CN-121385399-B - CVT error calculation method and medium based on ALS-MLE

Abstract

The invention provides an ALS-MLE-based CVT error calculation method and medium, relates to the technical field of electric power measurement, and aims to solve the problem that a ratio difference error and an angle difference error influence voltage measurement accuracy due to unsteady state characteristics of a CVT in an electric power system caused by temperature, humidity and load changes. The invention adopts a mixed algorithm combining ALS and MLE for online monitoring and calculating CVT error. Firstly, collecting secondary terminal voltage measurement data of the CVT at a plurality of time points, and analyzing amplitude and phase information. Based on the log-likelihood function, the maximum likelihood function is utilized to estimate, the ALS is combined to carry out iterative optimization, the CVT error parameter is solved, and finally, the amplitude and the phase error of the CVT are corrected simultaneously according to the solving result, so that the CVT is adapted to the dynamic error change in operation. The invention can effectively estimate the error coefficient in the capacitive voltage transformer and correct the error, thereby improving the accuracy of on-line monitoring voltage values.

Inventors

• XIAO XIONG
• WEI XIAOXING
• HOU MINGCHUN
• SUN YONG
• PENG XIANG
• ZHANG HUAICHENG
• LU WENHAO
• Jing Maoheng
• LI CHUN
• WANG JING
• LIN JIACHENG
• LIU JIXIANG
• TAN BINGYUAN
• LUO LEI
• YANG JUN
• WEI HONG
• YAN SHUAI
• XIAO DALI
• SUN XIANHE

Assignees

• 中国南方电网有限责任公司超高压输电公司电力科研院

Dates

Publication Date

20260505

Application Date

20251225

Claims (8)

1. 1. A CVT error calculation method based on ALS-MLE, characterized in that it is applied to at least three CVT of the same phase on the same bus, by calculating CVT error parameters comprising a specific difference and an angular difference by: S1, data acquisition and model construction, namely acquiring secondary voltage data containing amplitude and phase information of the CVT at a plurality of time points, and estimating an initial estimated value of real voltage at each time point and an initial estimated value of noise variance of each CVT based on the secondary voltage data; S2, iterative optimization of joint parameters, namely constructing an objective function based on maximum likelihood estimation by using a log likelihood function by taking an initial estimated value as a starting point, and carrying out iterative solution on a ratio difference value, an angle difference value and a real voltage estimated value of each time point of the CVT by adopting an alternative optimization framework; S3, iteration convergence and output, namely calculating the variation of CVT error parameters after each iteration, executing whether iteration judgment is terminated or not based on the variation, and outputting a final ratio difference value and an angle difference value; s4, error correction, namely performing amplitude and phase error compensation on the subsequent voltage measurement result of the corresponding CVT by utilizing the final ratio difference value and the angle difference value; The step S1 includes the steps of: S1.1, collecting secondary voltage data containing amplitude and phase information of a CVT at a plurality of time points, and calculating an initial estimated value of real voltage at each time point by adopting a tail-biting average method; s1.2, calculating an initial residual error based on an initial estimated value of the real voltage, and estimating an initial estimated value of each CVT noise variance based on a median absolute deviation of the initial residual error; The step S2 includes the steps of: S2.1, fixing a real voltage estimated value, estimating CVT error parameters, and solving a ratio difference value estimated value and an angle difference value estimated value through a minimized objective function, wherein the minimized objective function is as follows: ; s2.2, fixing CVT error parameters, estimating a real voltage estimated value, and updating the real voltage estimated value by solving a linear optimization function, wherein the linear optimization function is as follows: ; Wherein, the Is a ratio difference value estimation value; Is an angular difference value estimation value; is the ratio difference; is the angular difference; the secondary voltage data of the ith transformer at the moment t; The method is characterized in that the method is used for acquiring real voltage at the moment T, j is an imaginary unit, N is the number of in-phase CVT voltage amplitude values, and T represents the total number of time points for collecting data.
2. 2. The ALS-MLE-based CVT error calculation method according to claim 1, wherein in the step S1.1, the specific steps of the tail-biting average method are as follows: let t be measured Personal (S) The voltage amplitude of the same-phase CVT is arranged from small to large Setting the tail cutting proportion Obtaining the initial estimated value of the real voltage at the moment t as Wherein The true amplitude of the voltage at time t, The true phase of the voltage at time t.
3. 3. The CVT error calculation method based on ALS-MLE according to claim 2, wherein the step S1.2 specifically includes: Based on the initial estimated value of the real voltage at the time t, the initial residual error of each CVT is calculated, the noise standard deviation is estimated by using the median absolute deviation, and the initial estimated value of the noise variance of each CVT is calculated.
4. 4. A CVT error calculation method based on ALS-MLE according to claim 3, characterized in that in step S2.1, the solution of the ratio difference and angle difference is: ; Wherein, the As the weight factor of the weight factor, For complex voltage values comprising amplitude and phase, Is the ratio difference; and j is an imaginary number unit, and T is the total number of time points for collecting data.
5. 5. The ALS-MLE-based CVT error calculation method according to claim 4, wherein the step S3 includes the steps of: s3.1, initializing the maximum iteration number and the convergence threshold ; S3.2, updating noise variance estimation: ; Wherein, the The secondary voltage data of the ith transformer at the moment t; is the ratio difference; is the angular difference; J is an imaginary unit, and T represents the total number of time points for collecting data; S3.3, checking convergence, and calculating iterative variation of CVT error parameters If (if) Or the iteration number reaches the maximum iteration number, outputting the calculated ratio difference value Sum angle difference Otherwise, returning to the step S2.2 to continue iteration.
6. 6. The ALS-MLE-based CVT error calculation method according to claim 5, wherein in the step S3.1, the maximum number of iterations is 200-400, and the convergence threshold is 10 -6 ~10 -4 .
7. 7. The ALS-MLE-based CVT error calculation method according to claim 6, wherein in step S3.2, the amount of change is iterated The method comprises the following steps: ; Wherein, the Representing the estimated value of the voltage ratio difference calculated by the ith CVT in the latest iteration; representing the voltage ratio difference estimate calculated in the last iteration for the ith CVT; Representing the voltage angle difference estimated value calculated by the ith CVT in the latest iteration; representing the voltage angle difference estimate calculated in the last iteration of the ith CVT.
8. 8. A computer readable storage medium having stored thereon a computer program, which when executed by a processor, implements a CVT error calculation method based on ALS-MLE as claimed in any one of claims 1 to 7.

Description

CVT error calculation method and medium based on ALS-MLE Technical Field The invention relates to the technical field of electric power measurement, in particular to a CVT (capacitive voltage transformer) error calculation method and medium based on ALS (ADAPTIVE LEAST square method) and MLE (Maximum Likelihood Estimation ). Background In modern power systems, CVT is used as a key device and widely applied to important links such as voltage measurement, relay protection, electric energy measurement and the like. The measurement accuracy is directly related to the accuracy of safe and stable operation, economic dispatch and electric energy metering of the electric power system. However, the CVT is inevitably subject to errors during actual operation, which are affected by various factors. On the one hand, the CVT is complex in internal structure, and is composed of a plurality of components such as a capacitive voltage divider, an intermediate transformer and a compensation reactor, and performance parameters of the components are easy to change in a long-term operation process, so that overall measurement accuracy is affected. For example, the capacitance of the capacitive voltage divider may drift due to environmental factors such as temperature and humidity, and parameters such as core loss and winding resistance of the intermediate transformer may also change with increasing operation time, which may cause deviation between the CVT output voltage and the actual input voltage. On the other hand, the operation conditions of the power system are complex and changeable. The conditions of load fluctuation, system short circuit fault, harmonic interference and the like frequently occur, so that the electromagnetic environment where the CVT is located is extremely bad. Under these complex electromagnetic environments, the CVT may be subjected to various kinds of interference such as electromagnetic coupling and electromagnetic induction, further increasing the possibility of error generation. For example, the presence of harmonics can distort the measurement of the CVT and not accurately reflect the actual voltage value. The conventional CVT error calculation method is mostly based on a simple model and fixed parameters, and it is difficult to comprehensively consider the influence of the above complex factors. In practical application, the calculated error result has larger deviation from the true error, and the ever-increasing high-precision operation requirement of the power system cannot be met. For example, the conventional method may not accurately capture the error change of the CVT under the extreme working condition, thereby affecting the correct action of the relay protection device and threatening the safe and stable operation of the power system. Along with the continuous development of the power system towards large capacity, high voltage and intellectualization, higher requirements are put on the accuracy and the instantaneity of CVT error calculation. The accurate error calculation is not only beneficial to timely finding potential faults of the CVT, maintaining and replacing the CVT in advance and avoiding power failure accidents caused by equipment faults, but also can provide reliable data support for optimal scheduling of the power system and improve the operation efficiency and economic benefit of the power system. Therefore, developing a CVT error calculation method that can adapt to a complex operating environment, has high accuracy, and is highly real-time is an important problem to be solved in the power field. Disclosure of Invention The invention provides a CVT error calculation method and medium based on ALS-MLE, and aims to solve the problem that voltage measurement accuracy is affected by ratio difference error and angle difference error due to unsteady state characteristics of CVT generated by temperature, humidity and load changes in an electric power system. The technical scheme adopted by the invention for solving the technical problems is that the CVT error calculation method based on ALS-MLE is applied to at least three CVTs with the same phase on the same bus, and CVT error parameters comprising specific difference values and angular difference values are calculated through the following steps: S1, data acquisition and model construction, namely acquiring secondary voltage data containing amplitude and phase information of the CVT at a plurality of time points, and estimating an initial estimated value of real voltage at each time point and an initial estimated value of noise variance of each CVT based on the secondary voltage data; S2, iterative optimization of joint parameters, namely constructing an objective function based on maximum likelihood estimation by using a log likelihood function by taking an initial estimated value as a starting point, and carrying out iterative solution on a ratio difference value, an angle difference value and a real voltage estimated value o