A porcelain insulator defect detection method, device, equipment and medium

Abstract

The present application mainly relates to the field of artificial intelligence technology. The present application discloses a porcelain insulator defect detection method, device, equipment and medium, the method comprising: based on the timing temperature information of the insulator string to be measured, generating the temperature feature set corresponding to the insulator string to be measured, the temperature feature set including the first feature set and the second feature set; constructing a position prediction model, and training the position prediction model using the first feature set; inputting the second feature set into the trained position prediction model, and determining the defect position of the insulator string to be measured through the trained position prediction model to obtain an output result; generating a linear model according to the second feature set and the output result, and showing the relationship between the input features and the output result of the position prediction model through the linear model. The present application can detect porcelain insulators by using thermal imaging technology, and improve the accuracy and reliability of the detection result.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, specifically to a method, apparatus, equipment, and medium for detecting defects in porcelain insulators. Background Technology

[0002] Porcelain insulators, as an important component of electrical insulation and mechanical support in power equipment, are widely used in transmission lines and substations. Internal and external damage to porcelain insulators can lead to a decline in their insulation performance, subsequently causing short circuits in power equipment. Therefore, inspecting the internal and external surfaces of porcelain insulators is an important method for determining their insulation performance. With the continuous development of thermal imaging technology, it has become possible to inspect the internal structure of porcelain insulators using thermal imaging. However, the accuracy of thermal imaging inspections of porcelain insulators is low, and the lack of visualization methods further reduces the reliability of the results. Summary of the Invention

[0003] This invention provides a method, apparatus, equipment, and medium for detecting defects in porcelain insulators, which can utilize thermal imaging technology to detect porcelain insulators, thereby improving the accuracy and reliability of the detection results.

[0004] This invention provides a method for detecting defects in porcelain insulators, the method comprising:

[0005] Based on the time-series temperature information of the insulator string to be tested, a temperature feature set corresponding to the insulator string to be tested is generated, and the temperature feature set includes a first feature set and a second feature set;

[0006] Construct a location prediction model and train the location prediction model using the first feature set;

[0007] The second feature set is input into the trained position prediction model, and the defect location of the insulator string to be tested is determined by the trained position prediction model to obtain the output result.

[0008] A linear model is generated based on the second feature set and the output result, and the relationship between the input features and the output result of the location prediction model is demonstrated through the linear model.

[0009] Furthermore, the time-series temperature information includes the current temperature value of each insulator in the insulator string under test and the ambient temperature value of the environment in which each insulator is located;

[0010] Based on the time-series temperature information of the insulator string under test, a temperature feature set corresponding to the insulator string under test is generated. The temperature feature set includes a first feature set and a second feature set, including:

[0011] Obtain the timing temperature information of the insulator string under test;

[0012] Based on the time-series temperature information, the temperature change value of each insulator is determined, wherein the temperature change value of the insulator under test is the difference between the current temperature value of the insulator under test and the ambient temperature value;

[0013] Based on the temperature change value of each insulator, a temperature characteristic information factor of the insulator string under test is generated and used as the temperature feature of the insulator string under test;

[0014] Return to the step of obtaining the time-series temperature information of the insulator string to be tested. When multiple temperature features are obtained, generate a temperature feature set corresponding to the insulator string to be tested and divide it into a first feature set and a second feature set.

[0015] Furthermore, the method for detecting defects in porcelain insulators also includes:

[0016] The temperature feature set is divided into multiple data clusters, and each data cluster includes multiple temperature features;

[0017] Determine a first average distance and a second average distance for each temperature feature, wherein the first average distance is the average distance between the temperature feature and temperature features of the same data cluster, and the second average distance is the average distance between the temperature feature and temperature features of different data clusters;

[0018] Based on the functional relationship between the first average distance and the second average distance, the profile coefficient of each temperature feature is determined;

[0019] Based on the comparison between the contour coefficient of each temperature feature and a preset threshold, multiple target temperature features are determined, and a filtered temperature feature set is generated based on the multiple target temperature features.

[0020] Furthermore, the method for detecting defects in porcelain insulators also includes:

[0021] The temperature features in the temperature feature set are clustered into multiple classes;

[0022] Determine the dispersion values ​​of the multiple classes;

[0023] The target class is determined based on the comparison between the dispersion value and the preset dispersion value;

[0024] Each target class is used as a target temperature feature, and a clustered temperature feature set is generated based on the multiple target temperature features.

[0025] Furthermore, the construction of the location prediction model and the training of the location prediction model using the first feature set include:

[0026] Based on the prediction accuracy and model complexity of the location prediction model, the objective function of the location prediction model is generated.

[0027] The model parameters of the location prediction model are adjusted based on the current prediction accuracy of the location prediction model.

[0028] The current model complexity and current prediction accuracy of the location prediction model are updated based on the adjusted model parameters.

[0029] The objective function is updated based on the updated model complexity and the updated prediction accuracy;

[0030] When the location prediction model is trained using the first feature set, the updated prediction accuracy is used as the current prediction accuracy, and the process returns to the step of adjusting the model parameters of the location prediction model according to the current prediction accuracy of the location prediction model until the optimal objective function is obtained. The optimal objective function is an objective function that meets preset conditions.

[0031] Furthermore, the location prediction model includes at least one tree structure, and each tree structure includes a first node and a second node, where the first node and the second node are both model parameters;

[0032] The first node is used to generate a new first node or a new second node, and the second node is used to output the prediction result;

[0033] The step of adjusting the model parameters of the location prediction model based on the current prediction accuracy of the location prediction model includes:

[0034] Based on the current prediction accuracy, the number of the first and second nodes of the current tree structure, and the node value of each node, the structural score of the current tree structure is determined.

[0035] Based on the structural score, a new first node or a new second node is generated on the tree structure to adjust the model parameters of the location prediction model.

[0036] Furthermore, the method includes:

[0037] Samples are selected from the second feature set, and the selected samples are grouped into a sample subset;

[0038] When the selected sample is added to the sample subset, the marginal gain value of each sample in the second feature set is determined according to a preset function;

[0039] Based on the marginal gain value, a target sample is determined in the second feature set and used as the selected sample;

[0040] When the preset stopping condition is met, the selection of samples in the second feature set is stopped, and the subset of samples is used as the optimized second feature set.

[0041] The present invention also provides a defect detection device for porcelain insulators, the device comprising:

[0042] The acquisition module is used to generate a temperature feature set corresponding to the insulator string under test based on the time-series temperature information of the insulator string under test. The temperature feature set includes a first feature set and a second feature set.

[0043] A construction module is used to construct a location prediction model and train the location prediction model using the first feature set;

[0044] The training module is used to input the second feature set into the trained position prediction model, and to obtain the output result after determining the defect location of the insulator string to be tested through the trained position prediction model.

[0045] The display module is used to generate a linear model based on the second feature set and the output result, and to display the relationship between the input features and the output result of the location prediction model through the linear model.

[0046] The present invention also provides an electronic device, the electronic device including a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the porcelain insulator defect detection method as described in any of the preceding claims.

[0047] The present invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the porcelain insulator defect detection method as described in any of the preceding claims.

[0048] The present invention has at least the following beneficial effects:

[0049] This application's technical solution monitors the time-series temperature information of insulator strings in real time and quickly generates a temperature feature set. Then, using a trained location prediction model, it predicts the location of defects in the insulator strings based on the patterns learned from the temperature feature set, thereby achieving real-time detection of overheating defects in porcelain insulators. Simultaneously, based on a second feature set and the model's output, a linear model is generated. This linear model demonstrates the relationship between the input features and output of the location prediction model, helping to understand how the model makes predictions based on input features and identifying which features have the greatest impact on the prediction results. This allows for visualization of the location prediction model, intuitively showcasing its prediction logic and feature importance, thus enhancing the model's interpretability and reliability. Attached Figure Description

[0050] The accompanying drawings are provided to further understand the technical solutions of the present invention and constitute a part of the specification. They are used together with the embodiments of the present invention to explain the technical solutions of the present invention, and do not constitute a limitation on the technical solutions of the present invention.

[0051] Figure 1 This is a flowchart of the steps involved in the defect detection method for porcelain insulators;

[0052] Figure 2 This is a flowchart of step S101 in the method for detecting defects in porcelain insulators;

[0053] Figure 3 This is a flowchart of step S102 in the method for detecting defects in porcelain insulators;

[0054] Figure 4 This is a schematic diagram illustrating the working principle of the data processing process in implementing a defect detection method for porcelain insulators.

[0055] Figure 5 This is a schematic diagram illustrating the working principle of the model training process when implementing a defect detection method for porcelain insulators.

[0056] Figure 6 This is a schematic diagram illustrating the working principle of the secondary mode selection process in implementing the defect detection method for porcelain insulators;

[0057] Figure 7 This is a schematic diagram illustrating a visual result obtained through a defect detection method for porcelain insulators in a real-world scenario.

[0058] Figure 8 This is a schematic diagram illustrating another visual result obtained through the defect detection method of porcelain insulators in a real-world scenario;

[0059] Figure 9 This is a schematic diagram of a porcelain insulator defect detection device;

[0060] Figure 10This is a schematic diagram of the structure of an electronic device. Detailed Implementation

[0061] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0062] In related technical fields, porcelain insulators are widely used in transmission lines and substations as an important component of electrical insulation and mechanical support for power equipment. With prolonged operation, their insulation and mechanical properties tend to decline. Insulator failure can cause short circuits, leading to electrical safety accidents. According to relevant research, the main causes of decreased insulation performance in porcelain insulators are external damage and internal corrosion. External damage primarily manifests as edge breakage of the skirts and cracking of the steel feet, which can be diagnosed through aerial photography or visual inspection; the relevant technologies are very mature. On the other hand, internal corrosion manifests as rusted steel caps and internal breakdown. Since its appearance cannot be identified, other characteristics are required for detection. Although research on porcelain insulator testing technology is constantly developing, it currently remains at the laboratory stage.

[0063] With the continuous advancement of thermal imaging technology, the inspection of porcelain insulators has become possible. Our research has revealed that when porcelain insulators are in the early stage of overheating defects (low-value insulators), the temperature distribution of the faulty insulator and its adjacent areas exhibits a convex or peak-shaped pattern. Furthermore, thermal imaging technology, with its high sensitivity to temperature, provides a new solution for insulator inspection.

[0064] However, detection methods based on thermal imaging technology have shortcomings. First, the detection model lacks visualization methods, resulting in weak reliability of the detection results. Second, the detection model network is complex, and for tens of thousands or even hundreds of thousands of insulators to be inspected, the computing equipment requirements are extremely high, and it is impossible to perform fast calculations.

[0065] To address the aforementioned technical problems, this application provides a method, apparatus, equipment, and medium for detecting defects in porcelain insulators. This method utilizes thermal imaging technology to detect porcelain insulators, improving the accuracy and reliability of the detection results. The following are embodiments of the technical solution provided in this application.

[0066] Please refer to Figure 1 , Figure 1 This is a flowchart of the steps involved in the defect detection method for porcelain insulators.

[0067] This embodiment provides a method for detecting defects in porcelain insulators, the method including:

[0068] S101. Based on the time-series temperature information of the insulator string to be tested, generate a temperature feature set corresponding to the insulator string to be tested. The temperature feature set includes a first feature set and a second feature set.

[0069] S102. Construct a location prediction model and train the location prediction model using the first feature set.

[0070] S103. Input the second feature set into the trained position prediction model, and obtain the output result after determining the defect location of the insulator string to be tested through the trained position prediction model.

[0071] S104. Generate a linear model based on the second feature set and the output results, and use the linear model to demonstrate the relationship between the input features and the output results of the location prediction model.

[0072] In some embodiments, the time-series temperature information includes the current temperature value of each insulator in the insulator string under test and the ambient temperature value of the environment in which each insulator is located.

[0073] In step S101 of some embodiments, time series analysis methods, such as Fourier transform or wavelet transform, are used to extract frequency domain features from time-series temperature information to form a temperature feature set, or statistical analysis methods, such as calculating the mean, standard deviation, maximum and minimum values ​​of temperature data, are applied to form a temperature feature set.

[0074] In step S102 of some embodiments, a location prediction model is constructed using traditional machine learning algorithms, such as support vector machine (SVM) or random forest.

[0075] In step S104 of some embodiments, linear regression analysis is used to establish a linear model based on the input features and output results, and to analyze the linear relationship between the input features and the output results. Specifically, a locally interpretable model (such as LIME or SHAP) is adopted. Combined with the locally interpretable model, the specific contribution of each feature to the model prediction is provided to enhance the interpretability of the model, thereby explaining the linear relationship between the input features and the output results. Then, visualization tools, such as scatter plots, heatmaps, or parallel coordinate plots, are used to intuitively display the relationship between the input features and the output results.

[0076] Please refer to Figure 2 , Figure 2 This is a flowchart of step S101 in the method for detecting defects in porcelain insulators.

[0077] In some embodiments, step S101 includes:

[0078] S201. Obtain the timing temperature information of the insulator string to be tested.

[0079] S202. Based on the time-series temperature information, determine the temperature change value of each insulator, wherein the temperature change value of the insulator under test is the difference between the current temperature value of the insulator under test and the ambient temperature value.

[0080] S203. Based on the temperature change value of each insulator, generate temperature characteristic information factors of the insulator string to be tested and use them as temperature characteristics of the insulator string to be tested.

[0081] S204. Return to step S201. When multiple temperature features are obtained, generate the temperature feature set corresponding to the insulator string to be tested and divide it into the first feature set and the second feature set.

[0082] Understandably, research has shown that under slightly heated conditions, the temperature of defective areas in an insulator string exhibits a convex parabolic relationship with adjacent areas. Therefore, to focus the model on adjacent temperature variation components, the temperature data is highlighted. Specifically, a temperature gradient coefficient (i.e., a temperature characteristic information factor) is constructed using adjacent temperature data of the insulator, and then a normalization operation is performed.

[0083] One specific implementation of this embodiment is to calculate the temperature change of the insulator using the following formula:

[0084] (1)

[0085] in, The temperature change value of the i-th insulator in the insulator string to be tested. The field temperature value of the i-th insulator in the insulator string to be inspected. It is the reference temperature value (i.e., ambient temperature value) of the i-th insulator in the insulator string to be tested.

[0086] Based on formula (1), the temperature characteristic information factor of each insulator in the insulator string to be tested is obtained according to formula (2).

[0087] (2)

[0088] in, For the k-th relative temperature diagnostic information, The maximum value of the dataset of insulator strings to be inspected. This represents the minimum value of the dataset of insulator strings to be inspected.

[0089] Please refer to Figure 3 , Figure 3 This is a flowchart of step S102 in the method for detecting defects in porcelain insulators.

[0090] In some embodiments, step S102 includes:

[0091] S301. Based on the prediction accuracy and model complexity of the location prediction model, generate the objective function of the location prediction model.

[0092] S302. Adjust the model parameters of the location prediction model according to the current prediction accuracy of the location prediction model.

[0093] S303. Update the current model complexity and current prediction accuracy of the location prediction model based on the adjusted model parameters.

[0094] S304. Update the objective function based on the updated model complexity and the updated prediction accuracy.

[0095] S305. When training the location prediction model using the first feature set, the updated prediction accuracy is used as the current prediction accuracy, and the process returns to step S302 until the optimal objective function is obtained. The optimal objective function is an objective function that meets the preset conditions.

[0096] In this embodiment, the location prediction model can be a fusion of Markov and multi-class machine learning models. This model combines the Markov model with various machine learning algorithms, such as decision tree (DT), random forest (RF), k-nearest neighbors (kNN), and support vector machine (SVM). The Markov model predicts the probability of the next state based on the current state, while the machine learning model provides the ability to classify location data.

[0097] In some embodiments, the location prediction model employs an optimized decision tree model. The location prediction model includes at least one tree structure, each tree structure including a first node and a second node, both of which are model parameters; the first node is used to generate a new first node or a new second node, and the second node is used to output the prediction result.

[0098] In some embodiments, step S302 includes:

[0099] Based on the current prediction accuracy, the number of first and second nodes in the current tree structure, and the node value of each node, the structural score of the current tree structure is determined; based on the structural score, a new first node or a new second node is generated on the tree structure to adjust the model parameters of the location prediction model.

[0100] Please see Figure 4 In one specific embodiment, the objective function of the location prediction model can be expressed by formula (3):

[0101] (3)

[0102] Where l is the loss function between the true value and the predicted value of the i-th sample, and It is the complexity of the k-th tree. During the recursion, the k-th tree will be trained based on the (k-1)-th tree, thus we get (4).

[0103] (4)

[0104] in, It is a sample from the j-th tree.

[0105] When training the k-th tree, the previous k trees can be used as a reference. The sum of the model complexities of 1 tree is considered constant, resulting in formula (5):

[0106] (5)

[0107] In some embodiments, the objective function of the location prediction model is further optimized.

[0108] In one specific embodiment, the objective function is optimized by taking the second derivative using Taylor's technique.

[0109] Specifically, the second-order expansion of the Taylor series is shown in equation (6):

[0110] (6)

[0111] Based on equation (5), the first and second derivatives are performed as follows:

[0112] (7)

[0113] (8)

[0114] Substituting equations (6), (7), and (8) into equation (5), we obtain the Taylor series transformation equation (9):

[0115] (9)

[0116] Next, simplify equation (9) into equation (10):

[0117] (10)

[0118] The decision tree super-clustering algorithm uses a greedy algorithm to handle the selection of the optimal tree structure. It constructs a structure score using formula (11), dynamically selects whether to build tree nodes based on the structure score, and finally outputs the optimal tree structure of the model. Formula (11) is as follows:

[0119] (11)

[0120] In the formula , gi The sum of the left and right subtrees, , h respectively i The sum of the left and right subtrees, where γ and λ are parameters.

[0121] The model complexity is controlled by the number of leaf nodes, node values, and tree depth, as shown in Equation (12):

[0122] (12)

[0123] Where T is the number of leaf nodes and ω is the value of a leaf node.

[0124] Based on equation (10), and combined with equation (12), we obtain (13):

[0125] (13)

[0126] (14)

[0127] (15)

[0128] Substituting (14) and (15) into equation (13), equation (13) can be simplified to equation (16):

[0129] (16)

[0130] The optimal leaf node values ​​of the tree can be obtained from equation (16):

[0131] (17)

[0132] Based on the above formulas, the minimum value of the loss function can be derived as follows:

[0133] (18)

[0134] In some embodiments, a method for detecting defects in porcelain insulators further includes:

[0135] Multiple temperature features in the temperature feature set are clustered into multiple classes; the dispersion value of multiple classes is determined; the target class is determined based on the comparison between the dispersion value and the preset dispersion value; each target class is used as a target temperature feature, and a clustered temperature feature set is generated based on multiple target temperature features.

[0136] In some embodiments, a method for detecting defects in porcelain insulators further includes:

[0137] The temperature feature set is divided into multiple data clusters, each of which includes multiple temperature features. A first average distance and a second average distance are determined for each temperature feature, where the first average distance is the average distance between the temperature feature and temperature features in the same data cluster, and the second average distance is the average distance between the temperature feature and temperature features in different data clusters. Based on the functional relationship between the first average distance and the second average distance, the contour coefficient of each temperature feature is determined. Based on the comparison between the contour coefficient of each temperature feature and a preset threshold, multiple target temperature features are determined, and a filtered temperature feature set is generated based on the multiple target temperature features.

[0138] It is understandable that for tens of thousands or even hundreds of thousands of temperature characteristic information factors, achieving both training speed and accuracy is an important issue. Therefore, this embodiment first simplifies redundant features through a super-clustering algorithm based on the traditional decision tree model, and then establishes a method for predicting the location of deteriorated insulators based on decision tree super-clustering according to its temporal characteristics.

[0139] Specifically, the dispersion of each cluster can be measured by two metrics: s(i) and CH. For a dataset D of n temperature characteristic information factors (i.e., the temperature feature set), D is divided into k clusters. For each object o in D, the average distance a(o) between it and other objects in its cluster is calculated, and the minimum average distance b(o) between it and the clusters it does not belong to is calculated. The silhouette coefficient is calculated for each object in the dataset, and the average value is taken as a measure of cluster quality. The silhouette coefficient ranges from -1 to 1, with values ​​closer to 1 indicating better cluster quality and values ​​closer to -1 indicating worse cluster quality.

[0140] (19)

[0141] Calculate the average distance a from sample i to other samples in the same cluster. i a i The smaller the value, the more likely sample i should be clustered into that cluster. Let a... i The intra-cluster dissimilarity of sample i is calculated. The dissimilarity of sample i to other clusters C is calculated. j The average distance b of all samples i , referred to as sample i and cluster c j The dissimilarity is defined as the inter-cluster dissimilarity of sample i: b i = min{b i1 ,b i2 ,......b ik}, s i If the value is close to 1, it indicates that the clustering of sample i is reasonable, and s i If the value is close to -1, it indicates that sample i should be classified into a different cluster. If s iIf the value is approximately 0, it means that sample i is on the boundary between the two clusters.

[0142] (20)

[0143] Where n represents the number of clusters, k represents the current cluster, trB(k) represents the trace of the inter-cluster deviation matrix, and trW(k) represents the trace of the intra-cluster deviation matrix. A larger CH value indicates that the clusters are more compact and the clusters are more dispersed, resulting in a better clustering result.

[0144] In some embodiments, when a certain temperature feature has si>0.8 and CH>20, it can be considered a redundant feature, and similar feature parts can be removed.

[0145] In some embodiments, a method for detecting defects in porcelain insulators further includes:

[0146] Samples are selected from the second feature set and aggregated into a sample subset. When a selected sample is added to the sample subset, the marginal gain value of each sample in the second feature set is determined according to a preset function. Based on the marginal gain value, a target sample is determined in the second feature set and used as the selected sample. When a preset stopping condition is met, the selection of samples in the second feature set is stopped, and the sample subset is used as the optimized second feature set.

[0147] Understandably, existing porcelain insulator testing typically employs a "black box model" (i.e., trained via a network and directly outputting results). To prevent the model from exhibiting local fitting due to small samples, and considering that porcelain insulators are components in the power industry, the weights of the network's discrimination need to be visualized to ensure the reliability of the testing results.

[0148] Please see Figure 5 In one specific implementation of this embodiment, for any set The set of all samples (V is the second feature set) and any element ,have:

[0149] (twenty one)

[0150] Where F is the importance of the selected sample subset in explaining the model behavior, this part obtains the representative samples to obtain the sample subset through formula (21), such as Figure 5 As shown.

[0151] Based on the optimized second feature set described above, this embodiment generates multiple neighboring samples by randomly perturbing the original sample x, forming a set of neighboring sample sets X', and obtains the predicted values ​​y' of the neighboring samples in the DCT model (position prediction model for deteriorated insulators). Simultaneously, minimizing (21) yields the fitted linear model g:

[0152] (twenty two)

[0153] in, This is a model for predicting the location of deteriorated insulators; g is a locally linear model, and the coefficients of this model are positively correlated with the characteristics of the deteriorated insulator location prediction model mentioned in this invention. It is a measure The loss function for predicting the difference between g and g; For regularization terms; It is the weight function.

[0154] Understandably, this embodiment uses submodal selection to obtain a representative insulator dataset (the optimized second feature set), and then obtains a local linear model through equation (22). By analyzing the coefficients of the local linear model, the contribution of the input features of the location prediction model of deteriorated insulators to the prediction results is reflected. In addition to avoiding the selection of similar interpretation instances and iteratively increasing the samples with the highest marginal coverage gain, this method can also effectively reduce the construction time of the interpretation model.

[0155] Please see Figure 6 In one specific embodiment, the defect location of the porcelain insulator and the visualization effect of the model are tested according to the porcelain insulator defect detection method.

[0156] In this embodiment, to verify the superiority of the algorithm, temperature data from 1400 strings of degraded 110 kV insulators and 100 strings of normal insulators at different locations were used in the experiment. The ratio of the training set, test set, and validation set in this experiment was 8:1:1.

[0157] The hardware is divided into servers and edge devices. The servers are equipped with a 2.4 GHz dual-memory processor and an NVIDIA GeForce RTX 3090 graphics processor, capable of handling data storage and processing tasks effectively. The edge devices use the DJI Meridian M300 RTK as the platform and Cloudcrest iCrest 2.0 to build the edge processors.

[0158] To ensure that the system reaches its optimal state during operation, it is necessary to adjust the hyperparameters that play a major role in the direct torque control algorithm. We combine grid search and 5-fold cross-validation.

[0159] To test the performance of the SLIME algorithm, we conducted an ablation experiment to visualize the algorithm. We selected the first string of faulty insulators for the experiment and compared the performance of the random selection method and the random sampling method.

[0160] First, by calculating the gradient and difference values ​​of time-series temperature data, the differences in time-series temperature are highlighted, which accelerates the subsequent training iteration time. Then, a decision tree super-clustering DTC model is constructed, and redundant features are filtered through super-clustering. The prediction model is built with the decision tree of the greedy algorithm as the core. Finally, based on the decision tree super-clustering DTC model and the validation set data, a locally interpretable model is selected through secondary modeling for visualization analysis.

[0161] Finally, the visualization of classification features is as follows: Figure 7 and Figure 8 As shown, the relative temperature NO1 value has the greatest gain effect on pattern classification, followed by the relative temperatures NO9, NO12, and NO2 values, but these three values ​​have little impact on pattern classification. The value of NO13 has a negative effect on the classification of model effects. The figure shows that NO1 still accounts for the largest gain in model classification feature recognition, while the values ​​of NO12 and NO13 no longer contribute to classification recognition gain, although their influence increases.

[0162] By comparing the data from three iterations, it can be found that the porcelain insulator defect detection method using the SLIME algorithm in this invention is superior to the porcelain insulator defect detection method using the existing LIME algorithm in terms of model convergence speed and detection effect.

[0163] Any of the above embodiments can monitor the time-series temperature information of the insulator string in real time, quickly generate a temperature feature set, and predict the defect location of the insulator string based on the rules learned from the temperature feature set through a trained location prediction model, thereby realizing real-time detection of overheating defects in porcelain insulators. At the same time, a linear model is generated based on the second feature set and the output results of the model. This linear model can show the relationship between the input features and the output results of the location prediction model. It not only helps to understand how the location prediction model makes predictions based on the input features, but also identifies which features have the greatest impact on the prediction results, thereby enabling visualization of the location prediction model, intuitively showing the prediction logic and feature importance of the model, which is conducive to enhancing the interpretability and reliability of the model.

[0164] Please refer to Figure 9 , Figure 9 This is a schematic diagram of a porcelain insulator defect detection device.

[0165] This embodiment also provides a porcelain insulator defect detection device, which includes:

[0166] The acquisition module 401 is used to generate a temperature feature set corresponding to the insulator string under test based on the time-series temperature information of the insulator string under test. The temperature feature set includes a first feature set and a second feature set.

[0167] Module 402 is used to build a location prediction model and train the location prediction model using the first feature set.

[0168] The training module 403 is used to input the second feature set into the trained position prediction model, and to obtain the output result after determining the defect location of the insulator string to be tested through the trained position prediction model.

[0169] The display module 404 is used to generate a linear model based on the second feature set and the output results, and to display the relationship between the input features and the output results of the location prediction model through the linear model.

[0170] It will be understood by those skilled in the art that all or some of the steps and apparatuses in the methods disclosed above can be implemented as software, firmware, hardware, and suitable combinations thereof. Some or all of the physical components can be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit. Such software can be distributed on a computer-readable medium, which can include computer storage media (or non-transitory media) and communication media (or transient media). As is known to those skilled in the art, the term computer storage media includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disc (DVD) or other optical disc storage, magnetic cartridges, magnetic tape, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and is accessible to a computer. As is known to those skilled in the art, communication media typically contain computer-readable instructions, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium.

[0171] It is understood that the content of the above method embodiments is applicable to the present device embodiments. The specific functions implemented by the present device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0172] This application also provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement any of the above-mentioned methods for detecting defects in porcelain insulators.

[0173] refer to Figure 10 , Figure 10The hardware structure of an electronic device according to another embodiment is illustrated. The electronic device includes:

[0174] The processor 501 can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this application.

[0175] The memory 502 can be implemented as a read-only memory (ROM), static storage device, dynamic storage device, or random access memory (RAM). The memory 502 can store operating devices and other application programs. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program code is stored in the memory 502 and is called and executed by the processor 501 to execute the porcelain insulator defect detection method of the embodiments of this application.

[0176] The input / output interface 503 is used to implement information input and output;

[0177] The communication interface 504 is used to enable communication and interaction between this device and other devices. Communication can be achieved through wired means (such as USB, network cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.).

[0178] Bus 505 transmits information between various components of the device (e.g., processor 501, memory 502, input / output interface 503, and communication interface 504);

[0179] The processor 501, memory 502, input / output interface 503, and communication interface 504 are connected to each other within the device via bus 505.

[0180] It is understood that the content of the above method embodiments is applicable to the embodiments of this electronic device. The specific functions implemented by the embodiments of this electronic device are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0181] This application also provides a computer-readable storage medium storing a processor-executable program, which, when executed by a processor, is used to implement the porcelain insulator defect detection method as described in any of the above specific embodiments.

[0182] This application also discloses a computer program product, including a computer program or computer instructions, which are stored in a computer-readable storage medium. The processor of the computer device reads the computer program or computer instructions from the computer-readable storage medium and executes the computer program or computer instructions, causing the computer device to perform the porcelain insulator defect detection method as described in any of the preceding embodiments.

[0183] It is understood that the content of the above method embodiments is applicable to this storage medium embodiment. The specific functions implemented in this storage medium embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0184] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented, for example, in orders other than those illustrated or described herein. Furthermore, the terms “comprising” and “having,” and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, apparatus, product, or device that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices. It should be understood that in this application, “at least one” means one or more, and “more than one” means two or more.

[0185] In the several embodiments provided in this application, it should be understood that the disclosed apparatus, devices, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or units, and may be electrical, mechanical, or other forms.

[0186] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0187] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0188] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0189] Although the description of this application has been quite detailed and particularly focused on several of the described embodiments, it is not intended to limit itself to any of these details or embodiments or any particular embodiment. Rather, it should be considered as effectively covering the intended scope of this application by referring to the appended claims and taking into account the prior art, which provides for a broad possible interpretation of these claims. Furthermore, the foregoing description of this application with respect to embodiments foreseeable by the inventors is intended to provide a useful description, and non-substantial modifications to this application that have not yet been foreseen may still represent equivalent modifications.

Claims

1. A method for detecting defects in porcelain insulators, characterized in that, The method includes: Based on the time-series temperature information of the insulator string to be tested, a temperature feature set corresponding to the insulator string to be tested is generated, and the temperature feature set includes a first feature set and a second feature set; Construct a location prediction model and train the location prediction model using the first feature set; The second feature set is input into the trained position prediction model, and the defect location of the insulator string to be tested is determined by the trained position prediction model to obtain the output result. A linear model is generated based on the second feature set and the output result, and the relationship between the input features and the output result of the location prediction model is demonstrated through the linear model. The time-series temperature information includes the current temperature value of each insulator in the insulator string to be tested and the ambient temperature value of the environment where each insulator is located. Based on the time-series temperature information of the insulator string under test, a temperature feature set corresponding to the insulator string under test is generated. The temperature feature set includes a first feature set and a second feature set, including: Obtain the timing temperature information of the insulator string under test; Based on the time-series temperature information, the temperature change value of each insulator is determined, wherein the temperature change value of the insulator under test is the difference between the current temperature value of the insulator under test and the ambient temperature value; Based on the temperature change value of each insulator, a temperature characteristic information factor of the insulator string under test is generated and used as the temperature feature of the insulator string under test; Return to the step of obtaining the time-series temperature information of the insulator string to be tested. When multiple temperature features are obtained, generate a temperature feature set corresponding to the insulator string to be tested and divide it into a first feature set and a second feature set. The construction of the location prediction model and the training of the location prediction model using the first feature set include: Based on the prediction accuracy and model complexity of the location prediction model, the objective function of the location prediction model is generated. The model parameters of the location prediction model are adjusted based on the current prediction accuracy of the location prediction model. The current model complexity and current prediction accuracy of the location prediction model are updated based on the adjusted model parameters. The objective function is updated based on the updated model complexity and the updated prediction accuracy; When the location prediction model is trained using the first feature set, the updated prediction accuracy is used as the current prediction accuracy, and the process returns to the step of adjusting the model parameters of the location prediction model according to the current prediction accuracy of the location prediction model until the optimal objective function is obtained. The optimal objective function is an objective function that meets preset conditions.

2. The method for detecting defects in porcelain insulators according to claim 1, characterized in that, The method further includes: The temperature feature set is divided into multiple data clusters, and each data cluster includes multiple temperature features; Determine a first average distance and a second average distance for each temperature feature, wherein the first average distance is the average distance between the temperature feature and temperature features of the same data cluster, and the second average distance is the average distance between the temperature feature and temperature features of different data clusters; Based on the functional relationship between the first average distance and the second average distance, the profile coefficient of each temperature feature is determined; Based on the comparison between the contour coefficient of each temperature feature and a preset threshold, multiple target temperature features are determined, and a filtered temperature feature set is generated based on the multiple target temperature features.

3. The method for detecting defects in porcelain insulators according to claim 1, characterized in that, The method further includes: The temperature features in the temperature feature set are clustered into multiple classes; Determine the dispersion values ​​of the multiple classes; The target class is determined based on the comparison between the dispersion value and the preset dispersion value; Each target class is used as a target temperature feature, and a clustered temperature feature set is generated based on the multiple target temperature features.

4. The method for detecting defects in porcelain insulators according to claim 1, characterized in that, The location prediction model includes at least one tree structure, and each tree structure includes a first node and a second node, where the first node and the second node are both model parameters. The first node is used to generate a new first node or a new second node, and the second node is used to output the prediction result; The step of adjusting the model parameters of the location prediction model based on the current prediction accuracy of the location prediction model includes: Based on the current prediction accuracy, the number of the first and second nodes of the current tree structure, and the node value of each node, the structural score of the current tree structure is determined. Based on the structural score, a new first node or a new second node is generated on the tree structure to adjust the model parameters of the location prediction model.

5. The method for detecting defects in porcelain insulators according to claim 1, characterized in that, The method includes: Samples are selected from the second feature set, and the selected samples are grouped into a sample subset; When the selected sample is added to the sample subset, the marginal gain value of each sample in the second feature set is determined according to a preset function; Based on the marginal gain value, a target sample is determined in the second feature set and used as the selected sample; When the preset stopping condition is met, the selection of samples in the second feature set is stopped, and the subset of samples is used as the optimized second feature set.

6. A defect detection device for porcelain insulators, characterized in that, The device includes: The acquisition module is used to generate a temperature feature set corresponding to the insulator string under test based on the time-series temperature information of the insulator string under test. The temperature feature set includes a first feature set and a second feature set. A construction module is used to construct a location prediction model and train the location prediction model using the first feature set; The training module is used to input the second feature set into the trained position prediction model, and to obtain the output result after determining the defect location of the insulator string to be tested through the trained position prediction model. The display module is used to generate a linear model based on the second feature set and the output result, and to display the relationship between the input features and the output result of the location prediction model through the linear model; The time-series temperature information includes the current temperature value of each insulator in the insulator string to be tested and the ambient temperature value of the environment where each insulator is located. Based on the time-series temperature information of the insulator string under test, a temperature feature set corresponding to the insulator string under test is generated. The temperature feature set includes a first feature set and a second feature set, including: Obtain the timing temperature information of the insulator string under test; Based on the time-series temperature information, the temperature change value of each insulator is determined, wherein the temperature change value of the insulator under test is the difference between the current temperature value of the insulator under test and the ambient temperature value; Based on the temperature change value of each insulator, a temperature characteristic information factor of the insulator string under test is generated and used as the temperature feature of the insulator string under test; Return to the step of obtaining the time-series temperature information of the insulator string to be tested. When multiple temperature features are obtained, generate a temperature feature set corresponding to the insulator string to be tested and divide it into a first feature set and a second feature set. The construction of the location prediction model and the training of the location prediction model using the first feature set include: Based on the prediction accuracy and model complexity of the location prediction model, the objective function of the location prediction model is generated. The model parameters of the location prediction model are adjusted based on the current prediction accuracy of the location prediction model. The current model complexity and current prediction accuracy of the location prediction model are updated based on the adjusted model parameters. The objective function is updated based on the updated model complexity and the updated prediction accuracy; When the location prediction model is trained using the first feature set, the updated prediction accuracy is used as the current prediction accuracy, and the process returns to the step of adjusting the model parameters of the location prediction model according to the current prediction accuracy of the location prediction model until the optimal objective function is obtained. The optimal objective function is an objective function that meets preset conditions.

7. An electronic device, characterized in that, The electronic device includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the porcelain insulator defect detection method according to any one of claims 1 to 5.

8. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the porcelain insulator defect detection method according to any one of claims 1 to 5.

Images