Improved FMEA method based on double-layer hesitant triangular fuzzy semantic set and fuzzy exp-TODIM

Abstract

The invention discloses an improved FMEA method based on a double-layer hesitant triangular fuzzy semantic set and fuzzy exp-TODIM, and the method comprises the steps: determining n potential failure modes of an analysis object, and constructing a failure mode set; risk factors in the FMEA method are improved to be severity, occurrence degree, detection degree, repair degree and economic degree; establishing a double-layer hesitant triangular fuzzy semantic set; obtaining an expert weight through a consensus decision algorithm based on the social network trust degree and the fuzzy preference relationship consistency; defuzzifying the fuzzy failure mode evaluation matrix in two steps to respectively obtain a semi-fuzzy failure mode evaluation matrix and a clear failure mode evaluation matrix; using an MEREC method for the clear failure mode evaluation matrix to obtain a risk factor weight; and for the semi-fuzzy failure mode evaluation matrix, using a fuzzy exp-TODIM method based on a generalized mixed Hausdorff preference distance to sort the failure modes. According to the method, the risk severity of each failure mode of the IMS is determined through the reliability and risk evaluation analysis of the IMS, and the effectiveness and applicability of the improved FMEA method are verified.

Description

Technical Field

[0001] The invention relates to the technical field of reliability analysis, and in particular to an improved FMEA method based on a double-layer hesitant triangular fuzzy semantic set and fuzzy exp-TODIM. Background Art

[0002] Failure Modes and Effect Analysis (FMEA) is a forward-looking, systematic reliability analysis and safety assessment technology that can be used to identify potential technical risks in various aspects of systems, products, processes and services, evaluate the impact of different failure modes, and help improve the safety and reliability of the system. Since FMEA was officially adopted by the US Air Force, it has been widely used in aerospace, machinery manufacturing, medical industry, nuclear industry and other fields.

[0003] The implementation steps of traditional FMEA are generally as follows: (1) Experts (usually engineers or other relevant personnel) list potential failure modes and their potential impacts and potential causes based on their professional knowledge and experience of products and processes. (2) FMEA has three risk factors, namely severity (S), occurrence (Occurrence, O), and detection (D). Industry experts evaluate the levels of these three risk factors for each failure mode. (3) Calculate the risk priority number (RPN) and rank the failure modes according to the RPN value. (4) Managers propose improvement plans or corrective measures based on the ranking results. However, in practical applications, traditional FMEA still has obvious shortcomings, such as: (1) The relative importance between occurrence, severity and detection is not reflected; (2) The rationality of the RPN formula is questionable, and it is difficult to make an accurate assessment of risks based on the RPN value alone; (3) In reality, it is difficult for experts to accurately quantify their evaluation into an exact value. Summary of the invention

[0004] In view of the shortcomings of the prior art, the purpose of the present invention is to provide an improved FMEA method based on a double-layer hesitant triangular fuzzy semantic set and fuzzy exp-TODIM. Through the reliability and risk evaluation analysis of the IMS system, the risk severity of each failure mode of the IMS system is determined, and the effectiveness and applicability of the improved FMEA method are verified. In order to achieve the above-mentioned purpose and other advantages according to the present invention, an improved FMEA method based on a double-layer hesitant triangular fuzzy semantic set and fuzzy exp-TODIM is provided, comprising:

[0005] S1. Set up a FMEA team to analyze the operation process of the object;

[0006] S2, determine n potential failure modes of the analysis object and construct a failure mode set;

[0007] S3. Improve the risk factors in the traditional FMEA method to include severity, occurrence, detection, repair and economy;

[0008] S4, establishing a double-layer hesitant triangular fuzzy semantic set, including one-to-one correspondence between the double-layer hesitant triangular fuzzy language elements and the evaluation language, constructing corresponding evaluation criteria according to the double-layer hesitant triangular fuzzy semantic set and the double-layer hesitant triangular fuzzy language elements corresponding to the language variables, and converting the expert scoring table into a scoring matrix of L experts;

[0009] S5, obtaining expert weights through a consensus decision-making algorithm based on social network trust and consistency of fuzzy preference relations;

[0010] S6. According to the expert weights, the scoring matrices of the experts are weighted averaged to obtain a fuzzy failure mode evaluation matrix;

[0011] S7, defuzzifying the fuzzy failure mode evaluation matrix in two steps to obtain a semi-fuzzy failure mode evaluation matrix and a definite failure mode evaluation matrix;

[0012] S8. Use the MEREC (Method based on the Removal Effects of Criteria) method on the explicit failure mode evaluation matrix to obtain the risk factor weights;

[0013] S9. The semi-fuzzy failure mode evaluation matrix is ​​used to rank the failure modes using the fuzzy exp-TODIM method based on the generalized mixed Hausdorff preference distance, and relevant personnel formulate corresponding prevention and response measures based on the ranking.

[0014] Preferably, the establishment of the double-layer hesitant triangular fuzzy semantic set in step 4 includes the following steps:

[0015] S41. Define a bilingual terminology set;

[0016] S42. Define a set of double-layer hesitant fuzzy language terms;

[0017] S43, define triangular fuzzy number, fuzzy number It can be obtained by (a L , a M , a U ) determines (0≤a L ≤a M ≤a U ≤1), and the membership function (or characteristic function) is:

[0018]

[0019] Then it is called is a triangular fuzzy number; where a L is the most conservative or pessimistic estimate, a M is the most likely estimate, a U is the most optimistic evaluation value;

[0020] S44. Define a double-layer hesitant triangle fuzzy language term set.

[0021] Preferably, step S41 includes setting S={S t |t=--τ,…,-1,0,1,…,τ} and are two finite, completely ordered discrete language term sets, S and O are the first-level and second-level language term sets respectively and are independent of each other, then the two-level language term set in It is a two-layer language term, s t Indicates the first level language, o k Indicates the second level of language;

[0022] Preferably, step S42 includes setting X to be a non-empty set, based on the two-layer language term set S o Double-layer hesitant fuzzy language term set in,

[0023]

[0024]

[0025] Indicates S o A continuous list of possible values ​​in Embodying Elements x i Belongs to S o The membership degree of is a double-layer hesitant fuzzy linguistic element, L i express The number of language terms in ; Φ×Ψ represents the set of double-layer hesitant fuzzy language elements;

[0026] Preferably, step S44 includes assuming that X is a non-empty set, is a two-level language term set, then S o Double-layer hesitant triangle fuzzy language term set It is expressed as: in, is a three-element set: in All three elements in the set belong to S o , respectively represent x iIn the most pessimistic, most likely and most optimistic scenarios, is a double-level hesitant triangular fuzzy language element, T(Φ×Ψ) represents the set of double-level hesitant triangular fuzzy language elements;

[0027] Preferably, the evaluation criteria construction process based on the double-layer hesitant triangular fuzzy semantic set in step 4 is as follows:

[0028] S45. Give scoring examples and explanation of their significance: Under each failure mode, the scoring examples for a single risk factor are:

[0029] [(x1, Y1), (x2, Y2), (x3, y3)]

[0030] The scoring of individual risk factors is divided into three parts:

[0031] The first part: (x1, y1), represents the minimum rating, that is, the minimum rating of a very large indicator or the maximum rating of a very small indicator.

[0032] The second part: (x2, y2), represents the most likely rating,

[0033] The third part: (x3, y3), represents the highest rating, that is, the maximum rating for very large indicators or the minimum rating for very small indicators;

[0034] S46. Give an explanation of each part of the scoring: (x i ,y i ) is a two-level description of risk factors, dividing an exact value into “integer part (x i )" and "decimal part (y i )" two layers, x i +y i To provide practical evaluation criteria;

[0035] x i is the first layer, the integer part, the value range is {-2, -1, 0, 1, 2}, representing {very low (VL), low (L), medium (M), high (H), very high (VH)};

[0036] y i For the second level, the decimal part, as a pair of x i The supplement (adjective) of They represent {much lower, a little lower, exactly the same, a little higher, much higher} respectively;

[0037] S47. Give additional explanation: x i +y i For a more practical evaluation standard, (x i ,y i) form; secondly, x i +y i The value range is [-2, 2]. i =2(x i =-2), y i can only take non-positive (non-negative) numbers; finally, in the evaluation criteria, x i Corresponding to the φ in the double-layer hesitant triangular fuzzy language element l ,y i Corresponding to the double-layer hesitant triangle fuzzy language element because but

[0038] Preferably, the defuzzification method of the fuzzy failure mode evaluation matrix in step 7 is as follows:

[0039] S71. For each double-layer hesitant triangle fuzzy language element Elements in The double-level hesitant fuzzy linguistic element defuzzification method is used to transform the double-level hesitant triangular fuzzy linguistic element into triangular fuzzy number.

[0040] S72. Defuzzify the obtained triangular fuzzy number by using the K-th power weighted average method to obtain an accurate number.

[0041] Preferably, the fuzzy exp-TODIM method based on generalized Hausdorff preference distance in step 9 comprises the following steps:

[0042] S91, standardize the semi-fuzzy failure mode evaluation matrix using a linear standardization method to obtain a standardized matrix;

[0043] S92, under the kth risk factor, failure mode FM i With FM j The corresponding double-layer hesitant triangular fuzzy language elements are: Then a ik with a jk The generalized Hausdorff preference distance between is:

[0044]

[0045] in Pick

[0046] S93, under the kth risk factor, failure mode FM i (i=1,2,...,n) The exponential prospect function value relative to all other failure modes is

[0047]

[0048] where j = 1, 2, ..., n, j ≠ i; w k is the weight of the kth risk factor; for a ik with a jk The generalized Hausdorff preference distance between them; λ is the amplification parameter for adjusting the gain and loss effects to ensure that the impact of loss is greater than the gain; ρ is used to indicate the degree of influence of a single evaluation gap on the dominance;

[0049] S94, Calculate failure mode FM i Dominance relative to all other failure modes

[0050] S95, Calculate failure mode FM i The overall advantage

[0051] S96, according to ξ i The failure modes are sorted in order of decreasing value.

[0052] Compared with the prior art, the present invention has the following beneficial effects: an evaluation standard derived from the double-layer hesitant triangular fuzzy semantic set theory is used to score each risk factor of the failure mode, so as to solve the problem that the risk factor evaluation is difficult to quantify; secondly, the MEREC method is used to process the scoring matrix to obtain the relative weight of each risk factor, so as to solve the problem that the relative importance of risk factors is not considered; finally, the exp-TODIM method based on the generalized mixed Hausdorff preference distance is used to rank the failure modes, and then the influence of each failure mode on the system is evaluated. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 A flowchart of an improved FMEA method based on a double-layer hesitant triangular fuzzy semantic set and fuzzy exp-TODIM according to the present invention;

[0054] Figure 2 A triangular fuzzy number membership value function diagram of an improved FMEA method based on a double-layer hesitant triangular fuzzy semantic set and fuzzy exp-TODIM according to the present invention;

[0055] Figure 3 It is a comparison diagram of the ranking of failure modes of the IMS system according to the improved FMEA method based on the double-layer hesitant triangle fuzzy semantic set and fuzzy exp-TODIM of the present invention. DETAILED DESCRIPTION

[0056] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in 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.

[0057] Reference Figure 1-3 , an improved FMEA method based on double-layer hesitant triangular fuzzy semantic sets and fuzzy exp-TODIM, including:

[0058] S1. Set up a FMEA team to analyze the operation process of the object;

[0059] S2. The team determines n potential failure modes of the analysis object through literature review, brainstorming, expert interviews, etc.: FM1, FM2, ..., FM n , construct a set of failure modes;

[0060] S3. The team improved the risk factors in the traditional FMEA method through literature review, brainstorming, expert interviews, etc., adding two risk factors, namely repairability and economy, and expanding the original risk factors into five risk factors, namely severity (S), occurrence (Occurrence, O), detection (D), repairability (R) and economy (Economy, E); Severity S indicates the impact of the failure mode on the system, modules and customers; Occurrence O indicates the possibility of the failure mode; Detection D indicates the difficulty of the failure mode being discovered; Repairability R indicates the difficulty of repairing the failure mode; Economy E indicates the degree of saving of funds, manpower, etc. in the process of fault detection and maintenance;

[0061] S4. Establish a double-layer hesitant triangle fuzzy semantic set, correspond the evaluation language to the double-layer hesitant triangle fuzzy language element, the first-layer and second-layer language variable evaluation criteria and corresponding values ​​are shown in Table 1 and Table 2. The corresponding values ​​of the first-layer evaluation terms are [-2, -1, 0, 1, 2], and the corresponding values ​​of the second-layer evaluation terms are [-2, -1, 0, 1, 2]. The team constructs the corresponding evaluation criteria based on the double-layer hesitant triangle fuzzy semantic set. The team invites L relevant experts to score the risk factors of each failure mode using the evaluation criteria constructed by the team, and obtains the expert scoring table. Based on the double-layer hesitant triangle fuzzy language elements corresponding to the language variables, the expert scoring table is converted into a scoring matrix of L experts.

[0062] Table 1 Evaluation criteria and corresponding values ​​of the first-level language variables

[0063]

[0064] Table 2 Second level language variables and corresponding values

[0065]

[0066] S5. Use the consensus decision-making algorithm based on social network trust and fuzzy preference relationship consistency to obtain the expert weight w e ;

[0067] S6. According to the obtained expert weight w e The fuzzy failure mode evaluation matrix A is obtained by weighted averaging the scoring matrix of L experts. n*m , n is the number of failure modes, m is the number of risk factors, and a in the matrix ij They are all double-layer hesitant triangular fuzzy language elements:

[0068] S7. Fuzzy failure mode evaluation matrix A n*m Defuzzification is performed in two steps. First, A n*m Each element a in ij For each element in, the following defuzzification formula is used to perform double-layer hesitant fuzzy defuzzification to obtain the semi-fuzzy failure mode evaluation matrix A′ n*m , where the elements

[0069]

[0070] Secondly, we use the K-power weighted average method to convert a′ ij Consider it as a triangular fuzzy number defuzzified to get an exact number: After simplification, we get: Where Crisp(a′ ij ) is a′ ij The exact value after defuzzification, μ(x) is the membership value function, and k is the power. The obtained exact number constitutes the clear failure mode evaluation matrix B n*m ;

[0071] S8. Clear failure mode evaluation matrix B n*m Using the MEREC method, we get the weight w of the kth risk factor: k ;

[0072] ① Standardize the failure mode evaluation matrix B according to the following formula n*m :

[0073]

[0074] Get the standardized decision matrix N;

[0075] ②Calculate the overall performance S for each failure mode i :According to the normalized matrix N obtained in the above steps, combined with the nonlinear function f(x)=ln(1+|ln(x)|), a smaller Get greater performance S i ; The calculation formula for the overall performance of the failure mode is:

[0076] ③ Calculate the remaining performance of each failure mode by deleting any standard: For each failure mode, by reducing a certain evaluation index, calculate the remaining performance of the failure mode under the remaining evaluation index to evaluate the importance of the reduced index. ij It represents the residual performance of the failure mode when the i-th failure mode lacks the j-th evaluation index. The calculation formula is:

[0077] ④ Calculate the total deviation of each failure mode and normalize it: The total deviation calculation formula is: E j =∑ i |S′ ij -S i |; Normalize the total deviation to get the final result;

[0078] S9. Semi-fuzzy failure mode evaluation matrix A′ n*m The failure modes were ranked using the fuzzy exp-TODIM method based on generalized mixed Hausdorff preference distance, and relevant personnel formulated corresponding prevention and response measures based on the ranking.

[0079] ①Use the linear normalization method to transform the semi-fuzzy failure mode evaluation matrix A′ n*m For standardization, for extremely large indicators (severity, occurrence, detection, and repair), the standardization formula is: For the very small indicator (economy), the standardized formula is: Get the standardized matrix;

[0080] ② Calculate the failure mode FM under the kth risk factor i With FM j The generalized mixed Hausdorff preference distance between the corresponding two-level hesitant triangle fuzzy linguistic elements, failure mode FM i With FM j The double-layer hesitant triangular fuzzy linguistic elements are: Then a ik with a jk The generalized Hausdorff preference distance between is:

[0081]

[0082] in

[0083] ③Calculate the failure mode FM under the kth risk factor i (i=1,2…,n) relative to the exponential prospect function value of all other failure modes; under the kth risk factor, the failure mode FM i The exponential prospect function value relative to all other failure modes is:

[0084]

[0085] where j = 1, 2, ..., n, j ≠ i; w k is the weight of the k-th risk factor obtained in step 8 in the improved FMEA method based on a double-layer hesitant triangular fuzzy semantic set and a fuzzy exp-TODIM method of claim 1; for a ik with a jk The generalized Hausdorff preference distance between them; λ is an amplification parameter usually used to adjust the different effects of gain and loss to ensure that the impact of loss is greater than that of gain; ρ is used to reflect the degree of influence of a single evaluation gap on dominance;

[0086] ④ Calculate failure mode FM i Dominance relative to all other failure modes; Failure Mode FM i The advantage relative to all other failure modes is:

[0087] ⑤ Calculate failure mode FM i The overall superiority of:

[0088] ⑥According to ξ i The failure modes are sorted in order of decreasing value.

[0089] Example 1

[0090] In this embodiment, the FMEA team is composed of university scholar A, IMS information project senior manager B, IMS information consulting expert C and IMS implementation consultant D. After the team literature review, brainstorming, and expert interviews, considering the five hierarchical architectures of the intelligent manufacturing system (collaboration layer, enterprise layer, manufacturing execution system (MES), factory control layer, equipment layer and other possible failures such as natural disasters), the IMS system failure modes are determined to be 29. At the same time, for the convenience of expression, the above failure modes are numbered FM i (i=1, 2, ..., 29) as shown in Table 3.

[0091] Table 3 IMS failure mode scoring table

[0092]

[0093]

[0094]

[0095] According to the evaluation criteria reflected in Table 1 and Table 2, the evaluation experts scored the above 29 failure modes and obtained an expert scoring table. According to the double-layer hesitant triangular fuzzy linguistic elements corresponding to the linguistic variables, the expert scoring table was converted into an expert scoring matrix. The scoring matrix of expert A for the first five failure modes is shown in Table 4.

[0096] Table 4 Scoring matrix of university scholar A on IMS failure modes (top 5 failure modes)

[0097]

[0098] The expert weight is obtained using a consensus decision-making algorithm based on social network trust and fuzzy preference relationship consistency. The expert weight w e =[0.164721, 0.215103, 0.263050, 0.357126].

[0099] The fuzzy failure mode evaluation matrix A is obtained by weighted averaging the four scoring matrices according to the expert weights. n*m . n*m Perform two defuzzifications to obtain the semi-fuzzy failure mode evaluation matrix A′ n*m And clarify the failure mode evaluation matrix B n*m Matrix A′ n*m and B n*m The results of the first five failure modes are shown in Table 5 and Table 6 respectively.

[0100] Table 5 Semi-fuzzy failure mode evaluation matrix A n*m (Top 5 failure modes)

[0101]

[0102] Table 6 Clear failure mode evaluation matrix B n*m (Top 5 failure modes)

[0103] Severity S Occurrence rate Detection degree D Repair R Economy <![CDATA[FM1]]> 0.036068 0.216948 0.07472 0.07472 0.954832 <![CDATA[FM2]]> 0.29118 0.242386 0.44908 0.315953 0.761754 <![CDATA[FM3]]> 0.107133 0.055908 0.036068 0.022341 0.87914 <![CDATA[FM4]]> 0.068396 0.174556 0.119649 0.109576 0.822594 <![CDATA[FM5]]> 0.471574 0.389213 0.365747 0.361787 0.684047

[0104] According to the obtained clear failure mode evaluation matrix B n*m , using the MEREC method, the weights of the risk factors are calculated to be [0.21548, 0.27658, 0.19710, 0.25699, 0.05385].

[0105] The semi-fuzzy failure mode evaluation matrix is ​​standardized, and the results of the first five failure modes after standardization are shown in Table 7.

[0106] Table 7 Semi-fuzzy failure mode evaluation matrix A′ n*m Normalized results (top 5 failure modes)

[0107]

[0108] Calculate the failure mode FM under 5 risk factors i The exponential prospect function values ​​relative to all other failure modes are added together to obtain FM i The dominance of the first five failure modes relative to all other failure modes is shown in Table 8.

[0109] Table 8 Dominance of failure modes relative to other failure modes (top 5 failure modes)

[0110] Failure Mode <![CDATA[FM1]]> <![CDATA[FM2]]> <![CDATA[FM3]]> <![CDATA[FM4]]> <![CDATA[FM5]]> <![CDATA[FM1]]> 0.000000 -0.129345 -0.003310 -0.021036 -0.190984 <![CDATA[FM2]]> 0.057487 0.000000 0.063720 0.053403 -0.078499 <![CDATA[FM3]]> -0.025968 -0.143370 0.000000 -0.029744 -0.207070 <![CDATA[FM4]]> 0.004116 -0.120157 0.004667 0.000000 -0.184166 <![CDATA[FM5]]> 0.084882 0.027445 0.092031 0.081852 0.000000

[0111] Failure Mode FM i The failure mode FM is obtained by summing the advantages relative to the other failure modes. i The overall advantage of .

[0112] The overall superiority of each failure mode is standardized and ranked to obtain the improved FMEA ranking. The traditional FMEA ranking is obtained by calculating the RPN value scored by the experts and sorting the RPN value in descending order. Four experts brainstormed the failure modes and obtained the reference failure mode ranking. The overall superiority and improved FMEA ranking, traditional FMEA ranking, and reference ranking are shown in Table 9.

[0113] Table 9 Overall dominance of failure modes and comparison of three ranking methods

[0114] Ranking Overall dominance Improve FMEA sequencing Traditional FMEA ranking Reference sorting 1 1.000000 <![CDATA[FM 22 ]]> <![CDATA[FM 22 ]]> <![CDATA[FM 22 ]]> 2 0.930944 <![CDATA[FM 20 ]]> <![CDATA[FM 20 ]]> <![CDATA[FM 20 ]]> 3 0.848545 <![CDATA[FM 11 ]]> <![CDATA[FM 18 ]]> <![CDATA[FM 11 ]]> 4 0.844513 <![CDATA[FM 19 ]]> <![CDATA[FM 19 ]]> <![CDATA[FM 19 ]]> 5 0.843296 <![CDATA[FM 18 ]]> <![CDATA[FM 11 ]]> <![CDATA[FM 16 ]]> 6 0.816124 <![CDATA[FM 16 ]]> <![CDATA[FM 21 ]]> <![CDATA[FM 18 <!-- 12 -->]]> 7 0.794442 <![CDATA[FM 21 ]]> <![CDATA[FM 24 ]]> <![CDATA[FM 21 ]]> 8 0.728830 <![CDATA[FM 24 ]]> <![CDATA[FM7]]> <![CDATA[FM 24 ]]> 9 0.712438 <![CDATA[FM9]]> <![CDATA[FM9]]> <![CDATA[FM9]]> 10 0.704772 <![CDATA[FM6]]> <![CDATA[FM 16 ]]> <![CDATA[FM6]]> 11 0.689065 <![CDATA[FM7]]> <![CDATA[FM6]]> <![CDATA[FM7]]> 12 0.649753 <![CDATA[FM 28 ]]> <![CDATA[FM 13 ]]> <![CDATA[FM 13 ]]> 13 0.622569 <![CDATA[FM 13 ]]> <![CDATA[FM 12 ]]> <![CDATA[FM 12 ]]> 14 0.614189 <![CDATA[FM 12 ]]> <![CDATA[FM8]]> <![CDATA[FM8]]> 15 0.585387 <![CDATA[FM8]]> <![CDATA[FM 10 ]]> <![CDATA[FM 10 ]]> 16 0.571874 <![CDATA[FM 10 ]]> <![CDATA[FM 14 ]]> <![CDATA[FM 14 ]]> 17 0.534726 <![CDATA[FM 14 ]]> <![CDATA[FM 26 ]]> <![CDATA[FM 26 ]]> 18 0.398878 <![CDATA[FM 23 ]]> <![CDATA[FM5]]> <![CDATA[FM 23 ]]> 19 0.395583 <![CDATA[FM5]]> <![CDATA[FM 23 ]]> <![CDATA[FM5]]> 20 0.394925 <![CDATA[FM 26 ]]> <![CDATA[FM 17 ]]> <![CDATA[FM 27 ]]> 21 0.345950 <![CDATA[FM 27 ]]> <![CDATA[FM 27 ]]> <![CDATA[FM 17 ]]> 22 0.344858 <![CDATA[FM 17 ]]> <![CDATA[FM2]]> <![CDATA[FM 28 ]]> 23 0.267304 <![CDATA[FM2]]> <![CDATA[FM 15 ]]> <![CDATA[FM2]]> 24 0.151796 <![CDATA[FM 25 ]]> <![CDATA[FM 25 ]]> <![CDATA[FM 25 ]]> 25 0.141022 <![CDATA[FM 15 ]]> <![CDATA[FM 29 ]]> <![CDATA[FM 15 ]]> 26 0.116556 <![CDATA[FM 29 ]]> <![CDATA[FM 28 ]]> <![CDATA[FM 29 ]]> 27 0.048599 <![CDATA[FM4]]> <![CDATA[FM1]]> <![CDATA[FM4]]> 28 0.026925 <![CDATA[FM1]]> <![CDATA[FM4]]> <![CDATA[FM1]]> 29 0.000000 <![CDATA[FM3]]> <![CDATA[FM3]]> <![CDATA[FM3]]>

[0115] Summarize the traditional FMEA ranking results, the reference ranking results given by experts, and the improved FMEA ranking results Figure 3 In the figure, the horizontal axis represents the number of the failure mode and the vertical axis represents the ranking of the failure mode.

[0116] according to Figure 3 As shown in Table 9, the ranking results obtained by this method are highly consistent with the reference results given by experts, especially the failure modes ranked high and low, such as FM 22 ,FM 20 The rankings of failure modes such as FM1, FM2, FM3, FM1, and FM4 are exactly the same, which verifies the accuracy of the ranking method.

[0117] There is a slight gap between the ranking of some failure modes of the traditional FMEA method and the reference results given by this method and experts, such as FM7, FM 16 ,FM 18 This may be due to the differences between the improved FMEA method and the traditional FMEA method in terms of expert scoring method, relative weight of risk factors, and ranking method.

[0118] The ranking of some failure modes in the improved FMEA is improved compared with the traditional FMEA. 16 For example, FM 16 It ranks at the top in the reference ranking and improved FMEA ranking given by experts, but is not prominent in traditional FMEA. 16 The product quality is unqualified, with high severity but low occurrence. Since the traditional FMEA does not consider the relative importance of risk factors, it is ranked at the bottom. The improved FMEA proposed in the present invention considers 5 risk factors and their relative importance in the ranking, and assigns different weights to the 5 risk factors, which is more in line with the actual situation, making FMEA 16 Ranked higher.

[0119] The ranking of some failure modes in the improved FMEA is lower than that of the traditional FMEA. Taking FM7 as an example, FM7 ranks low in both the reference ranking given by experts and the improved FMEA ranking, but is ranked high in the traditional FMEA. FM7 is an error in inventory management information. The scores of its various risk factors are not outstanding, but they are all relatively high. The use of RPN value sorting leads to FM7 ranking high. In the improved FMEA sorting, the exp-TODIM method is used to sort the failure modes, which is more inclined to avoid risks. Therefore, failure modes with higher risk factor scores will be relatively high, while the risk factor scores of FM7 are not outstanding and the ranking will be relatively low, which is more in line with the actual situation.

[0120] The number of devices and processing scales described here are used to simplify the description of the present invention, and the application, modification and variation of the present invention will be obvious to those skilled in the art.

[0121] Although the embodiments of the present invention have been disclosed as above, they are not limited to the applications listed in the specification and implementation modes. They can be fully applied to various fields suitable for the present invention. For those familiar with the art, additional modifications can be easily implemented. Therefore, without departing from the general concept defined by the claims and the scope of equivalents, the present invention is not limited to the specific details and the illustrations shown and described herein.

Claims

1. An improved FMEA method based on double-layer hesitant triangle fuzzy semantic set and fuzzy exp-TODIM, characterized in that: The following steps are involved: S1. Set up a FMEA team to analyze the operation process of the object; S2, determine n potential failure modes of the analysis object and construct a failure mode set; S3. Improve the risk factors in the traditional FMEA method to include severity, occurrence, detection, repair and economy; S4, establishing a double-layer hesitant triangular fuzzy semantic set, including one-to-one correspondence between the double-layer hesitant triangular fuzzy language elements and the evaluation language, constructing corresponding evaluation criteria according to the double-layer hesitant triangular fuzzy semantic set and the double-layer hesitant triangular fuzzy language elements corresponding to the language variables, and converting the expert scoring table into a scoring matrix of L experts; S5, obtaining expert weights through a consensus decision-making algorithm based on social network trust and consistency of fuzzy preference relations; S6. According to the expert weights, the scoring matrices of the experts are weighted averaged to obtain a fuzzy failure mode evaluation matrix; S7, defuzzifying the fuzzy failure mode evaluation matrix in two steps to obtain a semi-fuzzy failure mode evaluation matrix and a definite failure mode evaluation matrix; S8. Use the MEREC method on the explicit failure mode evaluation matrix to obtain the risk factor weights; S9. The semi-fuzzy failure mode evaluation matrix is ​​used to rank the failure modes using the fuzzy exp-TODIM method based on the generalized mixed Hausdorff preference distance, and relevant personnel formulate corresponding prevention and response measures based on the ranking.

2. The improved FMEA method based on double-layer hesitant triangular fuzzy semantic set and fuzzy exp-TODIM as claimed in claim 1, characterized in that: The establishment of the double-layer hesitant triangular fuzzy semantic set in step 4 includes the following steps: S41. Define a bilingual terminology set; S42. Define a set of double-layer hesitant fuzzy language terms; S43, define triangular fuzzy number, fuzzy number It can be obtained by (a L , a M , a U ) determines (0≤a L ≤a M ≤a U ≤1), and the membership function (or characteristic function) is: Then it is called is a triangular fuzzy number; where a L is the most conservative or pessimistic estimate, a M is the most likely estimate, a U is the most optimistic evaluation value; S44. Define a double-layer hesitant triangle fuzzy language term set.

3. The improved FMEA method based on double-layer hesitant triangle fuzzy semantic set and fuzzy exp-TODIM as claimed in claim 2, characterized in that: Step S41 includes setting S = {s t |t=-τ,…,-1,0,1,…,τ} and are two finite, completely ordered discrete language term sets, S and O are the first-level and second-level language term sets respectively and are independent of each other, then the two-level language term set in It is a two-layer language term, s t Indicates the first level language, o k Indicates the second level language.

4. The improved FMEA method based on double-layer hesitant triangular fuzzy semantic set and fuzzy exp-TODIM as claimed in claim 3, characterized in that: Step S42 includes setting X to be a non-empty set, based on the two-layer language term set S o Double-layer hesitant fuzzy language term set in, Indicates S o A continuous list of possible values ​​in Embodying Elements x i Belongs to S o The membership degree of is a double-layer hesitant fuzzy linguistic element, L i express is the number of language terms in the sentence; Φ×Ψ represents the set of double-layer hesitant fuzzy language elements.

5. The improved FMEA method based on double-layer hesitant triangular fuzzy semantic set and fuzzy exp-TODIM as claimed in claim 4, characterized in that: Step S44 includes assuming that X is a non-empty set, is a two-level language term set, then S o Double-layer hesitant triangle fuzzy language term set It is expressed as: in, For a three-element set: in All three elements in the set belong to S o , respectively represent x i In the most pessimistic, most likely and most optimistic scenarios, is a double-layer hesitant triangular fuzzy language element, and T(Φ×Ψ) represents the set of double-layer hesitant triangular fuzzy language elements.

6. The improved FMEA method based on double-layer hesitant triangular fuzzy semantic set and fuzzy exp-TODIM as claimed in claim 1, characterized in that: The construction process of the evaluation criteria based on the double-layer hesitant triangular fuzzy semantic set in step 4 is as follows: S45. Give scoring examples and explanations of their significance. For each failure mode, the scoring examples for a single risk factor are as follows: [(x1, y1), (x2, y2), (x3, y3)]; The scoring of individual risk factors is divided into three parts: The first part: (x1, y1), represents the minimum rating, that is, the minimum rating of a very large indicator or the maximum rating of a very small indicator. The second part: (x2, y2), represents the most likely rating, The third part: (x3, y3), represents the highest rating, that is, the maximum rating for very large indicators or the minimum rating for very small indicators; S46. Give an explanation of each part of the scoring: (x i ,y i ) is a two-level description of risk factors, dividing an exact value into "integer part (x i )" and "decimal part (y i )" two layers, x i +y i To provide practical evaluation criteria; x i is the first layer, the integer part, the value range is {-2, -1, 0, 1, 2}, representing {very low (VL), low (L), medium (M), high (H), very high (VH)}; y i For the second level, the decimal part, as a pair of x i The supplement (adjective) of They represent {much lower, a little lower, exactly the same, a little higher, much higher} respectively; S47. Provide additional explanation: (1) x i +y i For a more practical evaluation standard, (x i ,y i ) form; secondly, x i +y i The value range is [-2, 2]. i =2(x i =-2), y i can only take non-positive (non-negative) numbers; finally, in the evaluation criteria, x i Corresponding to the φ in the double-layer hesitant triangular fuzzy language element l ,y i Corresponding to the double-layer hesitant triangle fuzzy language element because but 7. The improved FMEA method based on double-layer hesitant triangular fuzzy semantic set and fuzzy exp-TODIM as claimed in claim 1, characterized in that: The defuzzification method of the fuzzy failure mode evaluation matrix in step 7 is as follows: S71. For each double-layer hesitant triangle fuzzy language element Elements in The double-level hesitant fuzzy linguistic element defuzzification method is used to transform the double-level hesitant triangular fuzzy linguistic element into triangular fuzzy number. S72. Defuzzify the obtained triangular fuzzy number by using the K-th power weighted average method to obtain an accurate number.

8. The improved FMEA method based on double-layer hesitant triangular fuzzy semantic set and fuzzy exp-TODIM as claimed in claim 1, characterized in that: The fuzzy exp-TODIM method based on generalized Hausdorff preference distance in step 9 includes the following steps: S91, standardize the semi-fuzzy failure mode evaluation matrix using a linear standardization method to obtain a standardized matrix; S92, under the kth risk factor, failure mode FM i With FM j The corresponding double-layer hesitant triangular fuzzy language elements are: Then a ik with a jk The generalized Hausdorff preference distance between is: Where λ>0,0≤a l , b≤1(l=1,2,...,L), Pick S93, under the kth risk factor, failure mode FM i (i=1,2,...,n) The exponential prospect function value relative to all other failure modes is where j = 1, 2, ..., n, j ≠ i; w k is the weight of the kth risk factor; for a ik with a jk The generalized Hausdorff preference distance between them; λ is the amplification parameter for adjusting the gain and loss effects to ensure that the impact of loss is greater than the gain; ρ is used to indicate the degree of influence of a single evaluation gap on the dominance; S94, Calculate failure mode FM i Dominance relative to all other failure modes S95, Calculate failure mode FM i The overall advantage S96, according to ξ i The failure modes are sorted in order of decreasing value.

Images