CN-122021273-A - Grounding grid corrosion detection method based on improved teaching and learning optimization algorithm

CN122021273ACN 122021273 ACN122021273 ACN 122021273ACN-122021273-A

Abstract

The invention discloses a grounding grid corrosion detection method based on an improved teaching and learning optimization algorithm, relates to the technical field of grounding grid corrosion detection, and solves the problems that early micro corrosion or complex corrosion forms are difficult to reliably identify and high-precision and high-reliability detection results of power grid operation and maintenance cannot be met in the prior art. According to the invention, an updating mechanism based on dominant discipline reservation is adopted to carry out fine cross reservation on dominant dimensions of optimal and suboptimal individuals, so that the local development capability is enhanced. In addition, the invention introduces a self-adaptive t distribution variation strategy, dynamically adjusts the searching step length according to the iterative process, strengthens global exploration in the early stage, and locally refines the focus in the later stage so as to realize the self-adaptive balance of the searching precision. The multi-strategy collaborative optimization framework not only improves the convergence stability and the solving precision of the inversion algorithm, but also can accurately identify the slight electrical change of the shallow low-resistance mutant, and provides reliable technical guarantee for accurately positioning the corrosion defect of the grounding grid.

Inventors

BIAN MEIHUA
LU ZHANQIANG
DENG SHANQUAN
HE YUYIN
ZHU JUNWEI
CHEN HENG
PENG JIANING
ZHANG XINGSEN
LIU GUICHAN
LI JUNHUA

Assignees

广西电网有限责任公司电力科学研究院

Dates

Publication Date: 20260512
Application Date: 20260119

Claims (9)

1. The grounding grid corrosion detection method based on the improved teaching and learning optimization algorithm is characterized by comprising the following steps of: S1, constructing a one-dimensional forward model of a grounding grid based on a transient electromagnetic method, and obtaining theoretical electromagnetic response data by injecting pulse current into a transmitting coil and receiving an induced electromotive force time domain response signal; S2, constructing an improved teaching and learning optimization algorithm, and performing inversion optimization by taking the fitting error minimization of transient electromagnetic observation data and forward data as a target; s3, positioning the corrosion fracture position of the grounding grid and quantifying the corrosion degree according to the resistivity distribution characteristics obtained by inversion; wherein the improved teaching and learning optimization algorithm comprises the following improvement strategies: (1) Introducing an updating mechanism based on dominant discipline reservation, and performing cross operation on different dimensions of the optimal fitness and suboptimal individuals; (2) And when the optimal fitness is continuously iterated for a plurality of times without lifting, the self-adaptive t distribution variation strategy is adopted, disturbance is applied to the population by utilizing the t distribution of which the degree of freedom parameter is adaptively changed along with the iteration times, and the global and local search precision is balanced.
2. The method for detecting corrosion of a ground network based on an improved teaching and learning optimization algorithm of claim 1, wherein the improved teaching and learning optimization algorithm uses HAMMERSLEY low-variance sequences and an improved reverse learning strategy for population initialization, specifically as follows: For the optimization problem that the D-dimension decision variable and the population size are Np, the first dimension is generated through uniform distribution, and the other dimensions are generated by adopting Van der Corput sequences based on a prime number base, so that initial population individuals with uniform distribution are obtained.
3. The method for detecting corrosion of a ground network based on an improved teaching and learning optimization algorithm of claim 2, wherein HAMMERSLEY sequences are initialized and elite screening is performed: For each individual X i of the initial population, calculating the inverse solution of the initial population in the range of the upper and lower boundaries of the parameters, wherein the calculation formula of the inverse solution is as follows: In the formula, In order to solve the problem in the reverse direction, Is a random number obeying uniform distribution between [1,5], and [ lb, ub ] is the upper and lower bounds of each variable.
4. The method for detecting corrosion of a ground network based on an improved teaching and learning optimization algorithm of claim 1, wherein the update mechanism based on dominant discipline preservation comprises: In each iteration, sequentially performing cross operation on each dimension component of the fitness optimal individual Xbest and the suboptimal individual Xsecond to generate a cross individual Xcross (j), wherein the j dimension component is taken from the j dimension of Xsecond, and the rest dimensions are taken from Xbest; And calculating the fitness of the crossed individuals, if the fitness is superior to the original optimal individuals, updating the optimal individuals by using the crossed individuals, otherwise, keeping the original optimal individuals unchanged.
5. The method for detecting corrosion of a ground network based on an improved teaching and learning optimization algorithm of claim 1, wherein the adaptive t-distribution variation strategy comprises: when the optimal fitness is detected to be not improved in three continuous iterations, activating a mutation operation, wherein a mutation formula is as follows: Where t (iter) represents a t-distributed random number with a degree of freedom parameter related to the current iteration number iter.
6. The method for detecting corrosion of a ground network based on an improved teaching and learning optimization algorithm of claim 1, wherein the fitness function of inversion optimization is: Wherein, the The average absolute percentage error is used for measuring the relative error of the forward value and the observed value; And (3) a data standard deviation term for increasing the weight of the transient electromagnetic early signal.
7. The method for detecting the corrosion of the grounding grid based on the improved teaching and learning optimization algorithm according to claim 1, wherein the one-dimensional forward model is based on a horizontal layered ground model, and the frequency domain electromagnetic field response is solved through a hank transformation numerical value and is converted into a time domain induced electromotive force response through cosine transformation and a broken line approximation method.
8. A computer-readable storage medium, characterized in that the computer-readable storage medium comprises a stored program, wherein the program, when run, controls a device in which the computer-readable storage medium is located to perform the method for detecting corrosion of a ground network based on an improved teaching and learning optimization algorithm according to any one of claims 1 to 7.
9. A processor for running a program, wherein the program when run performs the method for detecting corrosion of a ground network based on the improved teaching and learning optimization algorithm of any one of claims 1 to 7.

Description

Grounding grid corrosion detection method based on improved teaching and learning optimization algorithm Technical Field The invention relates to the technical field of grounding grid corrosion detection, in particular to a grounding grid corrosion detection method based on an improved teaching and learning optimization algorithm. Background The grounding grid is used as a key safety facility of the power system, plays an important role of providing a discharge channel for fault current, lightning current and operation overvoltage, and is a foundation for guaranteeing the safety of personnel and equipment and maintaining the stable operation of the power grid. However, the grounding grid is buried in a complex soil environment for a long time, is subjected to combined action of multiple factors such as physical and chemical properties of soil, stray current and electrochemical corrosion, and the grounding conductor is extremely easy to corrode, so that the sectional area is reduced, the structural strength is reduced, even the grounding conductor is broken, the leakage capacity is seriously weakened, and the reliability of an electric power system is directly threatened. While the traditional detection method is visual but low in efficiency and high in cost, the non-excavation method such as electric network analysis and electrochemical analysis has the problems of power failure operation, complex flow, incapability of accurate positioning and the like. In recent years, a transient electromagnetic method is taken as a geophysical exploration technology which does not need excavation and has no interference to operating equipment, has obvious advantages in shallow underground target body detection, and provides a new technical path for ground network corrosion detection. According to the method, pulse current is applied through the transmitting coil, secondary magnetic field response generated by the induction eddy current of the underground conductor is received, and then the resistivity distribution is inverted to judge the corrosion condition, so that the method is particularly suitable for online detection requirements under the condition that the power grid is not in a power failure state. Although the transient electromagnetic method has application potential in the detection of the grounding grid, the limitation of the inversion algorithm severely limits the detection precision and engineering practicability. The traditional linear inversion algorithm such as least square method, OCCAM inversion, gaussian Newton method and the like is a local optimization method which is essentially dependent on an initial model, and is easy to trap into a local optimal trap by approaching an optimal solution through mathematical iteration. In the method, regularization constraint is added in the calculation process to improve the pathogenicity, but for special targets with shallow burial depths and low resistance mutation such as grounding grids, the inversion effect is often not ideal, and the boundary morphology and the quantized corrosion degree of the corrosion defects are difficult to accurately describe. More importantly, the inversion results are highly sensitive to the choice of initial model, and different initial guesses may lead to distinct inversion conclusions, lacking robustness and stability. The fundamental defect makes the prior art difficult to reliably identify early micro corrosion or complex corrosion in engineering practice, and cannot meet the urgent requirements of power grid operation and maintenance on high-precision and high-reliability detection results, and the urgent need is to break through the constraint of the traditional algorithm framework. In view of this, there is a need for a method of detecting corrosion of a ground network based on improved teaching and learning optimization algorithms. Disclosure of Invention Aiming at the problems that early micro corrosion or complex corrosion forms are difficult to reliably identify in the prior art and the high-precision and high-reliability detection result of the operation and maintenance of a power grid cannot be met, the invention provides a grounding grid corrosion detection method based on an improved teaching and learning optimization algorithm, which can greatly improve the convergence stability and the solving precision of an inversion algorithm based on an optimization framework with multiple strategies and can accurately identify the fine electrical changes of shallow low-resistance mutants, thereby providing reliable technical guarantee for accurate positioning and degree quantification of the corrosion defects of the grounding grid. The specific technical scheme is as follows: A grounding grid corrosion detection method based on an improved teaching and learning optimization algorithm comprises the following steps: S1, constructing a one-dimensional forward model of a grounding grid based on a transient electromagnetic method,