Early warning method, system and equipment for grounding risk of distribution line and storage medium

Abstract

The invention discloses a power distribution line grounding risk early warning method, system and device and a storage medium, and relates to the technical field of power operation management and control. The method comprises the following steps: acquiring monitoring data of a tested distribution line; performing empirical mode decomposition on the residual current signal to obtain ground fault signal features; performing residual current signal reconstruction to obtain a reconstructed residual current signal distortion degree; under the condition that the reconstructed residual current signal distortion degree is greater than the preset reconstructed signal distortion degree, determining the tested distribution line as an abnormal distribution line; predicting the grounding resistance of the abnormal distribution line to obtain a grounding resistance prediction value; calculating a grounding risk value through a Bayesian network; and carrying out grounding risk early warning on the abnormal distribution line. According to the invention, the problems of false alarm or missing alarm and incapability of dynamic early warning of current grounding risk early warning are solved, the accuracy of fault detection is improved, and the risk of faults of the distribution line is reduced.

Description

Technical Field

[0001] This invention relates to the field of power operation control technology, specifically to a method, system, device, and storage medium for early warning of grounding risks in power distribution lines. Background Technology

[0002] As a crucial component of the power system, power distribution lines are responsible for transmitting electrical energy from substations to various end-users. Because distribution lines are often spread throughout urban and rural areas, they are frequently affected by environmental and climatic factors, leading to frequent line faults, especially grounding faults, which seriously impact the safe and stable operation of the power system. A grounding fault occurs when current leakage in a power line due to equipment insulation failure or external factors causes the system to ground, potentially leading to short circuits, equipment damage, or even fires and other serious safety accidents. Therefore, timely detection and early warning of grounding faults are critical tasks for ensuring the safe operation of power distribution lines.

[0003] Currently, the detection and early warning technology for grounding faults in power distribution lines mainly relies on traditional current and voltage monitoring techniques. These monitoring technologies typically include grounding protection devices, fault indicators, and monitoring systems installed on the power distribution lines. They determine the presence of grounding faults by collecting parameters such as current and voltage in real time. When the current value is abnormal, an alarm can be issued. Furthermore, the application of modern technologies, such as smart grids and sensor networks, is gradually providing more means for the detection and early warning of grounding faults. By deploying various sensors to collect line status data in real time, and combining this with big data analysis and machine learning techniques, intelligent systems can predict line fault trends and provide early warnings of grounding fault occurrences, enhancing the accuracy and timeliness of fault diagnosis.

[0004] However, current grounding risk early warning methods are susceptible to external interference, resulting in an incomplete reflection of the actual condition of the line, leading to false alarms or missed alarms and affecting the accuracy of fault detection. Furthermore, current grounding risk early warning methods often rely on static threshold settings, lacking dynamic early warning capabilities and the ability to predict changes in grounding resistance in real time. This makes it difficult to take timely and effective preventative measures, thereby increasing the risk of power distribution line faults. Summary of the Invention

[0005] This invention addresses the shortcomings of current grounding risk warning methods, which are susceptible to external interference and lack dynamic warning capabilities. It provides a grounding risk warning method, system, device, and storage medium for power distribution lines, solving the problems of false alarms or missed alarms and the inability to provide dynamic warnings. This improves the accuracy of fault detection and reduces the risk of power distribution line failures.

[0006] The present invention is achieved through the following technical solution.

[0007] In a first aspect, the present invention provides a method for early warning of grounding risk in power distribution lines. The method includes: acquiring monitoring data of the power distribution line under test, wherein the monitoring data includes residual current signals, environmental data, static covariates, equipment health data, and external event data; performing empirical mode decomposition on the residual current signals to obtain ground fault signal characteristics; reconstructing the residual current signals based on the ground fault signal characteristics to obtain the distortion degree of the reconstructed residual current signals; determining the power distribution line under test as an abnormal power distribution line if the distortion degree of the reconstructed residual current signals is greater than a preset reconstructed signal distortion degree; predicting the grounding resistance of the abnormal power distribution line using a time-series fusion transformer based on the environmental data and static covariates to obtain a predicted grounding resistance value; calculating a grounding risk value using a Bayesian network based on the monitoring data and the predicted grounding resistance value; and providing early warning of grounding risk for the abnormal power distribution line based on the grounding risk value.

[0008] In some embodiments, empirical mode decomposition (EMD) is performed on the residual current signal to obtain the ground fault signal characteristics, including: adding multiple sets of Gaussian white noise to the residual current signal to obtain multiple sets of residual current and noise signals; decomposing each set of residual current and noise signals into multiple intrinsic mode function (IMF) components and a single residual component through the EMD; averaging the IMF components and residual components at the same time in the multiple sets of residual current and noise signals to obtain multiple stable IMF components and a single stable residual component; and using the stable IMF component corresponding to the frequency band related to the ground fault among the multiple stable IMF components and the single stable residual component as the ground fault signal characteristics.

[0009] In some embodiments, reconstructing the residual current signal based on the characteristics of the ground fault signal to obtain the distortion degree of the reconstructed residual current signal includes: reconstructing the residual current signal based on the characteristics of the ground fault signal and the stable residual component, and extracting the maximum amplitude change and the maximum frequency change between the reconstructed residual current signal and the residual current signal at the corresponding time; and calculating the distortion degree of the reconstructed residual current signal based on the maximum amplitude change and the maximum frequency change, the maximum amplitude and the maximum frequency in the residual current signal, the amplitude change weight and the frequency change weight.

[0010] In some embodiments, based on the environmental data and the static covariates, a time-series fusion transformer is used to predict the grounding resistance of the abnormal power distribution line to obtain a predicted grounding resistance value. This includes: determining multiple key features of the input residual current signal using a gated residual network based on the environmental data and the static covariates, and determining the weights of each key feature using a first preset activation function; processing the multiple key features and their weights using a Long Short-Term Memory (LSTM) network to generate a hidden state for the predicted grounding resistance value; passing the hidden state to a multi-head self-attention layer, determining the key time steps affecting the predicted grounding resistance value by weighting the hidden state, and outputting an enhanced hidden state; and inputting the enhanced hidden state to the output layer of the LSTM to output the predicted grounding resistance value.

[0011] In some embodiments, calculating the grounding risk value using a Bayesian network based on the monitoring data and the predicted grounding resistance value includes: using the predicted grounding resistance value and the monitoring data as nodes, and connecting each node with directed edges according to the causal relationship between each node to construct a Bayesian network including multiple parent nodes and multiple child nodes; determining the probability distribution of the multiple child nodes under different conditions based on the state of the multiple parent nodes in the Bayesian network using an expert evaluation method; constructing a conditional probability table based on the probability distribution of the multiple child nodes; performing Bayesian inference based on the conditional probability table to calculate the posterior probability of the multiple child nodes; and calculating the grounding risk value of the abnormal power distribution line based on the posterior probability of the multiple child nodes.

[0012] In some embodiments, grounding risk warning is provided for the abnormal power distribution line based on the grounding risk value, including: determining the grounding risk level of the abnormal power distribution line according to the grounding risk value, wherein the grounding risk level includes minor risk, medium risk and severe risk; and providing grounding risk warning for the abnormal power distribution line according to the grounding risk level.

[0013] Secondly, the present invention provides a power distribution line grounding risk early warning system, the system comprising: a monitoring data acquisition module, configured to: acquire monitoring data of the tested power distribution line, wherein the monitoring data includes residual current signal, environmental data, static covariates, equipment health data, and external event data; a ground fault signal feature acquisition module, configured to: perform empirical mode decomposition on the residual current signal to obtain ground fault signal features; a reconstructed residual current signal distortion calculation module, configured to: reconstruct the residual current signal based on the ground fault signal features to obtain the reconstructed residual current signal distortion; and an abnormal power distribution line determination module. The system comprises the following modules: a ground resistance prediction module, used to: determine the tested power distribution line as an abnormal power distribution line when the distortion degree of the reconstructed residual current signal is greater than a preset reconstructed signal distortion degree; a grounding resistance prediction calculation module, used to: predict the grounding resistance of the abnormal power distribution line using a time-series fusion transformer based on the environmental data and static covariates to obtain a predicted grounding resistance value; a grounding risk value calculation module, used to: calculate the grounding risk value using a Bayesian network based on the monitoring data and the predicted grounding resistance value; and a grounding risk warning module, used to: provide a grounding risk warning for the abnormal power distribution line based on the grounding risk value.

[0014] Thirdly, the present invention provides a power distribution line grounding risk early warning device, the device comprising: at least one processor; and at least one memory, the at least one memory being coupled to the at least one processor and storing instructions for execution by the at least one processor, the instructions implementing the method described above when executed by the at least one processor.

[0015] Fourthly, the present invention provides a computer-readable storage medium storing instructions that, when executed by a computer, cause the computer to perform the methods described above.

[0016] Fifthly, the present invention provides a computer program product comprising instructions that, when executed by a computer, cause the computer to perform the methods described above.

[0017] Compared with existing technologies, this invention has the following advantages and beneficial effects: By using empirical mode decomposition to extract the ground fault signal characteristics of residual current signals in the monitoring data of each power distribution line, external noise can be effectively filtered out, reducing false alarms and missed alarms; at the same time, signal reconstruction and distortion calculation further optimize the judgment of fault signals, avoid the influence of external interference, and improve the accuracy of fault detection; by using TFT technology to dynamically predict the grounding resistance, the changing trend of grounding resistance can be tracked in real time, breaking the limitations of static thresholds; combined with Bayesian networks, grounding risks can be dynamically assessed and early warnings can be issued in a timely manner, ensuring that effective preventive measures can be taken when the grounding resistance changes. Attached Figure Description

[0018] To more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly described below. It should be understood that the following drawings only show some embodiments of the present invention and should not be considered as limiting the scope. For those skilled in the art, other related drawings can be obtained from these drawings without creative effort.

[0019] Figure 1 This is a flowchart of a power distribution line grounding risk early warning method according to an embodiment of the present invention.

[0020] Figure 2 This is a flowchart illustrating the acquisition of ground fault signal characteristics according to an embodiment of the present invention.

[0021] Figure 3 This is a flowchart illustrating the process of obtaining the distortion degree of the reconstructed residual current signal according to an embodiment of the present invention.

[0022] Figure 4 This is a flowchart for obtaining the predicted value of grounding resistance according to an embodiment of the present invention.

[0023] Figure 5 This is a flowchart for calculating the grounding risk value according to an embodiment of the present invention.

[0024] Figure 6 This is a flowchart illustrating grounding risk warning according to an embodiment of the present invention.

[0025] Figure 7 This is a structural block diagram of a power distribution line grounding risk early warning system according to an embodiment of the present invention.

[0026] Figure 8 This is a schematic diagram of a power distribution line grounding risk early warning device according to an embodiment of the present invention. Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to embodiments and accompanying drawings. The illustrative embodiments and descriptions of this invention are for illustrative purposes only and are not intended to limit the invention.

[0028] In view of the fact that current grounding risk early warning methods are easily affected by external interference and lack dynamic early warning capabilities, this invention provides a grounding risk early warning method for power distribution lines. Figure 1 This is a flowchart illustrating a power distribution line grounding risk early warning method according to an embodiment of the present invention. (Reference) Figure 1 The method for early warning of grounding risks in power distribution lines includes: S1 to S7.

[0029] In S1, the monitoring data of the power distribution line under test is acquired.

[0030] In one implementation, the monitoring data may include: residual current signals, environmental data, equipment health data, static covariates, and external event data.

[0031] Environmental data include soil resistivity, soil moisture, and temperature.

[0032] Equipment health data includes equipment aging status, equipment maintenance records, and equipment failure history.

[0033] Static covariates include the geographical location of the power distribution line and the tower structure type.

[0034] External event data includes lightning strike data, storm data, and power grid fault records.

[0035] Residual current signal refers to the current flowing through grounding resistance or incompletely closed loops in a power system due to grounding faults or asymmetrical faults. It is usually caused by asymmetrical current or current leakage in the system and can lead to safety problems such as damage to electrical equipment, fire, or electrical hazards. Monitoring and analyzing residual current signals is a crucial part of power distribution systems, enabling timely detection of grounding faults or other types of current leakage to ensure system safety.

[0036] In this embodiment of the invention, the multi-dimensional and comprehensive nature of the monitoring data provides sufficient basis for early warning of grounding risks in power distribution lines. By collecting and analyzing this data, a comprehensive understanding of the operating status of power distribution lines and changes in the external environment can be achieved, thereby improving the accuracy and timeliness of the grounding risk early warning system.

[0037] In S2, empirical mode decomposition is performed on the residual current signal to obtain ground fault signal characteristics. For example, empirical mode decomposition is used to extract ground fault signal characteristics from the residual current signal in the monitoring data of the tested power distribution line.

[0038] Empirical Mode Decomposition (EMD) is a signal processing method mainly used to analyze and decompose nonlinear and non-stationary signals. It adaptively decomposes complex signals into a set of inherent oscillation modes called intrinsic mode functions (IMFs), which can reflect the local characteristics of the signal.

[0039] Ground fault signal characteristics refer to the characteristics of current, voltage, or other power signals monitored when a ground fault occurs in a power distribution system. Ground fault signal characteristics can be expressed in various ways, such as current, frequency, and amplitude, and can reflect the abnormal state of the system.

[0040] Figure 2 This is a flowchart illustrating the acquisition of ground fault signal characteristics according to an embodiment of the present invention. In one embodiment, in S2, acquiring ground fault signal characteristics includes: S201 to S204.

[0041] In S201, multiple sets of Gaussian white noise are added to the residual current signal to obtain multiple sets of residual current and noise signals:

[0042]

[0043] Where, x I (t) represents the residual current and noise signal at time t, x(t) represents the residual current signal at time t, and n I (t) represents the random Gaussian white noise added at time t.

[0044] In this embodiment of the invention, the purpose of adding random Gaussian white noise to the original signal is to enhance the diversity of the signal and simulate noise interference that may occur in the signal. In this way, the signal can be made closer to the situation in the real environment. The addition of Gaussian white noise can help the model process and adapt to noise in the signal, avoiding erroneous fault detection due to noise in practical applications.

[0045] In S202, through the empirical mode decomposition, each set of residual current and noise signals in the multiple sets of residual current and noise signals is decomposed into multiple intrinsic mode function (IMF) components and a single residual component:

[0046]

[0047] Where n represents the total number of IMF components, J Id (t) represents the d-th IMF component at time t, r I (t) represents the residual component at time t.

[0048] In S203, the intrinsic mode function (IMF) components and residual components at the same time in multiple sets of residual current and noise signals are averaged to obtain multiple stable IMF components and a single stable residual component:

[0049]

[0050] Among them, J d (t) represents the stable IMF component of the d-th IMF component at time t, N represents the degree of empirical mode decomposition, and J ld r(t) represents the IMF component of the d-th IMF component in the l-th empirical mode decomposition at time t, and r(t) represents the stable residual component at time t. l (t) represents the residual component in the l-th empirical mode decomposition at time t.

[0051] In this embodiment of the invention, averaging the results of multiple decompositions can reduce the interference of random noise and obtain more accurate IMF components and residual components. Each decomposition may result in slight fluctuations in the IMF and residual components; averaging helps to remove these fluctuations, enhancing the stability and consistency of the decomposition results, thereby improving the accuracy of subsequent analyses.

[0052] In S204, the stable IMF component corresponding to the frequency band related to the ground fault in the multiple stable IMF components and the single stable residual component is used as the ground fault signal feature.

[0053] It should be noted that the frequency bands related to ground faults are determined using wavelet transform. When a ground fault occurs, the signal often exhibits transient characteristics, which traditional Fourier transform may not be able to fully capture. Wavelet transform can analyze the signal at different time scales and extract the frequency band features related to ground faults.

[0054] In this embodiment of the invention, by selecting a frequency band related to grounding faults, key signal features associated with the fault can be effectively extracted, avoiding interference from redundant information and improving the accuracy of fault detection. This allows focus on the most representative part of the signal, thereby diagnosing grounding faults more accurately.

[0055] In S3, residual current signal is reconstructed based on ground fault signal characteristics to obtain the distortion degree of the reconstructed residual current signal.

[0056] Reconstructed signal distortion refers to the degree of difference or distortion between the original signal and the reconstructed signal during the signal reconstruction process. The greater the distortion, the more severe the signal reconstruction distortion.

[0057] Figure 3 A flowchart illustrating the process of obtaining the distortion degree of the reconstructed residual current signal according to an embodiment of the present invention. (See reference...) Figure 3 In one implementation, in S3, the distortion degree of the reconstructed residual current signal is obtained, including S301 and S302.

[0058] In S301, the residual current signal is reconstructed based on the characteristics of the ground fault signal and the stable residual components, and the maximum amplitude change and maximum frequency change between the reconstructed residual current signal and the residual current signal at the corresponding time are extracted:

[0059]

[0060]

[0061] Where ΔA represents the magnitude change, and max() represents the maximum value function. This indicates taking the absolute value, where t represents time, and x represents the time interval. re x(t) represents the reconstructed residual current signal at time t, and ΔF represents the frequency change. This represents the spectrum of the reconstructed residual current signal at time t after Fourier transform, when converted from the time domain to the frequency domain. This represents the spectrum of the residual current signal after being transformed from the time domain to the frequency domain using Fourier transform at time t.

[0062] In S302, the distortion of the reconstructed residual current signal is calculated based on the maximum amplitude change and the maximum frequency change, the maximum amplitude and the maximum frequency in the residual current signal, the amplitude change weight, and the frequency change weight.

[0063]

[0064] Among them, D distortion The reconstructed signal distortion is represented by α, where α represents the weight of amplitude changes and β represents the weight of frequency changes. max F represents the maximum amplitude value in the residual current signal. max This represents the maximum spectrum in the residual current signal.

[0065] It should be noted that the weights of amplitude changes and frequency characteristic changes can be determined experimentally so that they can be flexibly adjusted according to the needs of actual applications.

[0066] Specifically, by using existing historical monitoring data, different weight combinations of amplitude and frequency characteristic changes are simulated. The data is divided into training and validation sets, and cross-validation is used to evaluate model performance under different weight settings to find the optimal weight combination. The model's performance under different weight settings is evaluated by calculating accuracy and recall on the validation set. Real-world fault data is used for validation to ensure the model's effectiveness and accuracy in real-world environments.

[0067] In this embodiment of the invention, combining amplitude and frequency response variations allows for a more comprehensive capture of signal changes, effectively reflecting both amplitude fluctuations and frequency component variations. By simultaneously considering amplitude and frequency characteristics, the reconstructed signal retains as much key information as possible from the original signal, improving the accuracy of fault detection. Furthermore, the introduction of maximum amplitude and spectrum as standards considers the global characteristics of the signal rather than just local differences. This global standard avoids excessive influence of local errors on distortion calculation, making the distortion calculation more representative and accurate throughout the entire signal reconstruction process.

[0068] In S4, if the distortion of the reconstructed residual current signal is greater than the preset distortion of the reconstructed signal, the tested power distribution line is identified as an abnormal power distribution line, and the process proceeds to S5. Otherwise, the tested power distribution line is identified as a normal power distribution line, and the process returns to S1.

[0069] It should be noted that those skilled in the art can set the preset reconstructed signal distortion degree according to actual needs, and this invention does not limit it.

[0070] In this embodiment of the invention, by determining whether the distortion degree of the reconstructed signal is greater than a preset threshold, the power distribution line can be accurately identified as an abnormal power distribution line when the signal shows a large deviation. Only when an abnormal power distribution line is identified will the next step be taken, reducing the computational complexity. Simultaneously, by setting a reasonable preset distortion degree threshold, false alarms caused by small, irrelevant changes in the signal can be effectively reduced. Since the distortion degree is only considered abnormal when it exceeds the preset threshold, frequent false alarms are avoided, while simultaneously improving the stability and reliability of the system.

[0071] In S5, based on environmental data and static covariates, a time-series fusion converter is used to predict the grounding resistance of abnormal power distribution lines to obtain the predicted grounding resistance value.

[0072] TFT (Temporal Fusion Transformers) is a deep learning model for time series data prediction, particularly suitable for processing sequential data with multiple time steps and effectively capturing dependencies over long time spans. TFT is a variant based on the Transformer architecture, designed to address some of the problems in traditional sequence prediction tasks, such as the complexity of time series data with multiple patterns, heterogeneous features, and multidimensional data.

[0073] Figure 4 This is a flowchart illustrating the process of obtaining a predicted grounding resistance value according to an embodiment of the present invention. (Refer to...) Figure 4 In one embodiment, in S5, the predicted value of the grounding resistance is obtained, including S501 to S504.

[0074] In step S501, based on environmental data and static covariates, a gated residual network is used to determine multiple key features of the input residual current signal, and a first preset activation function is used to determine the weights of each key feature. For example, the gated residual network processes the monitoring data to determine the feature weights of each input data point in the monitoring data, thereby filtering out multiple key features related to predicting the grounding resistance of abnormal power distribution lines.

[0075]

[0076] in, Represents the feature weight of the j-th feature at time t, Softmax represents the Softmax activation function, and CRN v ( ) indicates a gating mechanism that selects key features, z t Let represent the environmental data observed at time t, and s represent the static covariates.

[0077] In this embodiment of the invention, the CRNv (Correlated Random Network with variation) mechanism is used to select the features most relevant to the grounding resistance prediction, avoiding redundant information from interfering with the prediction results. This feature selection method enhances the model's ability to focus on key information, making the prediction more accurate.

[0078] In S502, a Long Short-Term Memory (LSTM) network is used to process multiple key features and their weights to generate a hidden state for predicting the grounding resistance. For example, the time-series relationships between the key features are processed using an LSTM network to generate the hidden state for predicting the grounding resistance.

[0079]

[0080] Among them, i t Let W represent the output of the input gate at time t, σ represent the sigmoid activation function, and W represent the output of the input gate at time t. i X represents the weight matrix in the input gate that acts on the transformation of key feature data. t U represents the key feature data of the input at time t. i h represents the weight matrix in the input gate that affects the hidden state transition. t-1 Let b represent the hidden state at time t-1. i f represents the bias of the input gate. t W represents the output of the forget gate at time t. f U represents the weight matrix applied to the current input data in the forget gate. f The weight matrix b represents the effect of the forget gate on the hidden state. f c represents the bias of the forget gate. t This represents the cell state at time t. c represents element-wise multiplication. t-1 Let W represent the cell state at time t-1, tanh represent the tanh activation function, and W represent the cell state at time t-1. c U represents the weight matrix applied to the current input data during cell state updates. c b represents the weight matrix applied to the hidden state during unit state update. c o represents the bias in the cell state update. t W represents the output of the output gate at time t. o U represents the weight matrix applied to the current input data in the output gate. o Let b represent the weight matrix that acts on the hidden state in the output gate. o h represents the bias of the output gate. t This represents the hidden state at time t.

[0081] In S503, the hidden state is passed to the multi-head self-attention layer. By weighting the hidden state, the key time steps affecting the prediction of grounding resistance are determined, and an enhanced hidden state is output.

[0082]

[0083] Where Attention represents the attention mechanism, Q represents the query vector, K represents the key vector, and V represents the value vector. T D represents the transpose operation. K This represents the dimension of the key vector.

[0084] In this embodiment of the invention, the TFT uses a combination of LSTM and multi-head self-attention mechanism to capture complex time series dependencies. It not only considers the impact of past time points on the current prediction, but also uses the self-attention mechanism to identify the most critical time steps for the prediction. By weighting the hidden states of different time steps, the TFT can automatically learn which time steps are most important for the prediction of ground resistance, thereby improving the accuracy and reliability of the prediction.

[0085] In S504, the enhanced hidden state is input to the output layer of the LSTM, and the predicted ground resistance value is output:

[0086]

[0087] in, W represents the predicted grounding resistance value at time t+τ, q represents the predicted quantile, and W represents the predicted grounding resistance value. q This represents the weight matrix of the output layer. Let b represent the enhanced hidden state at time t+τ. q This indicates the bias of the output layer.

[0088] In S6, the grounding risk value is calculated using a Bayesian network based on monitoring data and predicted grounding resistance.

[0089] A Bayesian network is a graphical model used to represent and reason about uncertainty. It uses a directed acyclic graph (DAG) to represent the conditional dependencies between random variables. The core idea of ​​Bayesian networks is to reason and make decisions based on Bayes' theorem, and they are widely used in probabilistic reasoning, decision support, machine learning, diagnostics, and data mining.

[0090] Figure 5 This is a flowchart illustrating the calculation of grounding risk values ​​according to an embodiment of the present invention. (See reference...) Figure 5 In one implementation, in S6, the grounding risk value is calculated, including S601 to S605.

[0091] In S601, the predicted grounding resistance value and monitoring data are used as nodes, and each node is connected by directed edges according to the causal relationship between them, thus constructing a Bayesian network including multiple parent nodes and multiple child nodes.

[0092] In S602, based on the states of multiple parent nodes in the Bayesian network, the probability distribution of multiple child nodes under different conditions is determined through expert evaluation:

[0093]

[0094] Among them, P(C e ) represents the e-th node C eThe probability distribution, where k represents the total number of child nodes, C e Indicates child node C e ,pa(C e ) represents child node C e The set of parent nodes, Indicates child node C e The probability distribution that occurs given the set of parent nodes.

[0095] In S603, a conditional probability table is constructed by combining the probability distributions of multiple child nodes.

[0096] In S604, Bayesian inference is performed based on the conditional probability table to calculate the posterior probabilities of multiple child nodes:

[0097]

[0098] in, This represents the probability that a ground fault A will occur in an abnormal power distribution line given that monitoring data B has occurred. Let P(A) represent the probability of observing monitoring data B given that ground fault A has occurred in the abnormal power distribution line, and let P(B) represent the total probability of monitoring data B occurring.

[0099] In S605, the grounding risk value of abnormal distribution lines is calculated based on the posterior probabilities of multiple child nodes:

[0100]

[0101]

[0102] Among them, R g Indicates the grounding risk value, w e Let represent the weight of the e-th child node, and k represent the total number of child nodes. C represents the e-th child node. e Given the state of the parent node, the posterior probability, μ represents the risk correction factor, ρ represents the adjustment coefficient, and G... pred G represents the predicted value of grounding resistance. th This indicates the safety standard value for grounding resistance.

[0103] It should be noted that the adjustment coefficient is determined based on experience.

[0104] In this embodiment of the invention, combining the posterior probabilities of each sub-node to comprehensively calculate the grounding risk value ensures that the influence of multiple variables in the power distribution line is taken into account, including the predicted grounding resistance, environmental data, and equipment condition. This method, by weighting the posterior probabilities of each sub-node, can comprehensively assess grounding risk and avoid over-reliance on a single factor in risk value assessment. Simultaneously, the risk correction factor can be dynamically adjusted based on real-time data and different operating conditions to ensure that the system's prediction results are more accurate and practical.

[0105] Furthermore, by constructing a grounding risk value model using Bayesian networks, the predicted grounding resistance value and monitoring data can be organically combined to form a systematic risk assessment framework. Bayesian networks can capture the causal relationships between various factors while considering multiple variables, thus providing a more accurate risk value assessment. This method can comprehensively assess the potential risks of grounding faults, avoiding the limitations of traditional methods that simply rely on a single feature.

[0106] In S7, grounding risk warnings are issued for abnormal power distribution lines based on grounding risk values.

[0107] Figure 6 This is a flowchart illustrating a grounding risk warning system according to an embodiment of the present invention. (Reference) Figure 6 In one implementation, grounding risk warning is performed in S7, including S701 and S702.

[0108] In S701, the grounding risk level of the tested power distribution line is determined based on the grounding risk value. Grounding risk levels include minor risk, medium risk, and severe risk.

[0109]

[0110] Where L represents the risk level and r represents the grounding risk value.

[0111] In S702, grounding risk warnings are issued for abnormal power distribution lines based on the grounding risk level.

[0112] In one implementation, grounding risk warning for abnormal power distribution lines includes: not issuing a warning when the grounding risk level of the abnormal power distribution line is minor; sending a stroke risk notification to management personnel when the grounding risk level of the abnormal power distribution line is medium, reminding management personnel to conduct regular inspections; and triggering an emergency warning when the grounding risk level of the abnormal power distribution line is severe, arranging maintenance personnel to conduct a comprehensive inspection and repair on-site, checking whether there are serious faults in the line grounding system, and replacing or repairing damaged equipment.

[0113] In this embodiment of the invention, by determining different risk levels based on grounding risk values, corresponding countermeasures can be taken at different risk levels. This tiered early warning method ensures that excessive resources and time are not wasted in low-risk situations, while in high-risk situations, an emergency response can be quickly triggered to avoid potential disasters. Meanwhile, as the scale and complexity of power systems continue to increase, the management and monitoring of distribution lines become particularly important. This tiered risk early warning system can be flexibly expanded according to different system scales and needs, ensuring continuous monitoring of the management of various power equipment and lines.

[0114] The beneficial effects of the technical solution provided by the embodiments of the present invention include at least the following: by extracting the ground fault signal characteristics of the residual current signal in the monitoring data of each power distribution line through empirical mode decomposition, external noise can be effectively filtered, reducing false alarms and missed alarms; at the same time, signal reconstruction and distortion calculation further optimize the judgment of fault signals, avoid the influence of external interference, and improve the accuracy of fault detection; by dynamically predicting the grounding resistance through TFT technology, the changing trend of the grounding resistance can be tracked in real time, breaking the limitation of static threshold; combined with Bayesian network, grounding risk can be dynamically assessed and early warning can be issued in a timely manner, ensuring that effective preventive measures can be taken when the grounding resistance changes.

[0115] On the other hand, the present invention provides a power distribution line grounding risk early warning system. Figure 7 This is a structural block diagram of a power distribution line grounding risk early warning system according to an embodiment of the present invention. (Reference) Figure 7 The power distribution line grounding risk early warning system includes: a monitoring data acquisition module, a ground fault signal characteristic acquisition module, a reconstructed residual current signal distortion calculation module, an abnormal power distribution line identification module, a grounding resistance prediction value calculation module, a grounding risk value calculation module, and a grounding risk early warning module.

[0116] The monitoring data acquisition module is used to acquire monitoring data of the tested power distribution line. The monitoring data includes residual current signals, environmental data, static covariates, equipment health data, and external event data.

[0117] The ground fault signal feature acquisition module is used to: perform empirical mode decomposition on the residual current signal to obtain ground fault signal features.

[0118] The module for calculating the distortion degree of reconstructed residual current signal is used to: reconstruct the residual current signal based on the characteristics of ground fault signals in order to obtain the distortion degree of the reconstructed residual current signal.

[0119] The abnormal power distribution line determination module is used to determine the tested power distribution line as an abnormal power distribution line when the distortion of the reconstructed residual current signal is greater than the preset distortion of the reconstructed signal.

[0120] The grounding resistance prediction calculation module is used to predict the grounding resistance of abnormal power distribution lines based on environmental data and static covariates using a time-series fusion converter, so as to obtain the predicted grounding resistance value.

[0121] The grounding risk value calculation module is used to calculate the grounding risk value using a Bayesian network based on monitoring data and predicted grounding resistance values.

[0122] The grounding risk early warning module is used to provide grounding risk warnings for abnormal power distribution lines based on grounding risk values.

[0123] In some embodiments, the ground fault signal feature acquisition module is further configured to: add multiple sets of Gaussian white noise to the residual current signal to obtain multiple sets of residual current and noise signals; decompose each set of residual current and noise signals into multiple intrinsic mode function (IMF) components and a single residual component through empirical mode decomposition; average the IMF components and residual components at the same time in the multiple sets of residual current and noise signals to obtain multiple stable IMF components and a single stable residual component; and use the stable IMF component corresponding to the frequency band related to the ground fault in the multiple stable IMF components and the single stable residual component as the ground fault signal feature.

[0124] In some implementations, the residual current signal distortion calculation module is further used to: reconstruct the residual current signal based on the characteristics of the ground fault signal and the stable residual components, and extract the maximum amplitude change and the maximum frequency change between the reconstructed residual current signal and the residual current signal at the corresponding time; and calculate the distortion of the reconstructed residual current signal based on the maximum amplitude change and the maximum frequency change, the maximum amplitude and the maximum frequency in the residual current signal, the amplitude change weight and the frequency change weight.

[0125] In some implementations, the ground resistance prediction calculation module is further configured to: determine multiple key features of the input residual current signal using a gated residual network based on environmental data and static covariates, and determine the weights of each key feature using a first preset activation function; process the multiple key features and their weights using a Long Short-Term Memory (LSTM) network to generate a hidden state for the ground resistance prediction; pass the hidden state to a multi-head self-attention layer, determine the key time steps affecting the ground resistance prediction by weighting the hidden state, and output an enhanced hidden state; and input the enhanced hidden state to the output layer of the LSTM to output the ground resistance prediction.

[0126] In some implementations, the grounding risk value calculation module is further used to: construct a Bayesian network including multiple parent nodes and multiple child nodes by using the predicted grounding resistance value and monitoring data as nodes and connecting each node with directed edges according to the causal relationship between each node; determine the probability distribution of multiple child nodes under different conditions based on the state of multiple parent nodes in the Bayesian network through expert evaluation; construct a conditional probability table by combining the probability distribution of multiple child nodes; perform Bayesian inference based on the conditional probability table to calculate the posterior probability of multiple child nodes; and calculate the grounding risk value of the abnormal power distribution line based on the posterior probability of multiple child nodes.

[0127] In some implementations, the grounding risk warning module is further configured to: determine the grounding risk level of the abnormal power distribution line based on the grounding risk value, wherein the grounding risk level includes minor risk, medium risk and severe risk; and provide grounding risk warning for the abnormal power distribution line based on the grounding risk level.

[0128] In implementing the functions of the integrated modules described above in hardware, this embodiment of the invention provides a structure for the power distribution line grounding risk early warning device involved in the above embodiments. Figure 8 This is a schematic diagram of a power distribution line grounding risk early warning device according to an embodiment of the present invention. (Reference) Figure 8 The power distribution line grounding risk early warning device includes: at least one processor; and at least one memory. The at least one memory is coupled to the at least one processor and stores instructions for execution by the at least one processor, which, when executed by the at least one processor, implement the method described above.

[0129] In one implementation, the memory can exist independently of the processor. The memory can be connected to the processor via a bus and used to store instructions or program code. When the processor calls and executes the instructions or program code stored in the memory, it can implement the method provided in the embodiments of the present invention. In another implementation, the memory can also be integrated with the processor.

[0130] On the other hand, the present invention also provides a computer-readable storage medium (e.g., a non-transitory computer-readable storage medium) storing computer program instructions that, when executed on a computer, cause the computer to perform the method as described in any of the above embodiments.

[0131] This invention provides a computer program that, when run on a computer, causes the computer to perform the method of any of the above embodiments.

[0132] This invention provides a computer program product containing instructions that, when run on a computer, cause the computer to perform the method of any of the above embodiments.

[0133] The beneficial effects of the technical solution provided by the embodiments of the present invention include at least the following: by extracting the ground fault signal characteristics of the residual current signal in the monitoring data of each power distribution line through empirical mode decomposition, external noise can be effectively filtered, reducing false alarms and missed alarms; at the same time, signal reconstruction and distortion calculation further optimize the judgment of fault signals, avoid the influence of external interference, and improve the accuracy of fault detection; by dynamically predicting the grounding resistance through TFT technology, the changing trend of the grounding resistance can be tracked in real time, breaking the limitation of static threshold; combined with Bayesian network, grounding risk can be dynamically assessed and early warning can be issued in a timely manner, ensuring that effective preventive measures can be taken when the grounding resistance changes.

[0134] It should be understood that the processor in the embodiments of the present invention can be a central processing unit (CPU), or it can be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor.

[0135] It should also be understood that the memory in the embodiments of the present invention can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate synchronous DRAM (DDR SDRAM), enhanced synchronous DRAM (ESDRAM), synchronous linked DRAM (SLDRAM), and direct rambus RAM (DR RAM).

[0136] The above embodiments can be implemented, in whole or in part, by software, hardware (such as circuits), firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0137] It should be understood that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. A and B can be singular or plural. Additionally, the character " / " in this article generally indicates an "or" relationship between the preceding and following related objects, but it can also represent an "and / or" relationship. Please refer to the context for a more accurate understanding.

[0138] In this invention, "at least one" means one or more, and "more than one" means two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of a single item or a plurality of items. For example, at least one of a, b, or c can represent: a, b, c, ab, ac, bc, or abc, where a, b, and c can be a single item or multiple items.

[0139] It should be understood that, in various embodiments of the present invention, the order of the above-mentioned process numbers does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0140] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0141] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the devices, apparatuses, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0142] In the several embodiments provided by this invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0143] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0144] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0145] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0146] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for early warning of grounding risks in power distribution lines, characterized in that, The method includes: Acquire monitoring data of the power distribution line under test, wherein the monitoring data includes residual current signal, environmental data, static covariates, equipment health data and external event data; Empirical mode decomposition is performed on the residual current signal to obtain ground fault signal characteristics; The residual current signal is reconstructed based on the characteristics of the ground fault signal to obtain the distortion degree of the reconstructed residual current signal. If the distortion of the reconstructed residual current signal is greater than the preset distortion of the reconstructed signal, the tested power distribution line is identified as an abnormal power distribution line. Based on the environmental data and the static covariate, a time-series fusion converter is used to predict the grounding resistance of the abnormal power distribution line to obtain the predicted grounding resistance value. Based on the monitoring data and the predicted grounding resistance value, the grounding risk value is calculated using a Bayesian network; and Based on the grounding risk value, a grounding risk warning is issued for the abnormal power distribution line.

2. The method according to claim 1, characterized in that, Empirical mode decomposition (EMD) is performed on the residual current signal to obtain the characteristics of the ground fault signal, including: Multiple sets of Gaussian white noise are added to the residual current signal to obtain multiple sets of residual current and noise signals; Through the empirical mode decomposition, each set of residual current and noise signals in the multiple sets of residual current and noise signals is decomposed into multiple intrinsic mode function (IMF) components and a single residual component. The intrinsic mode function (IMF) components and residual components at the same time in the multiple sets of residual current and noise signals are averaged to obtain multiple stable IMF components and a single stable residual component; and The stable IMF component corresponding to the frequency band related to the ground fault among the multiple stable IMF components and the single stable residual component is used as the ground fault signal feature.

3. The method according to claim 2, characterized in that, Based on the characteristics of the ground fault signal, residual current signal reconstruction is performed to obtain the distortion degree of the reconstructed residual current signal, including: The residual current signal is reconstructed based on the characteristics of the ground fault signal and the stable residual components, and the maximum amplitude change and maximum frequency change between the reconstructed residual current signal and the residual current signal at the corresponding time are extracted; and The distortion degree of the reconstructed residual current signal is calculated based on the maximum amplitude change and the maximum frequency change, the maximum amplitude and maximum frequency in the residual current signal, the amplitude change weight and the frequency change weight.

4. The method according to any one of claims 1 to 3, characterized in that, Based on the environmental data and the static covariates, a time-series fusion transformer is used to predict the grounding resistance of the abnormal power distribution line to obtain a predicted grounding resistance value, including: Based on the environmental data and the static covariates, a gated residual network is used to determine multiple key features of the input residual current signal, and a first preset activation function is used to determine the weight of each key feature. The Long Short-Term Memory (LSTM) network is used to process the multiple key features and the weights of each key feature to generate a hidden state for the predicted grounding resistance value. The hidden state is passed to a multi-head self-attention layer. By weighting the hidden state, the key time steps affecting the predicted grounding resistance value are determined, and an enhanced hidden state is output. The enhanced hidden state is input to the output layer of the LSTM to output the predicted ground resistance value.

5. The method according to claim 4, characterized in that, Based on the monitoring data and the predicted grounding resistance value, the grounding risk value is calculated using a Bayesian network, including: The predicted grounding resistance value and the monitoring data are used as nodes, and each node is connected by directed edges according to the causal relationship between them to construct a Bayesian network including multiple parent nodes and multiple child nodes. Based on the states of the multiple parent nodes in the Bayesian network, the probability distribution of the multiple child nodes under different conditions is determined by expert evaluation. A conditional probability table is constructed by combining the probability distributions of the multiple child nodes; Based on the conditional probability table, Bayesian inference is performed to calculate the posterior probabilities of the multiple child nodes; and The grounding risk value of the abnormal power distribution line is calculated based on the posterior probabilities of the multiple sub-nodes.

6. The method according to claim 1, characterized in that, Based on the aforementioned grounding risk value, a grounding risk warning is issued for the abnormal power distribution line, including: Based on the grounding risk value, the grounding risk level of the abnormal power distribution line is determined, wherein the grounding risk level includes minor risk, medium risk, and severe risk; and Based on the grounding risk level, a grounding risk warning is issued for the abnormal power distribution line.

7. A power distribution line grounding risk early warning system, characterized in that, The system includes: The monitoring data acquisition module is used to acquire monitoring data of the tested power distribution line, wherein the monitoring data includes residual current signal, environmental data, static covariates, equipment health data, and external event data. The ground fault signal feature acquisition module is used to: perform empirical mode decomposition on the residual current signal to obtain ground fault signal features; The residual current signal distortion calculation module is used to: reconstruct the residual current signal based on the characteristics of the ground fault signal to obtain the distortion of the reconstructed residual current signal; An abnormal power distribution line determination module is used to determine the tested power distribution line as an abnormal power distribution line when the distortion degree of the reconstructed residual current signal is greater than the preset reconstructed signal distortion degree. The grounding resistance prediction calculation module is used to: predict the grounding resistance of the abnormal power distribution line based on the environmental data and static covariates, using a time-series fusion converter, so as to obtain the predicted grounding resistance value. The grounding risk value calculation module is used to: calculate the grounding risk value using a Bayesian network based on the monitoring data and the predicted grounding resistance value; and The grounding risk early warning module is used to: provide grounding risk early warning for the abnormal power distribution line based on the grounding risk value.

8. A grounding risk early warning device for power distribution lines, characterized in that, The device includes: At least one processor; and At least one memory coupled to the at least one processor and storing instructions for execution by the at least one processor, the instructions implementing the method of any one of claims 1 to 6 when executed by the at least one processor.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores instructions that, when executed by a computer, cause the computer to perform the method of any one of claims 1 to 6.

10. A computer program product, characterized in that, The computer program product includes instructions that, when executed by a computer, cause the computer to perform the method according to any one of claims 1 to 6.

Images