CN-122017708-A - Method and device for estimating ratio error of phase-adaptive switching voltage sensor

CN122017708ACN 122017708 ACN122017708 ACN 122017708ACN-122017708-A

Abstract

The application discloses a method and a device for estimating ratio errors of a voltage sensor in stage self-adaptive switching, and relates to the technical field of power transmission system state monitoring, wherein the method comprises the steps of obtaining a historical real voltage value of the voltage sensor to be subjected to ratio error estimation and a historical measurement voltage value of the voltage sensor; the method comprises the steps of constructing a multi-objective ratio error estimation problem model of the voltage sensor based on the historical real voltage value and the historical measured voltage value, solving the multi-objective ratio error estimation problem model by adopting an expensive constraint multi-objective intelligent optimization method based on stage self-adaptive switching to obtain a real voltage estimated value of the voltage sensor, and calculating based on the real voltage estimated value and the historical real voltage value to obtain a ratio error estimated result. The method can improve the accuracy of the ratio error estimation of the voltage sensor.

Inventors

LI QIAN
LIU JIANCHANG
TAN SHUBIN
LIU YUANCHAO

Assignees

东北大学

Dates

Publication Date: 20260512
Application Date: 20260313

Claims (10)

1. A method for estimating ratio error of a phase-adaptively switched voltage sensor, comprising: Acquiring a historical real voltage value of a voltage sensor to be subjected to ratio error estimation and a historical measurement voltage value of the voltage sensor; constructing a multi-objective ratio error estimation problem model of the voltage sensor based on the historical real voltage values and the historical measured voltage values; Solving the multi-objective ratio error estimation problem model by adopting an expensive constraint multi-objective intelligent optimization method based on stage self-adaptive switching to obtain a real voltage estimation value of the voltage sensor; and calculating based on the real voltage estimated value and the historical real voltage value to obtain a ratio error estimated result.
2. The method of claim 1, wherein the constructing a multi-objective ratio error estimation problem model of the voltage sensor based on the historical true voltage values and the historical measured voltage values, comprises: Calculating based on the historical real voltage value and the historical measured voltage value to obtain a historical ratio error of the voltage sensor; Calculating based on the historical ratio error to obtain a ratio error variation; performing model construction based on the historical real voltage value and the historical ratio error to obtain a first objective function aiming at minimizing the sum of time-varying ratio errors of the voltage sensor; Performing model construction based on the historical real voltage value and the ratio error variation to obtain a second objective function which aims at minimizing the variance of the ratio error variation of the voltage sensor; constructing constraint conditions, wherein the constraint conditions comprise a first topological constraint condition, a second topological constraint condition and a time sequence constraint condition; And carrying out model construction based on the first topological constraint condition, the second topological constraint condition, the time sequence constraint condition, the first objective function and the second objective function to obtain a multi-objective ratio error estimation problem model of the voltage sensor.
3. The method of claim 1, wherein the solving the multi-objective ratio error estimation problem model by using the expensive constraint multi-objective intelligent optimization method based on phase adaptive switching to obtain the actual voltage estimation value of the voltage sensor specifically comprises: Uniformly sampling in a preset decision space by using a Latin hypercube sampling method to generate a plurality of solutions, performing real evaluation by using the multi-objective ratio error estimation problem model, and storing the solutions in an archive set; Initializing algorithm parameters and initializing a current population, wherein the algorithm parameters comprise maximum iteration times, consumed function evaluation quantity and an initial evolution stage; judging the evolution type of the current evolution stage to obtain a judgment result; Step four, when the judgment result is that the evolution type is the target optimization stage, performing unconstrained evolution optimization on the current population in the current evolution stage by adopting an NSGA-III algorithm to obtain a current candidate population and a current interpolation point set; Fifthly, when the judgment result is that the evolution type is a constraint optimization stage, performing constraint evolution optimization on the current population in the current evolution stage by adopting an NSGA-III algorithm combined with constraint dominance criteria to obtain the current candidate population and the current interpolation point set; Step six, calculating the measurement index of each solution in the current interpolation point set to obtain a measurement index value, and updating an index vector based on the measurement index value to obtain a current index vector, wherein the measurement index comprises a convergence index, a constraint violation index and a feasible solution proportion index; Step seven, determining an evolution stage of the next iteration round based on the current index vector and a preset stage switching rule, wherein the evolution stage comprises a target optimization stage and a constraint optimization stage; step eight, acquiring a preset number of solutions from the archive set by using an environment selection method in an NSGA-III algorithm to obtain a first population; step nine, updating the current population based on the first population and the current interpolation point set; step ten, updating the archive set based on the current interpolation point set to obtain a current archive set; Step eleven, when the current iteration round is smaller than or equal to the preset iteration round threshold, determining the evolution phase as an optimization phase of the next round, and repeatedly executing the steps three to ten to update the current candidate population, the current interpolation point set, the current index vector, the current population and the current archive set until the updated non-dominant solution in the current archive set is determined as a target solution of an expensive constraint multi-target optimization problem under the condition that the current iteration round is larger than the preset iteration round threshold so as to obtain a real voltage estimated value of the voltage sensor; and step twelve, determining the updated non-dominant solution in the current archive set as a target solution of an expensive constraint multi-target optimization problem so as to obtain a real voltage estimated value of the voltage sensor.
4. The method of claim 3, wherein when the determination result is that the evolution type is the target optimization stage, performing unconstrained evolutionary optimization on the current population in the current evolution stage by using NSGA-III algorithm to obtain the current candidate population, and specifically comprising: step one, when the judgment result is that the evolution type is the target optimization stage, performing intersection and mutation operation on the current population to obtain a first offspring population; Step two, merging the current population and the first offspring population to obtain a first target population; step three, non-dominant sorting is carried out on the first target population, and a first multi-layer non-dominant solution set is obtained; Step four, extracting a preset number of solutions from the first layer of the first multi-layer non-dominant solution set to obtain a first new generation population; Step five, executing step six under the condition that the current iteration times are larger than a preset iteration times threshold, replacing the current population based on the first new generation population under the condition that the current iteration times are smaller than or equal to the preset iteration times threshold, repeatedly executing the step one to the step five to update the first new generation population until the current iteration times are larger than the preset iteration times threshold, and determining the updated first new generation population as the current candidate population; And step six, determining the first new generation population as the current candidate population.
5. The method of claim 4, wherein the performing unconstrained evolutionary optimization on the current population in the current evolutionary stage using NSGA-III algorithm to obtain the current interpolation point set, specifically comprises: generating a set of fixed, uniformly distributed first reference vectors within the predetermined decision space; associating the solutions in the current candidate population with the first reference vector to obtain a first active reference vector set; clustering the first active reference vector sets by adopting a K-means clustering method to obtain K first reference vector clusters; performing index calculation on the solutions in the first candidate population based on a preset convergence and diversity function, a preset convergence and diversity improvement function and a preset expected convergence and diversity improvement function to obtain expected convergence and diversity improvement indexes respectively corresponding to the solutions in the current candidate population; Selecting a solution with the maximum expected convergence and diversity improvement index from each first reference vector cluster to obtain the current interpolation point set; an expensive true evaluation of the solutions in the current set of interpolation points is performed using the first objective function and the second objective function.
6. The method of claim 3, wherein when the determination result is that the evolution type is a constrained optimization stage, performing constrained evolution optimization on the current population in the current evolution stage by using NSGA-III algorithm combined with constraint dominance criteria to obtain the current candidate population, specifically comprising: step one, when the judgment result is that the evolution type is a constraint optimization stage, performing intersection and mutation operation on the current population to obtain a second offspring population; step two, merging the current population and the second offspring population to obtain a second target population; Step three, non-dominant sorting is carried out on the second target population, and a second multi-layer non-dominant solution set is obtained; step four, extracting a preset number of solutions from the first layer of the second multi-layer non-dominant solution set to obtain a second new generation population; step five, executing step six under the condition that the current iteration times are larger than a preset iteration times threshold, replacing the current population based on the second new generation population under the condition that the current iteration times are smaller than or equal to the preset iteration times threshold, repeatedly executing step one to step five to update the second new generation population until the current iteration times are larger than the preset iteration times threshold, and determining the updated second new generation population as the current candidate population; And step six, determining the second new generation population as the current candidate population.
7. The method of claim 6, wherein the constrained evolution optimization is performed on the current population at a current evolutionary stage using an NSGA-III algorithm in combination with constraint governance criteria to obtain the current set of interpolation points, comprising: Generating a set of fixed, uniformly distributed second reference vectors within the predetermined decision space; Associating the solutions in the current candidate population with the second reference vector to obtain a second active reference vector set; Clustering the second active reference vector sets by adopting a K-means clustering method to obtain K second reference vector clusters; performing index calculation on the solutions in the current candidate population based on a preset convergence and diversity function, a preset convergence and diversity improvement function and a preset expected convergence and diversity improvement function to obtain expected convergence and diversity improvement indexes respectively corresponding to the solutions in the current candidate population; Performing index calculation on the solutions in the current candidate population based on a preset feasibility probability function to obtain feasibility probability indexes respectively corresponding to the solutions in the current candidate population; Multiplying the expected convergence and diversity improvement index and the feasibility probability index corresponding to the same solution in the current candidate population to obtain a comprehensive index corresponding to the same solution; Selecting a solution with the maximum comprehensive index from each second reference vector cluster to obtain the current interpolation point set; an expensive true evaluation of the solutions in the current set of interpolation points is performed using the first objective function and the second objective function.
8. A method according to claim 3, wherein said determining the evolutionary phase of the next iteration round based on said current index vector and a predetermined phase switching rule, in particular comprises: Sampling the current index vectors with different index dimensions by adopting a preset sliding window to obtain sampling vectors respectively corresponding to the different index dimensions; comparing the end points of the sampling vector by adopting an end point comparison method to obtain trend comparison results respectively corresponding to different index dimensions; When the trend comparison result is that the convergence index is reduced and the constraint violation index is reduced, determining the evolution stage of the next iteration round as a target optimization stage; When the trend comparison result is that the convergence index is reduced, the constraint violation index is reduced and the feasible solution proportion index is greater than zero, determining the evolution stage of the next iteration round as a target optimization stage; When the trend comparison result is that the convergence index is reduced, the constraint violation index is increased, and the feasible solution proportion index is equal to zero, determining the evolution stage of the next iteration round as a constraint optimization stage; When the trend comparison result is that the convergence index is increased and the constraint violation index is reduced, determining the evolution phase of the next iteration round as a constraint optimization phase; When the trend comparison result is that the convergence index grows, the constraint violation index grows and the feasible solution proportion index is larger than zero, determining the evolution stage of the next iteration round as a target optimization stage; and when the trend comparison result is that the convergence index grows, the constraint violation index grows and the feasible solution proportion index is equal to zero, determining the evolution stage of the next iteration round as a constraint optimization stage.
9. The method of claim 3, wherein the updating the current population based on the first population and the current set of interpolation points, in particular comprises: constructing a first set and a second set of which the initial solution set is empty; Calculating a first convergence index and a first feasibility index of the first population; calculating a second convergence index and a second feasibility index of the union of the first population and the current interpolation point set; When the second convergence index is smaller than the first convergence index or the solution in the current interpolation point set is a non-dominant solution, adding the solution in the current interpolation point set to a first set to update the solution in the first set; when the second feasibility index is smaller than the first feasibility index or the solution in the current interpolation point set is a feasible solution, adding the solution in the current interpolation point set to a second set to update the solution in the second set; Updating the current population based on the first population and the updated solutions in the first set under the condition that the current stage is a target optimization stage; and updating the current population based on the solutions in the first population and the updated second set under the condition that the current stage is a constraint optimization stage.
10. A phase-adaptively switched voltage sensor ratio error estimation device, comprising: The acquisition module is used for acquiring a historical real voltage value of the voltage sensor to be subjected to ratio error estimation and a historical measurement voltage value of the voltage sensor; the construction module is used for constructing a multi-objective ratio error estimation problem model of the voltage sensor based on the historical real voltage value and the historical measurement voltage value; The solving module is used for solving the multi-objective ratio error estimation problem model by adopting an expensive constraint multi-objective intelligent optimization method based on stage self-adaptive switching to obtain a real voltage estimation value of the voltage sensor; and the calculation module is used for calculating based on the real voltage estimated value and the historical real voltage value to obtain a ratio error estimated result.

Description

Method and device for estimating ratio error of phase-adaptive switching voltage sensor Technical Field The invention relates to the technical field of power transmission system state monitoring, in particular to a method and a device for estimating ratio error of a voltage sensor in stage self-adaptive switching. Background In the field of industrial power transmission, voltage sensors (Voltage Transformer, VT) are the basic equipment widely used for measuring voltage values in power transmission system substations. Wherein, the Ratio Error (RE) drift fault can directly affect the accuracy of downstream applications such as relay protection, metering, and the like. Therefore, the estimation of the ratio error for the voltage sensor plays a vital role in the power transmission system. In-situ calibration techniques and data-driven RE estimation methods are the primary RE estimation methods in modern power delivery systems. The field calibration technology is to manually compare the VT to be calibrated with the high-precision standard VT by switching off the power supply. The method is accurate under standard conditions, but needs standard equipment far higher than the VT to be measured, and repeated calibration of tens of VT in the station is high in cost and low in efficiency, and only discrete time-varying RE data can be obtained. Whereas data-driven estimation methods estimate RE by building VT specific models and exploiting electrical relationship constraints. Although the method is low in cost, accuracy and generality are limited by the model, and complex and changeable actual working conditions are difficult to meet. To alleviate the problems in the existing methods, the ratio error estimation problem is converted into an expensive constraint multi-objective optimization problem by considering the difference between the real voltage value and the VT measurement value, and the real voltage value is estimated using an intelligent optimization method. The method is suitable for VT of different types, can provide fault early warning in real time, provides early warning for potential fault VT, and has remarkable application value. Under the background, an efficient and reliable intelligent optimization method is constructed so as to find high-quality solutions meeting all constraint conditions within limited simulation evaluation times, and the solution becomes a core challenge for solving the estimation problem. To address this challenge, agent-assisted evolutionary algorithms (Surrogate-Assisted Evolutionary Algorithms, SAEAs) are the mainstream framework to guide searches by building inexpensive agent models instead of expensive simulations, in combination with multi-stage optimization frameworks. While SAEAs and multi-stage frameworks have met with significant success, most current advanced approaches suffer from a fundamental deficiency in optimization strategies, the lack of perceived and adaptive responses to real-time search conditions. The stage switching strategy of the existing method is mostly static preset (such as switching after fixed evaluation times). The static switching rules cannot be dynamically adjusted according to the real-time change of indexes such as convergence, feasibility and the like in the optimization process, so that the stage switching time is inaccurate, and the distribution efficiency of reducing the expensive evaluation times is reduced. In addition, the feasible domain structures of different optimization problems have huge morphological differences with Pareto fronts, and static switching rules are difficult to adapt to various problem scenes. Disclosure of Invention In view of this, the present invention provides a method and apparatus for estimating ratio error of a voltage sensor with phase adaptive switching, which mainly aims to solve the problem that the ratio error estimation of the voltage sensor with complex nonlinear constraint and unknown boundary structure is inaccurate at present. In order to solve the above problems, the present application provides a method for estimating a ratio error of a voltage sensor in a stage adaptive switching, comprising: Acquiring a historical real voltage value of a voltage sensor to be subjected to ratio error estimation and a historical measurement voltage value of the voltage sensor; constructing a multi-objective ratio error estimation problem model of the voltage sensor based on the historical real voltage values and the historical measured voltage values; Solving the multi-objective ratio error estimation problem model by adopting an expensive constraint multi-objective intelligent optimization method based on stage self-adaptive switching to obtain a real voltage estimation value of the voltage sensor; and calculating based on the real voltage estimated value and the historical real voltage value to obtain a ratio error estimated result. In order to solve the above problems, the present application provides a voltage