Rapid evaluation method for voltage tolerance index based on electric stress accumulation damage matrix

Abstract

The invention discloses a rapid evaluation method for a voltage tolerance index based on an electric stress cumulative damage matrix. The rapid evaluation method comprises the following steps: obtaining a short-time breakdown voltage or breakdown field strength of an insulating material film sample by using a rapid-rise test; selecting test parameters of a stepping electrical stress acceleration test based on three-parameter Weibull distribution according to external breakdown field intensity data of the thin film sample; accumulated damage DC is calculated, and an electric stress accumulated damage matrix DCp * q is constructed; solving a centralization matrix # imgabs0 # for the electric stress accumulation damage matrix DCp * q, performing SVD (singular value decomposition) on the centralization matrix # imgabs1 #, and calculating a singular value of the centralization matrix # imgabs2 #; and according to the singular value matrix [sigma] p * q, calculating the value of the voltage tolerance index heuristic function to obtain a proper voltage tolerance index n value. According to the method, the centralized matrix of the SVD decomposition damage matrix is utilized, the singular value is solved and serves as the basis for searching the optimal voltage tolerance index n, the voltage tolerance index n can be quickly and accurately found, and a feasible and reliable data processing method is provided for quickly evaluating the insulating property of a current material.

Description

Technical Field

[0001] The present invention relates to the technical field of voltage tolerance performance evaluation of insulating materials, and in particular to a method for quickly evaluating a voltage tolerance index based on singular values ​​of a covariance matrix of an electric stress cumulative damage matrix. Background Art

[0002] Spontaneous combustion of overhead cables may be caused by the continued heating caused by the degradation of insulation performance after the insulation material deteriorates. Heat accumulates at the hot spots and eventually ignites the insulation material. Therefore, in the aging process of insulation materials, it is necessary to correctly evaluate the voltage tolerance performance of the material.

[0003] The voltage withstand index is one of the parameters that characterizes the insulation performance of solid materials under long-term electric field, and plays an important role in the design of cable insulation thickness, service life design, and type / pre-qualification test withstand voltage level selection. This parameter is based on the classic inverse power model, and its mathematical form is:

[0004] t=DE -n

[0005] Where t is time, D is a constant, E is the average electric field strength, and n is the voltage tolerance index.

[0006] In order to obtain the voltage tolerance index of different insulating materials, an accelerated test with constant electrical stress is usually used, that is, the insulation failure time under different constant electric fields is tested, and the voltage tolerance index n is obtained by linear fitting of lnE~lnt. However, this method has a long test duration and the results are highly dispersed; in order to quickly obtain the voltage tolerance index of insulating materials, a more general method is to use a step electrical stress accelerated test based on the Nelson hypothesis, which can be obtained according to the following formula:

[0007] Where k is the number of steps; ΔE is the field strength increase in each step; Δt is the duration of each step; E0 is the initial field strength; Δt end The duration of the last step; D C is a constant, indicating the critical cumulative damage. For samples of the same shape and material, they should have the same critical cumulative damage under the same test environment. Generally, the voltage tolerance index obtained based on the step acceleration test is similar to the voltage tolerance index obtained based on the constant electrical stress acceleration test, but the step acceleration test shortens the time, so it is widely used in the industrial field.

[0008] The core content of the step electric stress accelerated test is to construct an electric stress cumulative damage matrix, that is, to construct a matrix with p rows and q columns. By selecting p groups of different test parameters and repeating the test q times for each group, the electric stress cumulative damage matrix D is obtained. C ,

[0009]

[0010] For specimens of the same shape and material, they should have the same critical cumulative damage under the same test environment, i.e. D C Ideally, all elements in D should be equal. However, considering the dispersion of weak point penetration, C Therefore, determine a suitable voltage tolerance index n so that D C The value of each element in is as equal as possible, which becomes the key basis for searching n. The common method is to calculate D C The standardized variance of all elements in (Patent Publication No. CN110007199A), find n so that the standardized variance is the smallest. This method has a potential assumption, that is, D C The mean of all elements in is considered to be the true critical loss value of the material. Obviously, this underlying assumption is a biased estimate when the number of test samples is small, that is, Because there are many influencing factors on the test equipment and test conditions and the accumulated value of electrical damage is distributed in a strip-shaped area, there must be non-negligible systematic errors in the measurement results.

[0011] In recent years, some scholars have proposed C The n-value search method of the eigenvalue, that is, using D under ideal conditions C The rank is 1, and the heuristic function is calculated m is the matrix dimension, eigenvalues ​​are arranged from small to large, λ i is the i-th eigenvalue. When I is the largest, the corresponding n value is what we are looking for. However, this method also has many shortcomings: First, this method is only applicable to D C For the case of a square matrix, since only square matrices have eigenvalues, and D C It is not necessarily a square matrix; secondly, D constructed through experiments C The eigenvalues ​​cannot be guaranteed to be non-negative real numbers. C When the eigenvalue has a conjugate complex pair or a negative number, the heuristic function I will not be used or will lose its physical meaning.

[0012] In view of this, this application is hereby filed. Summary of the invention

[0013] In order to solve the above technical problems, the present invention provides a rapid evaluation method for voltage tolerance index based on electrical stress cumulative damage matrix, which uses SVD to decompose the central matrix of the damage matrix and obtain singular values ​​as the basis for searching the optimal voltage tolerance index n. It can construct a critical cumulative damage D of any shape based on the results of electrical stress step acceleration test. C, quickly and accurately find out the voltage tolerance index n, which has the characteristics of clear physical meaning, simplicity, strong universality, and accurate evaluation results, providing a feasible and reliable data processing method for quickly evaluating the insulation performance of current materials.

[0014] The present invention is achieved through the following technical solutions:

[0015] In a first aspect, the present invention provides a method for quickly evaluating a voltage tolerance index based on an electrical stress cumulative damage matrix, comprising the following steps:

[0016] S1, using a rapid-rise test to obtain the short-time breakdown voltage or breakdown field strength of the insulating material film sample;

[0017] S2, according to the short-time breakdown field strength data of the thin film samples, the test parameters of the step electrical stress accelerated test were selected based on the three-parameter Weibull distribution;

[0018] S3, calculation of accumulated damage D C , construct the electrical stress cumulative damage matrix D C | p×q , where p represents the number of experimental schemes, and q represents the number of samples of repeated tests in the same experimental scheme;

[0019] S4, electrical stress cumulative damage matrix D C | p×q Find the centering matrix

[0020] S5, the centralization matrix Perform SVD decomposition and calculate the central matrix The singular values ​​of

[0021] S6, according to the singular value matrix [∑] p×q , calculate the voltage tolerance index heuristic function cosθ and obtain the appropriate voltage tolerance index n value.

[0022] In a specific implementation, step S1 is as follows:

[0023] Sample preparation: Prepare circular film samples with a diameter of 60 mm and a thickness of 0 to 3 mm. The number of samples should be no less than 5. Test the sample thickness and record it with an accuracy of 0.001 mm.

[0024] Test electrode: According to the recommendation of IEC 60243-1-2013, the test electrode is a pure copper cylinder of equal diameter, with a top / bottom diameter of 25mm and a height of about 25mm. The top / bottom and side chamfer radius R3 is 3mm;

[0025] Test environment: The ambient temperature is controlled at 23±2℃. The sample should be placed in an insulating medium to prevent flashover on the sample surface and edge during the boost process. Preferably, the environmental medium is No. 25 mineral insulating oil.

[0026] Excitation conditions: DC voltage is applied to the samples in sequence, and the voltage increases evenly from 0. The voltage increase rate should be sufficient to make most of the samples in the test break down within 10s to 20s to prevent the insulation failure time from being too long or too short. Considering that the voltage increase rate has a significant effect on the space charge in the sample, it can be the same as the voltage increase rate in the step voltage increase test to maintain the consistency of the test. The breakdown voltage is recorded, and the applied electrical stress of the film sample can be calculated based on the thickness.

[0027] In a specific implementation, step S2 is as follows:

[0028] The cumulative failure probability of the three-parameter Weibull distribution is:

[0029]

[0030] Among them, α is the scale parameter, β is the shape parameter, and γ is the location parameter. The three-parameter Weibull distribution fits the breakdown field data better than the two-parameter Weibull distribution fits the test data, so the parameter estimation based on the three-parameter Weibull distribution is more accurate.

[0031] The location parameter γ is first determined by the correlation coefficient method or grayscale theory, and the scale parameter α and shape parameter β are determined by common parameter evaluation methods such as maximum likelihood estimation, least squares method or iteration method;

[0032] At this point, two test parameters with clear physical meanings can be obtained, namely, the maximum field strength E with a breakdown probability of 0 0％ =γ, the characteristic breakdown voltage E is 63.2% breakdown probability 63.2％ =α+γ;

[0033] Thus, the test parameters of the step electric stress accelerated test are obtained:

[0034] The starting test voltage is

[0035] Step voltage in, is the nominal thickness or average thickness of the sample, 40% and 4% are empirical parameters; taken from the recommended value of 20sSST in IEC 60243-1-2013, it can also be modified according to the actual situation, but it should be ensured that Avoid disputes over whether there is cumulative electrical damage under low field strength; the voltage boost rate is 1 kV / s or other recommended rates in the IEC 60243-1-2013 standard, but should be consistent with step S1;

[0036] Step time Δt i The value range is [150,1500]s, i is the number of test groups, i≥3.

[0037] In a specific implementation, step S3 is as follows:

[0038] Prepare p×q test samples identical to those in step S1, where p is the number of test groups and q is the number of repeated samples in each group; according to U0, ΔU and Δt selected in step S2 i , perform a step electrical stress acceleration test on the jth sample in the i-th group (i∈[1,2,…,p], j∈[1,2,…,q]) and obtain the number of steps k i,j and the last step duration Calculate the electrical stress cumulative damage of the sample

[0039]

[0040] Where n is the voltage tolerance index. According to reports on cable insulation materials at home and abroad, the range of n is generally [9, 20]. Therefore, the estimated range of n can be selected as [5, 25]. By traversing all possible n values ​​in the estimated range with an accuracy of 0.1, that is, n∈[5.0, 5.1, 5.2, …, 24.9, 25], all possible electrical stress cumulative damage matrices D are calculated. C | n＝5.0 , D C | n＝5.1 Until D C | n＝25.0 , and obtain the cumulative damage matrix D C | p×q .

[0041] In a specific implementation, step S4 is as follows:

[0042] According to the D constructed in step S3 C | p×q Matrix, each row of q data represents a set of the same test parameters q different sample observations under C Each row represents a feature dimension, and each column represents a sample observation value. Therefore, it can be centered according to the row. Taking the i-th row as an example,

[0043]

[0044] in, Indicates D C The mean of the i-th row.

[0045] Principal component analysis requires The elements in are processed by zero mean to ensure the basic principle of maximum variance reconstruction (Maximum variation formulation) or minimum error reconstruction (Minimum error formulation).

[0046] In a specific implementation, step S5 is as follows:

[0047] According to matrix theory, Where U and V are p-dimensional and q-dimensional unitary matrices respectively; ∑ is a singular value matrix, the elements on the diagonal are singular values, and the values ​​of elements in other positions are 0;

[0048] right Transpose and decompose, but

[0049] Among them, the diagonal elements of ∑ are The square root of the matrix eigenvalue, so the diagonal elements of ∑ are all non-negative real numbers, which is a singular value matrix; U is The characteristic value ∑ 2 The matrix of the corresponding normalized eigenvectors; V is The matrix spanned by each normalized eigenvector is used to calculate the singular value matrix ∑ according to the above formula.

[0050] In a specific implementation, step S6 is specifically as follows:

[0051] According to the singular value matrix [Σ] p×q contains min{p,q} singular values, assuming that p<q, that is, the number of groups with different test parameters is less than the number of samples tested repeatedly in each group, and they are arranged in descending order as σ1≥σ2…≥σ p ;

[0052] Ideally,

[0053] 1) The voltage tolerance index n is not affected by the test parameters The influence of conditions, that is, fully satisfying the premise assumption that "for specimens of the same shape and material, they should have the same critical cumulative damage under the same test environment";

[0054] 2) There is no measurement error during the test;

[0055] Therefore the following relationship exists:

[0056] σ1=σ2=…=σ p =σideal ,

[0057] That is, in the first p-dimensional subspace spanned by U and V, the observation values ​​along each axis should be the same. This is because in an ideal case, the matrix D C Every element in is equal.

[0058] However, the actual situation is not ideal. Due to the distortion of the electric field distribution inside the sample caused by space charge and the randomness of weak point penetration, experimental errors are inevitable. C The observed value of each element in is different, but should be directed toward the ideal value σ ideal The tendency to gather.

[0059] Therefore, this patent uses D C The singular value angle between the observation matrix and the ideal matrix is ​​used as the search basis for the voltage tolerance index n, describing the D under different n values. C The difference between the observation matrix and the ideal matrix, the specific heuristic function is defined as follows:

[0060] Under ideal conditions, The singular value vector of is [1] 1×p =(1,1,…,1), in actual situations, The singular value vector of is [σ]| 1×p =(σ1,σ2,…,σ p );

[0061] For n∈[5.0,5.1,5.2,...,24.9,25.0], find max{f| n＝5.0 ,f n＝5.1 ,…,f| n＝25.0} corresponding to the value of n, the value of n should make the actual vector [σ]| 1×p Close to the ideal vector [1] 1×p , that is, the angle between the two vectors θ→0, then the heuristic function is as follows:

[0062]

[0063] When the value of n makes cosθ maximum (close to 1), the required voltage tolerance index n value is obtained.

[0064] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0065] The present invention proposes a rapid evaluation method for voltage tolerance index based on electrical stress cumulative damage matrix. The central matrix of the damage matrix is ​​decomposed by SVD to obtain singular values ​​as the basis for searching the optimal voltage tolerance index n. The critical cumulative damage D of any shape constructed based on the electrical stress step acceleration test results can be obtained. C, quickly and accurately find out the voltage tolerance index n, which has the characteristics of clear physical meaning, simplicity, strong universality, and accurate evaluation results, providing a feasible and reliable data processing method for quickly evaluating the insulation performance of current materials. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] In order to more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0067] Figure 1 A flow chart of a method for quickly evaluating a voltage tolerance index provided by an embodiment of the present invention;

[0068] Figure 2 A schematic diagram of the calculation principle of the singular value of the cumulative damage matrix provided in an embodiment of the present invention;

[0069] Figure 3 Comparison of the experimental data fitting of the three-parameter Weibull distribution provided by the embodiment of the present invention and the experimental data fitting of the two-parameter Weibull distribution;

[0070] Figure 4 Comparison of the evaluation results of the embodiments of the present invention and the comparative examples;

[0071] Figure 5 A relationship diagram between the heuristic function and the voltage tolerance index provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0072] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with embodiments. The illustrative embodiments of the present invention and their description are only used to explain the present invention and are not intended to limit the present invention.

[0073] In the following description, a large number of specific details are set forth in order to provide a thorough understanding of the present invention. However, it is apparent to those of ordinary skill in the art that these specific details do not have to be employed to practice the present invention. In other embodiments, in order to avoid confusing the present invention, well-known structures are not specifically described.

[0074] Throughout the specification, references to "one embodiment," "an embodiment," "an example," or "an example" mean that a particular feature, structure, or characteristic described in conjunction with the embodiment or example is included in at least one embodiment of the present invention. Therefore, the phrases "one embodiment," "an embodiment," "an example," or "an example" appearing in various places throughout the specification do not necessarily all refer to the same embodiment or example. Furthermore, particular features, structures, or characteristics may be combined in one or more embodiments or examples in any appropriate combinations and / or subcombinations.

[0075] In the description of the present invention, the directions or positional relationships indicated by terms such as “front”, “rear”, “left”, “right”, “up”, “down”, “vertical”, “horizontal”, “high”, “low”, “inside” and “outside” are based on the directions or positional relationships shown in the accompanying drawings and are only for the convenience of describing the present invention and simplifying the description. They do not indicate or imply that the device or element referred to must have a specific direction, be constructed and operated in a specific direction. Therefore, they should not be understood as limiting the scope of protection of the present invention.

[0076] Example 1

[0077] like Figure 1-Figure 5 As shown, a method for rapid evaluation of voltage tolerance index based on the singular values ​​of the covariance matrix of cumulative damage of electric stress is provided. The method obtains the cumulative damage matrix through a stepping electric stress accelerated test, thereby obtaining the corresponding covariance matrix, and constructs an heuristic function based on the singular values ​​of the covariance matrix to achieve the purpose of accurately finding the voltage tolerance index. This method overcomes the shortcomings of the existing methods for searching the voltage tolerance index, such as the heuristic function design is not objective, the physical meaning is not clear, or the calculation method has limitations. Taking a certain LDPE cable material without additives as an example, the implementation steps of the present invention are described as follows:

[0078] S1, rapid voltage rise test to obtain the short-time breakdown voltage or breakdown field strength of the sample, including:

[0079] Sample preparation: Prepare circular film samples with a diameter of 60 mm and a thickness of 0 to 3 mm. The number of samples should be no less than 5. Test the sample thickness and record it with an accuracy of 0.001 mm. Preferably, the number of samples in this test is N = 10.

[0080] Test electrode: According to the recommendation of IEC 60243-1-2013, the test electrode is a pure copper cylinder of equal diameter, with a top / bottom diameter of 25mm and a height of about 25mm. The top / bottom and side chamfer radius R3 is 3mm;

[0081] Test environment: The ambient temperature is controlled at 23±2℃. The sample should be placed in an insulating medium to prevent flashover on the sample surface and edge during the boost process. Preferably, the environmental medium is No. 25 mineral insulating oil.

[0082] Excitation conditions: DC voltage is applied to the samples in sequence. The voltage increases evenly from 0. The voltage increase rate should be the same as that in the subsequent step-by-step voltage increase test to maintain the consistency of the test, that is, the voltage increase rate is preferably 1 kV / s. The breakdown voltage and the applied electric field strength are recorded, as shown in Table 1 below. The applied electric stress on the film sample can be calculated based on the thickness.

[0083] Table 1

[0084]

[0085] S2, according to the breakdown field strength data of the film samples, the test parameters of the step electrical stress accelerated test were selected based on the three-parameter Weibull distribution: the cumulative failure probability of the three-parameter Weibull distribution is Among them, α is the scale parameter, β is the shape parameter, and γ is the location parameter. According to step S1, N=10, i represents the sequence number from small to large according to the breakdown field strength, and the value is the first row of data in the table of step S1. The position parameter γ is first determined by the correlation coefficient method or grayscale theory, and then the scale parameter α and shape parameter β are determined by common parameter evaluation methods such as maximum likelihood estimation, least squares method or iteration method. There are many parameter estimation methods, which will not be repeated here. The parameter results are γ=376, α=63, β=1.47. Therefore, the maximum field strength E with a breakdown probability of 0 can be calculated 0％ =γ = 376 kV / mm and a characteristic breakdown voltage E with a breakdown probability of 63.2% 63.2％ =α+γ=439kV / mm; Preferably, in this test, the starting test voltage of the step electrical stress acceleration test is Step voltage in, For the nominal or average thickness of the sample, the coefficients of 61.6% and 5.1% are guaranteed Avoid disputes over whether there is cumulative electrical damage under low field strength; voltage increase rate is 1kV / s; step time Δt i The value is in the range of [150, 1500] s, i is the test group number, i ≥ 3. Preferably, in this test, three groups of tests are taken, Δt1 = 300 s, Δt2 = 900 s and Δt3 = 1200 s, and the three groups of step times are separated as much as possible to enhance the accuracy of the evaluation results.

[0086] S3, construct the electrical stress cumulative damage matrix [D C ] p×q: Prepare p×q test samples identical to those in step S1, where p is the number of test groups and q is the number of repeated samples in each group; preferably, in this test, p=3 and q=5. Based on U0, ΔU and Δt selected in step S2 i , perform a step electrical stress acceleration test on the jth sample of the i-th group (i∈[1,2,…,p], j∈[1,2,…,q]) and obtain the number of steps k i,j and the last step duration And record it in Table 2 below.

[0087] Table 2

[0088]

[0089] According to Table 2, the electrical stress cumulative damage of the sample can be calculated

[0090]

[0091] Where n is the voltage tolerance index. According to reports on cable insulation materials at home and abroad, the range of n is generally [9, 20]. Therefore, the estimated range of n can be selected as [5, 25]. By traversing all possible n values ​​in the estimated range, the accuracy is taken as 0.1, that is, n∈[5.0, 5.1, 5.2, ..., 24.9, 25]., all possible electrical stress cumulative damage matrices D are calculated. C | n＝5.0 , D C | n＝5.1 Until D C | n＝25.0 .

[0092] For example, when n = 5.0,

[0093]

[0094] S4, electrical stress cumulative damage matrix D C Find the centering matrix According to the D constructed in step S3 C Matrix, each row of q data represents a set of the same test parameters q different sample observations under C Each row represents a feature dimension, and each column represents a sample observation. Therefore, we can center it by row. Taking the i-th row as an example, the common method is:

[0095]

[0096] in, Indicates D C The mean of the i-th row.

[0097] Principal component analysis requires The elements in are processed by zero mean to ensure the basic principle of maximum variance reconstruction (Maximum variation formulation) or minimum error reconstruction (Minimum error formulation).

[0098] For example, when n = 5.0,

[0099]

[0100] S5, calculate the centralization matrix The singular value of is to perform SVD decomposition on it. According to matrix theory, for Can be written as Where U and V are p-dimensional and q-dimensional unitary matrices respectively; ∑ is a singular value matrix, the elements on the diagonal are singular values, and the values ​​of the elements in other positions are 0; Transpose and factorize but Among them, the diagonal elements of ∑ are The square root of the matrix eigenvalue, so the diagonal elements of ∑ are all non-negative real numbers, called the singular value matrix; U is The characteristic value ∑ 2 The matrix of the corresponding normalized eigenvectors; similarly, V is The matrix of each normalized eigenvector. Based on the above description, the singular value matrix ∑ can be calculated.

[0101] For example, when n = 5.0,

[0102]

[0103] S6, according to the singular value matrix [∑] p×q contains min{p,q} singular values. Here, we assume that p<q, that is, the number of groups with different test parameters is less than the number of samples tested repeatedly in each group. In descending order, it can be recorded as σ1≥σ2…≥σ p Ideally,

[0104] 1) The voltage tolerance index n is not affected by the test parameters The influence of conditions, that is, fully satisfying the premise assumption that "for specimens of the same shape and material, they should have the same critical cumulative damage under the same test environment"

[0105] 2) There is no measurement error during the test,

[0106] The following relationship exists:

[0107] σ1=σ2=…=σ p=σ ideal ,

[0108] That is, in the first p-dimensional subspace spanned by U and V, the observation values ​​along each axis should be the same. This is because in an ideal case, the matrix D C Every element in is equal.

[0109] However, the actual situation is not ideal. Not only is the test error inevitable, but also the distortion of the electric field distribution inside the sample caused by space charge and the randomness of weak point penetration will weaken the premise assumption to "for samples of the same shape and material, under the same test environment, the critical cumulative damage is distributed in a band shape". Therefore, D C The value of each element in is different, but should be directed to σ ideal The tendency to gather.

[0110] Therefore, the degree of aggregation of observed values ​​to ideal values ​​can be used as a heuristic function for searching the voltage tolerance index n. Under ideal conditions, The singular value vector of is [1] 1×p =(1,1,…,1); in actual situations, The singular value vector of is [σ]| 1×p =(σ1,σ2,…,σ p ).

[0111] For n∈[5.0,5.1,5.2,...,24.9,25.0], find max{f| n＝5.0 ,f| n＝5.1 ,…,f| n＝25.0}, the required value of n should make the actual vector [σ]| 1×p As close as possible to the ideal vector[1] 1×p , that is, the angle between the two vectors θ→0. Then the heuristic function can be written as:

[0112]

[0113] When the value of n makes cosθ maximum (close to 1), the required value of n is obtained.

[0114] For example, as shown in the figure, n = 12.3, at this time cosθ = 0.99595, the actual vector [σ] | 1×p Almost the same as the ideal vector [1] 1×p The n value of the constant pressure accelerated aging test is n=12.0, which is slightly different from the evaluation result of the method proposed in this patent, verifying the accuracy of the method proposed in this patent.

[0115] Comparative Example 1

[0116] Based on the data in Table 2, Figure 5 As shown, using the traditional D-basedC The matrix normalized variance algorithm (such as patent publication number CN110007199A), n = 13.3, is obviously different from the reference value of n = 12.0, which further verifies the superiority of the patented method.

[0117] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A rapid evaluation method for voltage tolerance index based on electrical stress cumulative damage matrix, characterized in that: The steps include: S1, using the rapid voltage rise test to obtain the short-time breakdown voltage or breakdown field strength of the insulating material film sample; S2, according to the short-time breakdown field strength data of the thin film samples, the test parameters of the step electrical stress accelerated test were selected based on the three-parameter Weibull distribution; S3, calculation of accumulated damage D C , construct the electrical stress cumulative damage matrix D C | p×q , where p represents the number of experimental schemes, and q represents the number of samples of repeated tests in the same experimental scheme; S4, electrical stress cumulative damage matrix D C | p×q Find the centering matrix S5, the centralization matrix Perform SVD decomposition and calculate the central matrix The singular values ​​of S6, according to the singular value matrix [∑] p×q , calculate the voltage tolerance index heuristic function and obtain the appropriate voltage tolerance index n value.

2. The rapid evaluation method of voltage tolerance index based on electrical stress cumulative damage matrix according to claim 1 is characterized in that: Step S1 is specifically as follows: Sample preparation: Prepare circular film samples, the number of samples is not less than 5, test the sample thickness and record it; Test electrode: The test electrode is a pure copper cylinder of equal diameter; Test environment: The ambient temperature is controlled at 23±2℃, and the sample is placed in an insulating medium; Excitation conditions: DC voltage is applied to the samples in sequence. The voltage increases evenly from 0. The voltage increase rate is sufficient to make most samples breakdown within 10s to 20s. The breakdown voltage is recorded and the applied electrical stress on the sample is calculated based on the thickness.

3. The rapid evaluation method of voltage tolerance index based on electrical stress cumulative damage matrix according to claim 1 is characterized in that: Step S2 is specifically as follows: The cumulative failure probability of the three-parameter Weibull distribution is: Among them, α is the scale parameter, β is the shape parameter, and γ is the position parameter; the position parameter γ is first determined by the correlation coefficient method or grayscale theory, and the scale parameter α and shape parameter β are determined by parameter evaluation means; That is, the maximum field strength E with a breakdown probability of 0 0％ =γ, the characteristic breakdown voltage E is 63.2% breakdown probability 63.2％ =α+γ; Thus, the test parameters U0, ΔU and Δt of the step electrical stress acceleration test are obtained. i .

4. The rapid evaluation method of voltage tolerance index based on electrical stress cumulative damage matrix according to claim 3 is characterized in that: The starting test voltage is Step voltage in, is the nominal thickness or average thickness of the sample, 40% and 4% are empirical parameters; Step time Δt i The value range is [150,1500]s, i is the number of test groups, i≥3.

5. The rapid evaluation method of voltage tolerance index based on electrical stress cumulative damage matrix according to claim 4 is characterized in that: Step S3 is as follows: Prepare p×q test samples identical to those in step S1, where p is the number of test groups and q is the number of repeated samples in each group; according to U0, ΔU and Δt selected in step S2 i , perform a step electrical stress acceleration test on the jth sample in the i-th group (i∈[1,2,...,p], j∈[1,2,...,q]) and obtain the number of steps k i,j and the last step duration Calculate the electrical stress cumulative damage of the sample Where n is the voltage tolerance index. The estimated range of n is [5, 25]. By traversing all possible n values ​​in the estimated range with an accuracy of 0.1, that is, n∈[5.0, 5.1, 5.2, ..., 24.9, 25.0], all possible electrical stress cumulative damage matrices D are calculated. C | n＝5.0 , D C | n＝5.1 Until D C | n＝25.0 , and obtain the cumulative damage matrix D C | p×q .

6. The rapid evaluation method of voltage tolerance index based on electrical stress cumulative damage matrix according to claim 5 is characterized in that: Step S4 is specifically as follows: According to the D constructed in step S3 C | p×q Matrix, each row of q data represents a set of the same test parameters q different sample observations under C Each row represents a feature dimension, and each column represents a sample observation value. Centering is performed by row or median to obtain a centered matrix.

7. The rapid evaluation method of voltage tolerance index based on electrical stress cumulative damage matrix according to claim 6 is characterized in that: The method of centering by row is, taking the i-th row as an example, in, Indicates D C The mean of the i-th row.

8. The rapid evaluation method of voltage tolerance index based on electrical stress cumulative damage matrix according to claim 6 is characterized in that: Step S5 is specifically as follows: According to matrix theory, Where U and V are p-dimensional and q-dimensional unitary matrices respectively; ∑ is a singular value matrix, the elements on the diagonal are singular values, and the values ​​of elements in other positions are 0; right Transpose and decompose, but Among them, the diagonal elements of Σ are The square root of the matrix eigenvalue, so the diagonal elements of Σ are all non-negative real numbers, which is a singular value matrix; U is The characteristic value Σ 2 The matrix of the corresponding normalized eigenvectors; V is The matrix spanned by each normalized eigenvector is used to calculate the singular value matrix ∑ according to the above formula.

9. The rapid evaluation method of voltage tolerance index based on electrical stress cumulative damage matrix according to claim 8 is characterized in that: Step S6 is specifically as follows: According to the singular value matrix [∑] p×q contains min{p,q} singular values, assuming that p<q, that is, the number of groups with different test parameters is less than the number of samples tested repeatedly in each group, and they are arranged in descending order as σ1≥σ2…≥σ p ; The degree of convergence of the observed values ​​to the ideal value is used as the heuristic function f for searching the voltage tolerance index n. Under ideal conditions, The singular value vector of is [1] 1×p =(1,1,…,1), in actual situations, The singular value vector of is [σ]| 1×p =(σ1,σ2,…,σ p ); For n∈[5.0,5.1,5.2,…,24.9,25.0], find max{f n＝5.0 ,f n＝5.1 ,…,f n＝25.0 } corresponding to the value of n, the value of n should make the actual vector [σ]| 1×p Close to the ideal vector [1] 1×p , that is, the angle between the two vectors θ→0, then the heuristic function is as follows: When the value of n makes cosθ maximum, the required voltage tolerance index n value is obtained.

Images