Fault prediction method and system for power transmission line under severe convective weather, and strategy generation method and system

Abstract

The application discloses a kind of strong convection weather under transmission line fault prediction, strategy generation method and system, obtain the grid meteorological forecast data of target area, extract strong convection weather parameter, and based on this, calculate the meteorological threat index of transmission line, divide transmission line into multiple line units, extract the disaster resistance resilience attribute of each line unit, fusion meteorological threat index of the meteorological grid point where line unit is located, predict the failure rate of each line unit by pre-trained machine learning model, determine the risk level of each line unit according to failure rate, generate and push corresponding coping strategy for each line unit in combination with pre-set risk level-disposal rule mapping library.The application realizes the accurate quantitative prediction of individual line unit failure rate by using pre-trained machine learning model, greatly improves the early warning accuracy, strategy executability and operation resource use efficiency, and provides effective technical support for ensuring the safe and stable operation of power grid under extreme weather.

Description

Technical Field

[0001] This invention relates to the field of smart grid technology, specifically to a method and system for predicting and generating strategies for transmission line faults under severe convective weather. Background Technology

[0002] With the intensification of global climate change, the frequency and intensity of severe convective weather events are on the rise, posing a serious threat to the safe and stable operation of power grids, especially transmission lines exposed to the elements. Extreme weather events such as strong winds, torrential rains, and hail can easily lead to faults such as line galloping, tower collapse, line breaks, and flashovers, causing widespread power outages and resulting in huge socio-economic losses.

[0003] Currently, the industry's commonly used early warning and response methods primarily rely on regional severe convective weather warnings issued by public meteorological departments. Based on these macro-level warnings and their own experience, maintenance personnel assess the risk level of lines within specific areas and formulate inspection and protection strategies accordingly. This approach is essentially an experience-based, extensive management model relying on regional early warnings and manual judgment.

[0004] Existing regional weather warning technologies cannot reflect weather differences at small geographical scales, nor can they distinguish the huge risk differences between different power transmission lines in the same area due to their own design, materials, and health conditions. This results in an overly large warning range, making it impossible to accurately locate the real high-risk points, and consequently, the generated response strategies lack specificity. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a method and system for predicting and generating strategies for transmission line faults under severe convective weather, with the aim of solving the problems in the background technology.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a method for predicting and generating strategies for transmission line faults under severe convective weather, comprising:

[0007] Step S10: Obtain grid meteorological forecast data in the target area containing the transmission line to be monitored, and extract severe convective weather parameters for multiple meteorological grid points in the grid meteorological forecast data within a future preset time period;

[0008] Step S20: Calculate the meteorological threat index of each meteorological grid point to the transmission line fault based on the severe convective weather parameters;

[0009] Step S30: Divide the transmission line into multiple line units, extract the disaster resilience attribute corresponding to each line unit, the disaster resilience attribute is determined based on the inherent attribute parameters of the line unit, the inherent attribute parameters include tower type, insulator type and number of strings, and line design wind speed level;

[0010] Step S40: Based on the disaster resilience attribute, the meteorological threat index of the meteorological grid at the location of the line unit is fused, and a pre-trained machine learning model is used to predict the unit failure rate of each line unit. The unit failure rate is then mapped to a predefined standardized failure rate range to obtain a standardized failure rate that uniformly measures the risk level of different line units.

[0011] Step S50: Determine the risk level of each line unit based on the standardized failure rate;

[0012] Step S60: Based on the preset risk level-disposal rule mapping library, generate and push the corresponding response strategy for each line unit. Detect whether there is a resource allocation conflict among the response strategies generated by multiple adjacent or related line units. When a conflict is detected, dynamically adjust the resource allocation to generate the optimal resource allocation scheme.

[0013] Furthermore, the specific process of step S40 is as follows:

[0014] The geometric coordinates of the line unit are spatially matched with the meteorological grid using the power grid geographic information system to determine at least one target meteorological grid point covering the line unit.

[0015] Identify the meteorological threat index of the target meteorological grid point, and fuse the meteorological threat index with the disaster resilience attribute of the line unit to form a feature vector of the comprehensive risk status of the line unit;

[0016] The feature vector is input into a pre-trained machine learning model, and the pre-trained machine learning model predicts and outputs the unit failure rate of the line unit. The unit failure rate is then mapped to a predefined standardized failure rate range to obtain the standardized failure rate.

[0017] The pre-trained machine learning model simultaneously quantifies the uncertainty of the unit failure rate based on the comprehensive risk status of the line unit corresponding to the feature vector, and outputs the confidence level of the failure prediction.

[0018] The meteorological threat index of the target meteorological grid point is identified, and the meteorological threat index is fused with the disaster resilience attribute of the line unit to form a feature vector of the comprehensive risk status of the line unit. The specific process is as follows:

[0019] When the line unit is covered by multiple target meteorological grid points, the meteorological threat index of each target meteorological grid point is assigned a weight based on the overlapping area or center point distance between each target meteorological grid point and the line unit. The weighted average algorithm is used to calculate the weighted average of the meteorological threat index assigned to the line unit as the final corrected meteorological threat index.

[0020] The non-numerical parameters in the disaster resilience attribute are quantified and encoded.

[0021] The final corrected meteorological threat index and all the disaster resilience attribute parameters that have been quantified and encoded are spliced ​​and fused together according to a preset feature order, and the spliced ​​and fused feature sequence is standardized to form a feature vector of the comprehensive risk status of the line unit.

[0022] Furthermore, the specific process of step S10 is as follows:

[0023] The target area containing the transmission line to be monitored is obtained, the target area is divided into grids to obtain multiple sub-grids, and then the meteorological forecast data of each sub-grid is obtained, that is, the grid meteorological forecast data.

[0024] Extract severe convective weather parameters for multiple meteorological grid points in the grid weather forecast data within a preset future time period.

[0025] Furthermore, the specific process of step S20 is as follows:

[0026] From the parameters of severe convective weather, key parameters corresponding to the fault indicators of transmission lines are selected;

[0027] Among them, the fault indicators of transmission lines include line wind deflection risk, lightning tripping probability, and tower overturning risk;

[0028] The key parameters corresponding to the fault indicators of transmission lines are normalized to obtain the parameters of each key disaster-causing factor.

[0029] The normalized parameters of each key disaster-causing factor are input into a preset nonlinear function model for fusion calculation to obtain the output result of the nonlinear fusion model.

[0030] A comprehensive, dimensionless meteorological threat index is generated based on the output of the nonlinear fusion model.

[0031] Furthermore, the specific process of step S30 is as follows:

[0032] Based on the topology of the power grid geographic information system, adjacent towers in the real physical location of the power grid are identified through the power grid geographic information system. The independent span between adjacent towers is used as a basic line unit, and the transmission line is divided into multiple line units.

[0033] After the line unit division is completed, the inherent disaster resilience attributes of each line unit are extracted. The disaster resilience attributes also include the years of operation and historical defect records.

[0034] Furthermore, the specific process of step S50 is as follows:

[0035] Define a risk probability range, which includes low risk, medium risk, high risk, and extremely high risk;

[0036] Among them, [0,0.01) corresponds to low risk, [0.01,0.05) corresponds to medium risk, [0.05,0.2) corresponds to high risk, and [0.2,1] corresponds to extremely high risk;

[0037] The unit failure rate is automatically compared with the preset risk probability range. If the value of the unit failure rate falls into the risk probability range, the unit failure rate is automatically assigned the corresponding risk level.

[0038] Furthermore, the specific process of step S60 is as follows:

[0039] After determining the risk level of each line unit based on the unit failure rate, a corresponding response strategy will be generated for each line unit based on the preset risk level-handling rule mapping library. The response strategy includes the inspection priority of each line unit by the power grid operation and maintenance unit, the key points of the specific handling measures for each line unit, and the list of required resource suggestions.

[0040] Based on the power grid geographic information system, the geometric coordinates of the line unit and the planned inspection path length are obtained. Combined with the material and manpower consumption corresponding to the resource suggestion list, and with reference to the weight allocation standard of the risk level-disposal rule mapping library, the mileage cost, material and manpower consumption cost corresponding to the inspection path length are weighted and calculated to obtain the path cost of resource scheduling.

[0041] Based on a pre-defined risk level-disposal rule mapping library, it detects whether there are resource allocation conflicts in the response strategies generated by multiple adjacent or related line units;

[0042] If no conflict is detected, the generated response strategy will be pushed out.

[0043] When a conflict is detected, the resource allocation in the response strategy is dynamically adjusted using an optimization algorithm based on the risk level of the line unit, the confidence level of the fault prediction, and the path cost of resource scheduling, so as to generate a globally optimized optimal resource allocation scheme.

[0044] Furthermore, when a conflict is detected, based on the risk level of the line unit, the confidence level of the fault prediction, and the path cost of resource scheduling, the resource allocation in the response strategy is dynamically adjusted using an optimization algorithm to generate a globally optimized optimal resource allocation scheme. The specific process is as follows:

[0045] A multi-objective optimization function is constructed with the objectives of minimizing the path cost of resource scheduling and maximizing the handling priority of high-risk line units;

[0046] The total amount of all allocated resources in the power line operation and maintenance scenario, the upper limit of resources allocated to line units whose fault prediction confidence is lower than the preset threshold, and the deadline for operation and maintenance tasks carried out for line units with fault risks are respectively used as constraints.

[0047] The multi-objective optimization function is solved using a genetic algorithm or a particle swarm optimization algorithm to obtain the optimal resource allocation scheme. The response strategies of the relevant line units are then updated based on the optimal resource allocation scheme, and the updated response strategies are pushed out.

[0048] A system for predicting and generating strategies for transmission line faults under severe convective weather conditions, applied to methods for predicting and generating strategies for transmission line faults under severe convective weather conditions, including:

[0049] The data acquisition module is used to acquire grid meteorological forecast data in the target area containing the transmission line to be monitored, and extract severe convective weather parameters for multiple meteorological grid points in the grid meteorological forecast data within a future preset time period.

[0050] The first processing module is used to calculate the meteorological threat index of each meteorological grid point to the transmission line fault based on the severe convective weather parameters.

[0051] The second processing module is used to divide the transmission line into multiple line units and extract the disaster resilience attribute corresponding to each line unit. The disaster resilience attribute is determined based on the inherent attribute parameters of the line unit, including tower type, insulator type and number of strings, and line design wind speed level.

[0052] The third processing module is used to fuse the meteorological threat index of the meteorological grid points at the location of the line unit based on the disaster resilience attribute, use a pre-trained machine learning model to predict the unit failure rate of each line unit, and map the unit failure rate to a predefined standardized failure rate range to obtain a standardized failure rate that uniformly measures the risk level of different line units.

[0053] The fourth processing module is used to determine the risk level of each line unit based on the standardized failure rate;

[0054] The fifth processing module is used to generate and push response strategies for each line unit based on a preset risk level-disposal rule mapping library, detect whether there are resource allocation conflicts among the response strategies generated by multiple adjacent or related line units, and dynamically adjust the resource allocation to generate the optimal resource allocation scheme when a conflict is detected.

[0055] A computer-readable storage medium having computer instructions stored thereon, which are executed by a processor, comprising steps for predicting transmission line faults and generating strategies under severe convective weather.

[0056] Compared with existing technologies, the present invention has the following advantages:

[0057] The method described in this invention divides the power grid into line units, calculates a gridded meteorological threat index, and integrates disaster resilience attributes such as tower type and insulator configuration. Using a pre-trained machine learning model, it achieves accurate quantitative prediction of the failure rate of individual line units, enabling risk warnings to be precise from a broad perspective to specific details. Furthermore, this invention automatically matches risk levels to response strategies that include specific inspection paths, handling measures, and resource lists, generating an overall optimal grid-based integrated work order. This significantly improves the accuracy of early warnings, the executability of strategies, and the efficiency of operation and maintenance resource utilization, providing effective technical support for ensuring the safe and stable operation of the power grid under extreme weather conditions. Attached Figure Description

[0058] Figure 1 This is a flowchart of a method for predicting and generating transmission line faults under severe convective weather, as shown in an embodiment of the present invention.

[0059] Figure 2 This is a structural block diagram of a power transmission line fault prediction and strategy generation system under severe convective weather, as shown in an embodiment of the present invention. Detailed Implementation

[0060] To make the objectives, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Several embodiments of the present invention are shown in the drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided so that the disclosure of the present invention will be more thorough and complete.

[0061] It should be noted that when a component is said to be "fixed to" another component, it can be directly on the other component or there may be an intervening component. When a component is said to be "connected to" another component, it can be directly connected to the other component or there may be an intervening component. The terms "vertical," "horizontal," "left," "right," and similar expressions used in this document are for illustrative purposes only.

[0062] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the description of the invention is for the purpose of particularly describing embodiments only and is not intended to be limiting of the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0063] Example 1

[0064] Please see Figure 1 The first embodiment of the present invention provides a method for predicting transmission line faults and generating strategies under severe convective weather, including:

[0065] Step S10: Obtain grid meteorological forecast data in the target area containing the transmission line to be monitored, and extract severe convective weather parameters for multiple meteorological grid points in the grid meteorological forecast data within a future preset time period.

[0066] Specifically, in this embodiment, by acquiring grid meteorological forecast data in the target area containing the transmission line to be monitored, it is possible to reflect weather differences at a small geographical scale; at the same time, by extracting severe convective weather parameters of multiple meteorological grid points in the future preset time period from the grid meteorological forecast data, it is possible to achieve refined and spatiotemporal capture of meteorological conditions in the target area, providing accurate and comprehensive raw meteorological data support for subsequent accurate calculation of the meteorological threat index.

[0067] Step S20: Calculate the meteorological threat index of each meteorological grid point to the transmission line fault based on the severe convective weather parameters.

[0068] Specifically, in this embodiment, by analyzing and calculating the parameters of severe convective weather, a corresponding quantitative index of fault threat is assigned to each meteorological grid point, thereby realizing the quantitative expression of meteorological disaster risk and establishing a quantitative correlation between meteorological conditions and transmission line fault risk.

[0069] Step S30: Divide the transmission line into multiple line units and extract the disaster resilience attribute corresponding to each line unit. The disaster resilience attribute is determined based on the inherent attribute parameters of the line unit, including tower type, insulator type and number of strings, and line design wind speed level.

[0070] It should be noted that step S30 is the fine division and inherent attribute extraction of the transmission line, which realizes the disassembly of the transmission line from the macroscopic whole to the microscopic unit, and at the same time completes the quantitative characterization of the disaster resistance capability of the line unit.

[0071] Specifically, in this embodiment, the first step is to divide the transmission line into multiple line units, enabling fault prediction and risk assessment to be accurate down to specific line segments. Secondly, the disaster resilience attribute is determined based on the inherent attribute parameters of the line unit, including at least the tower type, insulator type and number of strings, and line design wind speed level. These inherent attribute parameters are the core factors that determine the line unit's ability to withstand severe convective weather damage, directly affecting the line unit's failure probability under the same meteorological threat. Extracting the disaster resilience attribute enables an accurate characterization of the line unit's own disaster resistance capability, providing inherent attribute support for subsequent comprehensive risk assessment by integrating the meteorological threat index.

[0072] Step S40: Based on the disaster resilience attribute, the meteorological threat index of the meteorological grid at the location of the line unit is fused, and a pre-trained machine learning model is used to predict the unit failure rate of each line unit. The unit failure rate is then mapped to a predefined standardized failure rate range to obtain a standardized failure rate that uniformly measures the risk level of different line units.

[0073] Step S50: Determine the risk level of each line unit based on the standardized failure rate.

[0074] Specifically, in this embodiment, based on the numerical value of the standardized failure rate, a corresponding risk level is matched for each line unit according to preset rules, realizing the conversion of the standardized failure rate from numerical value to level. This enables power grid operation and maintenance personnel to intuitively and quickly identify the risk level of the line unit, establishing a link between risk level and disposal measures for subsequent targeted response strategies, and ensuring that the formulation of disposal strategies matches the actual risk level of the line unit.

[0075] Step S60: Based on the preset risk level-disposal rule mapping library, generate and push the corresponding response strategy for each line unit. Detect whether there is a resource allocation conflict among the response strategies generated by multiple adjacent or related line units. When a conflict is detected, dynamically adjust the resource allocation to generate the optimal resource allocation scheme.

[0076] Specifically, firstly, a corresponding response strategy is generated for each line unit based on a preset risk level-handling rule mapping library to ensure the strategy's relevance and practicality. Simultaneously, a resource allocation conflict detection and dynamic adjustment mechanism is added to detect resource conflicts between strategies of adjacent or related line units. When a conflict is detected, resource allocation is dynamically adjusted to generate the optimal resource allocation scheme. This solves the problem of traditional strategy generation methods only considering a single line unit and ignoring global resource optimization, achieving reasonable allocation and optimal global utilization of operation and maintenance resources, and improving the executability of strategies and the efficiency of operation and maintenance resource utilization.

[0077] It is important to note that, in this embodiment, the specific process of step S40 is as follows:

[0078] The geometric coordinates of the line unit are spatially matched with the meteorological grid using the power grid geographic information system to determine at least one target meteorological grid point covering the line unit.

[0079] Identify the meteorological threat index of the target meteorological grid point, and fuse the meteorological threat index with the disaster resilience attribute of the line unit to form a feature vector of the comprehensive risk status of the line unit;

[0080] The feature vector is input into a pre-trained machine learning model, and the pre-trained machine learning model predicts and outputs the unit failure rate of the line unit. The unit failure rate is then mapped to a predefined standardized failure rate range to obtain the standardized failure rate.

[0081] The pre-trained machine learning model simultaneously quantifies the uncertainty of the unit failure rate based on the comprehensive risk status of the line unit corresponding to the feature vector, and outputs the confidence level of the failure prediction.

[0082] The meteorological threat index of the target meteorological grid point is identified, and the meteorological threat index is fused with the disaster resilience attribute of the line unit to form a feature vector of the comprehensive risk status of the line unit. The specific process is as follows:

[0083] When the line unit is covered by multiple target meteorological grid points, the meteorological threat index of each target meteorological grid point is assigned a weight based on the overlapping area or center point distance between each target meteorological grid point and the line unit. The weighted average algorithm is used to calculate the weighted average of the meteorological threat index assigned to the line unit as the final corrected meteorological threat index.

[0084] The non-numerical parameters in the disaster resilience attribute are quantified and encoded.

[0085] The final corrected meteorological threat index and all the disaster resilience attribute parameters that have been quantified and encoded are spliced ​​and fused together according to a preset feature order, and the spliced ​​and fused feature sequence is standardized to form a feature vector of the comprehensive risk status of the line unit.

[0086] It is easy to understand that step S40 achieves deep integration of meteorological threat index and disaster resilience attributes of line unit, completes accurate prediction of unit failure rate through pre-trained machine learning model, and realizes standardization of failure rate and quantification of prediction uncertainty.

[0087] The specific process of this step is as follows: First, based on the power grid geographic information system, the geometric coordinates of the line unit are spatially matched with the meteorological grid to determine the target meteorological grid points covering the line unit. This solves the problem of spatial scale matching between the target meteorological grid points and the line unit, ensuring that each line unit is matched with the corresponding accurate meteorological threat index. Then, the meteorological threat index of the target meteorological grid points is identified and fused with the disaster resilience attribute to form a feature vector of comprehensive risk status. The construction of this feature vector realizes the comprehensive representation of external meteorological threats and internal disaster resistance capabilities, and fully reflects the comprehensive failure risk of the line unit.

[0088] In the construction of feature vectors, if a line unit is covered by multiple target meteorological grid points, weights are assigned based on the overlapping area or center point distance of the target meteorological grid points, and a weighted average is used to obtain a corrected meteorological threat index. This operation further improves the matching accuracy between the meteorological threat index and the line unit, avoiding the one-sidedness of single grid point data. At the same time, non-numerical disaster resilience parameters (including attributes that exist in the form of categories, such as tower type, insulator type and number of strings) are quantized and encoded. Then, the corrected meteorological threat index is spliced ​​and fused with the quantized non-numerical disaster resilience parameters and standardized to ensure that all parameters in the feature vector are on the same dimension and numerical scale, avoiding the pre-trained machine learning model from over-focusing on parameters with large values ​​and ensuring the effectiveness of the feature vector. The feature vectors are then input into a pre-trained machine learning model, which outputs the unit failure rate and maps it to a predefined standardized failure rate range to obtain the standardized failure rate. This standardization operation eliminates the differences in basic failure rates of lines in different regions and at different voltage levels, achieving a unified scale for measuring the failure risk of all line units and making the risk levels of different line units comparable. At the same time, the pre-trained machine learning model can also quantify the uncertainty of the unit failure rate and output the confidence level of the failure prediction, realizing the assessment of the reliability of the unit failure rate. This provides an important reference for the resource allocation of subsequent strategy generation and solves the problem that traditional prediction methods only output a single failure rate and do not have a reliability assessment.

[0089] Compared with existing technologies, the method for predicting transmission line faults and generating strategies under severe convective weather as shown in this embodiment has the following advantages:

[0090] By dividing transmission lines into line units, calculating a gridded meteorological threat index, and integrating disaster resilience attributes such as tower type and insulator configuration, a pre-trained machine learning model is used to achieve accurate quantitative prediction of individual line unit failure rates, enabling risk warnings to be precise from a broad perspective to specific details. Based on this, the invention automatically matches risk levels with response strategies that include specific inspection paths, handling measures, and resource lists, generating an overall optimal grid-based integrated work order. This significantly improves the accuracy of early warnings, the executability of strategies, and the efficiency of operation and maintenance resource utilization, providing effective technical support for ensuring the safe and stable operation of the power grid under extreme weather conditions.

[0091] Example 2

[0092] The second embodiment of the present invention also provides a method for predicting transmission line faults and generating strategies under severe convective weather, which is basically similar to the method shown in the first embodiment, except that:

[0093] The specific process of step S10 is as follows:

[0094] The target area containing the transmission line to be monitored is obtained, the target area is divided into grids to obtain multiple sub-grids, and then the meteorological forecast data of each sub-grid is obtained, that is, the grid meteorological forecast data.

[0095] Extract severe convective weather parameters for multiple meteorological grid points in the grid weather forecast data within a preset future time period.

[0096] Specifically, in this embodiment, the target area containing the transmission line to be monitored is first obtained, then the target area is divided into multiple sub-grids, and finally the meteorological forecast data of each sub-grid is obtained. This process realizes the gridded decomposition of the meteorological forecast data of the target area, ensuring that each sub-grid has corresponding accurate meteorological forecast data, and avoiding the ambiguity of traditional regional meteorological forecasts.

[0097] Secondly, after acquiring the grid-based meteorological forecast data, the process further clarifies how to extract severe convective weather parameters for multiple meteorological grid points within a predetermined time period. This sub-step echoes the grid division operation, enabling refined data extraction layer by layer from region to sub-grid to meteorological grid point. This ensures that the extracted severe convective weather parameters can accurately correspond to specific spatial locations, providing precise meteorological data support for subsequent spatial matching with line units.

[0098] The specific process of step S20 is as follows:

[0099] From the parameters of severe convective weather, key parameters corresponding to the fault indicators of transmission lines (such as wind speed that induces line wind deflection, lightning intensity that causes lightning tripping, and ice thickness that causes tower overturning) are screened out.

[0100] Among them, the fault indicators of transmission lines include line wind deflection risk, lightning tripping probability, and tower overturning risk;

[0101] The key parameters corresponding to the fault indicators of transmission lines are normalized to obtain the parameters of each key disaster-causing factor.

[0102] The normalized parameters of each key disaster-causing factor are input into a preset nonlinear function model for fusion calculation to obtain the output result of the nonlinear fusion model.

[0103] A comprehensive, dimensionless meteorological threat index is generated based on the output of the nonlinear fusion model.

[0104] Specifically, in this embodiment, the first step is to screen key parameters. Specifically, key parameters corresponding to transmission line fault indicators are selected from severe convective weather parameters. The core of this operation is to eliminate irrelevant meteorological parameters and retain disaster-causing parameters that have a direct impact on transmission line fault indicators. This avoids irrelevant parameters interfering with the calculation results, ensures the relevance and effectiveness of subsequent calculations, and establishes a direct correlation between meteorological parameters and transmission line fault indicators, making the calculation of the meteorological threat index more consistent with the actual fault patterns of transmission lines.

[0105] Then, normalization is performed on the selected key parameters to obtain key disaster-causing factor parameters. Since different severe convective weather parameters have different dimensions and numerical ranges, such as wind speed in m / s and rainfall in mm, direct fusion calculation will lead to distorted results due to differences in dimensions and numerical ranges. Normalization can transform all key parameters into dimensionless data with the same dimension and numerical range, eliminating the influence of dimensions and numerical ranges, laying a data foundation for subsequent fusion calculations, and ensuring that each key disaster-causing factor parameter has an equal weight in the fusion calculation.

[0106] Furthermore, nonlinear fusion calculations are performed by inputting the normalized parameters of key disaster-causing factors into a pre-set nonlinear function model for fusion calculation, yielding the output results of the nonlinear fusion model. Under severe convective weather, the destructive effects of multiple disaster-causing factors on transmission lines are not simply linear superpositions, but rather exhibit a compound enhancement effect. For example, the destructive force on lines when strong winds are accompanied by heavy rainfall is far greater than the sum of the effects of a single factor. The pre-set nonlinear function model can accurately characterize this compound enhancement effect of multiple factors working together. Compared with traditional linear fusion methods, it better reflects the actual disaster-causing patterns of severe convective weather, ensuring that the fusion calculation results can truly reflect the meteorological threat level under the combined action of multiple disaster-causing factors.

[0107] Finally, and most importantly, the generation of the meteorological threat index involves generating a comprehensive, dimensionless meteorological threat index based on the output of the nonlinear fusion model. Through the final standardization of the output of the nonlinear fusion model, the generated dimensionless index enables direct comparison of the threat level between different meteorological grid points. At the same time, it makes the meteorological threat index a standardized quantitative indicator that can be directly used for subsequent fault prediction. This realizes the transformation from multi-factor fusion calculation results to a single, intuitive, and comparable meteorological threat index, providing a standardized meteorological threat index for subsequent fusion with the disaster resilience attributes of line units.

[0108] The specific process of step S30 is as follows:

[0109] Based on the topology of the power grid geographic information system, adjacent towers in the real physical location of the power grid are identified through the power grid geographic information system. The independent span between adjacent towers is used as a basic line unit, and the transmission line is divided into multiple line units.

[0110] After the line unit division is completed, the inherent disaster resilience attributes of each line unit are extracted. The disaster resilience attributes also include the years of operation and historical defect records.

[0111] Specifically, in this embodiment, the specific criteria and process for dividing line units are clarified: based on the topology of the power grid geographic information system, adjacent towers in the actual physical location of the power grid are identified, and the independent span between adjacent towers is used as a basic line unit to divide the transmission line into multiple line units.

[0112] The topology of the power grid geographic information system can accurately reflect the actual physical location and tower distribution of transmission lines. Based on this, the division of line units is ensured to be consistent with the actual physical structure of the transmission lines. The independent span of adjacent towers is used as the basic unit, so that each line unit has relative independence in both physical and electrical aspects, which facilitates independent risk assessment and operation and maintenance management.

[0113] Secondly, after completing the division of line units, the inherent disaster resilience attributes of each line unit are extracted. These attributes also include the service life and historical defect records. For example, clearly defined disaster resilience attributes include at least the tower type, insulator type and number of strings, and the line design wind speed level. These are inherent design attributes of the line unit. However, the service life and historical defect records are attributes formed during the actual operation of the line unit, directly reflecting its actual operating status and health. For example, line units with longer service lives are prone to equipment aging, and line units with historical defects will have reduced disaster resilience. Including the service life and historical defect records in the scope of disaster resilience attributes achieves a comprehensive characterization of the design and operational attributes of the line unit's disaster resilience, which is more in line with the actual disaster resilience of the line unit. This makes subsequent fault prediction based on the meteorological threat index more accurate and more in line with the actual operating patterns of transmission lines.

[0114] In a specific embodiment, step S50 is performed as follows:

[0115] Define a risk probability range, which includes low risk, medium risk, high risk, and extremely high risk;

[0116] Among them, [0,0.01) corresponds to low risk, [0.01,0.05) corresponds to medium risk, [0.05,0.2) corresponds to high risk, and [0.2,1] corresponds to extremely high risk;

[0117] The unit failure rate is automatically compared with the preset risk probability range. If the value of the unit failure rate falls into the risk probability range, the unit failure rate is automatically assigned the corresponding risk level.

[0118] Specifically, in this embodiment, the risk probability range is divided into four levels: low risk, medium risk, high risk, and extremely high risk, thus constructing a four-level risk level system. This classification method meets the actual needs of power grid operation and maintenance, and can accurately distinguish the risk level corresponding to different fault probabilities. This enables operation and maintenance personnel to quickly identify the risk level of line units, providing a clear risk basis for subsequent targeted handling strategies.

[0119] The system then automatically compares the unit failure rate with a preset risk probability range. The risk level is automatically assigned to the corresponding line unit based on which risk probability range the unit failure rate falls into. This automates risk level determination, avoiding the subjectivity and errors of manual comparison. It ensures consistent risk level determination results for the same unit failure rate across different scenarios, improving the efficiency and accuracy of risk level assessment. Furthermore, it seamlessly integrates the determination process with subsequent strategy generation, laying the foundation for automated strategy generation.

[0120] Furthermore, the specific process of step S60 is as follows:

[0121] After determining the risk level of each line unit based on the unit failure rate, a corresponding response strategy will be generated for each line unit based on the preset risk level-handling rule mapping library. The response strategy includes the inspection priority of each line unit by the power grid operation and maintenance unit, the key points of the specific handling measures for each line unit, and the list of required resource suggestions.

[0122] Based on the power grid geographic information system, the geometric coordinates of the line unit and the planned inspection path length are obtained. Combined with the material and manpower consumption corresponding to the resource suggestion list, and with reference to the weight allocation standard of the risk level-disposal rule mapping library, the mileage cost, material and manpower consumption cost corresponding to the inspection path length are weighted and calculated to obtain the path cost of resource scheduling.

[0123] Based on a pre-defined risk level-disposal rule mapping library, it detects whether there are resource allocation conflicts in the response strategies generated by multiple adjacent or related line units;

[0124] If no conflict is detected, the generated response strategy will be pushed out.

[0125] When a conflict is detected, the resource allocation in the response strategy is dynamically adjusted using an optimization algorithm based on the risk level of the line unit, the confidence level of the fault prediction, and the path cost of resource scheduling, so as to generate a globally optimized optimal resource allocation scheme.

[0126] Specifically, in this embodiment, after determining the risk level of a line unit, a response strategy is generated for each line unit based on a preset risk level-handling rule mapping library. The response strategy includes inspection priorities, key points of specific handling measures, and a resource suggestion list. The risk level-handling rule mapping library is built based on historical experience and actual needs of power grid operation and maintenance, achieving precise matching between risk levels and handling strategies, ensuring that line units with different risk levels can receive targeted handling. It also clarifies the three core elements of the response strategy, making the strategy not only include "what to do," but also "what to do first" and "what resources are needed," enabling operation and maintenance personnel to directly carry out operation and maintenance work according to the strategy, greatly improving the executability of the strategy.

[0127] Next, the resource scheduling path cost is calculated. Specifically, the geometric coordinates of the line units and the planned inspection path length are obtained from the power grid geographic information system. Combined with the material and manpower consumption in the resource recommendation list and referring to the weight allocation standard, the mileage cost, material and manpower consumption cost are weighted and calculated to obtain the resource scheduling path cost. This step combines geospatial information (geometric coordinates of line units and planned inspection path length) with resource consumption (material and manpower consumption in the resource recommendation list) to achieve quantitative calculation of resource scheduling costs, providing a quantitative cost basis for subsequent resource allocation conflict adjustment. At the same time, by comprehensively considering various cost factors through weighted calculation, it ensures that the cost calculation results can truly reflect the actual cost of resource scheduling, making subsequent resource optimization and allocation more in line with actual operation and maintenance scenarios.

[0128] Finally, resource allocation conflict detection is performed. Specifically, based on a pre-defined risk level-handling rule mapping library, it checks whether there are resource allocation conflicts in the response strategies generated by multiple adjacent or related line units. Adjacent or related line units share some operational resources during maintenance. If a strategy is generated only for a single line unit, conflicts such as overlapping resource requirements and insufficient resource allocation are very likely to occur. This step, by detecting conflicts in advance, can promptly identify problems in resource configuration, laying the foundation for subsequent dynamic adjustments and preventing maintenance work from being unable to proceed due to resource conflicts.

[0129] Specifically, when no resource allocation conflict is detected, the generated response strategy is pushed directly to ensure that the strategy can be implemented quickly; when a conflict is detected, the resource allocation scheme will be dynamically adjusted using optimization algorithms based on the risk level of the line unit, the confidence level of the fault prediction, and the path cost of resource scheduling, to generate the globally optimized optimal resource allocation scheme.

[0130] More specifically, when a conflict is detected, the resource allocation in the response strategy is dynamically adjusted using an optimization algorithm based on the risk level of the line unit, the confidence level of the fault prediction, and the path cost of resource scheduling, in order to generate a globally optimized optimal resource allocation scheme. The specific process is as follows:

[0131] A multi-objective optimization function is constructed with the objectives of minimizing the path cost of resource scheduling and maximizing the handling priority of high-risk line units;

[0132] The total amount of all allocated resources in the power line operation and maintenance scenario, the upper limit of resources allocated to line units whose fault prediction confidence is lower than the preset threshold, and the deadline for operation and maintenance tasks carried out for line units with fault risks are respectively used as constraints.

[0133] The multi-objective optimization function is solved using a genetic algorithm or a particle swarm optimization algorithm to obtain the optimal resource allocation scheme. The response strategies of the relevant line units are then updated based on the optimal resource allocation scheme, and the updated response strategies are pushed out.

[0134] Specifically, in this embodiment, the objectives are to minimize the path cost of resource scheduling and maximize the handling priority of high-risk line units. This objective setting takes into account the two core operation and maintenance principles of "optimal cost" and "risk priority," which can achieve low-cost utilization of operation and maintenance resources while ensuring that high-risk line units are handled with priority, thus meeting the actual needs of power grid operation and maintenance. At the same time, the two core objectives are integrated into a multi-objective optimization function, realizing the synergistic optimization of multiple objectives and avoiding the problem of neglecting one aspect for another caused by single-objective optimization, ensuring that the optimization result can meet the comprehensive needs of power grid operation and maintenance.

[0135] And it clarifies the three major constraints of the multi-objective optimization function:

[0136] The total amount constraint on resource allocation ensures that the resource allocation plan does not exceed the actual available resource range and is practically feasible;

[0137] The upper limit constraint on the resources allocated to line units with a fault prediction confidence level lower than a preset threshold avoids allocating too many resources to line units with low reliability of unit failure rate, ensuring that the total amount of resources allocated is tilted towards line units with more reliable unit failure rate, thereby improving resource utilization efficiency.

[0138] The deadline constraints for maintenance and repair tasks carried out on line units with fault risks ensure that the resource allocation plan can complete the repair work within the specified time and meet the timeliness requirements of transmission line maintenance and repair under severe convective weather.

[0139] The three constraints restrict the optimization process from three dimensions: resources, reliability, and time, to ensure that the optimization results not only conform to actual operation and maintenance capabilities, but also take into account timeliness and reliability.

[0140] Finally, the multi-objective optimization function is solved using a genetic algorithm or particle swarm optimization algorithm to obtain the optimal resource allocation scheme. Based on this scheme, the response strategies of the relevant line units are updated, and the updated response strategies are pushed out.

[0141] It should be noted that genetic algorithms and particle swarm optimization algorithms are mature intelligent optimization algorithms that can efficiently solve multi-objective optimization problems and quickly find the global optimum. Compared with traditional manual adjustment methods, they have the characteristics of high solution efficiency and better optimization results. At the same time, the complete process from solving for the optimal solution to updating the strategy and then pushing the strategy is clearly defined, realizing the seamless integration of resource allocation optimization with strategy generation and push. This ensures that the optimized optimal resource allocation scheme can be implemented quickly, resolve resource conflict problems, achieve the global optimal utilization of operation and maintenance resources, and improve the efficiency and effectiveness of transmission line operation and maintenance under severe convective weather.

[0142] Example 3

[0143] Please see Figure 2 The third embodiment of the present invention provides a transmission line fault prediction and strategy generation system under severe convective weather, applied to the transmission line fault prediction and strategy generation method under severe convective weather described in any embodiment, including:

[0144] The data acquisition module is used to acquire grid meteorological forecast data in the target area containing the transmission line to be monitored, and extract severe convective weather parameters for multiple meteorological grid points in the grid meteorological forecast data within a future preset time period.

[0145] The first processing module is used to calculate the meteorological threat index of each meteorological grid point to the transmission line fault based on the severe convective weather parameters.

[0146] The second processing module is used to divide the transmission line into multiple line units and extract the disaster resilience attribute corresponding to each line unit. The disaster resilience attribute is determined based on the inherent attribute parameters of the line unit, including tower type, insulator type and number of strings, and line design wind speed level.

[0147] The third processing module is used to fuse the meteorological threat index of the meteorological grid points at the location of the line unit based on the disaster resilience attribute, use a pre-trained machine learning model to predict the unit failure rate of each line unit, and map the unit failure rate to a predefined standardized failure rate range to obtain a standardized failure rate that uniformly measures the risk level of different line units.

[0148] The fourth processing module is used to determine the risk level of each line unit based on the standardized failure rate;

[0149] The fifth processing module is used to generate and push response strategies for each line unit based on a preset risk level-disposal rule mapping library, detect whether there are resource allocation conflicts among the response strategies generated by multiple adjacent or related line units, and dynamically adjust the resource allocation to generate the optimal resource allocation scheme when a conflict is detected.

[0150] Example 4

[0151] The fourth embodiment of the present invention provides a computer-readable storage medium storing computer instructions thereon, which, when executed by a processor, implement the steps of the transmission line fault prediction and strategy generation method under severe convective weather as described in any embodiment.

[0152] Example 5

[0153] The fifth embodiment of the present invention provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the steps of the method for predicting transmission line faults and generating strategies under severe convective weather as described in any embodiment.

[0154] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

Claims

1. A method for predicting and generating strategies for transmission line faults under severe convective weather, characterized in that, include: Step S10: Obtain grid meteorological forecast data in the target area containing the transmission line to be monitored, and extract severe convective weather parameters for multiple meteorological grid points in the grid meteorological forecast data within a future preset time period; Step S20: Calculate the meteorological threat index of each meteorological grid point to the transmission line fault based on the severe convective weather parameters; From the parameters of severe convective weather, key parameters corresponding to the fault indicators of transmission lines are selected; Among them, the fault indicators of transmission lines include line wind deflection risk, lightning tripping probability, and tower overturning risk; key parameters include wind speed that induces line wind deflection, lightning intensity that causes lightning tripping, and ice thickness that causes tower overturning. The key parameters corresponding to the fault indicators of transmission lines are normalized to obtain the parameters of each key disaster-causing factor. The normalized parameters of each key disaster-causing factor are input into a preset nonlinear function model for fusion calculation to obtain the output result of the nonlinear fusion model. A comprehensive, dimensionless meteorological threat index is generated based on the output of the nonlinear fusion model. Step S30: Based on the topology of the power grid geographic information system, adjacent towers in the actual physical location of the power grid are identified through the power grid geographic information system. The independent span between adjacent towers is used as a basic line unit. The transmission line is divided into multiple line units. The disaster resilience attribute corresponding to each line unit is extracted. The disaster resilience attribute is determined based on the inherent attribute parameters of the line unit. The inherent attribute parameters include tower type, insulator type and number of strings, and line design wind speed level. Step S40: Spatial matching of the geometric coordinates of the line unit with the meteorological grid is performed through the power grid geographic information system to determine at least one target meteorological grid point covering the line unit. Based on the disaster resilience attribute, the meteorological threat index of the meteorological grid point of the target meteorological grid point at the location of the line unit is fused to form a feature vector of the comprehensive risk status of the line unit. A pre-trained machine learning model is used to predict the unit failure rate of each line unit, and the unit failure rate is mapped to a predefined standardized failure rate range to obtain a standardized failure rate that uniformly measures the risk level of different line units. Step S50: Determine the risk level of each line unit based on the standardized failure rate; Step S60: After determining the risk level of each line unit based on the unit failure rate, a corresponding response strategy for each line unit will be generated based on the preset risk level-handling rule mapping library. The response strategy includes the inspection priority of each line unit by the power grid operation and maintenance unit, the key points of the specific handling measures for each line unit, and the list of required resource suggestions. Based on the power grid geographic information system, the geometric coordinates of the line unit and the planned inspection path length are obtained. Combined with the material and manpower consumption corresponding to the resource suggestion list, and with reference to the weight allocation standard of the risk level-disposal rule mapping library, the mileage cost, material and manpower consumption cost corresponding to the inspection path length are weighted and calculated to obtain the path cost of resource scheduling. Based on a pre-defined risk level-disposal rule mapping library, it detects whether there are resource allocation conflicts in the response strategies generated by multiple adjacent or related line units; If no conflict is detected, the generated response strategy will be pushed out. When a conflict is detected, the resource allocation in the response strategy is dynamically adjusted using an optimization algorithm based on the risk level of the line unit, the confidence level of the fault prediction, and the path cost of resource scheduling, so as to generate a globally optimized optimal resource allocation scheme.

2. The method for predicting and generating strategies for transmission line faults under severe convective weather as described in claim 1, characterized in that: The specific process of step S40 is as follows: The geometric coordinates of the line unit are spatially matched with the meteorological grid using the power grid geographic information system to determine at least one target meteorological grid point covering the line unit. Identify the meteorological threat index of the target meteorological grid point, and fuse the meteorological threat index with the disaster resilience attribute of the line unit to form a feature vector of the comprehensive risk status of the line unit; The feature vector is input into a pre-trained machine learning model, and the pre-trained machine learning model predicts and outputs the unit failure rate of the line unit. The unit failure rate is then mapped to a predefined standardized failure rate range to obtain the standardized failure rate. The pre-trained machine learning model simultaneously quantifies the uncertainty of the unit failure rate based on the comprehensive risk status of the line unit corresponding to the feature vector, and outputs the confidence level of the failure prediction. The meteorological threat index of the target meteorological grid point is identified, and the meteorological threat index is fused with the disaster resilience attribute of the line unit to form a feature vector of the comprehensive risk status of the line unit. The specific process is as follows: When the line unit is covered by multiple target meteorological grid points, the meteorological threat index of each target meteorological grid point is assigned a weight based on the overlapping area or center point distance between each target meteorological grid point and the line unit. The weighted average algorithm is used to calculate the weighted average of the meteorological threat index assigned to the line unit as the final corrected meteorological threat index. The non-numerical parameters in the disaster resilience attribute are quantified and encoded. The final corrected meteorological threat index and all the disaster resilience attribute parameters that have been quantified and encoded are spliced ​​and fused together according to a preset feature order, and the spliced ​​and fused feature sequence is standardized to form a feature vector of the comprehensive risk status of the line unit.

3. The method for predicting and generating strategies for transmission line faults under severe convective weather as described in claim 1, characterized in that, The specific process of step S10 is as follows: The target area containing the transmission line to be monitored is obtained, the target area is divided into grids to obtain multiple sub-grids, and then the meteorological forecast data of each sub-grid is obtained, that is, the grid meteorological forecast data. Extract severe convective weather parameters for multiple meteorological grid points in the grid weather forecast data within a preset future time period.

4. The method for predicting and generating transmission line faults under severe convective weather as described in claim 3, characterized in that, The specific process of step S30 is as follows: Based on the topology of the power grid geographic information system, adjacent towers in the real physical location of the power grid are identified through the power grid geographic information system. The independent span between adjacent towers is used as a basic line unit, and the transmission line is divided into multiple line units. After the line unit division is completed, the inherent disaster resilience attributes of each line unit are extracted. The disaster resilience attributes also include the years of operation and historical defect records.

5. The method for predicting and generating strategies for transmission line faults under severe convective weather as described in claim 4, characterized in that, The specific process of step S50 is as follows: Define a risk probability range, which includes low risk, medium risk, high risk, and extremely high risk; Among them, [0,0.01) corresponds to low risk, [0.01,0.05) corresponds to medium risk, [0.05,0.2) corresponds to high risk, and [0.2,1] corresponds to extremely high risk; The unit failure rate is automatically compared with the preset risk probability range. If the value of the unit failure rate falls into the risk probability range, the unit failure rate is automatically assigned the corresponding risk level.

6. The method for predicting and generating transmission line faults under severe convective weather as described in claim 5, characterized in that, When a conflict is detected, the resource allocation in the response strategy is dynamically adjusted using an optimization algorithm based on the risk level of the line unit, the confidence level of the fault prediction, and the path cost of resource scheduling, in order to generate a globally optimized optimal resource allocation scheme. The specific process is as follows: A multi-objective optimization function is constructed with the objectives of minimizing the path cost of resource scheduling and maximizing the handling priority of high-risk line units. The total amount of all allocated resources in the power line operation and maintenance scenario, the upper limit of resources allocated to line units whose fault prediction confidence is lower than the preset threshold, and the deadline for operation and maintenance tasks carried out for line units with fault risks are respectively used as constraints. The multi-objective optimization function is solved using a genetic algorithm or a particle swarm optimization algorithm to obtain the optimal resource allocation scheme. The response strategies of the relevant line units are then updated based on the optimal resource allocation scheme, and the updated response strategies are pushed out.

7. A system for predicting and generating strategies for transmission line faults under severe convective weather, applied to the method for predicting and generating strategies for transmission line faults under severe convective weather as described in any one of claims 1-6, characterized in that, include: The data acquisition module is used to acquire grid meteorological forecast data in the target area containing the transmission line to be monitored, and extract severe convective weather parameters for multiple meteorological grid points in the grid meteorological forecast data within a future preset time period. The first processing module is used to calculate the meteorological threat index of each meteorological grid point to the transmission line fault based on the severe convective weather parameters. The second processing module is used to divide the transmission line into multiple line units and extract the disaster resilience attribute corresponding to each line unit. The disaster resilience attribute is determined based on the inherent attribute parameters of the line unit, including tower type, insulator type and number of strings, and line design wind speed level. The third processing module is used to fuse the meteorological threat index of the meteorological grid points at the location of the line unit based on the disaster resilience attribute, use a pre-trained machine learning model to predict the unit failure rate of each line unit, and map the unit failure rate to a predefined standardized failure rate range to obtain a standardized failure rate that uniformly measures the risk level of different line units. The fourth processing module is used to determine the risk level of each line unit based on the standardized failure rate; The fifth processing module is used to generate and push response strategies for each line unit based on a preset risk level-disposal rule mapping library, detect whether there are resource allocation conflicts among the response strategies generated by multiple adjacent or related line units, and dynamically adjust the resource allocation to generate the optimal resource allocation scheme when a conflict is detected.

8. A computer-readable storage medium storing computer instructions thereon, characterized in that, When executed by the processor, this instruction implements the steps of the method for predicting and generating transmission line faults under severe convective weather as described in any one of claims 1-6.

Images