Distribution transformer state monitoring method and system based on power distribution room economic operation control device

Abstract

The invention relates to a distribution transformer state monitoring method and system based on a power distribution room economic operation control device. By collecting transformer state monitoring variables and combining a subjective and objective weighting method and an LSTM neural network model, the running health state of the distribution transformer is monitored. Through real-time monitoring and intelligent diagnosis, potential problems can be found in advance, the power failure time caused by faults is shortened, the power supply reliability is improved, the unnecessary overhaul frequency is effectively reduced, and the maintenance cost is reduced; power-off maintenance can be arranged more reasonably, and excessive loss or premature maintenance of equipment is avoided; sudden faults of equipment can be effectively prevented, and long-term safe operation of the transformer is ensured; potential problems can be found and processed in time, early damage of equipment caused by hidden dangers is reduced, and therefore the service life of the equipment is prolonged. The active and reactive power loss in equipment operation can be reduced, and the energy utilization efficiency of the whole power grid system is improved.

Description

Technical Field

[0001] The invention relates to a distribution transformer state monitoring method and system based on an economic operation control device of a distribution room. Background Art

[0002] Condition maintenance of power transmission and transformation equipment is an important measure to ensure the safe operation of the power grid. Through online monitoring and live detection technology, the status information of the equipment can be obtained in real time or regularly. These data, combined with historical records and preventive tests, can accurately identify equipment anomalies, conduct a comprehensive assessment of its operating status, and carry out targeted maintenance accordingly.

[0003] Transformers are one of the most critical and expensive equipment in the power grid. Their failure may bring huge economic losses and power outage risks. Therefore, how to effectively monitor the operating status of transformers is not only the focus of the power sector, but also a hot research topic in academia. Due to the complex structure and numerous components of transformers, different fault forms require the development of multiple online monitoring methods.

[0004] By applying these monitoring technologies, the power sector can better understand the status of transformers, prevent serious faults, and arrange maintenance work reasonably to ensure the safe operation of the power grid and efficient management of equipment.

[0005] Accurately assessing transformer status and rationally applying online monitoring technology is a major challenge facing the power sector and is also a current research hotspot. By summarizing the research results of transformer status assessment and the practical application of online monitoring, the following main problems are found:

[0006] 1. Although a variety of transformer condition assessment models have been developed, there are still few models that can be applied in practice due to various limitations. In addition, most of the existing fault diagnosis methods are based on offline dissolved gas data, ignoring the time variation trend of the data and failing to make full use of online monitoring data.

[0007] 2. The online monitoring device itself is also a part of the power equipment. When carrying out the intelligent transformation of power transmission and transformation equipment, its configuration and planning need to be reasonably designed. At the same time, after the monitoring device is put into use, how to arrange its operation and maintenance is also an urgent problem to be solved by the power department.

[0008] Therefore, accurately evaluating the state of the transformer and optimizing the configuration and operation of the online monitoring device accordingly are of great significance for ensuring the long-term stable operation of the transformer. Reasonable monitoring applications can not only ensure the safe operation of the power grid, but also bring greater benefits. The present invention aims to propose effective solutions by exploring transformer state assessment technology, fault diagnosis methods, and planning and operation adjustments of online monitoring devices. This topic not only has important academic value, but is also closely related to the actual problems of power production and has extremely high engineering application value. Summary of the invention

[0009] The purpose of the present invention is to overcome the problems existing in the prior art and to provide a distribution transformer status monitoring method and system based on an economic operation control device of a distribution room.

[0010] To achieve the above-mentioned purpose, the technical solution of the present invention is: a distribution transformer status monitoring method based on the economic operation control device of the distribution room, which realizes the distribution transformer operation health status monitoring by collecting transformer status monitoring variables and combining subjective and objective weighting method and LSTM neural network model.

[0011] In one embodiment of the present invention, the transformer state monitoring variables include transformer oil dissolved gas variable x1, electrical characteristic variable x2, and insulating oil characteristic variable x3; wherein,

[0012] Transformer oil dissolved gas variable x1 includes CH4 content x 11 、C2H4 content x 12 、CO content x 13 、CO2 content x 14 、H2 content x 15 、C2H6 content x 16 ;

[0013] The electrical characteristic variable x2 includes the core current x 21 , Winding DC resistance unbalance rate x 22 , winding absorption ratio x 23 、Winding dielectric loss x 24 ;

[0014] The insulating oil characteristic variable x3 includes the trace water content x in the oil 31 、Oil dielectric loss x 32 .

[0015] In one embodiment of the present invention, the indicators in transformer oil dissolved gas variable x1, electrical characteristic variable x2, and insulating oil characteristic variable x3 are weighted by subjective and objective weighting method, that is, subjective weighting is combined with objective weighting.

[0016] In one embodiment of the present invention, the subjective weighting of the transformer oil dissolved gas variable x1 is specifically implemented as follows:

[0017] Construct a judgment matrix of different dissolved gases in transformer oil dissolved gas variable x1, and normalize each column of the judgment matrix:

[0018]

[0019] Sum each row of the normalized matrix:

[0020]

[0021] Normalize the rows and vector v to get the weight vector w:

[0022]

[0023] Among them, a ij represents the i-th sample value of the j-th indicator, b ij Indicates a ij The normalized value, v i Indicates that the i-th row b ij The sum of the values ​​of Represented as a vector v i The weight value obtained after normalization.

[0024] In one embodiment of the present invention, objective weighting of transformer oil dissolved gas variable x1 is performed using entropy increase method, which is specifically implemented as follows:

[0025] Normalize the data to eliminate the dimension effect

[0026]

[0027] Among them, x ij represents the i-th sample value of the j-th indicator, X ij represents the normalized value, x jmax represents the maximum value of the jth index, x jmin represents the minimum value of the jth index;

[0028] Calculating Probability

[0029]

[0030] Where m is the number of samples, X ij is the normalized i-th sample value of the j-th indicator;

[0031] Calculating entropy

[0032]

[0033] When p ij = 0, define pij lnp ij =0;

[0034] Calculate the coefficient of variation

[0035] d i =1-e i

[0036] Calculating objective weights

[0037]

[0038] The above formula is used to solve the sum of the difference coefficients of the six gases.

[0039] In one embodiment of the present invention, the subjective and objective comprehensive weighting calculation formula for weighting different dissolved gas components of transformer oil dissolved gas variable x1 is as follows:

[0040]

[0041] in is the subjective weight, is the objective assignment weight, w i is the final weight result.

[0042] In one embodiment of the present invention, an LSTM neural network model is used to implement the health status monitoring of the distribution transformer, which is specifically implemented as follows:

[0043] Based on m historical samples, each sample includes three characteristic variables, namely transformer oil dissolved gas variable x1, electrical characteristic variable x2, and insulating oil characteristic variable x3;

[0044] The input matrix of the LSTM neural network model is

[0045]

[0046] Among them, x ij represents the i-th sample value of the j-th indicator;

[0047] The target output Y is the status of the transformer, ranging from 0 to 10, indicating the health of the transformer;

[0048] The output vector matrix is

[0049]

[0050] Among them, y i Represents the transformer state quantity output after the i-th historical sample is predicted by LSTM, i∈[1,m].

[0051] In one embodiment of the present invention, the specific structure of the LSTM neural network model is as follows:

[0052] (1) Input layer

[0053] Each sample has 3 features, and the output of the input layer is the input feature vector x i :

[0054]

[0055] (2) Hidden Layer

[0056] The hidden layer performs a linear transformation and then applies an activation function; assuming that the kth layer has h neurons, the calculation formula for the kth layer is

[0057]

[0058] Where W k is the weight matrix, b k is the bias vector, is the linear output of the previous layer, is the k-1th input vector;

[0059] Then apply the activation function ReLU to get the activated output

[0060]

[0061] in is the output after activation;

[0062] (3) Output layer

[0063] The output layer is a single node, and the output is:

[0064]

[0065] Among them, W L is the weight matrix of the output layer, b L is the bias vector of the output layer, is the predicted value for the i-th sample;

[0066] (4) Loss Function

[0067] The mean square error MSE is used as the loss function. For N samples, the loss function is defined as:

[0068]

[0069] according to The output results determine the transformer operating health status.

[0070] In one embodiment of the present invention, based on the transformer operating health status, the following operations are performed:

[0071] 1) When In normal state, the state quantity is stable and within the warning value and caution value specified in the regulations;

[0072] 2) When When the state is in progress, pay attention to the state quantity change trend towards the standard limit value, and strengthen operation monitoring;

[0073] 3) When When the power is turned off, it is abnormal, the state quantity changes greatly, and is close to or slightly exceeds the standard limit, so power outage and maintenance should be arranged in time;

[0074] 4) When When it is in a serious state, the state quantity seriously exceeds the standard limit and a power outage for maintenance needs to be arranged as soon as possible.

[0075] The present invention also provides a distribution transformer state monitoring system based on the economic operation control device of the distribution room, which executes the method steps as described above, including:

[0076] Acquisition module, collecting transformer status monitoring variables;

[0077] The subjective and objective weighting module is used to perform subjective and objective weighting on transformer state monitoring variables;

[0078] The neural network prediction module realizes the operation health status monitoring of distribution transformers through the LSTM neural network model.

[0079] Compared with the prior art, the present invention has the following beneficial effects:

[0080] 1. Improve power supply reliability: Through real-time monitoring and intelligent diagnosis, potential problems can be discovered in advance, the power outage time caused by faults can be reduced, and the power supply reliability can be improved.

[0081] 2. Reduce maintenance costs: Rationally arrange the configuration and operation of online monitoring devices to effectively reduce unnecessary maintenance times and reduce maintenance costs.

[0082] 3. Optimize maintenance arrangements: Based on accurate transformer status assessment results, power outage maintenance can be arranged more reasonably to avoid excessive equipment wear or premature maintenance.

[0083] 4. Enhance equipment safety: Through refined status monitoring and diagnosis, sudden equipment failures can be effectively prevented and the long-term safe operation of the transformer can be ensured.

[0084] 5. Extend the service life of equipment: Accurately diagnose the equipment status, promptly discover and deal with potential problems, reduce early damage to equipment caused by hidden dangers, and thus extend the service life of equipment.

[0085] 6. Improve energy efficiency: By optimizing the operation mode of the transformer, reduce the active and reactive power losses in equipment operation, and improve the energy efficiency of the entire power grid system. BRIEF DESCRIPTION OF THE DRAWINGS

[0086] Figure 1 It is the transformer status monitoring variable of the present invention.

[0087] Figure 2 This is a schematic diagram of the prediction of the LSTM neural network model of the present invention.

[0088] Figure 3 The figure is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0089] The technical solution of the present invention is described in detail below in conjunction with the accompanying drawings.

[0090] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present application. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present application belongs.

[0091] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present application. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprise" and / or "include" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0092] The present invention provides a distribution transformer state monitoring method based on the economic operation control device of the distribution room. The method collects transformer state monitoring variables and combines the subjective and objective weighting method and the LSTM neural network model to realize the distribution transformer operation health state monitoring.

[0093] The present invention also provides a distribution transformer state monitoring system based on the economic operation control device of the distribution room, which executes the method steps as described above, including:

[0094] Acquisition module, collecting transformer status monitoring variables;

[0095] The subjective and objective weighting module is used to perform subjective and objective weighting on transformer state monitoring variables;

[0096] The neural network prediction module realizes the operation health status monitoring of distribution transformers through the LSTM neural network model.

[0097] The following is the specific implementation process of the present invention.

[0098] like Figure 1 As shown, the present invention selects the following indicators as transformer status monitoring variables.

[0099] 1. Dissolved gas in oil

[0100] The transformer uses an oil-paper insulation structure inside. When the transformer fails or ages, different gases will be produced, which will dissolve in the insulating oil. The type and content of the gas can reflect the type and severity of the fault. For example, during normal operation, the oil and insulating paper will produce a small amount of gas; but when overheating or discharge faults occur, the amount of gas will increase sharply. By detecting the composition and content of dissolved gas in the oil, the degree of aging of the transformer's internal insulation can be evaluated, potential faults can be identified and their severity can be judged. This is an important basis for judging the health status of the transformer.

[0101] 2. Core current

[0102] A common fault of the iron core is multi-point grounding. When a multi-point grounding fault occurs, a circulation will be formed, causing local overheating, decomposing the insulating oil, and even burning the iron core in severe cases. Although this fault can be detected by detecting dissolved gas in the oil, if the overheating caused by the circulation is relatively mild, it is difficult to detect through gas detection. By monitoring the core grounding current, the insulation condition of the core can be effectively judged, especially the current flowing through the grounding wire. If it exceeds the specified value (usually 100mA), there may be a fault.

[0103] 3. Winding DC resistance unbalance rate

[0104] The DC resistance of the winding can reflect the longitudinal insulation state of the transformer and the connection status of the current loop. This indicator can help determine whether there is a short circuit or broken strands in the winding, and can also detect the contact status of the tap changer. According to the standard, the unbalance rate of the resistance of each phase winding at the same temperature should be within 2%. In addition, the difference in the initial value of each phase resistance should not exceed ±2%. By measuring these resistance values, the health status of the transformer winding can be effectively evaluated.

[0105] 4. Winding absorption ratio

[0106] The winding absorption ratio is a key indicator for evaluating the main insulation condition of the transformer. It can reflect problems such as insulation paper being damp or insulation oil being aged. Compared with relying solely on insulation resistance, the absorption ratio or polarization index can more effectively determine the insulation condition of the winding. In practical applications, the absorption ratio is usually used to determine whether the winding is damp or aged. According to the standard, the absorption ratio shall not be less than 1.3.

[0107] 5. Winding dielectric loss

[0108] The winding dielectric loss factor (abbreviated as dielectric loss) reflects the insulation performance of the transformer. Combined with the dielectric loss values ​​of insulating oil and insulating paper, the insulation performance can be evaluated more comprehensively. When the insulation is damp, the oil quality deteriorates, a discharge fault or a local defect occurs, the dielectric loss value will change. In order to eliminate the influence of temperature on the measurement results, the measurement results are usually converted to the 20°C standard. The standard stipulates that the dielectric loss of transformer windings at voltage levels of 330kV and above should not be greater than 0.8%; 220kV and below voltage levels should not be greater than 0.5%.

[0109] 6. Micro-water in oil

[0110] Due to the production, transportation, storage and operation process, transformer oil will inevitably contain a small amount of water. During normal operation, the water content in the oil should be between 10-40ppm. When the water content exceeds this range, it means that the quality of the insulating oil has deteriorated, and problems such as moisture and water ingress may occur. The presence of water in the oil will increase dielectric loss and acid value, reduce breakdown voltage, and affect insulation performance. For transformers with voltage levels of 220kV and below, the water content in the oil should not exceed 25mg / L; for transformers with voltage levels of 330kV and above, the water content should not exceed 15mg / L.

[0111] 7. Oil loss

[0112] The dielectric loss factor of transformer oil is an important parameter for evaluating the insulation performance of oil. During operation, moisture, impurities, metal ions and microbial contamination in the oil will cause the oil dielectric loss to increase, which in turn reflects the deterioration of the oil quality. By monitoring the changes in oil dielectric loss, the operating conditions of other parts of the transformer can be indirectly evaluated. The standard stipulates that the oil dielectric loss value of transformers with voltage levels of 330kV and below should not be greater than 0.04; the oil dielectric loss value of transformers with voltage levels of 500kV and above should not be greater than 0.02.

[0113] These key indicators can fully reflect the operating status and potential fault conditions of the transformer, thus providing data support for equipment maintenance and inspection.

[0114] Assume that the gas dissolved in transformer oil is variable x1, the electrical characteristics are variable x2, and the insulating oil characteristics are variable x3.

[0115] (1) The transformer oil dissolved gas variable x1 includes the CH4 content, which is denoted as x 11 , the C2H4 content is recorded as x 12 , CO content is recorded as x 13 , CO2 content is recorded as x 14 , H2 content is recorded as x 15 , C2H6 content is recorded as x 16 .

[0116] (2) The electrical characteristic variable x2 includes the core current, denoted by x 21 , the unbalance rate of winding DC resistance is recorded as x 22 , the winding absorption ratio is recorded as x 23 , the winding dielectric loss is recorded as x 24 .

[0117] (3) The insulating oil characteristic variable x3 includes the trace water in the oil, which is denoted as x 31 , oil dielectric loss is recorded as x 32 .

[0118] 1. Variable Weight Assignment

[0119] 1. Subjective empowerment

[0120] 1.1 Construction of judgment matrix (taking the dissolved gas content of transformer oil as an example, see Table 1)

[0121] Table 1

[0122] <![CDATA[CH4]]> <![CDATA[C2H4]]> CO <![CDATA[CO2]]> <![CDATA[H2]]> <![CDATA[C2H6]]> <![CDATA[CH4]]> 1 1 / 2 3 2 5 1 <![CDATA[C2H4]]> 2 1 4 3 6 2 CO 1 / 3 1 / 4 1 1 / 2 3 1 / 2 <![CDATA[CO2]]> 1 / 2 1 / 3 2 1 4 1 <![CDATA[H2]]> 1 / 5 1 / 6 1 / 3 1 / 4 1 1 / 4 <![CDATA[C2H6]]> 1 1 / 2 2 1 4 1

[0123] The weight of each gas is calculated by normalizing the eigenvector and checking the consistency (CI / CR). If the consistency ratio (CR) is less than 0.1, the judgment matrix passes the consistency check and the weight is available.

[0124] 1.2 Column Normalization: First, each column of the judgment matrix is ​​normalized.

[0125]

[0126] 1.3 Row summation: Then, each row of the normalized matrix is ​​summed.

[0127]

[0128] 1.4 Normalize to get the weight vector: Finally, normalize the row and vector v to get the weight vector w.

[0129]

[0130] 2. Objective empowerment

[0131] The objective weighting method is carried out using the entropy increase method.

[0132] 2.1 Normalize the data first to eliminate the dimension effect

[0133]

[0134] Among them, x ij represents the i-th sample value of the j-th indicator, X ij Indicates the normalized value.

[0135] 2.2 Calculating Probability

[0136]

[0137] Where m is the number of samples, X ij is the normalized i-th sample value of the j-th indicator;

[0138] 2.3 Calculating Entropy

[0139]

[0140] When p ij = 0, define p ij lnp ij =0.

[0141] 2.4 Calculation of coefficient of variation

[0142] d i =1-e i

[0143] 2.5 Calculation of objective weights

[0144]

[0145] 3. Comprehensive empowerment of subjective and objective factors

[0146] Multiplicative synthesis normalization

[0147]

[0148] in is the subjective weight, is the objective assignment weight, w i is the final weight result.

[0149] 2. Using LSTM to diagnose and predict the health status of transformers (see Figure 2 )

[0150] Based on m historical samples, each sample includes three characteristic variables, namely transformer oil dissolved gas x1, electrical characteristics x2, and insulating oil characteristics x3.

[0151] Therefore, the input matrix of the neural network is

[0152]

[0153] Among them, x ij represents the i-th sample value of the j-th indicator

[0154] The target output Y is the status of the transformer, ranging from 0 to 10 to indicate the health of the transformer.

[0155] The output vector matrix is

[0156]

[0157] Input structure of a neural network

[0158] 1. Input layer

[0159] Assuming that each sample has m features (here m = 3), the output of the input layer is the input feature vector x i :

[0160]

[0161] 2. Hidden Layer

[0162] The hidden layer performs a linear transformation and then applies an activation function. Assuming that the kth layer has h neurons, the calculation formula for this layer is

[0163]

[0164] where w k is the weight matrix, b k is the bias vector, is the linear output of the previous layer.

[0165] Then apply an activation function (such as ReLU) to get the activated output

[0166]

[0167] in is the output after activation.

[0168] 3. Output layer

[0169] The output layer is a single node, and the output is:

[0170]

[0171] Among them, W L is the weight matrix of the output layer, b L is the bias vector of the output layer, is the predicted value for the i-th sample.

[0172] 4. Loss Function

[0173] The method of the present invention adopts mean square error (MSE) as the loss function. For N samples, the loss function is defined as:

[0174]

[0175] according to The output results determine the transformer operating health status.

[0176] 1) When In the normal state, the state quantity is stable and within the warning value and attention value specified in the regulations.

[0177] 2) When When monitoring the status, pay attention to the changing trend of the status quantity towards the standard limit and strengthen operation monitoring.

[0178] 3) When When the power is in abnormal state, the state quantity changes greatly and is close to or slightly exceeds the standard limit, arrange power outage and maintenance in time.

[0179] 4) When When it is in a serious state, the state quantity seriously exceeds the standard limit and a power outage for maintenance needs to be arranged as soon as possible.

[0180] Figure 3 The present invention is an overall implementation flow chart of the method.

[0181] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0182] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0183] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1A function specified in one or more boxes.

[0184] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0185] The above is only a preferred embodiment of the present invention, and does not limit the present invention in other forms. Any technician familiar with the profession may use the above disclosed technical content to change or modify it into an equivalent embodiment with equivalent changes. However, any simple modification, equivalent change and modification made to the above embodiment according to the technical essence of the present invention without departing from the technical solution of the present invention still belongs to the protection scope of the technical solution of the present invention.

Claims

1. A distribution transformer status monitoring method based on a distribution room economic operation control device, characterized in that: By collecting transformer status monitoring variables and combining subjective and objective weighting methods and LSTM neural network models, the operating health status of distribution transformers can be monitored.

2. A distribution transformer status monitoring method based on the economic operation control device of the distribution room according to claim 1, characterized in that: The transformer status monitoring variables include transformer oil dissolved gas variable x1, electrical characteristic variable x2, and insulating oil characteristic variable x3; wherein, Transformer oil dissolved gas variable x1 includes CH4 content x 11 、C2H4 content x 12 、CO content x 13 、CO2 content x 14 、H2 content x 15 、C2H6 content x 16 ; The electrical characteristic variable x2 includes the core current x 21 , Winding DC resistance unbalance rate x 22 , winding absorption ratio x 23 、Winding dielectric loss x 24 ; The insulating oil characteristic variable x3 includes the trace water content x in the oil 31 、Oil dielectric loss x 32 .

3. A distribution transformer status monitoring method based on the economic operation control device of the distribution room according to claim 2, characterized in that: Through the subjective and objective weighting method, that is, combining subjective weighting with objective weighting, weighted processing is performed on each index in transformer oil dissolved gas variable x1, electrical characteristic variable x2, and insulating oil characteristic variable x3.

4. A distribution transformer status monitoring method and system based on the economic operation control device of the distribution room according to claim 3, characterized in that: The specific implementation of subjective weighting for transformer oil dissolved gas variable x1 is as follows: Construct a judgment matrix of different dissolved gases in transformer oil dissolved gas variable x1, and normalize each column of the judgment matrix: Sum each row of the normalized matrix: Normalize the rows and vector v to get the weight vector w: Among them, a ij represents the i-th sample value of the j-th indicator, b ij Indicates a ij The normalized value, v i Indicates that the i-th row b ij The sum of the values ​​of Represented as a vector v i The weight value obtained after normalization.

5. A distribution transformer status monitoring method based on the economic operation control device of the distribution room according to claim 4, characterized in that: The objective weighting of transformer oil dissolved gas variable x1 is performed using the entropy increase method, which is specifically implemented as follows: Normalize the data to eliminate the dimension effect Among them, x ij represents the i-th sample value of the j-th indicator, X ij represents the normalized value, x jmax represents the maximum value of the jth index, x jmin represents the minimum value of the jth index; Calculating Probability Where m is the number of samples, X ij is the normalized i-th sample value of the j-th indicator; Calculating entropy When p ij = 0, define p ij lnp ij =0; Calculate the coefficient of variation d i =1-e i Calculating objective weights The above formula is used to solve the sum of the difference coefficients of the six gases.

6. A distribution transformer status monitoring method based on the economic operation control device of the distribution room according to claim 5, characterized in that: Through the subjective and objective weighting method, the subjective and objective comprehensive weighting calculation formula for weighting the different dissolved gas components of transformer oil dissolved gas variable x1 is as follows: in is the subjective weight, is the objective assignment weight, w i is the final weight result.

7. A distribution transformer status monitoring method based on the economic operation control device of a distribution room according to claim 1, characterized in that: The LSTM neural network model is used to monitor the health status of distribution transformers. The specific implementation is as follows: Based on m historical samples, each sample includes three characteristic variables, namely transformer oil dissolved gas variable x1, electrical characteristic variable x2, and insulating oil characteristic variable x3; The input matrix of the LSTM neural network model is Among them, x ij represents the i-th sample value of the j-th indicator; The target output Y is the status of the transformer, ranging from 0 to 10, indicating the health of the transformer; The output vector matrix is Among them, y i Represents the transformer state quantity output after the i-th historical sample is predicted by LSTM, i∈[1,m].

8. A distribution transformer status monitoring method based on the economic operation control device of a distribution room according to claim 7, characterized in that: The specific structure of the LSTM neural network model is as follows: (1) Input layer Each sample has 3 features, and the output of the input layer is the input feature vector x i : (2) Hidden Layer The hidden layer performs a linear transformation and then applies an activation function; assuming that the kth layer has h neurons, the calculation formula for the kth layer is Where W k is the weight matrix, b k is the bias vector, is the linear output of the previous layer, is the k-1th input vector; Then apply the activation function ReLU to get the activated output in is the output after activation; (3) Output layer The output layer is a single node, and the output is: Among them, W L is the weight matrix of the output layer, b L is the bias vector of the output layer, is the predicted value for the i-th sample; (4) Loss Function The mean square error MSE is used as the loss function. For N samples, the loss function is defined as: according to The output results determine the transformer operating health status.

9. A distribution transformer status monitoring method based on the economic operation control device of a distribution room according to claim 8, characterized in that: Based on the transformer health status, perform the following operations: 1) When In normal state, the state quantity is stable and within the warning value and caution value specified in the regulations; 2) When When the state is in progress, pay attention to the state quantity change trend towards the standard limit value, and strengthen operation monitoring; 3) When When the power is turned off, it is abnormal, the state quantity changes greatly, and is close to or slightly exceeds the standard limit, so power outage and maintenance should be arranged in time; 4) When When it is in a serious state, the state quantity seriously exceeds the standard limit and a power outage for maintenance needs to be arranged as soon as possible.

10. A distribution transformer status monitoring system based on a distribution room economic operation control device, characterized in that: Executing any of the method steps as claimed in claims 1 to 9, comprising: Acquisition module, collecting transformer status monitoring variables; The subjective and objective weighting module is used to perform subjective and objective weighting on transformer state monitoring variables; The neural network prediction module realizes the operation health status monitoring of distribution transformers through the LSTM neural network model.

Images