Power line risk dynamic inversion and fault tracing method, system, medium and equipment

Abstract

The embodiment of the invention discloses a power line risk dynamic inversion and fault tracing method and system, a medium and equipment. The method comprises the following steps: acquiring multi-dimensional data along a power line; constructing a forward model based on a physical law, representing a mapping relationship between the multi-dimensional data and the internal risk factor, further determining an inverse problem model, taking the multi-dimensional data as known data, and taking the internal risk factor as data to be solved; and after multi-dimensional data is input, solving the inverse problem model through a preset regularization objective function, outputting a risk factor estimation value and determining a risk index. And after the fault occurs, obtaining a historical risk index before the fault, multi-dimensional data and fault moment multi-dimensional data, inputting the historical risk index, the multi-dimensional data and the fault moment multi-dimensional data into a preset Bayesian network for probabilistic reasoning, and outputting a final fault root cause. According to the method, the historical risk data and the real-time fault information are combined, the fault source is accurately diagnosed, the operation safety and stability of the power line are remarkably improved, and the maintenance cost and time after the fault are effectively reduced.

Description

Technical Field

[0001] This invention relates to the field of power line safety technology, and in particular to a method, system, medium, and equipment for dynamic risk inversion and fault tracing of power lines. Background Technology

[0002] Power lines are a critical component of modern social infrastructure, used for power transmission to ensure the normal operation of economic activities. However, due to the diversity and complexity of the operating environment of power lines, these lines have long faced various risks. These risks include problems caused by internal factors such as insulation aging and joint defects, as well as hidden dangers caused by external factors such as third-party construction and environmental erosion. Therefore, although current technologies for ensuring the safety of power lines have made some progress, such as acquiring continuous data along the line through advanced monitoring methods like distributed fiber optic sensing, there are still significant shortcomings that cannot meet the requirements for real-time performance, proactive monitoring, and accuracy.

[0003] Existing monitoring technologies, such as distributed fiber optic sensing systems, while providing continuous data on temperature and vibration along the power line, are essentially reactive, responding to past conditions rather than proactively predicting potential risks. They primarily rely on path threshold alarms to report anomalies such as temperature abnormalities and acoustic disturbances. However, they cannot continuously quantify and assess progressive, deeper risk factors, such as insulation aging, cumulative mechanical stress at joints, and slow intrusion of minute moisture. Furthermore, existing risk assessment frameworks often rely on infrequent qualitative updates, failing to promptly process and analyze high-frequency, high-resolution data streams from advanced sensing systems. This results in current risk assessments being often phased, unable to fully reflect the actual operating conditions of power lines, and thus unable to support effective operation and maintenance decisions.

[0004] Existing fault diagnosis and tracing technologies are typically triggered passively after a fault occurs, analyzing data only captured at the moment of the fault or in the short period following it, failing to provide long-term operational information prior to the fault. For example, in analyzing an insulation breakdown accident, traditional methods struggle to distinguish whether it was caused by long-term thermal aging or a recent transient overvoltage. This singular post-fault analysis not only affects the accuracy of fault diagnosis but also leads to prolonged troubleshooting time and increased costs. Therefore, an improved method is needed that can perform real-time risk assessment before a fault occurs and accurate, context-aware root cause analysis after the fault occurs, thereby providing more reliable risk management and efficient fault diagnosis. Summary of the Invention

[0005] Based on this, it is necessary to propose a method, system, medium, and equipment for dynamic inversion of power line risks and fault tracing to address the above problems.

[0006] A method for dynamic risk inversion and fault source tracing of power lines, the method comprising: Obtain multidimensional data along the power line.

[0007] A forward model is established based on physical laws to characterize the mapping relationship between multidimensional data and internal risk factors. An inverse problem model is determined based on the forward model, in which multidimensional data is used as known data and internal risk factors of power lines are used as data to be solved.

[0008] The multidimensional data is input into the inverse problem model, and the inverse problem model is solved based on a preset regularization objective function to output the estimated value of the internal risk factors of the power line.

[0009] The risk index of the power line is determined based on the estimated values ​​of the risk factors.

[0010] After a failure occurs, the historical risk index and multidimensional data before the failure, as well as the multidimensional data at the time of the failure, are obtained. Based on a preset Bayesian network, the historical risk index before the failure and the multidimensional data at the time of the failure are used as inputs to the Bayesian network to output the final root cause of the failure.

[0011] Specifically, the step of inputting the multidimensional data into the inverse problem model, solving the inverse problem model based on a preset regularized objective function, and outputting the final estimated value of the internal risk factors of the power line includes: according to Define the regularization objective function, where F is the positive operator, dobs are multidimensional data, and ||F(m) dobs||22 is the residual term, ||L(m) mprior)||22 is the regularization term, m is the risk factor, mprior is the prior estimate of the risk factor, α is the regularization parameter, and L is the regularization operator.

[0012] The multidimensional data is input into the inverse problem model, and the inverse problem model is solved based on the regularized objective function to output the estimated values ​​of the internal risk factors of the power line. The inverse problem model is as follows: , where mest is the estimated value of the risk factor and J(m) is the regularization objective function.

[0013] Specifically, determining the risk index of the power line based on the estimated value of the risk factor includes: The estimated values ​​of the risk factors are normalized.

[0014] according to Determine the risk index of the power line, where DRI is the risk index, m^i(x,t) is the estimated value of the i-th risk factor after normalization at location x and time t, and wi is the weighting coefficient.

[0015] Specifically, after a fault occurs, the process involves acquiring historical risk indices and multidimensional data prior to the fault, as well as multidimensional data at the moment of the fault. Based on a pre-defined Bayesian network, the historical risk indices prior to the fault and the multidimensional data at the moment of the fault are used as inputs to the Bayesian network to output the final root cause of the fault. This process includes: Construct a Bayesian network, which includes hypothesis nodes, evidence nodes, and directed edges connecting the hypothesis nodes and evidence nodes. Each evidence node corresponds to a conditional probability, and each hypothesis node represents the root cause of a fault that occurs during the operation of a power line.

[0016] Obtain historical risk indices and multidimensional data before the failure occurs, as well as multidimensional data at the moment of the failure.

[0017] The prior probability of the hypothetical node is determined based on the historical risk index.

[0018] Substitute the historical risk index and multidimensional data before the failure, as well as the multidimensional data at the time of the failure, into the evidence node, and call the conditional probability corresponding to the evidence node.

[0019] The conditional probability of the evidence node is transmitted along the network to the hypothesis node through the directed edge. The prior probability of the hypothesis node is adjusted according to the conditional probability to determine the posterior probability of the hypothesis node.

[0020] The final root cause of the failure is determined based on the posterior probability.

[0021] Specifically, determining the prior probability of the hypothetical node based on the historical risk index includes: according to Determine the prior probability of the hypothetical node, where P(Hj) is the prior probability, and g( ) is a mapping function. This is a historical risk index. Location of the fault. It is the starting time for risk accumulation calculation. The time when the fault occurred. In the time interval [ , A time variable that changes within a given period.

[0022] Specifically, determining the final root cause of the failure based on the posterior probability includes: The posterior probabilities of each hypothetical node are sorted in descending order, and the cause of failure corresponding to the hypothetical node with the highest posterior probability is selected as the final root cause of failure.

[0023] The process of establishing a forward model based on physical laws to characterize the mapping relationship between multidimensional data and internal risk factors, and determining the inverse problem model based on the forward model, wherein the multidimensional data is treated as known data and the internal risk factors of the power line are treated as data to be solved in the inverse problem model, further includes: The multidimensional data is sequentially subjected to filtering and noise reduction, normalization, time synchronization, and feature extraction to obtain preprocessed multidimensional data.

[0024] A power line risk dynamic inversion and fault source tracing system, the system comprising: The multidimensional data acquisition module is used to acquire multidimensional data along the power line.

[0025] The inverse problem model determination module is used to establish a forward model based on physical laws to characterize the mapping relationship between multidimensional data and internal risk factors, and to determine the inverse problem model based on the forward model. In the inverse problem model, the multidimensional data is used as known data, and the internal risk factors of the power line are used as data to be solved.

[0026] The risk factor estimation module is used to input the multidimensional data into the inverse problem model, solve the inverse problem model based on a preset regularized objective function, and output the estimated values ​​of the internal risk factors of the power line.

[0027] The risk index determination module for power lines is used to determine the risk index of power lines based on the estimated values ​​of the risk factors.

[0028] The root cause determination module is used to acquire historical risk index and multidimensional data before the fault occurred, as well as multidimensional data at the time of the fault, after the fault occurs. Based on a preset Bayesian network, the historical risk index before the fault occurred and the multidimensional data at the time of the fault are used as inputs to the Bayesian network to output the final root cause of the fault.

[0029] A computer-readable storage medium storing a computer program that, when executed by a processor, causes the processor to perform the steps of the method described above.

[0030] A computer device includes a memory and a processor, the memory storing a computer program that, when executed by the processor, causes the processor to perform the steps of the method described above.

[0031] The embodiments of the present invention have the following beneficial effects: This invention establishes a forward model based on physical laws to characterize the mapping relationship between multidimensional data and internal risk factors. Based on the forward model, an inverse problem model is determined. In the inverse problem model, the multidimensional data is treated as known data, and the internal risk factors of the power line are treated as data to be solved. The multidimensional data is input into the inverse problem model, and the model is solved based on a preset regularized objective function. The estimated values ​​of the internal risk factors of the power line are output. A risk index is generated in real time based on the final estimated values ​​of the risk factors, thereby accurately identifying potential risks in specific sections and providing precise basis for operation and maintenance decisions. This invention achieves risk quantification of multi-source data through forward and inverse modeling, eliminating the shortcomings of existing technologies in qualitative analysis and delayed early warning of single physical phenomena. Furthermore, it can identify and assess potential risk factors in real time, enabling early warning before faults occur, thus improving the overall safety and stability of operation.

[0032] Furthermore, after a fault occurs, the system can output the final root cause of the fault based on a pre-defined Bayesian network, combined with historical risk indices before the fault and multi-dimensional data at the time of the fault. By combining historical risk indices and real-time fault information, it can perform probabilistic reasoning on various potential fault causes and provide accurate diagnostic reports. This significantly improves the safety and stability of power line operation and reduces maintenance costs and time after a fault occurs. Attached Figure Description

[0033] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0034] in: Figure 1 This is a flowchart illustrating an embodiment of a method for dynamic inversion of power line risks and fault tracing provided by the present invention. like Figure 2 As shown, Figure 2 This is a flowchart illustrating another embodiment of the power line risk dynamic inversion and fault tracing method provided by the present invention; Figure 3 This is a schematic diagram of an embodiment of a power line risk dynamic inversion and fault tracing system provided by the present invention; Figure 4 A schematic diagram of the structure of an embodiment of the device provided by the present invention; Figure 5 A schematic diagram of the structure of an embodiment of the medium provided by the present invention. Detailed Implementation

[0035] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0036] like Figure 1 As shown, Figure 1 This is a flowchart illustrating an embodiment of a power line risk dynamic inversion and fault source tracing method provided by the present invention. The method includes: S101: Obtain multidimensional data along the power line.

[0037] For example, a multi-source heterogeneous sensor network captures operational status information along power lines comprehensively and from multiple dimensions. Distributed fiber optic sensors are installed along the power line using an integrated optical cable, which serves as the sensing medium for both distributed temperature sensing and distributed acoustic / vibration sensing. The distributed temperature sensing provides a continuous temperature profile along the line, monitoring the power line itself and the ambient temperature in real time. The distributed acoustic / vibration sensing monitors acoustic and vibration signals caused by third-party construction, environmental disturbances (such as wind loads or foreign object impacts), and line displacement. Discrete point sensors are deployed at critical locations along the power line (such as joints, corners, low-lying areas, or support points) as integrated early warning sensor nodes. These nodes can measure local environmental parameters, such as water level, air humidity, smoke, or combustible gas concentration, to cover specific risk points along the line. Electrical quantity sensors, located in substations or switching stations connected to the power line, utilize intelligent electronic devices and protective relays to collect real-time electrical operating parameters such as voltage, current, and frequency, and can record high-resolution fault waveform data when a fault occurs.

[0038] S102: Establish a forward model based on physical laws to characterize the mapping relationship between multidimensional data and internal risk factors. Determine the inverse problem model based on the forward model. In the inverse problem model, multidimensional data is used as known data, and internal risk factors of power lines are used as data to be solved.

[0039] For example, the collected multidimensional data is cleaned (e.g., filtered and denoised), normalized (uniformed in units), and time-synchronized to ensure data consistency and comparability. Then, feature extraction is performed, such as extracting temperature gradient, hotspot amplitude, and duration from distributed temperature sensing data, performing spectrum analysis on distributed acoustic / vibration sensing data to identify vibration fingerprints of specific events such as excavation, impact, water flow, or wind vibration, and analyzing load cycle, harmonic content, and transient events from electrical operating parameters.

[0040] Furthermore, a positive model representing the mapping relationship between multidimensional data and internal risk factors is established based on physical laws, as shown in the following equation: ; Here, 'm' represents internal risk factors that cannot be directly observed. These factors include, but are not limited to, the insulation health index. H ins (Related to material aging, moisture content, partial discharge activity, etc.), the cumulative mechanical stress index σ of the joint. acc Sheath integrity index I sh Proximity of third-party intrusion threats P thi The vector form is shown in the following equation: ; d represents observable multidimensional data, i.e., physical quantities collected by sensors, such as the temperature distribution T(x) along the line and the vibration energy spectrum of a specific frequency band. E ( ω Local humidity H ( x i ), load current I ( t ), etc., in vector form as shown in the following formula: ; F stands for positive operator, which describes how risk factor m leads to the generation of multidimensional data d. This operator is constructed based on explicit physical laws and can be calibrated and enhanced using historical data through machine learning methods.

[0041] For example, a positive model of the temperature distribution T(x) can be constructed based on the steady-state heat conduction equation, where the heat source term Qc is related to the insulation health index. H ins and load current I ( t Related to, as in equation (3).

[0042] ; in, R 0 is the initial resistance, f( H ins F is a function describing the additional losses caused by insulation degradation. The entire positive operator F combines a coupled physical model of electromagnetics, thermodynamics, and materials science.

[0043] Furthermore, the forward model is transformed into an inverse problem model, where the multidimensional data is taken as known data and the internal risk factors of the power line are taken as data to be solved, as shown in the following equation: ; in, d obs For multidimensional data, m est This is an estimate of the risk factor.

[0044] S103: Input multidimensional data into the inverse problem model, solve the inverse problem model based on the preset regularized objective function, and output the estimated value of the internal risk factors of the power line.

[0045] For example, by employing regularization techniques, the inverse problem is transformed into an optimization problem by minimizing a regularization objective function that includes a data fitting term and a regularization term. J ( m To find a stable and physically reasonable solution. The transformed inverse problem model is shown in the following equation: ; in, m est Let J(m) be the estimated value of the risk factor, and J(m) be the regularization objective function.

[0046] The regularization objective function is shown in the following equation: ; in, F For positive operators, d obs For multidimensional data, || F ( m ) d obs ||2 2 The residual term measures the difference between the model's predicted data and the actual observed data. L ( m m prior )||2 2 It is a regularization term used to introduce prior knowledge about the solution as a constraint. m As a risk factor, m prior For prior estimation of risk factors, such as based on historical data or initial state, L is a regularization operator, such as the identity matrix to constrain the size of the solution, or the gradient operator to constrain the smoothness of the solution, and α is a regularization parameter used to balance the weights of the data fitting term and the regularization term. By adjusting α, a balance can be achieved between the accuracy of the data fitting and the stability of the solution.

[0047] Multidimensional data is input into the inverse problem model, and the inverse problem model is solved based on a regularized objective function, outputting estimated values ​​of internal risk factors of power lines. By introducing prior knowledge (such as the possible range and rate of change of risk factors) as constraints during the solution process, the unique solution that best conforms to physical reality is selected from many possible solutions.

[0048] S104: Determine the risk index of power lines based on the estimated values ​​of risk factors.

[0049] For example, the inverse problem solution process is executed continuously at a high frequency of minutes, and the calculated estimates of the risk factors constitute a risk factor time series. m est ( t Through a weighted fusion algorithm, a comprehensive dynamic risk index is integrated. First, the estimated values ​​of the risk factors are normalized; then, the risk index of the power line is determined according to the following formula: ; DRI stands for Risk Index. m ^ i ( x , t () represents the estimated value of the i-th risk factor after normalization at location x and time t. w i The weighting coefficient is set based on its importance to the overall security of the system, and ∑wi=1.

[0050] when DRI ( x , t ) or its rate of change DRI / t Predictive alerts are issued when preset dynamic thresholds are exceeded. This index not only reflects the overall risk level of the entire power line but can also drill down to any specific section. The DRI and its various sub-risk factors are visualized in the application presentation layer as trend curves, heatmaps, etc. When any risk factor shows an accelerating deterioration trend, a predictive alert is issued, even if no sensor reading reaches the traditional hard alarm threshold at this time. This aligns with the core ideas of continuous monitoring and proactive management advocated by dynamic risk assessment theory, generating a real-time dynamic risk profile.

[0051] Specifically, a certain power line section is located at... x 0. Due to its low-lying terrain, the ambient humidity has been consistently high. Observational data. d obs Data collected by distributed temperature sensors T (x 0) The displayed temperature is 55℃, slightly higher than the normal value of 50℃, but not reaching the alarm threshold of 65℃. (Point-type humidity sensor) H ( x 0) Reading is 95%RH. Distributed acoustic / vibration sensing data. S ( x 0, ω , t No obvious third-party construction vibrations were detected.

[0052] Establish temperature distribution based on physical laws T ( x The positive model F(m) of 0) indicates that, under the current load current I ( t If the insulation health index is below ) H ins If the value is ideal (e.g., 1.0), the predicted temperature should be 50°C.

[0053] Furthermore, the multidimensional data is input into the temperature distribution... T ( x The inverse problem model of 0) is solved based on a preset regularized objective function, and the insulation health index H is output. ins The estimated value. To make the residual term of the inverse problem model || F ( m ) d obs ||2 2 To minimize the temperature to match the observed 55°C, H must be reduced. ins The value of the risk factor at that point is then derived. m est In the middle, H ins A significant drop to 0.7 indicates a 30% decrease in insulation performance, while the sheath integrity index also... I sh The effect is also reduced due to the high humidity environment. Although no single sensor triggers a traditional alarm, the system uses the inverted H... ins and I sh Calculate the section DRI ( x 0, t If the value is high-risk, a predictive alarm is issued, indicating that there is a risk of insulation degradation at location x0 due to a high humidity environment. Preventive inspection is recommended, thus achieving effective early warning before the fault occurs.

[0054] S105: After a fault occurs, acquire the historical risk index and multidimensional data before the fault occurred, as well as the multidimensional data at the time of the fault. Based on a preset Bayesian network, use the historical risk index before the fault and the multidimensional data at the time of the fault as inputs to the Bayesian network, and output the final root cause of the fault.

[0055] For example, a Bayesian network is constructed, which includes hypothesis nodes, evidence nodes, and directed edges connecting the hypothesis nodes and evidence nodes. Each evidence node corresponds to a conditional probability, and the parameters of the conditional probability can be calibrated based on historical fault data, domain expert knowledge, and system simulation. Each hypothesis node represents the root cause of a fault occurring during the operation of the power line.

[0056] Furthermore, historical risk indices and multidimensional data before the failure occurred, as well as multidimensional data at the time of the failure, are obtained. These are then substituted into the evidence nodes, and the conditional probabilities corresponding to the evidence nodes are invoked. Additionally, the prior probability of the hypothetical node is determined based on the historical risk index. For example, for the hypothetical node of insulation thermal aging breakdown... H j Its prior probability P ( H j The risk index of historical thermal stress at the failure point will be a function of the risk index, as shown in the following formula: ; Where P(Hj) is the prior probability, This is a historical risk index. Location of the fault. This is the starting time for risk accumulation calculation, ranging from one month to one year. The time when the fault occurred. In the time interval [ , The time variable that changes within the period. g ( ) is a mapping function that transforms the accumulated thermal risk history into a probability value. If the historical dynamic risk profile of a certain section shows that it has been under high thermal stress for a long time, the prior probability of the failure hypothesis related to thermal aging will be significantly increased.

[0057] Furthermore, the conditional probabilities of the evidence nodes are propagated along the network to the hypothetical nodes via directed edges. The prior probabilities of the hypothetical nodes are adjusted based on the conditional probabilities to determine their posterior probabilities. The final root cause of the failure is then determined based on the posterior probabilities.

[0058] As described above, this invention establishes a forward model based on physical laws to characterize the mapping relationship between multidimensional data and internal risk factors. Based on the forward model, an inverse problem model is determined. In the inverse problem model, multidimensional data is treated as known data, and internal risk factors of the power line are treated as data to be solved. The multidimensional data is input into the inverse problem model, and the model is solved based on a preset regularized objective function. The estimated values ​​of the internal risk factors of the power line are output, and a risk index is generated in real time based on the final estimated values ​​of the risk factors. This accurately identifies potential risks in specific sections, providing precise basis for operation and maintenance decisions. This invention achieves risk quantification of multi-source data through forward and inverse modeling, eliminating the shortcomings of existing technologies in qualitative analysis and delayed early warning of single physical phenomena. Furthermore, it can identify and assess potential risk factors in real time, enabling early warning before faults occur, thereby improving the overall safety and stability of operation.

[0059] Furthermore, after a fault occurs, the system can output the final root cause of the fault based on a pre-defined Bayesian network, combined with historical risk indices before the fault and multi-dimensional data at the time of the fault. By combining historical risk indices and real-time fault information, it can perform probabilistic reasoning on various potential fault causes and provide accurate diagnostic reports. This significantly improves the safety and stability of power line operation and reduces maintenance costs and time after a fault occurs.

[0060] like Figure 2 As shown, Figure 2 This is a flowchart illustrating another embodiment of the power line risk dynamic inversion and fault source tracing method provided by the present invention. The power line risk dynamic inversion and fault source tracing method includes: S201: Obtain multidimensional data along the power line.

[0061] S202: Establish a forward model based on physical laws to characterize the mapping relationship between multidimensional data and internal risk factors. Determine the inverse problem model based on the forward model. In the inverse problem model, multidimensional data is used as known data, and internal risk factors of power lines are used as data to be solved.

[0062] S203: Input multidimensional data into the inverse problem model, solve the inverse problem model based on the preset regularized objective function, and output the estimated value of the internal risk factor of the power line.

[0063] S204: Determine the risk index of power lines based on the estimated values ​​of risk factors.

[0064] It should be noted that steps S01-S204 are in Figure 1 The implementation scenarios shown have been discussed in detail and will not be repeated here.

[0065] S205: Construct a Bayesian network. The Bayesian network includes hypothesis nodes, evidence nodes, and directed edges connecting hypothesis nodes and evidence nodes. Each evidence node corresponds to a conditional probability, and each hypothesis node represents the root cause of a fault that occurs during the operation of a power line.

[0066] For example, a Bayesian network is constructed, which includes hypothesis nodes, evidence nodes, and directed edges connecting the hypothesis nodes and evidence nodes. Each evidence node corresponds to a conditional probability, and the parameters of the conditional probability can be calibrated based on historical fault data, domain expert knowledge, and system simulation. Each hypothesis node represents the root cause of a fault occurring during the operation of the power line.

[0067] A Bayesian network is a directed acyclic graph, as shown in the following equation: ; Where V represents a network node, indicating a random variable (the root cause of the failure), and E represents a directed edge, indicating a direct causal relationship between the variables. According to the chain rule of Bayesian networks, the joint probability distribution of all variables in the system can be decomposed as shown in the following equation: ; in, parents ( V i ) is a node V i The set of parent nodes. Each node V i They are all associated with a conditional probability table (CPT), which quantifies P ( V i | parents ( V i In a Bayesian network, which is a directed acyclic graph, nodes are categorized into two types: hypothesis nodes and evidence nodes. Hypothesis nodes, also known as root cause nodes, represent potential, undiagnosed root causes of faults. Examples include A-phase insulation breakdown, B-phase joint failure, external force damage, and sheath damage leading to water ingress. The probabilities of these nodes are the ultimate goal of tracing the source. Evidence nodes, also known as observation nodes, represent multidimensional data captured in previous steps, including the fault moment and the moment before the fault occurred. For example, the fault current waveform exhibits typical metallic short-circuit characteristics; distributed acoustic / vibration sensing recorded excavation vibration signals at a distance of 150 meters 10 minutes before the fault; and historical data collected by distributed temperature sensing showed a long-term abnormal temperature hotspot near 152 meters.

[0068] S206: Obtain historical risk indices and multidimensional data before the failure occurred, as well as multidimensional data at the moment of the failure.

[0069] For example, when a protective relay or intelligent electronic device detects a fault signal, such as an overcurrent or differential protection trip, it captures the following key data: (1) High-resolution fault recording data (current) from relays i ( t ),Voltage v ( t (waveform) (2) The last multidimensional data collected by all sensors just before the fault occurred.

[0070] (3) A complete historical risk index covering the faulty section.

[0071] S207: Determine the prior probability of the hypothetical node based on the historical risk index.

[0072] For example, the prior probability of a hypothetical node is determined based on a historical risk index. For instance, for the hypothetical node of insulation thermal aging breakdown. H j Its prior probability P ( H j The risk index of historical thermal stress at the failure point will be a function of the risk index, as shown in the following formula: ; Where P(Hj) is the prior probability and DRI is the risk index. For, the range of t0 is, for, g ( ) is a mapping function that transforms the accumulated thermal risk history into a probability value. If the historical dynamic risk profile of a certain section shows that it has been under high thermal stress for a long time, the prior probability of the failure hypothesis related to thermal aging will be significantly increased.

[0073] S208: Substitute the historical risk index before the failure and the multidimensional data at the time of the failure into the evidence node, and call the conditional probability corresponding to the evidence node.

[0074] S209: The conditional probability of the evidence node is transmitted along the network to the hypothesis node through directed edges. The prior probability of the hypothesis node is adjusted according to the conditional probability to determine the posterior probability of the hypothesis node.

[0075] For example, the historical risk index before the failure, the time of failure, and multidimensional data before the failure are substituted into the evidence node, and the conditional probability corresponding to the evidence node is invoked. Through directed edges, the conditional probability of the evidence node is passed along the network to the hypothesis node. The prior probability of the hypothesis node is adjusted according to the conditional probability to determine the posterior probability of the hypothesis node.

[0076] S210: Sort the posterior probabilities of each hypothetical node in descending order, and select the cause of failure corresponding to the hypothetical node with the highest posterior probability as the final root cause of failure.

[0077] For example, the posterior probabilities of each hypothetical node are sorted in descending order to obtain a detailed fault tracing report. The fault cause corresponding to the hypothetical node with the highest posterior probability is selected as the final root cause of the fault.

[0078] like Figure 3 As shown, Figure 3 This is a schematic diagram of an embodiment of a power line risk dynamic inversion and fault tracing system provided by the present invention. A power line risk dynamic inversion and fault tracing system 10, the system includes: The multidimensional data acquisition module 11 is used to acquire multidimensional data along the power line.

[0079] The inverse problem model determination module 12 is used to establish a forward model based on physical laws to characterize the mapping relationship between multidimensional data and internal risk factors, and to determine the inverse problem model based on the forward model. In the inverse problem model, the multidimensional data is used as known data, and the internal risk factors of the power line are used as data to be solved.

[0080] The risk factor estimation module 13 is used to input multidimensional data into the inverse problem model, solve the inverse problem model based on the preset regularized objective function, and output the estimated values ​​of the internal risk factors of the power line.

[0081] The risk index determination module 14 for power lines is used to determine the risk index of power lines based on the estimated values ​​of risk factors.

[0082] The root cause determination module 15 is used to obtain the historical risk index before the fault occurred and the multidimensional data at the time of the fault after the fault occurs. Based on the preset Bayesian network, the historical risk index before the fault occurred and the multidimensional data at the time of the fault are used as input to the Bayesian network to output the final root cause of the fault.

[0083] For example, in the multidimensional data acquisition module 11, multidimensional data along the power line is acquired. In the inverse problem model determination module 12, a forward model representing the mapping relationship between multidimensional data and internal risk factors is established based on physical laws. The inverse problem model is determined based on the forward model, where the multidimensional data is used as known data and the internal risk factors of the power line are used as data to be solved.

[0084] In module 13 for determining risk factor estimates, the regularization objective function is defined according to the following formula: ; in,F For positive operators, d obs For multidimensional data, || F ( m ) d obs ||2 2 For the residual term, || L ( m m prior )||2 2 It is a regularization term. m Let be the risk factor, mprior be the prior estimate of the risk factor, α be the regularization parameter, and L be the regularization operator.

[0085] Furthermore, the multidimensional data is input into the inverse problem model, and the inverse problem model is solved based on the regularized objective function to output the estimated values ​​of the internal risk factors of the power line. The inverse problem model is as follows: ; in, m est Let J(m) be the estimated value of the risk factor, and J(m) be the regularization objective function.

[0086] In the risk index determination module 14 for power lines, the estimated values ​​of risk factors are normalized, and the risk index of the power lines is determined according to the following formula: ; DRI stands for Risk Index. m ^ i ( x , t () represents the estimated value of the i-th risk factor after normalization at location x and time t. w i These are the weighting coefficients.

[0087] In the root cause determination module 15, a Bayesian network is constructed. The Bayesian network includes hypothesis nodes, evidence nodes, and directed edges connecting hypothesis nodes and evidence nodes. Each evidence node corresponds to a conditional probability, and each hypothesis node represents the root cause of a fault occurring during the operation of the power line. Historical risk indices and multidimensional data before the fault occurred, as well as multidimensional data at the time of the fault, are obtained. The prior probability of the hypothesis node is determined based on the historical risk index. The historical risk index before the fault occurred and the multidimensional data at the time of the fault are substituted into the evidence node, and the conditional probability corresponding to the evidence node is called. The conditional probability of the evidence node is passed along the network to the hypothesis node through the directed edges. The prior probability of the hypothesis node is adjusted according to the conditional probability to determine the posterior probability of the hypothesis node. The final root cause of the fault is determined based on the posterior probability.

[0088] like Figure 4 As shown, Figure 4 This is a schematic diagram of an embodiment of the device provided by the present invention. The device 20 includes a memory 21 and a processor 22. The memory 21 stores a computer program, and the processor 22 executes the computer program during operation to achieve, for example... Figure 1 and Figure 2 The method shown.

[0089] The specific technical details of the power line risk dynamic inversion and fault tracing method implemented by the above-mentioned device 20 when executing the computer program have been discussed in detail in the above method steps, so they will not be repeated here.

[0090] like Figure 5 As shown, Figure 5 This is a schematic diagram of the structure of an embodiment of the medium provided by the present invention. The medium 30 stores at least one computer program 31, which is executed by the processor 22 to perform the following... Figure 1 and Figure 2 The method shown is detailed above and will not be repeated here. In one embodiment, the medium 30 can be a storage chip, hard disk, portable hard disk, USB flash drive, optical disk, or other read / write storage device, or even a server, etc.

[0091] Furthermore, the processes depicted in the accompanying drawings do not necessarily have to be performed in the specific or sequential order shown to achieve the desired result. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0092] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the embodiments of apparatus, devices, and non-volatile computer-readable storage media are basically similar to the method embodiments, and therefore described more simply; relevant parts can be referred to the descriptions of the method embodiments.

[0093] The apparatus, device, non-volatile computer-readable storage medium and method provided in the embodiments of this specification are corresponding. Therefore, the apparatus, device and non-volatile computer storage medium also have similar beneficial technical effects as the corresponding method. Since the beneficial technical effects of the method have been described in detail above, the beneficial technical effects of the corresponding apparatus, device and non-volatile computer storage medium will not be repeated here.

[0094] The systems, devices, modules, or units described in the above embodiments can be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, a computer can be, for example, a personal computer, laptop computer, cellular phone, camera phone, smartphone, personal digital assistant, media player, navigation device, email device, game console, tablet computer, wearable device, or any combination of these devices.

[0095] For ease of description, the above apparatus is described by dividing it into various functional units. Of course, in implementing this specification, the functions of each unit can be implemented in one or more software and / or hardware components. Those skilled in the art will understand that the embodiments of this specification can be provided as methods, systems, or computer program products. Therefore, the embodiments of this specification can take the form of entirely hardware embodiments, entirely software embodiments, or embodiments combining software and hardware aspects. Furthermore, the embodiments of this specification can take the form of computer program products implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0096] This specification is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this specification. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create a machine for implementing the flowchart illustrations and / or block diagrams. Figure 1One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0097] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0098] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0099] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0100] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0101] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information by any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0102] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0103] This specification can be described in the general context of computer-executable instructions that are executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform a specific task or implement a specific abstract data type. This specification can also be practiced in distributed computing environments, where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0104] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments.

[0105] The above description discloses only preferred embodiments of the present invention and should not be construed as limiting the scope of the present invention. Therefore, equivalent variations made in accordance with the claims of the present invention are still within the scope of the present invention.

Claims

1. A method for dynamic inversion of power line risks and fault source tracing, characterized in that, The method includes: Acquire multidimensional data along power lines; A forward model is established based on physical laws to characterize the mapping relationship between multidimensional data and internal risk factors. An inverse problem model is determined based on the forward model. In the inverse problem model, the multidimensional data is taken as known data and the internal risk factors of the power line are taken as data to be solved. The multidimensional data is input into the inverse problem model, and the inverse problem model is solved based on the preset regularization objective function to output the estimated value of the internal risk factor of the power line. The risk index of the power line is determined based on the estimated values ​​of the risk factors. After a failure occurs, the historical risk index and multidimensional data before the failure, as well as the multidimensional data at the time of the failure, are obtained. Based on a preset Bayesian network, the historical risk index before the failure and the multidimensional data at the time of the failure are used as inputs to the Bayesian network to output the final root cause of the failure.

2. The method for dynamic inversion of power line risks and fault tracing according to claim 1, characterized in that, The process of inputting the multidimensional data into the inverse problem model, solving the inverse problem model based on a preset regularized objective function, and outputting the final estimated value of the internal risk factors of the power line specifically includes: according to Define the regularization objective function, where, F For positive operators, d obs For multidimensional data, || F ( m ) d obs ||2 2 For the residual term, || L ( m m prior )||2 2 It is a regularization term. m Let be the risk factor, mprior be the prior estimate of the risk factor, α be the regularization parameter, and L be the regularization operator; The multidimensional data is input into the inverse problem model, and the inverse problem model is solved based on the regularized objective function to output the estimated values ​​of the internal risk factors of the power line. The inverse problem model is as follows: ,in, m est Let J(m) be the estimated value of the risk factor, and J(m) be the regularization objective function.

3. The method for dynamic inversion of power line risks and fault tracing according to claim 1, characterized in that, The determination of the risk index of the power line based on the estimated value of the risk factor specifically includes: Normalize the estimated values ​​of the risk factors; according to Determine the risk index of power lines, where DRI is the risk index. m ^ i ( x , t () represents the estimated value of the i-th risk factor after normalization at location x and time t. w i These are the weighting coefficients.

4. The method for dynamic inversion of power line risks and fault tracing according to claim 1, characterized in that, After a fault occurs, the system acquires historical risk indices and multidimensional data prior to the fault, as well as multidimensional data at the moment of the fault. Based on a pre-defined Bayesian network, the system uses the historical risk indices prior to the fault and the multidimensional data at the moment of the fault as inputs to the Bayesian network, and outputs the final root cause of the fault. Specifically, this includes: Construct a Bayesian network, which includes hypothesis nodes, evidence nodes, and directed edges connecting the hypothesis nodes and evidence nodes. Each evidence node corresponds to a conditional probability, and each hypothesis node represents the root cause of a fault that occurs during the operation of a power line. Obtain historical risk indices and multidimensional data before the failure occurred, as well as multidimensional data at the moment of the failure; The prior probability of the hypothetical node is determined based on the historical risk index. Substitute the historical risk index and multidimensional data before the failure occurred, as well as the multidimensional data at the time of the failure, into the evidence node, and call the conditional probability corresponding to the evidence node. The conditional probability of the evidence node is transmitted along the network to the hypothesis node through the directed edge. The prior probability of the hypothesis node is adjusted according to the conditional probability to determine the posterior probability of the hypothesis node. The final root cause of the failure is determined based on the posterior probability.

5. The method for dynamic inversion of power line risks and fault tracing according to claim 4, characterized in that, The step of determining the prior probability of the hypothetical node based on the historical risk index specifically includes: according to Determine the prior probability of the hypothetical node, where P(Hj) is the prior probability, and g( ) is a mapping function. This is a historical risk index. Location of the fault. It is the starting time for risk accumulation calculation. The time when the fault occurred. In the time interval [ , A time variable that changes within a given period.

6. The method for dynamic inversion of power line risks and fault tracing according to claim 4, characterized in that, The determination of the final root cause of the failure based on the posterior probability specifically includes: The posterior probabilities of each hypothetical node are sorted in descending order, and the cause of failure corresponding to the hypothetical node with the highest posterior probability is selected as the final root cause of failure.

7. The method for dynamic inversion of power line risks and fault tracing according to claim 1, characterized in that, The process of establishing a forward model based on physical laws to characterize the mapping relationship between multidimensional data and internal risk factors, and determining the inverse problem model based on the forward model, further includes the following steps before the multidimensional data is treated as known data and the internal risk factors of the power line are treated as data to be solved: The multidimensional data is sequentially subjected to filtering and noise reduction, normalization, time synchronization, and feature extraction to obtain preprocessed multidimensional data.

8. A dynamic risk inversion and fault tracing system for power lines, characterized in that, The system includes: The multidimensional data acquisition module is used to acquire multidimensional data along the power line. The inverse problem model determination module is used to establish a forward model based on physical laws to characterize the mapping relationship between multidimensional data and internal risk factors, and to determine the inverse problem model based on the forward model. In the inverse problem model, the multidimensional data is used as known data, and the internal risk factors of the power line are used as data to be solved. The risk factor estimation module is used to input the multidimensional data into the inverse problem model, solve the inverse problem model based on a preset regularized objective function, and output the estimated values ​​of the internal risk factors of the power line. A risk index determination module for power lines is used to determine the risk index of power lines based on the estimated values ​​of the risk factors. The root cause determination module is used to acquire historical risk index and multidimensional data before the fault occurred, as well as multidimensional data at the time of the fault, after the fault occurs. Based on a preset Bayesian network, the historical risk index before the fault occurred and the multidimensional data at the time of the fault are used as inputs to the Bayesian network to output the final root cause of the fault.

9. A computer-readable storage medium storing a computer program that, when executed by a processor, causes the processor to perform the steps of the method as claimed in any one of claims 1 to 7.

10. A computer device comprising a memory and a processor, the memory storing a computer program that, when executed by the processor, causes the processor to perform the steps of the method as claimed in any one of claims 1 to 7.

Images