicians in this field that Figure 1 The structure shown is only for illustration and does not limit the structure of the above terminal. Figure 1 More or fewer components as shown, or with Figure 1 Different configurations shown.
[0087] The memory can be used to store computer programs, such as the computer program corresponding to the online monitoring and early warning method for gas in transformer oil in this embodiment. The processor executes various functional applications and data processing by running the computer program stored in the memory, that is, realizing the above method. The memory may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage systems, flash memory, or other non-volatile solid-state memory. In some instances, the memory may further include a memory remotely arranged relative to the processor, and these remote memories may be connected to the terminal via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.
[0088] The transmission device is used to receive or send data via a network. The above-mentioned network includes a wireless network provided by a communication provider of the terminal. In one example, the transmission device includes a network adapter (Network Interface Controller, referred to as NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device can be a radio frequency (Radio Frequency, referred to as RF) module, which is used to communicate with the Internet wirelessly.
[0089] like Figure 2 As shown, the embodiment of the present invention provides an online monitoring and early warning method for gas in transformer oil, and the method is applied to Figure 1 The terminal in is used as an example to illustrate, including the following steps:
[0090] S1: The transformer oil is transported to the oil-gas separation device and the separated gas composition data is detected. At the same time, the oil pump speed is adjusted according to the real-time detected gas content data and transformer load data.
[0091] S1.1: Use gas sensors to monitor the gas composition in transformer oil in real time and collect gas concentration data;
[0092] The gas sensor includes a plurality of gas detection sensors for detecting different types of gases, and the plurality of gas detection sensors for detecting different types of gases are arranged in an array;
[0093] The gas detection sensors include hydrogen detection sensors, carbon monoxide detection sensors, carbon dioxide detection sensors, methane detection sensors, acetylene detection sensors, ethylene detection sensors and ethane detection sensors.
[0094] S1.2: Based on the real-time detected gas content data and transformer load data, taking into account gas attenuation and normalization factors, the real-time detected gas content data and transformer load data are calculated to obtain improved gas content data and improved transformer load data;
[0095] The calculation formula for improving gas content data is as follows:
[0096]
[0097] in, To improve the gas content data, S(τ) is the instantaneous gas concentration detected at time τ, t is the time, T g is the gas concentration detection time window, α is the gas attenuation coefficient;
[0098] The calculation formula for improving transformer load data is as follows:
[0099]
[0100] in, To improve transformer load data, β k is the weighting coefficient, N is the number of sampling points, Δt is the sampling interval, P(τ) is the load power of the transformer at time τ, and k is a discrete point in time;
[0101] Calculate the optimal oil pump speed according to the improved gas content data and the improved transformer load data, and send the adjustment signal to the oil pump control center;
[0102] The formula for calculating the optimum oil pump speed is as follows:
[0103]
[0104] Among them, Z is the normalization constant, γ is the feedback coefficient, δ n is the historical adjustment coefficient, M is the number of historical feedback points, V(t) is the oil pump speed at time t, and n is the index of historical feedback points;
[0105] The oil pump control center adjusts the speed of the oil pump according to the received regulation signal, and accurately controls the operation of the oil pump through the frequency converter to ensure that the separation efficiency of the oil-gas separation device reaches the optimal state;
[0106] Regularly calibrate gas detection sensors to ensure the accuracy of detection data, record the results and time of each calibration, and generate a calibration log.
[0107] S2: The historical load data and gas content data of the transformer are dimensionlessly processed, and a gas monitoring and early warning model is established by adopting a multi-layer LSTM model and integrating the attention mechanism, and the gas monitoring and early warning model is trained and verified.
[0108] S2.1: Collect historical load data and gas content data of transformers and perform dimensionless processing to ensure data quality;
[0109]
[0110] Among them, X t is the dimensionless data point, μ x and σ x are the mean and standard deviation of the original data respectively;
[0111] Special consideration will be given to time during the data preprocessing phase;
[0112] Processing of sequence data, using sliding window technology and smoothing to eliminate noise and outliers in the data;
[0113] Divide the cleaned data into training set, validation set and test set;
[0114] The training set is used for model training, the validation set is used for model tuning, and the test set is used for model performance evaluation;
[0115] To better simulate real-time scenarios, the test set will contain data from real-time detection and will be dynamically updated according to the working status of the transformer;
[0116] Build a neural network model suitable for time series data and use a multi-layer LSTM model to better capture the time series characteristics of the transformer working status;
[0117] The calculation formula of the multi-layer LSTM model is as follows:
[0118]
[0119] Among them, LSTM l represents the LSTM unit of the lth layer, L is the number of LSTM layers, and are the hidden state and cell state of the lth layer at time t respectively;
[0120] At the same time, the attention mechanism is introduced to strengthen the model's attention to key moments to improve the accuracy and stability of predictions;
[0121] The calculation formula for the attention output of the multi-layer LSTM model is as follows:
[0122]
[0123] Among them, W q , W k , W v and b q , b k , b v are the weight matrix and bias term, A t is the output after the attention weights are applied, Q t is the query vector at time t, K tis the key vector at time t, V t is the value vector at time t, d k is the key vector K t Dimensions;
[0124] The final warning model output is calculated based on the attention output of the multi-layer LSTM model to obtain the predicted probability of transformer failure. The calculation formula is as follows:
[0125]
[0126] in, is the predicted probability of transformer failure, W o and b o are the weight matrix and bias term of the output layer respectively.
[0127] S2.2: Tune and validate the model on the validation set, use time series cross-validation and optimize time series data to ensure the generalization ability and stability of the model.
[0128] Divide the original time series data X into N subsequences X i .
[0129] First, for each subsequence X i , perform sliding window segmentation processing to obtain the segmented time series subset X i,j , where j = 1, 2, ..., M, X i,j represents the jth segment of the i-th subsequence;
[0130] For each X i,j , using the gas monitoring and early warning model to make predictions and obtain prediction results;
[0131] Calculate the prediction error based on the prediction results;
[0132] The cross-validation error is obtained by weighted summation of all prediction errors. The specific calculation formula is as follows:
[0133]
[0134] Where CV(X) is the cross validation error, w i,j Corresponding to X ij The weight, E i,j To calculate the prediction error, is the prediction result of the gas monitoring and early warning model, X i,j is the jth segment of the i-th subsequence;
[0135] Optimize the time series data. When the cross validation error CV(X) reaches the minimum value, the optimized time series data is obtained.
[0136] S3: Analyze the detected gas component content and transformer oil quality in real time, compensate and correct the gas monitoring and early warning results, and obtain compensated prediction results.
[0137] S3.1: Collect real-time monitoring data of gas component content and transformer oil quality, and perform data cleaning and preprocessing to ensure data quality;
[0138] Clean and preprocess the collected gas component content and transformer oil quality data to remove possible noise and outliers;
[0139] The cleaned and pre-processed data is input into the compensation and correction model to compensate and correct the gas monitoring warning results. The calculation formula is as follows:
[0140]
[0141] in, is the original gas monitoring warning result, x is the real-time analysis compensation data, m is the compensation weight factor, b is the compensation bias term, To compensate for the prediction results.
[0142] S4: When the compensation prediction result exceeds the safety range, an alarm signal is issued according to the compensation prediction result to complete the early warning.
[0143] S4.1: Compare the compensation prediction results with the set safety threshold. When the compensation prediction results exceed the safety range, identify potential failure risks and issue a warning signal when necessary.
[0144] The alarm signal includes sound and light alarm devices and remote alarm functions to ensure that relevant personnel can receive alarm information in a timely manner;
[0145] Record detailed information of all alarm events, including alarm time, alarm reason, gas content data and load data, etc., to form an alarm log;
[0146] Regularly analyze the frequency and causes of alarm events based on alarm logs, optimize the early warning model and system parameters, and improve the reliability and accuracy of the early warning system;
[0147] Formulate an emergency response plan. After receiving the alarm signal, the operator will perform corresponding inspection and maintenance measures according to the plan to ensure the safe operation of the transformer;
[0148] The emergency response plan includes:
[0149] The first-level alarm response includes the operator immediately stopping the operation of the transformer after receiving the first-level alarm signal to avoid further damage;
[0150] Quickly check the key parts of the transformer, including oil temperature, oil level and gas content, record all data and upload to the central system;
[0151] The secondary alarm response includes that if the secondary alarm signal is triggered, the operator will conduct local inspection and maintenance according to the specific alarm cause to ensure that it does not affect the operation of the entire system;
[0152] Record all inspection data in detail, especially gas content and load changes, to ensure that the cause of the alarm can be traced;
[0153] The third-level alarm includes the third-level alarm signal. The operator continuously observes the operating status of the transformer through the remote monitoring system and records abnormal conditions in a timely manner;
[0154] Regular maintenance and service of transformers ensures that potential problems can be detected in advance before alarm signals are issued.
[0155] Example 2
[0156] As an embodiment of the present invention, there is provided an online monitoring and early warning system for gas in transformer oil, comprising: a data acquisition module, an early warning modeling module, a data correction module, an alarm processing module, and an emergency response module;
[0157] The data acquisition module is used to transport transformer oil to the oil-gas separation device, detect gas composition data and monitor gas content data and transformer load data in real time, and adjust the oil pump speed;
[0158] The early warning modeling module is used to perform dimensionless processing on historical load data and gas content data, build a multi-layer LSTM model, and integrate the attention mechanism to establish a gas monitoring and early warning model;
[0159] The data correction module is used to analyze the detected gas component content and transformer oil quality data in real time, and to compensate and correct the gas monitoring and early warning results;
[0160] The alarm processing module is used to compare the compensation prediction result with the set safety threshold. When the compensation prediction result exceeds the safety range, an alarm signal is issued to complete the early warning;
[0161] The emergency response module is used to formulate emergency response plans and implement corresponding inspection and maintenance measures according to the emergency response plans to ensure that abnormal situations are handled in a timely manner.
[0162] The specific definition of the online monitoring and early warning system for gas in transformer oil can be found in the above definition of the online monitoring and early warning method for gas in transformer oil, which will not be repeated here. Each module in the above-mentioned online monitoring and early warning system for gas in transformer oil can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory in the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.
[0163] Example 3
[0164] Reference Figure 3 , which is the third embodiment of the present invention. Based on the first two embodiments, the embodiment of the present invention provides a computer device, which may be a server, and its internal structure diagram may be as shown in FIG. Figure 3 As shown. The computer device includes a processor, a memory and a network interface 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, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium.
[0165] The database of the computer device is used to store action detection data. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, the steps in any of the above-mentioned sparse tensor operation acceleration method embodiments are implemented.
[0166] Those skilled in the art will understand that Figure 3 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.
[0167] In one embodiment, an embodiment of the present invention provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps in any of the above-mentioned sparse tensor operation acceleration method embodiments are implemented.
[0168] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory may include read-only memory (ROM), magnetic tape, floppy disk, flash memory or optical memory, etc. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).
[0169] Example 4
[0170] As an embodiment of the present invention, a method for online monitoring and early warning of gas in transformer oil is provided. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.
[0171] In order to verify the effectiveness of the invention, we designed an experiment to simulate the online monitoring and early warning of gas content in transformer oil. The experiment used a 100MVA transformer, which was equipped with a special oil-gas separation device, a gas sensor array, and a data acquisition and processing system.
[0172] First, we put the transformer into operation and continuously monitor the gas components separated from the transformer oil through the oil-gas separation device. The gas sensor array includes seven gas detection sensors: hydrogen, carbon monoxide, carbon dioxide, methane, acetylene, ethylene and ethane. The sensor array collects the concentration data of various gases in real time and transmits the data to the data acquisition system.
[0173] At the same time, we installed transformer load monitoring equipment to record the transformer's operating load data in real time. Based on the real-time detected gas content data and transformer load data, we improved the original data and obtained improved gas content data and improved transformer load data. The specific calculation formula is as follows:
[0174] Improved gas content data:
[0175] Improvements to transformer load data:
[0176] The parameters are set as follows: gas concentration detection time window Tg=10min, gas attenuation coefficient α=0.95, weighting coefficient β=0.7, sampling interval Δt=1s, and number of sampling points N=600.
[0177] Next, based on the improved data, we calculated the optimal oil pump speed and sent the adjustment signal to the oil pump control center to control the real-time operating speed of the oil pump. The formula for calculating the optimal oil pump speed is as follows:
[0178] Among them, the normalization constant Z=1000, the feedback coefficient γ=0.8, the historical adjustment coefficient δn=0.2, and the historical feedback points M=10.
[0179] During the experiment, we calibrated the gas detection sensors every month to ensure the accuracy of the data and recorded the calibration log.
[0180] Table 1 Data record table
[0181]
[0182] It can be seen from the above table that by adopting the invention, various indicators of gas monitoring in transformer oil have been significantly improved. Specifically:
[0183] In terms of gas content deviation, the deviation of the prior art is as high as 8.2%, while the inventive schemes 1-5 control the deviation to 2.1%, 1.6%, 1.2%, 0.9% and 0.7% respectively, which greatly reduces the detection error of the gas content.
[0184] In terms of load fluctuation, the load fluctuation of the prior art is as high as 6.5MW, while the inventive schemes 1-5 control the load fluctuation within 2.8MW, 2.2MW, 1.9MW, 1.5MW and 1.1MW respectively, effectively reducing the impact of load fluctuation.
[0185] In terms of alarm response time, the response time of the prior art is as long as 45 minutes, while the invention schemes 1-5 shorten the response time to 12 minutes, 9 minutes, 7 minutes, 5 minutes and 3 minutes respectively, greatly improving the timeliness of fault warning.
[0186] In terms of false alarm rate, the false alarm rate of the prior art is as high as 15%, while the invention schemes 1-5 reduce the false alarm rate to 5%, 3%, 2%, 1% and 0.5% respectively, effectively reducing the interference caused by false alarms.
[0187] In terms of accuracy, the accuracy of the existing technology is only 78%, while the invention schemes 1-5 increase the accuracy to 91%, 93%, 95%, 97% and 98% respectively, greatly improving the reliability of the early warning model.
[0188] In terms of scalability, the scalability of the existing technology is poor, while the scalability of the invention solutions 2-5 ranges from good to excellent, indicating that the solution can adapt to more complex application scenarios.
[0189] In terms of oil pump efficiency, the oil pump efficiency of the prior art is only 72%, while the invention schemes 1-5 increase the efficiency to 82%, 86%, 89%, 92% and 94% respectively, effectively reducing energy consumption and improving the operating efficiency of the system.
[0190] In summary, by adopting this invention, we have successfully built an efficient and accurate online monitoring and early warning system for gas in transformer oil. The system integrates advanced gas detection technology, intelligent modeling methods and real-time optimization control strategies, significantly improving the accuracy and reliability of gas monitoring, while reducing system operating costs and energy consumption. Compared with the existing technology, this invention has demonstrated excellent performance and innovation in all key indicators, providing a strong guarantee for the safe operation of the transformer.
[0191] 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 online monitoring and early warning of gas in transformer oil, It is characterized in that include: The transformer oil is transported to the oil-gas separation device and the separated gas composition data is detected. At the same time, the oil pump speed is adjusted according to the real-time detected gas content data and transformer load data; The historical load data and gas content data of the transformer are dimensionless, and a multi-layer LSTM model is used to integrate the attention mechanism to establish a gas monitoring and early warning model. The gas monitoring and early warning model is trained and verified to obtain the predicted probability of transformer failure. The detected gas component content and transformer oil quality are analyzed in real time, and the gas monitoring and early warning results are compensated and corrected to obtain the compensation prediction results; When the compensation prediction result exceeds the safety range, an alarm signal is issued according to the compensation prediction result to complete the early warning.
2. The on-line monitoring and early warning method for gas in transformer oil according to claim 1, Features: The gas composition data detected and separated include: Use gas sensors to monitor the gas composition in transformer oil in real time and collect gas concentration data; The gas sensor comprises a plurality of gas detection sensors for detecting different types of gases, and the plurality of gas detection sensors for detecting different types of gases are arranged in an array; The gas detection sensors include a hydrogen detection sensor, a carbon monoxide detection sensor, a carbon dioxide detection sensor, a methane detection sensor, an acetylene detection sensor, an ethylene detection sensor and an ethane detection sensor.
3. The on-line monitoring and early warning method for gas in transformer oil as claimed in claim 2, Features: The adjusting the oil pump speed comprises: According to the real-time detected gas content data and transformer load data, the real-time detected gas content data and transformer load data are calculated to obtain improved gas content data and improved transformer load data; the calculation formula of the improved gas content data is as follows: in, To improve the gas content data, S(τ) is the instantaneous gas concentration detected at time τ, t is the time, T g is the gas concentration detection time window, α is the gas attenuation coefficient; The calculation formula for improving transformer load data is as follows: in, To improve transformer load data, β k is the weighting coefficient, N is the number of sampling points, Δt is the sampling interval, P(τ) is the load power of the transformer at time τ, and k is a discrete point in time; Calculate the optimal oil pump speed according to the improved gas content data and the improved transformer load data, and send the adjustment signal to the oil pump control center; The formula for calculating the optimum oil pump speed is as follows: Among them, Z is the normalization constant, γ is the feedback coefficient, δ n is the historical adjustment coefficient, M is the number of historical feedback points, V(t) is the oil pump speed at time t, and n is the index of historical feedback points; The oil pump control center adjusts the speed of the oil pump according to the received regulation signal and accurately controls the operation of the oil pump through the frequency converter; Regularly calibrate gas detection sensors to ensure the accuracy of detection data, record the results and time of each calibration, and generate a calibration log.
4. The on-line monitoring and early warning method for gas in transformer oil as claimed in claim 3, Features: The establishment of a gas monitoring and early warning model comprises: The historical load data and gas content data of the transformer are collected and dimensionless processed. The dimensionless calculation formula is as follows: Among them, X t is the dimensionless data point, μ x and σ x are the mean and standard deviation of the original data respectively; Use sliding window techniques and smoothing to eliminate noise and outliers in the data; Divide the cleaned data into a training set and a validation set; Train the model on the training set; The model is tuned and verified on the validation set, which includes real-time detection data and is dynamically updated according to the working status of the transformer; Construct a neural network model suitable for time series data, using a multi-layer LSTM model. The calculation formula of the multi-layer LSTM model is as follows: Among them, LSTM l represents the LSTM unit of the lth layer, L is the number of LSTM layers, and are the hidden state and cell state of the lth layer at time t respectively; At the same time, the attention mechanism is introduced to strengthen the model's attention to key moments. The calculation formula for the attention output of the layer LSTM model is as follows: Among them, W q , W k , W v and b q 、b k 、b v are the weight matrix and bias term, A t is the output after the attention weights are applied, Q t is the query vector at time t, K t is the key vector at time t, V t is the value vector at time t, d k is the key vector K t Dimensions; The final warning model output is calculated based on the attention output of the multi-layer LSTM model to obtain the predicted probability of transformer failure. The calculation formula is as follows: in, is the predicted probability of transformer failure, W o and b o are the weight matrix and bias term of the output layer respectively.
5. The on-line monitoring and early warning method for gas in transformer oil as claimed in claim 4, Features: The tuning and verification of the model on the validation set includes: Use time series cross validation and optimize time series data to ensure the generalization ability and stability of the model; Divide the original time series data X into N subsequences X i ; For each subsequence X i , perform sliding window segmentation processing to obtain the segmented time series subset X i,j , where j = 1, 2, ..., M; For each X i,j Use the gas monitoring and early warning model to make predictions and obtain prediction results; Calculate the prediction error based on the prediction results; The cross-validation error is obtained by weighted summation of all prediction errors. The specific calculation formula is as follows: Where CV(X) is the cross validation error, w i,j Corresponding to X i,j The weight, E i,j To calculate the prediction error, is the prediction result of the gas monitoring and early warning model, X i,j is the jth segment of the i-th subsequence; Adjust the time series data, and when the cross validation error CV(X) reaches the minimum value, the optimized time series data is obtained.
6. The on-line monitoring and early warning method for gas in transformer oil as claimed in claim 5, Features: The compensation and correction of the gas monitoring warning results include: Collect real-time monitoring data of gas component content and transformer oil quality, and perform data cleaning and preprocessing; Clean and preprocess the collected gas component content and transformer oil quality data to remove noise and outliers; The cleaned and pre-processed data is input into the compensation and correction model to compensate and correct the gas monitoring warning results. The calculation formula is as follows: in, is the original gas monitoring warning result, x is the real-time analysis compensation data, m is the compensation weight factor, b is the compensation bias term, To compensate for the prediction results.
7. The on-line monitoring and early warning method for gas in transformer oil according to claim 6, Features: The issuing of an alarm signal according to the compensation prediction result comprises: Compare the compensation prediction results with the set safety threshold. When the compensation prediction results exceed the safety range, identify potential failure risks and issue early warning signals. The alarm signal includes sound and light alarm devices and remote alarm functions to ensure that relevant personnel can receive alarm information in a timely manner; Record detailed information of all alarm events, including alarm time, alarm reason, gas content data and load data, and form an alarm log with detailed information of all alarm events; Regularly analyze the frequency and causes of alarm events based on alarm logs, and optimize early warning models and system parameters; Develop an emergency response plan. When operators receive an alarm signal, they will perform corresponding inspection and maintenance measures according to the emergency response plan; The emergency response plan includes, Level 1 alarm response includes when the operator receives a level 1 alarm signal, immediately stopping the operation of the transformer to avoid further damage, quickly checking the key parts of the transformer, including oil temperature, oil level and gas content, recording all data and uploading it to the central control center; Level 2 alarm response includes when the level 2 alarm signal is triggered, the operator will conduct local inspection and maintenance according to the specific alarm cause to ensure that it does not affect the operation of the entire system, record all inspection data in detail, including gas content and load changes, and track the cause of the alarm; The third-level alarm includes that when the third-level alarm signal is triggered, the operator continuously observes the operating status of the transformer through the remote monitoring system, records abnormal conditions, and regularly performs maintenance and service on the transformer.
8. A system using the online monitoring and early warning method for gas in transformer oil as claimed in any one of claims 1 to 7, It is characterized in that include: Data acquisition module, early warning modeling module, data correction module, alarm processing module, emergency response module; The data acquisition module is used to transport transformer oil to the oil-gas separation device, detect gas composition data and monitor gas content data and transformer load data in real time, and adjust the oil pump speed; The early warning modeling module is used to perform dimensionless processing on historical load data and gas content data, build a multi-layer LSTM model, and integrate the attention mechanism to establish a gas monitoring early warning model; The data correction module is used to analyze the detected gas component content and transformer oil quality data in real time, and to compensate and correct the gas monitoring and early warning results; The alarm processing module is used to compare the compensation prediction result with the set safety threshold, and when the compensation prediction result exceeds the safety range, an alarm signal is issued to complete the early warning; The emergency response module is used to formulate an emergency response plan and execute corresponding inspection and maintenance measures according to the emergency response plan to ensure timely handling of abnormal situations.
9. A computer device comprising a memory and a processor, wherein the memory stores a computer program. It is characterized in that When the processor executes the computer program, the steps of the online monitoring and early warning method for gas in transformer oil according to any one of claims 1 to 7 are implemented.
10. A computer-readable storage medium having a computer program stored thereon, It is characterized in that When the computer program is executed by a processor, the steps of the on-line monitoring and early warning method for gas in transformer oil according to any one of claims 1 to 7 are implemented.
Citation Information
Cited By
Oil leakage diagnosis and gas concentration compensation method and system for oil and gas monitoring equipment
CN119880298A
An oil leakage diagnosis and gas concentration compensation method and system for an oil and gas monitoring device
CN119880298B
Transformer oil quality comprehensive monitoring method and system and storage medium
CN120385812A
Transformer oil level on-line monitoring and early warning method and system and medium
CN121163627A
Multi-source infrared telemetry gas identification optimization method and system
CN121256187A
