Superconducting cable and defect detection method thereof

Abstract

The invention relates to the technical field of cable detection, solves the technical problem that the omission ratio and false defect misjudgment rate of superconducting cable defect detection are improved due to the fact that corresponding adaptation is not carried out for the particularity of a superconducting cable in the prior art, and particularly relates to a superconducting cable and a defect detection method thereof. A multi-modal sensor array is spirally arranged in the circumferential direction and the axial direction of the superconducting cable, multi-modal data of all layers of the superconducting cable are collected and preprocessed, then feature extraction is carried out on the multi-modal data to obtain multi-modal features including electromagnetic features, temperature field features and structural vibration features, quantum bit coding is carried out on the multi-modal features, and quantum bit coding is carried out on the multi-modal features. And cross-modal association weights among the multi-modal features are calculated through the entanglement entropy. According to the method, the accuracy of superconducting cable defect detection is improved, different types of defects such as superconducting layer cracks, insulating layer corrosion and shielding layer damage are accurately distinguished, the omission ratio of key defects is greatly reduced, and the reliability of a superconducting cable detection result is improved.

Description

Technical Field

[0001] This invention relates to the field of cable testing technology, and in particular to a method for detecting defects in superconducting cables. Background Technology

[0002] Superconducting cables are a type of cable made using superconductors. They offer advantages such as high capacity, low loss, energy saving, and environmental friendliness, and are widely used in the power industry. Currently, defect detection in superconducting cables mostly employs a multi-modal data acquisition, feature processing, and defect identification process. However, this method only separates noise from valid signals based on signal frequency, without comprehensively considering the layered structure characteristics, critical parameters, and low-temperature, high-electromagnetic operating environment of superconducting cables. This means existing detection methods are not adapted to the specific characteristics of superconducting cables, affecting the multi-dimensional distribution of layered defects and the expansion patterns of defects at low temperatures. Consequently, the false negative rate and misjudgment rate of defect detection increase, reducing the accuracy of defect detection. This results in a disconnect between the detection results and the actual operating performance and defect severity of the superconducting cable, failing to meet the high-reliability operation requirements of superconducting cables. Summary of the Invention

[0003] To address the shortcomings of existing technologies, this invention provides a method for detecting superconducting cables and their defects. This method solves the technical problem that existing technologies have not been adapted to the specific characteristics of superconducting cables, which leads to increased false defect detection rates and misjudgment rates. This invention aims to improve the accuracy of superconducting cable defect detection.

[0004] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a defect detection method for superconducting cables, the method comprising the following steps: S1. A multi-mode sensor array is arranged in a spiral pattern along the circumferential and axial directions of the superconducting cable to collect and preprocess the multi-mode data of each layer of the superconducting cable. Then, feature extraction is performed on the data to obtain multi-mode features including electromagnetic features, temperature field features and structural vibration features. S2. Encode the multimodal features with qubits, calculate the cross-modal correlation weights between each multimodal feature through entanglement entropy, introduce an optimized multimodal feature contrast adaptive enhancement coefficient to normalize each feature, and calculate the weighted feature aggregation value. S3. Construct an initial individual vector based on the weighted feature aggregation value, combine it with each layer of the superconducting cable to construct a multi-dimensional search space, and generate a chaotic individual vector covering the multi-dimensional search space based on the chaotic mapping. Merge it with the initial individual vector to form an initial particle swarm and transform it into a quantum particle swarm. S4. Introduce critical parameters of superconducting cables to construct local optimization constraint terms to calculate the fitness value of particles, and based on the preset safety threshold, select the optimal individual vectors of each layer to generate a defect candidate set. S5. Map the defect candidate set to a three-dimensional defect image, enhance the key defect region features to transform it into an enhanced defect image, and introduce an attenuation correction factor to calculate the feature normalization value of each pixel and integrate it into a feature normalization matrix. S6. Calculate the scale eigenvalues, spatial gradient amplitudes, directional values, and defect significance values ​​of each layer of the superconducting cable based on the feature normalization matrix, and establish a superconducting defect specific criterion based on the unique signal pattern of superconducting defects, classifying the enhanced defect images into several defect sets.

[0005] Furthermore, each layer of the superconducting cable includes a superconducting layer, an insulating layer, and a shielding layer. The multimodal data includes electromagnetic data, temperature field data, and structural vibration data. Step S2 specifically includes the following steps: S21. Map the electromagnetic features, temperature field features, and structural vibration features to a preset quantum Hilbert space, convert them into qubit codes, and establish the correspondence between multimodal feature values ​​and qubit states. S22. Based on entanglement entropy quantization, the quantum correlation strength of any two features in a multimodal feature is calculated, and the cross-modal correlation weight of each feature and the entanglement entropy of the quantum state of each feature are calculated. The calculation formula is as follows:

[0006]

[0007] In the above formula, Indicates the first The first feature and the second Cross-modal association weights of each feature, Indicates the first The first feature and the second Entanglement entropy of a characteristic quantum state Indicates the first The first feature and the second The reduced density matrix of each characteristic quantum state, Represents the matrix trace operation. This represents the maximum value of the entanglement entropy for each feature; S23. Based on the preset critical current and critical temperature of the superconducting cable, calculate the adaptive enhancement coefficient used to optimize the contrast of multimodal features. The calculation formula is as follows:

[0008] In the above formula, Indicates the first The adaptive enhancement coefficients of each feature Indicates the first The rate of change of superconducting layer resistance corresponding to each characteristic Indicates the first The dielectric constant deviation of the insulating layer corresponding to each characteristic This indicates the critical current of the superconducting cable. This indicates the critical temperature of the superconducting cable. This indicates the real-time operating temperature of the superconducting cable. Indicates the first The standard deviation of each feature This represents the average of the standard deviations of each characteristic. Represents the natural exponential function; S24. The enhanced multimodal features are concatenated into tensors for each layer to form a high-dimensional feature tensor. Based on multi-scale median filtering and the superconducting property normalization function, the normalized eigenvalues ​​of each feature used to eliminate dimensional differences in the multimodal features are calculated. The calculation formula is as follows:

[0009] In the above formula, Indicates the first Normalized eigenvalues ​​of each feature This indicates a multi-scale median filtering operation. This indicates a tensor splicing operation. Indicates the first Electromagnetic data corresponding to each feature Indicates the first Temperature field data corresponding to each feature Indicates the first Structural vibration data corresponding to each feature This represents the minimum value of each feature after enhancement. This represents the maximum value of the enhanced feature; S25. Based on the signal-to-noise ratio and critical parameters of the superconducting cable, a dual-constraint model is constructed to calculate the fusion weights of each feature. The calculation formula is as follows:

[0010]

[0011] In the above formula, Indicates the first The fusion weights of each feature Indicates the first The signal-to-noise ratio of each feature Indicates the first The real-time operating current of the superconducting cable corresponding to each characteristic Indicates the first The power of the effective signal in each feature, Indicates the first The power of noise in each feature, Represents the logarithmic function with base 10; S26. Calculate the weighted feature aggregation value based on the cross-modal correlation weight and the fusion weight of the dual constraints. The calculation formula is as follows:

[0012] In the above formula, Indicates the first The weighted feature aggregation value of each feature. Indicates the first The fusion weights of each feature Indicates the first Normalized eigenvalues ​​of each feature.

[0013] Furthermore, step S3 specifically includes the following steps: S31. Based on the weighted feature aggregation value as the initial individual basis, each weighted feature aggregation value is mapped to the defect feature dimension of the corresponding layer of the superconducting cable to calculate the initial individual vector. The calculation formula is as follows:

[0014] In the above formula, Indicates the first The initial individual vector corresponding to each feature Indicates the first The weighted feature aggregation value of each feature is mapped to the defect feature components of the superconducting layer. Indicates the first The weighted feature aggregation value of each feature is mapped to the defect feature components of the insulation layer. Indicates the first The weighted feature aggregation value of each feature is mapped to the defect feature components of the shielding layer. This represents the vector transpose operation; S32. Based on the physical characteristics and defect types of each layer of the superconducting cable, independent search boundaries are set for the defect feature components of each layer, and a multi-dimensional search space with layer dimension and defect parameter dimension is constructed, as shown in the following expression:

[0015] In the above formula, Indicates a hierarchical index. Indicates the first The lower boundary of the search space for each layered defect feature component Indicates the first The upper boundary of the search space for each hierarchical defect feature component. Indicates the first The lower boundary correction coefficient for each layer, Indicates the first The upper boundary correction coefficient of each layer, Indicates the first The minimum detectable defect parameter value for each layer. Indicates the first The maximum detectable defect parameter value for each layer; S33. Based on chaotic mapping, the superconducting critical temperature is incorporated into the chaotic parameters, and the chaotic individual vectors covering each feature of the multi-dimensional search space are calculated. The calculation formula is as follows:

[0016]

[0017] In the above formula, Indicates the first A chaotic individual vector, Indicates the first Output value of the next chaotic iteration Indicates the first Output value of the next chaotic iteration Indicates the first One chaotic parameter; S34. Merge the initial individuals and chaotic individuals into an initial particle swarm, and transform it into a quantum particle swarm adapted for superconducting cable defect detection. Calculate the quantum state probability amplitude vector and the probability density of each dimension. The calculation formula is as follows:

[0018]

[0019] In the above formula, Indicates the first The quantum state probability magnitude vector of each particle. Indicates the first The particle in the first Quantum probability amplitude in each hierarchical dimension Indicates the first The particle in the first The probability density of each hierarchical dimension Indicates the first The particle in the first Feature values ​​of each hierarchical dimension.

[0020] Furthermore, step S4 specifically includes the following steps: S41. Based on the global search principle of quantum particle swarm optimization, a local optimization constraint term is constructed by introducing critical parameters of the superconducting cable, including critical current and critical temperature. The correlation between defect signal intensity and superconducting performance attenuation is quantified to calculate the fitness value of the particles. The calculation formula is as follows:

[0021]

[0022]

[0023] In the above formula, Indicates the first The fitness value of each particle. Indicates the defect signal strength weight. This indicates the correlation weight between the degradation of superconducting performance and its degree of degradation. Indicates the first The defect signal intensity corresponding to each particle Indicates the first The correlation of superconducting performance degradation for each particle Indicates the first The defect signal amplitude corresponding to each particle This represents the amplitude of the background signal when there are no defects. This represents the signal amplitude corresponding to the largest known defect. This indicates the lowest temperature at which the superconducting cable can operate stably. This indicates the critical current of the superconducting cable. Indicates the first The real-time operating current of the superconducting cable corresponding to each characteristic This indicates the critical temperature of the superconducting cable. This indicates the real-time operating temperature of the superconducting cable; S42. Based on the dynamic adjustment ratio mechanism between the current iteration number and the maximum iteration number, a safety threshold coupled with the defect propagation rate under low temperature conditions is constructed and calculated, and effective particle individuals with fitness values ​​greater than the safety threshold are retained. The calculation formula for the safety threshold is as follows:

[0024]

[0025] In the above formula, Indicates the first The rate of low-temperature defect propagation. This represents the defect propagation rate constant. The activation energy represents the defect propagation. Represents the Boltzmann constant. Indicates the first Real-time operating temperature of the superconducting cable. Indicates the first The safety threshold for each step. Indicates the basic security threshold. Indicates the maximum number of iterations. Indicates the current iteration number; S43. Determine whether the current iteration count has reached the preset maximum iteration count; If so, then select all valid particle individuals from the iterations as the optimal individual vector for each layer, and proceed to the next step; If not, return to step S41; S44. The optimal individual vectors selected from each layer are spliced ​​together as tensors, and the defect information of each layer is integrated to form a defect candidate set that includes the defect parameters of the entire layer and the impact assessment of superconducting performance.

[0026] Furthermore, the three-dimensional defect image includes information on each layer of the superconducting layer, insulating layer, and shielding layer. Step S5 specifically includes the following steps: S51. Map the defect features of each layer of the optimal individual vector in the defect candidate set to independent cross-modal channels. Based on the influence weights of each layer of the superconducting cable, weighted superposition of the multimodal channel features of the same layer is performed to calculate the layered multimodal fusion features corresponding to the optimal individual vector of each layer. The calculation formula is as follows:

[0027]

[0028] In the above formula, Indicates the first The optimal individual vector at the th th th A layered multimodal fusion feature Indicates the first The performance impact weight of each layer Indicates the first The optimal individual vector at the th ... The first layer Feature vectors of each modal channel Indicates the first The optimal individual vector at the th th th Each layer of defect feature value, Indicates the first The modal channel in the first... Each layer has its own adaptation coefficient; S52. By expanding the multimodal channel dimension, the three-layer fused features are stitched together into a three-dimensional defect image tensor. Gaussian pyramid and Laplacian pyramid decomposition is performed on each layer of the three-dimensional defect image tensor. The top few scale features with the highest attention weight are retained, and the key defect region features are enhanced to obtain the enhanced feature layer. S53. Map the enhanced feature layer to RGB pixel values, incorporate the critical parameters and real-time operating parameters of the superconducting cable, and calculate the RGB pixel values ​​that reflect the intensity of the defect features and the associated superconducting performance state to transform the enhanced feature layer into an enhanced defect image. The calculation formula is as follows:

[0029] In the above formula, Indicates the first The defect image is in the first Each layer of RGB pixel values, Indicates the first The defect image is in the first A layered enhancement feature layer, Indicates the first The defect image is in the first The maximum value of each hierarchical enhancement feature layer Indicates the first The real-time dielectric constant of the insulating layer corresponding to each defect image. Indicates the standard dielectric constant of the insulating layer. This indicates the critical current of the superconducting cable. Indicates the first The real-time operating current of the superconducting cable corresponding to each characteristic This indicates the critical temperature of the superconducting cable. This indicates the real-time operating temperature of the superconducting cable; S54. Select the target pixel in the enhanced defect image, extract its neighboring pixels, and introduce an attenuation correction factor to correct for low-temperature attenuation. Calculate the neighborhood feature value after low-temperature correction using the following formula:

[0030] In the above formula, Indicates the first The defect image is in the first Each layer, target pixel The neighborhood feature values, Represents target pixel The set of neighboring pixels, Represents neighboring pixels grayscale value, Represents target pixel grayscale value, Indicates the normal temperature reference temperature; S55. Based on the statistical distribution of neighborhood feature values, combined with real-time temperature and electromagnetic interference intensity, calculate the standardized feature values ​​of all pixels in the defect image and integrate them into a feature standardization matrix. The calculation formula is as follows:

[0031] In the above formula, Indicates the first The defect image is in the first Each layer, target pixel The characteristic standardized value, Indicates the first The average of all neighborhood feature values ​​in each layer Indicates the first The standard deviation of all neighborhood feature values ​​in each layer Indicates the real-time electromagnetic interference intensity. Indicates the electromagnetic interference correction factor. This indicates the interval translation adjustment term.

[0032] Furthermore, the unique signal modes of the superconducting defects include electromagnetic signal abrupt changes corresponding to superconducting layer cracks, dielectric constant anomalies corresponding to insulating layer corrosion, and structural vibration signal distortion corresponding to shielding layer damage. Step S6 specifically includes the following steps: S61. Perform convolution operation between the preset multi-scale Gaussian difference kernel and the feature normalization matrix to extract the scale feature values ​​of each layer, and calculate the scale feature value matrix. The calculation formula is as follows:

[0033] In the above formula, Indicates the first The defective image, the first Each layer, scale The scale eigenvalue matrix, Indicates the first The defective image, the first A hierarchical feature standardization matrix, This represents a two-dimensional convolution operation. Indicates the first Each layer, scale Superconducting-specific Gaussian difference kernel; S62. Based on the critical parameters and real-time operating parameters of the superconducting cable, calculate the superconducting performance attenuation weight for each defect image. The calculation formula is as follows:

[0034] In the above formula, Indicates the first The superconducting performance degradation weight of each defect image This indicates the critical current of the superconducting cable. Indicates the first The real-time operating current of the superconducting cable corresponding to each characteristic This indicates the critical temperature of the superconducting cable. This indicates the real-time operating temperature of the superconducting cable; S63. Based on the scale feature values ​​of each layer, calculate the spatial gradient magnitude reflecting the intensity of the defect edge and the directional value reflecting the direction of defect extension. The calculation formulas are as follows:

[0035]

[0036] In the above formula, Indicates the first The defective image, the first Each layer, scale The weighted gradient magnitude, Indicates the first The defective image, the first Each layer, scale The gradient direction value, Represents the arctangent function in the four quadrants. The gradient response matrix represents the horizontal or vertical direction; S64. Combining the weighted gradient magnitude, gradient direction consistency, and superconductivity sensitivity coefficient, the defect significance value is calculated using the following formula:

[0037]

[0038] In the above formula, Indicates the first The defective image, the first Each layer of defect significance value, Indicates the total number of scales. Indicates the first Each layer, scale Gradient direction consistency This represents the variance of the gradient direction values. Indicates the sensitivity coefficient for superconducting properties; S64. Based on the unique signal modes of superconducting defects, establish a superconducting defect specificity criterion and calculate the defect specificity criterion threshold. The calculation formula is as follows:

[0039] In the above formula, These represent the specific criterion thresholds for superconducting layer cracks, insulating layer corrosion, and shielding layer damage, respectively. They represent the specificity criterion coefficients, Indicates the first The electromagnetic mode signal sequence corresponding to each defect image. Indicates the first The temperature field mode signal sequence corresponding to each defect image. Indicates the first The structural vibration mode signal sequence corresponding to each defect image. This represents variance calculation. This represents entropy value operations. This represents the root mean square operation. Indicates the first The real-time dielectric constant of the insulating layer corresponding to each defect image. Indicates the standard dielectric constant of the insulating layer; S65. Construct a cross-modal verification mechanism for defect saliency and modal signals to determine whether the defect saliency value is greater than the specificity criterion threshold of the corresponding layer and whether the modal signal pattern matches the defect type. If so, the verification result is correct, proceed to the next step; If not, then the significance of the defect is eliminated; S66. Combining the specificity criteria and cross-modal verification results, the enhanced defect images are classified, and the verified defects are assigned to the corresponding layer defect sets. The defect sets include the superconducting layer crack defect set, the insulating layer corrosion defect set, and the shielding layer damage defect set.

[0040] A superconducting cable includes a superconducting layer covered with a high-voltage insulation layer, a high-temperature superconducting shielding layer covered outside the high-voltage insulation layer, and an outer protective layer covered outside the high-temperature superconducting shielding layer.

[0041] By employing the above technical solution, the present invention provides a method for detecting superconducting cables and their defects, which has at least the following beneficial effects: 1. This invention employs a spiral multimodal sensor array arranged along the circumference and axis of a superconducting cable to precisely adapt to the differences in the layered structure of the superconducting layer, insulation layer, and shielding layer of the superconducting cable, as well as the propagation characteristics of defect signals. Combined with targeted data preprocessing and modal feature extraction, it can effectively capture electromagnetic, temperature field, and structural vibration defect signals of each layer, thereby improving the accuracy of superconducting cable defect detection. This solves the problem of inaccurate capture of layered defect signals by sensors in existing technologies and significantly reduces the missed detection rate of critical defects.

[0042] 2. This invention maps multimodal features to quantum space through qubit encoding, uses entanglement entropy to accurately quantify cross-modal nonlinear correlation weights, and combines adaptive enhancement coefficients with a dual-constraint fusion weight model to achieve nonlinear deep aggregation of multimodal features. This not only unifies data distribution but also highlights key features related to defects, strengthens deep coupling of cross-modal features, improves the effectiveness of defect features, and avoids the feature dilution problem caused by traditional linear fusion.

[0043] 3. This invention constructs a multi-dimensional search space and, based on the fusion optimization of chaotic mapping and quantum particle swarm optimization, combines local constraint terms constructed from superconducting critical parameters with a dynamic safety threshold coupled with the low-temperature defect propagation rate to achieve precise screening of the initial particle swarm. This effectively eliminates false defects with strong signals but no impact on superconducting performance, optimizes the screening accuracy of the defect candidate set, reduces interference from false defects, ensures that the defect candidate set focuses on high-hazard defects, and improves the reliability of superconducting cable detection results.

[0044] 4. This invention enhances the characteristics of key defect areas, corrects low-temperature signal attenuation, and performs environmentally adaptive standardization processing, enabling the characteristic values ​​to remain stable even in the extreme operating environment of superconducting cables under low temperature and strong electromagnetic fields. This improves the stability and environmental adaptability of defect characteristics, making it suitable for more operating scenarios of superconducting cables. It solves the problem of poor characteristic stability in extreme environments of traditional detection methods and ensures the consistency of detection results under different operating conditions.

[0045] 5. This invention utilizes the unique signal patterns of superconducting defects, pre-set multi-scale Gaussian difference kernels with layered adaptation, and combines gradient information with the significance value calculation of superconducting characteristics to establish specific criteria and cross-modal verification mechanisms. This allows for the precise differentiation of different types of defects, such as superconducting layer cracks, insulation layer corrosion, and shielding layer damage, enabling accurate classification of superconducting cable defects. This provides precise defect type and location information for layered maintenance and risk assessment of superconducting cables, reducing operation and maintenance costs. Attached Figure Description

[0046] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings: Figure 1 This is a flowchart of the detection method of the present invention; Figure 2 This is a schematic diagram of the structure of the superconducting cable of the present invention. Detailed Implementation

[0047] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. This will allow for a full understanding of how the present application uses technical means to solve technical problems and achieve technical effects, and to facilitate its implementation.

[0048] Due to the layered structure, critical parameters, and low-temperature, high-electromagnetic operating environment of superconducting cables, it is difficult to precisely adapt the detection methods to the specific characteristics of each layer when detecting defects. This results in low detection accuracy, affecting the actual operating performance of the superconducting cable and shortening its service life. To improve the accuracy of superconducting cable defect detection and reduce the missed detection rate of critical defects, this invention proposes a defect detection method for superconducting cables. This method employs an improved particle swarm optimization algorithm to achieve accurate identification of defects in each layer of the superconducting cable, solving the problems of increased missed detection rate and false defect misjudgment rate in existing technologies for superconducting cable defect detection. This method meets the detection requirements for high-reliability operation of superconducting cables. Figure 1 As shown, it includes the following steps: S1. A multimodal sensor array is arranged in a spiral pattern along the circumferential and axial directions of the superconducting cable. The multimodal sensor array consists of a quantum interference device, a low-temperature infrared sensor, and an electromagnetic induction sensor to accurately capture the defect signals of each layer, avoiding the missed detection of key layer signals caused by the traditional fixed pitch arrangement. Multimodal data of each layer of the superconducting cable, including the superconducting layer, the insulation layer, and the shielding layer, are collected and preprocessed. The multimodal data includes electromagnetic data, temperature field data, and structural vibration data. Then, feature extraction is performed on the multimodal data to obtain multimodal features containing electromagnetic features, temperature field features, and structural vibration features.

[0049] S2. Encode the multimodal features using qubits, calculate the cross-modal correlation weights between each multimodal feature using entanglement entropy. These cross-modal correlation weights are used to accurately characterize the nonlinear correlation strength between different modal features in superconducting cable defect detection, and are deeply adapted to the characteristics of the superconducting scenario. An adaptive enhancement coefficient is introduced to optimize the contrast of multimodal features to unify data distribution and eliminate interference from different modal data. Then, a dual-constraint fusion weight model is constructed to calculate the weighted feature aggregation value, realizing the nonlinear deep aggregation of multimodal features. This outputs a weighted feature aggregation value with cross-modal correlation, data consistency, and superconducting defect specificity. The specific steps include the following: S21. Map the electromagnetic features, temperature field features, and structural vibration features to a preset quantum Hilbert space, convert them into qubit codes, and establish the correspondence between multimodal feature values ​​and qubit states. S22. Based on entanglement entropy, quantify the quantum correlation strength between cross-modal features, calculate the cross-modal correlation weight of each feature and the entanglement entropy of the quantum states of each feature, and use the following formulas:

[0050]

[0051] In the above formula, Indicates the first The first feature and the second Cross-modal association weights of each feature, Indicates the first The first feature and the second Entanglement entropy of a characteristic quantum state Indicates the first The first feature and the second The reduced density matrix of each characteristic quantum state, Represents the matrix trace operation. This represents the maximum value of the entanglement entropy for each feature; S23. Based on the preset critical current and critical temperature of the superconducting cable, calculate the adaptive enhancement coefficient used to optimize the contrast of multimodal features. The calculation formula is as follows:

[0052] In the above formula, Indicates the first The adaptive enhancement coefficients of each feature Indicates the first The rate of change of superconducting layer resistance corresponding to each characteristic Indicates the first The dielectric constant deviation of the insulating layer corresponding to each characteristic This indicates the critical current of the superconducting cable. This indicates the critical temperature of the superconducting cable. This indicates the real-time operating temperature of the superconducting cable. Indicates the first The standard deviation of each feature This represents the average of the standard deviations of each characteristic. Represents the natural exponential function; S24. The enhanced multimodal features are concatenated into tensors according to each layer to form a high-dimensional feature tensor. Based on multi-scale median filtering and the superconducting property normalization function, the normalized eigenvalues ​​of each feature are calculated to eliminate the dimensional differences of the multimodal features, thereby unifying the data distribution. The calculation formula is as follows:

[0053] In the above formula, Indicates the first Normalized eigenvalues ​​of each feature This indicates a multi-scale median filtering operation. This indicates a tensor splicing operation. Indicates the first Electromagnetic data corresponding to each feature Indicates the first Temperature field data corresponding to each feature Indicates the first Structural vibration data corresponding to each feature This represents the minimum value of each feature after enhancement. This represents the maximum value of the enhanced feature; S25. Based on the signal-to-noise ratio and critical parameters of the superconducting cable, a dual-constraint model is constructed to calculate the fusion weights of each feature. The calculation formula is as follows:

[0054]

[0055] In the above formula, Indicates the first The fusion weights of each feature Indicates the first The signal-to-noise ratio of each feature Indicates the first The real-time operating current of the superconducting cable corresponding to each characteristic Indicates the first The power of the effective signal in each feature, Indicates the first The power of noise in each feature, Represents the logarithmic function with base 10; S26. Calculate the weighted feature aggregation value based on the cross-modal correlation weight and the fusion weight of the dual constraints. The calculation formula is as follows:

[0056] In the above formula, Indicates the first The weighted feature aggregation value of each feature. Indicates the first The fusion weights of each feature Indicates the first Normalized eigenvalues ​​of each feature.

[0057] S3. Construct an initial individual vector based on weighted feature aggregation values. Combine this with the various layers of the superconducting cable to build a multi-dimensional search space. This multi-dimensional search space is a quantization space adapted to the layered structure and defect detection requirements of the superconducting cable. Specifically, it represents a high-dimensional vector space composed of defect feature parameters of the superconducting layer, insulation layer, and shielding layer of the superconducting cable. Each dimension corresponds to a quantifiable defect characteristic. Each point in the space uniquely maps to a complete set of defect state combinations. Based on chaotic mapping, a chaotic individual vector covering the multi-dimensional search space is generated to supplement the diversity of the initial individual vector and avoid the search getting stuck in local optima and becoming limited. The chaotic individual vector and the initial individual vector are merged into an initial particle swarm, which is then transformed into a quantum particle swarm. This ensures that the defect search process not only conforms to the layered defect distribution law of the superconducting cable but also accurately captures individuals related to high-hazard defects. The specific steps include the following: S31. Based on the weighted feature aggregation value as the initial individual basis, each weighted feature aggregation value is mapped to the corresponding layer's defect feature dimension to calculate the initial individual vector. The calculation formula is as follows:

[0058] In the above formula, Indicates the first The initial individual vector corresponding to each feature Indicates the first The weighted feature aggregation value of each feature is mapped to the defect feature components of the superconducting layer. Indicates the first The weighted feature aggregation value of each feature is mapped to the defect feature components of the insulation layer. Indicates the first The weighted feature aggregation value of each feature is mapped to the defect feature components of the shielding layer. This represents the vector transpose operation; S32. Based on the physical characteristics and defect types of each layer of the superconducting cable, independent search boundaries are set for the defect feature components of each layer, and a multi-dimensional search space with layer dimension and defect parameter dimension is constructed, as shown in the following expression:

[0059] In the above formula, Indicates a hierarchical index. Indicates the first The lower boundary of the search space for each layered defect feature component Indicates the first The upper boundary of the search space for each hierarchical defect feature component. Indicates the first The lower boundary correction coefficient for each layer, Indicates the first The upper boundary correction coefficient of each layer, Indicates the first The minimum detectable defect parameter value for each layer. Indicates the first The maximum detectable defect parameter value for each layer; S33. Based on chaotic mapping, the superconducting critical temperature is incorporated into the chaotic parameters, and the chaotic individual vectors covering each feature of the multi-dimensional search space are calculated. The calculation formula is as follows:

[0060]

[0061] In the above formula, Indicates the first A chaotic individual vector, Indicates the first Output value of the next chaotic iteration Indicates the first Output value of the next chaotic iteration Indicates the first One chaotic parameter; S34. Based on the three-layer structure of the superconducting cable (superconducting layer, insulation layer, and shielding layer) and the multi-dimensional search space boundary constraints, the process proceeds sequentially: structural verification, effectiveness screening, proportional fusion, deduplication optimization, superconducting characteristic adaptation, and quantum initialization. Initial individuals and chaotic individuals are merged into an initial particle swarm. Then, constrained by the layered structure, critical characteristics, and defect signal modes of the superconducting cable, the initial particle swarm is transformed into a quantum particle swarm adapted for superconducting cable defect detection through a progressive logic of layered quantum characterization, superconducting parameter coupling calibration, quantum state dynamic adjustment, and particle swarm attribute integration. The quantum state probability amplitude vector and the probability density of each layer dimension are calculated using the following formulas:

[0062]

[0063] In the above formula, Indicates the first The quantum state probability magnitude vector of each particle. Indicates the first The particle in the first Quantum probability amplitude in each hierarchical dimension Indicates the first The particle in the first The probability density of each hierarchical dimension Indicates the first The particle in the first Feature values ​​of each hierarchical dimension.

[0064] S4. Based on the quantum particle swarm optimization, local optimization constraints are constructed for the critical parameters of the superconducting cable to calculate the fitness value of the particles. A safety threshold coupled with the defect propagation rate in the low-temperature environment and used to screen fitness values ​​is constructed and calculated. The optimal individual vectors of each layer are selected to generate a defect candidate set. This can achieve accurate screening of the initial particle swarm, optimize the screening accuracy of the defect candidate set, and improve the reliability of the superconducting cable detection results. The specific steps include the following: S41. Based on the global search principle of quantum particle swarm optimization, a local optimization constraint term is constructed by introducing critical parameters of the superconducting cable, including critical current and critical temperature. The correlation between defect signal intensity and superconducting performance attenuation is quantified to calculate the fitness value of the particles. The calculation formula is as follows:

[0065]

[0066]

[0067] In the above formula, Indicates the first The fitness value of each particle. Indicates the defect signal strength weight. This indicates the correlation weight between the degradation of superconducting performance and its degree of degradation. Indicates the first The defect signal intensity corresponding to each particle Indicates the first The correlation of superconducting performance degradation for each particle Indicates the first The defect signal amplitude corresponding to each particle This represents the amplitude of the background signal when there are no defects. This represents the signal amplitude corresponding to the largest known defect. This indicates the lowest temperature at which the superconducting cable can operate stably. This indicates the critical current of the superconducting cable. Indicates the first The real-time operating current of the superconducting cable corresponding to each characteristic This indicates the critical temperature of the superconducting cable. This indicates the real-time operating temperature of the superconducting cable; S42. Based on the ratio of the current iteration count to the maximum iteration count, and using a dynamic adjustment mechanism for the iteration process, the iteration stage is first clearly defined. Then, the discovery ratio, warning ratio, and elite ratio are dynamically adjusted in a differentiated manner, while incorporating superconducting critical parameter correction. This ensures that the ratio adjustment adapts to the optimization algorithm process. A safety threshold coupled with the defect propagation rate under low-temperature conditions is constructed and calculated, and effective particle individuals with fitness values ​​greater than the safety threshold are retained. The formula for calculating the safety threshold is as follows:

[0068]

[0069] In the above formula, Indicates the first The rate of low-temperature defect propagation. This represents the defect propagation rate constant. The activation energy represents the defect propagation. Represents the Boltzmann constant. Indicates the first Real-time operating temperature of the superconducting cable. Indicates the first The safety threshold for each step. Indicates the basic security threshold. Indicates the maximum number of iterations. Indicates the current iteration number; S43. Determine whether the current iteration count has reached the preset maximum iteration count; if yes, then select all valid particle individuals in the iterations as the optimal individual vectors for each layer and proceed to the next step; if no, return to step S41.

[0070] S44. The optimal individual vectors selected from each layer are spliced ​​together as tensors, and the defect information of each layer is integrated to form a defect candidate set that includes the defect parameters of the entire layer and the impact assessment of superconducting performance.

[0071] S5. The defect candidate set is mapped to a 3D defect image containing information about each layer of the superconducting layer, insulating layer, and shielding layer. Key defect region features are enhanced to obtain an enhanced feature layer. This enhanced feature layer is mapped to RGB pixel values, and low-temperature corrected neighborhood feature values ​​are extracted. Standardized feature values ​​are calculated and integrated into a feature standardization matrix, improving the stability and environmental adaptability of the defect features and making them suitable for more operating scenarios of superconducting cables. Specifically, the steps include: S51. Map the defect features of each layer of the optimal individual vector in the defect candidate set to independent cross-modal channels. Based on the influence weights of each layer of the superconducting cable, weighted superposition of the multimodal channel features of the same layer is performed to calculate the layered multimodal fusion features corresponding to the optimal individual vector of each layer. The calculation formula is as follows:

[0072]

[0073] In the above formula, Indicates the first The optimal individual vector at the th th th A layered multimodal fusion feature Indicates the first The performance impact weight of each layer Indicates the first The optimal individual vector at the th th th The first layer Feature vectors of each modal channel Indicates the first The optimal individual vector at the th th th Each layer of defect feature value, Indicates the first The modal channel in the first... Each layer has its own adaptation coefficient; S52. By expanding the multimodal channel dimension, the three-layer fused features are stitched together into a three-dimensional defect image tensor. Gaussian pyramid and Laplacian pyramid decompositions are then performed on each layer of the three-dimensional defect image tensor. Based on the Gaussian pyramid, and with the Laplacian pyramid calculated through coupled decomposition generated by the difference between adjacent layers of the Gaussian pyramid, the Laplacian pyramid layers are calculated using the following formula:

[0074] In the above formula, Indicates the first The defective image, the first Each layer, scale The Laplace pyramid layers, Indicates the first The defective image, the first Each layer, scale Gaussian pyramid layers; Furthermore, the top few scale features with the highest attention weights are retained, and the features of key defect regions are enhanced to obtain an enhanced feature layer. The formula for calculating the enhanced feature layer is as follows:

[0075] In the above formula, Indicates the first The defect image is in the first A layered enhancement feature layer, Indicates the first Each layer in scale Attention weights under, Indicates the first The optimal defect scale for each layer; S53. Map the enhanced feature layer to RGB pixel values, incorporate the critical parameters and real-time operating parameters of the superconducting cable, and calculate the RGB pixel values ​​that reflect the intensity of the defect features and the associated superconducting performance state to transform the enhanced feature layer into an enhanced defect image. The calculation formula is as follows:

[0076] In the above formula, Indicates the first The defect image is in the first Each layer of RGB pixel values, Indicates the first The defect image is in the first The maximum value of each hierarchical enhancement feature layer Indicates the first The real-time dielectric constant of the insulating layer corresponding to each defect image. Indicates the standard dielectric constant of the insulating layer. This indicates the critical current of the superconducting cable. Indicates the first The real-time operating current of the superconducting cable corresponding to each characteristic This indicates the critical temperature of the superconducting cable. This indicates the real-time operating temperature of the superconducting cable; S54. Select the target pixel in the enhanced defect image, extract its neighboring pixels, and introduce an attenuation correction factor to correct for low-temperature attenuation. Calculate the neighborhood feature value after low-temperature correction using the following formula:

[0077] In the above formula, Indicates the first The defect image is in the first Each layer, target pixel The neighborhood feature values, Represents target pixel The set of neighboring pixels, Represents neighboring pixels grayscale value, Represents target pixel grayscale value, Indicates the normal temperature reference temperature; S55. Based on the statistical distribution of neighborhood feature values, combined with real-time temperature and electromagnetic interference intensity, calculate the standardized feature values ​​of all pixels in the defect image and integrate them into a feature standardization matrix. The calculation formula is as follows:

[0078] In the above formula, Indicates the first The defect image is in the first Each layer, target pixel The characteristic standardized value, Indicates the first The average of all neighborhood feature values ​​in each layer Indicates the first The standard deviation of all neighborhood feature values ​​in each layer Indicates the real-time electromagnetic interference intensity. Indicates the electromagnetic interference correction factor. This indicates the interval translation adjustment term.

[0079] S6. Based on the unique signal patterns of superconducting defects, including electromagnetic signal abrupt changes corresponding to superconducting layer cracks, dielectric constant anomalies corresponding to insulation layer corrosion, and structural vibration signal distortions corresponding to shielding layer damage, a multi-scale Gaussian difference kernel with layered adaptation is preset. The scale feature values ​​of each layer are extracted, and the gradient amplitude and direction values, as well as the defect significance value, are calculated. Then, a superconducting defect specificity criterion is established. Combined with cross-modal verification results, the enhanced defect images are classified to generate a defect set, accurately distinguishing different types of defects such as superconducting layer cracks, insulation layer corrosion, and shielding layer damage, thus achieving accurate classification of superconducting cable defects. Specifically, the steps include the following: S61. Perform convolution operation between the preset multi-scale Gaussian difference kernel and the feature normalization matrix to extract the scale feature values ​​of each layer, and calculate the scale feature value matrix. The calculation formula is as follows:

[0080] In the above formula, Indicates the first The defective image, the first Each layer, scale The scale eigenvalue matrix, Indicates the first The defective image, the first A hierarchical feature standardization matrix, This represents a two-dimensional convolution operation. Indicates the first Each layer, scale Superconducting-specific Gaussian difference kernel; S62. Based on the critical parameters and real-time operating parameters of the superconducting cable, calculate the superconducting performance attenuation weight for each defect image. The calculation formula is as follows:

[0081] In the above formula, Indicates the first The superconducting performance degradation weight of each defect image This indicates the critical current of the superconducting cable. Indicates the first The real-time operating current of the superconducting cable corresponding to each characteristic This indicates the critical temperature of the superconducting cable. This indicates the real-time operating temperature of the superconducting cable; S63. Based on the scale feature values ​​of each layer, calculate the spatial gradient magnitude reflecting the intensity of the defect edge and the directional value reflecting the direction of defect extension. The calculation formulas are as follows:

[0082]

[0083] In the above formula, Indicates the first The defective image, the first Each layer, scale The weighted gradient magnitude, Indicates the first The defective image, the first Each layer, scale The gradient direction value, Represents the arctangent function in the four quadrants. The gradient response matrix represents the horizontal or vertical direction; S64. Combining the weighted gradient magnitude, gradient direction consistency, and superconductivity sensitivity coefficient, the defect significance value is calculated using the following formula:

[0084]

[0085] In the above formula, Indicates the first The defective image, the first Each layer of defect significance value, Indicates the total number of scales. Indicates the first Each layer, scale Gradient direction consistency This represents the variance of the gradient direction values. Indicates the sensitivity coefficient for superconducting properties; S64. Based on the unique signal modes of superconducting defects, establish a superconducting defect specificity criterion and calculate the defect specificity criterion threshold. The calculation formula is as follows:

[0086] In the above formula, These represent the specific criterion thresholds for superconducting layer cracks, insulating layer corrosion, and shielding layer damage, respectively. They represent the specificity criterion coefficients, Indicates the first The electromagnetic mode signal sequence corresponding to each defect image. Indicates the first The temperature field mode signal sequence corresponding to each defect image. Indicates the first The structural vibration mode signal sequence corresponding to each defect image. This represents variance calculation. This represents entropy value operations. This represents the root mean square operation. Indicates the first The real-time dielectric constant of the insulating layer corresponding to each defect image. This represents the standard dielectric constant of the insulating layer.

[0087] S65. Construct a cross-modal verification mechanism for defect saliency and modal signals, and determine whether the defect saliency value is greater than the specificity criterion threshold of the corresponding layer and whether the modal signal pattern matches the defect type. If yes, the verification result is correct; otherwise, the defect saliency is removed.

[0088] S66. Combining the specificity criteria and cross-modal verification results, the enhanced defect images are classified, and the verified defects are assigned to the corresponding layered defect sets, including the superconducting layer crack defect set, the insulating layer corrosion defect set, and the shielding layer damage defect set.

[0089] This defect detection method collects multimodal data from each layer of a superconducting cable, performs preprocessing and feature extraction, encodes the multimodal features into qubits, calculates the cross-modal correlation weights between the multimodal features using entanglement entropy, introduces an adaptive enhancement coefficient to optimize the contrast of the multimodal features to unify the data distribution, constructs a fusion weight model to calculate the weighted feature aggregation value and generate an initial individual vector, then combines each layer of the superconducting cable to construct a multidimensional search space, generates chaotic individual vectors based on chaotic mapping, and merges them with the initial individual vectors to form an initial particle swarm, which is then transformed into a quantum particle swarm. For quantum particle swarm optimization, critical parameters of superconducting cables are introduced to construct local optimization constraints and calculate the fitness value of particles. A safety threshold coupled with the defect propagation rate in the low-temperature environment is constructed and calculated. The optimal individual vectors of each layer are selected to generate a defect candidate set, which is then mapped to a three-dimensional defect image. The features of key defect regions are enhanced to obtain an enhanced feature layer, which is then mapped to RGB pixel values. The neighborhood feature values ​​after low-temperature correction are extracted to calculate the feature normalization value, which is then integrated into a feature normalization matrix. The scale feature values ​​of each layer are extracted using a multi-scale Gaussian difference kernel. The gradient magnitude and direction of the scale feature values, as well as the defect significance value, are calculated to establish a superconducting defect specificity criterion. The defect set is generated by combining the cross-modal verification results.

[0090] The present invention also provides a superconducting cable, such as Figure 2 As shown, a defect detection method is used to detect defects such as superconducting layer cracks, insulation layer corrosion, and shielding layer damage in a superconducting cable. The superconducting cable includes a superconducting layer 2 covered with a high-voltage insulation layer 1, a high-temperature superconducting shielding layer 3 covered with the outside of the high-voltage insulation layer 1, and an outer protective layer 4 covered with the outside of the high-temperature superconducting shielding layer 3.

[0091] The superconducting layer 2 of the superconducting cable of the present invention serves as the core functional layer of the superconducting cable, used to achieve low-loss transmission of large currents. The high-voltage insulation layer 1 is used to provide reliable electrical insulation and adapt to the superconducting low-temperature environment. The high-temperature superconducting shielding layer 3 is used to shield electromagnetic interference and divert fault current. The outer protective layer 4 serves as the outermost layer of the superconducting cable, used to provide all-round physical protection and adapt to the laying environment.

[0092] Those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, this application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0093] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. Since the above embodiments are substantially similar to the method embodiments, they are described relatively simply; relevant parts can be referred to the descriptions of the method embodiments.

[0094] The above embodiments provide a detailed description of the present invention. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for detecting defects in superconducting cables, characterized in that, The method includes the following steps: S1. A multi-mode sensor array is arranged in a spiral pattern along the circumferential and axial directions of the superconducting cable to collect and preprocess the multi-mode data of each layer of the superconducting cable. Then, feature extraction is performed on the data to obtain multi-mode features including electromagnetic features, temperature field features and structural vibration features. S2. Encode the multimodal features with qubits, calculate the cross-modal correlation weights between each multimodal feature through entanglement entropy, introduce an optimized multimodal feature contrast adaptive enhancement coefficient to normalize each feature, and calculate the weighted feature aggregation value. S3. Construct an initial individual vector based on the weighted feature aggregation value, combine it with each layer of the superconducting cable to construct a multi-dimensional search space, and generate a chaotic individual vector covering the multi-dimensional search space based on the chaotic mapping. Merge it with the initial individual vector to form an initial particle swarm and transform it into a quantum particle swarm. S4. Introduce critical parameters of superconducting cables to construct local optimization constraint terms to calculate the fitness value of particles, and based on the preset safety threshold, select the optimal individual vectors of each layer to generate a defect candidate set. S5. Map the defect candidate set to a three-dimensional defect image, enhance the key defect region features to transform it into an enhanced defect image, and introduce an attenuation correction factor to calculate the feature normalization value of each pixel and integrate it into a feature normalization matrix. S6. Calculate the scale eigenvalues, spatial gradient amplitudes, directional values, and defect significance values ​​of each layer of the superconducting cable based on the feature normalization matrix, and establish a superconducting defect specific criterion based on the unique signal pattern of superconducting defects, classifying the enhanced defect images into several defect sets.

2. The detection method according to claim 1, characterized in that, The superconducting cable comprises a superconducting layer, an insulating layer, and a shielding layer, and the multimodal data includes electromagnetic data, temperature field data, and structural vibration data.

3. The detection method according to claim 1, characterized in that, Step S2 specifically includes the following steps: S21. Map the electromagnetic features, temperature field features, and structural vibration features to a preset quantum Hilbert space, convert them into qubit codes, and establish the correspondence between multimodal feature values ​​and qubit states. S22. Based on entanglement entropy quantization, the quantum correlation strength of any two features in a multimodal feature is calculated, and the cross-modal correlation weight of each feature and the entanglement entropy of the quantum state of each feature are calculated. The calculation formula is as follows: In the above formula, Indicates the first The first feature and the second Cross-modal association weights of features, Indicates the first The first feature and the second The entanglement entropy of a characteristic quantum state Indicates the first The first feature and the second The reduced density matrix of each characteristic quantum state, Represents the matrix trace operation. This represents the maximum value of the entanglement entropy for each feature; S23. Based on the preset critical current and critical temperature of the superconducting cable, calculate the adaptive enhancement coefficient used to optimize the contrast of multimodal features. The calculation formula is as follows: In the above formula, Indicates the first The adaptive enhancement coefficients of each feature Indicates the first The rate of change of superconducting layer resistance corresponding to each characteristic Indicates the first The dielectric constant deviation of the insulating layer corresponding to each characteristic This indicates the critical current of the superconducting cable. This indicates the critical temperature of the superconducting cable. This indicates the real-time operating temperature of the superconducting cable. Indicates the first The standard deviation of each feature This represents the average of the standard deviations of each characteristic. Represents the natural exponential function; S24. The enhanced multimodal features are concatenated into tensors for each layer to form a high-dimensional feature tensor. Based on multi-scale median filtering and the superconducting property normalization function, the normalized eigenvalues ​​of each feature used to eliminate dimensional differences in the multimodal features are calculated. The calculation formula is as follows: In the above formula, Indicates the first Normalized eigenvalues ​​of each feature This indicates a multi-scale median filtering operation. This indicates a tensor splicing operation. Indicates the first Electromagnetic data corresponding to each feature Indicates the first Temperature field data corresponding to each feature Indicates the first Structural vibration data corresponding to each feature This represents the minimum value of each feature after enhancement. This represents the maximum value of the enhanced feature; S25. Based on the signal-to-noise ratio and critical parameters of the superconducting cable, a dual-constraint model is constructed to calculate the fusion weights of each feature. The calculation formula is as follows: In the above formula, Indicates the first The fusion weights of each feature Indicates the first The signal-to-noise ratio of each feature Indicates the first The real-time operating current of the superconducting cable corresponding to each characteristic Indicates the first The power of the effective signal in each feature, Indicates the first The power of noise in each feature, Represents the logarithmic function with base 10; S26. Calculate the weighted feature aggregation value based on the cross-modal correlation weight and the fusion weight of the dual constraints. The calculation formula is as follows: In the above formula, Indicates the first The weighted feature aggregation value of each feature. Indicates the first The fusion weights of each feature Indicates the first Normalized eigenvalues ​​of each feature.

4. The detection method according to claim 1, characterized in that, Step S3 specifically includes the following steps: S31. Based on the weighted feature aggregation value as the initial individual basis, each weighted feature aggregation value is mapped to the defect feature dimension of the corresponding layer of the superconducting cable to calculate the initial individual vector. The calculation formula is as follows: In the above formula, Indicates the first The initial individual vectors corresponding to each feature Indicates the first The weighted feature aggregation value of each feature is mapped to the defect feature components of the superconducting layer. Indicates the first The weighted feature aggregation value of each feature is mapped to the defect feature components of the insulation layer. Indicates the first The weighted feature aggregation value of each feature is mapped to the defect feature components of the shielding layer. This represents the vector transpose operation; S32. Based on the physical characteristics and defect types of each layer of the superconducting cable, independent search boundaries are set for the defect feature components of each layer, and a multi-dimensional search space with layer dimension and defect parameter dimension is constructed, as shown in the following expression: In the above formula, Indicates a hierarchical index. Indicates the first The lower boundary of the search space for each layered defect feature component Indicates the first The upper boundary of the search space for each hierarchical defect feature component. Indicates the first The lower boundary correction coefficient for each layer, Indicates the first The upper boundary correction coefficient of each layer, Indicates the first The minimum detectable defect parameter value for each layer. Indicates the first The maximum detectable defect parameter value for each layer; S33. Based on chaotic mapping, the superconducting critical temperature is incorporated into the chaotic parameters, and the chaotic individual vectors covering each feature of the multi-dimensional search space are calculated. The calculation formula is as follows: In the above formula, Indicates the first A chaotic individual vector, Indicates the first Output value of the next chaotic iteration Indicates the first Output value of the next chaotic iteration Indicates the first One chaotic parameter; S34. Merge the initial individuals and chaotic individuals into an initial particle swarm, and transform it into a quantum particle swarm adapted for superconducting cable defect detection. Calculate the quantum state probability amplitude vector and the probability density of each dimension. The calculation formula is as follows: In the above formula, Indicates the first The quantum state probability magnitude vector of each particle. Indicates the first The particle in the first Quantum probability amplitude in each hierarchical dimension Indicates the first The particle in the first The probability density of each hierarchical dimension Indicates the first The particle in the first Feature values ​​of each hierarchical dimension.

5. The detection method according to claim 1, characterized in that, Step S4 specifically includes the following steps: S41. Based on the global search principle of quantum particle swarm optimization, a local optimization constraint term is constructed by introducing critical parameters of the superconducting cable, including critical current and critical temperature. The correlation between defect signal intensity and superconducting performance attenuation is quantified to calculate the fitness value of the particles. The calculation formula is as follows: In the above formula, Indicates the first The fitness value of each particle. Indicates the defect signal strength weight. This indicates the correlation weight between the degradation of superconducting performance and its degree of degradation. Indicates the first The defect signal intensity corresponding to each particle Indicates the first The correlation of superconducting performance degradation for each particle Indicates the first The defect signal amplitude corresponding to each particle This represents the amplitude of the background signal when there are no defects. This represents the signal amplitude corresponding to the largest known defect. This indicates the lowest temperature at which the superconducting cable can operate stably. This indicates the critical current of the superconducting cable. Indicates the first The real-time operating current of the superconducting cable corresponding to each characteristic This indicates the critical temperature of the superconducting cable. This indicates the real-time operating temperature of the superconducting cable; S42. Based on the dynamic adjustment ratio mechanism between the current iteration number and the maximum iteration number, a safety threshold coupled with the defect propagation rate under low temperature conditions is constructed and calculated, and effective particle individuals with fitness values ​​greater than the safety threshold are retained. The calculation formula for the safety threshold is as follows: In the above formula, Indicates the first The rate of low-temperature defect propagation. This represents the defect propagation rate constant. The activation energy represents the defect propagation. Represents Boltzmann's constant. Indicates the first Real-time operating temperature of the superconducting cable. Indicates the first The safety threshold for each step. Indicates the basic security threshold. Indicates the maximum number of iterations. Indicates the current iteration number; S43. Determine whether the current iteration count has reached the preset maximum iteration count; If so, then select all valid particle individuals from the iterations as the optimal individual vector for each layer, and proceed to the next step; If not, return to step S41; S44. The optimal individual vectors selected from each layer are spliced ​​together as tensors, and the defect information of each layer is integrated to form a defect candidate set that includes the defect parameters of the entire layer and the impact assessment of superconducting performance.

6. The detection method according to claim 1, characterized in that, The three-dimensional defect image contains information about each layer of the superconducting layer, insulating layer, and shielding layer. Step S5 specifically includes the following steps: S51. Map the defect features of each layer of the optimal individual vector in the defect candidate set to independent cross-modal channels. Based on the influence weights of each layer of the superconducting cable, weighted superposition of the multimodal channel features of the same layer is performed to calculate the layered multimodal fusion features corresponding to the optimal individual vector of each layer. The calculation formula is as follows: In the above formula, Indicates the first The optimal individual vector at the th th th A layered multimodal fusion feature Indicates the first The performance impact weight of each layer Indicates the first The optimal individual vector at the th th th The first layer Feature vectors of each modal channel Indicates the first The optimal individual vector at the th th th Each layer of defect feature value, Indicates the first The modal channel in the first... Each layer has its own adaptation coefficient; S52. By expanding the multimodal channel dimension, the three-layer fused features are stitched together into a three-dimensional defect image tensor. Gaussian pyramid and Laplacian pyramid decomposition is performed on each layer of the three-dimensional defect image tensor. The top few scale features with the highest attention weight are retained, and the key defect region features are enhanced to obtain the enhanced feature layer. S53. Map the enhanced feature layer to RGB pixel values, incorporate the critical parameters and real-time operating parameters of the superconducting cable, and calculate the RGB pixel values ​​that reflect the intensity of the defect features and the associated superconducting performance state to transform the enhanced feature layer into an enhanced defect image. The calculation formula is as follows: In the above formula, Indicates the first The defect image is in the first Each layer of RGB pixel values, Indicates the first The defect image is in the first A layered enhancement feature layer, Indicates the first The defect image is in the first The maximum value of each hierarchical enhancement feature layer Indicates the first The real-time dielectric constant of the insulating layer corresponding to each defect image. Indicates the standard dielectric constant of the insulating layer. This indicates the critical current of the superconducting cable. Indicates the first The real-time operating current of the superconducting cable corresponding to each characteristic This indicates the critical temperature of the superconducting cable. This indicates the real-time operating temperature of the superconducting cable; S54. Select the target pixel in the enhanced defect image, extract its neighboring pixels, and introduce an attenuation correction factor to correct for low-temperature attenuation. Calculate the neighborhood feature value after low-temperature correction using the following formula: In the above formula, Indicates the first The defect image is in the first Each layer, target pixel The neighborhood feature values, Represents target pixel The set of neighboring pixels, Represents neighboring pixels grayscale value, Represents target pixel grayscale value, Indicates the normal temperature reference temperature; S55. Based on the statistical distribution of neighborhood feature values, combined with real-time temperature and electromagnetic interference intensity, calculate the standardized feature values ​​of all pixels in the defect image and integrate them into a feature standardization matrix. The calculation formula is as follows: In the above formula, Indicates the first The defect image is in the first Each layer, target pixel The characteristic standardized value, Indicates the first The average of all neighborhood feature values ​​in each layer Indicates the first The standard deviation of all neighborhood feature values ​​in each layer Indicates the real-time electromagnetic interference intensity. Indicates the electromagnetic interference correction factor. This indicates the interval translation adjustment term.

7. The detection method according to claim 1, characterized in that, In step S6, the unique signal modes of the superconducting defects include electromagnetic signal abrupt changes corresponding to superconducting layer cracks, dielectric constant abnormalities corresponding to insulation layer corrosion, and structural vibration signal distortion corresponding to shielding layer damage.

8. The detection method according to claim 1, characterized in that, Step S6 specifically includes the following steps: S61. Perform convolution operation between the preset multi-scale Gaussian difference kernel and the feature normalization matrix to extract the scale feature values ​​of each layer, and calculate the scale feature value matrix. The calculation formula is as follows: In the above formula, Indicates the first The defective image, the first Each layer, scale The scale eigenvalue matrix, Indicates the first The defective image, the first A hierarchical feature standardization matrix, This represents a two-dimensional convolution operation. Indicates the first Each layer, scale Superconducting-specific Gaussian difference kernel; S62. Based on the critical parameters and real-time operating parameters of the superconducting cable, calculate the superconducting performance attenuation weight for each defect image. The calculation formula is as follows: In the above formula, Indicates the first The superconducting performance degradation weight of each defect image This indicates the critical current of the superconducting cable. Indicates the first The real-time operating current of the superconducting cable corresponding to each characteristic This indicates the critical temperature of the superconducting cable. This indicates the real-time operating temperature of the superconducting cable; S63. Based on the scale feature values ​​of each layer, calculate the spatial gradient magnitude reflecting the intensity of the defect edge and the directional value reflecting the direction of defect extension. The calculation formulas are as follows: In the above formula, Indicates the first The defective image, the first Each layer, scale The weighted gradient magnitude, Indicates the first The defective image, the first Each layer, scale The gradient direction value, Represents the arctangent function in the four quadrants. The gradient response matrix represents the horizontal or vertical direction; S64. Combining the weighted gradient magnitude, gradient direction consistency, and superconductivity sensitivity coefficient, the defect significance value is calculated using the following formula: In the above formula, Indicates the first The defective image, the first Each layer of defect significance value, Indicates the total number of scales. Indicates the first Each layer, scale Gradient direction consistency This represents the variance of the gradient direction values. Indicates the sensitivity coefficient for superconducting properties; S64. Based on the unique signal patterns of superconducting defects, establish a superconducting defect specificity criterion and calculate the defect specificity criterion threshold. S65. Construct a cross-modal verification mechanism for defect saliency and modal signals to determine whether the defect saliency value is greater than the specificity criterion threshold of the corresponding layer and whether the modal signal pattern matches the defect type. If so, then the verification result is correct; If not, then the significance of the defect is eliminated; S66. Combining the specificity criteria and cross-modal verification results, the enhanced defect images are classified, and the verified defects are assigned to the corresponding layer defect sets. The defect sets include the superconducting layer crack defect set, the insulating layer corrosion defect set, and the shielding layer damage defect set.

9. The detection method according to claim 8, characterized in that, The formula for calculating the defect specificity criterion threshold is as follows: In the above formula, These represent the specific criterion thresholds for superconducting layer cracks, insulating layer corrosion, and shielding layer damage, respectively. They represent the specificity criterion coefficients, Indicates the first The electromagnetic mode signal sequence corresponding to each defect image. Indicates the first The temperature field mode signal sequence corresponding to each defect image. Indicates the first The structural vibration mode signal sequence corresponding to each defect image. This represents variance calculation. This represents entropy value operations. This represents the root mean square operation. Indicates the first The real-time dielectric constant of the insulating layer corresponding to each defect image. This represents the standard dielectric constant of the insulating layer.

10. A superconducting cable, characterized in that, The defect is detected using the defect detection method according to any one of claims 1-9; The superconducting cable includes a superconducting layer (2) covered with a high-voltage insulation layer (1), a high-temperature superconducting shielding layer (3) covered with the outside of the high-voltage insulation layer (1), and an outer protective layer (4) covered with the outside of the high-temperature superconducting shielding layer (3).

Images