Distribution line single-phase earth fault section prediction method and system based on transient signals

Abstract

The invention discloses a distribution line single-phase earth fault section prediction method and system based on transient signals. And carrying out multi-scale analysis on the transient current signal or the transient voltage signal by using wavelet transform, extracting a wavelet coefficient reflecting fault features and energy thereof as feature values, inputting the feature values into a pre-trained neural network model, carrying out distributed training, and predicting to obtain a fault section. According to the method, the fault positioning accuracy is remarkably improved, false alarm and missing alarm are reduced, the fault positioning time is greatly shortened, the fault response speed of the power system is improved, and the power failure time and economic loss are reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of power distribution network line faults, and in particular to a method and system for predicting a single-phase grounding fault section of a power distribution line based on transient signals. Background Art

[0002] The distribution network refers to the power grid that receives electric energy from the transmission network or regional power plants and distributes it locally or step by step according to voltage to various users through distribution facilities. It is composed of overhead lines, cables, towers, distribution transformers, disconnectors, reactive power compensators and some ancillary facilities, and plays an important role in distributing electric energy in the power grid.

[0003] Single-phase grounding faults in distribution lines are a common type of fault in power systems, and their rapid and accurate positioning is crucial to ensuring the continuity and stability of power supply. Traditional fault section prediction methods mainly rely on steady-state signal analysis, fault recording devices, traveling wave methods and other technical means. However, these methods often have problems such as insufficient positioning accuracy, slow response speed, and poor adaptability when faced with complex network structures, transient fault characteristics, and environmental factors, which prolongs the troubleshooting time and affects the quality of power supply services. Summary of the invention

[0004] The purpose of the present invention is to provide a method and system for predicting the single-phase grounding fault section of a distribution line based on transient signals. The present invention significantly improves the accuracy of fault location and reduces false alarms and missed alarms through the powerful nonlinear fitting ability of the deep learning model and the fine features provided by wavelet analysis. The present invention has good fault section prediction accuracy when facing complex network structures, transient fault characteristics and environmental factors.

[0005] To achieve this purpose, the present invention provides a method for predicting a single-phase grounding fault section of a power distribution line based on transient signals, which comprises the following steps:

[0006] Collect transient current signals or transient voltage signals generated by a distribution line when a single-phase grounding fault occurs;

[0007] The collected transient current signal or transient voltage signal is subjected to wavelet transform multi-scale analysis to obtain the wavelet coefficients of the transient current signal or transient voltage signal at each wavelet transform scale and time axis or space axis position;

[0008] Calculate the energy of the wavelet coefficients of the transient current signal or transient voltage signal at each scale and time axis or space axis position, and determine and select the maximum energy in all wavelet transform scales as the main eigenvalue of the transient current signal or transient voltage signal;

[0009] The main characteristic values ​​of transient current signals or transient voltage signals are input into multiple trained neural network models for distributed training to obtain corresponding multiple fault section prediction results, and the multiple fault section prediction results are weighted averaged to obtain the fused fault section prediction results.

[0010] A distribution line single-phase grounding fault section prediction system based on transient signals, which includes a signal acquisition module, a wavelet analysis module, a feature calculation module and multiple neural network modules:

[0011] A signal acquisition module, used to acquire transient current signals or transient voltage signals generated when a single-phase grounding fault occurs in a distribution line;

[0012] A wavelet analysis module, used to perform wavelet transform multi-scale analysis on the transient current signal or transient voltage signal collected by the signal acquisition module using wavelet transform, to obtain the wavelet coefficients of the transient current signal or transient voltage signal at each wavelet transform scale and time axis or space axis position;

[0013] A feature calculation module, used to calculate the energy of the wavelet coefficients of the transient current signal or transient voltage signal obtained by the wavelet analysis module at each scale and time axis or space axis position at each wavelet transform scale, and determine and select the maximum energy among all wavelet transform scales as the main feature value of the transient current signal or transient voltage signal;

[0014] Multiple neural network modules, each neural network module is used to receive the main characteristic values ​​of the transient current signal or the transient voltage signal calculated by the characteristic calculation module, and use the trained neural network model in the corresponding neural network module to perform distributed training, and each neural network module obtains a corresponding fault section prediction result.

[0015] Beneficial effects of the present invention:

[0016] A method and system for predicting single-phase grounding fault sections of distribution lines based on transient signals, using wavelet transform to perform multi-scale analysis on transient current signals or transient voltage signals, and extracting wavelet coefficients and their energy that reflect fault characteristics. This method can effectively capture the transient characteristics of the fault moment, and can more accurately identify high-frequency fluctuation information unique to the fault than traditional analysis methods based on steady-state signals, and construct a deep learning model, such as a deep feedforward neural network, to learn and predict the characteristic values ​​obtained by wavelet analysis, and output the fault section. Through the powerful nonlinear fitting ability of the deep learning model, combined with the fine features provided by wavelet analysis, the accuracy of fault location is significantly improved, and false alarms and missed alarms are reduced;

[0017] Adopting a distributed processing architecture, multiple intelligent terminals process transient signals in parallel, and the central processor performs weighted average fusion of the prediction results of multiple terminals. The intelligent terminals independently complete data processing and prediction, which reduces the burden on the central processor. The central processor only needs to fuse the prediction results, simplifies the system structure, greatly shortens the time of fault location, improves the fault response speed of the power system, and helps to reduce power outage time and economic losses;

[0018] The central processing unit dynamically adjusts the weight of the intelligent terminal according to its historical prediction accuracy to achieve adaptive fusion of prediction results. It builds a fusion mechanism that can dynamically adjust the weight according to the terminal prediction performance, ensuring that the fusion results can fully utilize the advantages of each terminal, enhancing the system's adaptability to complex fault scenarios and prediction stability. Even when the prediction performance of some terminals declines, the overall fault location performance can be ensured by adjusting the weight. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 It is a block diagram of a method for predicting a single-phase grounding fault section of a power distribution line based on transient signals of the present invention;

[0020] Figure 2 It is a system structure diagram of the present invention. DETAILED DESCRIPTION

[0021] The present invention is further described in detail below with reference to the accompanying drawings and specific embodiments:

[0022] like Figure 1 A method for predicting a single-phase grounding fault section of a distribution line based on transient signals is shown, which comprises the following steps:

[0023] S1, collecting transient current signals or transient voltage signals generated by a distribution line when a single-phase grounding fault occurs;

[0024] S2, performing multi-scale analysis on the transient current signal or transient voltage signal collected in step S1 by using wavelet transform, and obtaining wavelet coefficients of the transient current signal or transient voltage signal at each wavelet transform scale and time axis or space axis position;

[0025] S3, calculating the energy of the wavelet coefficients of the transient current signal or transient voltage signal obtained in step S2 at each scale and time axis or space axis position at each wavelet transform scale, determining and selecting the maximum energy at the wavelet transform scale as the main eigenvalue of the transient current signal or transient voltage signal;

[0026] S4, inputting the main characteristic values ​​of the transient current signal or transient voltage signal obtained in step S3 into the trained multi-layer feedforward neural network model to predict the fault section;

[0027] S5, the trained multi-layer feedforward neural network model adopts distributed training to obtain the corresponding multiple fault section prediction results, and the multiple fault section prediction results are weighted averaged to obtain the fused fault section prediction result.

[0028] A method for predicting a single-phase grounding fault section of a distribution line based on transient signals, wherein the step S1 collects a transient current signal or a transient voltage signal generated by the distribution line when a single-phase grounding fault occurs, and the method is as follows:

[0029] High-precision transient signal sensors, such as current transformers (CT) or voltage transformers (PT), are installed in the distribution lines to monitor and record the transient current signals or transient voltage signals generated at the moment of single-phase grounding faults in real time. These sensors should have fast response capabilities to capture the transient process of fault occurrence. The collected transient current signals or transient voltage signals are stored in digital form for subsequent computer processing.

[0030] The purpose of step S1 is that when a single-phase grounding fault occurs in the distribution line, the system needs to use a transient signal sensor to promptly capture the transient current or voltage signal generated. This design ensures a rapid response and records these transient information, providing raw data for subsequent analysis.

[0031] A method for predicting single-phase grounding fault section of a distribution line based on transient signals, wherein the wavelet transform in step S2 is implemented as follows:

[0032] (1) Select wavelet basis functions suitable for transient signal analysis, such as Morlet wavelet, Daubechies wavelet or other wavelets with time-frequency localization characteristics;

[0033] (2) Setting the scale sequence of wavelet transform to ensure that the frequency domain range of transient signals can be covered;

[0034] The scale sequence of wavelet transform is to ensure that the entire frequency domain of transient signals can be covered, including all frequency components contained in the signal. There is no fixed numerical range for the setting of specific scale sequence, because it depends on the characteristics of the actual transient signal, the wavelet basis function used, and the specific needs of the analysis.

[0035] The methods to ensure coverage of the frequency domain range of transient signals include: ① analyzing the spectral characteristics of transient signals of distribution lines to understand the range of their main frequency components; ② selecting basis functions with time-frequency localization characteristics; ③ setting the scale parameter a (scale factor) of the wavelet transform according to the spectral distribution of the signal and the selected wavelet basis function, so that it gradually increases from a small value to cover a wide range from high frequency to low frequency; ④ calculating the results after wavelet transform to verify whether all key frequency components of the signal can be effectively captured, and the scale sequence can also be adjusted until all key frequency bands are included.

[0036] (3) Perform discrete wavelet transform on the transient signal collected in step 1 to obtain wavelet coefficients at various scales and positions.

[0037] A method for predicting single-phase grounding fault sections of distribution lines based on transient signals. In S2, wavelet transform is used to perform multi-scale analysis on the transient current signal or transient voltage signal collected in step 1 to obtain the wavelet coefficient of the transient current signal or transient voltage signal. The formula is expressed as follows:

[0038]

[0039] Where x(t) represents the transient current signal or transient voltage signal, ψ * (t) is the complex conjugate of the wavelet basis function, a is the scale parameter, and b is the translation parameter.

[0040] The specific scale selection is determined by the characteristics of the transient signal and the frequency range to be analyzed. Generally, some values ​​of the scale parameter a are selected to cover all or key frequency bands of the signal. These scale values ​​can be continuous or discrete and are determined according to the actual application scenario and signal characteristics. For example, a scale sequence that gradually increases from small to large is selected so that everything from high-frequency details to low-frequency trends can be analyzed. There is no fixed value for the specific scale selection, which is determined by the signal analysis requirements and the characteristics of the wavelet transform.

[0041] A method for predicting a single-phase grounding fault section of a distribution line based on transient signals, which calculates the energy of the wavelet coefficients of the transient current signal or transient voltage signal obtained in step S2 at each scale and time axis or space axis position, is expressed as follows:

[0042] Ej=l∑|X[j,l]| 2 (2)

[0043] Among them, Ej is the energy of the jth scale, and X[j,l] is the wavelet coefficient of the corresponding scale and position;

[0044] The purpose of step S2 is to perform wavelet transform analysis, use wavelet transform to perform multi-scale analysis on the collected transient signal, decompose the signal at different scales, and extract fault characteristics. This design can effectively separate the high-frequency fluctuation information unique to the fault, help distinguish fault modes, and improve the accuracy and pertinence of the analysis.

[0045] In the present invention, real-time meteorological data (such as temperature, humidity, wind speed, etc.), line load conditions, equipment health status and other information are introduced as auxiliary features during the fault diagnosis process. These environmental and state parameters can be obtained in real time through the Internet of Things technology and input into the neural network model together with the transient signal characteristics, so that the prediction model can more accurately reflect the impact of environmental factors on fault performance and location.

[0046] A method for predicting a single-phase grounding fault section of a distribution line based on transient signals. In step S3, the specific method for judging and selecting the energy on the largest scales as the main characteristic values ​​of the transient current signal or the transient voltage signal is:

[0047] Calculate the scale energy and sort it: First, calculate the energy at each scale according to formula (2), then sort the energies of each scale in descending order, and select the largest energy values ​​Ej from the sorted energy sequence. max As the main characteristic value of transient current signal or transient voltage signal.

[0048] The purpose of step S3 is to calculate the eigenvalue, which is the energy of the wavelet coefficients obtained in step 2 at each scale, as the eigenvalue of the transient signal. This design can extract the information that best characterizes the fault characteristics and provide discriminative input for subsequent machine learning models, reflecting the intensity distribution and frequency composition of the signal, especially the energy concentration of the frequency band where the fault characteristics are located.

[0049] A method for predicting a single-phase grounding fault section of a distribution line based on transient signals. In step S4, the neural network model adopts a multi-layer feedforward neural network, including at least one hidden layer and an output layer, wherein the output layer is used to predict the fault section. The training data set of the neural network model includes transient signal feature values ​​of various types of single-phase grounding faults and corresponding fault section labels. The training of the neural network model adopts a back propagation algorithm, and the optimization goal is to minimize the prediction error.

[0050] A method for predicting a single-phase grounding fault section of a distribution line based on transient signals, in step S4, the formula is expressed as:

[0051] y=f(W2.σ(W1.Ej max +b1)+b2), (3)

[0052] Among them, W1 is the weight matrix of the hidden layer, W2 is the weight matrix of the output layer, b1 and b2 are the corresponding bias items, b1 is the bias item of the hidden layer weight matrix, b2 is the bias item of the output layer weight matrix, σ is the activation function, f is the output layer activation function, and y is the predicted fault section.

[0053] The purpose of step S4 is to use neural networks to predict the fault section. The characteristic values ​​calculated in step 3 are input into the trained multi-layer feedforward neural network model. The model predicts the specific location of the fault by learning the relationship between the fault characteristics and the fault section. Through distributed training and weighted fusion of prediction results, the positioning speed and accuracy are improved, the false alarm and missed alarm rates are reduced, and the system's adaptability to complex scenarios is optimized. This design combines the efficient feature extraction of wavelet analysis with the precise prediction advantages of deep learning models to achieve fast and accurate prediction of fault sections.

[0054] A method for predicting a single-phase grounding fault section of a distribution line based on transient signals, wherein in step S5, in the distributed processing module, the central processor dynamically adjusts the weight wi of each intelligent terminal according to the historical prediction accuracy of each intelligent terminal, and the central processor performs weighted average fusion on multiple prediction results to obtain a fused prediction result;

[0055] When performing weighted average fusion, the specific formula can be expressed as:

[0056]

[0057] in, represents the prediction result after fusion, N is the number of intelligent terminals involved in the fusion, and w i is the weight of the i-th intelligent terminal, y i is the prediction result of the i-th smart terminal;

[0058] Prediction result y i , for the single-phase grounding fault section prediction system of the distribution line, it usually includes the location information of the fault section, such as the distance measurement value of the fault point, the fault type judgment, or more specific geographical location information. The specific information content depends on the specific design and application requirements of the system, but the core goal is to determine the exact location of the fault for rapid response and maintenance. Therefore, the prediction result of each intelligent terminal may be one or a group of values, representing a quantitative estimate or classification label of the fault section. By weighted fusion of these prediction results, the system aims to obtain a more accurate and reliable fault location output.

[0059] In the present invention, a complete fault diagnosis and decision support system is constructed, including functions such as fault alarm, fault level assessment, fault cause reasoning, fault repair suggestions, etc. For example, based on the predicted fault section, combined with historical fault data, equipment information and operation and maintenance knowledge base, possible fault causes (such as insulation aging, external damage, bird damage, etc.) are inferred, and targeted fault handling suggestions and preventive measures are given.

[0060] In addition to current or voltage transient signals, other types of fault-related signals can also be collected and fused, such as partial discharge signals, acoustic emission signals, optical radiation signals, etc. These multi-source signals can often reflect fault characteristics from different angles. Combined with wavelet transform and other signal processing technologies, multi-modal feature fusion can be achieved to improve the comprehensiveness and reliability of fault identification.

[0061] like Figure 2 As shown, a distribution line single-phase grounding fault section prediction system based on transient signals includes a signal acquisition module, a wavelet analysis module, a feature calculation module and a neural network module:

[0062] A signal acquisition module, used to acquire transient current signals or transient voltage signals generated when a single-phase grounding fault occurs in a distribution line;

[0063] A wavelet analysis module is used to perform wavelet transform multi-scale analysis on the transient current signal or transient voltage signal collected by the signal collection module to obtain the wavelet coefficients of the transient current signal or transient voltage signal at each scale and time axis or space axis position;

[0064] A feature calculation module, used to calculate the energy of the wavelet coefficients of the transient current signal or transient voltage signal obtained by the wavelet analysis module at each scale and time axis or space axis position, and to determine and select the largest energy at several scales as the main feature values ​​of the transient current signal or transient voltage signal;

[0065] The neural network module includes at least one hidden layer and an output layer, and is used to receive the main characteristic values ​​of the transient current signal or the transient voltage signal calculated by the characteristic calculation module, and predict the fault section.

[0066] A single-phase grounding fault section prediction system for distribution lines based on transient signals, which also includes a distributed processing module, the distributed processing module includes multiple intelligent terminals and a central processing unit, the multiple intelligent terminals respectively process transient current signals or transient voltage signals and send their respective prediction results to the central processing unit, and the central processing unit performs weighted average fusion on multiple prediction results to obtain a fused prediction result.

[0067] The contents not described in detail in this specification belong to the prior art known to professional and technical personnel in this field.

Claims

1. A method for predicting single-phase grounding fault section of distribution line based on transient signal, characterized in that: The following steps are involved: Collect transient current signals or transient voltage signals generated by distribution lines when single-phase grounding fault occurs; The collected transient current signal or transient voltage signal is subjected to wavelet transform multi-scale analysis to obtain the wavelet coefficients of the transient current signal or transient voltage signal at each scale and time axis or space axis position; Calculate the energy of the wavelet coefficients of the transient current signal or transient voltage signal at each scale and time axis or space axis position, and determine and select the maximum energy in all scales as the main characteristic value of the transient current signal or transient voltage signal; The main characteristic values ​​of transient current signals or transient voltage signals are input into multiple trained neural network models for distributed training to obtain corresponding multiple fault section prediction results, and the multiple fault section prediction results are weighted averaged to obtain the fused fault section prediction results.

2. A method for predicting single-phase grounding fault section of a distribution line based on transient signals according to claim 1, characterized in that: The specific method of wavelet transform is: Select wavelet basis functions suitable for transient current signal or transient voltage signal analysis; Set the scale sequence of wavelet transform to ensure that the frequency domain range of transient signals can be covered; The collected transient current or voltage signal is subjected to discrete wavelet transform to obtain wavelet coefficients at various scales and positions.

3. A method for predicting single-phase grounding fault section of a distribution line based on transient signals according to claim 2, characterized in that: The specific method of using wavelet transform to perform multi-scale analysis on the collected transient current signal or transient voltage signal to obtain the wavelet coefficient X(a, b) of the transient current signal or transient voltage signal at each scale and position on the time axis or space axis is: Where x(t) represents the transient current signal or transient voltage signal, ψ * (t) is the complex conjugate of the wavelet basis function, a is the scale parameter, and b is the translation parameter.

4. A method for predicting single-phase grounding fault section of a distribution line based on transient signals according to claim 3, characterized in that: The specific method for calculating the energy of the wavelet coefficients of the transient current signal or transient voltage signal at each scale and time axis or space axis position is: Ej=l∑∣X[j,l]∣ 2 (2) Among them, Ej is the energy of the jth scale, and X[j,l] is the wavelet coefficient of the corresponding scale and position.

5. A method for predicting single-phase grounding fault section of a distribution line based on transient signals according to claim 4, characterized in that: The specific method for judging and selecting the maximum energy among all scales as the main characteristic value of the transient current signal or transient voltage signal is: Calculate the energy at each scale according to formula (2), and select the maximum energy value Ej among all scales max , as the main characteristic value of transient current signal or transient voltage signal.

6. A method for predicting single-phase grounding fault section of a power distribution line based on transient signals according to claim 5, characterized in that: The neural network model adopts a multi-layer feedforward neural network, including at least one hidden layer and an output layer, wherein the output layer is used to predict the fault section. The training data set of the neural network model contains the main characteristic values ​​of transient current signals or transient voltage signals of various types of single-phase grounding faults and the corresponding fault section labels. The training of the neural network model adopts a back propagation algorithm, and the optimization goal is to minimize the prediction error.

7. A method for predicting single-phase grounding fault section of a power distribution line based on transient signals according to claim 6, characterized in that: The calculation formula for predicting the fault section in the neural network model is expressed as: y=f(W2.σ(W1.Ej max +b1)+b2), (3) Where W1 is the weight matrix of the hidden layer, W2 is the weight matrix of the output layer, b1 and b2 are the corresponding bias items, b1 is the bias item of the hidden layer weight matrix, b2 is the bias item of the output layer weight matrix, σ is the activation function, f is the output layer activation function, and y is the predicted fault section.

8. A method for predicting single-phase grounding fault section of a power distribution line based on transient signals according to claim 7, characterized in that: During the distributed training process, the weight w of the prediction results fusion is dynamically adjusted according to the historical prediction accuracy of each trained neural network model. i , the prediction results of each trained neural network model are calculated according to the weight w i Perform weighted average fusion to obtain the fused prediction result. When performing weighted average fusion, the specific formula is: in, represents the prediction result of the fault section after fusion, N is the number of trained neural network models involved in the fusion, and w i is the weight of the i-th trained neural network model, y i is the fault section prediction result of the i-th trained neural network model.

9. A distribution line single-phase grounding fault section prediction system based on transient signals, characterized in that: It includes signal acquisition module, wavelet analysis module, feature calculation module and multiple neural network modules: A signal acquisition module, used to acquire transient current signals or transient voltage signals generated when a single-phase grounding fault occurs in a distribution line; A wavelet analysis module, used to perform wavelet transform multi-scale analysis on the transient current signal or transient voltage signal collected by the signal acquisition module using wavelet transform, so as to obtain the wavelet coefficients of the transient current signal or transient voltage signal at each scale and time axis or space axis position; A feature calculation module, used to calculate the energy of the wavelet coefficients of the transient current signal or transient voltage signal obtained by the wavelet analysis module at each scale and time axis or space axis position, and to determine and select the maximum energy among all scales as the main feature value of the transient current signal or transient voltage signal; Multiple neural network modules, each neural network module is used to receive the main characteristic values ​​of the transient current signal or the transient voltage signal calculated by the characteristic calculation module, and use the trained neural network model in the corresponding neural network module to perform distributed training, and each neural network module obtains a corresponding fault section prediction result.

10. A distribution line single-phase grounding fault section prediction system based on transient signals according to claim 9, characterized in that: It also includes distributed processing modules; The distributed processing module performs weighted average fusion on all fault section prediction results to obtain a fused fault section prediction result.

Images