Current transformer error online monitoring method and system based on elastic network regression

Abstract

The invention discloses a current transformer error online monitoring method and system based on elastic network regression, and belongs to the technical field of current transformer error monitoring, and the method comprises the steps: collecting m groups of secondary fundamental current phasor values of n current transformers, and constructing an m * n-dimensional secondary current data matrix; establishing an n-dimensional overdetermined solution regression equation set based on the secondary current data matrix, the to-be-solved mutual inductor error parameter matrix and the residual vector; constructing an elastic network hybrid regularization loss function; solving the overdetermined regression equation set by adopting a coordinate descent method to obtain a mutual inductor error parameter matrix; and calculating the ratio difference and the angular difference of each mutual inductor according to the mutual inductor error parameter matrix. The technical problems that the error identification precision of an existing on-line monitoring device of the mutual inductor is not high enough, and multi-mutual inductor monitoring cannot be completed are solved.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of current transformer error monitoring, and particularly relates to a current transformer error online monitoring method and system based on elastic network regression. BACKGROUND

[0002] In recent years, with the rapid advancement of smart grid construction, smart substations are accelerating the realization of full-station information digitization, communication networking and information sharing standardization. Under this background, electric energy metering problems have never been valued as much, and higher standards have been put forward for their accuracy and error online monitoring capability.

[0003] However, the existing digitalized metering system still has obvious shortcomings: the traditional error detection method mainly relies on manual inspection. The operation and maintenance personnel periodically go to the scene to visually inspect the current transformer and other equipment, compare instrument readings, measure with handheld detection instruments, and judge whether the equipment has abnormal problems according to the operation and maintenance experience. This method needs to access the test line under the condition of power-off or isolation, and does not have the real-time error capturing capability in the continuous operation process, which is not only low in efficiency and high in cost, but also difficult to meet the requirements of real-time monitoring and accurate diagnosis of smart substations; at the same time, the existing high-voltage substation transformer error detection device compares the measured transformer with the standard signal by introducing a reference standard transformer or built-in precision current source, and obtains the ratio difference and angle difference parameters by using the sampling and calculation module. However, such devices generally rely on external standard equipment, and the detection process lacks intelligent online monitoring scheme, making it difficult to realize automatic diagnosis in the running state; the patent "Substation current transformer relative error verification method, device and electronic equipment" gives a method for monitoring the error of multiple current transformers, but this method uses the least square method to solve the parameters, although it can provide unbiased optimal linear estimation, but it is extremely sensitive to the linear correlation between the explanatory variables. When there is a high correlation (i.e. multicollinearity) between multiple transformer secondary current signals, the transpose product of the design matrix in the least square method is close to singular, resulting in numerical instability in the solving process, the parameter estimation variance increases significantly, and even significant distortion occurs. This phenomenon makes the loss function model fitted by the least square method show problems such as decreased identification accuracy, large fluctuation of results, and inability to effectively identify subtle error shifts in fault identification, which seriously affects the online monitoring capability of transformer error.

[0004] Therefore, the existing method is difficult to meet the actual needs of centralized metering devices in terms of intelligence and monitoring accuracy. In view of this situation, it is urgent to build a new type of current transformer error monitoring method, which needs to make full use of centralized metering data and has excellent error identification accuracy, and can complete the monitoring of multiple transformers online, thereby effectively guaranteeing the reliable operation of the power metering system. SUMMARY

[0005] The application aims to provide an online monitoring method and system for errors of current transformers based on elastic network regression, so as to solve the technical problem that the error identification accuracy of existing online monitoring devices for transformers is not high enough and multiple transformer monitoring cannot be completed.

[0006] In order to achieve the above-mentioned purpose, the application adopts the following technical solutions: In a first aspect, the application provides an online monitoring method for errors of current transformers based on elastic network regression, comprising: Collecting m groups of secondary fundamental current phasor values of n current transformers to construct an m*n-dimensional secondary current data matrix; establishing an n-dimensional overdetermined solution regression equation group based on the secondary current data matrix, a to-be-solved transformer error parameter matrix and a residual error vector; Constructing an elastic network mixed regularization loss function; Solving the n-dimensional overdetermined solution regression equation group by using a coordinate descent method to obtain the transformer error parameter matrix; Calculating the ratio error and the angle error of each transformer according to the transformer error parameter matrix.

[0007] Further, the n-dimensional overdetermined solution regression equation group is: ; wherein is the input secondary current data matrix, is the mth group of secondary fundamental current phasor values of the nth transformer; is the to-be-solved transformer error parameter matrix, the elements in the to-be-solved transformer error parameter matrix are complex numbers, is the error parameter of the first transformer, is the error parameter of the nth transformer, , is the ratio error of the ith transformer, is the angle error of the ith transformer, is the residual error of the nth transformer.

[0008] Further, the elastic network mixed regularization loss function is:

[0009] wherein, is the original least square loss, is the L1 norm of the matrix A, is the L2 norm of the matrix A, is the regularization strength, is the mixing coefficient; the loss function Q E is the minimum, the regression iteration outputs the optimal parameter matrix A , and then the optimal ratio error and angle error are obtained.

[0010] Further, the coordinate descent method comprises the following steps: inputting a quadratic current data matrix I2, a preselected regularization strength γ and a mixing coefficient α; initializing a regression parameter matrix A; iteratively updating parameters: fixing other parameters, element by element solving a value of aᵢ that minimizes the elastic network mixed regularization loss function; when if the value is less than a set threshold, stopping iteration.

[0011] Further, the preselection of the regularization strength γ and the mixing coefficient α adopts a 5-fold cross-validation method.

[0012] Further, the preselection of the regularization strength γ and the mixing coefficient α comprises the following steps: dividing data in the quadratic current data matrix I2 into 5 parts; for each candidate (γ, α) combination, alternately selecting 4 parts of data to train a model, and calculating a mean square error with the remaining 1 part of data; selecting (γ, α) at the time when the average mean square error is the smallest as the selected regularization strength and mixing coefficient.

[0013] In a second aspect, the present application provides an online current transformer error monitoring system based on elastic network regression, comprising: a modeling module, configured to collect m groups of secondary fundamental current phasor values of n current transformers, and construct a m×n-dimensional secondary current data matrix; and establish an n-dimensional overdetermined solution regression equation group based on the secondary current data matrix, a to-be-solved transformer error parameter matrix and a residual error vector; a loss function construction module, configured to construct an elastic network mixed regularization loss function; a solution module, configured to solve the overdetermined solution regression equation group by using a coordinate descent method, and obtain a transformer error parameter matrix; a calculation module, configured to calculate a ratio error and an angle error of each current transformer according to the transformer error parameter matrix.

[0014] In a third aspect, the present application provides an electronic device, comprising: at least one processor; and a memory in communication connection with the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the online current transformer error monitoring method based on the elastic network regression according to any one of the first aspect of the present application.

[0015] Fourthly, the present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the online monitoring method for current transformer errors based on elastic network regression as described in any one of the first aspects of the present invention.

[0016] Fifthly, the present invention provides a computer program product, comprising a computer program, characterized in that, when the computer program is executed by a processor, it implements the steps of the online monitoring method for current transformer errors based on elastic network regression as described in the first aspect of the invention.

[0017] Compared with the prior art, the present invention has at least the following beneficial technical effects: 1. Low sensitivity to initial error values ​​of current transformers: By using the elastic network regression method to process the loss function, this method can still effectively converge to the optimal solution even when the initial parameter values ​​are far from the true values, reducing the dependence on the initial value setting and exhibiting low initial value sensitivity. This characteristic makes the method more adaptable and reliable in practical engineering applications.

[0018] 2. It eliminates the need to rely on maintenance experience to determine if there are any abnormal issues with the equipment. Therefore, it eliminates the need for manual inspection of test lines during power outages or isolation. It has the ability to capture errors in real time during continuous operation, which improves maintenance efficiency and security.

[0019] 3. Obtaining the absolute value of current transformer error: Addressing the potential multiple solutions to the current transformer error model, the elastic network regression method, by designing a strictly convex objective function, guarantees a theoretically unique solution, thereby obtaining the absolute value of the current transformer error. Specifically, the elastic network regression objective function incorporates L1 and L2 regularization terms, maintaining the sparsity of model parameters while enhancing stability. Its strictly convex nature ensures that the objective function has only one global minimum point in the parameter space, theoretically guaranteeing the uniqueness of the error model parameter estimation.

[0020] 4. High Accuracy of Current Transformer Error Identification: This method achieves high identification accuracy by compressing the error of fault-free current transformers to zero. The maximum relative error identification accuracy can reach: ratio difference 0.1%, angle difference 0.001 rad (3.4'). Currently, the highest accuracy class of current transformers used in power systems is 0.2, i.e., ratio difference measurement accuracy is ±0.2%, and angle difference measurement accuracy is ±0.002 rad (±7'). 0.1 class current transformers are mainly used for calibration. Therefore, the ratio difference and angle difference deviations of faulty current transformers are generally much greater than the measurement accuracy of 0.2 class current transformers. The identification accuracy of the elastic network regression algorithm meets all requirements. Attached Figure Description

[0021] Figure 1 A schematic diagram of the traditional architecture for metering and monitoring of substation busbars; Figure 2 Flowchart of the online error monitoring method for current transformers provided by the present invention; Figure 3 A simulation model for a multi-channel current transformer; Figure 4a A comparison chart of error diagnosis simulation results for one embodiment - error distribution of the comparison parameter; Figure 4b A comparison chart of error diagnosis simulation results for one embodiment - error distribution of angle difference parameter; Figure 5 This is a schematic diagram of a three-transformer experimental system. Figure 6 This is a structural block diagram of an online error monitoring system for current transformers based on elastic network regression provided in an embodiment of the present invention. Figure 7 This is a block diagram of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0022] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0023] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0024] The present invention will now be described in further detail with reference to the accompanying drawings: Example 1 Please see Figure 1 , Figure 1A schematic diagram of the traditional architecture for substation busbar metering and monitoring is given. Current transformers (CTs) and voltage transformers (TVs) are connected to lines 1 through 2, and then to an analog input merging unit. The analog input merging unit merges and synchronizes the electrical quantities transmitted from the transformers. Data observations are integrated into message data according to a standard protocol and sent to a switch. In the switch, the received information is regenerated and amplified. The switch receives message data from multiple merging units and sends it to a centralized metering device via optical fiber. The method provided in this invention is applied to the monitoring and analysis unit in the centralized metering device. The centralized metering device includes an energy metering unit and a monitoring and analysis unit. The energy metering unit collects raw energy signals, and the monitoring and analysis unit calculates and converts these signals and compares them with a set threshold to determine anomalies.

[0025] Reference Figure 2 An online error monitoring method for current transformers based on elastic network regression includes the following steps: Step 1: Establish a current transformer error diagnosis model: Actual operating current transformers exhibit two types of errors: ratio error and phase error. By collecting and calculating the phasor values ​​of the secondary fundamental current of n transformers, an n-dimensional overdetermined regression equation set is established. This n-dimensional overdetermined regression equation set constitutes the current transformer error diagnosis model. The n-dimensional overdetermined regression equation set is as follows: (1) in For input Secondary current data matrix For the m-th group of secondary fundamental current phasors of the n-th current transformer, there is one secondary fundamental current phasor value for each group; Let be the matrix of error parameters of the mutual inductor to be determined, and let the elements of the matrix be complex numbers. The error parameter of the first current transformer. Let n be the error parameter of the nth current transformer. Let be the residual of the nth mutual inductor.

[0026] And:

[0027] in Let the ratio difference be that of the i-th mutual inductor. Let be the angle difference of the i-th mutual inductor.

[0028] Step 2: Elastic Network Hybrid Regularization: By compressing some regression coefficients through L1 regularization terms, the current transformer error diagnosis model becomes simpler, more optimized, and still accurate. The L1 regularization term compresses the error of non-faulty CTs to zero, while the L2 regularization term addresses ill-posed problems, i.e., the problem of multicollinearity and non-unique solutions. Elastic network regression balances the constraints of L1-norm and L2-norm regularization terms, taking into account the advantages of both, and can achieve more accurate regression of the current transformer error diagnosis model.

[0029] Introducing elastic network regression to construct a hybrid regularized loss function Q E :

[0030] in, For the original least squares loss, Let A be the L1 norm of matrix A. Let A be the L2 norm of matrix A. For regularity, The mixing coefficients are used. The loss function Q is... E When the minimum value is reached, the regression iteration yields the optimal matrix of the current transformer error parameters to be determined. A Then, the optimal ratio difference and angle difference can be obtained.

[0031] Step 3: Use coordinate descent to adjust the loss function Q. E Optimization: Based on the principle of elastic network regression, an optimal set of mutual inductor error parameter matrices is selected. A This makes the above regularization loss function Q E Minimum. The specific algorithm steps are as follows: 1) Input current transformer secondary current data matrix I2 ( Where m is the number of samples collected, and n is the number of mutual inductors; 2) Perform parameter adjustments gamma and α Pre-selection and cross-validation. A 5-fold cross-validation method is used: after dividing the fundamental current phasor of the secondary side of the transformer into 5 data subsets, each pre-selected... gamma and α Four data subsets were used to train the model using the coordinate descent method, and one data subset was used to calculate the MSE (Mean Square Error). The one data subset included N sets of sample data, and each set of sample data included n phasor values ​​of the secondary fundamental current of the current transformer.

[0032]

[0033] in, This represents the predicted value of the error parameter matrix A for the t-th sample group. Let A represent the actual value of the error parameter matrix A for the t-th sample group; Repeatedly perform data subset The above process is repeated, and finally, the average of the five MSE values ​​is calculated. The smaller the average value, the more accurate the model's prediction. The parameters in the loss function are determined based on this principle. gamma and α ; 3) Assign initial values ​​to the error parameter matrix. (n×1), It should be a complex matrix; 4) Use the coordinate descent method to solve the regression problem, that is, only one parameter in the error parameter matrix A is considered each time. Update as follows: In other parameters ( Given a fixed loss function, find the one that makes the loss function... Q E smallest Value. Once all n parameters have been updated, one iteration is considered complete, yielding a new parameter matrix. A 1; 5) Repeat step 4) until... Stop iteration ( j (Number of iterations); Step 4: Calculate the ratio difference Sum of angle difference And output the ratio difference The ratio difference matrix formed and the difference of each angle The angle difference matrix formed The calculation formula is: .

[0034] During simulation, the phasor values ​​of the secondary fundamental current are acquired by establishing a multi-channel current transformer simulation model. Please refer to [link / reference]. Figure 3 , Figure 3 A simulation model of multiple current transformers is demonstrated. Taking a 5-line circuit as an example, a simulation model of 5 current transformers is established to simulate the connection relationship of the bus current transformers. By adjusting the values ​​of shunt resistors R1 to R5, the primary current values ​​of CT1 to CT5 can be changed. One or more current transformers from CT1 to CT5 are randomly selected as faulty current transformers, and the offset of ratio and phase angle is generated by adjusting the parameter values ​​of the Transformer module (CT1 to CT5). Generally, the failure rate of current transformers will not exceed 50%. If the total number of current transformers is n, the number of faulty transformers should be less than n / 2; for a 5-line circuit, this should be less than or equal to 2. By continuously adjusting the values ​​of R1 to R5, multiple sets of data can be obtained for subsequent algorithm calculations.

[0035] Each set of data includes: the phasor values, ratio difference, and angle difference of the secondary fundamental current of transformers CT1 to CT5.

[0036] Please see Figure 4a and Figure 4b , Figure 4a and Figure 4b A comparison of simulation results for error diagnosis in one embodiment is shown. TA2 and TA4 are designated as faulty transformers, with the following error offsets: TA2: ratio error 1.56%, angle error 0.0278 rad; TA4: ratio error 3.75%, angle error 0.0156 rad. The two figures compare the error distributions of ratio error and angle error under different regression problems and algorithms.

[0037] As can be seen in the simulation, the elastic network regression achieves the highest identification accuracy by compressing the error of the fault-free current transformer to 0.

[0038] Please see Figure 5 , Figure 5 The schematic diagram of the three-current transformer experimental system provided by this invention is shown. The output of the programmable power frequency current source is divided into three branch currents after passing through three shunts. Each branch has a current transformer (TA1~TA3) connected in series, which is used to simulate the line transformers on a transmission bus. Their primary currents satisfy the KCL equation. The secondary outputs of all transformers are sampled by an A / V conversion circuit and then sent to the terminal to generate SV message data. An adjustable impedance Z is connected in parallel with a switch in the branch containing TA1. When the switch is closed, a portion of the current in TA1 will be diverted through the adjustable impedance Z, thereby causing proportional and phase shift errors in the secondary current of TA1. By replacing shunts 2 and 3 with different specifications, the primary current in the current transformers can be changed, thereby obtaining multiple sets of experimental sample data for regression analysis.

[0039] Each set of experimental sample data includes: secondary side fundamental current phasor, ratio difference, and angle difference.

[0040] Please refer to Table 1, which shows a comparison of the ratio shift experimental results using gradient descent based on least squares regression and coordinate descent after applying elastic network regression to the original loss function. By connecting an impedance Z in parallel with a switch in the branch containing TA1, a ratio shift is generated in the secondary current of TA1. Since the impedance of the shunt in the branch containing TA1 remains constant at 1mΩ, connecting a resistor of 50mΩ in parallel can generate a 2% ratio shift in TA1.

[0041] Please refer to Table 2, which shows a comparison of the angle shift experimental results using the gradient descent method based on least squares regression and the coordinate descent method applied after applying elastic network regression to the original loss function. By connecting an impedance Z in parallel with a switch in the branch containing TA1, an angle shift is generated in the secondary current of TA1. Since the impedance value of the shunt in the branch containing TA1 is constant at 1mΩ, connecting an inductor with a parallel value of 0.157mH can generate an angle shift of 0.02rad in TA1.

[0042] Table 1. Estimation Results of Experimental Data for 2% Bias Migration

[0043] Table 2. Estimation Results of Experimental Data for 0.02 rad Angular Deviation

[0044] As can be seen in the experiment, under the two error conditions mentioned above, applying coordinate descent after elastic network regression to the original loss function can more accurately identify the error of the mutual inductor than gradient descent based on least squares regression.

[0045] The algorithm proposed in this invention is designed to meet the practical needs of online monitoring of current transformer errors in current power systems, and is suitable for high-precision diagnosis of transformer errors in high-voltage centralized metering systems. Furthermore, the established error modeling and regression solution method has good adaptability to transformer topologies and can be extended to other systems with multi-transformer structures and the ability to acquire secondary data, enabling error analysis and fault identification.

[0046] The following are system embodiments of the present invention, which can be used to execute the method embodiments of the present invention. For details not disclosed in the system embodiments, please refer to the method embodiments of the present invention.

[0047] Example 2 Please see Figure 6 In this embodiment, an online error monitoring system for current transformers based on elastic network regression is provided, including: The modeling module is used to collect m sets of secondary fundamental current phasor values ​​from n current transformers and construct an m×n dimensional secondary current data matrix; based on the secondary current data matrix, the error parameter matrix of the transformer to be determined, and the residual vector, an n-dimensional overdetermined regression equation system is established. The loss function building module is used to construct the hybrid regularization loss function for elastic networks; The solution module is used to solve the overdetermined regression equations using the coordinate descent method to obtain the mutual inductor error parameter matrix. The calculation module is used to calculate the ratio difference and angle difference of each transformer based on the transformer error parameter matrix.

[0048] All relevant content of each step involved in the aforementioned embodiments of the online monitoring method for current transformer errors based on elastic network regression can be referenced to the functional description of the corresponding functional module of the online monitoring system for current transformer errors based on elastic network regression in the embodiments of the present invention, and will not be repeated here.

[0049] Example 3 Reference Figure 7 This embodiment provides an electronic device including a processor and a memory, with the processor and memory connected via a bus. The memory stores a computer program, which includes program instructions. The processor executes the program instructions stored in the computer storage medium. The processor may be a Central Processing Unit (CPU), or 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, suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions in the computer storage medium to achieve corresponding method flows or corresponding functions. The processor described in this embodiment can be used for the operation of an online monitoring method for current transformer errors based on elastic network regression. The bus may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. Buses can be categorized into address buses, data buses, control buses, etc. For ease of representation, Figure 7 The symbol is represented by only one line, but this does not mean that there is only one bus or one type of bus.

[0050] Example 4 This embodiment provides a storage medium, specifically a computer-readable storage medium (Memory), which is a memory device in an electronic device used to store programs and data. It is understood that the computer-readable storage medium here can include both the built-in storage medium in the electronic device and extended storage media supported by the electronic device. The computer-readable storage medium provides storage space that 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 computer programs (including program code). It should be noted that the computer-readable 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 computer-readable storage medium to implement the corresponding steps of the online monitoring method for current transformer errors based on elastic network regression in the above embodiment.

[0051] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application 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.

[0052] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. 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... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0053] 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.

[0054] 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.

[0055] Example 5 This embodiment provides a computer program product, including a non-volatile computer-readable storage medium storing the computer program product. When the computer program is executed by a processor, it implements the steps of the methods described in various embodiments of this application.

[0056] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0057] 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 it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for online monitoring of current transformer errors based on elastic network regression, characterized in that, include: Collect m sets of secondary fundamental current phasor values ​​from n current transformers to construct an m×n dimensional secondary current data matrix; An n-dimensional overdetermined regression equation system is established based on the secondary current data matrix, the error parameter matrix of the transformer to be determined, and the residual vector. Construct a hybrid regularized loss function for elastic networks; The coordinate descent method is used to solve the n-dimensional overdetermined regression equation system to obtain the mutual inductor error parameter matrix. Calculate the ratio error and angle error of each transformer based on the transformer error parameter matrix.

2. The online error monitoring method for current transformers based on elastic network regression according to claim 1, characterized in that, The n-dimensional overdetermined regression equation system is as follows: ;in For input Secondary current data matrix The phasor value of the m-th group of secondary fundamental current of the n-th current transformer; Let be the error parameter matrix of the current transformer to be determined, where the elements are complex numbers. The error parameter of the first current transformer. Let n be the error parameter of the nth current transformer. , Let the ratio difference be that of the i-th mutual inductor. Let i be the angle difference of the i-th mutual inductor. Let be the residual of the nth mutual inductor.

3. The online error monitoring method for current transformers based on elastic network regression according to claim 1, characterized in that, The hybrid regularization loss function of the elastic network is: in, For the original least squares loss, Let A be the L1 norm of matrix A. Let A be the L2 norm of matrix A. For regularity, The mixing coefficients are used; the loss function is Q. E When the minimum is reached, the optimal parameter matrix is ​​obtained through regression iteration. A Then, the optimal ratio difference and angle difference can be obtained.

4. The online error monitoring method for current transformers based on elastic network regression according to claim 1, characterized in that, The coordinate descent method includes the following steps: Input the secondary current data matrix I2, the pre-selected canonical intensity γ, and the mixing coefficient α; Initialize the regression parameter matrix A; Iterative parameter update: With other parameters fixed, solve for the value of aᵢ that minimizes the hybrid regularization loss function of the elastic network element by element; when The iteration stops when the value is less than the set threshold.

5. The online error monitoring method for current transformers based on elastic network regression according to claim 3, characterized in that, The pre-selection of the regularization intensity γ and the mixing coefficient α is carried out using the 5-fold cross-validation method.

6. The online error monitoring method for current transformers based on elastic network regression according to claim 1, characterized in that, The pre-selection of the canonical intensity γ and the mixing coefficient α includes the following steps: Divide the data in the secondary current data matrix I2 into 5 equal parts; For each candidate (γ, α) combination, four sets of data are selected in turn to train the model, and the mean squared error is calculated from the remaining set. The (γ, α) value with the smallest mean square error is selected as the chosen regularization strength and mixing coefficient.

7. An online error monitoring system for current transformers based on elastic network regression, characterized in that, include: The modeling module is used to collect m sets of secondary fundamental current phasor values ​​from n current transformers and construct an m×n dimensional secondary current data matrix. An n-dimensional overdetermined regression equation system is established based on the secondary current data matrix, the error parameter matrix of the transformer to be determined, and the residual vector. The loss function building module is used to construct the hybrid regularization loss function for elastic networks; The solution module is used to solve the overdetermined regression equations using the coordinate descent method to obtain the mutual inductor error parameter matrix. The calculation module is used to calculate the ratio difference and angle difference of each transformer based on the transformer error parameter matrix.

8. An electronic device, characterized in that, include: At least one processor; as well as, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, which, when executed by the at least one processor, enables the at least one processor to perform the online monitoring method for current transformer errors based on resilient network regression as described in any one of claims 1 to 6.

9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the online monitoring method for current transformer errors based on elastic network regression as described in any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the online monitoring method for current transformer errors based on elastic network regression as described in any one of claims 1 to 6.

Images