Method and system for evaluating comprehensive aging of submarine cable insulation material based on time scaling equivalent axis

Abstract

The application discloses a kind of based on time scaling equivalent shaft's submarine cable insulating material comprehensive aging evaluation method and system, and relates to submarine cable insulating material aging evaluation technical field.The existing method cannot realize the problem of acceleration and natural aging data equivalent alignment, multi-index evaluation contradiction, long evaluation period is solved.This application includes the following steps: obtaining multiple preset performance parameters of submarine cable insulating material after electric-thermal-power multi-physical field accelerated aging test, synthesizing comprehensive aging index after homogenization mapping and robust standardization processing, establishing the equivalent mapping of acceleration and natural service time by linear scaling coefficient, outputting life evaluation result or reliability warning information after consistency test.This technical scheme can accurately and efficiently evaluate the aging state and service life of submarine cable insulating material, and provide support for material selection, life evaluation and safety decision of submarine cable engineering.

Description

Technical Field

[0001] This invention relates to the field of aging evaluation technology for submarine cable insulation materials, and in particular to a comprehensive aging evaluation method and system for submarine cable insulation materials based on a time-scaling equivalent axis. Background Technology

[0002] Polyethylene and other submarine cable insulation materials, while widely used in engineering practice due to their excellent electrical insulation and molding stability, inevitably experience the coupled effects of electric fields, temperature fields, and mechanical loads during long-term service. Therefore, a systematic study of artificially accelerated aging under electro-thermal-mechanical multi-physics conditions is needed to establish an equivalent calibration and evaluation system that can be aligned with the natural service process. This is of great significance and application value for revealing the aging mechanism of polyethylene insulation, improving its environmental adaptability, and extending its service life. This method integrates multi-source data using a unified time benchmark and degradation indicators. By constructing equivalent electro-thermal-mechanical doses and performing statistical alignment between reference and experiment, it generates stress program suggestions for experimental optimization, while simultaneously outputting the remaining life and risk range. This method not only provides a standardized tool for the life assessment and reliability design of polyethylene insulation materials but also lays the methodological foundation for cross-platform comparison and decision support in submarine cable engineering under multiple operating conditions and batches.

[0003] Current research on aging assessment of cable insulation materials mainly focuses on three types of methods: First, methods based on multi-index comprehensive evaluation. For example, CN114966338A proposes using the TOPSIS model to comprehensively score cable insulation materials with different voltage-stabilizing components while simultaneously considering electrical and thermal characteristics, for screening or comparing the comprehensive insulation performance of material formulations. Second, prediction methods based on empirical / semi-empirical life equations. For example, CN112949099B constructs a life prediction equation that can simultaneously characterize electrical aging and thermal aging and their combined effects, aiming to achieve real-time life assessment of power cables without destructive testing. Third, grading and judgment methods based on uncertainty knowledge representation and environment-driven approaches. For example, CN111339679B uses fuzzy recognition technology to establish an aging level assessment model for LDPE under long-term natural environments using environmental parameters. The above-mentioned schemes are representative in terms of "multi-index comprehensive evaluation", "life equation extrapolation" and "environmental parameter-aging level mapping", but they are generally based on static evaluation or unidirectional extrapolation. They lack a generalized process for unified equivalent scaling of electro-thermal-mechanical stress and alignment of reference-experiment distribution, and do not provide stress sequence suggestions and uncertainty quantification at the method level under limited material index conditions. They are difficult to directly support reproducible experimental design and alignment evaluation across platforms and batches.

[0004] Existing problems:

[0005] 1. Insufficient characterization of multi-indicator correlation and common evolution: Existing studies mostly focus on comparison and evaluation based on a single indicator or a small number of indicators, lacking systematic modeling of the correlation structure and common evolutionary factors among key parameters. The direction and time sequence of changes in different indicators during the aging process are not consistent (there is even a non-monotonic phenomenon of first rising and then falling). If their correlation and synergistic changes are not explicitly addressed, it is easy to cause incomparability of results across batches and platforms, making it difficult to form a unified "deterioration indicator".

[0006] 2. Lack of objectivity in the weighting and importance allocation of comprehensive evaluations: Many comprehensive evaluation or grading models rely on empirical weighting or single-time statistical weighting, failing to fully consider the dynamic changes in the relative contributions of each indicator under different aging stages and stress combinations. When the weighting settings do not match the actual degradation mechanism, the comprehensive score (or grade) may deviate from the actual aging state of the material, resulting in contradictory conclusions such as "high evaluation but poor durability" or "low evaluation but still usable," affecting the accuracy of screening and decision-making.

[0007] 3. Insufficient robustness of data preprocessing to anisotropic trends and negative values: In the general evaluation process, different indicators change in opposite directions and have significant differences in dimensions. Common practices cannot avoid problems such as negative values, conflicting directions, or incomparable dimensions. There is a lack of unified processing constraints for non-monotonic indicators (such as certain crystallinity and high-temperature characteristics) to "monotonic with aging", which leads to unstable model input, high sensitivity of evaluation results to preprocessing details, and insufficient repeatability and reproducibility.

[0008] 4. Lack of data equivalence and alignment mechanisms for short-term artificially accelerated aging: To improve efficiency, there is a greater preference for short-term artificially accelerated aging data; however, existing life equations or environment-driven models are mostly based on long-term or complex operating conditions, and do not adequately consider the equivalence conversion and statistical alignment from short-term to long-term. With limited sample size, discrete stress loading, and large combination differences, model extrapolation is easily distorted by distribution shifts, making it difficult to stably reflect the actual service evolution and life trend under electro-thermal-mechanical coupling. Summary of the Invention

[0009] The technical problem to be solved and the technical task proposed by this invention is to improve and refine existing technical solutions, and to provide a comprehensive aging evaluation method for submarine cable insulation materials based on a time-scaling equivalent axis, with the aim of achieving equivalent alignment between accelerated and natural aging data and improving the accuracy and efficiency of the evaluation. To this end, this invention adopts the following technical solution.

[0010] A comprehensive aging evaluation method for submarine cable insulation materials based on time-scaling equivalent axes includes the following steps:

[0011] S1. Obtain multiple preset performance parameters of submarine cable insulation material after accelerated aging test in multiple physical fields of electro-thermal-mechanical fields, and obtain similar performance parameters of similar materials under natural service conditions.

[0012] S2. Map each performance parameter under accelerated aging and natural service in the same direction and convert it into a non-negative degradation amount that increases monotonically with the aging process.

[0013] S3. Perform robust standardization on each degradation quantity obtained in step S2 to eliminate differences in dimensions and scales, and mark outlier data points.

[0014] S4. Weight the standardized degradation amounts processed in step S3 to form a comprehensive aging index.

[0015] S5. Using the natural service time as the equivalent time axis, the accelerated aging time is mapped through a single linear scaling factor. By minimizing the difference between the combined aging index of accelerated aging and natural service on the equivalent time axis, the optimal scaling factor is solved, thereby establishing the mapping relationship between accelerated aging time and the equivalent time axis.

[0016] S6. On the equivalent time axis, conduct a consistency test on the distribution pattern and numerical differences of the comprehensive aging index of accelerated aging and natural service, and obtain the consistency test results.

[0017] S7. Based on the mapping relationship and the consistency test results, output evaluation information; wherein, if the consistency test results meet the preset conditions, output the time evaluation results; otherwise, output the reliability warning information.

[0018] This technical solution abandons the traditional method that requires a complex definition of "equivalent dose," and instead employs a time-scaling equivalent axis. By solving for a single linear scaling factor, it directly maps accelerated aging time to an equivalent axis based on natural service time, achieving conceptual intuitiveness and model simplification. Furthermore, an optimization algorithm is used to optimize this mapping relationship. This enables quantitative benchmarking between short-term test data and long-term service behavior. It solves the problems of incomparability between accelerated and natural aging data, unclear equivalence relationships, and lack of credibility verification in existing technologies, thus improving the scientific rigor, practicality, and interpretability of aging evaluation results.

[0019] As a preferred technical means: In step S2, for performance parameters that do not change monotonically during the aging process, an ordinal regression method is used for homogeneous mapping.

[0020] This technical solution addresses non-monotonic parameters, such as crystallinity, which may initially increase and then decrease. It employs order-preserving regression to process these parameters, transforming them into a sequence that monotonically increases with aging while respecting the overall trend of the original data. This effectively solves the problem that directly involving non-monotonic indicators in modeling can lead to fluctuations in the comprehensive index and affect the stability of time series analysis. It ensures the consistency of the direction of all input indicators, laying a solid foundation for the subsequent construction of a reliable comprehensive aging index and accurate time alignment.

[0021] As a preferred technical means: In step S3, a robust standardization method based on the median and interquartile range is adopted.

[0022] This technical solution uses the median and interquartile range for standardization, a robust statistical method insensitive to outliers. Compared to traditional mean-standard deviation standardization, this method effectively reduces the impact of extreme outliers on the standardization results, resulting in a more stable and representative standardization degradation. This enhances the robustness and repeatability of the subsequent index synthesis and model fitting, improving the overall evaluation method's resistance to unavoidable fluctuations in experimental data.

[0023] As a preferred technical means: In step S4, the weights used to synthesize the comprehensive aging index are determined by a data-driven method, and the weights are non-negative and sum to 1.

[0024] This technical solution abandons the practice of relying on subjective experience to allocate weights, and adopts a data-driven approach to automatically determine the contribution weight of each indicator in the comprehensive aging index. It also constrains the weights to be non-negative and summed to 1, ensuring the objectivity and interpretability of the weights. This allows the comprehensive index to more accurately reflect the overall aging state characterized by multiple indicators. The non-negative constraint avoids the misleading effect of mutual cancellation between indicators, while the sum of 1 makes the comprehensive index suitable for comparisons across batches and materials.

[0025] As a preferred technical means: in step S5, the scaling factor is solved by minimizing the mean square error between the combined aging index of accelerated aging and natural service.

[0026] This technical solution transforms the problem of finding the optimal time scaling factor into an optimization problem aimed at minimizing the mean square error. Using this optimization criterion, the scaling factor that best matches the accelerated aging index curve and the natural service index curve on the equivalent time axis can be automatically and quantitatively found. This achieves objectification and optimization of the alignment process, providing a clear mathematical basis and solution method for establishing accurate equivalence relationships.

[0027] As a preferred technical means: In step S6, the consistency check includes: comparing the median offset and quantile overlap rate of the comprehensive aging index of accelerated aging and natural service, and calculating the alignment error index.

[0028] This technical solution can assess whether the center positions of two aging trend lines are aligned by comparing the median offset; it can assess whether the discrete range and shape of the data distributions are similar by calculating the quantile overlap rate; and it can quantify the overall goodness of fit by combining alignment error indicators (such as mean square error). This method conducts a comprehensive test from the central trend and distribution range to the overall error, which can more scientifically and rigorously evaluate the acceptability of the time scaling alignment results and provide reliable quantitative criteria for the decision in step S7.

[0029] As a preferred technical means, the preset performance parameters include: elastic modulus, dielectric constant, breakdown strength, crystallinity, high-temperature characteristics, and melting temperature.

[0030] The preset performance parameters cover the mechanical, electrical, thermal, and microstructural properties of the material, and are typical and key indicators for characterizing the aging state of submarine cable insulation materials such as polyethylene. This ensures that the evaluation method can comprehensively capture the degradation information of the material under the electro-thermal-mechanical coupling field from multiple perspectives, making the final comprehensive aging index and life assessment results more comprehensive and representative, and in line with the actual concerns of engineering projects.

[0031] Another technical solution of the present invention is:

[0032] A comprehensive aging evaluation system for submarine cable insulation materials based on a time-scaling equivalent axis includes:

[0033] The data acquisition module is used to acquire multiple preset performance parameters of submarine cable insulation materials after accelerated aging tests in multiple physical fields of electricity, heat and force, and to acquire similar performance parameters of similar materials under natural service conditions.

[0034] The data preprocessing module is used to map the performance parameters under accelerated aging and natural service in the same direction, convert them into non-negative degradation quantities that monotonically increase with the aging process, and perform robust standardization on each degradation quantity to eliminate differences in dimensions and scales.

[0035] The comprehensive aging index construction module is used to weight and combine the standardized deterioration quantities into a comprehensive aging index.

[0036] The time scaling and alignment module is used to map accelerated aging time to the equivalent time axis using natural service time as the equivalent time axis. By minimizing the difference between the combined aging index of accelerated aging and natural service on the equivalent time axis, the optimal scaling factor is solved, thereby establishing the mapping relationship between accelerated aging time and the equivalent time axis.

[0037] The consistency verification module is used to verify the consistency of the distribution pattern and numerical differences of the comprehensive aging index of accelerated aging and natural service on the equivalent time horizontal axis, and obtain the consistency verification results.

[0038] The results output module is used to output evaluation information of submarine cable insulation materials based on the mapping relationship and the consistency test results. The evaluation information includes: the life assessment result output when the consistency test results meet the preset conditions, and the reliability warning information output when the consistency test results do not meet the preset conditions.

[0039] The data acquisition module in this technical solution focuses on two scenarios: "accelerated aging test in multiple physical fields of electricity, heat and force" and "natural service condition". It also acquires "similar performance parameters" in a targeted manner, which ensures the correlation and comparability of the data. Compared with data acquisition under a single operating condition, it covers the data needs of the entire scenario of artificial acceleration and actual service, avoids evaluation bias caused by one-sided data sources, and lays a comprehensive and reliable data foundation for subsequent equivalent mapping and aging analysis.

[0040] The data preprocessing module addresses the core pain points of multi-indicator evaluation through "uniform mapping + robust standardization": On the one hand, it converts performance parameters into "non-negative degradation quantities that monotonically increase with the aging process," unifying the expression logic of degradation trends and avoiding evaluation confusion caused by inconsistent changes in different indicators (such as some indicators first increasing and then decreasing); on the other hand, it eliminates differences in dimensions and scales through robust standardization, giving previously scattered and incomparable multi-dimensional parameters (such as elastic modulus, dielectric constant, etc.) a unified quantitative benchmark, providing stable and comparable input data for the subsequent construction of comprehensive indices.

[0041] The comprehensive aging index construction module weights and synthesizes multiple standardized deterioration quantities into a single comprehensive aging index, which simplifies the dimension from "multiple indicators to single index". This avoids the potential contradictions in conclusions that may occur when multiple indicators are evaluated separately (such as some indicators showing light aging and others showing heavy aging). It makes the judgment of aging degree more intuitive and consistent, which not only reduces the difficulty of evaluation, but also reflects the overall aging state of the material in a centralized manner, thereby improving the efficiency and accuracy of evaluation.

[0042] The time scaling alignment module uses natural service time as the baseline and establishes a mapping relationship between accelerated aging time and equivalent time through a linear scaling factor, solving the key industry problem of "how to equivalently replace long-term natural aging data with short-term accelerated test data". It eliminates the need to wait for a long natural service cycle, allowing the derivation of long-term material aging patterns from short-term accelerated data, significantly shortening the evaluation cycle and reducing testing costs. Simultaneously, the design of a single linear scaling factor balances the simplicity and practicality of the mapping logic, ensuring the stability and attainability of the equivalent relationship.

[0043] The consistency verification module avoids potential deviations that may occur after forced alignment of accelerated data and natural data by conducting dual verification of "distribution pattern + numerical difference". It ensures the equivalence and consistency of the two through quantitative verification, prevents evaluation errors caused by mapping distortion, significantly improves the credibility of the final evaluation results, and provides rigorous and reliable support for engineering decisions.

[0044] As a preferred technical means: the data preprocessing module includes a homogenization unit, which is used to perform homogenization mapping on performance parameters that do not change monotonically during the aging process by using ordinal regression processing; the preset performance parameters include elastic modulus, dielectric constant, breakdown strength, crystallinity, high temperature characteristics and melting temperature; the robust standardization processing adopts a method based on median and interquartile range.

[0045] The data preprocessing module incorporates a homogenization unit to handle non-monotonic indicators, enhancing the system's ability to process complex data. A robust standardization method is employed to improve the system's stability and reliability when processing data containing anomalies or fluctuations. When dealing with real and complex aging data of submarine cable insulation materials, the system exhibits stronger practicality and anti-interference capabilities.

[0046] As a preferred technical approach: In the comprehensive aging index construction module, the weights used to synthesize the comprehensive aging index are determined through a data-driven method, and the weights are non-negative and sum to 1; the time scaling alignment module solves for the scaling coefficient by minimizing the mean square error between the comprehensive aging indices of accelerated aging and natural service; the consistency verification module performs consistency verification including: comparing the median offset and quantile overlap rate of the comprehensive aging indices of accelerated aging and natural service, and calculating the alignment error index.

[0047] The median offset refers to the difference between the medians of two aging index curves, and the quantile overlap rate refers to the proportion of overlap between the quantile intervals of the two curves at the same time point. This technical solution further optimizes the core algorithm and verification logic of the system. It specifies that the index construction module uses an objective data-driven approach to determine weights, ensuring the objectivity and adaptability of the system evaluation. It clarifies that the time alignment module uses the minimization of mean square error as the optimization criterion, providing a stable and efficient parameter solution algorithm for the system. It refines the specific verification content of the consistency verification module, making the system's reliability judgment based on evidence and rich in dimensions. The combination of these technical features improves the accuracy, automation level, and decision support capability of the system's core functions (index synthesis, time alignment, and result verification), making the evaluation information output by the entire system more scientific and reliable.

[0048] Beneficial Effects: This technical solution employs a time-scaling equivalent axis. By solving for a single linear scaling factor, accelerated aging time is mapped onto an equivalent axis based on natural service time. This achieves conceptual intuitiveness and model simplification. Furthermore, optimization algorithms are used to optimize this mapping relationship, enabling quantitative benchmarking between short-term test data and long-term service behavior. This solves the problems of incomparability between accelerated and natural aging data, unclear equivalence relationships, and lack of credibility verification in existing technologies, thus improving the scientific rigor, practicality, and interpretability of aging evaluation results. Attached Figure Description

[0049] Figure 1 This is a diagram of the present invention. Detailed Implementation

[0050] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings.

[0051] Example 1:

[0052] This embodiment provides a comprehensive aging evaluation method for submarine cable insulation materials based on time-scaling equivalent axes, which includes the following steps:

[0053] S1. Obtain multiple preset performance parameters of submarine cable insulation material after accelerated aging test in multiple physical fields of electro-thermal-mechanical fields, and obtain similar performance parameters of similar materials under natural service conditions.

[0054] The preset performance parameters include: elastic modulus, dielectric constant, breakdown strength, crystallinity, high-temperature properties, and melting temperature.

[0055] Elastic modulus is one of the most important and fundamental properties of polymers, representing the strength and ability of a material to retain its deformation under tension. The specific measurement process is as follows: the test is conducted according to the international standard Cigré TB 852, and the mechanical properties are tested using a Zwick / Roell Z020 universal testing machine.

[0056] Dielectric constant testing was performed according to standards IEC 60247-2014 and ASTM D924-15, using a Novocontrol Concept 80 wideband wide-temperature dielectric spectrometer. Both sides of the sample were sputtered with gold using a VPI SD-900M ion sputtering system for 60 seconds at a current of 20 mA. The gold-sputtered sample was then placed between two copper electrodes with diameters of 20 mm and 30 mm, and the sample was fixed in the testing apparatus for measurement. The test frequency range was 1–10 Hz. 6 Hz.

[0057] Breakdown field strength is a technical indicator reflecting the DC breakdown resistance of submarine cable insulation materials. The specific measurement process is as follows: A ball-and-cylinder electrode is used to test the breakdown electric field of the submarine cable insulation sample under DC voltage. A rapidly increasing DC voltage is applied to the sample at room temperature until the sample breaks down; the voltage increase rate is 1 kV / s. The test is repeated 15 times to obtain the Weibu 11 distribution map of the DC breakdown field strength of the sample. The breakdown field strength at the 63.2% probability point is taken as the statistical breakdown result.

[0058] The thermal decomposition temperature was analyzed using an SDT-Q600 DSC-TG thermogravimetric analyzer. The sample mass was 5.0~6.0 mg. Under a protective nitrogen atmosphere, the sample was heated from room temperature (30℃) to 700℃ at a rate of 10℃ / min to obtain thermogravimetric curves under different aging time conditions, thereby obtaining the high-temperature characteristics.

[0059] The crystallinity and melting temperature properties of the sample were studied and analyzed using a TA-Q200 differential scanning calorimeter. The sample was heated from room temperature (30℃) to 200℃ at a rate of 10℃ / min under a protective argon atmosphere, and then the atmosphere was changed to oxygen and the temperature was increased to 250℃ at a rate of 10℃ / min.

[0060] S2. Map each performance parameter under accelerated aging and natural service in the same direction and convert it into a non-negative degradation amount that increases monotonically with the aging process.

[0061] The elastic modulus, dielectric constant, breakdown strength, crystallinity, high-temperature properties, and melting temperature from step S1 are set as variables that change with aging time. These indicators are mapped to a non-negative, monotonically increasing scale with degradation. The baseline value is the 25th percentile of the initial samples from the same batch. The mapping formula is:

[0062] , ;

[0063] , ;

[0064] For those that may rise first and then fall / non-monotonic , The monotonic mapping is constructed using order-preserving piecewise partitioning, and the formula is as follows:

[0065] , .

[0066] Finally, the set of non-negative degradation values ​​corresponding to all indicators is obtained. The output is:

[0067] ;

[0068] in, , , , , , : These represent the elastic modulus, dielectric constant, breakdown strength, crystallinity, high-temperature properties (such as thermal decomposition temperature or a certain characteristic temperature), and melting temperature measured at time t, respectively. , , , , , : The baseline value corresponding to the above parameters (such as the initial value or the typical value of the health status). , , , , , : These are the non-negative degradation values ​​obtained after the induction mapping of each indicator.

[0069] This indicates order-preserving regression, used to transform a non-monotonic sequence into a monotonically increasing sequence. Its input is a numerical value that changes with time series, and its output is a new sequence that is as close as possible to the original data but strictly monotonic (in this case, monotonically increasing). , To determine the original crystallinity X c A function that performs a preliminary linear scaling of H(t) and the high-temperature characteristic H(t). Let be the non-negative degradation of the k-th indicator.

[0070] S3. Perform robust standardization on each degradation quantity obtained in step S2 to eliminate differences in dimensions and scales, and mark outlier data points.

[0071] To eliminate the difference between dimensions and scale, for each Robust standardization is performed using the median and interquartile range:

[0072] ; ;

[0073] For sequence the median of For sequence The interquartile difference. For sequence The upper quartile, For sequence The lower quartile. Anomaly detection uses an improved boxplot with a threshold of [value missing]. By combining repeated measurements at the same stress level, the median absolute deviation is calculated, and points exceeding a preset threshold are marked. Outliers are only marked for tracing purposes and are not directly deleted to ensure the robustness of subsequent analyses.

[0074] S4. Weight the standardized degradation amounts processed in step S3 to form a comprehensive aging index.

[0075] A single aging intensity scale is synthesized on a uniform scale using non-negative weighting. The specific process is as follows:

[0076] , , ;

[0077] In the formula: The comprehensive aging index at time t; The weight of the k-th indicator; weight Determined through data-driven approaches, employing time discernibility / order preservation criteria, and normalized time... To fit the target, a penalty is added to the negative difference.

[0078] Under the condition of satisfying the constraints ( )and Under the premise of minimizing the following objective function, the weights are determined. :

[0079]

[0080] In the formula, For the first Each indicator at time The standardized value. Normalized time; This is a regularization parameter used to balance the weights of fitting error and monotonicity penalty; The comprehensive aging index; For the forward difference of the exponent; This function is in When the exponent decreases, the value is negative. Squaring produces a positive penalty, in When the index remains flat or rises, the value is set to 0, and no penalty is imposed.

[0081] S5. Using the natural service time as the equivalent time axis, the accelerated aging time is mapped through a single linear scaling factor. By minimizing the difference between the combined aging index of accelerated aging and natural service on the equivalent time axis, the optimal scaling factor is solved, thereby establishing the mapping relationship between accelerated aging time and the equivalent time axis.

[0082] Using the natural service time as the equivalent horizontal axis The time for accelerated aging will be proportionally increased. Stretching increases the overall aging index of both domains. Match as closely as possible on the same horizontal axis. Settings:

[0083] , , ;

[0084] The equivalent time of natural service; Accelerated aging equivalent time; s is a linear scaling factor.

[0085] Constructing interpolation functions using natural service data The objective function is minimum mean squared error alignment:

[0086] ;

[0087] Numerically, a bounded one-dimensional search is used. This can be reliably obtained.

[0088] The output includes: the optimal scaling factor. (Equivalent propulsion ratio); Alignment error MSE; Equivalent axis of acceleration domain Serving with nature .

[0089] S6. On the equivalent time axis, conduct a consistency test on the distribution pattern and numerical differences of the comprehensive aging index of accelerated aging and natural service, and obtain the consistency test results.

[0090] On the unified equivalent horizontal axis Above, the two domains are calculated separately based on shape consistency. The median and Numerical bands: Observe the degree of overlap between the quantile bands of the two regions and the median offset:

[0091] ;

[0092] The maximum / average offset is calculated; the quantile overlap rate is recorded, and the difference is quantized and reported as MSE; individual time points in the acceleration domain are deleted one by one, and the estimation is repeated. Compared with MSE, examine whether the fluctuation is acceptable, and perform statistical analysis. The proportion of negative slopes ensures that the overall synthetic index follows... It increases and rises.

[0093] S7. Based on the mapping relationship and the consistency test results, output evaluation information; wherein, if the consistency test results meet the preset conditions, output the time evaluation results; otherwise, output the reliability warning information.

[0094] This embodiment has the following characteristics:

[0095] 1. An integrated preprocessing approach combining homogenization and robust standardization is adopted: taking natural service t=0 as the health benchmark, the six indicators are mapped to non-negative quantities such that "the larger the value, the more severe the degradation", and robust standardization is performed using median / IQR to form a dimensionless feature set that is comparable across batches and domains (acc / nat).

[0096] 2. Construction of a comprehensive aging index driven by nonnegative least squares (NNLS): Under the condition that the time domains are normalized respectively, NNLS is used to adaptively learn the nonnegative weights of six indicators (sum=1) to obtain a comprehensive aging index that is monotonic over time and has a uniform scale. It also allows for the automatic application of sparse and unrecognizable indicator weights in small sample scenarios.

[0097] 3. Time scaling alignment and optimization criterion of the equivalent dose axis: In the absence of stress time series, the equivalent dose axis is simplified by minimizing... and The mean square error (MSE) is used to obtain a unique equivalent scaling factor, thereby achieving an equivalent mapping and quantitative comparability between the acceleration and reference.

[0098] 4. Alignment verification framework for quantile band consistency and error quantization: based on equivalent axis redrawing The median and quantile bands, along with error indicators such as MSE, provide intuitive and quantitative criteria for alignment adequacy.

[0099] 5. Outlier labeling strategy: The outlier labeling strategy uses a dual threshold rule of interquartile range (IQR) and median absolute deviation (MAD) to label each indicator as an outlier instead of forcibly deleting it. The influence weight of outliers is reduced in subsequent weight estimation and alignment to ensure robustness and traceability.

[0100] The following experimental data and experimental curve results are used to demonstrate the effectiveness of the present invention.

[0101] The test was conducted according to the test scheme in step S1 of the method provided by this invention, and the changes in six performance parameters of the submarine cable insulation material before and after aging under multi-physical fields were recorded. The performance measurement results are shown in Tables 1 and 2. The data objects are two groups of samples of the same material, with accelerated aging (acc) for 8 time points and natural service (nat) for 4 time points.

[0102] Table 1. Raw experimental data of aging index parameters

[0103]

[0104] Table 2. Raw data of natural service for aging index parameters.

[0105]

[0106] In the table, X1, X2, X3, X4, X5, and X6 represent elastic modulus (MPa), dielectric constant, breakdown strength (kV / mm), crystallinity (%), high-temperature properties (°C), and melting temperature (°C), respectively.

[0107] In step S2 of the method provided by this invention, the baseline is selected as the value of natural service t=0 as the health baseline, and the six indicators are mapped to non-negative degradation quantities. The results are shown in Tables 3 and 4.

[0108] Table 3 Experimental mapping values ​​of aging index parameters

[0109]

[0110] Table 4 Natural service mapping of aging index parameters

[0111]

[0112] In step S3 of the method provided by this invention, the two domains (acc+nat) are... Combine the calculations of the median and interquartile range (IQR), and standardize them according to the above formula to obtain the dimensionless value. Outliers were marked using the IQR / MAD rules (without forced deletion) to ensure robust fitting in subsequent iterations. The results are shown in Tables 5 and 6.

[0113] Table 5. Experimental Standardization Results of Aging Index Parameters

[0114]

[0115] Table 6. Natural service standardization results of aging index parameters

[0116]

[0117] In step S4 of the method provided by this invention, a single aging intensity scale is constructed, and the weighting solution involves normalizing the two time domains to [0, 1]. Fit a normalized time (non-negative least squares NNLS) to obtain non-negative weights and normalize them. This yields... It has a consistent scale across both domains and increases monotonically over time. The non-negative weights of the six indicators are obtained. As shown in Tables 7, 8, and 9, in this batch of data, E and The discernibility of Z is dominant. The fact that the weights of the other indicators are 0 in this small sample does not mean that they are meaningless, but rather that their marginal contribution to the time scale is insufficient under the current data volume.

[0118] Table 7 Non-negative weights of aging index parameters

[0119]

[0120] Table 8. Experimental Z(t) Results

[0121]

[0122] Table 9 Results of natural service Z(t)

[0123]

[0124] According to step S5 of the method provided by this invention, the optimal scaling factor in the alignment result is s = 10.1922; mean square error (MSE) = 0.145879. The equivalent mapping results of the two domains are shown in Tables 10 and 11.

[0125] Table 10 Equivalent Axis Mapping at the Experimental End

[0126]

[0127] Table 11 Equivalent Axis Mapping for Natural Service

[0128]

[0129] According to step S6 of the method provided by the present invention, the acceleration domain time is mapped to using the scaling factor s obtained in step S5. Reference domain In each relative to the reference domain Linear interpolation is obtained Calculate the residuals:

[0130]

[0131] Based on this, the median offset and quantile overlap rate were assessed (using the residual δ = 1.4826·MAD(ΔZ) as a robust tolerance for small samples, and statistically analyzing the proportion of |ΔZ|≤δ), and the stability of the scaling factor s and MSE was tested using the leave-one-out method; simultaneously, statistical analysis was conducted. The proportion of negative slopes. Specific results are shown in Table 12:

[0132] Table 12 Shape Consistency Indicators

[0133]

[0134] The stability results of leave-one reestimation are shown in Table 13:

[0135] Table 13 Stability Results of Leave-One-Out Revaluation

[0136]

[0137] This is the average scaling factor calculated after deleting the i-th data point using the leave-one-out method; The mean squared error (MSE) is the average value obtained after removing the i-th data point using the leave-one-out method. Table 13 shows that the fluctuation range of s and MSE is limited (e.g., s is between 8.33 and 10.42), proving that the scaling factor is not driven by a specific data point, but is based on the stable characteristics of the overall data. This method has strong anti-interference ability and repeatability.

[0138] After completing the alignment calculation and consistency verification of the accelerated aging time axis and the natural service time axis in steps S5 and S6, the accelerated aging test time is equivalently converted based on the obtained optimal time scaling factor, thereby uniformly mapping the aging process under different accelerated aging conditions to the natural service time axis. Through this mapping relationship, the correspondence between accelerated aging samples and natural service samples at the same aging stage can be characterized on the equivalent time axis. The correspondence results of accelerated aging time and natural service equivalent time obtained based on the above method are shown in Table 14.

[0139] Table 14 Comparison of Accelerated Aging Time and Natural Service Equivalent Time

[0140]

[0141] Example 2:

[0142] This embodiment provides a comprehensive aging evaluation system for submarine cable insulation materials based on a time-scaling equivalent axis. The comprehensive aging evaluation system for submarine cable insulation materials includes:

[0143] The data acquisition module is used to acquire multiple preset performance parameters of submarine cable insulation materials after accelerated aging tests in multiple physical fields (electric-thermal-mechanical), and to acquire similar performance parameters of similar materials under natural service conditions.

[0144] The data preprocessing module is used to perform isotropic mapping on various performance parameters under accelerated aging and natural service, converting them into non-negative degradation quantities that monotonically increase with the aging process, and performing robust standardization on each degradation quantity to eliminate dimensional and scale differences. The data preprocessing module includes an isotropic mapping unit, which is used to perform isotropic mapping on performance parameters that do not change monotonically during the aging process using ordinal regression. The preset performance parameters include elastic modulus, dielectric constant, breakdown strength, crystallinity, high-temperature characteristics, and melting temperature. The robust standardization process uses a method based on median and interquartile range.

[0145] The comprehensive aging index construction module is used to weight and synthesize the standardized deterioration quantities into a comprehensive aging index. In the comprehensive aging index construction module, the weights used to synthesize the comprehensive aging index are determined by a data-driven method, and the weights are non-negative and sum to 1.

[0146] The time scaling and alignment module is used to map accelerated aging time to the equivalent time axis using natural service time as the equivalent time axis. It solves for the optimal scaling factor by minimizing the difference between the combined aging index of accelerated aging and natural service on the equivalent time axis, thereby establishing the mapping relationship between accelerated aging time and the equivalent time axis. The time scaling and alignment module solves for the scaling factor by minimizing the mean square error between the combined aging index of accelerated aging and natural service.

[0147] The consistency verification module is used to perform consistency verification on the distribution pattern and numerical differences of the comprehensive aging index of accelerated aging and natural service on the equivalent time horizontal axis, and obtain the consistency verification results. The consistency verification performed by the consistency verification module includes: comparing the median offset and quantile overlap rate of the comprehensive aging index of accelerated aging and natural service, and calculating the alignment error index.

[0148] The results output module is used to output evaluation information of submarine cable insulation materials based on the mapping relationship and the consistency test results. The evaluation information includes: the life assessment result output when the consistency test results meet the preset conditions, and the reliability warning information output when the consistency test results do not meet the preset conditions.

[0149] It is understood that the detailed functional implementation of each of the above modules can be found in the description of the aforementioned method embodiments, and will not be elaborated further here.

[0150] The above-described specific embodiments of the present invention demonstrate the substantial features and advancements of the present invention. Equivalent modifications can be made to these embodiments based on actual usage needs and the guidance of the present invention are all within the scope of protection of this solution.

Claims

1. A comprehensive aging evaluation method for submarine cable insulation materials based on time-scaling equivalent axes, characterized in that: Includes the following steps: S1. Obtain multiple preset performance parameters of submarine cable insulation material after accelerated aging test in multiple physical fields of electro-thermal-mechanical fields, and obtain similar performance parameters of similar materials under natural service conditions. S2. Map each performance parameter under accelerated aging and natural service in the same direction and convert it into a non-negative degradation amount that increases monotonically with the aging process. S3. Perform robust standardization on each degradation quantity obtained in step S2 to eliminate differences in dimensions and scales, and mark outlier data points. S4. Weight the standardized degradation amounts processed in step S3 to form a comprehensive aging index. S5. Using the natural service time as the equivalent time axis, the accelerated aging time is mapped through a single linear scaling factor. By minimizing the difference between the combined aging index of accelerated aging and natural service on the equivalent time axis, the optimal scaling factor is solved, thereby establishing the mapping relationship between accelerated aging time and the equivalent time axis. S6. On the equivalent time axis, conduct a consistency test on the distribution pattern and numerical differences of the comprehensive aging index of accelerated aging and natural service, and obtain the consistency test results. S7. Based on the mapping relationship and the consistency test results, output the evaluation information; wherein, if the consistency test results meet the preset conditions, output the time evaluation results. Otherwise, output a reliability warning message.

2. The comprehensive aging evaluation method for submarine cable insulation materials according to claim 1, characterized in that, In step S2, for performance parameters that do not change monotonically during the aging process, an ordinal regression method is used for homogeneous mapping.

3. The comprehensive aging evaluation method for submarine cable insulation materials according to claim 1, characterized in that, In step S3, a robust standardization method based on the median and interquartile range is adopted.

4. The comprehensive aging evaluation method for submarine cable insulation materials according to claim 1, characterized in that, In step S4, the weights used to synthesize the comprehensive aging index are determined by a data-driven approach, and the weights are non-negative and sum to 1.

5. The comprehensive aging evaluation method for submarine cable insulation materials according to claim 1, characterized in that, In step S5, the scaling factor is solved by minimizing the mean square error between the combined aging index of accelerated aging and natural service.

6. The comprehensive aging evaluation method for submarine cable insulation materials according to claim 1, characterized in that, In step S6, the consistency check includes: comparing the median offset and quantile overlap rate of the combined aging index of accelerated aging and natural service, and calculating the alignment error index.

7. The comprehensive aging evaluation method for submarine cable insulation materials according to claim 1, characterized in that, The preset performance parameters include: elastic modulus, dielectric constant, breakdown strength, crystallinity, thermal decomposition temperature, and melting temperature.

8. A comprehensive aging evaluation system for submarine cable insulation materials based on time-scaling equivalent axes, characterized in that, include: The data acquisition module is used to acquire multiple preset performance parameters of submarine cable insulation materials after accelerated aging tests in multiple physical fields of electricity, heat and force, and to acquire similar performance parameters of similar materials under natural service conditions. The data preprocessing module is used to map the performance parameters under accelerated aging and natural service in the same direction, convert them into non-negative degradation quantities that monotonically increase with the aging process, and perform robust standardization on each degradation quantity to eliminate differences in dimensions and scales. The comprehensive aging index construction module is used to weight and combine the standardized deterioration quantities into a comprehensive aging index. The time scaling and alignment module is used to map accelerated aging time to the equivalent time axis using natural service time as the equivalent time axis. By minimizing the difference between the combined aging index of accelerated aging and natural service on the equivalent time axis, the optimal scaling factor is solved, thereby establishing the mapping relationship between accelerated aging time and the equivalent time axis. The consistency verification module is used to verify the consistency of the distribution pattern and numerical differences of the comprehensive aging index of accelerated aging and natural service on the equivalent time horizontal axis, and obtain the consistency verification results. The results output module is used to output evaluation information of submarine cable insulation materials based on the mapping relationship and consistency test results. The evaluation information includes: the lifetime assessment results output when the consistency test results meet the preset conditions, and the reliability warning information output when the consistency test results do not meet the preset conditions.

9. A comprehensive aging evaluation system for submarine cable insulation materials based on time-scaling equivalent axes as described in claim 8, characterized in that: The data preprocessing module includes a homogenization unit, which is used to homogenize and map performance parameters that do not change monotonically during aging using ordinal regression. The preset performance parameters include elastic modulus, dielectric constant, breakdown strength, crystallinity, thermal decomposition temperature, and melting temperature. Robust normalization is performed using a method based on median and interquartile range.

10. A comprehensive aging evaluation system for submarine cable insulation materials based on time-scaling equivalent axes as described in claim 8, characterized in that: In the comprehensive aging index construction module, the weights used to synthesize the comprehensive aging index are determined through a data-driven approach, and the weights are non-negative and sum to 1. The time scaling alignment module solves for the scaling factor by minimizing the mean square error between the combined aging index of accelerated aging and natural service. The consistency verification module performs the following consistency verifications: comparing the median offset and quantile overlap rate of the combined aging index of accelerated aging and natural service, and calculating the alignment error index.

Images