Cvt error evaluation method, device, error evaluation apparatus, and storage medium

Abstract

The present application relates to a kind of CVT error evaluation method, device, error evaluation equipment and storage medium, belong to mutual inductor error identification technical field, wherein, the CVT error evaluation method includes: based on the voltage proportionality coefficient of multiple CVTs, first data set is restored to primary voltage level and obtains second data set, and difference matrix is constructed based on second data set;Based on difference matrix, likelihood function is constructed, and likelihood function is converted into posterior probability density function based on Bayes formula;From second data set, the data corresponding to the data of the extraction of the preset number of CVT is constructed third data set, and prior probability density is determined based on third data set;Based on the multi-chain MCMC algorithm of fusion plant rhizome growth and posterior probability density function, when the posterior probability density maximum, the error value corresponding to multiple CVTs is determined.The present application effectively guarantees the accuracy of CVT error evaluation, also improves the efficiency of error evaluation.

Description

Technical Field

[0001] This invention relates to the field of transformer error identification technology, and in particular to a CVT error assessment method, apparatus, error assessment device and storage medium. Background Technology

[0002] Capacitor voltage transformers (CVTs) are important measuring devices in power systems, used to convert high voltage to low voltage for energy metering and protection control.

[0003] To ensure the accuracy of electricity metering, the error of the CVT (current transformer) must be kept within specified limits. According to the verification regulations for power transformers, periodic power outages are required to verify their accuracy class against standards. However, this traditional power outage verification method has many problems, such as difficulties in coordinating power outages, numerous operating devices, and heavy on-site workload. This can lead to the failure to promptly detect potentially inaccurate CVTs, thus affecting the accuracy of electricity trade settlement.

[0004] Existing CVT error assessment methods have high requirements for data and operating conditions, resulting in poor accuracy. Therefore, improving the accuracy of CVT error assessment has become an urgent technical problem to be solved. Summary of the Invention

[0005] In view of this, it is necessary to provide a CVT error assessment method, apparatus, error assessment device and storage medium to solve the problem of low accuracy of existing CVT error assessment schemes.

[0006] To address the aforementioned problems, in a first aspect, the present invention provides a CVT error assessment method, comprising:

[0007] The sampled amplitudes of multiple CVTs at multiple measurement times are obtained as the first dataset. The first dataset is restored to the primary voltage level based on the voltage scaling factor of multiple CVTs to obtain the second dataset. A difference matrix is ​​constructed based on the second dataset.

[0008] A likelihood function is constructed based on the difference matrix, and the likelihood function is transformed into a posterior probability density function based on Bayes' theorem.

[0009] A third dataset is constructed by extracting a predetermined number of data corresponding to CVTs from the second dataset, and the prior probability density is determined based on the third dataset.

[0010] The initial population is generated based on the prior probability density as the starting point of multiple Markov chains. Based on the multi-chain MCMC algorithm that integrates plant root and stem growth and the posterior probability density function, the error values ​​corresponding to multiple CVTs when the posterior probability density is maximized are determined.

[0011] The construction of the dissimilarity matrix based on the second dataset includes:

[0012] Determine the mean vector corresponding to the second dataset, where each element in the mean vector corresponds to the mean of the primary voltage level for each CVT.

[0013] A difference matrix is ​​constructed based on the difference between every two elements in the mean vector.

[0014] In one possible implementation, constructing the likelihood function based on the difference matrix includes:

[0015] Based on the difference matrix, a likelihood function is constructed for the mean of the primary voltage level corresponding to multiple CVTs, given that the mean of the difference between every two CVTs is known, where the mean of the difference between every two CVTs follows a normal distribution.

[0016] In one possible implementation, determining the prior probability density based on a third dataset includes:

[0017] Singular value decomposition is performed on the covariance matrix corresponding to the third dataset, and the mean of the change is determined based on the results of the singular value decomposition.

[0018] The initial value of the change is determined based on the third dataset, and the prior probability density is determined based on the initial value of the change and the mean of the change.

[0019] In one possible implementation, the initial population generated based on prior probability density serves as the starting point for multiple Markov chains. Based on a multi-chain MCMC algorithm that integrates plant root and stem growth and a posterior probability density function, the error values ​​corresponding to multiple CVTs when the posterior probability density is maximized are determined, including:

[0020] An initial population is generated based on prior probability density as the starting point for multiple Markov chains, and fitness values ​​are determined based on the fitness function.

[0021] Individual mutation is performed based on fitness value and plant root and stem growth algorithm. The crossover probability factor is used to determine whether to accept mutated individuals, and the acceptance probability is used to determine whether to accept the mutation crossover operation, until multiple Markov chains converge.

[0022] In one possible implementation, the fitness function is determined based on the posterior probability density function.

[0023] In one possible implementation, the individual variation based on fitness values ​​and plant root and stem growth algorithms includes:

[0024] Using each Markov chain as the primary root, a mathematical model of fiber root growth is constructed based on fitness values, and individual variation is achieved through nutrient absorption and diffusion growth strategies.

[0025] On the other hand, the present invention also provides a CVT error assessment device, comprising:

[0026] The acquisition module is used to acquire the sampled amplitude values ​​of multiple CVTs at multiple measurement times as the first dataset, restore the first dataset to the primary voltage level based on the voltage proportional coefficient of multiple CVTs to obtain the second dataset, and construct a difference matrix based on the second dataset.

[0027] The transformation module is used to construct a likelihood function based on the difference matrix and convert the likelihood function into a posterior probability density function based on Bayes' theorem.

[0028] The first determining module is used to extract a preset number of data corresponding to CVT from the second dataset to construct a third dataset, and to determine the prior probability density based on the third dataset.

[0029] The second determination module is used to generate an initial population based on the prior probability density as the starting point of multiple Markov chains, and to determine the error values ​​corresponding to multiple CVTs when the posterior probability density is maximized based on the multi-chain MCMC algorithm that integrates plant root and stem growth and the posterior probability density function.

[0030] The construction of the dissimilarity matrix based on the second dataset includes:

[0031] Determine the mean vector corresponding to the second dataset, where each element in the mean vector corresponds to the mean of the primary voltage level for each CVT.

[0032] A difference matrix is ​​constructed based on the difference between every two elements in the mean vector.

[0033] Secondly, the present invention also provides an error assessment device, including a memory and a processor, wherein,

[0034] The memory is used to store programs;

[0035] The processor, coupled to the memory, is used to execute the program stored in the memory to implement the steps in the CVT error evaluation method described in any of the above implementations.

[0036] Thirdly, the present invention also provides a computer-readable storage medium for storing a computer-readable program or instructions, which, when executed by a processor, can implement the steps in the CVT error evaluation method described in any of the above implementations.

[0037] The beneficial effects of this invention are as follows: The CVT error assessment method, apparatus, error assessment device, and storage medium provided by this invention first determine the error relationship of multiple CVTs under time-in-phase conditions by acquiring the sampling amplitude of multiple CVTs at multiple measurement times. Then, a likelihood function and a posterior probability density function are constructed based on the error relationship of multiple CVTs under time-in-phase conditions. Next, the errors of CVTs at different measurement times are fused into the error assessment process to improve the accuracy of error assessment. Finally, the final error assessment result is determined by fusing the multi-chain MCMC of plant root and stem growth. This effectively ensures the accuracy of CVT error assessment and also improves the efficiency of error assessment. Attached Figure Description

[0038] Figure 1 This is a schematic flowchart of an embodiment of the CVT error evaluation method provided by the present invention;

[0039] Figure 2 This is a schematic flowchart of an embodiment of the CVT error evaluation process provided by the present invention;

[0040] Figure 3 A schematic diagram of an embodiment of the topology of the CVT error assessment and verification system provided by the present invention;

[0041] Figure 4 A schematic diagram of an embodiment of the CVT evaluation ratio error change during the verification process provided by the present invention;

[0042] Figure 5 A schematic diagram of an embodiment of the CVT error evaluation device provided by the present invention;

[0043] Figure 6 A schematic diagram of an embodiment of the error assessment device provided by the present invention. Detailed Implementation

[0044] 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 a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0045] In the description of the embodiments of the present invention, unless otherwise stated, "multiple" means two or more. "And / or" describes the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can represent three situations: A exists alone, A and B exist simultaneously, and B exists alone.

[0046] The terms "first," "second," etc., used in the embodiments of this invention are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, a technical feature defined with "first" or "second" may explicitly or implicitly include at least one of that feature.

[0047] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0048] This invention provides a CVT error assessment method, apparatus, error assessment device, and storage medium, which are described below.

[0049] Figure 1 This is a schematic flowchart of an embodiment of the CVT error evaluation method provided by the present invention, as shown below. Figure 1 As shown, the CVT error assessment methods include:

[0050] S101. Obtain the sampling amplitude of multiple CVTs at multiple measurement times as the first dataset. Based on the voltage scaling factor of multiple CVTs, restore the first dataset to the primary voltage level to obtain the second dataset. Construct a difference matrix based on the second dataset.

[0051] It should be noted that the CVT error assessment method provided by this invention can be applied to the assessment of CVT operating errors, especially the assessment of operating errors of multiple CVTs in a power system.

[0052] When performing CVT error assessment, the error assessment device (such as a desktop or portable computer) first acquires the sampled amplitude values ​​of multiple CVTs at multiple measurement times as the first dataset. It's important to note that the sampled amplitude values ​​of multiple CVTs at the same measurement time are in phase, thus effectively reflecting the differences in error information. Next, based on the voltage proportionality coefficients of the multiple CVTs (i.e., the ratio between the secondary voltage and the primary voltage of the CVT), the first dataset is restored to the primary voltage level to obtain the second dataset. Then, a difference matrix is ​​constructed using the second dataset. In the difference matrix, each element represents the mean difference between any two CVTs. For example, the element in the second row and first column of the difference matrix represents the mean difference between the second CVT and the first CVT, while all diagonal elements of the difference matrix are 0. The difference matrix can serve as an evaluation index of the CVT error status, providing a data foundation for subsequent error assessment.

[0053] S102. Construct a likelihood function based on the difference matrix, and convert the likelihood function into a posterior probability density function based on Bayes' theorem.

[0054] It should be noted that after constructing the dissimilarity matrix, a likelihood function can be further constructed using the dissimilarity matrix. Based on the dissimilarity function, it can be further determined that the error of each CVT within the in-phase population follows a mean of [missing value]. The standard deviation is The normal distribution. Furthermore, the error calculation process for CVT under in-phase time can be defined as a function. , will the error As the vector of parameters to be estimated, the following likelihood function is constructed:

[0055]

[0056] in, Represents the likelihood function. This indicates the sampling amplitude of the CVT. Indicates error. For the number of CVT units, This represents the error calculation function. The standard deviation represents the error.

[0057] The likelihood function can then be transformed into a posterior probability density function using Bayes' theorem.

[0058]

[0059] in, Denotes the posterior probability density function. This represents the prior probability density.

[0060] S103. Extract a preset number of data corresponding to CVT from the second dataset to construct a third dataset, and determine the prior probability density based on the third dataset.

[0061] It should be noted that, to further determine the prior density, a third dataset can be constructed by extracting a predetermined number of data points corresponding to CVTs from the second dataset. This third dataset is used to determine the prior probability density. This prior probability density reflects the error of a CVT at different measurement times. By combining the errors of different CVTs at the same sampling time with the errors of a specific CVT at different measurement times, the accuracy of CVT error assessment can be further improved.

[0062] S104. Based on the prior probability density, an initial population is generated as the starting point of multiple Markov chains. Based on the multi-chain MCMC algorithm that integrates plant root and stem growth and the posterior probability density function, the error values ​​corresponding to multiple CVTs when the posterior probability density is maximized are determined.

[0063] It should be noted that after determining the prior probability density, an initial population can be generated using the prior probability density as the starting point for multiple Markov chains. Then, a multi-chain Markov chain Monte Carlo (MCMC) algorithm, which incorporates a plant root and stem growth algorithm, is used in conjunction with the posterior probability density function to determine the error values ​​corresponding to multiple CVTs when the posterior probability density is maximized, thus obtaining the final error evaluation result. In the multi-chain MCMC algorithm, the incorporated plant root and stem growth algorithm mainly serves to update individuals.

[0064] In summary, the CVT error assessment method provided by this invention first determines the error relationship of multiple CVTs under time-in-phase conditions by acquiring the sampling amplitudes of multiple CVTs at multiple measurement times. Then, it constructs a likelihood function and a posterior probability density function based on the error relationship of multiple CVTs under time-in-phase conditions. Next, it integrates the errors of CVTs at different measurement times into the error assessment process to improve the accuracy of error assessment. Finally, it determines the final error assessment result by integrating the multi-chain MCMC of plant root and stem growth. This effectively ensures the accuracy of CVT error assessment and also improves the efficiency of error assessment.

[0065] In some embodiments of the present invention, constructing the dissimilarity matrix based on the second dataset includes:

[0066] Determine the mean vector corresponding to the second dataset, where each element in the mean vector corresponds to the mean of the primary voltage level for each CVT.

[0067] A difference matrix is ​​constructed based on the difference between every two elements in the mean vector.

[0068] It should be noted that when constructing the difference matrix based on the second dataset, the mean vector corresponding to the second dataset can be determined first. Each element in the mean vector corresponds to the mean of the primary voltage level for each CVT. Then, the difference matrix is ​​constructed using the differences between every two elements in the mean vector. The specific construction process is as follows:

[0069] For the second dataset (m is the number of sampling times, n is the number of CVTs), and its mean vector is: ,in , Construct a difference matrix:

[0070]

[0071] In some embodiments of the present invention, the construction of the likelihood function based on the difference matrix includes:

[0072] Based on the difference matrix, a likelihood function is constructed for the mean of the primary voltage level corresponding to multiple CVTs, given that the mean of the difference between every two CVTs is known, where the mean of the difference between every two CVTs follows a normal distribution.

[0073] It should be noted that when constructing the likelihood function based on the difference matrix, the difference matrix can be used to construct a likelihood function of the mean of the primary voltage levels corresponding to multiple CVTs, given that the mean of the difference between any two CVTs is known. When constructing the likelihood function, the mean of the difference between any two CVTs follows a normal distribution. Based on the difference function, it can be further determined that the error of each CVT within the in-phase group follows a mean of... The standard deviation is The normal distribution. Furthermore, the error calculation process for CVT under in-phase time can be defined as a function. , will the error As the vector of parameters to be estimated, the following likelihood function is constructed:

[0074]

[0075] in, Represents the likelihood function. This indicates the sampling amplitude of the CVT. Indicates error. For the number of CVT units, This represents the error calculation function. The standard deviation of the error.

[0076] In some embodiments of the present invention, determining the prior probability density based on a third dataset includes:

[0077] Singular value decomposition is performed on the covariance matrix corresponding to the third dataset, and the mean of the change is determined based on the results of the singular value decomposition.

[0078] The initial value of the change is determined based on the third dataset, and the prior probability density is determined based on the initial value of the change and the mean of the change.

[0079] It should be noted that when determining the prior probability density based on the third dataset, singular value decomposition can be performed on the covariance matrix corresponding to the third dataset. The mean of the change can be determined based on the singular value decomposition result, and then the initial value of the change can be determined through the third dataset. Finally, the prior probability density can be determined through the initial value of the change and the mean of the change.

[0080] In some embodiments of the present invention, the step of generating an initial population based on prior probability density as the starting point for multiple Markov chains, and determining the error values ​​corresponding to multiple CVTs when the posterior probability density is maximized based on a multi-chain MCMC algorithm that integrates plant root and stem growth and a posterior probability density function, includes:

[0081] An initial population is generated based on prior probability density as the starting point for multiple Markov chains, and fitness values ​​are determined based on the fitness function.

[0082] Individual mutation is performed based on fitness values ​​and plant root and stem growth algorithms. The crossover probability factor is used to determine whether to accept mutated individuals, and the acceptance probability is used to determine whether to accept the mutation crossover operation, until multiple Markov chains converge.

[0083] It should be noted that when determining the error values ​​corresponding to multiple CVTs, an initial population can first be generated based on the prior probability density as the starting point of multiple Markov chains. Then, the fitness value is determined through the fitness function. Next, individual mutation is performed through the fitness value and the plant root and stem growth algorithm. The crossover probability factor is used to determine whether to accept mutated individuals, and the acceptance probability is used to determine whether to accept the mutation crossover operation, until multiple Markov chains converge.

[0084] In some embodiments of the present invention, the fitness function is determined based on the posterior probability density function.

[0085] It should be noted that the fitness function used to determine fitness can be determined by the posterior probability density function.

[0086] In some embodiments of the present invention, the individual variation based on fitness value and plant root and stem growth algorithm includes:

[0087] Using each Markov chain as the primary root, a mathematical model of fiber root growth is constructed based on fitness values, and individual variation is achieved through nutrient absorption and diffusion growth strategies.

[0088] It should be noted that when performing individual variation based on fitness values ​​and plant root and stem growth algorithms, each Markov chain can be regarded as the main root. A mathematical model of fiber root growth can be constructed through fitness values, and individual variation can be performed through nutrient absorption and diffusion growth strategies.

[0089] Combination Figure 2 Error assessment specifically includes the following steps:

[0090] 1. Analyze the physical correlation of capacitive voltage transformers in the same phase in high-voltage systems, and construct a time-dependent in-phase measurement model based on the consistency of primary voltage fluctuations at the same measurement point.

[0091] 2. Construct a Bayesian error measurement model, using the spatial in-phase error state assessment results as prior knowledge. At the same time, integrate small sample observation data into the temporal in-phase measurement model, construct a likelihood function, and use Bayesian methods to achieve the fusion of spatial and temporal in-phase models, so as to improve the accuracy of real-time assessment of CVT error state.

[0092] 3. Solve for the error under the fusion of spatial in-phase and temporal in-phase models by using the Bayesian posterior probability density of the error of the multi-chain MCMC sampling capacitive voltage transformer.

[0093] 4. In the multi-chain MCMC sampling posterior probability density, sampling is performed on the multi-chain subspace using the mechanism of the plant root and stem growth algorithm to improve the accuracy and efficiency of error calculation, thereby quickly approximating the Bayesian posterior probability density of the error.

[0094] 5. Collect the secondary operation data of the newly calibrated and put into operation capacitive voltage transformer, solve for the error through steps 1-4, and compare it with the error value in the power outage verification report. Calculate the method error value of the multi-chain MCMC Bayesian capacitive voltage transformer error assessment method. In normal assessment, solve for the error of the secondary operation data of the normally operating capacitive voltage transformer using the same steps. Based on this, correct the solution error under normal operation using the error value in the power outage verification report and the method error value, and use this as the assessment error of the capacitive voltage transformer to determine whether it exceeds the tolerance.

[0095] In step 1, an in-phase measurement model regarding time is constructed, specifically including:

[0096] In time-in-phase measurements, the primary voltage fluctuation information contained in the secondary output of a CVT at the same measurement point is essentially the same; the differences mainly lie in the differences in error information. Assume the number of CVTs is n, the number of sampling points is m, and the sampling dataset is denoted as... If the measured amplitude of the CVT is restored to the primary voltage level according to the nameplate scaling factor, while keeping the phase unchanged, then the corresponding dataset is: The mean vector of Y is: ,in, , i=1,2,…,n.

[0097] At this point, the difference matrix can be defined as:

[0098]

[0099] The mean of the absolute differences between each CVT and other CVTs in the same phase group is used as the evaluation index of the error state of that CVT, specifically: ki is the CVT nameplate scaling factor, i=1,2,…,n.

[0100] In step 2, a Bayesian error measurement model that fuses spatial and temporal in-phase models is constructed, specifically including:

[0101] Assuming the measurement sample matrix y is collected from the data matrix Y, y can be written as... Where k = 1, 2, ..., m, and the time-in-phase model in step 1 is f(y). Let be the vector of parameters to be estimated, representing the errors of each CVT within the same phase group. They follow a mean of . The standard deviation is Given a normal distribution, the likelihood function is constructed as follows:

[0102]

[0103] According to Bayes' theorem Normalizing the constant term in the denominator to 1, we obtain the posterior probability density as shown in the following equation:

[0104]

[0105] In step 2, the error of spatial in-phase solution is used as the prior probability, specifically including:

[0106] Extract data from data matrix Y to construct a modeling dataset. For modeling datasets Singular value decomposition is performed on the covariance matrix R, the number of principal components is selected, and the loading matrix of the residual subspace is determined. Calculate the Q value: .

[0107] Calculation process value A:

[0108]

[0109] for Let E be the mean vector function of the i-th row vector. Let Y be the standard deviation matrix of the dataset at each time step.

[0110] Error change value Change in statistical value The mapping relationship is as follows:

[0111]

[0112] Collect the ratio error value of the first inspection report of CVT As the initial value for error change, calculate the initial Q value. By averaging the changes Add to initial value The above yields the evaluation ratio error. :

[0113]

[0114] The input variable (error) is determined based on the spatial in-phase error state assessment results. The prior probability density. Assume the prior probability density of each input is... Their joint probability density is , as the prior probability for mutual inductor error assessment.

[0115] In step 3, the error value corresponding to the maximum Bayesian posterior probability density under the fusion of spatial and temporal in-phase models is solved using multi-chain MCMC, specifically including:

[0116] (1) The evaluation result of the spatial in-phase error state is used as the prior distribution of Bayes to generate an initial population ε(0), which serves as the starting point of N Markov chains, where ε is the population individual, M is the number of population individuals, and N is the dimension of the population individual.

[0117] (2) Find the best fitness value based on the fitness function.

[0118] (3) In the multi-chain algorithm, individual mutation is performed, and the mutated sample individuals are... .

[0119] (4) Determine whether to accept variant individuals based on the crossover probability factor CR. If u ≤ 1 - CR, then it is not accepted. Conversely, if the conditions are not met, then accept. The crossover probability factor CR∈[0,1], and u is a random number generated according to a uniform distribution from 0 to 1.

[0120] (5) Calculate the acceptance probability This is used to determine whether to accept the mutated crossover sample.

[0121] (6) According to Determine whether to accept mutated crossover samples , randomly generate u from [0, 1], if Then accept Otherwise, reservations will not be accepted. .

[0122] (7) Remove useless chains according to the interquartile range method.

[0123] (8) Determine convergence based on quantitative convergence criteria. When the parameter converges to a stable posterior distribution, the calculation ends; otherwise, steps (2)-(7) are repeated to continue evolving the parallel chains. The SR calculation method is shown in the following formula:

[0124]

[0125] In the formula, Let be the number of generations in each Markov chain; The number of Markov chains used for evaluation; for The variance of the mean of a Markov chain; for The average variance of a Markov chain.

[0126] In step 3, the expression for the fitness function is:

[0127]

[0128] in, It is a constant. Assume that the number of individuals in each population is [missing information]. The state of the Markov chain in generation t is Calculate the fitness value of each individual in the population. The minimum fitness value is the best fitness value, denoted as . The corresponding set is .

[0129] In step 4, the multi-chain MCMC strategy based on the fusion of plant root and stem growth specifically includes:

[0130] Based on the growth characteristics of taproot plants, lateral roots grow from the taproot, and fibrous roots also proliferate from both the taproot and lateral roots. According to this characteristic, in the variation sampling of the error solution of multi-chain MCMC Bayesian optimization, the main chain is considered the main root, and the remaining chains are considered fibrous roots. Based on this, the variation is optimized according to the mathematical model of taproot plant growth. By simulating the process of roots searching for nutrients in the soil, a combination of global search and local optimization is achieved to realize rapid and accurate optimization variation, thereby improving the efficiency and accuracy of sampling. The specific implementation is as follows:

[0131] The mathematical model for fiber root growth is as follows:

[0132]

[0133] in, This indicates the t-th growth of the i-th fiber root. This indicates that the best fiber root found so far has been discovered in history. This represents a fiber root randomly selected from the N fiber roots grown during the t-th growth cycle. A random number in the interval (-0.5, 1.5) is generated through... Random search ensures the best historical fiber roots It plays a dominant role in guiding all fiber roots to grow in the optimal direction, while also ensuring that some fiber roots grow randomly. This guides the remaining fiber roots to perform random searches, thus reducing the probability of the search getting trapped in local optima.

[0134] Assuming the taproot plant has w lateral roots, then the M-1 fibrous roots are randomly divided into w groups, and the mathematical model for each lateral root is as follows:

[0135]

[0136] in, This represents the amount of nutrients absorbed by the lateral roots of each group during the t-th growth cycle. This represents the average nutrient absorption of all fiber roots in each group during the t-th growth cycle. This represents a fiber root randomly selected from the M-1 fiber roots grown in the t-th cycle. A random number in the interval (-1, 1) is generated by... Randomly select a balance between global and local search.

[0137] Randomly select from w lateral root seeds2 The group, as a sampling subspace during the mutation process, is updated as follows:

[0138]

[0139] Where i = 1, 2, ..., M, v represents the mutant individuals, and seed2 represents the model parameters. Represents the number of candidate samples. is the scaling factor, where d represents the dimension of mutation. It is generated by a symmetrical distribution with very small variance. and .

[0140] To verify the effectiveness of the method, an experimental verification was conducted at a newly connected 220kV substation. This substation has a 3 / 2 connection, with two sets of Potential Transformers (PTs) installed on the busbar (one for each of the three phases, referred to as one set). A total of six CVTs are installed on the outgoing lines. The PTs have an accuracy class of 0.2, and the CVTs have an accuracy class of 0.5. A high-precision data acquisition device is deployed within the substation, with a sampling rate of 10kHz, outputting one feature quantity per second. The secondary outputs of each normal A-phase CVT are collected as the experimental dataset. The corresponding topology is as follows: Figure 3 As shown, the specific verification steps are as follows:

[0141] (1) Collect the secondary operation data of the newly calibrated and put into operation capacitive voltage transformer, evaluate it using the method provided by this invention, and compare the evaluation result with the error value of the power outage verification report as the method error.

[0142] (2) In normal evaluation, the secondary operation data of the normally operating capacitive voltage transformer is solved for error using the method provided by the present invention. Based on this, the error value of the power outage verification report and the method error value are used to correct the solution error under normal operation.

[0143] (3) Use this error as the evaluation error of the capacitive voltage transformer to determine whether it is out of tolerance.

[0144] 130 days of operational data were collected, and a set of datasets corresponding to each day was randomly selected. The same ratio error (+0.2%) was added, and the deviation between the ratio error and phase error obtained from each day's evaluation was used as the evaluation index to determine whether the magnitude of the deviation showed an increasing trend over time, thus demonstrating the long-term effectiveness of the method provided by this invention. The evaluation error was as follows: Figure 4 As shown.

[0145] Depend on Figure 4 It can be seen that during the 130-day monitoring period, the evaluation ratio error of the six CVTs was around 0.2% of the added human error, and the evaluation error deviation was within 0.05%. Moreover, the evaluation error deviation did not increase significantly over time, which verifies that the error evaluation method provided by the present invention has good long-term effectiveness.

[0146] To better implement the CVT error assessment method in the embodiments of the present invention, based on the CVT error assessment method, correspondingly, as follows: Figure 5 As shown, this embodiment of the invention also provides a CVT error assessment device, the CVT error assessment device 500 comprising:

[0147] The acquisition module 501 is used to acquire the sampled amplitude values ​​of multiple CVTs at multiple measurement times as the first dataset, restore the first dataset to the primary voltage level based on the voltage proportional coefficient of multiple CVTs to obtain the second dataset, and construct a difference matrix based on the second dataset.

[0148] Transformation module 502 is used to construct a likelihood function based on the difference matrix and convert the likelihood function into a posterior probability density function based on Bayes' theorem.

[0149] The first determining module 503 is used to extract a preset number of data corresponding to CVT from the second dataset to construct a third dataset, and to determine the prior probability density based on the third dataset.

[0150] The second determining module 504 is used to generate an initial population based on the prior probability density as the starting point of multiple Markov chains, and to determine the error values ​​corresponding to multiple CVTs when the posterior probability density is maximized based on the multi-chain MCMC algorithm that integrates plant root and stem growth and the posterior probability density function.

[0151] The construction of the dissimilarity matrix based on the second dataset includes:

[0152] Determine the mean vector corresponding to the second dataset, where each element in the mean vector corresponds to the mean of the primary voltage level for each CVT.

[0153] A difference matrix is ​​constructed based on the difference between every two elements in the mean vector.

[0154] The CVT error assessment device 500 provided in the above embodiments can realize the technical solutions described in the above CVT error assessment method embodiments. The specific implementation principles of each module or unit can be found in the corresponding content in the above CVT error assessment method embodiments, and will not be repeated here.

[0155] like Figure 6 As shown, the present invention also provides an error assessment device 600. The error assessment device 600 includes a processor 601, a memory 602, and a display 603. Figure 6 Only some components of the error assessment device 600 are shown; however, it should be understood that implementation of all shown components is not required, and more or fewer components may be implemented instead.

[0156] In some embodiments, processor 601 may be a central processing unit (CPU), microprocessor, or other data processing chip, used to run program code stored in memory 602 or process data, such as the CVT error assessment method of the present invention.

[0157] In some embodiments, processor 601 may be a single server or a group of servers. The server group may be centralized or distributed. In some embodiments, processor 601 may be local or remote. In some embodiments, processor 601 may be implemented on a cloud platform. In one embodiment, the cloud platform may include a private cloud, public cloud, hybrid cloud, community cloud, distributed cloud, internal cloud, multi-cloud, etc., or any combination thereof.

[0158] In some embodiments, memory 602 may be an internal storage unit of error assessment device 600, such as a hard disk or memory of error assessment device 600. In other embodiments, memory 602 may also be an external storage device of error assessment device 600, such as a pluggable hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., provided on error assessment device 600.

[0159] Furthermore, the memory 602 may include both internal storage units of the error assessment device 600 and external storage devices. The memory 602 is used to store the application software and various types of data installed on the error assessment device 600.

[0160] In some embodiments, display 603 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an organic light-emitting diode (OLED) touchscreen. Display 603 is used to display information from the error assessment device 600 and to display a user interface for visualization. Components 601-603 of the error assessment device 600 communicate with each other via a system bus.

[0161] In one embodiment, when processor 601 executes the CVT error evaluation program in memory 602, the following steps can be implemented:

[0162] The sampled amplitudes of multiple CVTs at multiple measurement times are obtained as the first dataset. The first dataset is restored to the primary voltage level based on the voltage scaling factor of multiple CVTs to obtain the second dataset. A difference matrix is ​​constructed based on the second dataset.

[0163] A likelihood function is constructed based on the difference matrix, and the likelihood function is transformed into a posterior probability density function based on Bayes' theorem.

[0164] A third dataset is constructed by extracting a predetermined number of data corresponding to CVTs from the second dataset, and the prior probability density is determined based on the third dataset.

[0165] The initial population is generated based on the prior probability density as the starting point of multiple Markov chains. Based on the multi-chain MCMC algorithm that integrates plant root and stem growth and the posterior probability density function, the error values ​​corresponding to multiple CVTs when the posterior probability density is maximized are determined.

[0166] The construction of the dissimilarity matrix based on the second dataset includes:

[0167] Determine the mean vector corresponding to the second dataset, where each element in the mean vector corresponds to the mean of the primary voltage level for each CVT.

[0168] A difference matrix is ​​constructed based on the difference between every two elements in the mean vector.

[0169] It should be understood that when the processor 601 executes the CVT error evaluation program in the memory 602, in addition to the functions mentioned above, it can also perform other functions, as can be found in the description of the corresponding method embodiments above.

[0170] Furthermore, this embodiment of the invention does not specifically limit the type of the error assessment device 600 mentioned. The error assessment device 600 can be a portable electronic device such as a mobile phone, tablet computer, personal digital assistant (PDA), wearable device, or laptop computer. Exemplary embodiments of portable electronic devices include, but are not limited to, portable electronic devices running iOS, Android, Microsoft, or other operating systems. The aforementioned portable electronic devices can also be other portable electronic devices, such as laptop computers with touch-sensitive surfaces (e.g., touch panels). It should also be understood that in some other embodiments of the invention, the error assessment device 600 may not be a portable electronic device, but rather a desktop computer with a touch-sensitive surface (e.g., a touch panel).

[0171] Accordingly, this application also provides a computer-readable storage medium for storing a computer-readable program or instruction. When the program or instruction is executed by a processor, it can implement the steps or functions in the CVT error evaluation method provided in the above-described method embodiments.

[0172] Those skilled in the art will understand that all or part of the processes of the methods described in the above embodiments can be implemented by a computer program instructing related hardware (such as a processor, controller, etc.), and the computer program can be stored in a computer-readable storage medium. The computer-readable storage medium may be a disk, optical disk, read-only memory, or random access memory, etc.

[0173] The CVT error assessment method, apparatus, error assessment device, and storage medium provided by the present invention have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A CVT error assessment method, characterized in that, include: The sampled amplitudes of multiple CVTs at multiple measurement times are obtained as the first dataset. The first dataset is restored to the primary voltage level based on the voltage scaling factor of multiple CVTs to obtain the second dataset. A difference matrix is ​​constructed based on the second dataset. A likelihood function is constructed based on the difference matrix, and the likelihood function is transformed into a posterior probability density function based on Bayes' theorem. A third dataset is constructed by extracting a predetermined number of data corresponding to CVTs from the second dataset, and the prior probability density is determined based on the third dataset. The initial population is generated based on the prior probability density as the starting point of multiple Markov chains. Based on the multi-chain MCMC algorithm that integrates plant root and stem growth and the posterior probability density function, the error values ​​corresponding to multiple CVTs when the posterior probability density is maximized are determined. The construction of the dissimilarity matrix based on the second dataset includes: Determine the mean vector corresponding to the second dataset, where each element in the mean vector corresponds to the mean of the primary voltage level for each CVT. A difference matrix is ​​constructed based on the difference between every two elements in the mean vector.

2. The CVT error assessment method according to claim 1, characterized in that, The construction of the likelihood function based on the difference matrix includes: The error of each CVT within the same phase group follows the mean value. The standard deviation is The normal distribution is used to define the error solution process of CVT under time-in-phase conditions as a function. , will the error As the vector of parameters to be estimated, the following likelihood function is constructed: in, Represents the likelihood function. This indicates the sampling amplitude of the CVT. Indicates error. For the number of CVT units, This represents the error calculation function. The standard deviation represents the error; The likelihood function is transformed into a posterior probability density function using Bayes' theorem. in, Denotes the posterior probability density function. This represents the prior probability density.

3. The CVT error assessment method according to claim 1, characterized in that, The determination of prior probability density based on a third dataset includes: Singular value decomposition is performed on the covariance matrix corresponding to the third dataset, and the mean of the change is determined based on the results of the singular value decomposition. The initial value of the change is determined based on the third dataset, and the prior probability density is determined based on the initial value of the change and the mean of the change.

4. The CVT error assessment method according to claim 1, characterized in that, The initial population generated based on prior probability density serves as the starting point for multiple Markov chains. Based on a multi-chain MCMC algorithm that integrates plant root and stem growth and a posterior probability density function, the error values ​​corresponding to multiple CVTs when the posterior probability density is maximized are determined, including: An initial population is generated based on prior probability density as the starting point for multiple Markov chains, and fitness values ​​are determined based on the fitness function. Individual mutation is performed based on fitness value and plant root and stem growth algorithm. The crossover probability factor is used to determine whether to accept mutated individuals, and the acceptance probability is used to determine whether to accept the mutation crossover operation, until multiple Markov chains converge.

5. The CVT error assessment method according to claim 4, characterized in that, The fitness function is determined based on the posterior probability density function.

6. The CVT error assessment method according to claim 4, characterized in that, The individual variation based on fitness value and plant root and stem growth algorithm includes: Using each Markov chain as the primary root, a mathematical model of fiber root growth is constructed based on fitness values, and individual variation is achieved through nutrient absorption and diffusion growth strategies.

7. A CVT error assessment device, characterized in that, include: The acquisition module is used to acquire the sampled amplitude values ​​of multiple CVTs at multiple measurement times as the first dataset, restore the first dataset to the primary voltage level based on the voltage proportional coefficient of multiple CVTs to obtain the second dataset, and construct a difference matrix based on the second dataset. The transformation module is used to construct a likelihood function based on the difference matrix and convert the likelihood function into a posterior probability density function based on Bayes' theorem. The first determining module is used to extract a preset number of data corresponding to CVT from the second dataset to construct a third dataset, and to determine the prior probability density based on the third dataset. The second determination module is used to generate an initial population based on the prior probability density as the starting point of multiple Markov chains, and to determine the error values ​​corresponding to multiple CVTs when the posterior probability density is maximized based on the multi-chain MCMC algorithm that integrates plant root and stem growth and the posterior probability density function. The construction of the dissimilarity matrix based on the second dataset includes: Determine the mean vector corresponding to the second dataset, where each element in the mean vector corresponds to the mean of the primary voltage level for each CVT. A difference matrix is ​​constructed based on the difference between every two elements in the mean vector.

8. An error assessment device, characterized in that, Including memory and processor, among which, The memory is used to store programs; The processor, coupled to the memory, is used to execute the program stored in the memory to implement the steps in the CVT error assessment method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, Used to store computer-readable programs or instructions, which, when executed by a processor, can implement the steps in the CVT error assessment method according to any one of claims 1 to 6.

Images