Method, application, system and program product for determining fusion threshold for new product reliability evaluation demand

Abstract

The invention relates to the field of reliability engineering, and relates to a method, application, system and program product for determining a fusion threshold for a new product reliability evaluation demand. The method comprises the following steps: constructing a feature set of a new product and an old product, and defining a similar element set; calculating the similarity and the weight of the similar elements, and calculating the similarity of new and old products based on the weight; establishing a minimum fusion cost mathematical model, and solving a fusion threshold value; and comparing the product similarity with a fusion threshold value, and when the similarity is greater than the fusion threshold value, performing fusion by using the data of the similar products so as to determine the range in which the similar products can be used. According to the method, the validity of the fusion data of the new product and the similar product is judged by a clear index, the credibility of the fusion data is improved, and the method can be widely applied to reliability analysis of complex products.

Description

Technical Field

[0001] The present invention relates to the field of reliability engineering, and in particular to a method, application, system and program product for determining a convergence threshold for reliability assessment of new products. Background Art

[0002] With the rapid development of industrial technology and the increasing complexity of products, the design update and iteration speed of new products has significantly accelerated. Especially in the fields of aerospace, automobile manufacturing, electronic equipment, etc., many new products have the characteristics of long life and high reliability, which makes the reliability field data that can be obtained in the actual working environment relatively scarce, resulting in the reliability data showing the characteristics of "multi-source and sparse". This data characteristic brings great challenges to the reliability evaluation of new products.

[0003] Traditional reliability assessment methods usually rely on complete experimental data or long-term field operation data of new products. However, in the absence of sufficient data, these methods cannot effectively evaluate the reliability of new products. To solve this problem, researchers began to try to use the data of similar products as a reference, by comparing the characteristics of new products with similar products, and fusing the data of similar products, thereby indirectly inferring the reliability of new products. However, existing methods have the following problems when fusing similar product data:

[0004] 1. Blindly use similar data: The definition of similar products is broad, and the degree of similarity between different products may range from 0.01 to 0.99. Product data with different degrees of similarity will have a great impact on reliability evaluation. Using similar product data may lead to distorted evaluation results.

[0005] 2. Lack of quantitative fusion mechanism: Existing methods lack a systematic quantitative model when fusing similar product data and are unable to effectively weigh the feature relationship between new products and similar products.

[0006] 3. The scope of similar products that can be used is unclear: During the fusion process, there are no clear indicators to judge the validity of similar product data, which limits the credibility of the fused data. Summary of the invention

[0007] In order to solve the above problems, the present invention proposes a method for determining a fusion threshold. The method quantifies the relationship between similar products and new products by constructing a mathematical model, calculates the similarity and solves the fusion threshold, thereby clarifying the range of similar products that can be used, and effectively solving the problem of similar product data abuse in a multi-source sparse data environment.

[0008] In order to achieve the above-mentioned purpose, the present invention adopts the following technical solutions:

[0009] A method for determining a range of similar products that can be used by using a convergence threshold, the method comprising the following steps:

[0010] 1) Construct feature sets for new and old products, and group features with the same physical meaning into similar element sets;

[0011] 2) Calculate the similarity and weight of similar elements based on the quantity similarity Q n And the numerical similarity Q u Calculate the similarity Q between new and old products;

[0012] 3) Establish a mathematical model for solving the fusion threshold; the mathematical model is as follows:

[0013]

[0014] Where φ: the value of the objective function, which represents the sum of the fusion costs, and the goal is to minimize it; let N = {1,2}, M = {1,2,…,n s}; D = {d i |i∈N} is the set of original products and new products; V={v j |j∈M} is the set of product similar elements; η ij For product d i In similar element v j The eigenvalues ​​on ; is the adjusted characteristic value of the product; is the characteristic value of the integrated product; c j is the unit feature deviation cost when the similar element feature value changes; ε j is the allowable range of characteristic value deviation between the characteristic value of the integrated product and the characteristic value of the adjusted product; ω j is the weight of the similar element in the population; γ is the fusion threshold, which indicates the minimum utility value required for similar product data to be fused and used; δ is a minimum constant to avoid the denominator being zero;

[0015] 4) Solve and verify the mathematical model to obtain the fusion threshold γ, compare the relationship between the similarity Q of the product and the fusion threshold γ, and when the similarity Q is greater than the fusion threshold γ, the data of the similar product can be effectively used.

[0016] Preferably, the step 1) comprises the following steps:

[0017] a) Let the margin equation of the system be m=g(z1,z2,…,z ns ), each parameter in the margin equation is used as a characteristic element to form a characteristic set;

[0018] b) Define the feature set of the new product as: A = {z1, z2, ..., z n}

[0019] c) The feature set of the old product is defined as: A′={z1′,z2′,...,z n′ ′}

[0020] d) The characteristics with the same physical meaning in the two characteristic sets are combined to form similar elements, and the similar element set is set as: U = {u1, u2, .. ., u ns}.

[0021] Preferably, the calculation formula of the similarity of similar elements in step 2) is as follows:

[0022]

[0023] The relative sensitivity calculation formula is as follows:

[0024]

[0025] The relative sensitivity weight calculation formula is as follows:

[0026]

[0027] The formula for calculating product similarity is as follows:

[0028]

[0029] Preferably, the objective function minΦ in the mathematical model is replaced by To determine the optimal threshold, set the objective function to:

[0030]

[0031] The constraints are:

[0032]

[0033] Where: x i and i is the known data of new products and old products in similar elements; a i and b i are the adjusted data respectively; η i is the data of the integrated product, ω1 and ω2 are the weights of the new and old products in the integrated product, c i and d i For x i and i The weight of the performance margin m.

[0034] Preferably, the setting ω1=ω2=0.5;

[0035]

[0036] Furthermore, the present invention also provides a method for evaluating whether data of similar products can be used in a sparse data environment.

[0037] Furthermore, the present invention also provides a verification system for solving the fusion threshold and whether similar product data can be used, and the system implements the method described, including:

[0038] Feature set building module, used to extract characteristic parameters from new and old products; combining characteristics with the same physical meaning into similar element sets;

[0039] The similarity and weight calculation module is used to: calculate the similarity of each similar element; calculate the overall similarity Q between the new product and the old product based on the similar element characteristic parameters and their weights;

[0040] Mathematical modeling module, used to build a mathematical model for solving the fusion threshold,

[0041] The solution and verification module is used to solve the mathematical model and obtain the optimal fusion threshold γ; compare the size relationship between the product similarity Q and the fusion threshold γ; when Q>γ, the data of similar products can be used for fusion;

[0042] The estimation and output module is used to output the results; display the similarity Q, the fusion threshold γ and the characteristic parameters of the adjusted data.

[0043] Furthermore, the present invention also provides a computer-readable storage medium having a computer program or instruction stored thereon, and the method is implemented when the computer program or instruction is executed by a processor.

[0044] Furthermore, the present invention also provides a computer program product, comprising a computer program or instructions, which implement the method when executed by a processor.

[0045] The present invention adopts the above technical solution, quantifies the similarity and introduces the concept of integration utility function to solve the integration threshold, thereby determining the range of similar products that can be used.

[0046] In addition, the present invention proposes a mathematical model of minimum fusion cost based on the objective function, clarifies key parameters such as similar element eigenvalue deviation, weight distribution, and fusion utility, and solves the optimal fusion threshold through mathematical optimization methods, providing a quantitative theoretical basis for reliability assessment needs.

[0047] Furthermore, the present invention is suitable for determining the scope of use of similar products, especially when there are design improvements or generational changes between new products and similar products. It can flexibly adapt to different product characteristics and application scenarios, and provide a general evaluation framework for multi-field engineering applications. By minimizing the fusion cost objective function, the present invention optimizes the resource utilization efficiency in the fusion process of new and old product data while ensuring the accuracy of similar fusion, and improves the accuracy and reliability of the collected data.

[0048] Through the above technical effects, the present invention solves the problem of insufficient accuracy of traditional methods using similar product data in sparse data environments, and provides an efficient, accurate and scientific prior condition for the reliability assessment of complex products, which has important theoretical value and practical application significance. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 The distribution of new and old products and the distribution of integrated products.

[0050] Figure 2 This is a schematic diagram of a seed source circuit of an application example of the present invention. DETAILED DESCRIPTION

[0051] The following is a clear and complete description of the technical solutions in the embodiments of the present invention in conjunction with the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0052] First, the present invention proposes a method for solving the fusion threshold: establishing a minimum fusion cost model to solve the fusion threshold. s similar elements, let N = {1,2}, M = {1,2,…,n s}, let the set D = {d i |i∈N} is the set of original products and new products, V={v j |j∈M} is the set of product similar elements, η ij For product d i In similar element v j The eigenvalues ​​on is the characteristic value after product adjustment, is the characteristic value of the blended product, c j The unit cost generated when the characteristic value of the similar element changes is set to an acceptable ε j In, ω j As the weight of the similar element in the population, the following mathematical model can be established:

[0053]

[0054] The integration utility function is introduced. The integration utility function reflects the closeness between the similarity parameters of the new and old products after adjustment and the similarity parameters of the integrated products. A threshold γ is set. According to the relationship between the integration utility function and the threshold, it is judged whether the expected integration is achieved. The integration utility function is defined as:

[0055]

[0056] Here, δ=0.0001 is set to avoid the situation where the denominator in the above formula is 0.

[0057] The soft consistent weighted fusion model based on fusion utility constraints is obtained as follows:

[0058]

[0059] Based on the above model, the present invention further provides a method for solving the fusion threshold and determining the range of similar products that can be used, the method comprising the following steps:

[0060] 1) Construct feature sets for new and old products, and group features with the same physical meaning into similar element sets;

[0061] 2) Calculate the similarity and weight of similar elements, and calculate the similarity Q of new and old products based on the feature values ​​and weights;

[0062] 3) Establishing a mathematical model for solving the fusion threshold; the mathematical model is shown above;

[0063] 4) Based on the case, the mathematical model is solved and verified to obtain the fusion threshold γ. The relationship between the similarity Q of the products and the fusion threshold γ is compared. When the similarity Q is greater than the fusion threshold γ, similar product data can be used for fusion.

[0064] Wherein: step 1) includes the following steps:

[0065] a) Write the margin equation of the system as A feature set is formed by taking each parameter in the margin equation as a feature element;

[0066] b) Define the feature set of the new product as: A = {z1, z2, ..., z n};

[0067] c) The feature set of the old product is defined as: A′={z1′,z2′,...,z n′ '};

[0068] d) The characteristics with the same physical meaning in the two characteristic sets are combined to form similar elements, and the similar element set is set as:

[0069] The calculation formula of similarity of similar elements in step 2) is as follows:

[0070]

[0071] The relative sensitivity calculation formula is as follows:

[0072]

[0073] The relative sensitivity weight calculation formula is as follows:

[0074]

[0075] The formula for calculating product similarity is as follows:

[0076]

[0077] Furthermore, in step 3), the objective function is a linear algorithm, which means that the data deviation cost between the fused product and the new and old products is the smallest. However, in the actual working environment, it is difficult for the data of the new and old products to be completely consistent with the fused product. Therefore, it is only necessary to adjust the data of the new and old products a i and b i Integration with product data i The deviation is set within an acceptable range of ε i Within, set the integration threshold γ, requiring the integration utility function to be no less than the integration threshold. Since the value of γ is usually unknown, and it is necessary to find a threshold that minimizes the unit threshold cost, the objective function minΦ can be simply replaced by To determine the optimal threshold, the objective function can be set as:

[0078]

[0079] The constraints are:

[0080]

[0081] x i and i is the known data of new products and old products in similar elements; a i and b i are the adjusted data of each; η is the data of the integrated product, ω1 and ω2 are the weights of the new and old products in the integrated product, and here we simply set ω1=ω2=0.5; c i and d i For x i andi The weight of the performance margin m.

[0082] in,

[0083]

[0084] In the present invention, sparse data of a seed source circuit of a laser generator is used to solve the fusion threshold. The schematic diagram of a typical seed source circuit is shown in FIG. Figure 2 The margin equation is:

[0085]

[0086] Wherein k1 = R3 / (R3+R5), k2 = R6 / (R6+R8).

[0087] Table 1 Variable distribution parameter table

[0088] Variable Name Mean Standard Deviation Variable Name Mean Standard Deviation <![CDATA[R SENSE ]]> 0.0936Ω 3e-4Ω <![CDATA[R′ SENSE ]]> 0.0923Ω 3e-4Ω <![CDATA[R3]]> 1203.0Ω 0.24Ω <![CDATA[R′3]]> 1200.8Ω 0.24Ω <![CDATA[R5]]> 5.470kΩ 48Ω <![CDATA[R′5]]> 5.459kΩ 45Ω <![CDATA[R6]]> 1.510kΩ 7Ω <![CDATA[R′6]]> 1.498kΩ 9Ω <![CDATA[R7,R8]]> 1.312kΩ 12.4Ω <![CDATA[R′7,R′8]]> 1.296kΩ 11.8Ω <![CDATA[R9,R 10 ]]> 2.63kΩ 2.8Ω <![CDATA[R′9,R′ 10 ]]> 2.6kΩ 3.2Ω <![CDATA[U SEED ]]> 2258mV 31mV <![CDATA[U′ SEED ]]> 2200mV 29mV <![CDATA[U P ]]> 452mV 0.4mV <![CDATA[U′ P ]]> 459mV 0.5mV <![CDATA[U N ]]> 45mV 4.5mV <![CDATA[U′ N ]]> 46mV 5.2mV

[0089] According to the similarity calculation formula of similar elements, the similarity of each similar element is calculated as shown in the following table

[0090] Table 2 Similarity calculation results of similar elements

[0091] <![CDATA[Similar element (u i )]]> <![CDATA[Similarity q(u i )]]> <![CDATA[Similar element (u i )]]> <![CDATA[Similarity q(u i )]]> <![CDATA[Similar element (u i )]]> <![CDATA[Similarity q(u i )]]> <![CDATA[u1]]> 0.9930 <![CDATA[u2]]> 0.9991 <![CDATA[u3]]> 0.9990 <![CDATA[u4]]> 0.9960 <![CDATA[u5]]> 0.9939 <![CDATA[u6]]> 0.9939 <![CDATA[u7]]> 0.9943 <![CDATA[u8]]> 0.9943 <![CDATA[u9]]> 0.9870 <![CDATA[u 10 ]]> 0.9923 <![CDATA[u 11 ]]> 0.9890 <![CDATA[u 12 ]]> 1

[0092] According to the calculation formula of relative sensitivity and relative sensitivity weight, the calculation results of the two are as follows:

[0093] Table 3 Relative sensitivity calculation results

[0094] Variable Name <![CDATA[S1]]> <![CDATA[S2]]> <![CDATA[S3]]> <![CDATA[S4]]> Numeric 4.6935 3.8473 3.8473 2.4452 Variable Name <![CDATA[S5]]> <![CDATA[S6]]> <![CDATA[S7]]> <![CDATA[S8]]> Numeric 0.9971 2.4452 1.2099 2.2069 Variable Name <![CDATA[S9]]> <![CDATA[S 10 ]]> <![CDATA[S 11 ]]> <![CDATA[S 12 ]]> Numeric 3.4168 1.4301 0.1535 3.6935

[0095] Table 4 Calculation results of relative sensitivity weights

[0096] Variable Name <![CDATA[ω1]]> <![CDATA[ω2]]> <![CDATA[ω3]]> <![CDATA[ω4]]> Numeric 0.1545 0.1266 0.1266 0.0805 Variable Name <![CDATA[ω5]]> <![CDATA[ω6]]> <![CDATA[ω7]]> <![CDATA[ω8]]> Numeric 0.0328 0.0805 0.0398 0.0726 Variable Name <![CDATA[ω9]]> <![CDATA[ω 10 ]]> <![CDATA[ω 11 ]]> <![CDATA[ω 12 ]]> Numeric 0.1124 0.0471 0.0051 0.1216

[0097] According to the mean data given in the case and the calculation formula proposed in the previous article, the product similarity Q = 0.9952 is obtained.

[0098] The mean data is also used to verify the proposed mathematical model, that is, x = [0.0936, 1203, 5470, 1510, 1312, 1312, 2630, 2630, 2.258, 0.452, 0.045, 2.5]; y = [0.0923, 1200.8, 5459, 1498, 1296, 1296, 2600, 2600, 2.2, 0.459, 0.046, 2.5].

[0099] The above data and mathematical model are encoded in MATLAB. Since γ is unknown, the traversal method is used here to solve the γ value. The range of γ is set to 0 to 1, and the step size is 0.01. After solving, the minimum objective function value is 8.24×10 5 , the corresponding γ value is 0.92, a i and b i as follows:

[0100] Table 5 Results of mathematical model under original data

[0101] i <![CDATA[a i ]]> <![CDATA[b i ]]> 1 0.093127 0.091834 2 1209.076 1206.865 3 5442.374 5431.429 4 1502.374 1505.566 5 1305.374 1302.545 6 1305.374 1302.545 7 2616.717 2613.131 8 2616.717 2613.131 9 2.223788 2.211111 10 0.454283 0.456682 11 0.044773 0.045768 12 2.487374 2.487374

[0102] It is found that the product similarity Q = 0.9952> γ = 0.92, that is, the data of existing new products and similar products can be used for fusion analysis to obtain the reliability evaluation results of new products.

[0103] The above is a description of the embodiments of the present invention. Through the above description of the disclosed embodiments, professionals and technicians in the field can implement or use the present invention. Various modifications to these embodiments will be apparent to professionals and technicians in the field. The general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to these embodiments shown in this article, but will conform to the widest range consistent with the principles and novelties disclosed herein.

Claims

1. A method for solving a fusion threshold and clarifying a range of similar products that can be used, the method comprising the following steps: 1) Construct feature sets for new and old products, and group features with the same physical meaning into similar element sets; 2) Calculate the similarity and weight of similar elements, and calculate the similarity Q of new and old products based on the feature values ​​and weights; 3) Establish a mathematical model for solving the fusion threshold; the mathematical model is as follows: Where φ: the value of the objective function, which represents the sum of the fusion costs, and the goal is to minimize it; there are n new and old products. s similar elements; let N = {1,2}, M = {1,2,…,n s }; D = {d i |i∈N} is the set of original products and new products; V={v j |j∈M} is the set of product similar elements; η ij For product d i In similar element v j The eigenvalues ​​on ; is the adjusted characteristic value of the product; is the characteristic value of the integrated product; c j is the unit feature deviation cost when the similar element feature value changes; ε j is the allowable range of characteristic value deviation between the characteristic value of the integrated product and the characteristic value of the adjusted product; ω j is the weight of the similar element in the population; γ is the fusion threshold, which indicates the minimum utility value required for similar product data to be fused and used; δ is a minimum constant to avoid the denominator being zero; 4) Solve and verify the mathematical model, obtain the fusion threshold γ, compare the size relationship between the product similarity Q and the fusion threshold γ, when the similarity Q is greater than the fusion threshold γ, the data of the similar product can be used to obtain the fusion product.

2. The method according to claim 1, characterized in that Step 1) includes the following steps: a) Let the margin equation of the system be m=g(z1,z2,…,z ns ), each parameter in the margin equation is used as a characteristic element to form a characteristic set; b) Define the feature set of the new product as: A = {z1, z2, …, z n }; c) The feature set of the old product is defined as: A′={z1′,z2′,...,z n″ }; d) The characteristics with the same physical meaning in the two characteristic sets are combined to form similar elements, and the similar element set is set as:

3. The method according to claim 2, characterized in that The calculation formula of similarity of similar elements in step 2) is as follows: The relative sensitivity calculation formula is as follows: The relative sensitivity weight calculation formula is as follows: The formula for calculating product similarity is as follows:

4. The method according to claim 3, characterized in that The objective function minΦ is replaced by To determine the optimal threshold, set the objective function to: The constraints are: Where: x i and i is the known data of new products and old products in similar elements; a i and b i are the adjusted data respectively; η i is the data of the integrated product, ω1 and ω2 are the weights of the new and old products in the integrated product, c i and d i For x i and i The weight of the performance margin m.

5. The method according to claim 4, characterized in that Set ω1=ω2=0.

5.

6. The method according to claim 4, characterized in that 7. The method described in any one of claims 1 to 6 is used to solve the product similarity and the fusion threshold of the fusion of new and old products.

8. A system for determining the integration threshold and whether similar product data can be used, characterized in that: The system implements the method described in any one of claims 1 to 6, including: a) Feature set construction module, used to extract characteristic parameters from new and old products; combining characteristics with the same physical meaning into similar element sets; b) Similarity and weight calculation module, used to: calculate the similarity of each similar element; calculate the overall similarity Q between the new product and the old product based on the similar element characteristic parameters and their weights; c) Mathematical modeling module, used to construct a mathematical model for solving the fusion threshold, d) Solving and verification module, used to solve the mathematical model and obtain the optimal fusion threshold γ; compare the size relationship between product similarity Q and fusion threshold γ; when Q>γ, the data of similar products can be used for fusion; e) Estimation and output module, used to verify the results; display the similarity Q, fusion threshold γ and characteristic parameters of the adjusted data.

9. A computer-readable storage medium having a computer program or instruction stored thereon, characterized in that: When the computer program or instruction is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

10. A computer program product comprising a computer program or instructions, characterized in that When the computer program or instruction is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

Images