Method for determining process parameters for producing a memory cell and method for producing a memory cell

Abstract

The invention relates to a method (23) for determining process parameters (24, 25) for producing a memory cell (2) by means of an electronic computing device (19), comprising the steps of: determining (30) desirabilities (26, 27) for target variables (16, 17) of the memory cell (2) by means of desirability functions (28, 29), which each assign a number from a predetermined interval to a respective value (16a, 17a) of the respective target variables (16, 17), determining (34) predicted values ​​(16b, 17b) of the respective target variable (16, 17) by means of a prediction model (33) generated by regression analysis (32), which describes the target variables (16, 17) as a function of the process parameters (24, 25), and determining (35) predicted desirabilities (26a, 27a) for the target variables (16, 17) by means of the desirability functions (28, 29), determining (37) at least one respective optimized value (24b, 25b) of the respective process parameter (25,25) using an optimization method (36).,

Description

[0001] The invention relates to a method for determining process parameters for producing a storage cell for an electrical energy storage device by means of an electronic computing device according to claim 1. Furthermore, the invention relates to a method for producing a storage cell for an electrical energy storage device, in which the storage cell is produced by means of optimized values ​​of process parameters determined by means of such a method.

[0002] Storage cells for electrical energy storage devices, for example for motor vehicles, are already well known in the general state of the art. A manufacturing method for producing the storage cells can comprise a multitude of predeterminable or adjustable process parameters. These process parameters influence the production of the storage cell and thus in particular target parameters of the storage cell to be produced or produced. These target parameters can be understood in particular as quality properties of the storage cell. Relationships between the process parameters and the resulting target parameters of the storage cell can usually be particularly complex, as a result of which optimal process parameters for producing the storage cell may be unknown or particularly difficult to determine.

[0003] It is an object of the invention to provide a method for determining process parameters for producing a storage cell for an electrical energy storage device and a method for producing a storage cell for an electrical energy storage device, so that the storage cell can be produced with particularly advantageous properties with particularly low expenditure.

[0004] This object is achieved according to the invention by a method for determining process parameters for producing a storage cell for an electrical energy storage device having the features of patent claim 1 and by a method for producing a storage cell for an electrical energy storage device having the features of patent claim 10. Advantageous embodiments of the invention are the subject of the dependent patent claims and the description.

[0005] A first aspect of the invention relates to a method for determining, in particular for calculating, in particular optimized or ideal, process parameters for producing at least one storage cell for an electrical energy storage device using an electronic computing device. The method can be understood, in particular, as a method for determining optimized values ​​of the process parameters, i.e., for determining at least one respective optimized value of the respective process parameter.

[0006] Process parameters can be understood, in particular, as parameters that are adjustable and / or predeterminable during the manufacture of the memory cell, in particular in a targeted manner, or are adjusted and / or predeterminable in order to manufacture the memory cell, in particular depending on the process parameters. The memory cell is or will thus be influenced by the process parameters. The memory cell can therefore be manufactured depending on the process parameters. Parameters can be understood, in particular, as variables.

[0007] Target variables of the storage cell, in particular the one manufactured or to be manufactured, are or will be influenced, for example, by the process parameters. In other words, the target variables depend on the process parameters. Target variables can be understood, in particular, as quality properties or quality characteristics of the storage cell.

[0008] The electrical energy storage device is preferably intended for a motor vehicle. Thus, the electrical energy storage device is, for example, an electrical energy storage device for the motor vehicle or the motor vehicle. The motor vehicle is, for example, designed as a motor vehicle, in particular as a passenger car.

[0009] The electrical energy store can be understood in particular as a battery or an accumulator. For example, the electrical energy store has at least one housing element, which can in particular be referred to as a storage housing. The housing element, for example, delimits at least one receiving space at least partially, in particular predominantly or completely. Preferably, the at least one storage cell can be arranged or is arranged in the receiving space. The storage cell is preferably designed as a battery cell. In particular, the storage cell is designed to store or chemically bind electrical energy.

[0010] The motor vehicle is designed, for example, as an electrically driven motor vehicle. This means that the motor vehicle has, for example, at least one electrical machine by means of which the motor vehicle can be driven, in particular purely electrically. The motor vehicle is thus designed, for example, as a battery-electric vehicle or as a hybrid vehicle. For example, the electrical machine can be supplied with electrical energy, in particular stored or chemically bound in the electrical energy store, by means of the electrical energy store in order to drive the motor vehicle. For example, the electrical energy store is designed as a high-voltage store, in particular as a high-voltage battery.

[0011] In order to be able to produce the storage cell, and in particular the electrical energy storage device, with particularly advantageous properties with particularly low expenditure, at least the following steps are carried out in the method, in particular by means of the electronic computing device: - Determining, in particular calculating, in particular dimensionless, desirabilities for the target variables of the memory cell by means of desirability functions which assign a respective value of the respective target variable of the memory cell, in particular the memory cell to be produced, in each case a respective number, in particular a real number, from a predetermined interval.This means that at least one first desirability, in particular for the first target variable, is determined, in particular calculated, by means of a first desirability function by means of the electronic computing device, wherein the first desirability function assigns a number, in particular a real number, from the predetermined interval to the values ​​of a first target variable of the memory cell, and at least one second desirability, in particular for the second target variable, is determined, in particular calculated, by means of a second desirability function by means of the electronic computing device, wherein the second desirability function assigns a number, in particular a real number, from the predetermined interval to the values ​​of a second target variable of the memory cell as desirability.In other words, the target variables, which could be quality requirements, for example, are transformed into a dimensionless, standardized number space using mathematical models in the form of desirability functions. In this case, in particular by technologists, quality engineers, and developers, specification limits can be defined and the target variables described in a first step. Subsequently, mathematical models in the form of desirability functions can be assigned to the target variables. Using the desirability functions, the target variables can be translated into unitless metrics, which enables, for example, a simple, particularly mathematical, comparability of the transformed target variables or transformed values ​​of the target variables.Desirability functions can be understood, in particular, as transformation functions, whereby the respective desirability function can also be referred to as a respective transformation function. The respective desirability can be understood, in particular, as a respective desirability characterizing the respective target variable. - Determining, in particular calculating, predicted values ​​of the respective target variable by means of a prediction model generated, in particular at least partially, predominantly or completely, by regression analysis, which describes the target variables as a function of the process parameters. In other words, the process parameters, in particular values ​​of the process parameters, are used as input variables for the prediction model in order to determine, in particular calculate, the predicted values ​​of the respective target variable as output variables by means of the prediction model, in particular by means of the electronic computing device. The fact that the prediction model is generated by regression analysis can be understood in particular to mean that the prediction model is based at least partially on the regression analysis and in particular has been created or formed using the regression analysis.A prediction model can be understood, in particular, as a production model that maps the manufacture of the memory cell as a function of the process parameters. The prediction model describing the target variables as a function of the process parameters can be understood, in particular, as mapping the target variables as a function of the process parameters, i.e., predicting or simulating them. - Determining, in particular calculating, predicted desirabilities for the target variables by means of the desirability functions, which each assigns a respective one of the predicted values ​​of the respective target variable of the memory cell, in particular the one to be produced, a number, in particular a real number, from the predetermined interval.This means that at least one first predicted desirability, in particular for the first target variable, is determined, in particular calculated, by means of the first desirability function, by means of the electronic computing device, wherein the first desirability function assigns a number, in particular a real number, from the predetermined interval to the predicted values ​​of the first target variable, in particular determined, and at least one second predicted desirability, in particular for the second target variable, is determined, in particular calculated, by means of the electronic computing device, wherein the second desirability function assigns a number, in particular a real number, from the predetermined interval to the predicted values ​​of the second target variable, in particular determined.In other words, the desirability function used to transform the values ​​of the respective target variable into the respective desirability is used again to transform the respective values ​​of the respective target variable predicted by the prediction model into the respective predicted desirabilities. The respective predicted desirability can be understood, in particular, as a respective dimensionless predicted desirability. - Determining, in particular calculating, at least one respective optimized or ideal value of the respective process parameter by means of at least one optimization method by minimizing at least one deviation, in particular all deviations, between the respective, in particular determined, desirability and the respective, in particular determined, predicted desirability.This means that at least one respective optimized value is determined for each process parameter by means of the electronic computing device, i.e., for example, at least one first optimized value is determined for a first of the process parameters and, for example, at least one second optimized value is determined for a second of the process parameters. This determination is carried out by means of the electronic computing device using the optimization method, which is based on minimizing the deviation between the respective desirability and the respective predicted desirability, i.e., for example, the deviation between the first desirability and the first predicted desirability and the deviation between the second desirability and the second predicted desirability. The optimized values ​​can in particular be referred to as optimal values.

[0012] This can be understood in particular as the following: For example, the desirability is first determined as a first component using the electronic computing device, and the prediction model is generated as a second component. The two components result in a multi-criteria optimization system, which is optimized, i.e., solved or approximated, using the optimization method in order to determine, in particular calculate, the optimal values ​​of the process parameters using the electronic computing device. The optimization system can be used to generate an optimum for several quality properties, for example, capacity or weight, based on targeted adjustment of the process parameters.

[0013] The invention is based in particular on the following findings and considerations: The production of storage cells, which are for example lithium-ion battery cells (LIB), can have enormous potential for increasing sustainability and reducing production costs, which can be reflected not least in a significantly high scrap rate, of for example 5 to 12 percent. One reason for this high scrap rate is that it can sometimes be unclear how the process parameters on individual production plants for manufacturing the storage cells are to be set in order to meet the quality characteristics or target values, for example in the form of capacity or weight. In principle, it is conceivable to use statistical experimental planning, which can in particular be referred to as Design of Experiments (DoE), to optimize the process for manufacturing or developing the storage cell.Process parameters can be varied according to specific specifications to investigate their influence on quality characteristics. However, the effort required to determine the influence of process parameters on the target variables can be particularly high in this case. Currently established approaches, such as statistical experimental design, typically cannot pursue a multi-criteria approach. This means that the state of the art does not provide a method for simultaneously optimizing and weighting multiple quality characteristics, such as weight or capacity, since, for example, an increase in capacity may be more relevant than a change in weight.

[0014] The method according to the invention can overcome the aforementioned disadvantages. Optimized process parameters or their optimized values ​​can be determined using the method according to the invention in order to be able to develop or manufacture the memory cell in such a way that it has particularly advantageous properties and is, in particular, optimal or optimized with regard to the target parameters. This can be achieved in a particularly low-cost manner using the method according to the invention.

[0015] By applying the method according to the invention, for example, the effort required for testing and / or the effort required to build prototypes can be kept particularly low, since the optimized or optimal process parameters can be determined virtually using the electronic computing device. Furthermore, conflicting objectives in process optimization can be avoided because a holistic optimization approach can be applied. One such conflicting objective can be, for example, increasing battery cell capacity by introducing more cell material, since this can lead to an undesirably high battery cell weight. Overall, it can be seen that the method according to the invention can be used to provide a multi-criteria optimization system for production, in particular of storage cells, and this optimization system can in particular be solved or approximated.

[0016] In a further embodiment, it is provided that, in order to minimize the, in particular respective, deviation(s), the determination of the predicted values ​​of the respective target variable and the determination of the predicted desirabilities by varying the values ​​of the process variables are carried out iteratively, in particular by means of the electronic computing device. In other words, the optimization system is iteratively approximated or solved by means of the electronic computing device, in particular by means of the optimization method. This means that the respective optimized value of the respective process parameter is iteratively determined, in particular calculated, by means of the electronic computing device using the optimization method by minimizing the deviation between the respective desirability and the respective predicted desirability.As a result, the respective optimized value of the respective process parameter can be determined with particularly little effort and / or with particular precision, whereby the particularly advantageous properties of the storage cell can be realized in a particularly low-effort manner.

[0017] In a further embodiment, the method comprises the following step: generating, in particular calculating, the prediction model from at least one data set comprising a plurality of values ​​of the respective target variable and a plurality of values ​​of the respective process parameter, by means of regression analysis. This means that the prediction model is generated by means of regression analysis from the data set using the electronic computing device using the regression analysis. In other words, in particular as the second component, the prediction model is created using the electronic computing device, which is based on the regression analysis. This makes it possible to represent a particularly precise relationship between the process parameters and the target variables in a particularly low-effort manner, whereby the optimal values ​​of the process parameters can be determined particularly precisely.

[0018] In a further embodiment, the regression analysis includes a ridge regression. This means that ridge regression is performed as the regression analysis. In other words, the method provides for generating the prediction model from the data set using ridge regression, i.e., applying ridge regression. The predicted values ​​of the respective target variables can thus be determined, in particular calculated, using the prediction model generated by ridge regression. This allows the particularly precise relationship between the process parameters and the target variables to be determined with particularly little effort.

[0019] In a further embodiment, it is provided that a one-sided Harrington function, a two-sided Harrington function, or a Derringer-Suich function is used as the, in particular, respective desirability function for determining the respective desirability and for determining the respective predicted desirability. This means that the one-sided Harrington function, the two-sided Harrington function, or the Derringer-Suich function is used as the first desirability function for determining the first desirability and for determining the first predicted desirability, and that the one-sided Harrington function, the two-sided Harrington function, or the Derringer-Suich function is used as the second desirability function for determining the second desirability and for determining the second predicted desirability.In other words, the respective desirability function is the one-sided Harrington function, the two-sided Harrington function, or the Derringer-Suich function. This allows the desirability to be determined particularly precisely, with minimal effort, and especially tailored to the needs of the respective application.

[0020] In a further embodiment, the method comprises the following step: determining, in particular calculating, optimized values ​​of the respective target variables by back-transforming the values ​​of those of the respective, in particular determined, predicted desirability that exhibit the smallest deviation from the respective, in particular determined, desirability. In other words, the values ​​of the respective predicted desirability assigned to the determined optimized values ​​of the process parameters are back-transformed, for example by means of an inverse function of the respective desirability function. As a result, the target variables resulting for a memory cell produced using the optimized process parameters are available in dimensioned form, i.e. with units, and therefore not in dimensionless form, whereby the resulting target variables can be assessed with particularly little effort and / or with particular precision.

[0021] In a further embodiment, it is provided that a mean square deviation is used as the optimization method. In other words, the optimization method is the mean square deviation. This means that the determination of the respective optimal value of the respective process parameter is carried out using the mean square deviation by minimizing the deviation between the respective desirability and the respective predicted desirability. This allows the optimal values ​​to be determined particularly precisely and with particularly little effort. The mean square deviation can in particular be understood to mean an optimization method which can be referred to in particular as the expected square deviation, the mean square error or in English as the mean square error (MSE).

[0022] In a further development, it is intended that the respective target value is one of the following target values: - a charge quantity of the storage cell, - a total weight of the storage cell or - an electrolyte weight of the storage cell.

[0023] In other words, the first target variable is the charge quantity, the total weight or the electrolyte weight of the storage cell and the second target variable is the charge quantity, the total weight or the electrolyte weight of the storage cell, wherein the first and the second target variables are preferably different from one another, i.e. not identical. For example, at least three target variables are provided in the method, wherein the first target variable is the charge quantity of the storage cell, the second target variable is the total weight of the storage cell and the third target variable is the electrolyte weight of the storage cell. The electrolyte weight can be understood in particular as a weight or a mass of an electrolyte of the storage cell.By specifying these target values, the optimized process parameters can be determined based on particularly important quality criteria of the memory cell, whereby the memory cell can be manufactured in the desired quality in a particularly low-cost manner.

[0024] In a further embodiment, it is provided that the respective process parameter characterizes one of the following manufacturing steps, i.e. one of the following manufacturing steps, for producing the memory cell: - a mixing process for producing active material of the memory cell, - a coating process, - a drying process, - a cutting process, - a winding process, - an assembly process, - an electrolyte filling, - sealing the memory cell or - a formation process of the memory cell.

[0025] The fact that the respective process parameter characterizes the respective manufacturing step can be understood in particular to mean that the respective manufacturing step depends on the respective process parameter, i.e., that the execution of the respective manufacturing step, and in particular a result of the respective manufacturing step, depends, in particular significantly, on the respective process parameter. Thus, the method can be used to optimize process parameters or their values ​​that are particularly important for achieving the desired quality properties of the memory cell.

[0026] A second aspect of the invention relates to a method for producing at least one storage cell for an electrical energy storage device, in particular for a motor vehicle, in which the storage cell is produced using optimized values ​​of process parameters determined by means of a method according to the first aspect of the invention. In other words, in the method, the storage cell is produced using optimized values ​​of the process parameters, wherein these optimized values ​​are or have been determined by means of the method according to the first aspect of the invention. This means that the values ​​of the process parameters determined by means of the method according to the first aspect of the invention are, in particular, specifically adjusted and / or predetermined in order to produce the storage cell.Thus, the manufactured memory cell is a memory cell produced or manufactured using the determined process parameters, in particular using the respective determined values ​​of the process parameters. Advantages and advantageous embodiments of the first aspect of the invention are to be regarded as advantages and advantageous embodiments of the second aspect of the invention, and vice versa.

[0027] Further features of the invention emerge from the claims, the figures, and the description of the figures. The features and combinations of features mentioned above in the description, as well as the features and combinations of features mentioned below in the description of the figures and / or shown alone in the figures, can be used not only in the respective combinations specified, but also in other combinations or on their own.

[0028] The invention will now be explained in more detail using a preferred embodiment and with reference to the drawings. They show: Fig. 1 is a schematic diagram illustrating a method according to the invention for producing a memory cell; and Fig. 2 shows a schematic process diagram of a method according to the invention for determining process parameters for producing a memory cell; and Fig. 3 a schematic flow diagram of a method according to the invention for determining process parameters for producing a memory cell; and Fig. 4 is a schematic diagram of a desirability function of a method according to the invention for determining process parameters for producing a memory cell; and Fig. 5 is a schematic diagram of a further desirability function of a method according to the invention for determining process parameters for producing a memory cell; and Fig. 6 is a schematic diagram illustrating a data cleanup of a method according to the invention for producing process parameters for producing a memory cell; and Fig. 7 a schematic representation of an optical output of process parameters of a method according to the invention for determining process parameters for producing a memory cell; and Fig. 8 a schematic representation of an optical output of target variables of a method according to the invention for determining process parameters for producing a memory cell.

[0029] In the figures, identical or functionally identical elements are provided with the same reference symbols.

[0030] Fig. 1 shows a schematic representation of a method 1 for producing at least one storage cell 2 for an electrical energy storage device, in particular for a motor vehicle. The method 1 comprises, for example, the following production steps: a mixing process 3 for producing active material for the storage cell 2, a coating process 4, a drying process 5, a cutting process 6, a winding process 7, an assembly process 8, an electrolyte filling process 9, a sealing process 10 of the storage cell 2, and / or a forming process 11 of the storage cell 2.

[0031] In the mixing process 3, which can also simply be referred to as mixing, the active material of the storage cell 2 is produced or formed, for example, from at least two components, in particular by mixing.

[0032] In the coating process 4, which can also be referred to simply as coating, at least one paste is applied to at least one foil, which can in particular be referred to as a carrier foil. The foil is formed, for example, from metal, in particular from aluminum or copper.

[0033] In the drying process 5, which can also simply be referred to as drying, the paste applied to the film is dried, in particular by applying heat.

[0034] In the cutting process 6, which can also be referred to simply as cutting, the film is cut or trimmed, for example, especially after drying. The cutting process 6 includes, for example, both calendering and cutting. Calendering can be understood, in particular, as the compaction or pressing of an electrode.

[0035] During the winding process 7, which can also be referred to simply as winding, at least one foil coil is formed from the foil, for example. Furthermore, a winding process for a cathode and / or an anode, in particular with a separator, can be performed.

[0036] During assembly process 8, which can also be referred to simply as assembly, at least one housing element of the storage cell 2 is formed or assembled, for example. The active material or the foil wrap is arranged, for example, within the housing element.

[0037] During electrolyte filling 9, the housing element is filled, for example, with at least one electrolyte. This means that the electrolyte is arranged within the housing element, which can be referred to in particular as electrolyte filling.

[0038] During the formation process 11 of the memory cell 2, the memory cell 2, in particular the one manufactured, is charged and / or discharged, for example, in particular for the first time. The formation process 11 can in particular be referred to as formation.

[0039] To carry out the respective manufacturing step, for example, at least one respective process parameter, in particular a respective value of the respective process parameter, is specified or set, in particular assigned to the respective manufacturing step. The respective process parameter or its value influences the respective manufacturing step, i.e., a result of the respective manufacturing step. Thus, the respective process parameter influences the memory cell 2, in particular the memory cell that has been manufactured or is to be manufactured. In other words, the respective process parameter or its value characterizes the respective manufacturing step.

[0040] A process parameter characterizing the mixing process 3 can be a recipe 12 for producing or forming the active material, wherein this recipe 12 can include a quantitative ratio of the components for forming the active material. Thus, the process parameter of the mixing process 3 can be the recipe 12 or the quantitative ratio.

[0041] A process parameter characterizing the coating process 4 is, for example, a pressure and / or a speed when applying the paste to the film. Thus, the process parameter(s) of the coating process 4 can be the pressure and / or the speed of application. Furthermore, a basis weight 13 of the, in particular coated, film can be a process parameter in the coating process 4.

[0042] A process parameter characterizing the drying process 5 can be a temperature to which the film is exposed for drying. Thus, the process parameter of the drying process 5 can be the temperature, in particular during drying.

[0043] A process parameter characterizing cutting process 6 is, for example, the cutting pressure during cutting. This means that the process parameter of cutting process 6 can be the cutting pressure.

[0044] A process parameter characterizing the winding process 7 can be a winding thickness 14 and / or a winding length and / or a winding speed. This means that the process parameter(s) of the winding process 7 can be the winding thickness 14 and / or the winding length and / or the winding speed. The winding length is, for example, a length, in particular a wound length, of the anode, cathode, and in particular the separator.

[0045] Process parameters characterizing the assembly process include, for example, welding power during welding, in particular for assembling add-on parts and / or the housing element, and / or pressure for inserting components. Thus, the process parameter(s) of the assembly process 8 can be the welding power and / or the pressure.

[0046] Process parameters characterizing the electrolyte filling 9 are, for example, a quantity of electrolyte, in particular the electrolyte filled, and / or a pressure during filling the electrolyte. This means that the process parameter(s) of the electrolyte filling 9 can be the quantity of electrolyte and / or the pressure during filling. The quantity can be referred to in particular as electrolyte quantity 15.

[0047] A process parameter characterizing the sealing 10 of the storage cell 2 is, for example, a welding power, in particular for sealing 10 by welding. Thus, the process parameter for sealing 10 of the storage cell 2 can be the welding power.

[0048] Process parameters characterizing the forming process 11 can be a current and / or a voltage, in particular when charging the storage cell 2, and / or a time period when charging and / or discharging the storage cell 2. Thus, the process parameter(s) of the forming process 11 can be the current, the voltage, and / or the time period.

[0049] Each of the process parameters shown can influence quality properties of the storage cell 2, in particular the one manufactured or to be manufactured. The process parameters can thus influence target variables 16, 17, 18 of the storage cell 2 to be manufactured or to be manufactured. The target variables 16, 17, 18 can represent the quality properties. The target variables 16, 17, 18 can be, for example, a charge quantity 20 of the storage cell 2, a total weight 21 of the storage cell 2 and / or an electrolyte weight 22 of the storage cell 2. Since the quality properties or the target variables 16, 17, 18 can conflict with one another, multi-criteria optimization may be necessary or useful.

[0050] Fig. 2 shows a schematic process diagram of a method 23 for determining the process parameters, in particular those mentioned above, for producing the storage cell 2 for an electrical energy storage device, in particular of a motor vehicle, by means of an electronic computing device 19. The method 23 is thus carried out by means of the electronic computing device 19. In the method, at least two of the process parameters mentioned above can be determined, which means that more than two or all of the process parameters mentioned above can also be determined. In the exemplary embodiment, two of the process parameters 25, 26 are considered as examples. The fact that the process parameters 24, 25 are determined can be understood in particular to mean that the process parameters 24, 25, in particular values ​​of the process parameters 24, 25, are optimized by means of multi-criteria optimization.

[0051] The method 23 comprises several steps S1, S2, S3, S4, which are carried out by means of the electronic computing device 19. The steps S1 to S4 are shown in Fig. 3, in which a flowchart of the method 23 is shown.

[0052] By means of the electronic computing device 19, in particular in a first of the steps S1, desirabilities 26, 27 for the target variables 16, 17 are determined, in particular calculated, using desirability functions 28, 29. In the exemplary embodiment, the method 23 is explained by way of example using two of the target variables 16, 17, although the method 23 can of course be carried out with more than two of the target variables. The first step S1 can in particular be referred to as determining 30 the desirabilities 26, 27. The desirability functions 28, 29 each assign a respective value 16a, 17a of the respective target variable 16, 17 of the memory cell 2, in particular a dimensionless number from a predetermined interval.In other words, by means of the electronic computing device 19, a number from the predetermined interval is assigned to the respective value 16a, 17a of the respective target variable 16, 17 by means of the respective desirability function 28, 29, i.e. by applying the desirability function 28, 29, in order to determine the respective desirability, for example by inserting the respective values ​​16a, 17a of the respective target variable 16, 17 into the respective desirability function 28, 29. Thus, in the exemplary embodiment, a first desirability 26, which in particular characterizes the first target variable 16 or is assigned to the first target variable 16, is determined, in particular calculated, by means of a first desirability function 28, which assigns a number from the predetermined interval, in particular as desirability, to a respective value 16a of the first target variable 16.Furthermore, a second desirability 27, which in particular characterizes the second target variable 17 or is assigned to the second target variable 16, is determined, in particular calculated, by means of a second desirability function 29, which assigns a number from the predetermined interval, in particular as desirability, to a respective value 17a of the second target variable 17.

[0053] The interval, which can be referred to in particular as a value interval or number interval, preferably ranges from zero to one. This means that the number zero can be a lower limit of the interval and the number one can be an upper limit of the interval. Values ​​of the respective desirability 26, 27 thus run, in particular exclusively, between zero and one. The desirabilities 26, 27 or the desirability functions 28, 29 can be used to de-dimension the target variables 16, 17, thereby providing particularly good, in particular mathematical, comparability, which has a particularly advantageous effect on further steps of the method 23, in particular on step S4.

[0054] Fig. Figure 4 shows a schematic diagram illustrating the first desirability function 28 or the first desirability 26. The values ​​16a of the first target variable 16 are plotted on the ordinate. The first desirability 26 is plotted on the abscissa, which extends between zero and one according to the specified value interval. In the Fig. In the embodiment shown in Figure 4, the first desirability function 28 is a one-sided Harrington function, which can be denoted by d1: d1(Y')=e−e−Y'Y'=b0+b1Y(according to Harrington)

[0055] Y is the respective value 16a, 16b of the respective target variable 16, 17, where b0 and b1 are model parameters. By adjusting the model parameters b0, b1, a desired Harrington function curve can be generated.

[0056] Fig. Figure 5 shows a schematic diagram illustrating the second desirability function 29 or the second desirability 27. The values ​​17a of the second target variable 17 are plotted on the abscissa. The second desirability 27 is plotted on the ordinate according to the specified value interval between zero and one. In the Fig. In the embodiment shown in Figure 5, the second desirability function 29 is a Derringer-Suich function, which can be denoted by d3: d3(Y)={0,(Y−LSLT−LSL)m1, forfor Y <LSLLSL≤Y≤T(USL−YUSL−T)m2,0, fu¨rfu¨r T<Y<USLUSL<Y(nach Derringer−Suich)

[0057] Here, m1 and m2 are model parameters. T is a target value for the respective target variable 16, 17, i.e., a desired value of 16a, 17a for the respective target variable 16, 17. LSL is a lower specification limit for the values ​​16a, 17a for the respective target variable 16, 17. USL is an upper specification limit for the values ​​16a, 17a for the respective target variable 16, 17. Thus, a range for the values ​​16a, 17a for the respective target variable 16, 17 can be specified, i.e., a permissible deviation from the target variable 16, 17. By adjusting the model parameters m1, m2, the desirability curve can be adjusted, particularly in a targeted manner.

[0058] Alternatively, it is possible that the respective desirability function 28, 29, i.e. the first and / or the second desirability function 28, 29, is a two-sided Harrington function, which is in particular d2 can be described as: d2(Y')=e−|Y'|n0 <n<∞, mit Y'=2Y−(USL+LSL)USL−LSL(nach Harrington

[0059] Here, n is a model parameter by means of which a desirability curve can be adjusted, in particular in a targeted manner. Thus, the one-sided Harrington function, the two-sided Harrington function, or the Derringer-Suich function can be used to determine 30 the respective desirability 26, 27 as the respective desirability function 28, 29.

[0060] The specification limits LSL, USL, and the target value T can be determined, for example, from experience, particularly internal to the company, such as discussions with technologists, developers, and / or quality engineers. Thus, the respective desirability 26, 27 or the respective desirability function 28, 29 can be derived from discussions with technologists, developers, and / or quality engineers.

[0061] In the method 23, by means of the electronic computing device 19, in particular in a second of the steps S2 taking place before or after the first step S1, predicted values ​​16b, 17b of the respective target variables 16, 17 are determined, in particular calculated, using a prediction model 33 generated by regression analysis 32, which can in particular be referred to as a production model. The prediction model 33 describes or simulates the target variables 16, 17 as a function of the process parameters 24, 25. This means that values ​​24a, 24b of the process parameters 24, 25 can be specified in the prediction model 33, as a function of which the predicted values ​​16b, 17b of the target variables 16, 17 can be determined, in particular calculated.Predictions of the quality properties in the form of the target variables 16, 17 can therefore be calculated from the prediction model 33 by entering input parameters in the form of the process parameters 24, 25 or their values ​​24a, 25a. The second step S2 can in particular be referred to as determining 34 predicted values ​​16b, 17b of the respective target variable 16, 17. For example, the first predicted value 16b of the first target variable 16, in particular as a function of at least one of the values ​​24a, 24b of the process parameters 24, 25, and the second predicted value 17b of the second target variable 17, in particular as a function of at least one of the values ​​24a, 24b of the process parameters 24, 25, are determined by means of the prediction model 33 by means of the electronic computing device 19.

[0062] Furthermore, in the method, by means of the electronic computing device 19, in particular in a third step S3 taking place after the second step S2, predicted desirabilities 26a, 27a for the target variables 16, 17 are determined, in particular calculated, using the desirability functions 28, 29. The desirability functions 28, 29 assign, in particular equivalent to the application in determining the desirabilities 26, 27, each of the predicted values ​​16b, 17b of the respective target variable 16, 17, i.e., the predicted values ​​16b, 17b of the respective target variable 16, 17, a number, in particular a dimensionless number, from the predetermined interval.This means that in order to determine 35 the respective predicted desirability 26a, 27a intended for the respective target variable 16, 17 or characterizing the respective target variable 16, 17, the same desirability function 28, 29 is used which is used or has been used for determining 30 the respective desirability 26, 27 intended for the respective target variable 16, 17 or characterizing the respective target variable 16, 17.In other words, by means of the electronic computing device 19, by means of the respective desirability function 28, 29, i.e. by applying the respective desirability function 28, 29, a number from the predetermined interval is assigned to the respective predicted value 16b, 17b of the respective target variable 16, 17 in order to determine the respective predicted desirability 26a, 27a, for example by inserting the respective predicted values ​​16b, 17b of the respective target variable 16, 17 into the respective desirability function 28, 29.To determine a first predicted desirability 26a, which in particular characterizes the first target variable 16 or is assigned to the first target variable 16, the first desirability function 28 is used, for example, and to determine 35 a second predicted desirability 27a, which in particular characterizes the second target variable 17 or is assigned to the second target variable 17, the second desirability function 29 is used, for example. Thus, predictions of the quality properties in the form of the predicted values ​​16b, 17b of the respective target variable 16, 17 can be transformed into unitless and thus comparable characteristics by means of the respective desirability function 28, 29.

[0063] Furthermore, in the method 23, by means of the electronic computing device 19, in particular in a fourth step S4, at least one respective optimized value 24b, 25b of the respective process parameter 24, 25 is determined by means of at least one optimization method 36 by minimizing at least one deviation between the respective desirability 26, 27 and the respective predicted desirability 26a, 27a, in particular iteratively.This means that at least one optimized value 24b of the first process parameter 24 is determined, in particular calculated, by means of the electronic computing device 19 using the optimization method 36 by minimizing a deviation between the first desirability 26 and the first predicted desirability 26a, and at least one optimized value 25b of the second process parameter 25 is determined, in particular calculated, by means of the electronic computing device 19 using the optimization method 36 by minimizing a deviation, in particular referred to as the second deviation, between the second desirability 27 and the second predicted desirability 27a. The fourth step S4 can in particular be referred to as determining 37 the respective optimal value 24b, 25b of the respective process parameter 24, 25.

[0064] In the exemplary embodiment, it is provided that, in order to minimize the respective deviation, the determination 34 of the predicted values ​​16b, 17b of the respective target variable 16, 17 and the determination 35 of the predicted desirability 26a, 27a are carried out iteratively by varying the values ​​24a, 25a of the process parameters 24, 25. This means that the second and third steps S2, S3, in particular for carrying out the fourth step S4, are carried out iteratively, i.e., are repeated, until at least one target criterion, in particular all target criteria, of the optimization method 36 is met, wherein, for this target criterion, for example, the respective deviation is minimal.Thus, by means of the optimization method 36, it can be determined, in particular calculated, for which values ​​of the process parameters 24, 25 in the form of the optimized values ​​24b, 25b the courses of the target variables 16, 17 represented by the predicted desirabilities 26a, 27a come closest to the desired courses of the target variables 16, 17 in the form of the desirabilities 26, 27. Fulfillment of the target criterion or criteria can be understood in particular as complete or partial, i.e. sufficient, fulfillment. It is therefore not necessarily the case that one of the target criteria is optimized to 100 percent. It may therefore be possible that all target criteria are or will be optimized "only" to a value less than 100 percent, for example 95 percent.However, this can be better, meaning there can be a lower overall error, than if one of the target criteria is or will be optimized to 100 percent and the others only to 80 percent.

[0065] For example, an error measure, in particular a mean square deviation, is used as the optimization method 36. Thus, the values ​​or optimized values ​​24b, 25b of the process parameters 24, 25 or desirability values ​​can be optimized using the error measure, for example, by applying the mean square deviation, while varying the input parameters in the form of the values ​​25a, 25b of the process parameters 24, 25. The total error can therefore refer to the mean square deviation, whereby the total error can be referred to in particular as the total error MSE (mean squared error).

[0066] The generation 38 of the prediction model 33 may or may not be part of the method 23. The generation 38 can therefore be provided optionally. If the generation 38 of the prediction model 33 is part of the method 23, the method comprises, for example, the following step S2*: Generation 38 of the prediction model 33 from at least one data set 39, which comprises a plurality of values ​​16a, 17a of the respective target variable 16, 17 and a plurality of values ​​24a, 25a of the respective process parameter 24, 25. The data set 39 thus has a plurality of values ​​16a, 17a, 24a, 25a of the target variables 16, 17 and the process parameters 24, 25. The regression analysis 32 preferably comprises a ridge regression 40, which can be mathematically described as follows: β^ridge=arg minβ{12∑i=1n(Yi−β0−∑j=1pxijβj)2+λ∑j=1pβj2} y i is the respective value 16a, 17a of the target variable 16, 17 from the data set 32. x ijis the respective value 24a, 25a of the respective process parameter 24, 25 from data set 39. β0 is a model parameter. β j is also a model parameter. n is a number of data points from data set 39. p is a number of process parameters 24, 25. Thus, the quality characteristics can be modeled using production data from data set 39 and ridge regression 40.

[0067] Preferably, it is provided that the production data or the data set 39 is or has been cleaned in advance, that is, before the generation 38 of the prediction model 32. This is in Fig. 6 is illustrated in a schematic diagram. Values ​​16a, 16b of one of the target variables 16, 17 are plotted on the abscissa. Cleaning includes, for example, removing outliers, imputing missing data, and / or removing observations with missing values. For example, the data, specifically referred to as production data, are merged or combined into one dataset after cleaning.

[0068] For example, the electronic computing device 19 has at least one optical output device, or the electronic computing device 19 is coupled or can be coupled to at least one output device formed separately from the electronic computing device. For example, it is provided that the optimized values ​​24b, 25b of the process parameters 24, 25 are optically output by means of the optical output device, i.e., are graphically displayed for a person, for example in the form of a respective dashboard. This is Fig. 7, where Fig. 7 several dashboards 41 to 46 are shown. The Fig. The process parameters shown as examples in Figure 7 characterize the winding process 7. In a first dashboard 41, a temperature of the winding process 7 is shown as a process parameter. In a second dashboard 42, a dew point in the winding process 7 is shown. The temperature and the dew point are each shown in degrees Celsius. In a third dashboard 43, an actual length value of the anode is shown as a process parameter of the winding process 7. In a fourth dashboard 44, an actual length value of the cathode is shown as a process parameter of the winding process 7. In a fifth dashboard 45, an actual length value of the separator is shown as a process parameter of the winding process 7. The respective lengths are shown in millimeters. In a sixth dashboard 46, an angular velocity is shown as a process parameter of the winding process 7. In the Fig. In the exemplary embodiment illustrated in Figure 7, value ranges of the respective process parameter are visualized. This can be achieved, for example, by determining, particularly in the fourth step S4, not a single optimal value 24b, 25b for the respective process parameter 24, 25, but rather several optimal values ​​24a, 24b, for example, a value range of optimal values ​​24a, 24b.

[0069] For example, in method 23, 47 optimized values ​​of the respective target variables 16, 17 are determined by back-transforming 48 the values ​​of those of the respective predicted desirabilities 26a, 27a that exhibit the smallest deviation from the respective desirability 26, 27. Thus, optimized desirability values ​​can be back-transformed into a value space of the original quality properties, i.e., the original target variables 16, 17, using the desirability functions 28, 29 or inverse functions of the desirability functions 28, 29, particularly referred to as inverse mappings. This allows the optimized quality properties to be graphically represented in the form of optimized target variables 16, 17. Thus, it can be provided that the determined optimized values ​​16c, 17c of the target variables 16, 17 are output by means of the optical output device, i.e., for example, are graphically represented to the person.This can be done using dashboards, which is available in . Fig. 8, in which two dashboards 49, 50 are shown, which can be referred to in particular as the seventh and eighth dashboard 49, 50. In this case, Fig. 8, the seventh dashboard 49 illustrates the electrolyte weight 21, which can be referred to as the actual electrolyte weight, as a target value, and the eighth dashboard 50 illustrates the charge quantity 20 as a target value. The actual electrolyte weight 21 is shown in grams, and the charge quantity is shown in ampere-hours. As shown in Fig. 8, the respective target variables 16, 17 can be illustrated in a value range, which can in particular be referred to as an ideal range. For this purpose, when determining 47 the optimized values ​​of the target variables 16, 17, a respective optimized value range of the respective target variables 16, 17 can be determined.

[0070] Thus, the described multi-criteria optimization system can be visualized, for example, in the respective dashboard 41 to 46, 49, 50, wherein the optimization system or the respective dashboard 41 to 46, 49, 50 can be used to optimize the target variables 16, 17, for example, the charge quantity, referred to in particular as capacity, and the total weight, referred to in particular as weight, of the storage cell 2. The respective dashboard 41 to 46, 49, 50 can display optimal back-transformed values ​​of the quality characteristics (output) as well as process parameters (input) with which these optimal values ​​were achieved.

[0071] Overall, it can be seen that a variation of the process parameters for adjusting the quality characteristics of storage cell 2 can be realized using method 23. In this case, an adjustment of the product quality of storage cell 2 can be carried out using the respective dashboards 41 to 46, 49, 50. In particular, optimal process parameters and thus the achievement of CO2 and cost savings can be realized in a particularly low-effort manner.

[0072] Based on the optimized values ​​24b, 25b of the process parameters 24, 25 determined by means of the method 23, the memory cell 2 can actually be manufactured, in particular subsequently, whereby the memory cell 2 has in particular the optimized values ​​16c, 17c of the target variables 16, 17. In the method 1 for producing the memory cell 2, it is accordingly provided that the memory cell 2 is manufactured using a plurality of the process parameters 24, 25 determined by means of the method 23, in particular using values ​​24b, 25b of the process parameters 24, 25 optimized by means of the method 23. List of reference symbols 1 Method for producing a memory cell 2 memory cells 3 Mixing process 4 Coating process 5 Drying process 6 Cutting process 7 Winding process 8 Assembly process 9 Electrolyte filling 10 Sealing 11 Formation process 12 Recipe 13 Basis weight 14 winding thickness 15 Electrolyte amount 16 first target value 16a Value of the first target variable 16b predicted value of the first target variable 16c optimized value of the first target variable 17 second target value 17a Value of the second target variable 17b predicted value of the second target variable 17c optimized value of the second target variable 18 third target 19 electronic computing device 20 load quantity 21 Total weight 22 Electrolyte weight 23 Methods for determining process parameters 24 first process parameters 24a Value of the first process parameter 24b optimized value of the first process parameter 25 second process parameter 25a Value of the second process parameter 25b optimized value of the second process parameter 26 first desirability 26a first predicted desirability 27 second desirability 27a second predicted desirability 28 first desirability function 29 second desirability function 30 Determining the predicted values 32 Prediction model 33 Regression analysis 34 Determining predicted values 35 Determining predicted desirabilities 36 Optimization method 37 Determining optimized values 38 Creating the prediction model 39 data sets 40 Ridge Regression 41 first dashboard 42 second dashboard 43 third dashboard 44 fourth dashboard 45 fifth dashboard 46 sixth dashboard 47 Determining optimized values 48 Inverse transformation 49 seventh dashboard 50 eighth dashboard LSL lower specification limit T target value USL upper specification size

Claims

[1] Method (23) for determining process parameters (24, 25) for producing a storage cell (2) for an electrical energy storage device by means of an electronic computing device (19), comprising the steps: • Determining (30) desirabilities (26, 27) for target variables (16, 17) of the memory cell (2) by means of desirability functions (28, 29), which each assign a number from a predetermined interval to a respective value (16a, 17a) of the respective target variables (16, 17); • Determining (34) predicted values ​​(16b, 17b) of the respective target variable (16, 17) by means of a prediction model (33) generated by regression analysis (32), which describes the target variables (16, 17) as a function of the process parameters (24, 25); • Determining (35) predicted desirabilities (26a, 27a) for the target variables (16, 17) by means of the desirability functions (28, 29), which each assigns a number from the predetermined interval to a respective one of the predicted values ​​(16b, 17b) of the respective target variable (16, 17) of the memory cell (2), • Determining (37) at least one respective optimized value (24b, 25b) of the respective process parameter (25, 25) by means of an optimization method (36) by minimizing a deviation between the respective desirability (26, 27) and the respective predicted desirability (26a, 27a). [2] Method (23) according to claim 1, characterized by that in order to minimize the respective deviation, the determination (34) of the predicted values ​​(16b, 17b) of the respective target variable (16, 17) and the determination (35) of the predicted desirabilities (26a, 27a) are carried out iteratively by varying the values ​​(16a, 17a) of the process variables (24, 25). [3] Method (23) according to claim 1 or 2, characterized by that the procedure includes the following step: Generating (38) the prediction model (33) from a data set (39) which comprises a plurality of values ​​(16a, 16b) of the respective target variable (16, 17) and a plurality of values ​​of the respective process parameter (24, 25) by means of regression analysis (32). [4] Method (23) according to one of the preceding claims, characterized by that the regression analysis (32) includes a ridge regression. [5] Method (23) according to one of the preceding claims, characterized by that for determining (30) the respective desirability (26, 27) and for determining (34) the respective predicted desirability (26a, 27a) a one-sided Harrington function, a two-sided Harrington function or a Derringer-Suich function is used as the desirability function (28, 29). [6] Method (23) according to one of the preceding claims, characterized by that the method (23) comprises the following step: Determining optimized values ​​(16c, 17c) of the respective target variable (16, 17) by back-transforming (48) the values ​​(16b, 17b) of those of the respective predicted desirability (26a, 27a) which have the smallest deviation from the respective desirability (26, 27). [7] Method (23) according to one of the preceding claims, characterized by that a mean square deviation is used as the optimization method (36). [8] Method (23) according to one of the preceding claims, characterized by that the respective target variable (16, 17) is one of the following target variables (16, 17): • a load quantity (20), • a total weight (21) or • an electrolyte weight (22) of the storage cell (2). [9] Method (23) according to one of the preceding claims, characterized bythat the respective process parameter (24, 25) characterizes one of the following manufacturing steps for producing the memory cell (2): • a mixing process (3) for producing active material of the memory cell (2), • a coating process (4), • a drying process (5), • a cutting process (6), • a winding process (7), • an assembly process (8), • an electrolyte filling (9), • sealing (10) of the storage cell (2) or • a formation process (11) of the memory cell (2). [10] Method (1) for producing a storage cell (2) for an electrical energy storage device, in which the storage cell (2) is produced by means of optimized values ​​(24b, 25b) of process parameters (24, 25) determined by means of a method (23) according to one of the preceding claims.

Images