A comprehensive evaluation method for air equipment key material application verification

Abstract

The application relates to a comprehensive evaluation method for verifying key material application of aviation equipment, which comprises the following steps: determining the object to be evaluated, and if multiple evaluation objects are contained, the sampling, sample preparation, testing method and testing condition of the multiple evaluation objects should all be the same; establishing an evaluation index system; selecting or constructing a comprehensive evaluation model, calculating the comprehensive evaluation value of each system and giving the comprehensive evaluation result, and giving the comprehensive evaluation result or conclusion according to the calculation result of the mathematical model. The application can obtain the sequence of each verification material through quantitative calculation, get rid of the thinking set of intuitive experience evaluation, and verify the evaluation result through various models, so that the accuracy and fairness of the material application verification of aviation equipment are improved, and the application has high social benefits.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of comprehensive evaluation of the performance of materials for aviation equipment, and particularly relates to a comprehensive evaluation method for application verification of key materials for aviation equipment. BACKGROUND

[0002] The rapid development of new generation aviation equipment puts forward higher requirements for the function performance, quality reliability, environmental adaptability and stable connection of the basic products such as electronic components, key materials, basic mechanical and electrical products and key software and hardware. With the development of the field of basic products, new basic products with new design, new structure, new material and new technology are emerging. Whether these new products can meet the development and construction needs of new aviation equipment needs to be verified by application. Application verification refers to a series of targeted tests, analysis, evaluation and iterative improvement work carried out in stages and in categories after the completion of product development identification of basic products, around the actual standard and equipment use requirements, under the actual or close to actual use conditions, for the adaptability of basic products, environmental adaptability, quality reliability and the like. Since there is no ready-made experience to be referred to for the application verification work of basic products, there is a lack of mature theory, method, standard and management mode, and therefore it is necessary to establish an application verification system for aviation equipment materials for evaluation.

[0003] The key material application verification field is carried out according to the six steps of determining the application verification demand, establishing the application verification index, clarifying the application verification elements, establishing the application verification process, implementing the application verification test and giving the application verification evaluation. The evaluation step requires a fuzzy conclusion that whether the material is usable, conditionally usable or unusable, but there is no clear provision on how to evaluate. In the practical process of application verification, it is found that when multiple material suppliers and multiple evaluation indexes participate in comprehensive evaluation at the same time, the traditional intuitive judgment and analysis of dispersion degree evaluation method cannot determine the good or bad of the materials, which affects the quality of the results of the whole application verification work.

[0004] After investigation, the quantitative comprehensive evaluation in the field of aviation components has developed earlier, and there are a large number of reference literature and patents explaining the elements and evaluation methods of quantitative evaluation, but the method of quantitative evaluation using mathematical model in the field of aviation materials is still in a blank state. When evaluating the advantages and disadvantages of materials, many large testing institutions generally use intuitive or experience judgment, and when encountering evaluation with different dimensions, scattered data and large amount of data, it is often difficult to evaluate or cause wrong judgment. Therefore, a quantitative evaluation method based on mathematical model is needed to replace the intuitive and experience evaluation method and solve the corresponding problems. The problems to be solved in the evaluation work of aviation equipment material application verification include the following aspects:

[0005] (1) A method is needed to evaluate all verification indexes on the same latitude;

[0006] (2) A quantitative method is needed to evaluate the quality of all validation materials, including benchmark materials, and to make a comprehensive ranking.

[0007] (3) A quantitative method is needed to evaluate the gap between the various indicators of the verification material and the benchmark material;

[0008] (4) A quantitative method is needed to evaluate which verification material is most similar to the benchmark material.

[0009] These four questions correspond to different demand scenarios. (1) The question concerns how to compare data of various indicators at different orders of magnitude; (2) The question concerns evaluating which material has the best overall performance among several materials that meet the minimum usage requirements; (3) The question concerns which indicators of the verification material and the benchmark material are similar, which indicators require process improvement, and which indicators require strengthened testing; (4) The question concerns which verification material is closer to the benchmark material or the required performance indicators, with performance that is neither too low nor too high. Solving these four questions will provide very useful guidance for the conclusions of the verification and evaluation of air equipment materials. Summary of the Invention

[0010] (1) Technical problems to be solved

[0011] This invention provides a comprehensive evaluation method for the application verification of key materials in aerospace equipment, comprising: identifying the object to be evaluated, establishing an evaluation index system, determining the weight coefficients corresponding to each evaluation index, and selecting or constructing a comprehensive evaluation model. This invention derives the ranking of each verification material through quantitative calculation, breaking away from the mindset of intuitive experience-based evaluation. Furthermore, the evaluation results can be verified through multiple models, improving the accuracy and fairness of aerospace equipment material application verification and yielding significant social benefits.

[0012] (2) Technical solution

[0013] An embodiment of the present invention provides a comprehensive evaluation method for the application verification of key materials in aerospace equipment, comprising the following steps:

[0014] Step 1: Determine the object to be evaluated. If there are multiple objects to be evaluated, the sampling, sample preparation, testing methods, and testing conditions for all objects should be identical.

[0015] Step two: Establishing the evaluation index system, which includes first, second and third level evaluation indexes. The first level evaluation indexes include material basic performance, environmental adaptability, functional performance and process adaptability. The second level evaluation indexes include mechanics, physical performance, electric / thermal performance, welding performance, cementing process performance, weather resistance and vibration resistance. The third level evaluation indexes include tensile strength, compression strength, fatigue strength, density, conductivity, dielectric constant, cementing strength, welding strength, high temperature storage, damp heat aging and vibration test. After the index system is established, the corresponding detection data is ensured to have certain uniformity, and each evaluation index has the same data length and precision.

[0016] Step three: Determining the weight coefficients corresponding to each evaluation index, which include three levels of key, important and general.

[0017] Step four: Selecting or constructing a comprehensive evaluation model, calculating the comprehensive evaluation value of each system and giving the comprehensive evaluation result. The comprehensive evaluation result or conclusion is given according to the calculation result of the mathematical model.

[0018] Further, the uniformity processing method of the evaluation index includes:

[0019] Firstly, the evaluation index is defined as "extreme type" index, "minimum type" index, "intermediate type" index and "interval type" index. The extreme type index means that the larger the value of the index is, the better it is. The minimum type index means that the smaller the value of the index is, the better it is. The intermediate type index means that the value of the index should not be too large or too small, and the appropriate intermediate value is the best. The interval type index means that the value of the index should be in a certain determined interval.

[0020] Secondly, according to the definition, the evaluation index is processed as an extreme type index, and the calculation method is as follows:

[0021] Minimum type index: for a minimum type index x, by transforming

[0022]

[0023] or transforming

[0024] x' = M - x (2)

[0025] The minimum value index x is converted into a maximum value x', wherein M is the maximum value of the possible value of the index x, so that the index x is maximized.

[0026] Intermediate type index: for a certain intermediate type index x, by transforming

[0027]

[0028] where M and m are the maximum and minimum of the possible values of the index x, respectively, i.e. the intermediate index x can be maximized to x'.

[0029] Interval index: for an interval index x, it can be transformed to

[0030]

[0031] where [a, b] is the best stable interval of the index x, c = max{a - m, M - b}, M and m are the maximum and minimum of the possible values of the index x, respectively; i.e. the interval index x can be maximized to x'.

[0032] Further, the uniformity processing of the evaluation index further includes dimensionless processing, and the specific method includes a standard deviation method, an extreme difference method, or an efficacy coefficient method.

[0033] Further, the standard deviation method includes:

[0034] Let

[0035] where

[0036]

[0037]

[0038] Obviously, the mean and mean square error of the index x' ij are 0 and 1, respectively, i.e. is a dimensionless index, which is called the standard observation value of x ij .

[0039] Further, the extreme difference method includes:

[0040] Let

[0041] where

[0042]

[0043]

[0044] Then x' ij ∈ [0, 1] is a dimensionless index observation value.

[0045] Further, the efficacy coefficient method:

[0046] Let

[0047] wherein c, d are both determined constants.C represents 'translation amount', d represents 'rotation amount', i.e., 'enlargement' or'reduction' multiple, then x' ij ∈ [c, c+d].

[0048] Further, the method for determining the weight coefficient comprises: determining the weight coefficient based on the key degree of the evaluation index, and the specific method is as follows:

[0049] Firstly, the key index is defined as three-star level, the important index is defined as two-star level, and the general index is defined as one-star level.

[0050] Divide the star number n of the index key by the total star number As the weight coefficient of the corresponding index, the weight formula is calculated:

[0051]

[0052] Wherein ∑w ij =1.

[0053] Further, the comprehensive evaluation model comprises a T test model, an Euclidean distance model, a subjective linear weighted evaluation model, a Borda ordering model and an objective weighting model, the T test model takes each parameter as a single dimension sample, respectively uses T test comparison on each selection parameter and standard parameter, adopts single sample nonparametric test and finally determines the comprehensive similarity according to the similarity degree of each parameter, and the specific test method is as follows:

[0054] Single sample mean T test: test whether the average value of a single sample is equal to the target value, and the test steps are as follows:

[0055] Firstly, the sample standard error Se is calculated.

[0056] Standard error

[0057] Then, the sample T value is calculated.

[0058] Wherein the sample mean is the mean value of the test sample in a single index, the overall average value m is the expected value, and Se is the standard error.

[0059] Then, according to the T value, the P value is obtained by searching the T table.

[0060] Finally, the obtained P value is compared with the significance level a, when P < a, it is considered that there is significant difference; when P > a, it is considered that there is no significant difference, when T > 0, it is verified that the mean value of the material index data is higher than that of the benchmark material, when T < 0, it is lower, so four results of "positive equivalence", "negative equivalence", "positive significant difference" and "negative significant difference" can be obtained; wherein "negative significant difference" can represent that the mean value of the benchmark material is used as the qualified criterion, and the performance index of the verified material is almost unqualified no matter how to strengthen the screening, "positive significant difference" represents that the index is almost qualified no matter how to strengthen the screening; "positive equivalence" represents that the probability of being qualified is higher, and "negative significant difference" represents that the probability of being qualified is lower.

[0061] Further, the Euclidean distance model is used to comprehensively evaluate the similarity degree of the verified material and the benchmark material, and the specific evaluation method is as follows:

[0062] Suppose that the ideal point of the benchmark material index is For an evaluated object (x i1 ,x i2 ,…,x im ), the weighted distance between the two is defined as

[0063]

[0064] Where w j is the weight coefficient, is the distance between x ij and in a certain sense;

[0065] Under normal circumstances, the simple Euclidean distance can be taken, that is, taking

[0066]

[0067] Then the comprehensive evaluation function is

[0068]

[0069] After calculation, according to the size of y i (i = 1, 2, …, n), each evaluated scheme is sorted and optimized, obviously, the smaller the value is, the better the scheme is; especially, when a certain y i = 0, that is, the ideal point is reached, then the corresponding verified material is similar.

[0070] Further, the objective weighting model is used for comparison when the weight is not easy to determine or is not clear about the influence of the material verification index on the type product, and can also be used for verification of the evaluation results of other models, and the specific method includes:

[0071] Select i materials, j indexes, and x ijThe value of the jth index of the ith material;

[0072] The proportion of the ith material in the jth index is calculated:

[0073]

[0074] The entropy value of the jth index is calculated:

[0075]

[0076] Where k = 1 / ln(n) > 0, e j ≤0.

[0077] The information entropy redundancy is calculated:

[0078] d j =1-e j (23)

[0079] The weight of each index is calculated:

[0080]

[0081] The comprehensive score of each material is calculated:

[0082]

[0083] The higher the score, the better the material.

[0084] (3) Beneficial effects

[0085] The present application can obtain the ranking of each verification material through quantitative calculation, break away from the thinking set of intuitive experience evaluation, and verify the evaluation result through multiple models, thereby improving the accuracy and fairness of the application verification of aviation equipment materials, and having high social benefits. At the same time, the present application can provide targeted improvement suggestions for the materials, improve the efficiency of material process improvement, and has high economic benefits.

[0086] Additional aspects and advantages of the present application will be given in part in the following description, some of which will become apparent from the following description, or will be learned by practice of the present application. BRIEF DESCRIPTION OF DRAWINGS

[0087] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiments of the present application. Obviously, the drawings described below are only some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without creative labor on the basis of these drawings.

[0088] Figure 1is a flowchart of a comprehensive evaluation method for verifying key material applications of aviation equipment according to an embodiment of the present application.

[0089] Figure 2 is a matrix diagram for determining weight coefficients based on an expert evaluation weighting method according to an embodiment of the present application.

[0090] Figure 3 is a basic information diagram of an evaluation object and a comparison material according to an embodiment of the present application.

[0091] Figure 4 is an evaluation index diagram of an index system according to an embodiment of the present application.

[0092] Figure 5 is performance detection data of Axx brand samples according to an embodiment of the present application.

[0093] Figure 6 is performance detection data of Bxx brand samples according to an embodiment of the present application.

[0094] Figure 7 is performance detection data of Cxx brand samples according to an embodiment of the present application.

[0095] Figure 8 is a data diagram of Axx brand sample homogenization processing according to an embodiment of the present application.

[0096] Figure 9 is a data diagram of Bxx brand sample homogenization processing according to an embodiment of the present application.

[0097] Figure 10 is a data diagram of Cxx brand sample homogenization processing according to an embodiment of the present application.

[0098] Figure 11 is a data diagram of Axx brand sample dimensionless processing according to an embodiment of the present application.

[0099] Figure 12 is a data diagram of Bxx brand sample dimensionless processing according to an embodiment of the present application.

[0100] Figure 13 is a data diagram of Bxx brand sample dimensionless processing according to an embodiment of the present application.

[0101] Figure 14 is an index weight coefficient diagram according to an embodiment of the present application.

[0102] Figure 15 is a W test result diagram according to an embodiment of the present application.

[0103] Figure 16 is a T test calculation result diagram according to an embodiment of the present application.

[0104] Figure 17 is a saliency detection result map in an embodiment of the present application.

[0105] Figure 18 is a calculation process map of the ideal point approximation method in an embodiment of the present application.

[0106] Figure 19 is a calculation result map of the subjective weighting model in an embodiment of the present application.

[0107] Figure 20 is a data map of the Axx brand entropy value method in an embodiment of the present application.

[0108] Figure 21 is a data map of the Bxx brand entropy value method in an embodiment of the present application.

[0109] Figure 22 is a data map of the Cxx brand entropy value method in an embodiment of the present application.

[0110] Figure 23 is a weight distribution map of the entropy value method in an embodiment of the present application. DETAILED DESCRIPTION

[0111] The embodiments of the present application will be further described in detail below with reference to the accompanying drawings and embodiments. The detailed description and drawings of the following embodiments are used to exemplarily illustrate the principles of the present application, but cannot be used to limit the scope of the present application, that is, the present application is not limited to the described embodiments, and covers any modification, replacement and improvement of the parts, components and connection modes without departing from the spirit of the present application.

[0112] It should be noted that the embodiments in the present application and the features in the embodiments can be combined with each other without conflict.

[0113] The embodiments of the present application will be further described in detail below with reference to the accompanying drawings and embodiments. The detailed description and drawings of the following embodiments are used to exemplarily illustrate the principles of the present application, but cannot be used to limit the scope of the present application, that is, the present application is not limited to the described embodiments, and covers any modification, replacement and improvement of the parts, components and connection modes without departing from the spirit of the present application. Figure 1 - the accompanying drawings Figure 23 The present application will be described in detail below with reference to the accompanying drawings and embodiments.

[0114] Referring to the accompanying drawings Figure 1 shown, according to the comprehensive evaluation method for verification of key material application of aviation equipment according to an embodiment of the present application, the following steps are included:

[0115] Step one: determine the evaluation object: after the object to be evaluated is determined, the selection of the subsequent establishment of the index system and the evaluation model, for multiple evaluation objects, the sampling, sample preparation, test method and test conditions should all be the same.

[0116] Step two: Establishing evaluation index system: The index system includes first, second and third level indexes. The first evaluation index generally includes material basic performance, environmental adaptability, functional performance, process adaptability, etc. The second evaluation index can generally include mechanical property, physical property, electrical / thermal property, welding property, cementing process property, weather resistance, vibration resistance, etc. The third evaluation index can include tensile strength, compressive strength, fatigue strength, density, electrical conductivity, dielectric constant, cementing strength, welding strength, high temperature storage, damp heat aging, vibration test, etc. The index system needs to be formulated according to the actual process conditions and use conditions of the aviation equipment, and the minimum value of the index use requirement is formulated. After the index system is established, the corresponding detection data should ensure a certain uniformity, each evaluation index has the same data length and precision, and the data can be processed to ensure that the data meet the quantitative calculation model.

[0117] Step three: Determining the weight coefficient corresponding to each evaluation index: The weight of the corresponding evaluation index is formulated according to the service characteristics of the aviation equipment, which is generally divided into key, important and general, and the weight of all the evaluation indexes participating in the evaluation can also be determined according to the experience of experts.

[0118] Step four: Selecting or constructing a comprehensive evaluation model: Selecting or constructing an evaluation model, when one evaluation model cannot meet the demand, multiple models need to be combined for evaluation or mutual verification; then, the comprehensive evaluation value of each system is calculated and the comprehensive evaluation result is given, and the comprehensive evaluation result or conclusion is given according to the calculation result of the mathematical model.

[0119] According to another embodiment of the present application, the evaluation index is processed uniformly, specifically including:

[0120] Firstly, the evaluation index is defined as "max type" index, "min type" index, "intermediate type" index and "interval type" index:

[0121] (1) Max type index: always expects the value of the index to be as large as possible;

[0122] (2) Min type index: always expects the value of the index to be as small as possible;

[0123] (3) Intermediate type index: always expects the value of the index to be neither too large nor too small, i.e. the appropriate intermediate value is the best;

[0124] (4) Interval type index: always expects the value of the index to be in a certain determined interval.

[0125] According to the definition, the evaluation index is processed as a max type index, so that all the mathematical model calculations can be carried out according to the following method:

[0126] Minimizing index: for a minimizing index x, by transforming

[0127]

[0128] or transforming

[0129] x' = M - x (2)

[0130] By transforming the minimization index x into maximization index x', where M is the maximum value of the possible values of index x, the index x can be maximized.

[0131] Intermediate index: for an intermediate index x, by transforming

[0132]

[0133] By transforming the intermediate index x into maximization index x', where M and m are the maximum and minimum values of the possible values of index x, the intermediate index x can be maximized.

[0134] Interval index: for an interval index x, by transforming

[0135]

[0136] By transforming the interval index x into maximization index x', where [a, b] is the best stable interval of index x, c = max{a - m, M - b}, and M and m are the maximum and minimum values of the possible values of index x. The interval index x can be maximized.

[0137] According to another embodiment of the present application, for the evaluation index, there are often different units and orders of magnitude among the evaluation indexes in practice, which causes the incommensurability among the indexes, and brings difficulties to the comprehensive evaluation, especially the establishment of the index and the unreasonable ordering according to the size of the index. If the indexes are not processed by the dimensionless processing, the error result of "large number eating small number" will occur in the comprehensive evaluation process, thus leading to the wrong evaluation conclusion. The dimensionless processing is also called the standardization or normalization processing of the index data.

[0138] In the embodiment of the present application, the standard deviation method, the extreme difference method and the efficacy coefficient method can be used for the dimensionless processing of the evaluation index. Assuming that there are m evaluation indexes x1, x2, …, xm, and n sets of sample observation values x m , it is assumed that the different types of uniformization processing have been performed, and there are n sets of sample observation values x ij (i = 1, 2, …, m, j = 1, 2, …, n), which are processed by the dimensionless processing.

[0139] (1) Standard deviation method: let

[0140]

[0141] wherein

[0142]

[0143]

[0144] Obviously, the mean and mean square error of the index x' ij (i = 1, 2, …, m; j = 1, 2, …, n) are 0 and 1, respectively, i.e. is a dimensionless index, called x ij standard observation value.

[0145] (2) Extreme difference method: let

[0146]

[0147] wherein

[0148]

[0149]

[0150] then x' ij ∈ [0, 1] is a dimensionless index observation value.

[0151] (3) Efficacy coefficient method:

[0152] Let

[0153]

[0154] wherein c, d are both determined constants. c represents "translation amount", d represents "rotation amount", i.e. "enlargement" or "reduction" multiple, then x' ij ∈ [c, c + d]. For example, if c = 60 and d = 40, then x' ij ∈ [60, 100].

[0155] According to another embodiment of the present application, the method for determining the weight coefficient comprises: determining the weight coefficient based on the criticality of the evaluation index, and the specific method is as follows: the critical index can be considered as three-star, the important index as two-star, and the general index as one-star. The number of stars n of the critical index is divided by the total number of stars As the weight coefficient of the corresponding index, the weight calculation formula is:

[0156]

[0157] wherein ∑w ij = 1.

[0158] According to another embodiment of the present application, the method for determining the weight coefficient further comprises determining the weight coefficient based on an expert evaluation weighting method, which is used when the importance of the index needs to be subdivided, and the index is compared with each other according to the longitudinal coordinate index in the matrix formed as shown in the figure. The important index is filled with 1 and the unimportant index is filled with 0 in the figure, and finally the ratio of the total number of the index to the total number is used as the weight. Figure 2

[0159] According to another embodiment of the present application, the comprehensive evaluation model comprises a T test model, an Euclidean distance model, a subjective linear weighted evaluation model, a Borda ranking model and an objective weighting model. The probability of the target value of the target material is obtained by comparing the single evaluation index by using the T test. The model takes each parameter as a single dimension sample, respectively compares each selected parameter with the standard parameter by using the T test, adopts the single sample nonparametric test and finally determines the comprehensive similarity according to the ranking of the similarity of each parameter. The specific test method is shown as follows:

[0160] Single sample mean T test: test whether the average value of the single sample is equal to the target value, and the test steps are as follows:

[0161] (1) sample standard error Se:

[0162]

[0163] Standard error

[0164] (2) calculate the sample T value:

[0165]

[0166] Wherein the sample mean is the mean value of the test sample in the single index, the overall average value m is the expected value, and Se is the standard error.

[0167] According to the T value, find the T table to obtain the P value.

[0168] ​The obtained P value is compared with the significance level a, which can be 0.01 or 0.05 according to requirements, since the main influencing factor of the T value corresponding to the P value is the standard deviation, the T value can be understood as the distance of the sample mean value from the population mean value in units of sample standard deviation. When P < a, it is considered that there is a significant difference; when P > a, it is considered that there is no significant difference. When T > 0, it is verified that the mean value of the material index data is higher than that of the reference material, and when T < 0, it is lower. Therefore, four results of "positive equivalence", "negative equivalence", "positive significant difference" and "negative significant difference" can be obtained. Among them, "negative significant difference" can indicate that the mean value of the reference material is used as the qualified criterion, and the performance index of the verification material is almost unqualified no matter how the screening is strengthened, "positive significant difference" indicates that the index is almost qualified no matter how the screening is strengthened; "positive equivalence" indicates a higher probability of being qualified, and "negative significant" indicates a lower probability of being qualified. When doing raw material entry inspection, attention should be paid to "negative equivalence" and "positive equivalence" indexes, and for indexes with "positive significant difference", the corresponding acceptance conditions should be adjusted, and for indexes with "negative significant difference", the material supplier should optimize the process or enhance product quality control. Since the P value in the T test result represents the size of the probability, no specific quantitative analysis based on the P value is made here.

[0169] The Euclidean distance model is used for comprehensive evaluation of the similarity degree of the verification material and the reference material, and the Euclidean distance model is used for evaluation.

[0170] Suppose the ideal point (reference material index) is For an evaluated object (x i1 ,x i2 ,…,x im ), the weighted distance between the two is defined as

[0171]

[0172] where w j is the weight coefficient, is the distance between x ij and in some sense.

[0173] Under normal circumstances, the simple Euclidean distance can be taken, that is,

[0174]

[0175] Then the comprehensive evaluation function is

[0176]

[0177] After calculation, the evaluated schemes are sorted and optimized according to the size of y i (i = 1, 2, …, n), obviously the smaller the value is, the better the scheme is. In particular, when a certain yi When = 0, the ideal point is reached, and the corresponding verification material is similar. The Euclidean distance can be analyzed and judged by weight for different use conditions, and is suitable for the application evaluation of verification materials.

[0178] The subjective linear weighted evaluation model uses the index criticality of product use requirements as the weight evaluation, which can reflect the importance of different types of materials performance. Since the units of various indicators are not unified, they are standardized before calculating the comprehensive index, that is, the absolute value of the index is converted into a relative value

[0179] Linear weighted function:

[0180] As a comprehensive evaluation model, the comprehensive performance is better when y is larger. This model is suitable for comprehensive ranking and can be used for comprehensive comparison of verification materials and benchmark materials. Among them, j w is the pre-set weight, is the average value of the ith index.

[0181] The Borda ranking model can be suitable for simple comprehensive scoring and ranking based on single index comparison results, such as dispersion ranking and pass rate ranking. Assuming that the evaluation objects are n materials and m indexes, each index is ranked and scored, with the worst score being 1, the second being 2,..., and the highest score being n. The highest score is the best.

[0182] The objective weighting model is mainly used for comparison when the weight is not easy to determine or the influence of the material verification index on the product is not clear. It can also be used to verify the evaluation results of other models. Using the concept of information entropy, the index characteristics of all materials are calculated as a large sample to calculate the weight of the index, and the results are ranked, finally reducing the potential risks caused by incomplete consideration. This model can be used to verify the evaluation results of linear weighting and dispersion comprehensive ranking.

[0183] Select i materials and j indexes, x ij is the value of the jth index of the ith material.

[0184] Since the units of various indicators are not unified, they are standardized before calculating the comprehensive index, that is, the absolute value of the index is converted into a relative value. After standardization, the proportion of the ith material under the jth index to the index can be calculated:

[0185]

[0186] Calculate the entropy value of j indexes:

[0187]

[0188] wherein k = 1 / ln(n) > 0, satisfying e j ≤ 0.

[0189] Calculate the information entropy redundancy:

[0190] d j = 1-e j (24)

[0191] Calculate the weight of each option index:

[0192]

[0193] Calculate the comprehensive score of each material:

[0194]

[0195] The higher the score, the better the material.

[0196] The above embodiments of the present application can obtain the ranking of each verification material by quantitative calculation, get rid of the thinking set of intuitive experience evaluation, and verify the evaluation results through multiple models. For material manufacturers, the comprehensive evaluation can be used to evaluate which materials need to be improved in process, which is more clear compared to the traditional evaluation steps. Instead of only giving the conclusion of whether the material is usable, conditionally usable or unusable, the quantitative suggestion can be given to avoid the subjective bias caused by the evaluation method of only using intuitive judgment and analyzing dispersion. For material users, the comprehensive evaluation system can provide more effective and accurate data support for the application of materials, so that the evaluation results of materials are more scientific and fair. At the same time, it can be concluded which index of the evaluation object needs to be strengthened for retest, which can provide effective test for improving the overall quality of aviation equipment.

[0197] In summary, the above-mentioned quantitative application verification method has high social benefits in improving the accuracy and fairness of aviation equipment material application verification. At the same time, it can provide targeted improvement suggestions for materials, improve the efficiency of material process improvement, and has high economic benefits.

[0198] The following comprehensive evaluation analysis of Axx brand, Bxx brand and Cxx brand materials is an embodiment to illustrate the beneficial effects of the comprehensive evaluation method for aviation equipment key material application verification.

[0199] Step 1: Identify the object to be evaluated: In this embodiment of the invention, firstly, the differences between the standard materials and the benchmark materials are evaluated based on the comprehensive verification test data of three types of verification materials: Axx, Bxx, and Cxx. The evaluation results of whether the verification materials are "superior" or "inferior" to the benchmark materials are given; and the overall performance ranking of the verification materials is provided. The comparison materials are from the XXX series. Basic information on the evaluation object and benchmark materials is attached. Figure 3 As shown.

[0200] Step 2: Establish an evaluation index system

[0201] As required, the indicator system includes indicators such as basic material properties, process adaptability, functional performance, and environmental adaptability. See the appendix for details. Figure 4 As shown, importance is categorized into three levels: one star for general indicators, two stars for important indicators, and three stars for critical indicators. Critical indicators are those that must meet the minimum usage requirements of the product model; important indicators are those that may affect the model's performance; and general indicators are those that are not yet clearly defined or have little impact on the product model's performance. Within each indicator category, maximum indicators are better the higher the value; minimum indicators are better the lower the value; intermediate indicators are better the closer the value is to the minimum; and range indicators are better the data falls within a specific range. See the appendix for indicator requirements for each product model. Figure 4 As shown. (The attached text is incomplete and cannot be translated.) Figure 5 -Appendix Figure 7 for Test sample performance test data Processing diagram.

[0202] Then the data is normalized, according to the appendix. Figure 4 In the application verification index system, the definitions of each index are as follows: the maximum value index remains unchanged during the standardization process; the minimum value is maximized by taking the reciprocal of formula (1) based on the actual situation; and the interval value index is standardized according to formulas (5)-(7). During calculation, the data of the three materials are considered as a whole, meaning the range of the maximum and minimum values ​​is the data of all materials under the same index. The data processing procedure is shown in the appendix. Figure 8 -Appendix Figure 10 .

[0203] Next, the data is dimensionless. The extreme value difference method is used for dimensionless processing using formulas (11)-(13). During the calculation, the data for the three materials are considered as a whole, meaning the range of the maximum and minimum values ​​is the data for all materials under the same index. See the appendix for the data processing procedure. Figure 11 -Appendix Figure 13 .

[0204] Step 3: Determine the weighting coefficients corresponding to each evaluation indicator.

[0205] The importance of the indicators is defined according to the requirements. The calculation results of the weight coefficients of the 24 indicators involved in the evaluation are attached. Figure 14shown.

[0206] Step four: select or construct comprehensive evaluation model

[0207] First, the similarity evaluation with the benchmark material, evaluation index difference comparison analysis

[0208] The index data of the benchmark sample is used as a reference index, and the indexes of Axx, Bxx and Cxx materials are subjected to T test. Since the premise of single sample T test is that the sample conforms to normal distribution, it is necessary to detect whether the provided sample conforms to normal distribution. The Shapiro-Wilk test method (W test) recommended by the composite national standard GB4882 is used for calculation. A correlation coefficient can be obtained, which is closer to 1, indicating that the data and normal distribution are better fitted. In this hypothesis test, set:

[0209] H0: The sample data is not significantly different from the normal distribution.

[0210] HA: The sample data is significantly different from the normal distribution.

[0211]

[0212] The test steps are as follows:

[0213] (1) Reorder the homogenized data according to the numerical value, so that x1≤x2≤…≤xn;

[0214] (2) (∑a i x i ) 2 is the best linear unbiased estimate of (n-1)σ 2 , σ is the standard deviation of the normal distribution from which the sample comes, and the exact value of a is

[0215] a=(m T V -1 V -1 m) -1 / 2m T V -1 .

[0216] Where V is the covariance matrix of the order statistics of n normally distributed random variables, and m is the vector of the expected composition of these variables.

[0217] (3) Calculate the denominator of the above formula, is the sample mean;

[0218] (4) Calculate the value of α, which can be obtained by looking up the table;

[0219] (5) Calculate the test statistic W,

[0220] (6) Calculate the test statistic P, which can be obtained by looking up the table, and compare it with α to draw a conclusion. When P is greater than 0.01, it is considered to meet the normal distribution, and T test can be performed.

[0221] Since the calculation of W test is more complex, it is recommended to use data analysis tools such as R language for batch calculation. The test results are shown in the attached Figure 15 .

[0222] In the results of W test, when the confidence interval is set to 99%, the dielectric constant and the tangent of the loss angle do not meet the W test, the length direction size change rate of Axx and Bxx materials at 180℃ / 3h does not meet the W test, and the rest of the sample values meet the normal distribution. The samples that meet the W test are used to determine the difference between the indicators in the T test scheme and the indicators in the benchmark scheme, and the results are shown in the attached Figure 16 .

[0223] After obtaining the T value of each indicator in each material, the significance level P is calculated. When P>0.01, it is considered that the two are equivalent, and when P≤0.01, it is considered that the difference is significant. Here, α is taken as 0.01. The verification indicators that do not meet the W test are determined by the intuitive judgment method, and the results are shown in the attached Figure 17 .

[0224] Axx brand material is superior to the benchmark material in 4 indicators, and equivalent in 11 indicators; Bxx material is superior in 3 indicators and equivalent in 12 indicators; Cxx material is superior in 11 indicators and equivalent in 8 indicators. The manufacturer should improve the process to improve the corresponding performance for the negative significant difference items, and pay attention to the pass rate in the raw material inspection for the negative equivalent indicators. (This evaluation cannot use the number of superior and equivalent as the standard of material quality).

[0225] The ideal point method is used to process the dimensionless data of the three materials verification indicators, and the standard deviation and mean value are calculated using the total sample. The verification indicators of the benchmark product are processed by dimensionless and the similarity degree is evaluated according to the Euclidean distance formula. The calculation results are shown in the attached Figure 18 .

[0226] From the above calculation results, Axx brand material is more similar to the benchmark material, followed by Cxx brand material, but it does not mean that the material is superior to the benchmark material.

[0227] The subjective linear weighting model is calculated, and the attached Figure 11 to Figure 14 results are calculated using formula (21) to obtain the linear weighted results, as shown in the attached Figure 19 . The higher the total score of linear weighting indicates that the material is better.

[0228] From the linear weighting results, Cxx material has the highest total score, followed by Axx material, imported material, and Bxx material has the lowest total score.

[0229] Objective empowerment model calculation, for easy calculation, the data given according to the scheme statistics matrix. Because the density index after non-dimensionalization, all data are 1, therefore, not considered in the evaluation calculation range. In R language according to formula (22)-(25) defined a consistent function and entropy method to calculate the weight of the function for multiple, large number of calculations. Because of the calculation of entropy method function has the nature of conduction, in the use of entropy method to calculate the weight, the data can not appear 1 and 0, single data 0 value or error value will lead to the overall results returned failure. In the original data will be narrowed to 0.002-0.996 range.

[0230] In the use of entropy method to calculate the weight of each index, the entropy method function will return the score of each sample.

[0231] In three kinds of materials, all the negative indicators have been converted to positive indicators, so the score of all sample average can be used to represent the total score of the material, the material with the highest total score value as the final recommended material. The calculation process and results are shown in the attached Figure 20 -Appendix Figure 23 .

[0232] In the scoring results, Cxx material is the highest, Axx material second, Bxx material is the lowest.

[0233] Finally, the results are evaluated as follows:

[0234] (1) Axx material and the benchmark material comprehensive performance is closest, followed by Cxx material, Bxx material difference is the largest;

[0235] (2) product stability Cxx material is the best, Axx material second, Cxx material is the worst;

[0236] (3) Axx, Cxx material is better than the benchmark material, Cxx material comprehensive performance is better, Bxx worse than the benchmark material;

[0237] (4) three material sorting results for Cxx, Axx, Bxx.

Claims

1. An integrated evaluation method for aerospace equipment critical material application validation, characterized in that, It comprises the following steps: Step 1: Determine the object to be evaluated, and if multiple evaluation objects are included, the sampling, sample preparation, test method and test conditions of the multiple evaluation objects should all be the same; Step 2: Establish an evaluation index system, which includes first-level, second-level and third-level evaluation indexes. The first-level evaluation indexes include material basic performance, environmental adaptability, functional performance and process adaptability. The second-level evaluation indexes include mechanical properties, physical properties, electrical / thermal properties, welding properties, bonding process properties, weather resistance and vibration resistance. The third-level evaluation indexes include tensile strength, compressive strength, fatigue strength, density, electrical conductivity, dielectric constant, bonding strength, welding strength, high-temperature storage, damp heat aging and vibration test. After the index system is established, the corresponding detection data is ensured to have certain uniformity, and each evaluation index has the same data length and precision; Step 3: Determine the weight coefficients corresponding to each evaluation index, which include three levels of key, important and general; Step 4: Select or construct a comprehensive evaluation model, calculate the comprehensive evaluation value of each system and give the comprehensive evaluation result. According to the calculation result of the mathematical model, the comprehensive evaluation result or conclusion is given. The uniformity processing method of the evaluation index includes: Firstly, the evaluation index is defined as "max-type" index, "min-type" index, "intermediate-type" index and "interval-type" index. The max-type index means that the larger the value of the index is, the better it is. The min-type index means that the smaller the value of the index is, the better it is. The intermediate-type index means that the value of the index should not be too large or too small, and the appropriate intermediate value is the best. The interval-type index means that the value of the index should be within a certain determined interval. Secondly, according to the definition, the evaluation index is processed as a max-type index, and the calculation method is as follows: Min-type index: for a min-type index x, it is transformed by or transformed by Transforming the minimization index x into a maximization where M is the maximum value of the possible values of the index x, i.e. the index x is maximized; Intermediate-type index: for a intermediate-type index x, it is transformed by where M and m are the maximum and minimum values of the possible values of the index x, i.e. the intermediate index x can be maximized to ; Interval-type index: for a interval-type index x, it is transformed by where [a, b] is the best stable interval of the index x, M and m are the maximum and minimum values of the possible values of the index x, respectively; That is, the interval type index x can be maximized as .

2. A comprehensive evaluation method for validation of key materials for aerospace equipment applications according to claim 1, characterized in that, The uniformity processing of the evaluation index also includes dimensionless processing, and the specific method includes standard deviation method, extreme difference method or efficacy coefficient method.

3. A comprehensive evaluation method for validation of key materials for aerospace equipment applications as claimed in claim 2, wherein, The standard deviation method includes: Let wherein Clearly the mean and the mean square deviation of the indicator are 0 and 1, respectively, i.e. is a dimensionless indicator, called the x ij standardized observation.

4. The integrated evaluation method for the validation of key materials for aerospace equipment applications according to claim 2, characterized in that, The extreme difference method includes: Let , wherein then is a dimensionless index observation.

5. The integrated evaluation method for the validation of key materials for aerospace equipment applications according to claim 2, characterized in that, The efficacy coefficient method: Let where c, d are both determined constants. c represents "translation amount", d represents "rotation amount", i.e. "enlargement" or "reduction" multiple, then .

6. The integrated evaluation method for the validation of key materials for aerospace equipment applications according to claim 2, characterized in that, The determination method of the weight coefficient includes determining the weight coefficient based on the criticality of the evaluation index, and the specific method is as follows: Firstly, define the key index as three-star level, the important index as two-star level and the general index as one-star level. Divide the number of stars critical to the index, n, by the total number of stars As a weight coefficient of the corresponding index, the weight formula is calculated: ，（ i=1,2,...,m ）（15） wherein .

7. The integrated evaluation method for the validation of key materials for aerospace equipment applications according to claim 1, characterized in that, The comprehensive evaluation model includes T-test model, Euclidean distance model, subjective linear weighted evaluation model, Borda sorting model and objective weighting model. The T-test model takes each parameter as a single-dimensional sample, compares each selected parameter with the standard parameter using T-test, adopts single-sample non-parametric test and finally determines the comprehensive similarity according to the ranking of the similarity of each parameter. The specific test method is as follows: Single-sample mean T-test: test whether the average value of a single sample is equal to the target value. The test steps are as follows: First, the sample standard error Se is calculated: standard error Se= ( S (sample standard deviation) / ( (sample size) Next, the sample T value is calculated: where the sample mean is the mean of the test samples within a single index, the population mean m is the expected value, and Se is the standard error; Then, according to the T value, find the T table to get the P value; Finally, the obtained P value is compared with the significance level a, when P < a, it is considered that there is significant difference; when P > a, it is considered that there is no significant difference, when T > 0, it is verified that the mean value of the material index data is higher than that of the reference material, and when T < 0, it is lower, so four results of "positive equivalence", "negative equivalence", "positive significant difference" and "negative significant difference" can be obtained; Among them, "negative significant difference" can represent that the mean value of the reference material is used as the qualified criterion, and almost no qualified performance index of the verification material can be obtained no matter how the screening is strengthened, "positive significant difference" represents that almost all the qualified performance indexes can be obtained no matter how the screening is strengthened; "Positive equivalence" represents a higher probability of passing, and "negative significant" represents a lower probability of passing.

8. A comprehensive evaluation method for validation of key materials for aerospace equipment applications as claimed in claim 7, wherein, The Euclidean distance model is used to comprehensively evaluate the similarity between the verification material and the reference material, and the specific evaluation method is as follows: Assume the ideal point of the reference material index is For an object to be evaluated , define the weighted distance between the two as （18） wherein is a weight coefficient, is and a certain sense distance between Under normal circumstances, the simple Euclidean distance can be taken, that is, （19） Then the comprehensive evaluation function is （20） After calculation, according to The value of the size of the order of each evaluation scheme optimization, obviously the smaller the value of the program is better; in particular, when a certain time, that is, the ideal point is reached, the corresponding verification material is similar.

9. The integrated evaluation method for the validation of key materials for aerospace equipment applications according to claim 7, characterized in that, The objective weighting model is used for comparison when the weight is not easy to determine or the influence of the material verification index on the product type is not clear, and can also be used for verification of the evaluation results of other models, and the specific method includes: selecting i one material, j one index, x ij the value of the first i material for the first j index; Calculation Item j The first indicator i The proportion of each material in this indicator: ，i= 1 , 2 ,...,m （21） Computing j Entropy value of item indicator: （22） wherein k = 1 / ln(n) > 0 , satisfies e j ≤ 0。 Calculate the information entropy redundancy: d j =1 -e j （23） Calculate the weight of each option index: Calculate the comprehensive score of each material: The higher the score, the better the material.

Images