A method for risk analysis in lithium battery assembly

Abstract

This invention discloses a risk analysis method for lithium battery assembly, comprising the following steps: Step S1, using a two-layer hesitant fuzzy language terminology set to evaluate the severity, frequency of occurrence, and detectability of failure modes in lithium battery assembly; wherein, the two-layer hesitant fuzzy language terminology set includes a two-layer language terminology set and a hesitant fuzzy language terminology set, both composed of evaluation information from evaluators; Step S2, calculating evaluator weights based on similarity measures, and calculating the influencing factor weights of severity, frequency of occurrence, and detectability based on the principle of water injection; Step S3, aggregating the evaluation information of evaluators using a combined compromise ranking method to calculate the failure mode risk value; Step S4, determining the optimal evaluation information through multi-objective aggregation technology; This invention accurately describes the degree of risk by expressing basic semantics at the main layer and modifying the degree of modification at the sub-layer, thus avoiding misjudgment of extreme values. Simultaneously, semantic constraint rules are used to eliminate logical conflicts and improve expression efficiency.

Description

Technical Field

[0001] This invention relates to the field of quality management technology, and in particular to a method for risk analysis in lithium battery assembly. Background Technology

[0002] In recent years, with the increase in sales of new energy vehicles, the demand for lithium batteries has also increased significantly. To meet market demand, lithium battery manufacturers have been expanding their production capacity and improving production efficiency. Compared with other types of batteries, lithium battery manufacturing involves more automated processes and requires higher control precision. If a fault occurs in the lithium battery processing, it will seriously affect production quality and efficiency. Therefore, before installing a lithium battery production line, a quality analysis of the production line design is necessary to identify potential failure modes and take preventive measures against high-risk failure modes in a timely manner.

[0003] In existing technologies, the FMEA method is suitable for analyzing and evaluating potential failure modes during the product manufacturing stage. The FMEA method analyzes the product's structure tree and function tree to obtain potential failure modes and their consequences. Then, an evaluation team is formed to evaluate the severity (S), frequency (O), and detectability (D) of the failure modes using integers from 1 to 10. Finally, based on... FMEA calculates the RPN (Recovery Risk Number) value of each failure mode and then determines the risk level of the failure mode based on the numerical value. This demonstrates that FMEA is simple to operate and easy to implement, hence its widespread use across various industries such as manufacturing, healthcare, and logistics. However, with the development of information technology, FMEA applications have revealed some shortcomings in complex and uncertain environments:

[0004] First, its use of integers as the evaluation language is unrealistic. In actual production processes, it is impractical for evaluators to accurately assess the consequences of failure modes, especially during the product design phase. Therefore, using integers for risk assessment of failure modes cannot capture the hesitation of evaluators during the assessment process.

[0005] Second, the weighting of evaluators is often overlooked. FMEAs typically involve multiple evaluators from various departments with different roles, work experience, and educational backgrounds. This leads to varying perceptions of failure modes and consequently, different evaluation values. Therefore, assigning different weights to evaluators with varying qualifications is crucial to ensuring the final evaluation results accurately reflect reality.

[0006] Third, the weights of influencing factors should be evenly distributed. Traditional FMEA does not consider the weights of influencing factors when calculating RPN. However, in practice, for the same RPN value, higher severity failure modes are often prioritized for prevention. That is, in practice, severity is given higher weight. Therefore, evenly distributing the weights of influencing factors can lead to inaccurate risk assessment results.

[0007] Fourth, the effectiveness of RPN in guiding risk prevention is low. In traditional FMEA, RPN is the cumulative product of the evaluation values ​​of influencing factors. This can easily lead to multiple failure modes obtaining the same RPN value, meaning that multiple failure modes have the same level of risk. If preventative measures are required for all of these failure modes, then the purpose of FMEA is lost. Summary of the Invention

[0008] The technical problems solved by this invention are: limitations in evaluation semantics, unreasonable weight allocation, duplicate risk ranking, RPN value accumulation leading to identical scores for multiple failure modes, lack of verification methods, and lack of objective standards to compare the merits of different optimization methods.

[0009] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0010] A method for risk analysis of lithium battery assembly includes the following steps:

[0011] Step S1: Use a two-layer hesitant fuzzy language terminology set to evaluate the severity, frequency of occurrence, and detectability of failure modes in lithium battery assembly; wherein, the two-layer hesitant fuzzy language terminology set includes a two-layer language terminology set and a hesitant fuzzy language terminology set, both of which are composed of evaluation information from the evaluators.

[0012] Step S2: Calculate the weight of evaluators based on similarity measure, and calculate the weight of influencing factors of severity, frequency of occurrence, and detectability based on the water injection principle;

[0013] Step S3: Use the combined compromise ranking method to aggregate the evaluation information of the evaluators and calculate the failure mode risk value;

[0014] Step S4: Determine the optimal evaluation information using multi-objective aggregation technology.

[0015] As a preferred embodiment of the lithium battery assembly risk analysis method described in this invention, wherein:

[0016] In step S1, the two-layer language terminology set It includes the main-level language terminology set and the sub-level language terminology set, and its expression is as follows:

[0017] ;

[0018] The main language terminology set Used to describe basic semantics, including the following expressions:

[0019] ;

[0020] The sub-level language terminology set Expressions used to modify semantic strength include the following:

[0021] ;

[0022] in, Boundary parameters of the primary language term set. This is the index value of the main-level language term in the main-level language term set. For the boundary parameters of the sub-level language term set, This is the index value of the sub-level language term in the main-level language term set.

[0023] As a preferred embodiment of the lithium battery assembly risk analysis method described in this invention, wherein:

[0024] In step S1, the expression for the hesitant and ambiguous language terminology set is as follows:

[0025] ;

[0026] in, S3 is a finite, ordered set of continuous linguistic terms. As a set of linguistic terms, , Given a non-empty set, yes One of the elements;

[0027] exist middle, The element representing the set of hesitant and ambiguous language terms is expressed as follows:

[0028] ;

[0029] in, A finite and continuous set of linguistic terms;

[0030] The set of hesitant and fuzzy language terms is transformed into membership degrees through a transformation function.

[0031] As a preferred embodiment of the lithium battery assembly risk analysis method described in this invention, wherein:

[0032] In step S2, the calculation steps for the evaluator weights include:

[0033] Step S211: Obtain the membership degree of the evaluator among the failure mode influencing factors, calculate the non-membership degree, and generate the ideal solution;

[0034] Step S212: Calculate the Euclidean distance between the evaluator and the ideal solution;

[0035] Step S213: Calculate the similarity between the evaluator and the ideal solution;

[0036] Step S214: Calculate the weight of the evaluators based on similarity.

[0037] As a preferred embodiment of the lithium battery assembly risk analysis method described in this invention, wherein:

[0038] In step S2, the calculation steps for the weights of influencing factors include:

[0039] Step S221: Calculate the mean and variance of the evaluation matrix;

[0040] Step S222, calculate the horizontal plane;

[0041] Step S223: Calculate the weights of influencing factors based on the horizontal plane.

[0042] As a preferred embodiment of the lithium battery assembly risk analysis method described in this invention, wherein:

[0043] Step S3 includes the following steps:

[0044] Step S311: Based on the evaluator weights, aggregate the evaluation information using a simple intuitive fuzzy weighted geometric operator to obtain the membership and non-membership degrees of the evaluators;

[0045] Step S312: Calculate the weighted sum and weighted product of the evaluation information to obtain the integration result of the evaluation information;

[0046] Step S313: Calculate the intermediate risk value using three scoring strategies;

[0047] Step S314: Calculate the intermediate risk value obtained through the three scoring strategies to obtain the final risk assessment value.

[0048] As a preferred embodiment of the lithium battery assembly risk analysis method described in this invention, wherein:

[0049] In step S313, the formulas for the three scoring strategies are as follows:

[0050] ;

[0051] ;

[0052] ;

[0053] in, This represents the median risk value for the first scoring strategy. This represents the median risk value for the second scoring strategy. This represents the median risk value for the third scoring strategy. For weighted and integrated, For weighted product integration; It is a tendency factor.

[0054] As a preferred embodiment of the lithium battery assembly risk analysis method described in this invention, wherein:

[0055] In step S314, the formula for calculating the final risk assessment value is as follows:

[0056] ;

[0057] in, This is the final risk assessment value.

[0058] As a preferred embodiment of the lithium battery assembly risk analysis method described in this invention, wherein:

[0059] Step S4 includes the following steps:

[0060] Step S41: Summarize the methods of failure modes ranked in risk ranking, generate the corresponding set, and calculate the cumulative number of times the failure mode appears in risk ranking.

[0061] Step S42: Based on the different evaluation information of the evaluation personnel on the failure mode under the influencing factors, calculate the evaluation value of the corresponding evaluation information that actually participates in the final risk calculation.

[0062] Step S43: Set up the allocation matrix according to the weight of the risk ranking;

[0063] Step S44: Set up a multi-objective mathematical model, the formula for its objective function is shown below:

[0064] ;

[0065] ;

[0066] The constraints are as follows:

[0067] ;

[0068] in, Indicates the number of failure modes. ; Indicates the number of influencing factors. ; Indicates the number of evaluators. ; This indicates a risk ranking. ; To optimize the number of methods, ; Failure mode In risk ranking The cumulative number of times it appears above, of which, ; The weights for ranking by risk are represented as To assign elements in the matrix; Failure mode Ranked by risk The above set of methods, Failure mode In risk ranking The number of times it appears above; Representation optimization method Influencing factors The weight, Representation optimization method Intermediate evaluators The weight, For evaluators Failure modes Influencing factors The following is the evaluation information; ;and For evaluation information The scoring function; ;

[0069] By comparing the calculation results and optimal solutions of different failure mode optimization methods, the merits of the optimization methods can be judged, and the optimal evaluation information can be determined.

[0070] As a preferred embodiment of the lithium battery assembly risk analysis method described in this invention, wherein:

[0071] In step S44, the formula for calculating the degree of difference is as follows:

[0072] ;

[0073] in, Representation optimization method Failure modes The sorting, Failure modes for solving multi-objective models Optimal sorting Representation optimization method Failure modes Influencing factors Evaluation information, Failure modes for solving multi-objective models Influencing factors Optimal evaluation information for The maximum value, for The maximum value that can be obtained.

[0074] The beneficial effects of this invention are:

[0075] First, the lithium battery assembly risk analysis method provided by this invention accurately captures the hesitation of experts through the two-layer language structure of DHHFLTS (the main layer expresses basic semantics, and the secondary layer modifies the degree of modification). This allows for a more accurate description of the degree of risk and avoids misjudgments based on extreme values. Simultaneously, semantic constraint rules also help eliminate logical conflicts and improve expression efficiency.

[0076] Secondly, the lithium battery assembly risk analysis method provided by this invention thoroughly solves the problem of duplicate RPN values ​​through a three-strategy fusion. First, it strengthens global / local features through weighted summation and weighted product respectively; then, it uses cross-validation of the three scoring strategies to adapt the bias factors to different decision preferences; finally, it ensures the uniqueness of the ranking through a geometric-arithmetic hybrid calculation. This effectively distinguishes different failure modes and significantly reduces the conflict rate of calculated values. Attached Figure Description

[0077] Figure 1 A flowchart illustrating a lithium battery assembly risk analysis method provided by the present invention;

[0078] Figure 2 This is a schematic diagram of the structure of a two-layer language terminology set in a lithium battery assembly risk analysis method provided by the present invention;

[0079] Figure 3 Distribution diagram of the sub-level language terminology set in a lithium battery assembly risk analysis method provided by the present invention;

[0080] Figure 4 A schematic diagram of the risk analysis optimization process in a lithium battery assembly risk analysis method provided by the present invention;

[0081] Figure 5 This is a schematic diagram of the structure of a square lithium battery provided by the present invention;

[0082] Figure 6 A flowchart illustrating the manufacturing process of a square lithium battery provided by this invention;

[0083] Figure 7 A flowchart of the assembly workshop process for square lithium batteries provided by the present invention;

[0084] Figure 8 This is a schematic diagram of the functional tree of the ultrasonic welding station provided by the present invention;

[0085] Figure 9 This is a schematic diagram of the ultrasonic welding station failure mesh provided by the present invention;

[0086] Figure 10 A schematic diagram of the laser welding station and the battery adhesive applicator failure mesh provided by the present invention;

[0087] Figure 11 A schematic diagram of experimental data for solving the Pareto front of the multi-objective mathematical model provided by this invention;

[0088] Figure 12 This is a schematic diagram of experimental data for ranking failure modes under different tendency factors provided by the present invention. Detailed Implementation

[0089] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0090] Using integers as the evaluation language in FMEA is unrealistic. In actual production processes, it's impractical for evaluators to accurately assess the consequences of failure modes, especially during the product design phase. Therefore, using integers for failure mode risk assessment fails to capture the hesitation and uncertainty experienced by evaluators during the evaluation process.

[0091] Reference Figure 1 To address the above problems, this invention provides a lithium battery assembly risk analysis method, comprising the following steps:

[0092] Step S1: Use a two-layer hesitant fuzzy language terminology set to evaluate the severity, frequency of occurrence, and detectability of failure modes in lithium battery assembly; wherein, the two-layer hesitant fuzzy language terminology set includes a two-layer language terminology set and a hesitant fuzzy language terminology set, both of which are composed of evaluation information from the evaluators.

[0093] Step S2: Calculate the weight of evaluators based on similarity measure, and calculate the weight of influencing factors of severity, frequency of occurrence, and detectability based on the water injection principle;

[0094] Step S3: Use the combined compromise ranking method to aggregate the evaluation information of the evaluators and calculate the failure mode risk value;

[0095] Step S4: Determine the optimal evaluation information using multi-objective aggregation technology.

[0096] In step S1, the two-layer language terminology set includes a primary-layer language terminology set and a secondary-layer language terminology set, the expression of which is shown below:

[0097] ;

[0098] The main-level language terminology set is used to describe basic semantics, including the following expressions:

[0099] ;

[0100] The sub-layer language terminology set is used to modify semantic strength and includes the following expressions:

[0101] ;

[0102] in, Boundary parameters of the primary language term set. The index (subscript) of the main language term in the set of main language terms. For the boundary parameters of the sub-level language term set, This is the index (subscript) of the sub-level language term in the main-level language term set.

[0103] For example, this application order hour, , At this point, the diagram of the two-layer language terminology set is as follows: Figure 2 As shown, Figure 2 middle express and Two sets of language terms.

[0104] At the same time, to avoid semantic and logical problems, when the main layer language terminology set is Sub-level language terminology set is taken again To avoid semantic repetition, this application makes the following provisions regarding the use of the two-layer language terminology set, the distribution of which is illustrated in the diagram below. Figure 3 As shown.

[0105] 1. and hour, ;

[0106] 2. and hour, ;

[0107] 3. hour, ;

[0108] 4. hour, .

[0109] Furthermore, in step S1, the set of hesitant and ambiguous language terms The expression is as follows:

[0110] ;

[0111] in, Let S be a finite, ordered set of continuous linguistic terms. As a set of linguistic terms, , Given a non-empty set, yes One of the elements;

[0112] exist middle, The element representing the set of hesitant and ambiguous language terms is expressed as follows:

[0113] ;

[0114] in, A finite and continuous set of linguistic terms; for example, express Belonging to The probability is located at and between.

[0115] The hesitant and fuzzy language terminology set is transformed into membership degrees through a transformation function, which are then used in the calculation of the final evaluation results. For example, let... A collection of linguistic terms, For hesitant and blurred elements, Indicates the number of language term sets. If membership degree and In If they express the same semantics, they can be converted to each other using a conversion function. Assume... It is expressed as "very high," while It can also be expressed as "very high", so they can be converted into each other.

[0116] Specifically, the expression for the transformation function is as follows:

[0117] ;

[0118] ;

[0119] At the same time, the transformation function is extended to the entire language terminology set. With hesitant and ambiguous elements When the expression is as follows:

[0120] ;

[0121] ;

[0122] in, for Boundary parameters within, For hesitant and ambiguous language terms in The index value (subscript) in the text; The transformed membership value is a real number in the interval [0,1].

[0123] Finally, based on the two-layer language terminology set and the hesitant and ambiguous terminology set, the concept of DHHFLTS is formed, and its expression is as follows:

[0124] ;

[0125] The meanings of the symbols in the expression are the same as those in the expression above.

[0126] Furthermore, a transformation function is used to convert DHHFLTS into membership degrees, which are then used to determine the risk level of the final failure mode.

[0127] For example, let If it's a hesitant or ambiguous element, then in DHHFLTS, the conversion function expression is as follows:

[0128] ;

[0129]

[0130] Further extending the transformation function to the entire hesitant and fuzzy language element With hesitant and ambiguous elements The conversion function expression is as follows:

[0131] ;

[0132] ;

[0133] The meanings of the symbols within the expression are the same as those within the expression above.

[0134] When DHHFLTS participates in calculations, the calculation rules must be followed. and There are two DHHFLTS, and The expression for the algorithm is as follows:

[0135] ;

[0136] ;

[0137] ;

[0138] ;

[0139] in, and The meaning is as shown above. It is a constant. , and They are and Hesitant and ambiguous elements.

[0140] Furthermore, let's assume For evaluators Failure modes Influencing factors The evaluation information can be divided into two parts according to the SM-WFT calculation process: one part is to determine the weight of the evaluator, and the other part is to calculate the weight of the influencing factors.

[0141] Specifically, in step S2, the first part of the calculation steps regarding the weights of the evaluators includes:

[0142] Step S211: Obtain the membership degree of the evaluator among the failure mode influencing factors, calculate the non-membership degree, and generate the ideal solution;

[0143] Among them, the assumed evaluators In failure mode Influencing factors The membership degree in The specific calculation can be performed using the formula above. The calculated non-membership degree is as follows: Therefore, the evaluation matrix can be written as: .

[0144] The formula for calculating the ideal solution is as follows:

[0145] ;

[0146] in, Indicates the number of failure modes. ; Indicates the number of influencing factors. ; Indicates the number of evaluators. .

[0147] Step S212: Calculate the Euclidean distance between the evaluator and the ideal solution. Specifically, the formula for calculating the Euclidean distance between the evaluator and the ideal solution is as follows:

[0148] ;

[0149] Step S213: Calculate the similarity between the evaluator and the ideal solution. Specifically, the formula for calculating the similarity between the evaluator and the ideal solution is as follows:

[0150] ;

[0151] in, Represents the complement of the ideal solution. .

[0152] Step S214: Calculate the evaluator weights based on similarity. Specifically, the formula for calculating the weight of evaluators is as follows:

[0153] ;

[0154] The meanings of the symbols within the expression are the same as those within the expression above.

[0155] Specifically, the "water injection principle" refers to allocating higher communication power when optimizing communication power allocation if channel conditions are good, and lower communication power otherwise. In this application, when calculating the weights of influencing factors in FMEA, if the consensus among evaluators regarding the influencing factors of the same failure mode is high, a higher weight is assigned; otherwise, a lower weight is assigned. Therefore, in step S2, the second part of the calculation steps related to the weights of influencing factors includes:

[0156] Step S221, calculate the evaluation matrix mean and variance Specifically, the calculation formula is as follows:

[0157] ;

[0158] ;

[0159] in, Represents the scoring function. Failure mode Influencing factors The mean matrix of the evaluation information of the evaluators is then used. Indicating influencing factors The situation where the evaluators reach a consensus.

[0160] Step S222, further based on the water injection principle and the obtained mean value With variance The formula for calculating the horizontal plane is as follows:

[0161] ;

[0162] in, The remaining symbols have the same meaning as those in the previous text.

[0163] Step S223: Based on the horizontal plane, calculate the weights of the influencing factors. The calculation formula is as follows:

[0164] ;

[0165] in, The remaining symbols have the same meaning as those in the previous text.

[0166] Furthermore, the execution of step S3 includes the following steps:

[0167] Step S311: Based on the evaluator weights, aggregate the evaluation information using the Simple Intuitionistic Fuzzy Weighted Geometric (SIFWG) operator to obtain the membership and non-membership degrees of the evaluators. The advantage of using the SIFWG operator is that it ensures the consistency of the evaluators' evaluation information. The specific expression is shown below:

[0168] ;

[0169] in, For evaluators The weights are given by the formula above. Calculated. These represent the membership and non-membership of the matrix after aggregation by the SIFWG operator, respectively. Since this application can consider two integration methods, weighted sum and weighted product, and then accurately evaluate the final result through three scoring strategies, the possibility of the same failure mode score can be reduced.

[0170] Therefore, in step S312, the weighted sum and weighted product of the evaluation information are integrated to obtain the integrated result of the evaluation information; the weighted sum expression and the weighted product expression are as follows:

[0171] ;

[0172] ;

[0173] in, Represents the aggregated matrix The scoring function.

[0174] Step S313: Calculate the intermediate risk value using three scoring strategies; specifically, after obtaining the integrated evaluation information, [the following steps are taken]. The calculations for the three scoring strategies are as follows:

[0175] ;

[0176] ;

[0177] ;

[0178] in, The propensity factor represents the tendency of the people to adopt a certain attitude. When, it indicates that evaluators prefer a weighted and integrated approach; while when When this occurs, it indicates that evaluators prefer a weighted product aggregation method. Therefore, it is usually taken as... .

[0179] Step S314: Calculate the intermediate risk value obtained through the three scoring strategies. The calculation expression is as follows:

[0180] ;

[0181] Then, the final risk assessment value is obtained. , The larger the value, the more likely it is to fail. The higher the risk, the more proactive the company should be in preventing potential high-risk failure modes.

[0182] Since there is currently a lack of objective and efficient methods for comparing the merits of different FMEA optimization methods, this application proposes a new multi-objective aggregation technique to compare the merits of different FMEA optimization methods.

[0183] First, according to the Borda rule, the calculation results of different FMEA optimization methods are similar to voting on the ranking of failure modes in terms of risk. Therefore, the optimal ranking should be assigned to the failure mode that receives the most votes.

[0184] Then, according to the Copeland rule, the Euclidean distance between the evaluation information of different FMEA optimization methods and the optimal evaluation information is calculated sequentially. The optimal evaluation information is determined with the minimum distance as the objective.

[0185] Finally, based on the aggregation technique, a high weight is given to the ranking with high risk, and a multi-objective mathematical model is constructed with the minimum risk ranking error and the minimum Euclidean distance to achieve the goals of Borda rule and Copeland rule.

[0186] Therefore, the execution of step S4 includes the following steps:

[0187] Step S41: Summarize the methods of failure modes ranked in the risk ranking, generate the corresponding set, and calculate the cumulative number of times the failure mode appears in the risk ranking.

[0188] Among them, failure modes Ranked by risk The above method statistical set is That is, failure mode. In risk ranking The number of times it appears above is .and Failure mode In risk ranking The cumulative number of times it appears above The expression is as follows:

[0189] ;

[0190] in, , , .

[0191] Step S42: Based on the evaluators' different evaluation information on the failure mode under influencing factors, calculate the evaluation value of the corresponding evaluation information that actually participates in the final risk calculation; specifically, let... As evaluators in this application Failure modes Influencing factors The evaluation information is as follows: Under different FMEA optimization methods, the numerical calculation formulas for the evaluation information actually involved in the final risk calculation are as follows:

[0192] ;

[0193] in, Representation optimization method Influencing factors The weight, Representation optimization method Intermediate evaluators The weight, For evaluation information Scoring function

[0194] Step S43: Set up the allocation matrix according to the weight of the risk ranking; specifically, according to the risk ranking... weight , Configure the allocation matrix; where, when Indicates the failure mode Assigned to ranking Other cases .

[0195] Step S44: Set up a multi-objective mathematical model, the formula for its objective function is shown below:

[0196] ;

[0197] ;

[0198] The constraints are as follows:

[0199] ;

[0200] in, Indicates the number of failure modes. ; Indicates the number of influencing factors. ; Indicates the number of evaluators. ; This indicates a risk ranking. ; To optimize the number of methods, ; Indicates failure mode Influencing factors The optimal evaluation information is as follows.

[0201] By comparing the calculation results and optimal solutions of different failure mode optimization methods, the merits of the optimization methods can be judged, and the optimal evaluation information can be determined.

[0202] Specifically, in step S44, the formula for calculating the degree of difference is as follows:

[0203] ;

[0204] in, Representation optimization method Failure modes The sorting, Failure modes for solving multi-objective models Optimal sorting Representation optimization method Failure modes Influencing factors Evaluation information, Failure modes for solving multi-objective models Influencing factors Optimal evaluation information for The maximum value, for The maximum value that can be obtained.

[0205] Here, we provide an example of a battery manufacturing company planning to produce a square lithium battery: the company uses the method proposed in this application to conduct a risk analysis of potential failure modes in the square lithium battery assembly workshop. In accordance with the five-step FMEA methodology, the company assembled an FMEA team, which consisted of three evaluators: , , .

[0206] First, a structural analysis of the square lithium battery assembly production line is conducted, the structure of which is as follows: Figure 5 As shown, the processing flow is as follows: Figure 6 As shown, this invention uses a process flow diagram to visualize the assembly workshop, such as... Figure 7 As shown. Then, taking the ultrasonic welding station as an example, the functional analysis of this station is visualized through a function tree, as follows. Figure 8 As shown. Secondly, a failure mesh was used to perform failure analysis on the ultrasonic welding station, such as... Figure 9 As shown. Besides the ultrasonic welding station, the laser welding station and the battery adhesive applicator are also important processing equipment in the square lithium battery assembly workshop. Their failure network is as follows: Figure 10 As shown.

[0207] exist Figure 9 and Figure 10 In this context, FE represents failure effect, FM represents failure mode, and FC represents failure cause. According to... Figure 9 and Figure 10 The failure network summarizes some potential failure modes of the square lithium battery assembly production line, as shown in Table 1. According to... Figure 4 The FMEA risk analysis flowchart was improved, and three evaluators used DHHFLTS to evaluate the failure modes in Table 1, as shown in Table 2.

[0208]

[0209]

[0210]

[0211]

[0212] In this application, let , , .

[0213] Calculate the membership and non-membership degrees of the evaluators' evaluation information based on Table 2. (Based on the evaluators...) In failure mode Influencing factors Taking the evaluation information on the platform as an example, , Similarly, the membership and non-membership of the remaining failure mode evaluation information are calculated, as shown in Tables 3 and 4.

[0214] To calculate the weights of the evaluators, firstly, based on Tables 3 and 4 and the formula... Calculate the ideal solution, for example, the failure mode. Influencing factors The ideal solution is Similarly, the ideal solutions for the other failure modes are calculated, as shown in Table 5.

[0215] Then, based on Tables 3, 4, and 5 and the formula... Calculate the Euclidean distance between the evaluator's evaluation information and the ideal solution.

[0216] For example, evaluators In failure mode Influencing factors Upper Euclidean distance is The evaluator The Euclidean distance to the complement of the ideal solution is Similarly, the Euclidean distances of the remaining evaluators are calculated, as shown in Tables 6 and 7.

[0217] Finally, based on Tables 6 and 7, the weights of the evaluators are calculated as follows: , , .

[0218]

[0219]

[0220]

[0221] To calculate the weights of the influencing factors, firstly, refer to Tables 3 and 4 and the scoring function calculation formula. Determine the score value for the evaluation information.

[0222] For example, evaluators In failure mode Influencing factors superior Similarly, calculate the evaluation scores of the remaining evaluators, as shown in Table 8.

[0223] Then, according to and The calculation expression calculates the mean and variance of the aggregated evaluation information. , , , , and .

[0224] Secondly, according to The formula for calculating the horizontal plane. Finally, according to Formula for calculating the weight of influencing factors , , .

[0225]

[0226] The method of this invention for calculating failure mode risk values ​​involves a trade-off. First, based on Tables 3 and 4 and the weights of the evaluators... , , as well as The formula is used to obtain the evaluation information after weighted aggregation of the SIFWG operator.

[0227] For example, failure modes Influencing factors The weighted data is Similarly, the results of the weighted aggregation of the SIFWG operators for the remaining failure modes are shown in Table 9.

[0228] Then, based on the weights of the influencing factors , , Perform weighted sums and weighted products, for example, failure modes. The weighted sum is calculated as follows: That failure mode The weighted product is calculated as follows: Similarly, the weighted sum and weighted product of the remaining failure modes are calculated, as shown in Table 10.

[0229] Secondly, the results of failure modes under the three scoring strategies are calculated, as shown in Table 10, where in formula (30), the propensity factor... .

[0230] Finally, the failure mode risk assessment results were calculated. For example, failure modes Evaluation results Similarly, the assessment results for the remaining failure modes are calculated, as shown in Table 10. Based on the risk assessment results... The failure mode risk ranking is as follows: .

[0231]

[0232]

[0233] To demonstrate the superiority of the FMEA optimization method of this invention, the calculation results of the new method are compared and analyzed with those of three other methods. The new method is denoted as... The other three methods are respectively denoted as , and .in The FMEA is optimized using Regret Theory and Weighted Aggregates Sum Product Assessment (WASPAS). The FMEA was not optimized. The FMEA was optimized using Spherical Fuzzy Step-Wise Weight Assessment Ratio Analysis (SF-SWARA).

[0234] In the comparative analysis, to ensure the objectivity of the analysis process, all four methods used the evaluation data of this invention, namely Table 8. The calculated weight information for the four methods is shown in Table 11. In Table 11, Without considering the weight of the evaluators, No weighting information is considered.

[0235] However, in the calculation, The weight of the evaluator is set to 1. All weights are set to 1. This is achieved through the formula... The actual data values ​​involved in the calculation for each method are shown in Table 12. Then, the evaluation results and rankings of the failure modes are calculated using each method, as shown in Table 13. Next, based on Table 13, the set of methods that rank the failure modes according to different risk levels is compiled, as shown in Table 14. Then, based on Table 14, the cumulative number of times the failure mode appears in the risk ranking is calculated, as shown in Table 15. Finally, a multi-objective mathematical model is defined, as follows:

[0236]

[0237] This invention uses the Non-Dominated Sorting Whale Optimization Algorithm (NSWA), the Nondominated Sorting Genetic Algorithm II (NSGA-II), and the Multi-Objective Grey Wolf Optimizer (MOGWO) to solve the multi-objective mathematical model. The results are as follows: Figure 11 As shown. From Figure 11 As can be seen from the Pareto front, NSWOA provides a wider range of solutions and converges faster than the other two methods. Therefore, this invention uses the Pareto front obtained by NSWOA and selects the optimal solution. The optimal NSWOA solution is the middle solution of the Pareto front because FMEA prioritizes the risk ranking results when formulating failure mode prevention measures. Therefore, this invention selects... The intermediate solution is taken as the optimal solution, that is... Figure 11 The solutions are marked with a pentagram, and the optimal solutions are shown in Table 16.

[0238] Calculate the degree of difference based on Tables 12, 13, and 16. ,in , , , .

[0239] The difference indicates that, compared to other methods, the new method proposed in this invention... The difference from the optimal solution is minimized. Since no weighting factors were considered, the degree of difference was the greatest. Considering only the weights of influencing factors, its degree of difference is... and This demonstrates that weighting information has a significant impact on the final evaluation result in FMEA analysis. Therefore, it is essential for this invention to calculate the weighting information. Calculating weight information based on the dispersion of evaluation information does not align with the application scenario of FMEA, therefore its variance ratio is... Large. Through comparative analysis, the new method shows better applicability and objectivity.

[0240]

[0241]

[0242]

[0243]

[0244]

[0245]

[0246] Ultimately, this invention calculates... When introducing a tendency factor ,when At that time, the failure mode risk ranking results are as follows: Figure 12 As shown. From Figure 12 It can be seen that only the last two in the risk ranking... , Changes occur. When hour, Difference It is still far below , and Therefore, sensitivity analysis demonstrates that the newly proposed optimized FMEA exhibits strong robustness.

[0247] Those skilled in the art will understand that embodiments of the present invention can provide methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media containing computer-usable program code. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0248] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A lithium battery assembly risk analysis method, characterized by, The method comprises the following steps: Step S1, using a double-layer hesitant fuzzy language term set to evaluate the severity, occurrence frequency and detectability of failure modes in lithium battery assembly; wherein the double-layer hesitant fuzzy language term set comprises a double-layer language term set and a hesitant fuzzy language term set, and both are composed of evaluation information of evaluators; Step S2, calculating the weight of evaluators based on similarity measure, and calculating the weight of influence factors of severity, occurrence frequency and detectability based on water injection principle; The calculation steps of the weight of evaluators comprise: Step S211, obtaining the membership degree of evaluators in the influence factors of failure modes, calculating the non-membership degree, and generating an ideal solution; Step S212, calculating the Euclidean distance between the evaluators and the ideal solution; Step S213, calculating the similarity between the evaluators and the ideal solution; Step S214, calculating the weight of evaluators based on the similarity; Step S3, aggregating the evaluation information of evaluators by using a combination compromise ranking method, and calculating the risk value of failure modes; Step S3 comprises the following steps: Step S311, based on the weight of evaluators, aggregating the evaluation information by using a simple intuitionistic fuzzy weighted geometric operator to obtain the membership degree and the non-membership degree of evaluators; Step S312, calculating the weighted sum and weighted product set of the evaluation information to obtain the integrated result of the evaluation information; Step S313, calculating the intermediate risk value by using three scoring strategies; Step S314, calculating the final risk evaluation value by using the intermediate risk value calculated by using the three scoring strategies; Step S4, determining the optimal evaluation information by using a multi-objective aggregation technology; Step S4 comprises the following steps: Step S41, summarizing the method of ranking failure modes in the risk ranking, generating a corresponding set, and calculating the cumulative number of times of failure modes in the risk ranking; Step S42, calculating the evaluation value of the actual participation in the final risk calculation according to different evaluation information of evaluators on failure modes under influence factors; Step S43, setting a distribution matrix according to the weight of the risk ranking; Step S44, setting a multi-objective mathematical model, and the formula of the objective function is as follows: ； ； The constraint condition is as follows: ； in, Indicates the number of failure modes. ; Indicates the number of influencing factors. ; Indicates the number of evaluators. ; This indicates a risk ranking. ; To optimize the number of methods, ; Failure mode In risk ranking The cumulative number of times it appears above, of which, ; The weights for ranking by risk are represented as To assign elements in the matrix; Failure mode The set of methods ranked in risk ranking f. Failure mode In risk ranking The number of times it appears above; Representation optimization method Failure modes Influencing factors Evaluation information, Failure modes for solving multi-objective models Influencing factors Optimal evaluation information; ； Representation optimization method Influencing factors The weight, Representation optimization method Intermediate evaluators The weight, For the evaluator k, the failure modes Evaluation information under influencing factor j; and For evaluation information The scoring function; ; By comparing the difference between the calculation results of different failure mode optimization methods and the optimal solution, the advantages and disadvantages of the optimization methods are determined, and the optimal evaluation information is determined.

2. The lithium battery assembly risk analysis method of claim 1, wherein , In the step S1, the double-layer language terminology set includes a main layer language terminology set and a sub layer language terminology set, and its expression is as follows: ； The main layer language terminology set For describing the underlying semantics, including the following expressions: ； The secondary layer language term set For modifying the semantic strength, including the following expression: ； wherein, is a boundary parameter for the primary layer language term set, is an index value for the primary layer language term in the primary layer language term set, is a boundary parameter for the secondary layer language term set, is an index value for the secondary layer language term set in the primary layer language term set.

3. The lithium battery assembly risk analysis method of claim 1, wherein , In the step S1, the expression of the hesitant fuzzy language term set is as follows: ； wherein, is a set of finite, ordered, continuous language term sets on S3, is a language term set, , is a non-empty set, is is an element of In In represents an element in the set of hesitant fuzzy linguistic terms whose expression is given by: ； wherein is a finite and continuous set of language terms on S3; The hesitant fuzzy language term set is converted into membership degree by using a conversion function.

4. The lithium battery assembly risk analysis method of claim 1, wherein, In step S2, the calculation steps of the influence factor weight comprise: Step S221, calculating the mean and variance of the evaluation matrix; Step S222, calculating the water injection level; Step S223, calculating the influence factor weight based on the water injection level.

5. The lithium battery assembly risk analysis method of claim 1, wherein, In step S313, the formulas of the three scoring strategies are as follows: ； ； ； wherein, is the intermediate risk value for the first scoring strategy, is the intermediate risk value for the second scoring strategy, is the intermediate risk value for the third scoring strategy; is the weighted sum integration, is the weighted product integration; is the inclination factor.

6. The lithium battery assembly risk analysis method of claim 1, wherein, In step S314, the calculation formula of the final risk evaluation value is as follows: ； wherein, is the final risk assessment value.

7. The lithium battery assembly risk analysis method of claim 1, wherein, In the step S44, the calculation formula of the difference degree is as follows: ； wherein, denotes an optimization method ranking of failure modes , failure modes for a multi-objective model optimal ranking denotes an optimization method evaluation information of failure modes , failure modes for a multi-objective model evaluation information of failure modes optimal evaluation information , is a maximum value of , is a maximum value of .

Images