Insulator structure damage prediction method and system under adverse weather

Abstract

The application discloses a method and system for predicting damage of an insulator structure under severe weather, and relates to the field of power system fault detection and diagnosis. The method comprises the following steps: constructing a transition resistance dynamic time-varying model based on three stages of insulator crack breakdown, grain boundary breakdown and grain breakdown; obtaining historical severe environment data and historical damage data of the insulator, extracting environment features in the historical severe environment data and damage features in the historical damage data of the insulator, and constructing a correlation between the environment features and the damage features; obtaining a damage feature correction coefficient based on real-time environment data, the correlation between the environment features and the damage features, updating the transition resistance dynamic time-varying model according to the damage feature correction coefficient; and obtaining an insulator structure damage prediction result based on the updated transition resistance dynamic time-varying model and real-time insulator electric parameters. The application has the beneficial effects of improving the accuracy and efficiency of insulator damage prediction and damage positioning under severe environment.

Description

Technical Field

[0001] This application relates to the field of power system fault detection and diagnosis, and in particular to a method and system for predicting damage to insulator structures under severe weather conditions. Background Technology

[0002] Insulators, as primary insulation devices, are widely used in overhead power distribution lines. During operation, insulators are often subjected to electrical, mechanical, and thermal effects, leading to damage or aging. These defective insulators, whose insulation performance has deteriorated due to aging, are prone to breakdown when exposed to moisture or overvoltage. Modeling the dynamic characteristics of insulator breakdown faults is of great significance for experimental simulation and fault detection.

[0003] In related technologies, insulator condition assessment methods mainly rely on manual inspection, offline testing, and threshold judgments based on single electrical parameters (such as leakage current and insulation resistance). These methods have significant limitations: on the one hand, manual inspection is inefficient, costly, and difficult to detect early minor defects; on the other hand, judgments based on static thresholds cannot effectively adapt to the performance changes of insulators under different environmental conditions, easily leading to misjudgments or omissions. In addition, although early studies attempted to establish the correlation between environmental factors and insulator damage, they mostly used statistical methods, analyzing the relationship between the two only at the overall data level, ignoring the time-series characteristics of environmental changes and insulator performance degradation.

[0004] The patent "Method and Device for Damage Detection of Contact Network Insulators Based on Few-Sample Transfer Learning," publication number CN120032192A, published on May 23, 2025, specifically discloses a method that includes inputting an image of the contact network insulator to be detected into a two-level contrastive proposal coding network; an encoder performing feature encoding and feature extraction on the insulator image to extract the image's feature representation; an adaptive feature fusion module obtaining the optimal fusion weights through adaptive weight learning; a hierarchical feature extraction module receiving the output of the adaptive feature fusion module and extracting local position information of small objects in the shallow network of the encoder, global information of large objects in the deep network, and intermediate semantic information between local and global information in the intermediate network layer; and an inter-class and intra-class contrastive proposal coding module receiving the output of the hierarchical feature extraction module and calculating the semantic similarity between different categories, thereby achieving damage detection of contact network insulators in complex environments and high-speed motion scenarios. However, in harsh environments, the prediction and location of insulator damage still suffer from misjudgment or omission. Summary of the Invention

[0005] This application addresses the problems of low accuracy and efficiency in the prediction and location of insulator damage under harsh environments in existing technologies. It provides a method and system for predicting insulator structural damage under severe weather conditions. By modifying the dynamic time-varying model of transition resistance through the correlation between environmental characteristics and damage characteristics, the influence of harsh environments on the decreasing trend of transition resistance is incorporated. Furthermore, by time-varying the dynamic time-varying model of transition resistance, time-series-based prediction and location of insulator damage under harsh environments are achieved, thereby improving the accuracy and efficiency of insulator damage prediction and location.

[0006] To achieve the above technical objectives, this application provides a technical solution based on a method for predicting insulator structural damage under severe weather conditions, comprising the following steps: constructing a dynamic time-varying model of transition resistance based on the three stages of insulator crack breakdown, grain boundary breakdown, and grain breakdown; acquiring historical severe environmental data and historical damage data of the insulator, extracting environmental features from the historical severe environmental data and damage features from the historical damage data, and constructing a correlation between environmental features and damage features; obtaining damage feature correction coefficients based on real-time environmental data and the correlation between environmental features and damage features, updating the dynamic time-varying model of transition resistance according to the damage feature correction coefficients; and obtaining the insulator structural damage prediction result using the updated dynamic time-varying model of transition resistance and real-time insulator electrical parameters.

[0007] Furthermore, the construction of the dynamic time-varying model of transition resistance based on the three stages of insulator crack breakdown-grain boundary breakdown-grain breakdown includes: constructing a time-varying model of pore resistance based on the insulator crack breakdown stage; constructing a time-varying model of grain boundary resistance based on the insulator grain boundary breakdown stage; constructing a time-varying model of grain resistance based on the insulator grain breakdown stage; and constructing a dynamic time-varying model of transition resistance based on the time-varying models of pore resistance, grain boundary resistance, and grain resistance.

[0008] Furthermore, the acquisition of historical severe environmental data and historical damage data of insulators, the extraction of environmental features from the historical severe environmental data and damage features from the historical damage data, and the construction of the correlation between environmental features and damage features include: acquiring historical damage data of insulators based on the dynamic time-varying model of transition resistance and historical operating data of insulators; acquiring historical severe environmental data of insulators based on the severe environmental factor threshold and historical environmental data of insulators; acquiring environmentally damaged sets and non-environmentally damaged sets based on the time-series correspondence between historical damage data and historical severe environmental data of insulators; iteratively updating the initial time-series relationship of the dynamic time-varying model of transition resistance using the non-environmentally damaged set and the dynamic time-varying model of transition resistance; and constructing the correlation between environmental features and damage features based on the environmentally damaged set using a machine learning algorithm.

[0009] Furthermore, the acquisition of historical insulator damage data based on the dynamic time-varying model of transition resistance and historical insulator operation data includes: obtaining the insulator operation data and corresponding insulator breakdown states for the matching time sequence based on the matching relationship between the transition resistance at each time sequence in the historical insulator operation data and the transition resistance under different breakdown conditions in the dynamic time-varying model of transition resistance; and using the insulator operation data and corresponding insulator breakdown states for the matching time sequence as historical insulator damage data.

[0010] Furthermore, the step of obtaining the environmental damage set and the non-environmental damage set based on the time-series correspondence between the historical damage data of insulators and the historical severe environment data of insulators includes: taking the historical damage data of insulators that matches the time-series of the historical severe environment data of insulators as the environmental damage set; and taking the historical damage data of insulators that does not match the time-series of the historical severe environment data of insulators as the non-environmental damage set.

[0011] Furthermore, the iterative update of the initial time series relationship of the dynamic time-varying model of the transition resistance using the non-environmentally damaged set and the transition resistance dynamic time-varying model includes: constructing the transition resistance time series features based on the changes in the transition resistance at adjacent times in the non-environmentally damaged set; performing iterative optimization on the initial time series relationship of the dynamic time-varying model of the transition resistance based on the transition resistance time series features using an optimization algorithm, and updating the time series parameters of the dynamic time-varying model of the transition resistance.

[0012] Furthermore, the method of constructing the correlation between environmental features and damage features based on the environmental damage set using machine learning algorithms includes: obtaining the temporal relationship of damage features based on historical damage data of insulators in the environmental damage set; obtaining the temporal relationship of environmental features based on severe environmental data of insulators in the environmental damage set; and using machine learning algorithms to obtain the correlation between environmental features and damage features based on the temporal relationship between damage features and the temporal relationship between environmental features.

[0013] Furthermore, the step of obtaining the damage feature correction coefficient based on real-time environmental data, the correlation between environmental features and damage features, and updating the dynamic time-varying model of transition resistance based on the damage feature correction coefficient includes: obtaining environmental data for the prediction period and the trend of damage feature changes during the prediction period based on real-time environmental data, the correlation between environmental features and damage features; obtaining the damage feature correction coefficient based on the trend of damage feature changes during the prediction period, and updating the dynamic time-varying model of transition resistance.

[0014] Furthermore, the step of obtaining the insulator structure damage prediction result based on the updated dynamic time-varying model of transition resistance and real-time insulator electrical parameters includes: obtaining the transition resistance of the current time series based on the real-time insulator electrical parameters; obtaining the transition resistance of the prediction period based on the transition resistance of the current time series and the updated dynamic time-varying model of transition resistance; and obtaining the insulator structure damage prediction result based on the relationship between the transition resistance and the breakdown stage.

[0015] Another technical solution provided in this application is an insulator structure damage prediction system based on severe weather conditions, used to implement the above-mentioned method, comprising: a model building unit for constructing a dynamic time-varying model of transition resistance based on the three stages of insulator crack breakdown-grain boundary breakdown-grain breakdown; a data acquisition unit for acquiring historical severe environment data, historical damage data, real-time environmental data, and real-time insulator electrical parameters; a model updating unit for constructing the correlation between environmental features and damage features based on historical severe environment data and historical damage data, and updating the dynamic time-varying model of transition resistance according to the correlation between environmental features and damage features; and a damage prediction unit for obtaining the insulator structure damage prediction result based on the updated dynamic time-varying model of transition resistance and real-time insulator electrical parameters.

[0016] The beneficial effects of this application are as follows: By introducing a time variable, the nonlinear change in resistance from a high-resistivity state to a low-resistivity state during the three processes of crack breakdown, grain boundary breakdown, and grain breakdown is dynamically captured, thereby improving the accuracy of insulator internal breakdown assessment. Simultaneously, historical data on severe environmental conditions and damage to insulators are obtained, revealing potential patterns between environmental factors and insulator damage. Furthermore, the correction coefficients of the dynamic time-varying model of transition resistance are dynamically updated based on real-time environmental data, improving the accuracy of insulator damage prediction and adapting to damage differences under various weather conditions. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating the insulator structure damage prediction method based on severe weather conditions proposed in this application.

[0018] Figure 2 This is a schematic diagram of the time-varying curve of the transition resistance under one implementation of the insulator structure damage prediction method based on severe weather conditions in this application. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description of this application is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely one preferred embodiment of this application and are only used to explain this application. They do not limit the scope of protection of this application. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0020] like Figure 1 As shown in Embodiment 1 of this application, the method for predicting insulator structure damage under severe weather conditions includes the following steps: A dynamic time-varying model of transition resistance is constructed based on the three stages of insulator crack breakdown, grain boundary breakdown and grain breakdown. Acquire historical severe environmental data and historical damage data of insulators, extract environmental features from the historical severe environmental data and damage features from the historical damage data, and construct the correlation between environmental features and damage features; Based on real-time environmental data and the correlation between environmental features and damage features, the damage feature correction coefficient is obtained, and the dynamic time-varying model of transition resistance is updated according to the damage feature correction coefficient. The updated dynamic time-varying model of transition resistance and real-time insulator electrical parameters are used to obtain the predicted results of insulator structure damage.

[0021] In this embodiment, by introducing a time variable, the nonlinear change in resistance from a high-resistivity state to a low-resistivity state during the three processes of crack breakdown, grain boundary breakdown, and grain breakdown is dynamically captured, thereby improving the accuracy of insulator internal breakdown assessment. Simultaneously, historical severe environmental and damage data of the insulator are acquired to uncover potential patterns between environmental factors and insulator damage. Furthermore, the correction coefficients of the dynamic time-varying model of transition resistance are dynamically updated based on real-time environmental data, improving the accuracy of insulator damage prediction and adapting to damage differences under various weather conditions.

[0022] Specifically, the dynamic time-varying model of transition resistance based on the three stages of insulator crack breakdown, grain boundary breakdown, and grain breakdown includes: A time-varying model of pore resistance is constructed based on the insulator crack breakdown stage; A time-varying model of grain boundary resistance is constructed based on the grain boundary breakdown stage of insulators; A time-varying model of grain resistance is constructed based on the grain breakdown stage of insulators; A dynamic time-varying model of transition resistance is constructed based on the time-varying models of pore resistance, grain boundary resistance, and grain resistance.

[0023] During the insulator breakdown process, the transition resistance consists of three series-connected pore resistance, grain boundary resistance, and grain resistance. Therefore, time-varying models are constructed based on the time-varying characteristics of pore resistance, grain boundary resistance, and grain resistance in the three stages of crack breakdown, grain boundary breakdown, and grain breakdown, respectively.

[0024] The time-varying model of pore resistance based on the insulator crack breakdown stage is as follows: ; in, Indicates pore resistance. This indicates the pore resistance when the insulator has not broken down. This represents the pore resistance when the insulator breaks down through a channel, and t represents the timing. This indicates the fracture breakdown sequence.

[0025] Because the crack breakdown time is too short, a time-varying model of pore resistance is constructed based on the step function. Before the crack breakdown time sequence is reached, when no crack breakdown has occurred inside the insulator, the pore resistance is in an insulating state, and at this time the pore resistance is... When the fracture breakdown timing is reached, the fracture breaks down instantaneously, forming a conductive channel, and the pore resistance jumps to a certain value. At the instant the fracture breaks down, the pore resistance changes from an extremely high resistance state to a low resistance state. In this embodiment, , .

[0026] The time-varying model of grain boundary resistance based on the grain boundary breakdown stage of insulators is as follows: ; in, Indicates grain boundary resistance. This represents the grain boundary resistance when the insulator has no breakdown path. Indicates the grain boundary breakdown timing. This indicates the grain boundary breakdown time.

[0027] Before reaching the grain boundary breakdown timeline, the grain boundary resistance remains in an initial high-resistivity state, at which point the grain boundary resistance is... When the grain boundary breakdown time sequence is reached, the grain boundary gradually breaks down under the influence of an electric field or thermal effect, and the grain boundary resistance decreases over time. In this embodiment, , Indicates the duration of fracture breakdown. This indicates the time interval between fracture breakdown and grain boundary breakdown. .

[0028] Based on the insulator grain breakdown stage, a time-varying model of grain resistance is constructed as follows: ; in, Indicates grain resistance. This indicates the grain resistance when the insulator has no breakdown path. Indicates the grain breakdown sequence. This indicates the grain breakdown time.

[0029] Before reaching the grain breakdown timing, the grain resistance remains in an initial high-resistance state, at which point the grain resistance is... When the grain breakdown timing is reached, the grain gradually breaks down under the influence of an electric field or thermal effect, and the grain resistance decreases over time. In this embodiment, , This indicates the time interval between grain boundary breakdown and grain breakdown. .

[0030] The transition resistance is obtained by superimposing pore resistance, grain boundary resistance, and grain resistance in series, and a dynamic time-varying model of the transition resistance is constructed. For example... Figure 2 The figure shows the time-varying curves of the transition resistance in three stages, output by the dynamic time-varying model of the transition resistance under one implementation scenario. It can be understood that the pore resistance, grain boundary resistance, and grain resistance values, initial breakdown sequence, and initial breakdown duration of the insulator before breakdown can be obtained based on the actual insulator material and expert experience.

[0031] Acquire historical severe environmental data and historical damage data of insulators, extract environmental features from the historical severe environmental data and damage features from the historical damage data, and construct the correlation between environmental features and damage features, including: Historical damage data of insulators is obtained based on the dynamic time-varying model of transition resistance and historical operating data of insulators. Historical severe environmental data of insulators are obtained based on the threshold of severe environmental factors and historical environmental data of insulators. Based on the time-series correspondence between historical insulator damage data and historical severe environment data, we obtain the environmental damage set and the non-environmental damage set; The initial time series relationship of the transition resistance dynamic time-varying model is iteratively updated using the non-environmentally damaged set and the transition resistance dynamic time-varying model; Based on machine learning algorithms, the correlation between environmental features and damaged features is constructed according to the set of environmental damage.

[0032] During heavy rain, rainwater washes over the surface of insulators, dissolving dirt and forming a conductive water film, thus reducing surface resistance. Simultaneously, the high-speed falling raindrops exert mechanical impact on the insulator surface, potentially causing micro-cracks and increasing the risk of crack breakdown. In icy and snowy weather, the insulator surface becomes covered with ice and snow. These layers not only increase the mechanical load on the insulator but also alter its electric field distribution. As the ice layer thickens, air gaps and cracks may appear within it, easily becoming the starting points for breakdown under the influence of the electric field. In strong winds, insulators experience significant mechanical stress, which may cause loosening at the connection between the insulator hardware and the porcelain components, even leading to mechanical damage to the porcelain components and the propagation of internal cracks. Furthermore, the continuous impact of sand and dust particles on the insulator surface gradually wears down the glaze layer, compromising its insulation performance and making the insulator more susceptible to breakdown failures. Therefore, in this embodiment, the thresholds for severe environmental factors include at least the rainfall threshold, temperature threshold, humidity threshold, wind speed threshold, ice thickness threshold, and dust concentration threshold.

[0033] Historical damage data of insulators, obtained based on the dynamic time-varying model of transition resistance and historical operating data of insulators, includes: Based on the matching relationship between the transition resistance at each time sequence in the historical operation data of the insulator and the transition resistance under different breakdown conditions in the dynamic time-varying model of the transition resistance, the insulator operation data of the matching time sequence and the corresponding insulator breakdown state are obtained. The insulator operation data matched with the timing sequence and the corresponding insulator breakdown state are used as the historical damage data of the insulator.

[0034] In this embodiment, by constructing a breakdown mechanism model (dynamic time-varying model of transition resistance), the time series of breakdown events that have occurred in the historical operating data of the insulator are obtained. The insulator operating data and the corresponding insulator breakdown state under this time series are used as the historical damage data of the insulator. Based on the transition resistance characteristics of different stages such as crack breakdown, grain boundary breakdown, and grain breakdown in the dynamic time-varying model of transition resistance, the actual transition resistance of each time series in the historical operating data is compared with that of the insulator. When the trend of transition resistance change in the historical data matches a certain stage of the model, the breakdown state of the insulator at that moment can be immediately locked, thus eliminating the need for manual screening of damaged data and automatically updating the historical damage data of the insulator.

[0035] Meanwhile, based on whether there are severe environmental factors in the historical environmental data of insulators and whether the severe environmental factor threshold is met, the historical severe environmental data of insulators is obtained.

[0036] Based on the time-series correspondence between historical insulator damage data and historical severe environment data, the sets of environmentally damaged data and non-environmentally damaged data are obtained, including: The historical damage data of insulators, which is matched with the historical severe environmental data of insulators in a time series, and the historical severe environmental data of insulators are used as the environmental damage set; Historical damage data of insulators that do not match the time series of historical severe environmental data are considered as non-environmental damage data.

[0037] Damaged data that strictly matches harsh environmental conditions is categorized into an environmental damage set, improving the accuracy of constructing the relationship between environmental factors and insulator damage and avoiding the influence of other non-environmental factors. Simultaneously, by analyzing data from non-environmental damage sets, the aging and damage patterns of the insulator's internal structure are obtained, improving the efficiency and accuracy of fault diagnosis.

[0038] The initial time series relationship of the transition resistance dynamic time-varying model is iteratively updated using the non-environmentally damaged set and the transition resistance dynamic time-varying model, including: The temporal characteristics of transition resistance are constructed by analyzing the changes in transition resistance at adjacent time points in the set of non-environmentally damaged environments. Based on the optimization algorithm, the initial timing relationship of the dynamic time-varying model of the transition resistance is iteratively optimized according to the timing characteristics of the transition resistance, and the timing parameters of the dynamic time-varying model of the transition resistance are updated.

[0039] In this embodiment, by iteratively updating the initial time-series relationship based on the non-environmentally damaged set, the dynamic time-varying model of transition resistance can more accurately describe the change process of transition resistance of insulators under the influence of non-environmental factors such as aging and manufacturing defects, thereby improving the accuracy of the dynamic time-varying model of transition resistance. It is understood that when screening historical damage data of insulators based on the dynamic time-varying model of transition resistance under the initial time-series relationship, similarity of transition resistance and similarity of abrupt changes are used for matching. If similar transition resistances or similar abrupt changes exist, it is considered that there is an insulator breakdown phenomenon, without comparing the breakdown time series. Here, however, the time-series relationship of the dynamic time-varying model of transition resistance is updated using the differences in the change time series of insulator transition resistance under different time series in the non-environmentally damaged set. Damaged data is selected through relatively fixed transition resistances and abrupt changes, while the dynamic breakdown time series is updated through the non-environmentally damaged set, improving the adaptability of the breakdown time series of the dynamic time-varying model of transition resistance.

[0040] In this embodiment, the optimization algorithm can be gradient descent, genetic algorithm, etc.

[0041] Based on machine learning algorithms, the association between environmental features and damaged features is constructed from a set of environmental damage characteristics, including: The temporal relationship of damage characteristics is obtained based on the historical damage data of insulators in the environmental damage set; Temporal relationships of environmental characteristics are obtained from severe environmental data of insulators in the damaged environmental set; Machine learning algorithms are used to obtain the correlation between environmental features and damaged features based on the temporal relationship between damaged features and environmental features.

[0042] Based on the trend curves of transition resistance decreasing over time in the historical damage data of insulators within the environmental damage set, and the trend curves of environmental characteristic data changing over time in the severe environmental data of insulators within the environmental damage set, temporal relationships of damage features and environmental features are constructed respectively. These acquired temporal relationships are aligned and integrated. Using sequence modeling algorithms in machine learning, with temporal data of environmental features as input and temporal data of damage features as output labels, the model is trained to learn the mapping relationship between the two. By mining the temporal relationship between the environment and damage features, the dynamic causal link between environmental changes and insulator damage can be accurately captured, avoiding misjudgments due to ignoring the time factor. Through the correlation between environmental features and damage features, the short-term damage state of insulators can be predicted based on real-time environmental data, improving the efficiency of insulator damage prediction. In this embodiment, machine learning algorithms such as recurrent neural networks and gated recurrent units can be selected.

[0043] Based on real-time environmental data and the correlation between environmental characteristics and damage characteristics, damage characteristic correction coefficients are obtained. The dynamic time-varying model of transition resistance is then updated according to these damage characteristic correction coefficients, including: Based on real-time environmental data and the correlation between environmental characteristics and damage characteristics, environmental data for the prediction period and the trend of damage characteristics during the prediction period are obtained. Based on the trend of damage characteristics changes during the prediction period, the damage characteristic correction coefficient is obtained, and the dynamic time-varying model of transition resistance is updated.

[0044] Based on real-time environmental data and the established correlation between environmental and damage characteristics, the impact of environmental changes on insulator damage at different times is predicted in real time. Damage characteristic correction coefficients are obtained promptly based on this trend. For example, if the current environmental data corresponds to an accelerated decrease in transition resistance, the damage characteristic correction coefficient is obtained based on the acceleration ratio, and the dynamic time-varying model of transition resistance is updated to improve the accuracy of the dynamic time-varying model in predicting transition resistance and thus improve the accuracy of insulator damage prediction.

[0045] The predicted damage results of the insulator structure were obtained using the updated dynamic time-varying model of the transition resistance and real-time insulator electrical parameters, including: The transition resistance of the current time sequence is obtained based on the real-time electrical parameters of the insulator; The transition resistance for the predicted period is obtained based on the current transition resistance and the updated dynamic time-varying model of the transition resistance. The relationship between transition resistance and breakdown stage is used to obtain the prediction results of insulator structure damage.

[0046] By collecting insulator voltage, current, and other electrical parameters in real time and combining Ohm's law, the current time-series transition resistance is quickly calculated, instantly reflecting the insulator's current operating status. Then, based on the updated dynamic time-varying model of transition resistance, the damaged parts and conditions of the insulator in future periods are predicted, facilitating timely and accurate maintenance. It can be understood that the updated dynamic time-varying model of transition resistance can be a model updated based on damage characteristic correction coefficients, or it can be a model obtained by first updating the transition resistance dynamic time-varying model using a non-environmental damage set and the current dynamic time-varying model, and then updating the model based on the damage characteristic correction coefficients.

[0047] As a second embodiment of this application, the insulator structure damage prediction system under severe weather conditions includes: The model building unit is used to construct a dynamic time-varying model of transition resistance based on the three stages of insulator crack breakdown, grain boundary breakdown and grain breakdown. The data acquisition unit is used to acquire historical severe environmental data, historical damage data, real-time environmental data, and real-time electrical parameters of the insulators. The model update unit is used to construct the correlation between environmental features and damage features based on historical severe environmental data and historical damage data of insulators, and update the dynamic time-varying model of transition resistance according to the correlation between environmental features and damage features. The damage prediction unit is used to obtain the damage prediction results of the insulator structure based on the updated dynamic time-varying model of the transition resistance and the real-time electrical parameters of the insulator.

[0048] In this embodiment, the model building unit is connected to the model updating unit, the model updating unit is connected to the data acquisition unit, and the damage prediction unit is connected to both the data acquisition unit and the model updating unit.

[0049] The specific embodiments described above are preferred embodiments of the insulator structure damage prediction method and system under severe weather conditions in this application, and are not intended to limit the specific scope of this application. The scope of this application includes but is not limited to the specific embodiments described above. All equivalent changes made in accordance with the shape and structure of this application are within the protection scope of this application.

Claims

1. A method for predicting insulator structure damage under severe weather conditions, characterized by: Includes the following steps: A dynamic time-varying model of transition resistance is constructed based on the three stages of insulator crack breakdown, grain boundary breakdown and grain breakdown. Acquire historical severe environmental data and historical damage data of insulators, extract environmental features from the historical severe environmental data and damage features from the historical damage data, and construct the correlation between environmental features and damage features; Based on real-time environmental data and the correlation between environmental features and damage features, the damage feature correction coefficient is obtained, and the dynamic time-varying model of transition resistance is updated according to the damage feature correction coefficient. The damage prediction results of the insulator structure were obtained by using the updated dynamic time-varying model of the transition resistance and the real-time electrical parameters of the insulator. The process of acquiring historical severe environmental data and historical damage data of insulators, extracting environmental features from the historical severe environmental data and damage features from the historical damage data, and constructing the correlation between environmental features and damage features includes: Historical damage data of insulators is obtained based on the dynamic time-varying model of transition resistance and historical operating data of insulators. Historical severe environmental data of insulators are obtained based on the threshold of severe environmental factors and historical environmental data of insulators. Based on the time-series correspondence between historical insulator damage data and historical severe environment data, we obtain the environmental damage set and the non-environmental damage set; The initial time series relationship of the transition resistance dynamic time-varying model is iteratively updated using the non-environmentally damaged set and the transition resistance dynamic time-varying model; Based on machine learning algorithms, the correlation between environmental features and damaged features is constructed according to the set of environmental damage.

2. The method for predicting insulator structure damage under severe weather conditions as described in claim 1, characterized in that: The dynamic time-varying model of transition resistance based on the three stages of insulator crack breakdown, grain boundary breakdown, and grain breakdown includes: A time-varying model of pore resistance is constructed based on the insulator crack breakdown stage; A time-varying model of grain boundary resistance is constructed based on the grain boundary breakdown stage of insulators; A time-varying model of grain resistance is constructed based on the grain breakdown stage of insulators; A dynamic time-varying model of transition resistance is constructed based on the time-varying models of pore resistance, grain boundary resistance, and grain resistance.

3. The method for predicting insulator structure damage under severe weather conditions as described in claim 1, characterized in that: The acquisition of historical insulator damage data based on the dynamic time-varying model of transition resistance and historical insulator operating data includes: Based on the matching relationship between the transition resistance at each time sequence in the historical operation data of the insulator and the transition resistance under different breakdown conditions in the dynamic time-varying model of the transition resistance, the insulator operation data of the matching time sequence and the corresponding insulator breakdown state are obtained. The insulator operation data matched with the timing sequence and the corresponding insulator breakdown state are used as the historical damage data of the insulator.

4. The method for predicting insulator structure damage under severe weather conditions as described in claim 1, characterized in that: The method of obtaining the environmentally damaged set and the non-environmentally damaged set based on the time-series correspondence between historical insulator damage data and historical severe environment data includes: The historical damage data of insulators, which is matched with the historical severe environmental data of insulators in a time series, and the historical severe environmental data of insulators are used as the environmental damage set; Historical damage data of insulators that do not match the time series of historical severe environmental data are considered as non-environmental damage data.

5. The method for predicting insulator structure damage under severe weather conditions as described in claim 1, characterized in that: The iterative update of the initial time series relationship of the transition resistance dynamic time-varying model using the non-environmentally damaged set and the transition resistance dynamic time-varying model includes: The temporal characteristics of transition resistance are constructed by analyzing the changes in transition resistance at adjacent time points in the set of non-environmentally damaged environments. Based on the optimization algorithm, the initial timing relationship of the dynamic time-varying model of the transition resistance is iteratively optimized according to the timing characteristics of the transition resistance, and the timing parameters of the dynamic time-varying model of the transition resistance are updated.

6. The method for predicting insulator structure damage under severe weather conditions as described in claim 1, characterized in that: The method of constructing the association between environmental features and damaged features based on the set of environmental damage using machine learning algorithms includes: The temporal relationship of damage characteristics is obtained based on the historical damage data of insulators in the environmental damage set; Temporal relationships of environmental characteristics are obtained from severe environmental data of insulators in the damaged environmental set; Machine learning algorithms are used to obtain the correlation between environmental features and damaged features based on the temporal relationship between damaged features and environmental features.

7. The method for predicting insulator structure damage under severe weather conditions as described in claim 6, characterized in that: The process of obtaining damage feature correction coefficients based on real-time environmental data and the correlation between environmental features and damage features, and updating the dynamic time-varying model of transition resistance based on damage feature correction coefficients includes: Based on real-time environmental data and the correlation between environmental characteristics and damage characteristics, environmental data for the prediction period and the trend of damage characteristics during the prediction period are obtained. Based on the trend of damage characteristics changes during the prediction period, the damage characteristic correction coefficient is obtained, and the dynamic time-varying model of transition resistance is updated.

8. The method for predicting insulator structure damage under severe weather conditions as described in claim 1, characterized in that: The method of obtaining the insulator structure damage prediction results using the updated dynamic time-varying model of transition resistance and real-time insulator electrical parameters includes: The transition resistance of the current time sequence is obtained based on the real-time electrical parameters of the insulator; The transition resistance for the predicted period is obtained based on the current transition resistance and the updated dynamic time-varying model of the transition resistance. The relationship between transition resistance and breakdown stage is used to obtain the prediction results of insulator structure damage.

9. A system for predicting damage to insulator structures under severe weather conditions, used to implement the method described in any one of claims 1 to 8, characterized in that: include: The model building unit is used to construct a dynamic time-varying model of transition resistance based on the three stages of insulator crack breakdown, grain boundary breakdown and grain breakdown. The data acquisition unit is used to acquire historical severe environmental data, historical damage data, real-time environmental data, and real-time electrical parameters of the insulators. The model update unit is used to construct the correlation between environmental features and damage features based on historical severe environmental data and historical damage data of insulators, and to update the dynamic time-varying model of transition resistance according to the correlation between environmental features and damage features. The damage prediction unit is used to obtain the damage prediction results of the insulator structure based on the updated dynamic time-varying model of the transition resistance and the real-time electrical parameters of the insulator.

Images