Power distribution network fault prediction positioning method and system
By installing sensors on distribution network nodes, collecting and processing power grid data, and using fault prediction models and proboscoon optimization algorithms, accurate prediction and positioning of distribution network faults is achieved, solving the problems of inaccurate positioning and time-consuming in traditional methods, and improving fault handling efficiency.
Patent Information
- Application Number
- CN202411953700.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-27
- Publication Date
- 2025-06-06
AI Technical Summary
The traditional distribution network fault positioning method has problems such as inaccurate positioning and long-term consumption, which cannot meet the needs of modern power systems for fault prediction and positioning.
A method for fault prediction and positioning of distribution networks is designed. By installing sensors on distribution network nodes, grid data is collected in real time, preprocessing and feature extraction is performed, prediction and analysis is performed using fault prediction models, and fault location is determined using prolonged raccoon optimization algorithm.
It realizes accurate prediction and positioning of distribution network faults, improves fault handling efficiency, can detect faults in a timely manner and shortens fault handling time.
Smart Images
Figure CN120103046A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of power systems, and in particular to a method and system for predicting and locating faults in a distribution network. Background Art
[0002] The distribution network is the part of the power system that directly faces users. Its reliability and stability are crucial to the quality of electricity consumption of users. However, due to the complex structure of the distribution network and the harsh operating environment, faults occur from time to time. Traditional fault location methods mainly rely on manual line patrol and equipment such as fault indicators, which have problems such as inaccurate positioning and long time consumption. Therefore, a more accurate and rapid distribution network fault prediction and location method and system is needed. Summary of the invention
[0003] The purpose of the present invention is to solve the above problems and to design a distribution network fault prediction and location method and system.
[0004] A first aspect of the present invention provides a distribution network fault prediction and positioning method, the distribution network fault prediction and positioning method comprising the following steps:
[0005] Installing sensors at each node of the distribution network to collect initial grid data in real time, wherein the initial grid data includes at least voltage, current and power;
[0006] Preprocessing the collected initial power grid data, and extracting features from the preprocessed initial power grid data to obtain target power grid data;
[0007] Inputting the target power grid data into a fault prediction model, performing prediction analysis on the target power grid data through the fault prediction model, and determining whether there is a fault according to the prediction result;
[0008] If a fault exists, the coati optimization algorithm is used to determine the fault location, and the obtained fault information is sent to the terminal, wherein the fault information at least includes the fault type, the fault location and the fault time.
[0009] Optionally, in a first implementation of the first aspect of the present invention, preprocessing the collected initial power grid data includes:
[0010] Acquire the initial power grid data collected by the sensor, set the sampling frequency of the initial power grid data to s, use the cubic spline interpolation algorithm to process the initial power grid data, eliminate gross errors and supplement missing data;
[0011] The initial power grid data is processed in layers using the wavelet transform method to obtain the detailed signals of different layers of the initial power grid data and the trend signals of the monitoring points;
[0012] Calculate the frequency band range after wavelet transform stratification, and divide the frequency band into three types: high frequency, medium frequency, and low frequency:
[0013]
[0014] In the formula, f s represents the sampling frequency of the initial power grid data, j is the jth layer after wavelet transform decomposition;
[0015] The initial power grid data is restored by using inverse wavelet transform to obtain the restored initial power grid data. The gross errors in the restored initial power grid data are identified based on the 3 times mean error as the limit value. The gross error data are supplemented and repaired by using the neighbor interpolation method to obtain the new initial power grid data.
[0016] The new initial power grid data is layered again to obtain high-frequency, medium-frequency, and low-frequency transformation coefficients. The high-frequency coefficient part is filtered using the adaptive Kalman filter algorithm, the medium-frequency coefficient is corrected in the form of mean error, and the low-frequency coefficient is filtered using the interactive average algorithm to remove noise. After the transformation coefficients of each level are corrected, the corrected coefficients are used to restore the new initial power grid data to complete the preprocessing.
[0017] Optionally, in a second implementation of the first aspect of the present invention, the step of extracting features from the preprocessed initial power grid data to obtain target power grid data includes:
[0018] The preprocessed initial power grid data is obtained, and the m principal components in the preprocessed initial power grid data are formed into the following matrix:
[0019]
[0020] Where, X ij It represents the element in the i-th row and j-th column of the preprocessed initial power grid data. The m×n elements contained in the matrix indicate that the matrix is composed of n elements with m features.
[0021] Standardize the data in matrix X to get standardized data X * , using X rotate =Q T X for standardized data X * The correlation coefficient matrix of is decomposed by eigenvalue and the data is rotated and projected onto the principal component axis, where Q is the standardized data X * The orthogonal matrix obtained by eigenvalue decomposition of the correlation coefficient matrix, Q T is the transposed matrix of Q;
[0022] Calculate the cumulative contribution rate of the principal component, and assume that the eigenvalue obtained after eigenvalue decomposition is λ 1 ,λ 2,…,λ h , arranged in order from large to small, the cumulative contribution rate of the principal component C k The expression is:
[0023]
[0024] In the formula, k represents the number of principal components selected;
[0025] According to the calculated cumulative contribution rate of the principal components, the number of principal components k to be retained is determined, and the reduced-dimensional data matrix with a dimension of m×k is obtained, thereby obtaining the target power grid data.
[0026] Optionally, in a third implementation method of the first aspect of the present invention, the fault prediction model structure consists of an input layer, a membership function generation layer, an inference layer, a normalization layer and an output layer. The target power grid data is input into the fault prediction model, the target power grid data is fuzzy processed through the membership function generation layer, the fuzzy membership value corresponding to the target power grid data is calculated, the inference layer is used to infer the fuzzy processed target power grid data to obtain a fitness value for fault identification, the normalization layer is used to normalize the fitness value for fault identification, and the output layer is used to output the prediction result of the fault prediction model.
[0027] Optionally, in a fourth implementation of the first aspect of the present invention, the reasoning layer calculates a fitness value for fault identification according to an input membership value:
[0028]
[0029] In the formula, ω k represents the kth fitness value of the automatic identification of short-circuit faults in distribution networks; represents two different degrees of membership.
[0030] Optionally, in a fifth implementation of the first aspect of the present invention, if a fault exists, the method of using a coati optimization algorithm to determine the fault location and sending the obtained fault information to a terminal includes:
[0031] Randomly generate a group of initial coati individuals, each of which represents a possible fault location, and generate an initial population, including the population size, the maximum number of iterations, and the scope of the search space;
[0032] Individuals randomly explore new positions in the search space, move closer to the best individual in the current population, and exchange information with other individuals to share their positions and fitness values;
[0033] For each updated individual, calculate its corresponding objective function value. If the individual's objective function value is better than the current optimal solution, update the optimal solution and determine whether the iteration termination condition is met. If the termination condition is met, the iteration stops and outputs the optimal solution as the fault location. Otherwise, continue to the next iteration, where the iteration termination condition is reaching the maximum number of iterations.
[0034] The second aspect of the present invention provides a distribution network fault prediction and positioning system, the distribution network fault prediction and positioning system comprises a data acquisition module, a feature extraction module, a fault prediction module and a fault sending module, wherein:
[0035] A data acquisition module is used to install sensors on each node of the distribution network to collect initial grid data in real time, wherein the initial grid data at least includes voltage, current and power;
[0036] A feature extraction module is used to preprocess the collected initial power grid data and extract features from the preprocessed initial power grid data to obtain target power grid data;
[0037] A fault prediction module is used to input the target power grid data into a fault prediction model, perform prediction analysis on the target power grid data through the fault prediction model, and determine whether there is a fault according to the prediction result;
[0038] The fault sending module is used to determine the fault location by using the coati optimization algorithm if a fault exists, and send the obtained fault information to the terminal, wherein the fault information at least includes the fault type, fault location and fault time.
[0039] Optionally, in a first implementation of the second aspect of the present invention, the feature extraction module includes an acquisition submodule, a standardization processing submodule, a first calculation submodule and a second calculation submodule, wherein:
[0040] The acquisition submodule is used to obtain the preprocessed initial power grid data, and form the following matrix for the m principal components in the preprocessed initial power grid data:
[0041]
[0042] Where, X ij It represents the element in the i-th row and j-th column of the preprocessed initial power grid data. The m×n elements contained in the matrix indicate that the matrix is composed of n elements with m features.
[0043] The standardization processing submodule is used to standardize the data in the matrix X to obtain the standardized data X * , using X rotate =Q T X for standardized data X *The correlation coefficient matrix of is decomposed by eigenvalue and the data is rotated and projected onto the principal component axis, where Q is the standardized data X * The orthogonal matrix obtained by eigenvalue decomposition of the correlation coefficient matrix, Q T is the transposed matrix of Q;
[0044] The first calculation submodule is used to calculate the cumulative contribution rate of the principal component. The eigenvalue obtained after eigenvalue decomposition is λ 1 ,λ 2 ,…,λ h , arranged in order from large to small, the cumulative contribution rate of the principal component C k The expression is:
[0045]
[0046] In the formula, k represents the number of principal components selected;
[0047] The second calculation submodule is used to determine the number k of principal components to be retained according to the calculated cumulative contribution rate of the principal components, and obtain a data matrix after dimensionality reduction with a dimension of m×k, thereby obtaining target power grid data.
[0048] Optionally, in a second implementation of the second aspect of the present invention, the fault sending module includes a random generation submodule, an information exchange submodule and an iteration submodule, wherein:
[0049] The random generation submodule is used to randomly generate a group of initial coati individuals, each of which represents a possible fault location, and generate an initial population, including the population size, the maximum number of iterations, and the range of the search space;
[0050] The information exchange submodule is used for individuals to randomly explore new positions in the search space, move closer to the best individual in the current population, and exchange information with other individuals to share their positions and fitness values;
[0051] The iteration submodule is used to calculate the corresponding objective function value for each updated individual. If the objective function value of the individual is better than the current optimal solution, the optimal solution is updated and it is determined whether the iteration termination condition is met. If the termination condition is met, the iteration stops and the optimal solution is output as the fault location. Otherwise, the next iteration is continued, where the iteration termination condition is reaching the maximum number of iterations.
[0052] In the technical solution provided by the present invention, sensors are installed on each node of the distribution network to collect initial power grid data in real time; the collected initial power grid data is preprocessed, and feature extraction is performed on the preprocessed initial power grid data to obtain target power grid data; the target power grid data is input into a fault prediction model, and the target power grid data is predicted and analyzed by the fault prediction model, and whether a fault exists is determined according to the prediction result; if a fault exists, the long-nosed coati optimization algorithm is used to determine the fault location, and the obtained fault information is sent to a terminal; the present invention can accurately predict and locate faults, improve fault handling efficiency, detect faults in time, and shorten fault handling time. BRIEF DESCRIPTION OF THE DRAWINGS
[0053] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the following detailed description of the preferred embodiment.The drawings are only for the purpose of illustrating the preferred embodiments and are not to be construed as limiting the invention.
[0054] Figure 1 A schematic diagram of a first embodiment of a distribution network fault prediction and location method provided by an embodiment of the present invention;
[0055] Figure 2 A schematic diagram of a second embodiment of a distribution network fault prediction and location method provided by an embodiment of the present invention;
[0056] Figure 3 A schematic diagram of the structure of a distribution network fault prediction and positioning system provided in an embodiment of the present invention. DETAILED DESCRIPTION
[0057] The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, device, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.
[0058] For ease of understanding, the specific process of the embodiment of the present invention is described below. Figure 1 A schematic diagram of a first embodiment of a distribution network fault prediction and location method provided by an embodiment of the present invention, the method specifically comprises the following steps:
[0059] Step 101: Install sensors on each node of the distribution network to collect initial grid data in real time, wherein the initial grid data includes at least voltage, current and power;
[0060] Step 102: preprocessing the collected initial power grid data, and extracting features from the preprocessed initial power grid data to obtain target power grid data;
[0061] Step 103: input the target power grid data into the fault prediction model, perform prediction analysis on the target power grid data through the fault prediction model, and determine whether there is a fault based on the prediction result;
[0062] Step 104: If a fault exists, the coati optimization algorithm is used to determine the fault location, and the obtained fault information is sent to the terminal, wherein the fault information at least includes the fault type, fault location and fault time.
[0063] In this embodiment, the initial power grid data collected by the sensor is obtained, the sampling frequency of the initial power grid data is set to s, and the initial power grid data is processed by a cubic spline interpolation algorithm to remove gross errors and supplement missing data; the initial power grid data is hierarchically processed by a wavelet transform method to obtain detailed signals of different levels of the initial power grid data and trend signals of monitoring points;
[0064] Calculate the frequency band range after wavelet transform stratification, and divide the frequency band into three types: high frequency, medium frequency, and low frequency:
[0065]
[0066] In the formula, f s represents the sampling frequency of the initial power grid data, j is the jth layer after wavelet transform decomposition;
[0067] The initial power grid data is restored by using inverse wavelet transform to obtain the restored initial power grid data. The gross errors in the restored initial power grid data are identified based on the 3 times mean error as the limit value. The gross error data are supplemented and repaired by using the neighbor interpolation method to obtain the new initial power grid data.
[0068] The new initial power grid data is layered again to obtain high-frequency, medium-frequency, and low-frequency transformation coefficients. The high-frequency coefficient part is filtered using the adaptive Kalman filter algorithm, the medium-frequency coefficient is corrected in the form of mean error, and the low-frequency coefficient is filtered using the interactive average algorithm to remove noise. After the transformation coefficients of each level are corrected, the corrected coefficients are used to restore the new initial power grid data to complete the preprocessing.
[0069] In this embodiment, feature extraction is performed on the preprocessed initial power grid data to obtain target power grid data, including:
[0070] The preprocessed initial power grid data is obtained, and the m principal components in the preprocessed initial power grid data are formed into the following matrix:
[0071]
[0072] Where, X ij It represents the element in the i-th row and j-th column of the preprocessed initial power grid data. The m×n elements contained in the matrix indicate that the matrix is composed of n elements with m features.
[0073] Standardize the data in matrix X to get standardized data X * , using X rotate =Q T X for standardized data X * The correlation coefficient matrix of is decomposed by eigenvalue and the data is rotated and projected onto the principal component axis, where Q is the standardized data X * The orthogonal matrix obtained by eigenvalue decomposition of the correlation coefficient matrix, Q T is the transposed matrix of Q;
[0074] Calculate the cumulative contribution rate of the principal component, and assume that the eigenvalue obtained after eigenvalue decomposition is λ 1 ,λ 2 ,…,λ h , arranged in order from large to small, the cumulative contribution rate of the principal component C k The expression is:
[0075]
[0076] In the formula, k represents the number of principal components selected;
[0077] According to the calculated cumulative contribution rate of the principal components, the number of principal components k to be retained is determined, and the reduced-dimensional data matrix with a dimension of m×k is obtained, thereby obtaining the target power grid data.
[0078] In this embodiment, the fault prediction model structure consists of an input layer, a membership function generation layer, an inference layer, a normalization layer and an output layer. The target power grid data is input into the fault prediction model, the target power grid data is fuzzified by the membership function generation layer, the fuzzy membership value corresponding to the target power grid data is calculated, the inference layer is used to infer the fuzzified target power grid data to obtain the fitness value of fault identification, the normalization layer is used to normalize the fitness value of fault identification, and then the output layer is used to output the prediction result of the fault prediction model.
[0079] In this embodiment, the reasoning layer calculates the fitness value of fault identification according to the input membership value:
[0080]
[0081] In the formula, ω k represents the kth fitness value of the automatic identification of short-circuit faults in distribution networks; represents two different degrees of membership.
[0082] See also Figure 2 , a schematic diagram of a second embodiment of a distribution network fault prediction and location method provided by an embodiment of the present invention, the method comprising:
[0083] Step 201: randomly generate a group of initial coati individuals, each of which represents a possible fault location, and generate an initial population, including a population size, a maximum number of iterations, and a range of a search space;
[0084] Step 202: The individual randomly explores a new position in the search space, approaches the best individual in the current population, and exchanges information with other individuals to share their own position and fitness value;
[0085] Step 203: For each updated individual, calculate its corresponding objective function value. If the objective function value of the individual is better than the current optimal solution, update the optimal solution and determine whether the iteration termination condition is met. If the termination condition is met, the iteration stops and outputs the optimal solution as the fault location. Otherwise, continue to the next iteration, where the iteration termination condition is reaching the maximum number of iterations.
[0086] See also Figure 3 , a schematic diagram of the structure of a distribution network fault prediction and positioning system provided by an embodiment of the present invention, the system includes a data acquisition module, a feature extraction module, a fault prediction module and a fault sending module, wherein,
[0087] A data acquisition module is used to install sensors on each node of the distribution network to collect initial grid data in real time, wherein the initial grid data includes at least voltage, current and power;
[0088] A feature extraction module is used to preprocess the collected initial power grid data and extract features from the preprocessed initial power grid data to obtain target power grid data;
[0089] A fault prediction module is used to input the target power grid data into the fault prediction model, perform prediction analysis on the target power grid data through the fault prediction model, and determine whether there is a fault based on the prediction result;
[0090] The fault sending module is used to determine the fault location by using the coati optimization algorithm if a fault exists, and send the obtained fault information to the terminal, wherein the fault information at least includes the fault type, fault location and fault time.
[0091] In this embodiment, the feature extraction module includes an acquisition submodule, a standardization processing submodule, a first calculation submodule and a second calculation submodule, wherein:
[0092] The acquisition submodule is used to obtain the preprocessed initial power grid data, and form the following matrix for the m principal components in the preprocessed initial power grid data:
[0093]
[0094] Where, X ij It represents the element in the i-th row and j-th column of the preprocessed initial power grid data. The m×n elements contained in the matrix indicate that the matrix is composed of n elements with m features.
[0095] The standardization processing submodule is used to standardize the data in the matrix X to obtain the standardized data X * , using X rotate =Q T X for standardized data X * The correlation coefficient matrix of is decomposed by eigenvalue and the data is rotated and projected onto the principal component axis, where Q is the standardized data X * The orthogonal matrix obtained by eigenvalue decomposition of the correlation coefficient matrix, Q T is the transposed matrix of Q;
[0096] The first calculation submodule is used to calculate the cumulative contribution rate of the principal component. The eigenvalue obtained after eigenvalue decomposition is λ 1 ,λ 2 ,…,λ h , arranged in order from large to small, the cumulative contribution rate of the principal component C k The expression is:
[0097]
[0098] In the formula, k represents the number of principal components selected;
[0099] The second calculation submodule is used to determine the number k of principal components to be retained according to the calculated cumulative contribution rate of the principal components, and obtain a data matrix after dimensionality reduction with a dimension of m×k, thereby obtaining target power grid data.
[0100] In this embodiment, the fault sending module includes a random generation submodule, an information exchange submodule and an iteration submodule, wherein:
[0101] The random generation submodule is used to randomly generate a group of initial coati individuals, each of which represents a possible fault location, and generate an initial population, including the population size, the maximum number of iterations, and the range of the search space;
[0102] The information exchange submodule is used for individuals to randomly explore new positions in the search space, move closer to the best individual in the current population, and exchange information with other individuals to share their positions and fitness values;
[0103] The iteration submodule is used to calculate the corresponding objective function value for each updated individual. If the objective function value of the individual is better than the current optimal solution, the optimal solution is updated and it is determined whether the iteration termination condition is met. If the termination condition is met, the iteration stops and the optimal solution is output as the fault location. Otherwise, the next iteration is continued, where the iteration termination condition is reaching the maximum number of iterations.
[0104] Through the implementation of the above scheme, sensors are installed on each node of the distribution network to collect initial power grid data in real time; the collected initial power grid data is preprocessed, and the features of the preprocessed initial power grid data are extracted to obtain target power grid data; the target power grid data is input into a fault prediction model, and the target power grid data is predicted and analyzed by the fault prediction model, and it is determined whether a fault exists according to the prediction result; if a fault exists, the long-nosed coati optimization algorithm is used to determine the fault location, and the obtained fault information is sent to the terminal; the present invention can accurately predict and locate faults, improve fault handling efficiency, can timely discover faults, and shorten fault handling time.
[0105] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and descriptions are only preferred examples of the present invention and are not intended to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.
Claims
1. A distribution network fault prediction and location method, characterized in that: The distribution network fault prediction and positioning method comprises the following steps: Installing sensors at each node of the distribution network to collect initial grid data in real time, wherein the initial grid data includes at least voltage, current and power; Preprocessing the collected initial power grid data, and extracting features from the preprocessed initial power grid data to obtain target power grid data; Inputting the target power grid data into a fault prediction model, performing prediction analysis on the target power grid data through the fault prediction model, and determining whether there is a fault according to the prediction result; If a fault exists, the coati optimization algorithm is used to determine the fault location, and the obtained fault information is sent to the terminal, wherein the fault information at least includes the fault type, the fault location and the fault time.
2. A distribution network fault prediction and location method according to claim 1, characterized in that: The preprocessing of the collected initial power grid data includes: Acquire the initial power grid data collected by the sensor, set the sampling frequency of the initial power grid data to s, use the cubic spline interpolation algorithm to process the initial power grid data, eliminate gross errors and supplement missing data; The initial power grid data is processed in layers using the wavelet transform method to obtain the detailed signals of different layers of the initial power grid data and the trend signals of the monitoring points; Calculate the frequency band range after wavelet transform stratification, and divide the frequency band into three types: high frequency, medium frequency, and low frequency: In the formula, f s represents the sampling frequency of the initial power grid data, j is the jth layer after wavelet transform decomposition; The initial power grid data is restored by using inverse wavelet transform to obtain the restored initial power grid data. The gross errors in the restored initial power grid data are identified based on the 3 times mean error as the limit value. The gross error data are supplemented and repaired by using the neighbor interpolation method to obtain the new initial power grid data. The new initial power grid data is layered again to obtain high-frequency, medium-frequency, and low-frequency transformation coefficients. The high-frequency coefficient part is filtered using the adaptive Kalman filter algorithm, the medium-frequency coefficient is corrected in the form of mean error, and the low-frequency coefficient is filtered using the interactive average algorithm to remove noise. After the transformation coefficients of each level are corrected, the corrected coefficients are used to restore the new initial power grid data to complete the preprocessing.
3. A distribution network fault prediction and location method according to claim 1, characterized in that: The feature extraction of the preprocessed initial power grid data to obtain target power grid data includes: The preprocessed initial power grid data is obtained, and the m principal components in the preprocessed initial power grid data are formed into the following matrix: Where, X ij It represents the element in the i-th row and j-th column of the preprocessed initial power grid data. The m×n elements contained in the matrix indicate that the matrix is composed of n elements with m features. Standardize the data in matrix X to get standardized data X * , using X rotate =Q T X for standardized data X * The correlation coefficient matrix of is decomposed by eigenvalue and the data is rotated and projected onto the principal component axis, where Q is the standardized data X * The orthogonal matrix obtained by eigenvalue decomposition of the correlation coefficient matrix, Q T is the transposed matrix of Q; Calculate the cumulative contribution rate of the principal component, and assume that the eigenvalues obtained after eigenvalue decomposition are λ1,λ2,…,λ h , arranged in order from large to small, the cumulative contribution rate of the principal component C k The expression is: In the formula, k represents the number of principal components selected; According to the calculated cumulative contribution rate of the principal components, the number of principal components k to be retained is determined, and the reduced-dimensional data matrix with a dimension of m×k is obtained, thereby obtaining the target power grid data.
4. A distribution network fault prediction and location method according to claim 1, characterized in that: The fault prediction model structure consists of an input layer, a membership function generation layer, an inference layer, a normalization layer and an output layer. The target power grid data is input into the fault prediction model, the target power grid data is fuzzified by the membership function generation layer, the fuzzy membership value corresponding to the target power grid data is calculated, the inference layer is used to infer the fuzzified target power grid data to obtain the fitness value of fault identification, the normalization layer is used to normalize the fitness value of fault identification, and then the output layer is used to output the prediction result of the fault prediction model.
5. A distribution network fault prediction and location method as claimed in claim 4, characterized in that: The inference layer calculates the fitness value of fault identification based on the input membership value: In the formula, ω k represents the kth fitness value of the automatic identification of short-circuit faults in distribution networks; represents two different degrees of membership.
6. A distribution network fault prediction and location method according to claim 1, characterized in that: If a fault exists, the raccoon optimization algorithm is used to determine the fault location, and the obtained fault information is sent to the terminal, including: Randomly generate a group of initial coati individuals, each of which represents a possible fault location, and generate an initial population, including the population size, the maximum number of iterations, and the range of the search space; Individuals randomly explore new positions in the search space, move closer to the best individual in the current population, and exchange information with other individuals to share their positions and fitness values; For each updated individual, calculate its corresponding objective function value. If the individual's objective function value is better than the current optimal solution, update the optimal solution and determine whether the iteration termination condition is met. If the termination condition is met, the iteration stops and outputs the optimal solution as the fault location. Otherwise, continue to the next iteration, where the iteration termination condition is reaching the maximum number of iterations.
7. A distribution network fault prediction and positioning system, characterized in that: The distribution network fault prediction and positioning system includes a data acquisition module, a feature extraction module, a fault prediction module and a fault sending module, wherein: A data acquisition module is used to install sensors on each node of the distribution network to collect initial grid data in real time, wherein the initial grid data at least includes voltage, current and power; A feature extraction module is used to preprocess the collected initial power grid data and extract features from the preprocessed initial power grid data to obtain target power grid data; A fault prediction module is used to input the target power grid data into a fault prediction model, perform prediction analysis on the target power grid data through the fault prediction model, and determine whether there is a fault according to the prediction result; The fault sending module is used to determine the fault location by using the coati optimization algorithm if a fault exists, and send the obtained fault information to the terminal, wherein the fault information at least includes the fault type, fault location and fault time.
8. A distribution network fault prediction and positioning system as claimed in claim 7, characterized in that: The feature extraction module includes an acquisition submodule, a standardization processing submodule, a first calculation submodule and a second calculation submodule, wherein: The acquisition submodule is used to obtain the preprocessed initial power grid data, and form the following matrix for the m principal components in the preprocessed initial power grid data: Where, X ij It represents the element in the i-th row and j-th column of the preprocessed initial power grid data. The m×n elements contained in the matrix indicate that the matrix is composed of n elements with m features. The standardization processing submodule is used to standardize the data in the matrix X to obtain the standardized data X * , using X rotate =Q T X for standardized data X * The correlation coefficient matrix of is decomposed by eigenvalue and the data is rotated and projected onto the principal component axis, where Q is the standardized data X * The orthogonal matrix obtained by eigenvalue decomposition of the correlation coefficient matrix, Q T is the transposed matrix of Q; The first calculation submodule is used to calculate the cumulative contribution rate of the principal component. The eigenvalues obtained after eigenvalue decomposition are λ1,λ2,…,λ h , arranged in order from large to small, the cumulative contribution rate of the principal component C k The expression is: In the formula, k represents the number of principal components selected; The second calculation submodule is used to determine the number k of principal components to be retained according to the calculated cumulative contribution rate of the principal components, and obtain a data matrix after dimensionality reduction with a dimension of m×k, thereby obtaining target power grid data.
9. A distribution network fault prediction and positioning system as claimed in claim 7, characterized in that: The fault sending module includes a random generation submodule, an information exchange submodule and an iteration submodule, wherein: The random generation submodule is used to randomly generate a group of initial coati individuals, each of which represents a possible fault location, and generate an initial population, including the population size, the maximum number of iterations, and the range of the search space; The information exchange submodule is used for individuals to randomly explore new positions in the search space, move closer to the best individual in the current population, and exchange information with other individuals to share their positions and fitness values; The iteration submodule is used to calculate the corresponding objective function value for each updated individual. If the objective function value of the individual is better than the current optimal solution, the optimal solution is updated and it is determined whether the iteration termination condition is met. If the termination condition is met, the iteration stops and the optimal solution is output as the fault location. Otherwise, the next iteration is continued, where the iteration termination condition is reaching the maximum number of iterations.