A transformer condition assessment method, system, medium, equipment, and product.

Abstract

This invention discloses a method, system, medium, device, and product for transformer condition assessment, aiming to solve the problem of inaccurate assessment caused by the complexity of condition variables and the large amount of evidence in existing technologies. The method includes: preprocessing acquired multi-parameter optical sensor data of the transformer to obtain preprocessed data; extracting features from the preprocessed data using a multilayer perceptron (MLP) to obtain a multi-source data feature set; using each feature in the multi-source data feature set as independent evidence, performing multi-evidence fusion based on the dynamic threshold D-S evidence theory to obtain a transformer condition assessment probability; and assessing the transformer condition based on the transformer condition assessment probability. This invention combines a multilayer perceptron (MLP) and the dynamic threshold D-S evidence theory, enabling accurate transformer condition assessment and applicable to various data types acquired by multi-parameter optical sensors inside the transformer.

Description

Technical Field

[0001] This invention relates to the field of transformer condition assessment technology, and in particular to a transformer condition assessment method, system, medium, equipment and product. Background Technology

[0002] Multi-parameter optical sensors (temperature, partial discharge, pressure, etc.) inside transformers can collect multi-dimensional status information in real time. However, the effective fusion and status assessment of multi-source heterogeneous data still faces the following challenges:

[0003] Data heterogeneity: Data from different sensors vary greatly in dimensions, sampling frequency, and noise characteristics, making it difficult for traditional fusion methods to handle uniformly. Uncertainty: Sensor data contains noise and uncertainty, making it difficult for a single evaluation method to accurately reflect the transformer's condition. Multi-evidence fusion: Existing methods lack a systematic fusion and evaluation mechanism for multi-source evidence.

[0004] Existing transformer condition assessment methods are mainly based on single methods and deep learning methods. Single methods based on thresholds or statistical models are difficult to handle complex nonlinear relationships. Existing deep learning methods (such as LSTM and CNN) have limitations in multi-source data fusion and uncertainty handling. There has been no research combining multilayer perceptron (MLP) with DS evidence theory for transformer multi-parameter optical sensor data condition assessment. Summary of the Invention

[0005] The purpose of this invention is to overcome the challenges of data heterogeneity, uncertainty, and multi-evidence fusion in the prior art, and to provide a transformer condition assessment method, system, medium, device, and product. It extracts features of multi-parameter data through MLP, improves the DS evidence theory of dynamic threshold to achieve multi-evidence fusion and condition assessment, and through the dynamic threshold calculation module, receives environmental sensor data and historical assessment results in real time, outputs the adjusted threshold, and solves the misjudgment problem of traditional fixed threshold when transformer load fluctuates and seasonal changes, so as to accurately assess the transformer condition.

[0006] To solve the above-mentioned technical problems, the present invention is implemented using the following technical solution:

[0007] In a first aspect, the present invention provides a transformer condition assessment method, comprising:

[0008] The acquired multi-parameter optical sensing data of the transformer is preprocessed to obtain preprocessed data;

[0009] The preprocessed data is used to extract features using a multilayer perceptron to obtain a multi-source data feature set;

[0010] Each feature in the multi-source data feature set is used as an independent piece of evidence. Multi-evidence fusion is performed using the dynamic threshold DS evidence theory to obtain the transformer state assessment probability.

[0011] The transformer condition is assessed based on the transformer condition assessment probability.

[0012] Optionally, the data preprocessing of the acquired transformer multi-parameter optical sensing data includes:

[0013] Wavelet transform or adaptive filtering is used to denoise the acquired transformer multi-parameter optical sensing data to obtain denoised data;

[0014] The denoised data is normalized to obtain normalized data;

[0015] The normalized data is interpolated and aligned to ensure temporal consistency, resulting in preprocessed data.

[0016] The normalization of the denoised data is achieved using the following formula:

[0017] ,

[0018] in, This represents the normalized data. , This represents the data after denoising. This represents the minimum value of the data after denoising. This represents the maximum value of the data after noise reduction.

[0019] Optionally, the feature extraction of the preprocessed data using a multilayer perceptron is achieved through the following formula:

[0020] ;

[0021] ;

[0022] ,

[0023] in, Indicates the first hidden layer. The output of each hidden unit Indicates the first hidden layer. The weights of each hidden unit, Indicates the number of data after preprocessing. One parameter, This indicates the total number of parameters in the preprocessed data. Indicates the first hidden layer. The bias of each hidden unit. This represents the activation function. Indicates the Hidden layer The output of each hidden unit Indicates the Hidden layer The output of each hidden unit Indicates the The total number of hidden units in the hidden layer. Indicates the Hidden layer The weights of each hidden unit, Indicates the Hidden layer The bias of each hidden unit. Indicates the total number of hidden layers. Indicates the The total number of hidden units in the hidden layer. Indicates the Hidden layer The weights of each hidden unit, Indicates the Hidden layer The output of each hidden unit Indicates the The bias of the hidden layer, This represents a set of features from multiple data sources.

[0024] Optionally, the step of treating each feature in the multi-source data feature set as independent evidence and using DS evidence theory to perform multi-evidence fusion includes:

[0025] Based on the multi-source data feature set, a transformer condition assessment framework is established; the transformer condition assessment framework The expression is as follows:

[0026] ,

[0027] in, As a proposition, This indicates that the transformer is normal. This indicates that the transformer is overheating at a low temperature. This indicates that the transformer is overheating. This indicates that the transformer is overheating. This indicates partial discharge in the transformer. This indicates low-energy discharge in the transformer. This indicates high-energy discharge from the transformer. Indicates uncertainty;

[0028] The power set of the transformer state assessment framework is used to construct a set of propositions. The basic probability assignment function is obtained. The expression is:

[0029] ;

[0030] in, To represent the empty set, The basic probability assignment value of the empty set is 0. Expressing a proposition The basic probability allocation value, , and Expressing a proposition The hypothetical subset of the two pieces of evidence, express The basic probability allocation value, express The basic probability allocation value, express and The intersection is , express and The intersection is empty. Indicates conflict factor, Represents the set of propositions The sum of the basic probability assignments of all propositions in the equation is 1.

[0031] According to the basic probability allocation function Each feature in the multi-source data feature set is treated as an independent piece of evidence, and multi-evidence fusion is performed to obtain the transformer state assessment probability. ,in, Indicates the transformer status. .

[0032] Optionally, assessing the transformer state based on the transformer state assessment probability includes:

[0033] When the transformer state assessment probability satisfies the following rule, the output is... As a result of transformer condition assessment:

[0034] ,

[0035] in, This represents the function that takes the maximum value. express The proposition with the largest basic probability distribution value. express The second largest proposition in the basic probability assignment is... Indicates a dynamic threshold. The basic probability assignment value representing uncertainty;

[0036] The dynamic threshold It can be obtained through the following formula:

[0037] ;

[0038] in, This represents the standard deviation of the recent state probability. Indicates the baseline threshold. C This represents the environmental correction factor. The weights represent the standard deviations of recent state probabilities. The weight representing the baseline threshold, This indicates the weight of the environmental correction factor.

[0039] Optionally, when the evaluation result of the transformer status is low-temperature overheating, medium-temperature overheating, high-temperature overheating, partial discharge, low-energy discharge, or high-energy discharge, a corresponding early warning signal is triggered.

[0040] In a second aspect, the present invention provides a transformer condition assessment system, comprising:

[0041] The preprocessing module is used to: perform data preprocessing on the acquired transformer multi-parameter optical sensing data to obtain preprocessed data;

[0042] The feature extraction module is used to: extract features from the preprocessed data using a multilayer perceptron to obtain a multi-source data feature set;

[0043] The multi-evidence fusion module is used to: treat each feature in the multi-source data feature set as an independent piece of evidence, and perform multi-evidence fusion using the dynamic threshold DS evidence theory to obtain the transformer state assessment probability;

[0044] The status assessment module is used to assess the transformer status based on the transformer status assessment probability.

[0045] Thirdly, the present invention provides a computer-readable storage medium having computer instructions stored thereon, which, when executed by a processor, implement the steps of any of the transformer condition assessment methods described in the first aspect.

[0046] Fourthly, the present invention provides a computer device, comprising:

[0047] Memory, used to store computer instructions;

[0048] A processor for executing the computer instructions to implement the steps of the transformer condition assessment method as described in any one of the first aspects.

[0049] Fifthly, the present invention provides a computer program product including computer instructions, characterized in that, when the computer instructions are executed by a processor, they implement the steps of the transformer state assessment method described in any one of the first aspects.

[0050] Compared with the prior art, the beneficial effects achieved by the present invention are as follows:

[0051] 1. The transformer condition assessment method provided by this invention extracts features from multi-parameter data through MLP, which can effectively extract nonlinear features from multi-source data; it combines dynamic threshold DS evidence theory to achieve multi-evidence fusion and condition assessment. DS evidence theory can handle uncertainty and realize the systematic fusion of multi-source evidence, solving the misjudgment problem of traditional fixed thresholds when transformer load fluctuates and seasonal changes; combining MLP and DS evidence theory can accurately assess the transformer condition.

[0052] 2. The transformer condition assessment system provided by this invention, by setting up a preprocessing module, a feature extraction module, a multi-evidence fusion module and a condition assessment module, jointly realizes transformer condition assessment, solves the problems of data heterogeneity, uncertainty and multi-evidence fusion, and is applicable to various data types collected by multi-parameter optical sensors inside the transformer, such as temperature, pressure, partial discharge, etc., and has practical significance and good application prospects.

[0053] 3. The computer-readable storage medium, device, and product provided by the present invention can execute the steps of the transformer condition assessment method provided by the present invention. Attached Figure Description

[0054] Figure 1 This is a flowchart of a transformer condition assessment method provided according to an embodiment of the present invention;

[0055] Figure 2 This is a flowchart of feature extraction provided according to an embodiment of the present invention;

[0056] Figure 3 A flowchart illustrating the multi-evidence fusion process provided according to an embodiment of the present invention. Detailed Implementation

[0057] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present invention and the specific features in the embodiments are detailed descriptions of the technical solution of the present invention, rather than limitations thereof. In the absence of conflict, the embodiments of the present invention and the technical features in the embodiments can be combined with each other.

[0058] It should be noted that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.

[0059] Example 1:

[0060] This invention discloses a transformer condition assessment method, with reference to... Figure 1 As shown, the specific steps include the following:

[0061] S1, perform data preprocessing on the acquired transformer multi-parameter optical sensing data to obtain preprocessed data;

[0062] S2, use a multilayer perceptron to extract features from the preprocessed data to obtain a multi-source data feature set;

[0063] S3, each feature in the multi-source data feature set is used as an independent piece of evidence, and multi-evidence fusion is performed using the dynamic threshold DS evidence theory to obtain the transformer state assessment probability.

[0064] S4, assess the transformer status based on the transformer status assessment probability.

[0065] Specifically, in step S1, this embodiment collects multi-parameter optical sensing data of a 500kV voltage level transformer, including three types of data: temperature, pressure, and partial discharge; the acquired transformer multi-parameter optical sensing data is preprocessed, including:

[0066] The process involves three steps: noise reduction, normalization, and time alignment.

[0067] Wavelet transform or adaptive filtering is used to denoise the acquired transformer multi-parameter optical sensing data to obtain denoised data;

[0068] The denoised data is normalized to the [0,1] interval to obtain the normalized data:

[0069] ,

[0070] in, This represents the normalized data. , This represents the data after denoising. This represents the minimum value of the data after denoising. This represents the maximum value of the data after noise reduction;

[0071] The normalized data is interpolated and aligned to ensure temporal consistency, resulting in preprocessed data.

[0072] In step S2, a multilayer sensor is used to extract features from multiple parameters such as temperature, pressure, and partial discharge at locations such as transformer windings and oil surface. Hot spot temperature, maximum pressure, partial discharge amplitude, and the skewness and steepness of the partial discharge phase spectrum are selected as inputs to the multilayer sensor to reflect the response of multiple parameters such as temperature, pressure, and discharge to the state.

[0073] The multilayer perceptron introduces one or more hidden layers on top of a single-layer neural network. Each layer is fully connected, and the output of each hidden layer is transformed by an activation function. After calculating the weighted sum of the hidden units in each hidden layer, an activation function, usually the tangent hyperbolic function tanh, is applied to the result. The tanh function can transform the value of the element to between -1 and 1, approaching -1 when the input value is small and approaching +1 when the input value is large. The activation function can introduce nonlinear factors into the neurons, allowing the neural network to arbitrarily approximate any nonlinear function and be better applied to various nonlinear models.

[0074] The input of the multilayer perceptron is Feature extraction of the preprocessed data using a multilayer perceptron is achieved through the following formula:

[0075] ;

[0076] ;

[0077] ;

[0078] in, Indicates the Hidden layer The output of each hidden unit Indicates the Hidden layer The weights of each hidden unit, Indicates the number of data after preprocessing. One parameter, This indicates the total number of parameters in the preprocessed data. Indicates the Hidden layer The bias of each hidden unit. This represents the activation function. Indicates the Hidden layer The output of each hidden unit Indicates the Hidden layer The output of each hidden unit Indicates the The total number of hidden units in the hidden layer. Indicates the first Hidden layer The weights of each hidden unit, Indicates the first Hidden layer The bias of each hidden unit. Indicates the total number of hidden layers. Indicates the first The total number of hidden units in the hidden layer. Indicates the first Hidden layer The weights of each hidden unit, Indicates the first Hidden layer The output of each hidden unit Indicates the first The bias of the hidden layer, This represents a set of features from multiple data sources.

[0079] refer to Figure 2 As shown, this embodiment designs a two-layer perceptron model with n features as input; the first hidden layer has 3 hidden units, the second hidden layer has 3 hidden units, and the output features are... .

[0080] In step S3, the features extracted by MLP are used as evidence, and DS evidence theory is used to perform multi-evidence fusion to generate the transformer state assessment probability. When the number of evidences is greater than 2, this embodiment performs DS fusion on multiple pieces of evidence and obtains the transformer state assessment probability after multiple fusions.

[0081] Dempster's evidence theory is an algorithm for reasoning under uncertainty. It introduces probability assignment functions, confidence functions, and likelihood functions, reducing the requirements of traditional probability theory, which necessitates complete prior knowledge, conditional probabilities, and a unified identification framework. It can effectively express not only random uncertainty but also incomplete and subjective uncertainty. The Dempster combination rule in the theory satisfies the commutative and associative laws, allowing for the fusion of evidence even without prior conditions, effectively reducing the uncertainty of the system.

[0082] refer to Figure 3 As shown, the steps of multi-evidence fusion in DS evidence theory include: establishing a state assessment framework, selecting evidence and constructing a basic probability allocation, calculating dynamic thresholds, and multi-evidence fusion.

[0083] In DS evidence theory, the set of all possible outcomes for deciding a problem is called the identification framework; the identification framework (transformer condition assessment framework) in this embodiment includes seven types of transformer conditions: normal, low-temperature overheating, medium-temperature overheating, high-temperature overheating, partial discharge, low-energy discharge, and high-energy discharge. The expression is as follows:

[0084] ,

[0085] in, As a proposition, This indicates that the transformer is normal. This indicates that the transformer is overheating at a low temperature. This indicates that the transformer is overheating. This indicates that the transformer is overheating. This indicates partial discharge in the transformer. This indicates low-energy discharge in the transformer. This indicates high-energy discharge from the transformer. It represents uncertainty.

[0086] The power set of the transformer state assessment framework is used to construct a set of propositions. For any proposition Its basic probability allocation function The following conditions must be met:

[0087] ,

[0088] in, To represent the empty set, The basic probability assignment value of the empty set is 0. Expressing a proposition The basic probability assignment value, that is, the confidence level accurately assigned to A, represents the degree to which the evidence supports A; , Represents the set of propositions The sum of the basic probability assignments of all propositions in the equation is 1.

[0089] This example uses the output of transformer multi-parameter optical sensing data extracted by a multilayer sensor. As six independent pieces of evidence, they are transformed into basic probability assignments that satisfy the DS evidence theory; this embodiment exemplifies the fusion of six independent pieces of evidence; this embodiment sets... and The basic probability assignments for two independent pieces of evidence are respectively used. express and The new evidence after combination is expressed by Dempster's combination rule as follows:

[0090] ;

[0091] in, and Expressing a proposition The hypothetical subset of the two pieces of evidence, express The basic probability allocation value, express The basic probability allocation value, express and The intersection is , express and The intersection is empty. The conflict factor represents the degree of conflict between two pieces of evidence. After five fusions of six independent pieces of evidence, the probabilities of seven transformer states—normal, low-temperature overheating, medium-temperature overheating, high-temperature overheating, partial discharge, low-energy discharge, and high-energy discharge—are derived, which are the transformer state assessment probabilities. ,in, Indicates the transformer status. .

[0092] In step S4, the transformer state is assessed based on the transformer state assessment probability, including:

[0093] When the transformer state assessment probability satisfies the following three rules, the output is... As a result of transformer condition assessment:

[0094] Rule 1: ,

[0095] Rule 2: ,

[0096] Rule 3: ,

[0097] in, This represents the function that takes the maximum value. express The proposition with the largest basic probability distribution value. express The second largest proposition in the basic probability assignment is... Indicates a dynamic threshold. The basic probability assignment value represents the uncertainty; Rule 1 states that the proposition with the largest basic probability assignment value should be selected as the output; Rule 2 states that the basic probability assignment value of the judgment result must be greater than the basic probability assignment value of the uncertainty; and Rule 3 states... Assign the maximum basic probability value for a given judgment result type. The second largest basic probability allocation value is the difference between the basic probability allocation value of the judgment result and the basic probability allocation values ​​of other propositions. Only when all three rules mentioned above are met can the state evaluation result be output.

[0098] The dynamic threshold It can be obtained through the following formula:

[0099] ;

[0100] in, This represents the standard deviation of the recent state probability. Real-time load rate data of the transformer is obtained from the transformer condition monitoring SCADA system, and the standard deviation of the confidence level of the past 5 state assessment results is calculated. This represents the baseline threshold, which is a constant. C This represents the environmental correction factor, which is adjusted according to the transformer load. Its value ranges from 0 to 1, with 1 for full load and 0 for no load. It is obtained through linear mapping. The weights represent the standard deviations of recent state probabilities. The weight representing the baseline threshold, This indicates the weight of the environmental correction factor.

[0101] When the transformer condition assessment result is low temperature overheating, medium temperature overheating, high temperature overheating, partial discharge, low energy discharge, or high energy discharge, the corresponding early warning signal is triggered.

[0102] In summary, this embodiment proposes a transformer condition assessment method based on the combination of Multilayer Perceptron (MLP) and DS evidence theory. It extracts features from multi-parameter data through MLP and combines dynamic threshold DS evidence theory to achieve multi-evidence fusion and condition assessment, solving the challenges of data heterogeneity, uncertainty, and multi-evidence fusion. The specific process includes multi-parameter data preprocessing, MLP feature extraction, DS evidence theory multi-evidence fusion, transformer condition assessment, and early warning. This method solves the problem of inaccurate assessment caused by the complexity of state variables and the large amount of evidence in transformer condition assessment, enabling accurate assessment of transformer condition. It is applicable to various data types collected by multi-parameter optical sensors inside the transformer, such as temperature, pressure, and partial discharge.

[0103] Example 2:

[0104] Based on the same inventive concept as Embodiment 1, this embodiment of the invention discloses a transformer condition assessment system, comprising:

[0105] The preprocessing module is used to: perform data preprocessing on the acquired transformer multi-parameter optical sensing data to obtain preprocessed data;

[0106] The feature extraction module is used to: extract features from the preprocessed data using a multilayer perceptron to obtain a multi-source data feature set;

[0107] The multi-evidence fusion module is used to: treat each feature in the multi-source data feature set as an independent piece of evidence, and perform multi-evidence fusion using the dynamic threshold DS evidence theory to obtain the transformer state assessment probability;

[0108] The status assessment module is used to assess the transformer status based on the transformer status assessment probability.

[0109] The specific functions of each module described above are explained in the relevant content of the method in Embodiment 1, and will not be repeated here.

[0110] Example 3:

[0111] This embodiment provides a computer-readable storage medium storing computer instructions thereon, which, when executed by a processor, implement the steps of the transformer condition assessment method as described in any of the embodiments.

[0112] Example 4:

[0113] This embodiment provides a computer device, including:

[0114] Memory, used to store computer instructions;

[0115] A processor is configured to execute the computer instructions to implement the steps of the transformer condition assessment method described in any one of Embodiment 1.

[0116] Example 5:

[0117] This embodiment provides a computer program product, including computer instructions, characterized in that, when executed by a processor, the computer instructions implement the steps of the transformer condition assessment method as described in any one of Embodiment 1.

[0118] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0119] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A system that specifies functions in one or more boxes.

[0120] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including an instruction set implemented in a process. Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0121] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0122] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims. All of these forms are within the protection scope of the present invention.

Claims

1. A method for assessing the condition of a transformer, characterized in that, include: The acquired multi-parameter optical sensing data of the transformer is preprocessed to obtain preprocessed data; The preprocessed data is used to extract features using a multilayer perceptron to obtain a multi-source data feature set; Each feature in the multi-source data feature set is used as an independent piece of evidence. Multi-evidence fusion is performed using the dynamic threshold DS evidence theory to obtain the transformer state assessment probability. The transformer condition is assessed based on the transformer condition assessment probability. The step of treating each feature in the multi-source data feature set as independent evidence and using DS evidence theory to perform multi-evidence fusion includes: Based on the multi-source data feature set, a transformer condition assessment framework is established; the transformer condition assessment framework The expression is as follows: ， in, As a proposition, This indicates that the transformer is normal. This indicates that the transformer is overheating at a low temperature. This indicates that the transformer is overheating. This indicates that the transformer is overheating. This indicates partial discharge in the transformer. This indicates low-energy discharge in the transformer. This indicates high-energy discharge from the transformer. Indicates uncertainty; The power set of the transformer state assessment framework is used to construct a set of propositions. The expression for the basic probability assignment function m is obtained as follows: ， in, To represent the empty set, The basic probability assignment value for the empty set is 0. This represents the basic probability distribution value of proposition A. , and Denotes the subset of hypotheses representing two pieces of evidence for proposition A. express The basic probability allocation value, express The basic probability allocation value, express and The intersection of them is A. express and The intersection is empty. Indicates conflict factor, Represents the set of propositions The sum of the basic probability assignments of all propositions in the equation is 1. Based on the basic probability allocation function m, each feature in the multi-source data feature set is treated as independent evidence, and multi-evidence fusion is performed to obtain the transformer state assessment probability. ,in, Indicates the transformer status. .

2. The transformer condition assessment method according to claim 1, characterized in that, The data preprocessing of the acquired transformer multi-parameter optical sensing data includes: Wavelet transform or adaptive filtering is used to denoise the acquired transformer multi-parameter optical sensing data to obtain denoised data; The denoised data is normalized to obtain normalized data; The normalized data is interpolated and aligned to ensure temporal consistency, resulting in preprocessed data. The normalization of the denoised data is achieved using the following formula: ， in, This represents the normalized data. , This represents the data after denoising. This represents the minimum value of the data after denoising. This represents the maximum value of the data after noise reduction.

3. The transformer condition assessment method according to claim 1, characterized in that, The feature extraction of the preprocessed data using a multilayer perceptron is achieved through the following formula: ， in, Indicates the first hidden layer. The output of each hidden unit Indicates the first hidden layer. The weights of each hidden unit, Indicates the number of data after preprocessing. There are 10 parameters, where n represents the total number of parameters in the preprocessed data. Indicates the first hidden layer. The bias of each hidden unit. This represents the activation function. Represents the t-th hidden layer. The output of each hidden unit Indicates the first Hidden layer The output of each hidden unit This represents the total number of hidden units in the t-th hidden layer. Indicates the first Hidden layer The weight of each hidden unit, Represents the t-th hidden layer. The bias of each hidden unit, where T represents the total number of hidden layers. This represents the total number of hidden units in the T-th hidden layer. Indicates the T-th hidden layer. The weights of each hidden unit, Indicates the T-th hidden layer. The output of the hidden unit, b represents the bias of the T-th hidden layer. This represents a set of features from multiple data sources.

4. The transformer condition assessment method according to claim 1, characterized in that, The transformer condition is assessed based on the transformer condition assessment probability, including: When the transformer state assessment probability satisfies the following rule, the output is... As a result of transformer condition assessment: ， in, This represents the function that takes the maximum value. express The proposition with the largest basic probability distribution value. express The second largest proposition in the basic probability assignment is... Indicates dynamic threshold. The basic probability assignment value representing uncertainty; The dynamic threshold It can be obtained through the following formula: ； in, This represents the standard deviation of the recent state probability. C represents the baseline threshold, and C represents the environmental correction factor. The weights represent the standard deviations of recent state probabilities. The weight representing the baseline threshold, This indicates the weight of the environmental correction factor.

5. The transformer condition assessment method according to claim 1, characterized in that, When the assessment result of the transformer condition is low temperature overheating, medium temperature overheating, high temperature overheating, partial discharge, low energy discharge, or high energy discharge, the corresponding early warning signal is triggered.

6. A transformer condition assessment system, characterized in that, A method for implementing the transformer condition assessment method according to any one of claims 1 to 5 includes: The preprocessing module is used to: perform data preprocessing on the acquired transformer multi-parameter optical sensing data to obtain preprocessed data; The feature extraction module is used to: extract features from the preprocessed data using a multilayer perceptron to obtain a multi-source data feature set; The multi-evidence fusion module is used to: treat each feature in the multi-source data feature set as an independent piece of evidence, and perform multi-evidence fusion using the dynamic threshold DS evidence theory to obtain the transformer state assessment probability; The status assessment module is used to assess the transformer status based on the transformer status assessment probability.

7. A computer-readable storage medium storing computer instructions thereon, characterized in that, When the computer instruction is executed by the processor, it implements the steps of the transformer condition assessment method according to any one of claims 1 to 5.

8. A computer device, characterized in that, include: Memory, used to store computer instructions; A processor for executing the computer instructions to implement the steps of the transformer condition assessment method according to any one of claims 1 to 5.

9. A computer program product comprising computer instructions, characterized in that, When executed by a processor, the computer instructions implement the steps of the transformer condition assessment method according to any one of claims 1 to 5.

Images