Cable insulation defect detection method and system based on infrared imaging

Abstract

The present application belongs to the technical field of cable detection, and discloses a cable insulation defect detection method and system based on infrared imaging; comprising the following modules: a control test module, used for applying a constant current step-up voltage and a constant voltage step-up current electric excitation sequence test to the cable. The present application actively simulates the state of the cable under different operating conditions by setting the constant current step-up voltage and constant voltage step-up current double sequence electric excitation test method. For light load cables, the current sequence can effectively stimulate the heating of resistance type defects, and for various cables, the voltage sequence can effectively stimulate the heating of medium type defects. By analyzing the response of the defects in this dynamic excitation process and using the defect classification model to analyze the comprehensive feature vector, different nature defects such as partial discharge, dielectric loss and excessive contact resistance can be clearly distinguished, and the classification of the cable defect types is realized.

Description

Technical Field

[0001] This invention relates to the field of cable inspection technology, and more specifically, to a method and system for detecting cable insulation defects based on infrared imaging. Background Technology

[0002] Power cables are the main arteries of power transmission and distribution networks, and the quality of their insulation directly affects the safe and stable operation of the entire power grid. Over long-term operation, cable insulation ages due to factors such as electrical, thermal, mechanical, and environmental stresses, resulting in defects such as partial discharge, increased dielectric loss, insulation dampness, and oxidation at connection points. These defects can lead to localized overheating, and if not detected and addressed promptly, will cause insulation breakdown, resulting in significant economic losses and social impact.

[0003] Therefore, it is necessary to test the insulation performance of cables to detect insulation defects. The current detection method mainly uses thermal images to determine the temperature of the cable when it is energized, thereby identifying insulation defects. However, existing infrared imaging-based detection methods have obvious limitations: First, this method is heavily dependent on the cable's operating load. When the load current is low, the heat generated by the defect is small, and the temperature rise is not obvious, which can easily lead to missed detection and cannot meet the needs of early warning. Second, most existing methods rely only on the highest temperature or relative temperature difference in a single thermal image to make a judgment, lacking analysis of the dynamic characteristics of defects and making it difficult to effectively distinguish defect types. Summary of the Invention

[0004] To address the problems in the prior art, this invention proposes a method and system for detecting cable insulation defects based on infrared imaging.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a cable insulation defect detection system based on infrared imaging, comprising the following modules:

[0006] The control test module is used for electrical excitation sequence tests that apply constant current step-up and constant voltage step-up to the cable;

[0007] The data acquisition module is used to acquire thermal imaging images of the cable during the electrical excitation sequence test, and to obtain preliminary characteristic parameters of the abnormal temperature range of the cable based on the thermal imaging images.

[0008] The defect assessment module is used to generate defect scores based on preliminary characteristic parameters and classify defect levels in abnormal temperature ranges based on the defect scores.

[0009] The defect feature module is used to obtain the comprehensive feature vector of the abnormal temperature range based on the preliminary feature parameters of the abnormal temperature range.

[0010] The defect classification module acquires historical defect type data, constructs a defect classification model based on the historical defect type data, and classifies defects in abnormal temperature ranges based on the defect classification model.

[0011] Furthermore, the control test module includes a programmable excitation power supply unit and a sequence control unit;

[0012] The programmable excitation power supply unit can independently and precisely control the amplitude of the output voltage and current;

[0013] The sequence control unit is configured to execute the following two independent test sequences:

[0014] The first sequence is a constant current stepped boost: the excitation power supply is controlled to output a constant base current value, and then the output voltage is stepped up in a first preset step size until the rated voltage is reached.

[0015] The second sequence is a constant voltage stepped current ramp: the excitation power supply outputs the rated voltage, and then the output current is increased in a stepped manner with a second preset step size until the rated current is reached.

[0016] Furthermore, the data acquisition module includes an infrared thermal imager, an image processing unit, and a preliminary feature extraction unit;

[0017] The infrared thermal imager is used to collect infrared radiation data on the cable surface and generate a temperature field distribution image after each excitation step reaches a thermally stable state.

[0018] The image processing unit is used to filter and reduce noise, calibrate temperature, and stitch images on the temperature field distribution image to generate a panoramic thermal image of the cable.

[0019] The preliminary feature extraction unit is used to analyze the panoramic thermal image, automatically identify and locate one or more abnormal temperature range regions, and obtain preliminary feature parameters for each abnormal temperature range region. The preliminary feature parameters include: the highest temperature value, average temperature value, temperature difference relative to the normal range, temperature rise rate, and area of ​​the abnormal temperature range region.

[0020] Within the identified abnormal temperature range, the temperature values ​​of all pixels are traversed, and the maximum value is the highest temperature value, which directly reflects the extreme heating situation in the abnormal temperature range.

[0021] The arithmetic mean of the temperature values ​​of all pixels within the abnormal temperature range is used to obtain the average temperature value. The average temperature value reflects the overall heat generation level of the abnormal area and represents the total thermal power of the defect.

[0022] On the same cable, select a section that is far from the abnormal temperature area and has no abnormal behavior, calculate its average temperature to obtain the average temperature value of the normal section, and then subtract the average temperature value of the normal section from the average temperature value of the abnormal temperature area to obtain the temperature difference relative to the normal section. The temperature difference relative to the normal section directly represents the temperature rise caused by the defect itself.

[0023] Under the excitation ladder, from the initial application of excitation to the temperature stabilization, the temperature change curve over time is recorded. The temperature rise rate is the average slope or first derivative of the curve during the main rising phase. The temperature rise rate characterizes the activity and severity of the defect.

[0024] The boundary of the abnormal temperature region is determined by an image segmentation algorithm. The total number of pixels within the boundary is calculated. Spatial calibration is performed based on parameters such as the distance between the infrared thermal imager and the cable, and the focal length of the lens. The pixel area is converted into the actual physical area, which is the region area. The region area reflects the spatial scale of the abnormal temperature segment.

[0025] Furthermore, the process of generating a defect score based on preliminary feature parameters includes:

[0026] The defect score P is:

[0027]

[0028] In the formula, , , , and These are all weighting coefficients, obtained through training with a large amount of historical data. This represents the highest temperature value in the abnormal temperature range. The maximum temperature reference value is taken as the highest permissible temperature of the cable insulation material or the typical maximum value from historical data. This is the average temperature value. This is the average temperature reference value. This represents the temperature difference relative to the normal range. The maximum permissible temperature rise is a safety threshold set according to standards or experience. For the rate of temperature rise, This is a reference value for the maximum temperature rise rate, which can be set based on the maximum rate observed in historical data. For the area, This is the maximum area reference value.

[0029] Furthermore, the process of classifying the defect levels of abnormal temperature ranges based on defect scores includes:

[0030] Set the defect scoring range and assign each abnormal temperature range to the corresponding defect scoring range;

[0031] When P≤X, the abnormal temperature range is classified as normal.

[0032] When X < P ≤ Y, the abnormal temperature range is classified as an abnormal level.

[0033] When Y < P ≤ Z, the abnormal temperature range is classified as the second most severe level.

[0034] When P > Z, the abnormal temperature range is classified as severe.

[0035] Furthermore, the process of obtaining the comprehensive feature vector of the abnormal temperature range based on the preliminary feature parameters of the abnormal temperature range includes:

[0036] All potential abnormal temperature ranges are automatically selected in the panoramic thermal image of the cable. The defect score of all abnormal temperature ranges is obtained according to the defect scoring formula. The defect level of the abnormal temperature range is obtained according to the defect score. For each abnormal temperature range of abnormal or higher level, a specific excitation step has been applied to it and it has been kept for a sufficient time to reach thermal stability, so its corresponding single-order feature vector can be obtained.

[0037] The single-order feature vector refers to a set of preliminary feature parameters extracted under a specific and stable electrical excitation condition for a specific abnormal temperature range region, describing its current thermal state.

[0038] The preliminary characteristic parameters also include the current excitation voltage and current excitation current values ​​of the abnormal temperature range region in the excitation ladder;

[0039] The control test module completely executes all excitation steps of the two sequences of constant current step boost and constant voltage step current boost. Let there be a total of i excitation steps.

[0040] For an abnormal temperature range, under the i-th excitation step, its single-order eigenvector for:

[0041]

[0042] In the formula, Let be the temperature difference between the i-th excitation step and the normal segment. Let be the temperature rise rate of the i-th excitation step. Let be the area of ​​the region of the i-th excitation step. Let be the current excitation voltage value of the i-th excitation step. The current excitation current value for the i-th excitation step;

[0043] For the same abnormal temperature range, its corresponding single-order eigenvector under each excitation step is: , , ,..., ;

[0044] According to the time sequence of the excitation ladder, these i single-order feature vectors are concatenated end to end to obtain the comprehensive feature vector of the abnormal temperature range region.

[0045] The comprehensive feature vector refers to a longer, higher-dimensional vector formed by concatenating the single-order feature vectors obtained under each excitation step in the order of excitation for the same abnormal temperature range region after experiencing all different electrical excitation steps.

[0046] Comprehensive feature vector for:

[0047]

[0048] In the formula, This represents the temperature difference between the first excitation step and the normal segment. The temperature rise rate of the first excitation step. Let this be the area of ​​the region in the first incentive step. This represents the current excitation voltage value for the first excitation step. This is the current excitation current value for the first excitation step;

[0049] The comprehensive feature vector constitutes a digital representation of the dynamic response characteristics of this abnormal temperature range under multi-stage electrical excitation.

[0050] Furthermore, the process of acquiring historical defect type data and constructing a defect classification model based on this data includes:

[0051] The historical defect type data is historical case data of the cable. Each case data includes a comprehensive feature vector of an abnormal temperature range and its corresponding defect type label, which is confirmed by laboratory disassembly.

[0052] Historical defect type data of cables is obtained. For each cable defect segment, i.e. abnormal temperature segment area, its comprehensive feature vector is extracted. All extracted comprehensive feature vectors are preprocessed and combined with their corresponding defect type labels to form a training dataset. The preprocessed comprehensive feature vectors are used as input and the defect type labels are used as output. The gradient boosting decision tree algorithm is used to train the model to obtain the defect classification model.

[0053] Furthermore, the comprehensive feature vector of the abnormal temperature range of the cable at the abnormal level and above is extracted. The extracted comprehensive feature vector is subjected to the same data preprocessing as in the training stage. The preprocessed comprehensive feature vector is then input into the defect classification model to obtain the output of the defect classification model, that is, to obtain the defect type classification result of the abnormal temperature range.

[0054] The defect type classification results include: dielectric loss type defects, partial discharge type defects, resistive contact defects, and armor layer defects.

[0055] The method for detecting cable insulation defects based on infrared imaging includes the following steps:

[0056] S1: Execute a constant current step-up sequence, fix the current at a lower base value, gradually increase the voltage to the rated value, and then execute a constant voltage step-up sequence, fix the voltage at the rated value, and gradually increase the current to the rated value. Thermal stability must be waited for at each excitation step.

[0057] S2: In the thermal stability state of each excitation step, a panoramic thermal image of the cable is acquired using an infrared thermal imager. The abnormal temperature range is identified and located through image processing algorithms, and its preliminary characteristic parameters are calculated. The preliminary characteristic parameters include the highest temperature, average temperature, temperature difference relative to the normal range, temperature rise rate, area, current excitation voltage value, and current excitation current value.

[0058] S3: Based on the extracted preliminary feature parameters, substitute them into the defect scoring formula to calculate the defect score P, and compare them with the preset defect scoring range to initially classify the defects into levels;

[0059] S4: For abnormal temperature ranges assessed as abnormal or above, the single-order feature vector of the corresponding abnormal temperature range is obtained based on the preliminary feature parameters. The single-order feature vectors of all excitation steps are then concatenated in the test order to form a high-dimensional comprehensive feature vector.

[0060] S5: Input the constructed comprehensive feature vector into the pre-trained gradient boosting decision tree classification model, and output the most likely type classification result of defects in the abnormal temperature range.

[0061] The technical effects and advantages of the cable insulation defect detection method and system based on infrared imaging of this invention are as follows:

[0062] (1) By setting up a dual-sequence electric excitation test method of constant current step-up and constant voltage step-up current, the state of the cable under different operating conditions was actively simulated. For light-load cables, the current-up sequence can effectively excite the heating of resistive defects. For various types of cables, the voltage-up sequence can effectively excite the heating of dielectric defects. By analyzing the response of defects in this dynamic excitation process and using the defect classification model to analyze the comprehensive feature vector, it is possible to clearly distinguish defects of different natures such as partial discharge, dielectric loss, and excessive contact resistance, and realize the classification of cable defect types.

[0063] (2) By setting a defect score, the preliminary feature data collected by the panoramic thermal image includes the highest temperature value, average temperature value, temperature difference relative to the normal section, temperature rise rate and area. Based on the preliminary feature data, the corresponding defect score formula is derived. The preliminary feature parameters of the abnormal temperature section of the cable are substituted into the defect score formula to obtain the corresponding defect score. The defect level of the abnormal temperature section is classified according to the defect score. Attached Figure Description

[0064] Figure 1 This is a schematic diagram of the system structure of the present invention. Detailed Implementation

[0065] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0066] Reference Figure 1 The cable insulation defect detection system based on infrared imaging includes the following modules:

[0067] The control test module is used for electrical excitation sequence tests that apply constant current step-up and constant voltage step-up to the cable;

[0068] The data acquisition module is used to acquire thermal imaging images of the cable during the electrical excitation sequence test, and to obtain preliminary characteristic parameters of the abnormal temperature range of the cable based on the thermal imaging images.

[0069] The defect assessment module is used to generate defect scores based on preliminary characteristic parameters and classify defect levels in abnormal temperature ranges based on the defect scores.

[0070] The defect feature module is used to obtain the comprehensive feature vector of the abnormal temperature range based on the preliminary feature parameters of the abnormal temperature range.

[0071] The defect classification module acquires historical defect type data, constructs a defect classification model based on the historical defect type data, and classifies defects in abnormal temperature ranges based on the defect classification model.

[0072] It should be further noted that, in the specific implementation process, the control test module includes a programmable excitation power supply unit and a sequence control unit;

[0073] The programmable excitation power supply unit can independently and precisely control the amplitude of the output voltage and current;

[0074] The sequence control unit is configured to execute the following two independent test sequences:

[0075] The first sequence is a constant current stepped boost: the excitation power supply is controlled to output a constant base current value, and then the output voltage is stepped up in a first preset step size until the rated voltage is reached.

[0076] The base current value in the first sequence is 5% to 20% of the cable's rated current value;

[0077] The second sequence is constant voltage stepped current boost: control the excitation power supply to output the rated voltage, and then increase the output current in a stepped manner with the second preset step size until the rated current is reached;

[0078] The first preset step size and the second preset step size are 5%, 10%, or 20% of the rated value.

[0079] It should be further explained that, in the specific implementation process, the data acquisition module includes an infrared thermal imager, an image processing unit, and a preliminary feature extraction unit;

[0080] The infrared thermal imager is used to collect infrared radiation data on the cable surface and generate a temperature field distribution image after each excitation step reaches a thermally stable state.

[0081] The image processing unit is used to filter and reduce noise, calibrate temperature, and stitch images on the temperature field distribution image to generate a panoramic thermal image of the cable.

[0082] The preliminary feature extraction unit is used to analyze the panoramic thermal image, automatically identify and locate one or more abnormal temperature range regions, and obtain preliminary feature parameters for each abnormal temperature range region. The preliminary feature parameters include: the highest temperature value, average temperature value, temperature difference relative to the normal range, temperature rise rate, and area of ​​the abnormal temperature range region.

[0083] Within the identified abnormal temperature range, the temperature values ​​of all pixels are traversed, and the maximum value is the highest temperature value. The highest temperature value directly reflects the extreme heating situation in the abnormal temperature range. The highest temperature value is a key indicator for calculating the temperature rise and determining whether the material safety threshold is exceeded. Resistive defects usually result in a very high highest temperature value.

[0084] The arithmetic mean of the temperature values ​​of all pixels within the abnormal temperature range is used to obtain the average temperature value. The average temperature value reflects the overall heat generation level of the abnormal area and represents the total thermal power of the defect. When the area is similar, the defect with the higher average temperature value generates more heat and is usually more severe.

[0085] On the same cable, select a section that is far from the abnormal temperature area and has no abnormal behavior, calculate its average temperature to obtain the average temperature value of the normal section, and then subtract the average temperature value of the normal section from the average temperature value of the abnormal temperature area to obtain the temperature difference relative to the normal section. The temperature difference relative to the normal section directly represents the temperature rise caused by the defect itself.

[0086] Under the excitation ladder, from the initial application of excitation to the temperature stabilization, the temperature change curve over time is recorded. The temperature rise rate is the average slope or first derivative of the curve during the main rising phase. The temperature rise rate characterizes the activity and severity of the defect. A severe or sharp defect will generate heat rapidly, resulting in a large temperature rise rate, while a slowly developing defect will have a slower temperature rise rate.

[0087] The boundary of the abnormal temperature region is determined by an image segmentation algorithm. The total number of pixels within the boundary is calculated. Spatial calibration is performed based on parameters such as the distance between the infrared thermal imager and the cable, and the focal length of the lens. The pixel area is converted into the actual physical area, which is the region area. The region area reflects the spatial scale of the abnormal temperature segment.

[0088] It should be further explained that, in the specific implementation process, the process of generating a defect score based on preliminary characteristic parameters includes:

[0089] The defect score P is:

[0090]

[0091] In the formula, , , , and These are all weighting coefficients, with values ​​of 0.2, 0.2, 0.3, 0.1, and 0.2 respectively. This represents the highest temperature value in the abnormal temperature range. The maximum temperature reference value is taken as the highest permissible temperature of the cable insulation material or the typical maximum value from historical data, specifically 100. This is the average temperature value. This is an average temperature reference value, specifically 80. This represents the temperature difference relative to the normal range. The maximum permissible temperature rise is set as a safety threshold based on standards or experience, specifically 20. For the rate of temperature rise, This is a reference value for the maximum temperature rise rate, which can be set based on the maximum rate observed in historical data; specifically, it is 2. For the area, This is the maximum area reference value, specifically 200;

[0092] Let the data for a certain abnormal temperature range be: It is 65. It is 55. It is 15. If S is 0.8 and P is 50, then P is 0.5825.

[0093] It should be further explained that, in the specific implementation process, the process of classifying the defect level of the abnormal temperature range based on the defect score includes:

[0094] Set the defect scoring range and assign each abnormal temperature range to the corresponding defect scoring range;

[0095] When P≤X, the abnormal temperature range is classified as normal.

[0096] When X < P ≤ Y, the abnormal temperature range is classified as an abnormal level.

[0097] When Y < P ≤ Z, the abnormal temperature range is classified as the second most severe level.

[0098] When P > Z, the abnormal temperature range is classified as severe.

[0099] When P is 0.5825, X is 0.5, and Y is 0.6, the abnormal temperature range is determined to be of an abnormal level, and the abnormal temperature range of the cable is marked.

[0100] It should be further explained that, in the specific implementation process, the process of obtaining the comprehensive feature vector of the abnormal temperature range based on the preliminary feature parameters of the abnormal temperature range includes:

[0101] All potential abnormal temperature ranges are automatically selected in the panoramic thermal image of the cable. The defect score of all abnormal temperature ranges is obtained according to the defect scoring formula. The defect level of the abnormal temperature range is obtained according to the defect score. For each abnormal temperature range of abnormal or higher level, a specific excitation step has been applied to it and it has been kept for a sufficient time to reach thermal stability, so its corresponding single-order feature vector can be obtained.

[0102] The single-order feature vector refers to a set of preliminary feature parameters extracted under a specific and stable electrical excitation condition for a specific abnormal temperature range region, describing its current thermal state.

[0103] The preliminary characteristic parameters also include the current excitation voltage and current excitation current values ​​of the abnormal temperature range region in the excitation ladder;

[0104] The control test module completely executes all excitation steps of the two sequences of constant current step boost and constant voltage step current boost. Let there be a total of i excitation steps.

[0105] For an abnormal temperature range, under the i-th excitation step, its single-order eigenvector for:

[0106]

[0107] In the formula, Let be the temperature difference between the i-th excitation step and the normal segment. Let be the temperature rise rate of the i-th excitation step. Let be the area of ​​the region of the i-th excitation step. Let be the current excitation voltage value of the i-th excitation step. The current excitation current value for the i-th excitation step;

[0108] For the same abnormal temperature range, its corresponding single-order eigenvector under each excitation step is: , , ,..., ;

[0109] According to the time sequence of the excitation ladder, these i single-order feature vectors are concatenated end to end to obtain the comprehensive feature vector of the abnormal temperature range region.

[0110] The comprehensive feature vector refers to a longer, higher-dimensional vector formed by concatenating the single-order feature vectors obtained under each excitation step in the order of excitation for the same abnormal temperature range region after experiencing all different electrical excitation steps.

[0111] Comprehensive feature vector for:

[0112]

[0113] In the formula, This represents the temperature difference between the first excitation step and the normal segment. The temperature rise rate of the first excitation step. Let this be the area of ​​the region in the first incentive step. This represents the current excitation voltage value for the first excitation step. This is the current excitation current value for the first excitation step;

[0114] The comprehensive feature vector constitutes a digital representation of the dynamic response characteristics of this abnormal temperature range under multi-stage electrical excitation.

[0115] Each There are 5 features, and a total of i stimulus ladder tests were performed. It is A 50-dimensional vector, for example, 10 steps would be 50-dimensional;

[0116] It is not a static value, but a dynamic curve that records the heating characteristics of the abnormal temperature range and how it changes with external electrical excitation (voltage / current).

[0117] It should be further explained that, in the specific implementation process, the process of acquiring historical defect type data and constructing a defect classification model based on this data includes:

[0118] The historical defect type data is historical case data of the cable. Each case data includes a comprehensive feature vector of an abnormal temperature range and its corresponding defect type label, which is confirmed by laboratory disassembly.

[0119] Historical defect type data of cables is obtained. For each cable defect segment, i.e. abnormal temperature segment area, its comprehensive feature vector is extracted. All extracted comprehensive feature vectors are preprocessed and combined with their corresponding defect type labels to form a training dataset. The preprocessed comprehensive feature vectors are used as input and the defect type labels are used as output. The gradient boosting decision tree algorithm is used to train the model to obtain the defect classification model.

[0120] It should be further explained that, in the specific implementation process, the process of classifying defects in abnormal temperature ranges according to the defect classification model includes:

[0121] Extract the comprehensive feature vector of the abnormal temperature range of the cable at the abnormal level and above. Perform the same data preprocessing on the extracted comprehensive feature vector as in the training phase. Input the preprocessed comprehensive feature vector into the defect classification model to obtain the output of the defect classification model, that is, obtain the defect type classification result of the abnormal temperature range.

[0122] The defect type classification results include: dielectric loss type defects, partial discharge type defects, resistive contact defects, and armor layer defects, among which:

[0123] Resistive contact defects: their ΔT increases significantly in the current-boosting sequence, but remains almost unchanged in the voltage-boosting sequence;

[0124] Dielectric loss type defect: its ΔT increases significantly in the boost sequence, but does not change much in the current sequence.

[0125] The method for detecting cable insulation defects based on infrared imaging includes the following steps:

[0126] S1: Execute a constant current step-up sequence, fix the current at a lower base value, gradually increase the voltage to the rated value, and then execute a constant voltage step-up sequence, fix the voltage at the rated value, and gradually increase the current to the rated value. Thermal stability must be waited for at each excitation step.

[0127] S2: In the thermal stability state of each excitation step, a panoramic thermal image of the cable is acquired using an infrared thermal imager. The abnormal temperature range is identified and located through image processing algorithms, and its preliminary characteristic parameters are calculated. The preliminary characteristic parameters include the highest temperature, average temperature, temperature difference relative to the normal range, temperature rise rate, area, current excitation voltage value, and current excitation current value.

[0128] S3: Based on the extracted preliminary feature parameters, substitute them into the defect scoring formula to calculate the defect score P, and compare them with the preset defect scoring range to initially classify the defects into levels;

[0129] S4: For abnormal temperature ranges assessed as abnormal or above, the single-order feature vector of the corresponding abnormal temperature range is obtained based on the preliminary feature parameters. The single-order feature vectors of all excitation steps are then concatenated in the test order to form a high-dimensional comprehensive feature vector.

[0130] S5: Input the constructed comprehensive feature vector into the pre-trained gradient boosting decision tree classification model, and output the most likely type classification result of defects in the abnormal temperature range.

[0131] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0132] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A cable insulation defect detection system based on infrared imaging, characterized in that, Includes the following modules: The control test module is used for electrical excitation sequence tests that apply constant current step-up and constant voltage step-up to the cable; The data acquisition module is used to acquire thermal imaging images of the cable during the electrical excitation sequence test, and to obtain preliminary characteristic parameters of the abnormal temperature range of the cable based on the thermal imaging images. The defect assessment module is used to generate defect scores based on preliminary characteristic parameters and classify defect levels in abnormal temperature ranges based on the defect scores. The defect feature module is used to obtain the comprehensive feature vector of the abnormal temperature range based on the preliminary feature parameters of the abnormal temperature range. The defect classification module acquires historical defect type data, constructs a defect classification model based on the historical defect type data, and classifies defects in abnormal temperature ranges based on the defect classification model. The control test module includes a programmable excitation power supply unit and a sequence control unit; The programmable excitation power supply unit can independently and precisely control the amplitude of the output voltage and current; The sequence control unit is configured to execute the following two independent test sequences: The first sequence is a constant current stepped boost: the excitation power supply is controlled to output a constant base current value, and then the output voltage is stepped up in a first preset step size until the rated voltage is reached. The second sequence is a constant voltage stepped current ramp: the excitation power supply outputs the rated voltage, and then the output current is increased in a stepped manner with a second preset step size until the rated current is reached.

2. The cable insulation defect detection system based on infrared imaging according to claim 1, characterized in that, The data acquisition module includes an infrared thermal imager, an image processing unit, and a preliminary feature extraction unit; The infrared thermal imager is used to collect infrared radiation data on the cable surface and generate a temperature field distribution image after each excitation step reaches a thermally stable state. The image processing unit is used to filter and reduce noise, calibrate temperature, and stitch images on the temperature field distribution image to generate a panoramic thermal image of the cable. The preliminary feature extraction unit is used to analyze the panoramic thermal image, automatically identify and locate one or more abnormal temperature range regions, and obtain preliminary feature parameters for each abnormal temperature range region. The preliminary feature parameters include: the highest temperature value, average temperature value, temperature difference relative to the normal range, temperature rise rate, and area of ​​the abnormal temperature range region. Within the identified abnormal temperature range, the temperature values ​​of all pixels are traversed, and the maximum value is the highest temperature value, which directly reflects the extreme heating situation in the abnormal temperature range. The arithmetic mean of the temperature values ​​of all pixels within the abnormal temperature range is used to obtain the average temperature value. The average temperature value reflects the overall heat generation level of the abnormal area and represents the total thermal power of the defect. On the same cable, select a section that is far from the abnormal temperature area and has no abnormal behavior, calculate its average temperature to obtain the average temperature value of the normal section, and then subtract the average temperature value of the normal section from the average temperature value of the abnormal temperature area to obtain the temperature difference relative to the normal section. The temperature difference relative to the normal section directly represents the temperature rise caused by the defect itself. Under the excitation ladder, from the initial application of excitation to the temperature stabilization, the temperature change curve over time is recorded. The temperature rise rate is the average slope or first derivative of the curve during the main rising phase. The temperature rise rate characterizes the activity and severity of the defect. The boundary of the abnormal temperature region is determined by an image segmentation algorithm. The total number of pixels within the boundary is calculated. Spatial calibration is performed based on parameters such as the distance between the infrared thermal imager and the cable, and the focal length of the lens. The pixel area is converted into the actual physical area, which is the region area. The region area reflects the spatial scale of the abnormal temperature segment.

3. The cable insulation defect detection system based on infrared imaging according to claim 2, characterized in that, The process of generating a defect score based on preliminary feature parameters includes: The defect score P is: In the formula, , , , and These are all weighting coefficients, obtained through training with a large amount of historical data. This represents the highest temperature value in the abnormal temperature range. The maximum temperature reference value is taken as the highest permissible temperature of the cable insulation material or the typical maximum value from historical data. This is the average temperature value. This is the average temperature reference value. This represents the temperature difference relative to the normal range. The maximum permissible temperature rise is a safety threshold set according to standards or experience. For the rate of temperature rise, This is a reference value for the maximum temperature rise rate, which can be set based on the maximum rate observed in historical data. For the area, This is the maximum area reference value.

4. The cable insulation defect detection system based on infrared imaging according to claim 3, characterized in that, The process of classifying the defect level of abnormal temperature ranges based on defect scores includes: Set the defect scoring range and assign each abnormal temperature range to the corresponding defect scoring range; When P≤X, the abnormal temperature range is classified as normal. When X < P ≤ Y, the abnormal temperature range is classified as an abnormal level. When Y < P ≤ Z, the abnormal temperature range is classified as the second most severe level. When P > Z, the abnormal temperature range is classified as severe.

5. The cable insulation defect detection system based on infrared imaging according to claim 4, characterized in that, The process of obtaining the comprehensive feature vector of the abnormal temperature range based on the preliminary feature parameters of the abnormal temperature range includes: All potential abnormal temperature ranges are automatically selected in the panoramic thermal image of the cable. The defect score of all abnormal temperature ranges is obtained according to the defect scoring formula. The defect level of the abnormal temperature range is obtained according to the defect score. For each abnormal temperature range of abnormal or higher level, a specific excitation step has been applied to it and it has been kept for a sufficient time to reach thermal stability, so its corresponding single-order feature vector can be obtained. The single-order feature vector refers to a set of preliminary feature parameters extracted under a specific and stable electrical excitation condition for a specific abnormal temperature range region, describing its current thermal state. The preliminary characteristic parameters also include the current excitation voltage and current excitation current values ​​of the abnormal temperature range region in the excitation ladder; The control test module completely executes all excitation steps of the two sequences of constant current step boost and constant voltage step current boost. Let there be a total of i excitation steps. For an abnormal temperature range, under the i-th excitation step, its single-order eigenvector for: In the formula, Let be the temperature difference between the i-th excitation step and the normal segment. Let be the temperature rise rate of the i-th excitation step. Let be the area of ​​the region of the i-th excitation step. Let be the current excitation voltage value of the i-th excitation step. The current excitation current value for the i-th excitation step; For the same abnormal temperature range, its corresponding single-order eigenvector under each excitation step is: , , ,..., ; According to the time sequence of the excitation ladder, these i single-order feature vectors are concatenated end to end to obtain the comprehensive feature vector of the abnormal temperature range region. The comprehensive feature vector refers to a longer, higher-dimensional vector formed by concatenating the single-order feature vectors obtained under each excitation step in the order of excitation for the same abnormal temperature range region after experiencing all different electrical excitation steps. Comprehensive feature vector for: In the formula, This represents the temperature difference between the first excitation step and the normal segment. The temperature rise rate of the first excitation step. Let this be the area of ​​the region in the first incentive step. This represents the current excitation voltage value for the first excitation step. This is the current excitation current value for the first excitation step; The comprehensive feature vector constitutes a digital representation of the dynamic response characteristics of this abnormal temperature range under multi-stage electrical excitation.

6. The cable insulation defect detection system based on infrared imaging according to claim 5, characterized in that, The process of acquiring historical defect type data and building a defect classification model based on that data includes: The historical defect type data is historical case data of the cable. Each case data includes a comprehensive feature vector of an abnormal temperature range and its corresponding defect type label, which is confirmed by laboratory disassembly. Historical defect type data of cables is obtained. For each cable defect segment, i.e. abnormal temperature segment area, its comprehensive feature vector is extracted. All extracted comprehensive feature vectors are preprocessed and combined with their corresponding defect type labels to form a training dataset. The preprocessed comprehensive feature vectors are used as input and the defect type labels are used as output. The gradient boosting decision tree algorithm is used to train the model to obtain the defect classification model.

7. The cable insulation defect detection system based on infrared imaging according to claim 6, characterized in that, Extract the comprehensive feature vector of the abnormal temperature range of the cable at the abnormal level and above. Perform the same data preprocessing on the extracted comprehensive feature vector as in the training phase. Input the preprocessed comprehensive feature vector into the defect classification model to obtain the output of the defect classification model, that is, obtain the defect type classification result of the abnormal temperature range. The defect type classification results include: dielectric loss type defects, partial discharge type defects, resistive contact defects, and armor layer defects.

8. A cable insulation defect detection method based on infrared imaging, implemented based on the cable insulation defect detection system based on infrared imaging according to any one of claims 1-7, characterized in that, Includes the following steps: S1: Execute a constant current step-up sequence, fix the current at a lower base value, gradually increase the voltage to the rated value, and then execute a constant voltage step-up sequence, fix the voltage at the rated value, and gradually increase the current to the rated value. Thermal stability must be waited for at each excitation step. S2: In the thermal stability state of each excitation step, a panoramic thermal image of the cable is acquired using an infrared thermal imager. The abnormal temperature range is identified and located through image processing algorithms, and its preliminary characteristic parameters are calculated. The preliminary characteristic parameters include the highest temperature, average temperature, temperature difference relative to the normal range, temperature rise rate, area, current excitation voltage value, and current excitation current value. S3: Based on the extracted preliminary feature parameters, substitute them into the defect scoring formula to calculate the defect score P, and compare them with the preset defect scoring range to initially classify the defects into levels; S4: For abnormal temperature ranges assessed as abnormal or above, the single-order feature vector of the corresponding abnormal temperature range is obtained based on the preliminary feature parameters. The single-order feature vectors of all excitation steps are then concatenated in the test order to form a high-dimensional comprehensive feature vector. S5: Input the constructed comprehensive feature vector into the pre-trained gradient boosting decision tree classification model, and output the most likely type classification result of defects in the abnormal temperature range.

Images