Operation state evaluation method and system of distribution transformer, equipment and storage medium

Abstract

The embodiment of the invention discloses an operation state evaluation method and system of a distribution transformer, equipment and a storage medium, and the method comprises the steps: optimizing the operation data indexes of the distribution transformer through a digital twin model according to a quantum optimization weight vector, and carrying out the coupling analysis through a dynamic physical field weighted coupling algorithm, generating a stress weight vector; and based on the stress weight vector, predicting a health index of the distribution transformer, performing state classification, generating a state evaluation report, based on a quantum bit sequence evaluation model, performing quantum annealing calculation on quantum bit representation, and identifying a quantum optimization weight vector through a Fouli operator expected value analysis and probability normalization method. According to the method, global optimal weight search and statistical stability analysis are realized, the physical interpretability and dynamic adaptability of the weight vector are ensured, and accurate parameter support is provided for the digital twin model, so that the accuracy and reliability of health degree prediction are improved.

Description

Technical Field

[0001] This invention relates to the field of power systems and their automation technology, and in particular to a method, system, equipment and storage medium for evaluating the operating status of a distribution transformer. Background Technology

[0002] As a core piece of equipment in the power distribution network, the stability of the operating status of distribution transformers directly affects the safety and reliability of the power system. Traditional operating status assessment methods mainly rely on collecting operating parameters such as voltage, current, and temperature, and combining statistical analysis, empirical formulas, or rule-based threshold judgments for status classification. Based on this, signal preprocessing techniques such as low-pass filtering or moving averages are typically used to denoise the raw data. Then, normalization is employed to unify the dimensions, constructing health indicators to achieve a quantitative assessment of equipment status. This type of method is relatively mature in engineering practice and is widely used in power equipment condition monitoring and fault early warning systems.

[0003] However, traditional methods rely heavily on expert experience or linear regression models to determine feature weights, making it difficult to accurately characterize the nonlinear relationships between features. At the same time, they have limited ability to suppress noise during data preprocessing, especially when faced with high-frequency interference generated under complex working conditions, which can easily cause distortion of effective signals. These two aspects limit the robustness and accuracy of existing evaluation models in complex environments, and affect the precision and interpretability of state recognition. Summary of the Invention

[0004] Based on this, it is necessary to propose a method, system, equipment, and storage medium for evaluating the operating status of distribution transformers to address the above-mentioned problems.

[0005] A method for assessing the operating status of a distribution transformer, the method comprising, Collect the operating data indicators of the distribution transformer, and eliminate high-frequency noise and normalize the data indicators; The processed operational data indicators are converted into coded states of qubits using the rotating gate quantum coding method, and a qubit sequence evaluation model is constructed based on the coded states. Based on the qubit sequence evaluation model, quantum annealing calculation is performed on the encoded state of the qubit to generate multiple sampling results of the lowest energy state and read the quantum state to obtain the optimal solution of the lowest energy state and identify the quantum optimization weight vector. Based on the quantum-optimized weight vector, the operational data indicators are optimized through a digital twin model, and a dynamic physical field weighted coupling algorithm is used for coupling analysis to generate a stress weight vector. Based on the stress weight vector, the health index of the distribution transformer is predicted, and the state is classified according to the prediction results to generate a state assessment report.

[0006] Furthermore, the process of collecting and normalizing the operational data indicators, specifically including eliminating high-frequency noise, includes: The operational data indicators include voltage, current, active power, reactive power, frequency, winding temperature, oil temperature, harmonic content, load rate, and three-phase imbalance on both the high-voltage and low-voltage sides. The cutoff frequency and filter order are determined based on the operating data indicators. A Butterworth filter is defined using the bilinear transform method to perform zero-phase filtering on the operating data to eliminate high-frequency noise. The operational data indicators after eliminating high-frequency noise are normalized.

[0007] Furthermore, the process of converting the processed operational data indicators into coded states of qubits using a rotating gate quantum coding method specifically includes: The running data indicators are divided into blocks according to feature dimensions, with each block corresponding to the encoding task of a qubit. The encoding task of all qubits is rotated using the rotation gate quantum coding method to form the encoded state of the qubit.

[0008] Furthermore, the construction of the qubit sequence evaluation model based on the encoded state specifically includes: Using a Hamiltonian model that includes spin operators, the encoded states of qubits are arranged and combined in a feature-based manner through a qubit state sequence recombination method to generate qubit state sequences. The quantum bit state sequence is input into a quantum circuit to generate a multi-body entangled state. The lowest energy state of the Hamiltonian is obtained through the quantum tunneling effect, and the weight coefficients in the Hamiltonian are generated to construct a quantum bit sequence evaluation model.

[0009] Furthermore, the quantum bit sequence evaluation model performs quantum annealing calculations on the encoded states of the qubits, generates multiple sampling results of the lowest energy state and reads the quantum state, obtains the optimal solution of the lowest energy state, and identifies the quantum optimization weight vector, specifically including: Based on the qubit sequence evaluation model, quantum annealing calculations are performed on the encoded states of the qubits to generate the lowest energy state result; Based on multiple sampling results of the lowest energy state, the quantum state is read to obtain the optimal solution of the lowest energy state; Based on the optimal solution of the lowest energy state, the expected value is obtained and analyzed in reverse to generate weight coefficients and identify the quantum optimization weight vector; Among them, through The minimum energy state result generated by the annealing calculation is determined as follows: Where, Indicates the first Local weighting coefficients for each qubit; express Pauli-Z operator for 1 qubit, express Pauli-X operator with 1 qubit, Representing a quantum bit and The coupling weight coefficients, express Pauli-X operator with 1 qubit, Indicates the lowest energy state. This represents the Pauli-X operator. This represents the Pauli-Z operator. Indicates the number of qubits; The quantum-optimized weight vector is input into the digital twin model to adjust the parameters of the operational data indicators, and then... Determine the optimal parameter solution; in, This represents the optimal parameter solution. This represents the normalized operational data metrics. Represents the mapping and simulation functions of the digital twin model. Indicates by Constructed diagonal weighted matrix, Represents the quantum optimization weight vector. Indicates the trade-off coefficient. This indicates parameter regularization and physical deviation penalty; Based on the operational data indicators of the optimal parameter solution, the dynamic physical field weighted coupling algorithm is used to perform weighted coupling analysis on the interaction relationship between different physical fields, and generate stress weight vectors that affect the operational state. Among them, through Determine the stress weight vector In the formula, Represents the stress weight vector. ( ) represents the probability normalization operator. Indicates the first Feature projection and extraction matrix of each physical field This indicates the number of physical fields included in the coupling analysis. Indicates the first The field weights of each physical field. Indicates the first With the The coupling coefficient of a physical field. Indicates the first Feature projection and extraction matrix of each physical field.

[0010] Furthermore, the step of predicting the health index of the distribution transformer based on the stress weight vector, and classifying its state according to the prediction results to generate a state assessment report, specifically includes: The operational data indicators are transformed into a health index through normalization and weighted fusion; wherein, the health index is obtained through... Determine; in the formula, Indicates the health index, Hierarchical Analysis (AA) The resulting weight coefficient vector, Optimized operational data metrics.

[0011] A health function is constructed by weighting and summing the stress weight vectors using a nonlinear influence model; the expression for the health function is as follows: In the formula, Represents the health function, ( ) represents a nonlinear mapping. Represents the stress weight vector; Based on the health function, the health index of the distribution transformer is predicted and its status is classified by combining a time series prediction model with multiphysics coupling analysis, and a status assessment report is generated.

[0012] An operation status assessment system for the aforementioned distribution transformer operation status assessment method, the system comprising: The data acquisition module is used to collect the operating data indicators of the distribution transformer, and to eliminate high-frequency noise and normalize the data indicators. An evaluation model module is constructed to convert the processed running data indicators into the encoded states of qubits using the rotating gate quantum coding method, and to construct a qubit sequence evaluation model based on the encoded states. The weight vector identification module is used to evaluate the model based on the qubit sequence, perform quantum annealing calculations on the encoded state of the qubit, generate multiple sampling results of the lowest energy state and read the quantum state, obtain the optimal solution of the lowest energy state, and identify the quantum optimization weight vector. A stress weight generation module is used to optimize the running data indicators through a digital twin model based on the quantum optimized weight vector, and to perform coupling analysis using a dynamic physical field weighted coupling algorithm to generate a stress weight vector. The assessment report generation module is used to predict the health index of the distribution transformer based on the stress weight vector, and to classify the state according to the prediction results to generate a state assessment report.

[0013] A computer device includes a memory and a processor, the memory storing a computer program that, when executed by the processor, causes the processor to perform the following steps: Collect the operating data indicators of the distribution transformer, and eliminate high-frequency noise and normalize the data indicators; The processed operational data indicators are converted into coded states of qubits using the rotating gate quantum coding method, and a qubit sequence evaluation model is constructed based on the coded states. Based on the qubit sequence evaluation model, quantum annealing calculation is performed on the encoded state of the qubit to generate multiple sampling results of the lowest energy state and read the quantum state to obtain the optimal solution of the lowest energy state and identify the quantum optimization weight vector. Based on the quantum-optimized weight vector, the operational data indicators are optimized through a digital twin model, and a dynamic physical field weighted coupling algorithm is used for coupling analysis to generate a stress weight vector. Based on the stress weight vector, the health index of the distribution transformer is predicted, and the state is classified according to the prediction results to generate a state assessment report.

[0014] A computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to perform the following steps: Collect the operating data indicators of the distribution transformer, and eliminate high-frequency noise and normalize the data indicators; The processed operational data indicators are converted into coded states of qubits using the rotating gate quantum coding method, and a qubit sequence evaluation model is constructed based on the coded states. Based on the qubit sequence evaluation model, quantum annealing calculation is performed on the encoded state of the qubit to generate multiple sampling results of the lowest energy state and read the quantum state to obtain the optimal solution of the lowest energy state and identify the quantum optimization weight vector. Based on the quantum-optimized weight vector, the operational data indicators are optimized through a digital twin model, and a dynamic physical field weighted coupling algorithm is used for coupling analysis to generate a stress weight vector. Based on the stress weight vector, the health index of the distribution transformer is predicted, and the state is classified according to the prediction results to generate a state assessment report.

[0015] The embodiments of the present invention have the following beneficial effects: This invention is based on a qubit sequence evaluation model, performs quantum annealing calculations on the qubit representation, and identifies the quantum-optimized weight vector through Pauli operator expectation analysis and probability normalization methods. It achieves global optimal weight search and statistical stability analysis, ensuring the physical interpretability and dynamic adaptability of the weight vector, providing accurate parameter support for digital twin models, thereby improving the accuracy and reliability of health prediction. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] in: Figure 1 The present invention provides a flowchart of a method for evaluating the operating status of a distribution transformer.

[0018] Figure 2 An embodiment of the present invention also provides a connection block diagram of a power distribution transformer operation status assessment system. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] This invention provides a method for evaluating the operating status of a distribution transformer, such as... Figure 1 , 2 As shown, it includes the following steps: S1. Collect the operating data indicators of the distribution transformer, and eliminate high-frequency noise and normalize the data indicators. Specifically, the operating data indicators of the distribution transformer include voltage, current, active power, reactive power, frequency, winding temperature, oil temperature, harmonic content, load rate, and three-phase imbalance on the high-voltage side and low-voltage side. Based on the operating data indicators, the cutoff frequency and filter order of the operating conditions are defined. The Butterworth filter is initialized using the bilinear transform method, and the operating data of the distribution transformer is subjected to zero-phase filtering to eliminate high-frequency noise. It should be noted that, based on the fluctuation characteristics of parameters such as voltage, current, and temperature in the operating data indicators, the frequency domain characteristics of the signal are analyzed to determine the cutoff frequency and filter order suitable for the operating conditions; based on the determined cutoff frequency and filter order, the Butterworth filter model is initialized using the bilinear transform method, and the transfer function is constructed; the transfer function is used to perform zero-phase filtering on the original operating data to eliminate high-frequency noise.

[0021] The operational data indicators after eliminating high-frequency noise are normalized.

[0022] It should be noted that, firstly, the maximum and minimum values ​​of each feature dimension are calculated, which are called the maximum and minimum running data values, respectively. Next, the min-max normalization formula is applied to transform all running data indicators to the [0,1] interval, and the normalized running data indicators are output.

[0023] S2. The processed operating data indicators are converted into coded states of qubits using the rotating gate quantum coding method, and a qubit sequence evaluation model is constructed based on the coded states; Specifically, time series analysis is used to extract feature dimensions from the operational data indicators and divide them into blocks. Each block corresponds to the encoding task of a qubit. The encoding task of all qubits is rotated using the rotating gate quantum coding method to form the encoding state of the qubit. It should be noted that a time series analysis method is used to extract feature dimensions from the operational data indicators. The normalized operational data indicators are then divided according to the feature dimensions, with each feature dimension corresponding to an independent data subset, and the feature dimension data subset is output. A qubit encoding task is assigned to each feature dimension data subset, and a quantum circuit structure that matches it one-to-one is constructed. A rotation gate quantum encoding method is used to perform rotation operations around the Y-axis or Z-axis on the corresponding qubits based on the numerical information of each feature dimension data subset, generating single-qubit states with feature mapping relationships. All qubit states that have completed the rotation operations are combined to form complete multi-qubit states, and the encoded state of the qubits is output.

[0024] The task of encoding qubits refers to dividing the running data indicators into blocks according to feature dimensions, and converting the data of each feature dimension into the quantum state of the corresponding qubit using the rotating gate quantum coding method.

[0025] A Hamiltonian model containing spin operators is adopted, and a quantum bit state sequence is generated by arranging and combining the features of the encoded state through a quantum bit state sequence recombination method. pass Determine the sequence of states of a qubit; where, Represents the sequence of states of a quantum bit. Indicates according to the permutation rule Operators that rearrange the order of qubits Indicates circling A single-qubit rotation gate of the axis Around A single-qubit rotation gate of the axis Represents the tensor product notation. Represents the number of qubits. Indicates the first One quantum bit is orbiting Rotation angle parameters on a rotary door. Indicates the first One quantum bit is orbiting Rotation angle parameters on a rotary door. Indicates the length of the state sequence. Indicates the first The qubit index permutation corresponding to the sub-characteristic permutation and combination.

[0026] It should be noted that the Hamiltonian model, which includes spin operators, is adopted. First, the interaction strength between each qubit is obtained based on the coded state of the qubit, and the corresponding spin operator matrix representation is determined, outputting a set of spin operator matrices. Then, using the qubit state sequence recombination method, the coded states of the qubits are reordered and combined according to the importance of the feature dimensions to generate a new qubit arrangement order, outputting the reordered qubit coded states. Subsequently, based on the set of spin operator matrices and the reordered qubit coded states, a specific Hamiltonian expression is constructed, and the Hamiltonian expression is applied to further process the qubit states, ultimately generating a qubit state sequence that can reflect the characteristic relationships of the power distribution transformer's operating data.

[0027] It should be noted that the training process of the Hamiltonian model first defines a Hamiltonian expression containing spin operators based on the encoded states of the qubits and a subset of feature dimension data. This expression reflects the interaction strength between the feature dimensions. Next, the Hamiltonian parameters are initialized, and the encoded states of the qubits are loaded into the quantum circuit as input. Then, through a quantum-classical hybrid algorithm, the local weights and coupling weights in the Hamiltonian are iteratively adjusted multiple times under the guidance of a classical optimizer to generate the expected energy value. During this process, the best solution of each iteration is recorded, and the Hamiltonian parameters are gradually updated according to the best solution. Finally, the trained Hamiltonian model is output.

[0028] The quantum bit state sequence is input into the quantum circuit, a multi-body entangled state is generated through the CNOT gate, and the weighting coefficients in the Hamiltonian are generated through the quantum tunneling effect. pass Determine the weighting coefficients in the Hamiltonian; where, This represents the weighting coefficient in the Hamiltonian. Indicates the first Multiple entangled states This represents the problem Hamiltonian containing a spin operator. Indicates the first The lowest energy state obtained from secondary annealing and tunneling search. This represents the ridge regularity term, which suppresses ill-conditioned solutions and enhances statistical stability. This indicates the state sequence used and the number of sampling rounds. This represents the minimization operator.

[0029] Furthermore, the qubit state sequence is used as input to the quantum circuit, and CNOT gate operations are applied sequentially to adjacent qubits in the quantum circuit to generate multi-body entangled states with non-classical correlation properties. Based on the multi-body entangled states, a Hamiltonian model containing spin operators is introduced, and its evolution behavior during quantum annealing is simulated. The quantum tunneling effect is used to search for and determine the global lowest energy state, and the lowest energy state is output. According to the energy level distribution and quantum state configuration corresponding to the lowest energy state, the local weights and coupling weights describing the interaction strength between features in the Hamiltonian are deduced, and the weight coefficients in the Hamiltonian are generated.

[0030] pass Determine the multi-body entangled state; among which, Represents a multi-body entangled state. This refers to a chain of CNOT gates sequentially applied along a quantum circuit to adjacent qubits, causing non-classical correlations in the input states and forming many-body entanglement. Indicates that the control bit is the 1st The number of qubits, with the controlled bit being the 1st qubit. CNOT gate for 1 qubit Represents a sequence recombination operator, based on permutation. Rearrange the order of qubits. Indicates the first The qubit index permutation corresponding to the next recombination. This indicates that a single-bit rotation gate is applied independently to each qubit, and then the results are multiplied together to form a multi-bit operation. Indicates circling A single-qubit rotation gate of the axis Around A single-qubit rotation gate of the axis Indicates the first One quantum bit is orbiting Rotation angle parameters on a rotary door. Indicates the first One quantum bit is orbiting Rotation angle parameters on a rotary door. Indicates by The initial tensor product state composed of the ground states of 1 qubits This indicates the number of qubits.

[0031] A quantum bit sequence evaluation model is constructed based on the weighting coefficients in the Hamiltonian using a quantum-classical hybrid algorithm.

[0032] Furthermore, based on the weighting coefficients in the Hamiltonian and combined with a quantum-classical hybrid algorithm, a quantum-classical hybrid computing framework incorporating quantum evolution and classical feedback iteration is constructed. Based on this framework, the weighting coefficients in the Hamiltonian are first embedded as input parameters into the variational parameters of the quantum circuit to regulate the interaction strength between qubits. Subsequently, a classical optimizer is used to adjust the measurement results output by the quantum circuit, iteratively optimizing the rotation angle and control parameters in the quantum circuit, ultimately outputting a qubit sequence evaluation model.

[0033] It should be noted that the training process of the qubit sequence evaluation model first involves constructing a Hamiltonian model containing spin operators based on the weight coefficients in the Hamiltonian and the encoded state of the qubit. Next, the encoded state of the qubit is input into the quantum circuit, and a many-body entangled state is generated through a CNOT gate. The lowest energy state is obtained using the quantum tunneling effect, and the output result is called the lowest energy state. Then, based on the lowest energy state, the quantum state is read using quantum measurement and probabilistic statistical methods, and the optimal solution of the lowest energy state is obtained using the weight optimization inversion method. Subsequently, the Pauli operator expectation analysis method is used for further processing to generate weight coefficients reflecting the importance of features and their interactions. Finally, the weight coefficients are standardized using a probabilistic normalization method to generate the final quantum-optimized weight vector, and the qubit sequence evaluation model is trained based on this vector.

[0034] S3. Based on the qubit sequence evaluation model, perform quantum annealing calculation on the encoded state of the qubit, generate multiple sampling results of the lowest energy state and read the quantum state, obtain the optimal solution of the lowest energy state, and identify the quantum optimization weight vector. Specifically, based on the qubit sequence evaluation model, quantum annealing calculations are performed on the encoded states of the qubits to generate the lowest energy state, and the optimal solution of the lowest energy state is obtained through the weight optimization inversion method; The expression for calculating the minimum energy state generated by annealing is: ; in, Indicates the first Local weighting coefficients for each qubit; express Pauli-Z operator for 1 qubit, express Pauli-X operator with 1 qubit, Representing a quantum bit and The coupling weight coefficients, express Pauli-X operator with 1 qubit, Indicates the lowest energy state. This represents the Pauli-X operator. This represents the Pauli-Z operator. This indicates the number of qubits.

[0035] It should be noted that the encoded state of the qubit is used as the initial input and loaded into the quantum annealing computation framework. Parameters are controlled using a stepwise evolution method to simulate the quantum tunneling process, guiding the initial state to converge towards the lowest energy configuration. During annealing, the quantum state is sampled multiple times, and the probability distribution of the lowest energy state is recorded for each sampling result. Based on the sampling data, the quantum state with the highest frequency of occurrence is identified, and the lowest energy state is output. Subsequently, the probability distribution of the lowest energy state is statistically analyzed using quantum measurement methods to obtain the frequency information of each quantum state, and the quantum state measurement probability distribution is output. Then, the quantum state measurement probability distribution is normalized using probabilistic statistical methods to output the dominant quantum state. Finally, the dominant quantum state is substituted into the weight optimization inversion model, and the local weights and coupling weights are inversely derived using the Hamiltonian structure to output the optimal solution for the lowest energy state.

[0036] Based on the optimal solution of the lowest energy state, the expected value is obtained through the Pauli operator expectation value analysis method and analyzed and reversed to generate the weight coefficients of the quantum state. The quantum optimization weight vector is then identified through the probability normalization method.

[0037] Furthermore, firstly, the optimal solution for the lowest energy state is evaluated under the constraints of the Pauli-Z and Pauli-X operators in the state using the Pauli operator expectation analysis method, and its corresponding expectation value is calculated, outputting the Pauli operator expectation value. Next, the Pauli operator expectation value is inverted and analyzed, and combined with the weight term in the Hamiltonian expression, the original weight coefficients reflecting the importance and interaction of features are derived, and the output results are called the original weight coefficients. Subsequently, the original weight coefficients are standardized using the probability normalization method to make them comparable under a uniform scale, and finally the quantum optimization weight vector is identified.

[0038] S4. Based on the quantum-optimized weight vector, the running data indicators are optimized through a digital twin model, and a dynamic physical field weighted coupling algorithm is used for coupling analysis to generate a stress weight vector. Specifically, the quantum-optimized weight vector is input into the digital twin model to adjust the parameters of the operational data indicators, and through... Determine the optimal parameter solution; in, This represents the optimal parameter solution. This represents the normalized operational data metrics. Represents the mapping and simulation functions of the digital twin model. Indicates by Constructed diagonal weighted matrix, Represents the quantum optimization weight vector. Indicates the trade-off coefficient. This indicates parameter regularization and physical deviation penalty.

[0039] Based on the operational data indicators of the optimal parameter solution, a dynamic physical field weighted coupling algorithm is used to perform weighted coupling analysis on the interaction relationship between different physical fields, generating a stress weight vector that affects the operational state.

[0040] Among them, through Determine the stress weight vector In the formula, Represents the stress weight vector. ( ) represents the probability normalization operator. Indicates the first Feature projection and extraction matrix of each physical field This indicates the number of physical fields included in the coupling analysis. Indicates the first The field weights of each physical field. Indicates the first With the The coupling coefficient of a physical field. Indicates the first Feature projection and extraction matrix of each physical field.

[0041] It should be noted that the quantum optimization weight vector is imported into the digital twin model as an input parameter. Combined with the physical structure and operating mechanism of the distribution transformer, the parameters related to voltage, current and temperature in the digital twin model are dynamically weighted and adjusted, and the optimized digital twin model after parameter adjustment is output. Subsequently, based on the normalized operating data indicators as training input, the digital twin model after parameter adjustment is iteratively optimized using the physical constraint analysis method.

[0042] It should be noted that the construction process of the digital twin model first involves dynamically adjusting key operating parameters (such as voltage, current, and temperature) based on operational data indicators and combined with quantum-optimized weight vectors, outputting weighted operational data. Next, the weighted operational data is used to initialize the digital twin framework and establish a virtual model corresponding to the actual device. Then, through physical laws and rules (such as thermodynamics, electromagnetism, and mechanical stress analysis), each parameter in the weighted operational data is precisely simulated and verified to ensure that the behavior of the virtual model is consistent with the actual device. On this basis, multiphysics coupling analysis is further integrated to capture the interactions between different physical fields and optimize model performance. Finally, a complete digital twin model is output.

[0043] Based on the optimized operational data indicators, a dynamic physical field weighted coupling algorithm is used to perform weighted coupling analysis on the interaction between different physical fields, generating a stress weight vector that affects the operational status.

[0044] Furthermore, based on the operational data indicators output by the optimized digital twin model, a multiphysics model is first constructed to obtain the distribution characteristics of each physical field under different operating conditions. Then, a quantum-optimized weight vector is introduced as a weighting factor, and the interaction relationship between each physical field is weighted and fused through a dynamic physical field weighted coupling algorithm to quantify the influence intensity of different physical fields on the overall operating state of the equipment. Subsequently, key coupling features are extracted and normalized to generate a stress weight vector that reflects the complex internal interactions.

[0045] S5. Based on the stress weight vector, predict the health index of the distribution transformer, classify the state according to the prediction results, and generate a state assessment report.

[0046] Specifically, the operational data indicators are transformed into a health index through normalization and weighted fusion; wherein, the health index is obtained through... Determine; in the formula, Indicates the health index, Hierarchical Analysis (AA) The resulting weight coefficient vector, Optimized operational data metrics; A health function is constructed by weighting and summing the stress weight vectors using a nonlinear influence model; the expression for the health function is as follows: In the formula, Represents the health function, ( ) represents a nonlinear mapping. Represents the stress weight vector; It should be noted that the normalized operational data indicators are used as input to the multiphysics model. First, a weighted fusion method combined with quantum-optimized weight vectors is used to dynamically adjust each feature dimension, outputting the weighted fused operational data. Next, the analytic hierarchy process (AHP) is applied to integrate and analyze the stress weight vectors. By comparing the importance of each feature dimension pairwise, a judgment matrix is ​​constructed and weight coefficients are calculated, outputting the AHP weight coefficients. Then, the AHP weight coefficients are used to further weight and sum the weighted fused operational data to generate a health index reflecting the overall health status. Finally, a health function is constructed based on the health index and nonlinear mapping relationship.

[0047] Based on the health function, combined with time series prediction model and multiphysics coupling analysis, the operating status of distribution transformers is predicted and classified, and a status assessment report is generated.

[0048] It should be noted that, based on the health function, it is first input into the time series prediction model as the core assessment model, and the evolution trend of the health status of the distribution transformer is predicted by combining historical operating data, outputting a health index prediction sequence. Next, using the multiphysics coupling analysis method, the key parameters in the health index prediction sequence are dynamically corrected according to the interaction relationship between different physical fields such as electromagnetic field, thermal conduction field and mechanical stress field, outputting the multiphysics corrected health index. Then, the multiphysics corrected health index is graded and judged, and the status classification result is output. Finally, the above information is combined to generate a structured status assessment report.

[0049] It should be noted that the training process of the time series forecasting model (LSTM) first involves constructing a time series dataset based on historical health index data generated by the "health function". Next, a suitable time series forecasting algorithm (LSTM) is selected, and the time series dataset is divided into a training set and a validation set. Then, the time series forecasting model is trained on the training set, with model parameters adjusted to minimize prediction error. The validation set is used to evaluate model performance and ensure its generalization ability. During this process, rolling prediction or cross-validation techniques are used to further optimize model parameters and improve prediction accuracy. Finally, the trained time series forecasting model is output.

[0050] This invention also provides an operational status assessment system for distribution transformers, such as... Figure 2 As shown, the system includes: The data acquisition module is used to collect operational data indicators, eliminate high-frequency noise, and perform normalization processing. An evaluation model module is constructed to convert the running data indicators into the encoded states of qubits using the rotating gate quantum coding method, and to construct a qubit sequence evaluation model. The weight vector identification module is used to evaluate the model based on the qubit sequence, perform quantum annealing calculations on the encoded state of the qubit, generate multiple sampling results of the lowest energy state and read the quantum state, obtain the optimal solution of the lowest energy state, and identify the quantum optimization weight vector. The stress weight generation module is used to optimize the running data indicators based on the quantum optimization weight vector through the digital twin model, and to perform coupling analysis using the dynamic physical field weighted coupling algorithm to generate the stress weight vector. The assessment report generation module predicts the health index of distribution transformers and classifies their states based on stress weight vectors, generating a state assessment report.

[0051] This invention also provides a computer device applicable to the method for evaluating the operating status of distribution transformers, comprising: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the method for evaluating the operating status of distribution transformers as proposed in the above embodiments.

[0052] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0053] This embodiment also provides a storage medium storing a computer program, which, when executed by a processor, implements the method for evaluating the operating status of a distribution transformer as proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0054] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0055] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A method of evaluating an operating state of a distribution transformer, characterized by, The method comprises: Collecting operation data indicators of a power distribution transformer, eliminating high-frequency noise from the data indicators, and performing normalization processing on the data indicators; Converting the processed operation data indicators into an encoding state of a quantum bit through a rotation gate quantum encoding method, and constructing a quantum bit sequence evaluation model based on the encoding state; Based on the quantum bit sequence evaluation model, performing quantum annealing calculation on the encoding state of the quantum bit, generating multiple sampling results of the lowest energy state and reading the quantum state, obtaining the optimal solution of the lowest energy state, and identifying the quantum optimization weight vector; According to the quantum optimization weight vector, optimizing the operation data indicators through a digital twin model, and performing coupling analysis using a dynamic physical field weighted coupling algorithm to generate a stress weight vector; Based on the stress weight vector, predicting the health index of the power distribution transformer, and generating a state evaluation report according to the prediction result.

2. The method of operating state assessment of a distribution transformer according to claim 1, characterized in that, The collection of operation data indicators, the elimination of high-frequency noise from the data indicators, and the normalization processing of the data indicators specifically comprise: The operation data indicators include voltage, current, active power, reactive power, frequency, winding temperature, oil temperature, harmonic content, load rate, and three-phase imbalance on the high-voltage side and the low-voltage side; According to the operation data indicators, determine the cutoff frequency and filter order, define the Butterworth filter through the bilinear transformation method, perform zero-phase filter processing on the operation data to eliminate high-frequency noise; The operation data indicators after eliminating high-frequency noise are normalized.

3. The method of operating state assessment of a distribution transformer according to claim 1 or 2, characterized in that, The conversion of the processed operation data indicators into an encoding state of a quantum bit through a rotation gate quantum encoding method specifically comprises: The operation data indicators are divided into blocks according to feature dimensions, each block corresponds to an encoding task of a quantum bit, and quantum bit application rotation is performed on all quantum bit encoding tasks through a rotation gate quantum encoding method to form an encoding state of a quantum bit.

4. The method of operating state assessment of a distribution transformer according to claim 3, characterized in that, The construction of a quantum bit sequence evaluation model based on the encoding state specifically comprises: A Hamiltonian model containing spin operators is used to perform feature permutation and combination on the encoding state of the quantum bit through a quantum bit state sequence reorganization method to generate a quantum bit state sequence; The quantum bit state sequence is input into a quantum circuit to generate a many-body entangled state, and the lowest energy state of the Hamiltonian is obtained through quantum tunneling effect to generate weight coefficients in the Hamiltonian, and a quantum bit sequence evaluation model is constructed.

5. The method of operating condition assessment of a distribution transformer according to claim 4, characterized in that, Based on the quantum bit sequence evaluation model, performing quantum annealing calculation on the encoding state of the quantum bit, generating multiple sampling results of the lowest energy state and reading the quantum state, obtaining the optimal solution of the lowest energy state, and identifying the quantum optimization weight vector, specifically comprising: Based on the quantum bit sequence evaluation model, performing quantum annealing calculation on the encoding state of the quantum bit to generate the lowest energy state result; According to the multiple sampling results of the lowest energy state, read the quantum state, and obtain the optimal solution of the lowest energy state; Based on the optimal solution of the lowest energy state, obtain the expected value and perform analysis and backstepping to generate the weight coefficient and identify the quantum optimization weight vector; wherein the annealing computation generates a lowest energy state result by determining the annealing computation generates a lowest energy state result by representing a local weight coefficient of the quantum bit; representing a Pauli-Z operator of the quantum bit, representing a Pauli-X operator of the quantum bit, representing a coupling weight coefficient of the quantum bit and , representing a Pauli-X operator of the quantum bit, representing a lowest energy state, representing a Pauli-X operator, representing a Pauli-Z operator, representing a number of quantum bits.

6. The method of operating condition assessment of a distribution transformer according to claim 5, characterized in that, The stress weight vector is generated by optimizing the operation data index through a digital twin model according to the quantum optimization weight vector, and performing coupling analysis by using a dynamic physical field weighted coupling algorithm, specifically including: inputting the quantum optimization weight vector into a digital twin model, performing parameter adjustment on the operation data index, and determining an optimal parameter solution through determining an optimal parameter solution; wherein, represents the optimal parameter solution, represents the normalized operational data indicator, represents the mapping and simulation function of the digital twin model, represents the diagonal weighting matrix constructed from represents the quantum optimization weight vector, represents the trade-off coefficient, represents the parameter regularization and physical deviation penalty;​ Based on the optimal parameter solution of the operation data index, the stress weight vector affecting the operation state is generated by performing weighted coupling analysis on the action relationship between different physical fields through a dynamic physical field weighted coupling algorithm; Among them, through Determine the stress weight vector In the formula, Represents the stress weight vector. ( ) represents the probability normalization operator. Indicates the first Feature projection and extraction matrix of each physical field This indicates the number of physical fields included in the coupling analysis. Indicates the first The field weights of each physical field. Indicates the first With the The coupling coefficient of a physical field. Indicates the first Feature projection and extraction matrix of each physical field.

7. The method of operating condition assessment of a distribution transformer according to claim 6, characterized in that, The health degree index of the power distribution transformer is predicted based on the stress weight vector, and a state classification is performed according to the prediction result to generate a state evaluation report, specifically including: The operation data indexes are converted into health degree indexes through normalization processing and weighted fusion, wherein the health degree indexes are determined through ; wherein, represents the health degree index, represents a weight coefficient vector obtained through an analytic hierarchy process (AHP), and the operation data indexes are optimized. ​ The health degree function is constructed by performing weighted summation on the stress weight vector through a nonlinear influence model; wherein, the health degree function expression is: ; wherein, represents the health degree function, ( ) represents a nonlinear mapping, represents the stress weight vector; Based on the health degree function, the health degree index of the power distribution transformer is predicted and a state classification is performed by combining a time series prediction model with a multi-physical field coupling analysis to generate a state evaluation report.

8. An operation state evaluation system for use in the operation state evaluation method according to any one of claims 1 to 7, characterized by The system comprises: A data acquisition module for acquiring operation data indexes of the power distribution transformer, eliminating high-frequency noise from the data indexes, and performing normalization processing; A construction evaluation model module for converting the processed operation data indexes into an encoding state of a quantum bit by using a rotating gate quantum encoding method, and constructing a quantum bit sequence evaluation model based on the encoding state; An identification weight vector module for performing quantum annealing calculation on the encoding state of the quantum bit based on the quantum bit sequence evaluation model, generating multiple sampling results of the lowest energy state and reading the quantum state, obtaining the optimal solution of the lowest energy state, and identifying the quantum optimization weight vector; A stress weight generation module for optimizing the operation data index through a digital twin model according to the quantum optimization weight vector, and performing coupling analysis by using a dynamic physical field weighted coupling algorithm to generate a stress weight vector; An evaluation report generation module for predicting the health degree index of the power distribution transformer based on the stress weight vector, and performing a state classification according to the prediction result to generate a state evaluation report. 9.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is characterized in that: The processor executes the computer program to realize the steps of the power distribution transformer operation state evaluation method of any one of claims 1-7.

10. A computer readable storage medium having stored thereon a computer program, characterized in that: The computer program is executed by the processor to realize the steps of the power distribution transformer operation state evaluation method of any one of claims 1-7.

Images