Quantum computing-based power transformer measurement point optimization method and system

Abstract

The application provides a power transformer measurement point optimization method and system based on quantum computing, and relates to the technical fields of power equipment monitoring and quantum computing application. The method comprises the following steps: obtaining multi-source sensor monitoring data of a power transformer at each candidate measurement point, fusing the multi-source sensor monitoring data, and constructing a multi-dimensional feature vector of each candidate measurement point; mapping the multi-dimensional feature vector of each candidate measurement point into a corresponding feature quantum state to obtain a unified quantum state representation of multi-source high-dimensional features; based on the corresponding feature quantum state of each candidate measurement point, establishing a measurement point combination superposition state for the selection state of each candidate measurement point; using a quantum approximate optimization algorithm to evolve and solve the measurement point combination superposition state, obtaining the probability distribution of the candidate measurement point combination, and outputting a candidate measurement point combination scheme according to the probability distribution. The application obtains a more optimal measurement point layout scheme, and improves the coverage, sensitivity and diagnostic accuracy of transformer state parameter monitoring.

Description

Technical Field

[0001] This invention relates to the field of power equipment monitoring and quantum computing application technology, specifically to a method and system for optimizing power transformer measurement points based on quantum computing. Background Technology

[0002] With the deepening construction of new power systems, power grid operation exhibits characteristics such as high uncertainty on both the source and load sides, complex equipment operating conditions, and rapid dynamic changes in state variables. This places higher demands on the accuracy, timeliness, and intelligence level of data acquisition in substation operation and maintenance monitoring. Traditional monitoring methods, which rely on manual inspections and experience-based judgments, have limitations such as response lag, discontinuous data acquisition, and insufficient perception dimensions, making it difficult to meet the needs of modern power grids for refined, all-weather status perception of key equipment such as transformers. With the development of technologies such as artificial intelligence and big data analysis, intelligent monitoring methods are gradually becoming widespread in power systems. However, for complex equipment like power transformers, which have strong coupling, multiple physical field superposition, and highly time-varying parameters, existing intelligent monitoring points still rely on a large number of physical measuring points. The rationality of the measuring point layout directly determines whether the model can accurately capture key early features and support high-precision health assessment and fault prediction. Therefore, how to optimize the configuration of measuring point locations and quantities while ensuring comprehensive monitoring has become a key issue for the industry. At the same time, as a core hub device in the power grid, any potential fault in a power transformer can trigger a chain reaction of power outages, equipment damage, or even fires, posing a serious threat to the safety of the power system. To achieve transparent, digital, and intelligent perception of equipment status, the industry is accelerating the application of new technologies such as digital twins and proactive defense. Multi-parameter monitoring based on high-fidelity digital twin models requires deploying a limited number of measurement points on the equipment, maximizing their information value, to accommodate the complex coupling of internal field quantities within transformers.

[0003] However, existing intelligent monitoring of power transformers relies on experience or traditional numerical / heuristic algorithms. When faced with high-dimensional feature parameters and a large number of candidate measurement point combinations, they suffer from high computational complexity, low search efficiency, and a tendency to get trapped in local optima. It is difficult to obtain a globally optimal or near-optimal measurement point configuration scheme, resulting in insufficient monitoring coverage and inaccurate state perception. Summary of the Invention

[0004] To overcome the shortcomings of the prior art, this invention provides a method for optimizing power transformer measurement points based on quantum computing, comprising: Acquire multi-source sensor monitoring data of power transformer at each candidate monitoring point, perform fusion processing on the multi-source sensor monitoring data, and construct multi-dimensional feature vectors for each candidate monitoring point; Amplitude coding is used to map the multidimensional feature vectors of each candidate measurement point to the corresponding feature quantum states, thereby obtaining a unified quantum state representation of multi-source high-dimensional features; Based on the characteristic quantum states corresponding to each candidate test point, the quantum superposition principle is used to establish a superposition state of test point combination for the selected states of each candidate test point; A quantum approximation optimization algorithm is used to solve the evolution of the superposition state of the test point combination. The preset optimization objective function is used as the optimization objective to obtain the probability distribution of the candidate test point combination and output the candidate test point combination scheme according to the probability distribution.

[0005] Preferably, a quantum approximation optimization algorithm is used to evolve and solve the superposition state of the test point combination, with a preset optimization objective function as the optimization objective, to obtain the probability distribution of the candidate test point combination, including: A quantum approximation optimization algorithm based on parameterized quantum gates is adopted to perform multiple rounds of evolution on the superposition state of the test point combination. The preset optimization objective function is minimized or maximized by iteratively optimizing the parameters of the parameterized quantum gates. The evolution process includes the alternating application of a first evolution operator generated by the cost Hamiltonian and a second evolution operator generated by the mixer Hamiltonian. The cost Hamiltonian corresponds to a preset optimization objective function, and the mixer Hamiltonian is used to drive the quantum state to traverse and search in the solution space. The superposition state of the evolved measurement point combination is sampled and measured to obtain the optimized probability distribution of the candidate measurement point combination.

[0006] Preferably, the preset optimization objective function that minimizes or maximizes includes: minimizing the overall quantum distance or maximizing the overall quantum coincidence. The quantum coincidence of a single candidate measurement point is the square of the modulus of the inner product of the reference quantum state and the measurement point quantum state; whereby the reference quantum state is a quantum state constructed based on the characteristics of the ideal measurement point or typical fault characteristics; The quantum distance to a single candidate measurement point is 1 minus the square root of the modulus of the inner product of the reference quantum state and the measurement point quantum state; The overall quantum distance of the measurement point combination scheme is the sum of the quantum distances of each candidate measurement point, and the overall quantum compatibility is the sum or average of the quantum compatibility of each candidate measurement point; The formula for calculating the quantum coincidence degree of a single candidate measurement point is as follows:

[0007]

[0008] in, For reference quantum state, For the first A quantum state at a single measurement point For the first The degree of quantum coincidence between a single measurement point and the reference quantum state; The formula for calculating the quantum distance of a single candidate measurement point is:

[0009] in, For the first The quantum distance between a single measurement point and the reference quantum state. For reference quantum state, For the first A quantum state at a single measurement point.

[0010] Preferably, when the preset objective function for optimizing the measurement points is to minimize the quantum distance, after outputting candidate measurement point combination schemes based on the probability distribution, the method further includes: The step size parameter and vector angle parameter are determined based on the quantum distance. The step size parameter is a monotonically decreasing function of the quantum distance, which controls the magnitude of each adjustment of the measurement point position. The vector angle parameter is used to determine the direction of the measurement point position adjustment. Based on the step size parameter and vector angle parameter, the spatial coordinates of each measuring point in the candidate measuring point combination scheme are iteratively adjusted until the quantum distance is less than the preset threshold, and the adjusted measuring point position is determined as the final measuring point layout. The formula for calculating the step size parameter is as follows:

[0011] in, For the first The quantum distance between a single measurement point and the reference quantum state. This is the step scaling factor. The step size index; The formula for calculating the vector angle parameter is as follows:

[0012] in, For the first Vector angle parameters for a single measuring point For reference quantum state, For the first A quantum state at a single measurement point.

[0013] Preferably, the characteristic quantum states corresponding to the multidimensional eigenvectors of each candidate test point satisfy the following formula:

[0014] Let the total number of candidate measurement points be . , No. The multidimensional feature vectors of the candidate measurement points are , For the first Dimensional feature dimension This represents the total number of feature dimensions. For the first The candidate measurement points at the th Feature vectors of dimension 1 For the first The candidate measurement points at the th Feature vectors of dimension 1 For the first The characteristic quantum states of each candidate test point For the first The norm of the eigenvectors of the candidate test points, For the first The computational ground state carries the first... Dimensional feature information.

[0015] Preferably, based on the characteristic quantum states corresponding to each candidate test point, and utilizing the principle of quantum superposition, a superposition state of test point combination is established for the selected states of each candidate test point, including: Let the total number of candidate measurement points be Each candidate test point is assigned a qubit; Based on the characteristic quantum states corresponding to each candidate measurement point, a superposition state of measurement points consisting of M qubits is constructed using the principle of quantum superposition; wherein, each calculated ground state corresponds to one... A binary number, where each bit takes the value 0 or 1 to indicate whether the corresponding candidate measurement point is not selected or is selected, respectively; The calculation formula for the superposition state of the measurement point combination is as follows:

[0016] in, For a combination of M measuring points, the superposition state of the measuring point combinations is given. Let M be the computational ground state for M qubits.

[0017] Preferably, the multi-source sensor monitoring data of the power transformer at each candidate monitoring point is acquired, including: The data is obtained through simulation using a digital twin model of a power transformer or through historical measured data, wherein the multi-source sensor monitoring data includes at least two of the following: temperature data, vibration data, current data, partial discharge data, and online gas monitoring data. Multi-source sensor monitoring data are fused and processed to construct multi-dimensional feature vectors for each candidate measurement point, including: The monitoring data from multiple sources are processed by time alignment and interpolation, redundant data removal, feature extraction, and complementary fusion to obtain multidimensional feature vectors for each candidate measurement point. Feature extraction includes one or more of the following: statistical extraction, time-frequency feature extraction, and spatial feature extraction.

[0018] Based on the same inventive concept, this invention also provides a power transformer measurement point optimization system based on quantum computing, the system comprising: The multi-dimensional vector construction module is used to acquire multi-source sensor monitoring data of power transformers at each candidate measurement point, perform fusion processing on the multi-source sensor monitoring data, and construct multi-dimensional feature vectors for each candidate measurement point. The quantum state representation module is used to map the multidimensional feature vectors of each candidate measurement point to the corresponding feature quantum states using amplitude encoding, thereby obtaining a unified quantum state representation of multi-source high-dimensional features; The superposition state establishment module is used to establish a superposition state of the selected states of each candidate test point based on the characteristic quantum states corresponding to each candidate test point and by utilizing the principle of quantum superposition. The combination scheme output module is used to solve the evolution of the superposition state of the test point combination using a quantum approximation optimization algorithm. It obtains the probability distribution of the candidate test point combination with a preset optimization objective function as the optimization objective and outputs the candidate test point combination scheme according to the probability distribution.

[0019] Preferably, the combined solution output module is specifically used for: A quantum approximation optimization algorithm based on parameterized quantum gates is adopted to perform multiple rounds of evolution on the superposition state of the test point combination. The preset optimization objective function is minimized or maximized by iteratively optimizing the parameters of the parameterized quantum gates. The evolution process includes the alternating application of a first evolution operator generated by the cost Hamiltonian and a second evolution operator generated by the mixer Hamiltonian. The cost Hamiltonian corresponds to a preset optimization objective function, and the mixer Hamiltonian is used to drive the quantum state to traverse and search in the solution space. The superposition state of the evolved measurement point combination is sampled and measured to obtain the optimized probability distribution of the candidate measurement point combination.

[0020] Preferably, the preset optimization objective function that minimizes or maximizes includes: minimizing the overall quantum distance or maximizing the overall quantum coincidence. The quantum coincidence of a single candidate measurement point is the square of the modulus of the inner product of the reference quantum state and the measurement point quantum state; whereby the reference quantum state is a quantum state constructed based on the characteristics of the ideal measurement point or typical fault characteristics; The quantum distance to a single candidate measurement point is 1 minus the square root of the modulus of the inner product of the reference quantum state and the measurement point quantum state; The overall quantum distance of the measurement point combination scheme is the sum of the quantum distances of each candidate measurement point, and the overall quantum compatibility is the sum or average of the quantum compatibility of each candidate measurement point; The formula for calculating the quantum coincidence degree of a single candidate measurement point is as follows:

[0021]

[0022] in, For reference quantum state, For the first A quantum state at a single measurement point For the first The degree of quantum coincidence between a single measurement point and the reference quantum state; The formula for calculating the quantum distance of a single candidate measurement point is:

[0023] in, For the first The quantum distance between a single measurement point and the reference quantum state. For reference quantum state, For the first A quantum state at a single measurement point.

[0024] Preferably, when the preset objective function for optimizing the measurement points is to minimize the quantum distance, the system further includes a measurement point adjustment module, used for: The step size parameter and vector angle parameter are determined based on the quantum distance. The step size parameter is a monotonically decreasing function of the quantum distance, which controls the magnitude of each adjustment of the measurement point position. The vector angle parameter is used to determine the direction of the measurement point position adjustment. Based on the step size parameter and vector angle parameter, the spatial coordinates of each measuring point in the candidate measuring point combination scheme are iteratively adjusted until the quantum distance is less than the preset threshold, and the adjusted measuring point position is determined as the final measuring point layout. The formula for calculating the step size parameter is as follows:

[0025] in, For the first The quantum distance between a single measurement point and the reference quantum state. This is the step scaling factor. The step size index; The formula for calculating the vector angle parameter is as follows:

[0026] in, For the first Vector angle parameters for a single measuring point For reference quantum state, For the first A quantum state at a single measurement point.

[0027] Preferably, the characteristic quantum states corresponding to the multidimensional eigenvectors of each candidate test point satisfy the following formula:

[0028] Let the total number of candidate measurement points be . , No. The multidimensional feature vectors of the candidate measurement points are , For the first Dimensional feature dimension This represents the total number of feature dimensions. For the first The candidate measurement points at the th Feature vectors of dimension 1 For the first The candidate measurement points at the th Feature vectors of dimension 1 For the first The characteristic quantum states of each candidate test point For the first The norm of the eigenvectors of the candidate test points, For the first The computational ground state carries the first... Dimensional feature information.

[0029] Preferably, the superposition state establishment module is specifically used for: Let the total number of candidate measurement points be Each candidate test point is assigned a qubit; Based on the characteristic quantum states corresponding to each candidate measurement point, a superposition state of measurement points consisting of M qubits is constructed using the principle of quantum superposition; wherein, each calculated ground state corresponds to one... A binary number, where each bit takes the value 0 or 1 to indicate whether the corresponding candidate measurement point is not selected or is selected, respectively; The calculation formula for the superposition state of the measurement point combination is as follows:

[0030] in, For a combination of M measuring points, the superposition state of the measuring point combinations is given. Let M be the computational ground state for M qubits.

[0031] Preferably, the multidimensional vector construction module is specifically used for: The data is obtained through simulation using a digital twin model of a power transformer or through historical measured data, wherein the multi-source sensor monitoring data includes at least two of the following: temperature data, vibration data, current data, partial discharge data, and online gas monitoring data. Multi-source sensor monitoring data are fused and processed to construct multi-dimensional feature vectors for each candidate measurement point, including: The monitoring data from multiple sources are processed by time alignment and interpolation, redundant data removal, feature extraction, and complementary fusion to obtain multidimensional feature vectors for each candidate measurement point. Feature extraction includes one or more of the following: statistical extraction, time-frequency feature extraction, and spatial feature extraction.

[0032] Based on the same inventive concept, the present invention also provides an electronic device, comprising: at least one processor and a memory; wherein the memory and the processor are connected via a bus; The memory is used to store one or more programs; When the one or more programs are executed by the at least one processor, a quantum computing-based power transformer measurement point optimization method as described above is implemented.

[0033] Based on the same inventive concept, the present invention also provides a readable storage medium having an executable program stored thereon, which, when executed, implements the aforementioned method for optimizing power transformer measurement points based on quantum computing.

[0034] Compared with the closest existing technology, the present invention has the following beneficial effects: This invention provides a quantum computing-based method for optimizing measurement points of power transformers, comprising: acquiring multi-source sensor monitoring data of power transformers at each candidate measurement point; fusing the multi-source sensor monitoring data to construct multi-dimensional feature vectors for each candidate measurement point; using amplitude encoding to map the multi-dimensional feature vectors of each candidate measurement point to corresponding feature quantum states, thereby obtaining a unified quantum state representation of multi-source high-dimensional features; based on the feature quantum states corresponding to each candidate measurement point, using the principle of quantum superposition, establishing a superposition state of measurement point combinations for the selection states of each candidate measurement point; using a quantum approximation optimization algorithm to evolve and solve the superposition state of measurement point combinations, using a preset optimization objective function as the optimization objective, obtaining the probability distribution of candidate measurement point combinations, and outputting candidate measurement point combination schemes based on the probability distribution. This invention utilizes the principles of quantum computing, specifically by encoding the amplitude of the multidimensional eigenvectors of each candidate measurement point into characteristic quantum states. It then employs the principle of quantum superposition to construct a superposition state of measurement point combinations covering all selection schemes. Combined with a quantum approximation optimization algorithm, it performs evolutionary solutions to achieve parallel search and efficient global optimization of massive measurement point combinations. This significantly reduces the computational load of optimization, yields a better measurement point layout scheme, and thereby improves the coverage, sensitivity, and diagnostic accuracy of transformer condition parameter monitoring. Attached Figure Description

[0035] Figure 1 A flowchart illustrating the quantum computing-based power transformer measurement point optimization method provided by this invention; Figure 2 A schematic diagram of the quantum computing-based power transformer measurement point optimization method provided by the present invention; Figure 3 A flowchart illustrating a specific embodiment of the present invention; Figure 4 The structural diagram of the quantum computing-based power transformer measurement point optimization system provided by this invention is shown below. Figure 5 A schematic diagram of the structure of the quantum computing-based power transformer measurement point optimization system provided by this invention. Figure 6 A schematic diagram of the electronic device provided by the present invention. Detailed Implementation

[0036] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0037] Example 1: This invention provides a method for optimizing measurement points of power transformers based on quantum computing. Specifically, Figure 1 A flowchart illustrating the quantum computing-based power transformer measurement point optimization method provided in this embodiment of the invention is shown in the figure, including the following steps: S101: Acquire multi-source sensor monitoring data of the power transformer at each candidate measurement point, perform fusion processing on the multi-source sensor monitoring data, and construct a multi-dimensional feature vector for each candidate measurement point; S102: Amplitude coding is used to map the multidimensional feature vectors of each candidate measurement point to the corresponding feature quantum states, thereby obtaining a unified quantum state representation of multi-source high-dimensional features; S103: Based on the characteristic quantum states corresponding to each candidate test point, the quantum superposition principle is used to establish a superposition state of test point combination for the selected states of each candidate test point; S104: The quantum approximation optimization algorithm is used to solve the evolution of the superposition state of the test point combination. The preset optimization objective function is used as the optimization objective to obtain the probability distribution of the candidate test point combination and output the candidate test point combination scheme according to the probability distribution.

[0038] This invention utilizes the principles of quantum computing, specifically by encoding the amplitude of the multidimensional eigenvectors of each candidate measurement point into characteristic quantum states. It then employs the principle of quantum superposition to construct a superposition state of measurement point combinations covering all selection schemes. Combined with a quantum approximation optimization algorithm, it performs evolutionary solutions to achieve parallel search and efficient global optimization of massive measurement point combinations. This significantly reduces the computational load of optimization, yields a better measurement point layout scheme, and thereby improves the coverage, sensitivity, and diagnostic accuracy of transformer condition parameter monitoring.

[0039] The application of artificial intelligence technology has enabled the comprehensive monitoring and evaluation of the operating status of transformers, but it has also brought a large amount of real-time multidimensional data. Therefore, how to quickly identify abnormal data by fusing and analyzing multidimensional data, and then determine the location and cause of the fault, remains an unsolved problem.

[0040] This invention first selects multiple candidate measurement points on the surface of a power transformer and acquires multi-source sensor monitoring data at each candidate measurement point. The multi-source sensor monitoring data can be obtained through simulation using a digital twin model of the power transformer or through historical measured data. The multi-source sensor monitoring data includes, but is not limited to, at least two of the following: temperature data, vibration data, current data, partial discharge data, and online gas monitoring data. Each type of multi-source sensor monitoring data constitutes a one-dimensional feature dimension. If there are four types of multi-source sensor monitoring data, it corresponds to a four-dimensional feature.

[0041] Furthermore, the multi-source sensor monitoring data is fused to construct multi-dimensional feature vectors for each candidate monitoring point. Specifically, this includes time alignment and interpolation, redundant data removal, feature extraction, and complementary fusion processing of the multi-source sensor monitoring data to obtain multi-dimensional feature vectors for each candidate monitoring point. Feature extraction includes one or more of the following: statistical extraction (e.g., mean, variance, peak value), time-frequency feature extraction (e.g., spectral energy distribution), and spatial feature extraction. It is understood that the specific fusion processing method is related to the type of multi-source sensor monitoring data obtained, and is not limited here. Any feasible fusion processing method is within the scope of protection claimed in this application.

[0042] Specifically, let the total number of candidate measurement points be... , No. The multidimensional feature vectors of the candidate measurement points are , For the first Dimensional feature dimension This represents the total number of feature dimensions. For the first The candidate measurement points at the th Feature vectors of dimension 1 For the first The candidate measurement points at the th Feature vectors with a feature dimension.

[0043] After constructing the multidimensional feature vectors of each candidate test point, the multidimensional feature vectors of each candidate test point are mapped to the corresponding quantum states to obtain the unified quantum state representation of the multi-source high-dimensional features of each candidate test point.

[0044] The characteristic quantum states corresponding to the multidimensional eigenvectors of each candidate test point satisfy the following formula:

[0045] Let the total number of candidate measurement points be . , No. The multidimensional feature vectors of the candidate measurement points are , For the first Dimensional feature dimension This represents the total number of feature dimensions. For the first The candidate measurement points at the th Feature vectors of dimension 1 For the first The candidate measurement points at the th Feature vectors of dimension 1 For the first The characteristic quantum states of each candidate test point For the first The norm of the eigenvectors of the candidate test points, For the first The computational ground state carries the first... Dimensional feature information.

[0046] This mapping unifies sensor data of different dimensions and ranges into normalized feature quantum states in the same Hilbert space, realizing a unified quantum state representation of multi-source high-dimensional features, thereby achieving unified processing and comparability of cross-type sensor data.

[0047] After obtaining the unified quantum state representation of the multi-source high-dimensional features of each candidate test point, based on the feature quantum states corresponding to each candidate test point, the quantum superposition principle is used to establish a superposition state of test point combinations for the selected states of each candidate test point. Specifically, this includes: Let the total number of candidate measurement points be Each candidate test point is assigned a qubit; Based on the characteristic quantum states corresponding to each candidate test point, and utilizing the principle of quantum superposition, a superposition state of test point combinations consisting of M qubits is constructed. This involves establishing qubit superposition states for all combinations of candidate test points to form a quantum search space covering all test point selection schemes. Each computational ground state corresponds to a... A binary number, where each bit takes the value 0 or 1 to indicate whether the corresponding candidate measurement point is not selected or is selected, respectively; The formula for calculating the superposition state of measurement points, which is the initial space of a quantum state formed by combining M measurement points, is as follows:

[0048] in, For a combination of M measuring points, the superposition state of the measuring point combinations is given. The computational ground state for M qubits has coefficients. This is a normalization constant, ensuring that the initial probabilities of all ground states are equal. This superposition state achieves equality for all... Parallel representation of various measurement point selection schemes.

[0049] Furthermore, a quantum approximation optimization algorithm is employed to evolve and solve the superposition state of the test point combinations, using a pre-defined optimization objective function as the optimization target, to obtain the probability distribution of candidate test point combinations. Specifically, this includes: The quantum approximate optimization algorithm (QAOA) based on parameterized quantum gates is used to perform multiple rounds of evolution on the superposition state of the test point combination. The preset optimization objective function is minimized or maximized by iteratively optimizing the parameters of the parameterized quantum gate. The evolutionary process includes the alternating application of the cost Hamiltonian. The first evolution operator generated and Hamiltonian from the mixer The generated second evolution operator The cost Hamiltonian corresponds to the preset optimization objective function, and the mixer Hamiltonian is used to drive the quantum state to traverse and search in the solution space. The superposition state of the evolved measurement point combination is sampled and measured to obtain the optimized probability distribution of the candidate measurement point combination.

[0050] Among them, the first evolution operator In The cost evolution angle is the cost Hamiltonian. The corresponding evolution time parameters. The evolution intensity of the quantum state is controlled by the cost Hamiltonian, which is usually a real number. Its value is adjusted through iterative optimization so that the quantum state evolves in a direction with lower cost (i.e., a smaller value of the optimization objective function).

[0051] Second evolution operator In The mixing evolution angle is the mixer Hamiltonian. The corresponding evolution time parameters. The strength of the interaction between the mixer Hamiltonian and the quantum state was controlled by adjusting... It can increase the ability to explore quantum states and avoid getting trapped in local optima.

[0052] In some optional implementations, minimizing or maximizing a preset optimization objective function includes minimizing the overall quantum distance or maximizing the overall quantum coincidence. The quantum coincidence of a single candidate measurement point is the square of the modulus of the inner product of the reference quantum state and the measurement point quantum state; whereby the reference quantum state is a quantum state constructed based on the characteristics of the ideal measurement point or typical fault characteristics; The quantum distance to a single candidate measurement point is 1 minus the square root of the modulus of the inner product of the reference quantum state and the measurement point quantum state; The overall quantum distance of the measurement point combination scheme is the sum of the quantum distances of each candidate measurement point, and the overall quantum compatibility is the sum or average of the quantum compatibility of each candidate measurement point; The formula for calculating the quantum coincidence degree of a single candidate measurement point is as follows:

[0053]

[0054] in, For reference quantum state, For the first A quantum state at a single measurement point For the first The degree of quantum coincidence between a single measurement point and the reference quantum state; The formula for calculating the quantum distance of a single candidate measurement point is:

[0055] in, For the first The quantum distance between a single measurement point and the reference quantum state. For reference quantum state, For the first A quantum state at a single measurement point.

[0056] Preferably, when the preset objective function for optimizing the measurement points is to minimize the quantum distance, after outputting the candidate measurement point combination scheme according to the probability distribution, the method further includes: fine-tuning the spatial coordinates of each measurement point in the candidate measurement point combination scheme.

[0057] Specifically, this includes: determining the step size parameter and vector angle parameter based on the quantum distance. The step size parameter is a monotonically decreasing function of the quantum distance, controlling the magnitude of each adjustment of the measurement point position, and the vector angle parameter is used to determine the direction of the measurement point position adjustment. Based on the step size parameter and vector angle parameter, the spatial coordinates of each measuring point in the candidate measuring point combination scheme are iteratively adjusted until the quantum distance is less than the preset threshold, and the adjusted measuring point position is determined as the final measuring point layout. The formula for calculating the step size parameter is as follows:

[0058] in, For the first The quantum distance between a single measurement point and the reference quantum state. This is the step scaling factor. The step size index; The formula for calculating the vector angle parameter is as follows:

[0059] in, For the first Vector angle parameters for a single measuring point For reference quantum state, For the first A quantum state at a single measurement point.

[0060] In some optional implementations, the quantum optimization solution is mapped onto a digital twin model of the power transformer (including simulations of physical fields such as electric field, magnetic field, and temperature field) for simulation verification, and the quantum optimization parameters are adjusted based on the simulation results to achieve alternating iterations of quantum search and physical simulation.

[0061] In some optional implementations, a feature library is also included for long-term collection and management of feature quantum states, operating modes and typical fault features, so as to realize model self-learning and dynamic supplementation.

[0062] like Figure 2 The diagram illustrates the quantum computing-based power transformer measurement point optimization method provided by this invention. The specific process includes: obtaining the feature vector of the measurement point → quantum state → consistency degree → quantum distance → step size / direction → adjustment of the measurement point.

[0063] The following detailed description of the quantum computing-based power transformer measurement point optimization method provided by this invention, using a specific embodiment, illustrates this invention in detail. Figure 3 The diagram shown is a flowchart of this specific embodiment.

[0064] (1) Basic settings Taking a typical 110kV oil-immersed power transformer as an example, assume that the placement of four types of sensors (temperature, vibration, current, and voltage) needs to be optimized, corresponding to a 4-dimensional feature vector. Thirty-two candidate measurement points (numbered 1-32) are uniformly selected on the surface of the power transformer. At each candidate measurement point, temperature, vibration, current, and voltage data are obtained through digital twin simulation or historical measured data, forming a four-dimensional feature vector. For example, the feature vector of measurement point 1 is... .

[0065] (2) Amplitude coding The four-dimensional feature vector of each candidate measurement point is normalized. Taking measurement point 1 as an example, the norm is calculated as follows:

[0066] The normalized amplitude coefficients are obtained as follows:

[0067] Therefore, the characteristic quantum state of measurement point 1 is

[0068] Similarly, the same amplitude encoding was performed on the remaining 31 candidate measurement points, resulting in 32 characteristic quantum states. .

[0069] (3) Construct a superposition state of measurement points Let the total number of candidate measurement points be... Assign one qubit to each measurement point. Construct a 32-qubit uniform superposition state:

[0070] Each ground state in this superposition state (For example ) corresponds to a specific point selection scheme (1 means select, 0 means do not select).

[0071] Each ground state (e.g., a 32-bit binary number) only indicates which candidate measurement points have been selected, without distinguishing which type of sensor is installed at that location. The characteristic quantum state corresponding to each candidate measurement point location... The amplitude encoding already incorporates data from all types of sensors at that location (e.g., temperature, vibration, current, and voltage, totaling four dimensions). Therefore, when a location is selected (binary value 1), it means that data from all four types of sensors can be acquired at that location simultaneously, or that a composite sensor capable of acquiring these four physical quantities can be installed at that location.

[0072] A uniform superposition state is a mathematical formula that describes a parallel representation of all possible measurement point selection schemes. Specifically, It is the computational ground state of M qubits, each Corresponding to a unique M-bit binary number (e.g. hour, This indicates that the second measuring point is selected, while the first and third measuring points are not selected. (Coefficient) Ensure that the probability amplitude of all ground states is equal, that is, initially the probability of each choice is the same (all are equal). In quantum circuits, this superposition state is achieved by initializing each qubit to... This is obtained by taking the tensor product over all bits. This represents the initial state of the algorithm, where there is no bias and all possible point selections are treated equally. This is defined in... The normalized state vector in the Wid Hilbert space has a sum of squares of 1 for its components.

[0073] (4) Set the reference quantum state and the objective function Based on the ideal operating state or typical fault characteristics of the equipment, a reference feature vector is set, such as ideal temperature 80℃, ideal vibration 0.05g, ideal current 340A, and ideal voltage 110kV. This vector is then encoded into a reference quantum state using the same method. .

[0074] Define the quantum distance at a single measurement point, and the overall quantum distance as the sum of the quantum distances of all selected measurement points in the scheme. The optimization objective is to minimize the overall quantum distance.

[0075] (5) Quantum approximate optimization evolution solution A quantum approximation optimization algorithm based on parameterized quantum gates is employed to perform multi-round evolution of the superposition state. Each round alternates between the cost Hamiltonian (corresponding to the global quantum distance) and the mixer Hamiltonian. The parameters of the parameterized quantum gates (e.g., ...) are iteratively adjusted using a classical optimizer. This allows the evolved quantum state to collapse into the ground state with a high probability during measurement, where the overall quantum distance is smaller.

[0076] Assuming that after several rounds of influence, the measurement statistics show that: [The following is a separate, unrelated sentence: Selected measurement points] If the probability of a given point is the highest (e.g., over 80%), then output these 5 measurement points as candidate measurement point combination schemes.

[0077] (6) Spatial coordinate fine-tuning For each selected measurement point (e.g., measurement point 3), calculate the quantum distance. ,set up rice, Then step size Meters. Simultaneously calculate the vector angle. The measurement point coordinates are moved along the predetermined gradient direction, the feature data at that location is reacquired, and the quantum state is updated. This process is repeated iteratively until... Finally, the coordinates of five finely adjusted measurement points were obtained.

[0078] (7) Deployment and monitoring The final list of monitoring points is output to the engineering staff, who then install sensors at the corresponding locations on the transformer to achieve long-term online monitoring.

[0079] This invention aims to solve the technical problems in existing technologies, such as the reliance on experience for measurement point layout, insufficient monitoring coverage, and limited efficiency of optimization algorithms. Specifically, it includes: (1) The amount of multi-source heterogeneous sensing data is huge, and the features are difficult to extract quickly, resulting in insufficient timeliness of transformer condition diagnosis; (2) The layout of measuring points relies on experience and deterministic simulation, making it difficult to obtain the optimal position in complex multiphysics coupling environments; (3) Traditional optimization algorithms are prone to getting stuck in local optima in high-dimensional data spaces, making it difficult to achieve efficient search; (4) The characteristic correlation between complex state parameters cannot be effectively quantified, which affects the accuracy of anomaly identification and fault location.

[0080] The technical effects achieved by this invention are as follows: (1) Achieve rapid global optimization of measurement point layout in high-dimensional complex scenes, significantly reducing the amount of optimization computation; (2) Improve the coverage and sensitivity of the transformer multi-parameter monitoring system so that key operating characteristics can be captured more comprehensively and accurately; (3) Improve the ability to identify anomalies in multidimensional monitoring data and the accuracy of fault location, and enhance the reliability of transformer condition assessment and active defense; (4) Provide efficient and scalable measurement point optimization methods for intelligent substation operation and maintenance, and improve the overall performance of the monitoring system.

[0081] This invention introduces the principles of quantum computing into the field of power transformer measurement point optimization, achieving a breakthrough in traditional measurement point layout methods and providing a highly efficient new technical path for intelligent monitoring of complex equipment.

[0082] This invention utilizes quantum computing principles such as quantum state encoding, quantum distance calculation, and quantum state similarity discrimination to accelerate key processes such as multidimensional feature matching, measurement point adjustment, and optimal solution search, thereby improving the efficiency and accuracy of measurement point optimization.

[0083] This invention combines a quantum computing-based method for optimizing power transformer state parameter measurement points with current power equipment condition monitoring technology. This enables real-time, effective, accurate, and comprehensive condition monitoring of power transformers, allowing for timely detection of potential risks and ensuring their normal operation. By introducing quantum computing principles (such as quantum superposition, quantum parallelism, and quantum approximation optimization algorithms), the invention combines multi-physics characteristic quantization modeling of power transformers, measurement point selection optimization, and quantum combinatorial optimization algorithms to efficiently solve for the measurement point layout. This achieves intelligent, global, and efficient optimization of the power transformer state parameter measurement point layout, significantly improving optimization speed and enhancing global search capabilities. Ultimately, this yields a superior state parameter measurement point configuration scheme, improving the overall performance of the transformer condition sensing system.

[0084] This invention falls under the technical categories of power system state perception, intelligent monitoring, and quantum algorithm-assisted optimization. It can be applied to the intelligent operation and maintenance of power system substations, providing a scientific measurement point layout scheme for transformer condition monitoring systems. Combined with a digital twin platform, it enables automated and intelligent measurement point optimization, reducing reliance on manual experience and improving monitoring system performance.

[0085] The beneficial effects of this invention are: (1) Achieve global optimal or near-global optimal solution for the layout of measuring points. By leveraging the properties of quantum superposition and quantum parallelism, a large number of measurement point combinations are searched in parallel. Compared with traditional exhaustive search, greedy algorithms, or classical heuristic algorithms, it has a larger optimization space and stronger search capabilities, making the measurement point selection results closer to the global optimum.

[0086] (2) Significantly improves the computational efficiency of measurement point optimization The quantum approximation optimization algorithm is used to solve the problem of measuring point layout. It can achieve optimization solutions with low computational complexity under complex multi-constraint conditions. Compared with traditional algorithms, it can significantly reduce the computation time and is particularly suitable for scenarios with multiple state parameters and a large number of candidate measuring points.

[0087] (3) Improve the effective coverage of transformer status sensing by the deployment of measuring points. By simulating the multi-physics field of power transformers using a digital twin model, the task of selecting measurement points is linked to the sensitivity of key features, enabling the final point layout scheme to more accurately cover fault-sensitive areas and improve the quality of state parameter acquisition and monitoring integrity.

[0088] (4) Improve the accuracy of abnormal detection and fault location of power transformers. The optimized measurement point layout reduces data redundancy and increases information contribution, which is beneficial for improving the performance of subsequent analysis models such as anomaly identification, feature fusion, and intelligent diagnosis, thereby improving the accuracy and positioning precision of transformer fault identification.

[0089] (5) Reduce sensor deployment costs and maintenance complexity By using quantum optimization techniques to reduce the blind increase in the number of measurement points and achieve the optimal balance between the number of measurement points and the monitoring effect, the cost of sensor deployment, wiring difficulty and subsequent maintenance can be effectively reduced.

[0090] (6) It has high scalability and engineering adaptability. The optimization framework of this invention is based on the principle of quantum computing and can be adapted to different types of transformer structures, different state parameters and different monitoring requirements, and has good scalability. At the same time, the quantum algorithm can be replaced by simulated quantum computing or actual quantum hardware operation, which facilitates engineering deployment.

[0091] In summary, this invention achieves high efficiency, intelligence, and reliability in optimizing power transformer state parameter measurement points through a technical approach combining quantum computing and digital twins. It effectively improves the problems of unreasonable measurement point distribution, low optimization efficiency, and insufficient monitoring coverage in existing technologies, and has significant engineering application value.

[0092] This invention, based on quantum computing principles and combining the consistency between feature vectors, optimizes the measurement points by adjusting the measurement points. Building upon a digital twin model of a transformer, it optimizes the location of sensor devices at measurement points on the transformer. This method can help quickly and accurately identify abnormal states in power equipment, thereby determining the location and cause of faults, and ultimately enabling early warning of power equipment risks and providing maintenance recommendations.

[0093] Example 2: Based on the same inventive concept, this invention also provides a power transformer measurement point optimization system based on quantum computing, the structure of which is as follows: Figure 4 As shown, the system includes: The multi-dimensional vector construction module 401 is used to acquire multi-source sensor monitoring data of power transformers at each candidate measurement point, perform fusion processing on the multi-source sensor monitoring data, and construct multi-dimensional feature vectors for each candidate measurement point. The quantum state representation module 402 is used to map the multidimensional feature vectors of each candidate measurement point to the corresponding feature quantum states by using amplitude encoding, so as to obtain a unified quantum state representation of multi-source high-dimensional features; The superposition state establishment module 403 is used to establish a superposition state of the selected states of each candidate test point based on the characteristic quantum states corresponding to each candidate test point and by utilizing the principle of quantum superposition. The combination scheme output module 404 is used to use a quantum approximation optimization algorithm to evolve and solve the superposition state of the test point combination, with a preset optimization objective function as the optimization objective, to obtain the probability distribution of the candidate test point combination, and output the candidate test point combination scheme according to the probability distribution.

[0094] Preferably, the combined solution output module is specifically used for: A quantum approximation optimization algorithm based on parameterized quantum gates is adopted to perform multiple rounds of evolution on the superposition state of the test point combination. The preset optimization objective function is minimized or maximized by iteratively optimizing the parameters of the parameterized quantum gates. The evolution process includes the alternating application of a first evolution operator generated by the cost Hamiltonian and a second evolution operator generated by the mixer Hamiltonian. The cost Hamiltonian corresponds to a preset optimization objective function, and the mixer Hamiltonian is used to drive the quantum state to traverse and search in the solution space. The superposition state of the evolved measurement point combination is sampled and measured to obtain the optimized probability distribution of the candidate measurement point combination.

[0095] Preferably, the preset optimization objective function that minimizes or maximizes includes: minimizing the overall quantum distance or maximizing the overall quantum coincidence. The quantum coincidence of a single candidate measurement point is the square of the modulus of the inner product of the reference quantum state and the measurement point quantum state; whereby the reference quantum state is a quantum state constructed based on the characteristics of the ideal measurement point or typical fault characteristics; The quantum distance to a single candidate measurement point is 1 minus the square root of the modulus of the inner product of the reference quantum state and the measurement point quantum state; The overall quantum distance of the measurement point combination scheme is the sum of the quantum distances of each candidate measurement point, and the overall quantum compatibility is the sum or average of the quantum compatibility of each candidate measurement point; The formula for calculating the quantum coincidence degree of a single candidate measurement point is as follows:

[0096]

[0097] in, For reference quantum state, For the first A quantum state at a single measurement point For the first The degree of quantum coincidence between a single measurement point and the reference quantum state; The formula for calculating the quantum distance of a single candidate measurement point is:

[0098] in, For the first The quantum distance between a single measurement point and the reference quantum state. For reference quantum state, For the first A quantum state at a single measurement point.

[0099] Preferably, when the preset objective function for optimizing the measurement points is to minimize the quantum distance, the system further includes a measurement point adjustment module, used for: The step size parameter and vector angle parameter are determined based on the quantum distance. The step size parameter is a monotonically decreasing function of the quantum distance, which controls the magnitude of each adjustment of the measurement point position. The vector angle parameter is used to determine the direction of the measurement point position adjustment. Based on the step size parameter and vector angle parameter, the spatial coordinates of each measuring point in the candidate measuring point combination scheme are iteratively adjusted until the quantum distance is less than the preset threshold, and the adjusted measuring point position is determined as the final measuring point layout. The formula for calculating the step size parameter is as follows:

[0100] in, For the first The quantum distance between a single measurement point and the reference quantum state. This is the step scaling factor. The step size index; The formula for calculating the vector angle parameter is as follows:

[0101] in, For the first Vector angle parameters for a single measuring point For reference quantum state, For the first A quantum state at a single measurement point.

[0102] Preferably, the characteristic quantum states corresponding to the multidimensional eigenvectors of each candidate test point satisfy the following formula:

[0103] Let the total number of candidate measurement points be . , No. The multidimensional feature vectors of the candidate measurement points are , For the first Dimensional feature dimension This represents the total number of feature dimensions. For the first The candidate measurement points at the th Feature vectors of dimension 1 For the first The candidate measurement points at the th Feature vectors of dimension 1 For the first The characteristic quantum states of each candidate test point For the first The norm of the eigenvectors of the candidate test points, For the first The computational ground state carries the first... Dimensional feature information.

[0104] Preferably, the superposition state establishment module is specifically used for: Let the total number of candidate measurement points be Each candidate test point is assigned a qubit; Based on the characteristic quantum states corresponding to each candidate measurement point, a superposition state of measurement points consisting of M qubits is constructed using the principle of quantum superposition; wherein, each calculated ground state corresponds to one... A binary number, where each bit takes the value 0 or 1 to indicate whether the corresponding candidate measurement point is not selected or is selected, respectively; The calculation formula for the superposition state of the measurement point combination is as follows:

[0105] in, For a combination of M measuring points, the superposition state of the measuring point combinations is given. Let M be the computational ground state for M qubits.

[0106] Preferably, the multidimensional vector construction module is specifically used for: The data is obtained through simulation using a digital twin model of a power transformer or through historical measured data, wherein the multi-source sensor monitoring data includes at least two of the following: temperature data, vibration data, current data, partial discharge data, and online gas monitoring data. Multi-source sensor monitoring data are fused and processed to construct multi-dimensional feature vectors for each candidate measurement point, including: The monitoring data from multiple sources are processed by time alignment and interpolation, redundant data removal, feature extraction, and complementary fusion to obtain multidimensional feature vectors for each candidate measurement point. Feature extraction includes one or more of the following: statistical extraction, time-frequency feature extraction, and spatial feature extraction.

[0107] like Figure 5 The diagram shown is a structural schematic of the power transformer measurement point optimization system based on quantum computing provided by the present invention.

[0108] The power transformer status characterization module is used to acquire transformer electrical performance and operating status data; it receives multi-source sensing data periodically collected by sensors, such as temperature, vibration, current, voltage, partial discharge, and gas online monitoring data.

[0109] The data fusion module performs fusion processing on multi-source sensing data from different channels and sampling rates, including: redundant data removal, time alignment and interpolation, feature extraction (based on statistics, time-frequency features, spatial features, etc.), complementary fusion, and construction of multi-dimensional feature vectors.

[0110] The feature matching module uses quantum state encoding to map different feature vectors into quantum states, achieving a unified representation of multi-source high-dimensional features. It uses quantum state similarity (quantum distance) to measure the degree of matching between features, enabling: rapid comparison of features from multiple samples, rapid screening of abnormal features, accelerated quantum search for optimal measurement point directions, and adjustment of measurement point positions based on analysis results.

[0111] The power transformer simulation module, based on the digital twin model of the power transformer, maps the quantum matching results onto the equipment structure and performs: measurement point position adjustment, simulation verification (electric field, magnetic field, temperature field, etc.), and receives feedback data from the measurement point optimization module, realizing alternating iteration of quantum search and physical simulation.

[0112] The measurement point optimization module determines the step size and adjustment direction based on the quantum measurement results when the quantum similarity or quantum distance exceeds the set threshold, and outputs the optimized data for the next round to the power transformer simulation module.

[0113] The feature library module is used for long-term collection and management of quantum state features, operating modes, typical fault features, etc., to achieve model self-learning and dynamic supplementation.

[0114] By combining the above-mentioned optimization method for power transformer state parameter measurement points based on quantum computing principles with current power equipment condition monitoring technology, real-time, effective, accurate, and comprehensive condition monitoring of power transformers can be achieved, enabling timely detection of potential risks in power transformers and ensuring their normal operation.

[0115] Example 3: Based on the same inventive concept, such as Figure 6 As shown, the present invention also provides an electronic device, which may be a computer device, a microcontroller device, a smart mobile device, etc. The electronic device in this embodiment may include a processor, a memory, a transceiver component, etc. The memory, processor, and transceiver component are connected via a bus; the memory can be used to store executable programs, and an exemplary executable program may include instructions; the processor is used to execute the instructions stored in the memory. The memory can also be used to store data, which can be accessed and / or modified when instructions are executed.

[0116] The processor may be a Central Processing Unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing and control core of the terminal, and it is suitable for implementing one or more instructions. Specifically, it is suitable for loading and executing one or more instructions in a readable storage medium to implement the corresponding method flow or corresponding function, so as to implement the steps of the quantum computing-based power transformer measurement point optimization method in the above embodiments.

[0117] Example 4: Based on the same inventive concept, this invention also provides a readable storage medium, specifically an electronic device readable storage medium (Memory). This readable storage medium is a memory device within an electronic device used to store programs and data. It is understood that the readable storage medium here can include both the built-in storage medium within the electronic device and extended storage media supported by the electronic device. The storage medium provides storage space, which stores the terminal's operating system. Furthermore, this storage space also stores one or more instructions suitable for loading and execution by a processor. These instructions can be one or more executable programs (including program code). It should be noted that the storage medium here can be high-speed RAM or non-volatile memory, such as at least one disk storage device. The processor can load and execute one or more instructions stored in the storage medium to implement the steps of the quantum computing-based power transformer measurement point optimization method described in the above embodiments.

[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 device that provides the functions specified 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 instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1The 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 The steps of the function specified in one or more boxes.

[0122] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit its scope of protection. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that after reading the present invention, they can still make various changes, modifications or equivalent substitutions to the specific implementation of the application, but these changes, modifications or equivalent substitutions are all within the scope of protection of the claims pending approval.

Claims

1. A method for optimizing measurement points of a power transformer based on quantum computing, characterized in that, include: Acquire multi-source sensor monitoring data of power transformer at each candidate measurement point, perform fusion processing on the multi-source sensor monitoring data, and construct a multi-dimensional feature vector for each candidate measurement point; Amplitude coding is used to map the multidimensional feature vectors of each candidate measurement point to the corresponding feature quantum states, thereby obtaining a unified quantum state representation of multi-source high-dimensional features; Based on the characteristic quantum states corresponding to each candidate test point, the quantum superposition principle is used to establish a superposition state of test point combination for the selected states of each candidate test point; A quantum approximation optimization algorithm is used to evolve and solve the superposition state of the test point combination. The preset optimization objective function is used as the optimization objective to obtain the probability distribution of the candidate test point combination, and the candidate test point combination scheme is output according to the probability distribution.

2. The method according to claim 1, characterized in that, The process employs a quantum approximation optimization algorithm to evolve and solve the superposition state of the test point combinations, using a preset optimization objective function as the optimization target, to obtain the probability distribution of candidate test point combinations, including: A quantum approximation optimization algorithm based on parameterized quantum gates is used to perform multiple rounds of evolution on the superposition state of the test point combination, and the preset optimization objective function is minimized or maximized by iteratively optimizing the parameters of the parameterized quantum gates. The evolution process includes the alternating application of a first evolution operator generated by a cost Hamiltonian and a second evolution operator generated by a mixer Hamiltonian, wherein the cost Hamiltonian corresponds to the preset optimization objective function, and the mixer Hamiltonian is used to drive the quantum state to perform traversal search in the solution space. The evolved superposition state of the measurement point combination is sampled and measured to obtain the optimized probability distribution of the candidate measurement point combination.

3. The method according to claim 2, characterized in that, The preset optimization objective function that minimizes or maximizes includes: minimizing the overall quantum distance or maximizing the overall quantum coincidence. The quantum coincidence degree of a single candidate measurement point is the squared modulus of the inner product of the reference quantum state and the measurement point quantum state; wherein, the reference quantum state is a quantum state constructed based on the characteristics of the ideal measurement point or typical fault characteristics; The quantum distance to a single candidate measurement point is 1 minus the square root of the modulus of the inner product of the reference quantum state and the measurement point quantum state; The overall quantum distance of the measurement point combination scheme is the sum of the quantum distances of each candidate measurement point, and the overall quantum compatibility is the sum or average of the quantum compatibility of each candidate measurement point; The formula for calculating the quantum coincidence degree of a single candidate measurement point is as follows: in, For reference quantum state, For the first A quantum state at a single measurement point For the first The degree of quantum coincidence between a single measurement point and the reference quantum state; The formula for calculating the quantum distance of a single candidate measurement point is: in, For the first The quantum distance between a single measurement point and the reference quantum state. For reference quantum state, For the first A quantum state at a single measurement point.

4. The method according to claim 3, characterized in that, When the preset objective function for optimizing measurement points is to minimize the quantum distance, after outputting candidate measurement point combination schemes according to the probability distribution, the method further includes: The step size parameter and vector angle parameter are determined based on the quantum distance. The step size parameter is a monotonically decreasing function with respect to the quantum distance, which controls the magnitude of each adjustment of the measurement point position. The vector angle parameter is used to determine the direction of the measurement point position adjustment. Based on the step size parameter and vector angle parameter, the spatial coordinates of each measuring point in the candidate measuring point combination scheme are iteratively adjusted until the quantum distance is less than a preset threshold, and the adjusted measuring point position is determined as the final measuring point layout. The formula for calculating the step size parameter is as follows: in, For the first The quantum distance between a single measurement point and the reference quantum state. This is the step scaling factor. The step size index; The formula for calculating the vector angle parameter is as follows: in, For the first Vector angle parameters for a single measuring point For reference quantum state, For the first A quantum state at a single measurement point.

5. The method according to claim 1, characterized in that, The characteristic quantum states corresponding to the multidimensional eigenvectors of each candidate test point satisfy the following formula: Let the total number of candidate measurement points be . , No. The multidimensional feature vectors of the candidate measurement points are , For the first Dimensional feature dimension This represents the total number of feature dimensions. For the first The candidate measurement points at the th Feature vectors of dimension 1 For the first The candidate measurement points at the th Feature vectors of dimension 1 For the first The characteristic quantum states of each candidate test point For the first The norm of the eigenvectors of the candidate test points, For the first The computational ground state carries the first... Dimensional feature information.

6. The method according to claim 1, characterized in that, The step of establishing a superposition state of combined test points based on the characteristic quantum states corresponding to each candidate test point and utilizing the principle of quantum superposition includes: Let the total number of candidate measurement points be Each candidate test point is assigned a qubit; Based on the characteristic quantum states corresponding to each candidate measurement point, a superposition state of measurement points consisting of M qubits is constructed using the principle of quantum superposition; wherein, each calculated ground state corresponds to one... A binary number, wherein each bit of the binary number is 0 or 1, indicating that the corresponding candidate test point is not selected or is selected respectively; The calculation formula for the superposition state of the measurement point combination is as follows: in, For a combination of M measuring points, the superposition state of the measuring point combinations is given. Let M be the computational ground state for M qubits.

7. The method according to claim 1, characterized in that, The acquisition of multi-source sensor monitoring data of the power transformer at each candidate measurement point includes: The data is obtained through simulation using a digital twin model of a power transformer or through historical measured data, wherein the multi-source sensor monitoring data includes at least two of the following: temperature data, vibration data, current data, partial discharge data, and online gas monitoring data. The process of fusing the multi-source sensor monitoring data to construct a multi-dimensional feature vector for each candidate measurement point includes: The monitoring data from the multi-source sensors are processed by time alignment and interpolation, redundant data removal, feature extraction, and complementary fusion to obtain multi-dimensional feature vectors for each candidate measurement point; wherein, the feature extraction includes one or more of statistical extraction, time-frequency feature extraction, and spatial feature extraction.

8. A power transformer measurement point optimization system based on quantum computing, characterized in that, include: A multi-dimensional vector construction module is used to acquire multi-source sensor monitoring data of power transformers at each candidate measurement point, perform fusion processing on the multi-source sensor monitoring data, and construct multi-dimensional feature vectors for each candidate measurement point. The quantum state representation module is used to map the multidimensional feature vectors of each candidate measurement point to the corresponding feature quantum states using amplitude encoding, thereby obtaining a unified quantum state representation of multi-source high-dimensional features; The superposition state establishment module is used to establish a superposition state of the selected states of each candidate test point based on the characteristic quantum states corresponding to each candidate test point and by utilizing the principle of quantum superposition. The combination scheme output module is used to use a quantum approximation optimization algorithm to evolve and solve the superposition state of the test point combination, with a preset optimization objective function as the optimization objective, to obtain the probability distribution of the candidate test point combination, and output the candidate test point combination scheme according to the probability distribution.

9. An electronic device, characterized in that, include: At least one processor and memory; The memory and processor are connected via a bus; The memory is used to store one or more programs; When the one or more programs are executed by the at least one processor, the quantum computing-based power transformer measurement point optimization method as described in any one of claims 1 to 7 is implemented.

10. A readable storage medium, characterized in that, It contains an executable program, which, when executed, implements the quantum computing-based power transformer measurement point optimization method as described in any one of claims 1 to 7.

Images