A method for evaluating the influence of a battery production process step on battery consistency

Abstract

The application discloses a method for evaluating influence of a battery production process link on battery consistency, and belongs to the technical field of battery production process capability analysis. The method introduces the idea of the truncated mean method when constructing a fuzzy consistency matrix, processes data of collected expert score sheets, reduces influence of subjective factors of experts on the consistency analysis system and influence of extreme data in expert scoring, then converts the consistency index into a fitness function, constructs an initial particle swarm by using the modified fuzzy consistency matrix and factor weight, combines particle swarm optimization calculation to obtain the weight, realizes consistency matrix discriminant test, and establishes a power battery production process consistency analysis system. This process does not need repeated scoring of experts, and does not need repeated calculation of the consistency verification and matrix correction process, so that the calculation amount is greatly reduced, and the weight most meeting the consistency requirement is obtained by particle swarm optimization.

Description

TECHNICAL FIELD

[0001] The application relates to a method for evaluating the influence of a battery production process link on battery consistency, and belongs to the technical field of battery production process capability analysis. BACKGROUND

[0002] A power battery is a key component of a power device such as a new energy vehicle, and is often applied in a group form. The key of the group form is the consistency degree of the battery product. The consistency degree of the battery product refers to the convergence of important characteristic parameters of a group of power battery monomers, such as voltage, capacity and the like. Because the power battery is supplied in a group form, if the consistency of the power battery monomers is poor, the battery life will be attenuated too fast, the battery power supply will be unstable, and significant safety hazards will be brought.

[0003] The production process of the power battery involves 12 process links. Different process links involve the selection process of a variety of materials, the reaction process of a variety of chemical materials, and the production process under the influence of different temperatures and humidities. The complex process will inevitably cause differences between the power battery monomers. The influence degree of each process link on the battery consistency is not consistent. In order to optimize the battery production process, it is necessary to evaluate the influence degree of each process link on the battery consistency.

[0004] Initially, the evaluation of the influence degree of each process link on the battery consistency depends on the experience of technical personnel, but the experience varies from person to person and has great subjectivity, so it cannot be accurately evaluated, and the influence degree of each process link on the battery consistency cannot be quantitatively compared. With the progress of technology, at present, the methods proposed at home and abroad mainly include the analytic hierarchy process and its derivative methods. However, although the analytic hierarchy process has been widely used in the establishment of evaluation and analysis systems, the algorithm still has some drawbacks at this stage. For example, when the analytic hierarchy process and the fuzzy analytic hierarchy process are used, expert scoring is needed to establish a consistency judgment matrix. However, in the battery production process with multiple links, strong coupling and multiple variables, the consistency judgment matrix established does not necessarily meet the consistency requirements. At this time, the factor weight corresponding to the consistency judgment matrix does not have credibility, and the consistency judgment matrix needs to be modified. The so-called modification process requires experts to repeatedly score until the matrix meets the consistency. However, the consistency verification and matrix modification process has a large amount of calculation, and repeated modification will cause waste of calculation time. At the same time, the traditional expert scoring is easy to produce extreme data, which will also affect the credibility of the weight calculation result. SUMMARY

[0005] In order to solve the problem of battery production process consistency analysis, reduce the influence of extreme data in expert scoring, and avoid repeated solving of weight in the process of correcting fuzzy consistency judgment matrix, and reduce the significant calculation amount brought by the traditional consistency weight solving method, the application provides a method for evaluating the influence of battery production process on battery consistency, which is based on fuzzy swarm intelligence truncation scoring mechanism for analysis, and the idea of truncation mean method is introduced to process the data of the collected expert scoring table when constructing the fuzzy consistency matrix, so as to reduce the influence of subjective factors of experts on the consistency analysis system; Then the consistency index is converted into the fitness function, and the initial particle swarm is constructed by using the corrected fuzzy consistency matrix and the factor weight, and the weight is obtained by combining the particle swarm optimization calculation, the consistency matrix discrimination test is realized, and the power battery production process consistency analysis system is established. This process does not need repeated scoring of experts, and does not need repeated calculation of consistency verification and matrix correction process, so the calculation amount is greatly reduced, and the weight most consistent with the consistency requirement is obtained by particle swarm optimization, and the existing method only obtains the weight consistent with the consistency requirement, and does not further optimize the weight most consistent with the consistency requirement in the range of the weight consistent with the consistency requirement.

[0006] A method for evaluating the influence of battery production process on battery consistency, the method comprises:

[0007] Step1: constructing an n-dimensional fuzzy consistency matrix R according to the expert scoring table e , the expert scoring table is the preliminary evaluation data of the influence degree of factors in each process of battery production on battery consistency;

[0008] Step2: taking each element in the fuzzy consistency matrix R e as the initial particle swarm, setting the maximum iteration number T, setting the consistency optimization function CIF(n) as the fitness function of the particle swarm algorithm, setting the particle position value range [0, b] and the particle motion speed value range [0, c]; taking the corresponding weight w i of each level factor as the optimization variable, a total of (n-1)(n-2) / 2+n optimization variables, i=1, 2, …n;

[0009] Step3: taking the initialized particle as the global optimal particle, recording the fitness function value; updating the particle position according to the speed, and recording the fitness function value corresponding to the updated particle, and comparing with the fitness function value corresponding to the global optimal particle, if it is smaller, the current population particle is updated to the global optimal particle;

[0010] Step4: complete optimization; when the iteration number t=T, end the iteration, and output the global optimal particle, obtain the corrected fuzzy consistency matrix Y and the corresponding weight w i of each level factor;

[0011] Step5: calculate the combined weight of each level factor on the target layer, build a battery process consistency analysis system, that is, determine the weight value system of the influence degree of factors in each process link of battery production on battery consistency.

[0012] Optionally, the Step1 constructs an n-dimensional fuzzy consistency matrix R e , including:

[0013] The expert scoring table is preprocessed using the truncation mechanism:

[0014] Suppose that a total of w bits of expert scoring tables are received, remove p parts of scoring data with extremely large scores and q parts of scoring data with extremely small scores, and then perform an average operation to obtain the fuzzy consistency matrix R e :

[0015]

[0016] wherein r ij,e = 0.5 + a(w i -w j ), i, j = 1, 2, …, n; a is a measurement unit of the important degree difference between two elements w i and w j , and the value thereof satisfies

[0017] Optionally, the consistency optimization function CIF(n) is:

[0018]

[0019] Constraint condition:

[0020] min CIF(n)

[0021]

[0022] wherein Y = (y ij ) n×n is the modified fuzzy consistency matrix;

[0023] When CIF(n) < 0.1, it is determined that the fuzzy consistency matrix obtained by particle swarm optimization satisfies consistency, and the weight of the corresponding element obtained by optimization has credibility.

[0024] Optionally, in the Step3:

[0025] Suppose that in an N-dimensional target search space, there are m particles to form a group, wherein the position of the i th particle is represented as that is, the position of the i The positions of the particles in the N-dimensional search space are

[0026] Each space represents a solution, and the fitness value of each particle is obtained according to the fitness function.

[0027] Based on the constraints, the best position experienced by the individual particle, which has the lowest fitness, is denoted as . The best position experienced by all particles in the entire swarm is denoted as No. The velocity of each particle is denoted as .

[0028] The particles are updated according to the following update formula.

[0029]

[0030] in, d = 1, 2, ..., N; β is a non-negative number called the inertia factor; acceleration constants c1 and c2 are non-negative constants; r1 and r2 are random numbers that vary within the range [0, 1]; α is called the constraint factor, used to control the weight of the velocity;

[0031] That is, particles Flight speed By a maximum speed limit.

[0032] Optionally, Step 5: Calculating the combined weights of the influence of factors at each level on the target layer, including:

[0033] After obtaining the weights corresponding to the factors at each level, calculate the combined weights.

[0034]

[0035] Where l represents the number of levels and m represents the number of factors, thus obtaining the influence weight of each sub-level factor on the target level.

[0036] Optionally, a certain level of factor u in the expert scoring table i Compared to lower-level factors u j The criterion for comparing the importance of items uses the 0.9 scale.

[0037] Optionally, when building a battery process consistency analysis system, the method divides the battery production process into three process stages: the wafer fabrication process stage, the assembly process stage, and the sorting and assembly process stage.

[0038] Optionally, the tabletting process stage includes five process links of material pretreatment, material stirring, coating, rolling and slicing, and mainly functions to process the base material, then coat and roll to obtain a group of battery pole pieces, and finally slice to obtain the battery pole pieces according to the customized production needs of the battery.

[0039] Optionally, the assembly process stage includes five process links of tab welding, lamination, edge forming, vacuum drying and electrolyte injection, and mainly functions to complete the forming manufacturing of the battery on the basis of the base material to obtain the soft package battery.

[0040] Optionally, the sorting and assembling process stage includes two process links of formation and aging, and mainly functions to activate the formed battery, and then detect and sort the battery according to the consistency to obtain the qualified battery product.

[0041] The present application has the following beneficial effects:

[0042] By introducing the idea of the truncated mean method when constructing the fuzzy consistency matrix, the data collected from the expert scoring table is processed, the influence of the subjective factors of the experts on the consistency analysis system and the influence of the extreme data in the expert scoring are reduced; then the consistency index is converted into the fitness function, and the initial particle swarm is constructed by using the modified fuzzy consistency matrix and the factor weight, and the weight is calculated by combining the particle swarm optimization calculation, the consistency matrix modification and verification are realized, and the consistency analysis system of the power battery production process is established. The group intelligence algorithm is used for optimization, the modification of the consistency matrix, the consistency verification and the weight calculation are integrated, and the calculation process of the method is more simple, the manual calculation is greatly reduced, and the weight index obtained by using the optimization characteristics of the group intelligence algorithm has better consistency, that is, higher reliability and scientificity. BRIEF DESCRIPTION OF DRAWINGS

[0043] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0044] Figure 1 is a flowchart of a method for evaluating the influence of the battery production process link on the consistency of the battery based on group intelligence according to an embodiment of the present application.

[0045] Figure 2 is a specific soft package battery production process flowchart according to an embodiment of the present application. DETAILED DESCRIPTION

[0046] In order to make the objectives, technical solutions and advantages of the present application clearer, the embodiments of the present application will be further described in detail below with reference to the drawings.

[0047] Embodiment one:

[0048] The embodiment provides a method for evaluating the influence of a battery production process link on battery consistency based on swarm intelligence, please refer to Figure 1 , the method comprises:

[0049] Step 1: constructing an n-dimensional fuzzy consistency matrix R according to an expert scoring table e , the expert scoring table is the preliminary evaluation data of the influence degree of factors in each process link of battery production on battery consistency;

[0050] Step 2: taking each element in the fuzzy consistency matrix R e as an initial particle swarm, setting the maximum number of iterations T, setting the consistency optimization function CIF(n) as the fitness function of the particle swarm algorithm, setting the particle position value range [0, b] and the particle motion speed value range [0, c]; taking the corresponding weight w i of each level factor as an optimization variable, a total of (n-1)(n-2) / 2+n optimization variables, i=1, 2, …n;

[0051] Step 3: taking the initialized particle as the global optimal particle, recording the fitness function value; constantly updating the particle position according to the speed, and recording the fitness function value corresponding to the updated particle, and comparing it with the fitness function value corresponding to the global optimal particle, if it is smaller, then updating the current population particle to the global optimal particle;

[0052] Step 4: completing the optimization; when the iteration number t=T, ending the iteration, and outputting the global optimal particle, obtaining the corrected fuzzy consistency matrix Y and the corresponding weight w i of each level factor;

[0053] Step 5: calculating the combined weight of each level factor on the target layer, building a battery process consistency analysis system, that is, determining the weight value system of the influence degree of factors in each process link of battery production on battery consistency.

[0054] When the battery production process is optimized subsequently, the weight value system obtained in Step 5 can be referred to for optimizing the battery production process.

[0055] Embodiment two:

[0056] The embodiment provides a method for evaluating the influence of a battery production process link on battery consistency based on swarm intelligence, taking a soft package battery actual production process as an example for illustration, and the method comprises the following steps:

[0057] Step one: obtaining expert scoring data and constructing a fuzzy consistency matrix;

[0058] According to the knowledge reserve of experts and the operation experience of first-line engineers, professional knowledge and production experience are quantitatively represented by designing a scoring mechanism.

[0059] According to the grading method commonly used in the industry, the soft package battery actual production process is split into four levels, and the soft package battery actual production process can be divided into four levels of factor layers. The first level of factor layers includes tabletting process, assembly process and sorting and assembly process. The second level of factor layers includes pretreatment, material stirring, coating, rolling, slicing, tab welding, lamination, edge forming, vacuum drying, electrolyte injection, formation and aging, which can be referred to as Figure 2 The specific division of the third level of factor layers and the fourth level of factor layers can refer to the content in Table 4 below, which will not be described here. The embodiment will be described below by taking the 12 process links in the second level of factor layers as an example.

[0060] It is assumed that there are n factors corresponding to the factor layer, and a factor set U = {u1, u2, … u n} is formed. A factor in a certain layer and a next factor in the same level of the factor form a row and a column of a fuzzy consistency matrix R = {r ij} n×n , wherein r ij represents the importance of the level factor u i compared with other factors u j in the same level.

[0061] For example, for the second level of factor layers, three factor sets are constructed, the first factor set corresponds to n = 5, the second factor set corresponds to n = 5, and the third factor set corresponds to n = 2. For the second level of factor layers, r ij may represent the importance of “pretreatment” relative to “slurry stirring uniformity”.

[0062] In the scoring process, the comparison criteria of factor u i and factor u j adopt the 0.9 scale method, as shown in Table 1:

[0063] Table 1: 0.9 scale method

[0064]

[0065] According to the above scoring mechanism, the scoring table is issued to industry experts and experienced first-line engineers, and the corresponding scoring table is designed according to each layer factor, and a total of N copies are visible. N is also the number of scoring table styles required for system establishment. According to the obtained scoring table, the fuzzy consistent matrix R1, R2 and R3 for the second factor layer are constructed as follows:

[0066]

[0067] According to the constructed fuzzy consistent matrix R1, R2 and R3, the weight of each process in the second factor layer of the production process is calculated;

[0068] The method proposed in the application is used to calculate the weight of the corresponding factors of the production process, and compared with the row and normalization method and the fuzzy consistency inference method. The comparison results are shown in Table 3:

[0069] Table 3: Scoring table designed by taking the second factor layer of the power battery production process analysis system as an example

[0070]

[0071]

[0072] For the problem of battery process consistency analysis of target structure complex, the working experience of professionals is quantified by fuzzy analytic hierarchy process, so as to establish a scientific and reliable analysis system.

[0073] Considering that the establishment of the fuzzy consistent matrix R={r ij} n×n is influenced by personal subjective thought in the scoring process, individual extreme scores will greatly affect the credibility of the fuzzy consistent matrix. Therefore, the idea of truncated mean is introduced on the basis of the traditional scoring mechanism, which excludes the extreme influence on the basis of maintaining the diversity of samples, and pre-processes the same group of data.

[0074] Suppose that a total of w experts received the scoring table, remove p copies of scoring data with extremely large scores and q copies of scoring data with extremely small scores, and then perform the mean value operation to obtain a new fuzzy consistent matrix R e , that is,

[0075]

[0076] r ij,e =0.5+a(w i -w j ),i,j=1,2,…,n, (2)

[0077] wherein, the weight w i (i=1,2,…n) represents the relative importance of the corresponding layer factor to the upper level factor, that is, the influence. a is the two elements wi and w j The value of a satisfies The greater the value of a, the more the decision maker values the importance difference between factors. ij,e w represents the relative importance of each link in the battery process to consistency, and w represents the degree of influence of the corresponding process link or parameter on the consistency of the battery process.

[0078] Step two:

[0079] The production process of power batteries involves 12 complex processes and dozens of indicators. The entire production process involves many links and complex processes, and there is a large amount of coupling between related indicators. To further improve consistency and ensure the feasibility of the established process consistency analysis system, a fuzzy consistency swarm intelligence algorithm is proposed below to solve the fuzzy consistency matrix R e and the weight w i of the corresponding factor (i = 1, 2, … n) through the optimization ability of the swarm intelligence algorithm.

[0080] Considering the influence of human subjective factors and randomness in the actual battery production process scoring, it is generally difficult to meet the conditions of the fuzzy consistency matrix judgment described in formula (3), that is, r ij,e ≠ 0.5 + a (w i -w j ). For this case, it is generally considered acceptable when the deviation is small enough. The smaller the deviation, the higher the degree of acceptance. Thus, the consistency test index is:

[0081]

[0082] If the fuzzy consistency matrix R e does not satisfy the consistency, the consistency matrix obtained by the battery process consistency analysis in this case is not reliable, and the fuzzy consistency matrix R e needs to be corrected to obtain a matrix Y = (y ij ) n×n that satisfies the consistency after correction, and the corresponding factor weight is recalculated. The factor weight represents the degree of influence of the factor on the process consistency.

[0083] Assuming that the first row of the fuzzy consistency matrix R e before correction has the highest credibility, the corrected matrix Y = (y ij ) n×n is considered Then, z ij is determined by subtracting from the remaining rows of the Y matrix , and Solve for z ij The mean of the values. Then, by using the fact that the difference between any specified row and the corresponding elements of the remaining rows in the fuzzy consistency matrix is ​​a constant, the consistency test index can be obtained:

[0084]

[0085] The corrected matrix Y = (y ij ) n×n element y in ij and the corresponding factor weights w i (i = 1, 2, ..., n) are used as the scalars to be determined, and a particle population is constructed. Based on the above consistency check index, the consistency optimization function is obtained:

[0086]

[0087] Subsequently, the consensus optimization function CIF(n) is used as the fitness function of the fuzzy consensus swarm intelligence algorithm, and the minimum value of the fitness function is searched under constraints.

[0088]

[0089] The smaller the fitness function CIF(n), the better the corrected fuzzy consistency matrix Y = (y ij ) n×n The better the consistency, the more reliable the fuzzy consistency matrix obtained by particle swarm optimization is when CIF(n) < 0.1.

[0090] Step 3:

[0091] Solve for the corrected matrix Y = (y ij ) n×n middle y 1j =r 1j Furthermore, combining the unity and complementarity of fuzzy consistent matrices, we can know that Y = (y ij ) n×n The elements in the top right corner of the matrix, excluding the first row, are unknown optimization variables. Simultaneously, the corresponding weights w for each level of factors are assigned. i There are a total of (n-1)(n-2) / 2+n optimization variables, denoted as N = (n-1)(n-2) / 2+n.

[0092] Step Four:

[0093] Suppose there is a swarm of m particles in an N-dimensional target search space, where the nth particle is the first particle in the swarm. Particles The position is represented as That is, the first The positions of the particles in the N-dimensional search space are

[0094] Each space represents a solution, and the fitness value of each particle is obtained according to the fitness function.

[0095] Based on the constraints, the best position experienced by the individual particle, which has the lowest fitness, is denoted as . The best position experienced by all particles in the entire swarm is denoted as particle The speed is denoted as

[0096] The particles are updated according to the following update formula.

[0097]

[0098]

[0099] in, d = 1, 2, ..., N; β is a non-negative number called the inertia factor; the acceleration constants c1 and c2 are non-negative constants; r1 and r2 are random numbers that vary within the range [0, 1]; α is called the constraint factor, which is used to control the weight of the velocity.

[0100] That is, particles Flight speed By a maximum speed limit.

[0101] Step 5:

[0102] The iteration ends after the number of iterations t = T, and the globally optimal particle is output, yielding the corrected fuzzy consistency matrix Y and the corresponding weights w of the factors. i (i = 1, 2, ..., n).

[0103] Step Six:

[0104] After obtaining the weights corresponding to the factors at each level, calculate the combined weights.

[0105]

[0106] Where l represents the number of levels and m represents the number of factors, thus obtaining the influence weight of each sub-level factor on the target level.

[0107] The battery manufacturing process mainly includes 12 process steps, which can be divided into three stages: wafer fabrication, assembly, and sorting and assembly. The battery manufacturing process is as follows: Figure 2 As shown.

[0108] The tabletting process stage includes five process links of material pretreatment, material stirring, coating, rolling, and slicing, and mainly functions to process the base material, then coat and roll to obtain a group of battery pole pieces, and finally slice to obtain the battery pole pieces according to the customized production needs of the battery.

[0109] The assembly process stage includes five process links of tab welding, lamination, forming and edge sealing, vacuum drying, and electrolyte injection, and mainly completes the forming manufacturing of the battery on the basis of the base material to obtain the soft package battery. The sorting and assembly process stage includes two process links of formation and aging, and mainly completes the activation of the formed battery, and then detects and sorts the battery according to the consistency to obtain the qualified battery product. In the face of different production needs of the battery, in order to better manage the production process of the battery, adjust the process link, and design a scientific production scheme, the present application designs a set of battery consistency process link analysis system according to the actual production process and process link, combined with the technical points and process parameters, and the corresponding battery consistency analysis factor set is shown in Table 2:

[0110] Table 2: Analysis factor set in the battery consistency process link analysis system proposed in the present application

[0111]

[0112]

[0113] In the present example, within a predetermined time range, after steps one to six are performed, the battery production process consistency analysis system is shown in Table 4.

[0114] Table 3 is a comparison of the results of process factor weight calculation using different methods; and Table 4 is a quantified battery process analysis consistency system obtained by using the method of the present application.

[0115] Table 3: Comparison of results of process factor weight calculation using different methods

[0116]

[0117] As can be seen from Table 3, in the weight calculation of the same matrix, the method of calculating the weight and the corresponding CIF value is directly calculated, while the fuzzy consistency swarm intelligence algorithm proposed in the present application first automatically corrects the fuzzy consistent matrix, and then calculates the weight of the corrected matrix and the CIF value corresponding to the corrected matrix. The CIF obtained by the fuzzy consistency swarm intelligence algorithm is smaller, the consistency is better, and has higher credibility. At the same time, when the fuzzy consistent matrix does not satisfy the consistency, the process of manually correcting the matrix is avoided, and the calculation amount is reduced.

[0118] Table 4: Quantified battery process analysis consistency system using the method of the present application

[0119]

[0120]

[0121]

[0122] From Table 4, the influence weight of each process link and each process parameter involved in the process link on the consistency of the battery can be obtained. The greater the factor weight, the greater the influence of the factor on the final consistency of the battery. By combining the weight to calculate the influence of each process parameter on the consistency of the battery, the weight set of each level on the influence of the process consistency is obtained as follows:

[0123] The weight set w of the first level factor layer I = (0.576, 0.364, 0.060);

[0124] The weight set w of the second level factor layer II = (0.1037, 0.0904, 0.1682, 0.1192, 0.0945, 0.0419, 0.0724, 0.0844, 0.0728, 0.0921, 0.0210, 0.0390);

[0125] The weight set w of the third level factor layer III = (0.0206, 0.0830, 0.0452, 0.0447, 0.0642, 0.0592, 0.0358, 0.0341, 0.0494, 0.0944, 0.0232, 0.0187, 0.0724, 0.0242, 0.0603, 0.0456, 0.0272, 0.0139, 0.0156, 0.0203, 0.0424, 0.0210, 0.0390);

[0126] The weight set w of the fourth level factor layer IV= (0.0104, 0.0056, 0.0046, 0.0165, 0.0665, 0.0245, 0.0065, 0.0142, 0.0041, 0.0083, 0.0132, 0.0114, 0.0082, 0.0132, 0.0316, 0.0519, 0.0123, 0.0195, 0.0104, 0.0294, 0.0097, 0.0056, 0.0087, 0.0036, 0.0087, 0.0286, 0.0055, 0.0444, 0.0050, 0.0329, 0.0097, 0.0520, 0.0053, 0.0093, 0.0086, 0.0046, 0.0075, 0.0066, 0.0198, 0.0246, 0.0087, 0.0151, 0.0043, 0.0040, 0.0068, 0.0072, 0.0302, 0.0122, 0.0178, 0.0456, 0.0060, 0.0055, 0.0090, 0.0063, 0.0056, 0.0077, 0.0006, 0.0048, 0.0108, 0.0064, 0.0139, 0.0255, 0.0130, 0.0047, 0.0163, 0.0314, 0.0076).

[0127] From the combination weight set values, it can be found that coating, rolling and material pretreatment links in the battery production link are the top three process links affecting the consistency of the battery production, and in the actual battery production process, the coating area size precision, material selection and rolling quality are the most critical process indexes affecting the consistency.

[0128] In the actual production process, when a large batch of customized production problems are encountered, the fuzzy consistency swarm intelligence algorithm based on the truncated scoring mechanism can be used to establish a quantitative power battery production process consistency analysis system, and the production scheme is adjusted according to the obtained factor weight.

[0129] Part of the steps in the embodiments of the application can be realized by software, and the corresponding software program can be stored in a readable storage medium, such as an optical disc or a hard disk.

[0130] The above description is only the preferred embodiment of the present application, and is not intended to limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for evaluating the impact of battery manufacturing process stages on battery consistency, characterized in that, The method divides the battery production process into four levels of factors. It is assumed that a certain factor level corresponds to n factors, forming a factor set U = {u1, u2, ... u...} n The hierarchical factors in a certain factor layer and other factors under the same factor layer constitute a fuzzy consistency matrix R. e ={r ij } n×n The rows and columns of r, where r ij This indicates the hierarchical factor u of the factor layer. i Compared to other factors in this factor layer, u j The degree of importance; The four-level factor layer includes: Level I factor layer, the factors are composed of process stages, and each factor is represented as 1, 2, 3; Level II factor layer: the factors are composed of process links, and each factor is represented as a, b, ... l; Level III factor layer, the factors are composed of process link indicators, and each factor is represented as a. n ,…l n ; Level IV factor layer, the factors are composed of process step index parameters, each factor is represented as a nn ,…l nn ; The method includes: Step 1: Construct an n-dimensional fuzzy consistency matrix R based on the expert scoring table. e The expert scoring sheet is a preliminary assessment of the impact of factors in each process of battery production on battery consistency. Step 2: Transform the fuzzy consistency matrix R... e Each element is used as the initial particle swarm, the maximum number of iterations T is set, the consistency optimization function CIF(n) is set as the fitness function of the particle swarm algorithm, and the particle position range [0, b] and particle velocity range [0, c] are set; the corresponding weights w of each level factor are assigned. i There are a total of (n-1)(n-2) / 2+n optimization variables, i = 1, 2, ..., n; Step 3: Using the initial particle as the global optimal particle, record the fitness function value; continuously update the particle position according to the velocity, and record the fitness function value corresponding to the updated particle, and compare it with the fitness function value corresponding to the global optimal particle. If it is smaller, update the current population particle to the global optimal particle. Step 4: Complete the optimization; after the number of iterations t = T, end the iteration and output the globally optimal particle, obtaining the corrected fuzzy consistency matrix Y and the corresponding weights w of each level of factors. i ; Step 5: Calculate the combined weights of the factors at each level on the target layer, and build a battery process consistency analysis system, that is, determine the weight value system of the degree of influence of factors in each process of battery production on battery consistency.

2. The method according to claim 1, characterized in that, In Step 1, an n-dimensional fuzzy consistency matrix R is constructed based on the expert scoring table. e ,include: Preprocessing the expert scoring sheet using a truncation mechanism: Assuming a total of w expert rating tables are received, after removing p extremely high ratings and q extremely low ratings, the fuzzy consistency matrix R is obtained by averaging the results. e : Where, r ij，e =0.5+a(w i -w j ), i, j = 1, 2, ..., n; a is a two-element w i and w j The unit of measurement for the difference in importance between them, whose value satisfies 3. The method according to claim 2, characterized in that, The consistency optimization function CIF(n) is: Constraints: minCIF(n) Assume the uncorrected fuzzy consistency matrix R e The first row is the most reliable, therefore the corrected matrix Y = (y ij ) n×n middle Use again Subtract from the remaining rows of the Y matrix Determine z ij , For z ij The mean; When CIF(n) < 0.1, it is determined that the fuzzy consistency matrix obtained by particle swarm optimization satisfies consistency, and the weights of the corresponding elements obtained by optimization are reliable.

4. The method according to claim 3, characterized in that, In Step 3: Suppose there is a swarm of m particles in an N-dimensional target search space, where the nth particle is the first particle in the swarm. The position of each particle is represented as That is, the first The positions of the particles in the N-dimensional search space are Each space represents a solution, and the fitness value of each particle is obtained according to the fitness function. Based on the constraints, the best position experienced by the individual particle, which has the lowest fitness, is denoted as [position name missing]. The best position experienced by all particles in the entire swarm is denoted as No. The velocity of each particle is denoted as . The particles are updated according to the following update formula. in, d = 1, 2, ..., N; β is a non-negative number called the inertia factor; acceleration constants c1 and c2 are non-negative constants; r1 and r2 are random numbers that vary within the range [0, 1]; α is called the constraint factor, used to control the weight of the velocity; That is, particles Flight speed By a maximum speed limit.

5. The method according to claim 4, characterized in that, Step 5: Calculate the combined weights of the influence of factors at each level on the target layer, including: After obtaining the weights corresponding to the factors at each level, calculate the combined weights. Where l represents the number of levels and m represents the number of factors.

6. The method according to claim 5, characterized in that, The hierarchical factor u in a certain factor layer of the expert scoring table i Compared to other factors j The criterion for comparing the importance of items uses the 0.9 scale.

7. The method according to claim 6, characterized in that, The method divides the battery production process into three stages when building a battery process consistency analysis system: the wafer fabrication stage, the assembly stage, and the sorting and assembly stage.

8. The method according to claim 7, characterized in that, The electrode manufacturing process includes five steps: material pretreatment, material mixing, coating, rolling, and slicing. Its main function is to process the base material, then coat and roll it to obtain a group of battery electrode sheets, and finally slice it according to the customized production needs of the battery to obtain the battery electrode sheets.

9. The method according to claim 7, characterized in that, The assembly process includes five steps: electrode welding, stacking, molding and sealing, vacuum drying, and electrolyte injection. It mainly involves completing the molding and manufacturing of the battery based on the basic materials to obtain a soft-pack battery.

10. The method according to claim 7, characterized in that, The sorting and assembly process includes two steps: formation and aging. The main steps are to activate the formed batteries and then sort them according to their consistency to obtain qualified battery products.

Images