Method for setting a process parameter in a manufacturing process to optimize a component quality of a component

Abstract

The invention relates to a method for adjusting a process parameter in a manufacturing process to optimize a component quality of a component that is processed by the manufacturing process; the method comprising: - Providing (S1) quality data of the component, which are determined by a quality inspection of the component quality; - Extracting (S2) at least one dimensionless indicator from the quality data; - Assessment (S3) of the component quality based on at least one dimensionless key figure; and - Setting (S4) the process parameter based on the evaluation.

Description

[0001] The invention relates to a method and a device for adjusting a process parameter in a manufacturing process to optimize a component quality of a component that is processed by the manufacturing process. State of the art

[0002] To ensure the service life of components under stress, it is crucial to specifically adapt material properties such as hardness and microstructure. Inductive heat treatment offers various adjustable parameters for this purpose, such as power, frequency, temperature, time, and feed rate. The current approach to determining the relationship between these process parameters, material structure, and component service life is primarily based on experimental methods. This experimental approach requires a large number of tests, especially for new materials and modified material mixtures, which is both time- and resource-intensive.

[0003] DE 102022203835 A1 further discloses a method for controlling at least one process parameter of a production process, the method comprising the following steps: During execution of the production process, monitoring at least one process characteristic, determining a current tool state and a resulting component quality by one or more prediction models based on the monitored at least one process characteristic, determining optimal specifications for the at least one process parameter based on the at least one monitored process characteristic, the current tool state and the resulting component quality, and controlling the at least one process parameter based on the optimal specifications for the at least one process parameter.

[0004] Thus, there is still potential for development.

[0005] The invention is therefore based on the object of specifying an optimized method and / or an optimized device for setting a process parameter in a manufacturing process in order to optimize a component quality of a component.

[0006] The object is achieved by a method according to the features of patent claim 1. The object is achieved by a device according to the features of patent claim 11. Disclosure of the invention

[0007] According to a first aspect, a method for adjusting a process parameter in a manufacturing process to optimize the component quality of a component is provided, wherein the component is processed by the manufacturing process. The method comprises the steps: - Providing quality data of the component, which is determined by a quality inspection of the component quality; - Extracting at least one dimensionless metric from the quality data; - Assessment of component quality based on at least one dimensionless key figure; and - Setting the process parameter based on the evaluation.

[0008] It is understood that the steps according to the invention, as well as other optional steps, do not necessarily have to be performed in the order shown, but can also be performed in a different order. Furthermore, additional intermediate steps can be provided. The individual steps can also comprise one or more substeps without thereby departing from the scope of the method according to the invention.

[0009] In a second aspect, a device for adjusting a process parameter of a manufacturing process to optimize the component quality of a component is provided, wherein the component is processed by the manufacturing process. The device comprises an evaluation and computing device configured to perform the following steps: - Providing quality data of the component, which is determined by a quality inspection of the component quality; - Extracting at least one dimensionless metric from the quality data; - Assessment of component quality based on at least one dimensionless key figure; and - Setting the process parameter based on the evaluation.

[0010] The statements made for the method apply accordingly to the device and vice versa.

[0011] The method enables the more efficient identification of anomalies triggered by specific mechanisms. This leads to a significant reduction in the number of required experiments. The increased efficiency is based on the ability to reduce the number of independent variables through the application of unit normalization using dimensionless parameters. This approach thus offers a more efficient method for optimizing process parameters in inductive heat treatment, simplifying and accelerating the development and application of new materials.

[0012] The term "dimensionless" in the context of the key figure means "unitless." The key figure therefore has no physical unit. The key figure is preferably composed of several quality parameters, which can be linked together using mathematical operators to make the key figure unitless or dimensionless.

[0013] To search for optimal process parameter ranges, dimensionless key figures, such as pi factors, are preferably determined based on a suitable theorem, such as the Buckingham pi theorem. Instead of directly searching for anomalies in the parameter space of the multitude of quality parameters contained in the quality data, the dimensionless key figures are analyzed. In some cases, the multidimensional quality parameter space can be represented by a single dimensionless or unitless key figure, thus reducing the multidimensional parameter space to a single parameter, which positively influences the complexity of the evaluation and the detection of anomalies.

[0014] Since the dimensionless indicators are unit-normalized, i.e., unitless, indicators based on physical quality parameters, the dimensionless indicators represent physical interactions and thus allow a better assessment regarding the evaluation of a component anomaly. The units of the quality parameters are preferably provided in the quality data set and thus available for further analysis to generate the dimensionless indicator, which facilitates the application of the methodology. Furthermore, the relationship with material parameters can be easily mapped by relating them to the original units of the quality parameters that were used to construct the dimensionless indicator.

[0015] The component can be, for example, a gear, a toothing, a shaft, a camshaft, a bolt, or something similar that is to be machined and / or hardened. The component can be machined, for example, using inductive heat treatment.

[0016] The method and device offer a solution for converting a multidimensional parameter space of quality parameters into a parameter space of dimensionless key figures that is smaller than the initial, multidimensional parameter space, thus reducing the complexity of a process optimization task. In special cases, the multidimensional parameter space can be reduced to a monodimensional parameter space. This reduces the optimization effort required to find suitable setting changes, which can lead to time savings.

[0017] In a further aspect, the quality testing comprises a non-destructive and / or destructive testing of the component quality, in particular a component hardness and / or a quality of a component structure.

[0018] The features of the claim describing the quality testing of a component can be summarized as follows: Quality testing methods are either non-destructive or destructive. This means that some testing methods leave the component intact after testing, while others damage or destroy the component to evaluate its quality. The focus is on checking the quality of a component. The quality of the component can essentially refer to various quality aspects. In this case, for example, the hardness of the component and the quality of the component's microstructure are tested. Hardness tests evaluate how well a component offers resistance to penetration or deformation. This is particularly important for components that are subject to high loads or wear. The quality of the component's microstructure refers to the microscopic arrangement of atoms or molecules within the component. This quality can influence the mechanical properties, such as the strength, toughness, and durability of the material.

[0019] In a further aspect, the component is pre-machined prior to the quality inspection in the manufacturing process by at least one pre-processing step, in particular by bending and / or forming and / or annealing. Furthermore, the component is finished in the manufacturing process by at least one finishing step, in particular by hardening, particularly preferably by inductive heat treatment.

[0020] Before quality control, the component preferably undergoes at least one pre-processing step as part of the manufacturing process. This preparation serves, for example, to manufacture the component for its intended purpose.

[0021] Bending describes the mechanical deformation of a material to give it a specific shape. Forming describes a broader term for changing the shape of the component material, which can include various techniques such as deep drawing, pressing, or stretching. Annealing describes a heat treatment process intended to soften the material, reduce stresses, change the microstructure, or improve machinability. The finishing step(s) preferably occur after pre-processing and before quality inspection, with the component being further processed through at least one finishing step. These steps serve to complete the component and improve its properties for its final use. Hardening describes a process for increasing the hardness and strength of the material, often by heating followed by rapid cooling (quenching).Induction heat treatment describes a special type of hardening in which the material is heated by induction and then quenched to improve specific properties such as hardness, toughness, and wear resistance. This method is highlighted as particularly preferred, indicating its high efficiency or particular suitability for the component.

[0022] In a further aspect, the extraction of at least one dimensionless metric from the quality data is carried out by applying a Buckingham pi theorem.

[0023] Dimensionless quantities are mathematical quantities that have no unit and represent ratios or functional relationships between physical quantities. They are of interest because they allow for universal comparisons and conclusions that do not depend on specific units of measurement. The application of the Buckingham Pi Theorem describes a method used for extracting these quantities, based on the Buckingham Pi Theorem. This theorem is a fundamental principle of dimensional analysis that allows physical phenomena to be described using a reduced number of dimensionless quantities.The theorem states that any physically meaningful equation involving n physical variables can be transformed into an equivalent equation consisting of a set of p (p < n) dimensionless indices, where p is the difference between the number of variables and the number of fundamental dimensions (such as length, mass, time) in the problem.

[0024] An important statement of the Buckingham Pi theorem is that any dimension-bound equation can be expressed as an equation composed solely of dimensionless power products (and numerical constants). The significance of the theorem lies in its ability to make a statement about the functional relationship between dimensioned physical quantities, which may not be explicitly expressed in a formula. This applies to many complex situations. Since quantities can only appear in certain relationships, the Π factors, a useful reduction of the functional variables is achieved at the same time.

[0025] In a further aspect, the evaluation of the component quality based on the at least one dimensionless key figure comprises an evaluation of a component anomaly by comparing the at least one dimensionless key figure with a limit value or a limit interval, wherein the evaluation classifies the component anomaly as OK or not OK.

[0026] The evaluation can also be illustrated graphically, for example in a diagram in which the at least one dimensionless key figure is plotted over time.

[0027] In a further aspect, when several dimensionless key figures are extracted from the quality data, a dimensionless key figure that optimally matches the manufacturing process is determined by training a machine learning model.

[0028] If, for example, the key figures or pi factors derived from the pi theorem are not unique and, for example, several key figures can be used for anomaly detection, it is advantageous to analyze which of the key figures is most optimal based on the anomaly to be assessed and / or the process parameter to be adjusted. The starting point is therefore that several dimensionless key figures are extracted from the quality data of a manufacturing process. As previously mentioned, these key figures are quantities without units of measurement that are derived from the ratios or functional relationships between physical quantities. The aim is now to determine from the set of extracted dimensionless key figures the one that best fits a specific manufacturing process, a component anomaly to be detected, or a process parameter to be adjusted.In this context, "optimal fit" means that the selected metrics are best suited to describe the component's anomaly, ultimately allowing conclusions to be drawn about the quality and efficiency of the manufacturing process and to improve them. A machine learning model is trained to identify the optimally fitting dimensionless metrics. Machine learning (ML) refers to computer-based algorithms that learn and improve from data without being explicitly programmed. Training an ML model in this context involves using historical data, including previously extracted dimensionless metrics and the results or performance indicators of the manufacturing process, to identify patterns and relationships that enable prediction of the optimal metrics.

[0029] In a further aspect, training the machine learning model comprises determining a respective model performance for each of the plurality of dimensionless metrics, wherein determining the respective model performance is preferably preceded by minimizing a respective metric-dependent loss function.

[0030] The process involves training the machine learning model, meaning the algorithm learns from data to recognize patterns and make predictions or decisions. This training involves adjusting model parameters to maximize the accuracy of the model's predictions. Part of the training process is evaluating the machine learning model's performance specifically with respect to each of the several extracted dimensionless metrics. This means that for each metric, the model's performance is individually examined to determine how well it can make accurate predictions or assessments based on that metric. Before evaluating model performance, it is preferable to minimize a respective loss function. The loss function, also known as the cost function, measures the error between the actual values ​​and those predicted by the model. The goal is to minimize this error, which means improving model accuracy.This step is crucial because it directly influences the model's adaptation to the specific data and task at hand. Minimizing the loss function is preferably "metric-dependent." This implies that a separate loss function is defined for each dimensionless metric, tailored specifically to the specific characteristics and requirements of that metric. This allows the model to be more effectively optimized for each metric individually, leading to more precise analysis and better overall model performance. In another aspect, the process parameter includes a power, a frequency, a temperature, a machining time, and / or a feed rate of a processing machine.

[0031] Power refers to the output power of the processing machine, typically measured in kilowatts (kW). Power can affect the speed and efficiency with which material is processed or workpieces are manufactured. Frequency in the context of processing machines can refer to the frequency of the tools or machine parts used, for example, the speed of a spindle in revolutions per minute (rpm). Frequency can directly influence the machining quality and speed of the manufacturing process. Temperature can refer to the operating temperature of the machine itself or to the temperature of the workpiece or component during machining. Temperatures should be kept within certain limits to avoid material damage and to ensure machining precision. Machining time is the duration required to complete a specific machining task.Machining time is a factor in the efficiency of the manufacturing process and the planning of production sequences. The feed rate is the speed at which the workpiece or tool is moved relative to each other, typically expressed in millimeters per minute (mm / min) or a similar unit. The feed rate influences the machining quality, the surface finish of the workpiece, and the machining time.

[0032] In a further aspect, a processing machine for processing a component in a manufacturing process is claimed, wherein at least one process parameter of the processing machine can be adjusted by the present method in one of its aspects.

[0033] A machining center is a machine designed to process materials to produce components with specific dimensions, shapes, and surface finishes. Machining can involve various processes, such as milling, turning, drilling, grinding, etc. The component is machined as part of a comprehensive manufacturing process. This implies that the machining center is integrated into a production line or production flow aimed at transforming the component from a raw or pre-processed state into a finished product or component. The machining center can be configured or controlled so that at least one process parameter is adjustable. Process parameters can be diverse and include variables such as power, frequency, temperature, machining time, feed rate, etc.This adjustability is essential for adapting the machine to different materials, component geometries, or specific manufacturing process requirements. The process parameters can be adjusted using a specific procedure. This procedure can be manual, semi-automatic, or fully automatic and includes methods or algorithms that determine how and to what extent the process parameters should be adjusted for optimal results.

[0034] In a further aspect, a computer program with program code is also claimed for executing at least parts of the method according to the invention in one of its embodiments when the computer program is executed on a computer. In other words, the invention provides a computer program (product) comprising instructions that, when the program is executed by a computer, cause the computer to execute the method / steps of the method according to the invention in one of its embodiments.

[0035] In a further aspect, a computer-readable data carrier with program code of a computer program is also proposed for executing at least parts of the method according to the invention in one of its embodiments when the computer program is executed on a computer. In other words, the invention relates to a computer-readable (storage) medium comprising instructions which, when executed by a computer, cause the computer to execute the method / steps of the method according to the invention in one of its embodiments.

[0036] The described designs and further training courses can be combined as desired.

[0037] Further possible embodiments, developments and implementations of the invention also include combinations of features of the invention described previously or below with regard to the embodiments that are not explicitly mentioned. Short description of the drawings

[0038] The accompanying drawings are intended to provide a further understanding of embodiments of the invention. They illustrate embodiments and, in conjunction with the description, serve to explain principles and concepts of the invention.

[0039] Other embodiments and many of the aforementioned advantages will become apparent upon review of the drawings. The elements illustrated in the drawings are not necessarily drawn to scale.

[0040] They show: Fig. 1 a schematic flow diagram of the method according to the invention; and Fig. 2 a schematic block diagram of the method and device according to the invention.

[0041] In the figures of the drawings, the same reference symbols designate the same or functionally equivalent elements, parts or components, unless otherwise stated.

[0042] Fig. 1 shows a schematic flow diagram of a method for adjusting a process parameter in a manufacturing process to optimize a component quality of a component that is processed by the manufacturing process.

[0043] In any embodiment, the method can be carried out at least partially by a device 100, which for this purpose can comprise several components not shown in detail, for example, one or more provision devices and / or at least one evaluation and computing device. It is understood that the provision device can be designed jointly with the evaluation and computing device or can be different from it. Furthermore, the device 100 can comprise a storage device and / or an output device and / or a display device and / or an input device.

[0044] The computer-implemented method comprises at least the following steps: In a step S1, quality data of the component is provided, which is determined by a quality inspection of the component quality. In a step S2, at least one dimensionless key figure is extracted from the quality data. In a step S3, the component quality is assessed based on at least one dimensionless key figure. In step S4, the process parameter is adjusted based on the evaluation.

[0045] Fig. 2 shows a block diagram of an embodiment of the method and / or the device 100. The quality data 200 of the component is provided in order to be further processed according to the method. The quality data 200 includes, for example, information about a hardness test of the component or about a test of the component's microstructure (µ-structure). Other quality parameters are also conceivable. To provide the quality data 200, the component can be machined, for example, in a finishing step, in particular by inductive hardening. The finishing step is designated by 202. To manipulate the quality of the component after finishing, process parameters 204 of the finishing process can be set. Examples include a power, a frequency, a temperature, a processing time, and / or a feed rate of a processing machine.The finishing step 202 can be preceded by at least one pre-processing step 206, in particular bending and / or forming and / or annealing. Thus, an initial material of the component, which, for example, corresponds to a manufacturer's specification, can already undergo several process routes prior to the finishing step 202, which can influence the material properties, for example, a chemical composition, a hardness, and / or a grain size in the microstructure.

[0046] The quality data 200 are converted into at least one dimensionless number, using, for example, the Buckingham pi theorem 208.

[0047] An example of a key figure is: Pe=L⋅να where L is the hardening depth [m], v is the feed rate [m / s] and α is the thermal diffusivity [m 2 / s].

[0048] If multiple dimensionless metrics are extracted from the quality data 200, a dimensionless metric that optimally matches the manufacturing process is determined S5 by training a machine learning model. Training the machine learning model involves determining a respective model performance for each of the multiple dimensionless metrics, wherein determining the respective model performance is preceded by minimizing a respective metric-dependent loss function.

[0049] Subsequently, the component quality assessment S3 is performed based on the at least one dimensionless key figure. The component quality assessment S3 based on the at least one dimensionless key figure preferably includes an assessment S6 of a component anomaly by comparing the at least one dimensionless key figure with a limit value or a limit interval 210, wherein the assessment classifies the component anomaly as OK or not OK. Fig. Figure 2 shows an exemplary diagram of such an evaluation, with the indicator Pe plotted over time.

[0050] Following the evaluation, the process parameter S4 is adjusted based on the evaluation. Optimal parameter adjustment can be performed manually, under expert supervision, or automatically. The process parameters 204 can be a power, a frequency, a temperature, a processing time, and / or a feed rate of a processing machine. QUOTES CONTAINED IN THE DESCRIPTION

[0000] This list of documents submitted by the applicant was generated automatically and is included solely for the convenience of the reader. This list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions. Cited patent literature

[0000] DE 102022203835 A1

[0003]

Claims

[1] A method for adjusting a process parameter in a manufacturing process to optimize a component quality of a component that is processed by the manufacturing process; the method comprising: - Providing (S1) quality data of the component, which are determined by a quality inspection of the component quality; - Extracting (S2) at least one dimensionless indicator from the quality data; - Assessment (S3) of the component quality based on at least one dimensionless key figure; and - Setting (S4) the process parameter based on the evaluation. [2] Method according to claim 1, wherein the quality test comprises a non-destructive and / or a destructive test of the component quality, in particular a component hardness and / or a quality of a component structure. [3] Method according to claim 1 or 2, wherein the component is pre-machined before the quality inspection in the manufacturing process by at least one pre-machining step, in particular by bending and / or forming and / or annealing, and wherein the component is finished in the manufacturing process by at least one finishing step, in particular by hardening, particularly preferably by inductive heat treatment. [4] Method according to one of the preceding claims, wherein the extraction (S2) of the at least one dimensionless characteristic number from the quality data is carried out by applying a Buckingham pi theorem. [5] Method according to one of the preceding claims, wherein the evaluation (S3) of the component quality based on the at least one dimensionless characteristic number comprises an evaluation (S6) of a component anomaly by comparing the at least one dimensionless characteristic number with a limit value or a limit interval, wherein the evaluation classifies the component anomaly as OK or not OK. [6] Method according to one of the preceding claims, wherein, if several dimensionless key figures are extracted from the quality data, a determination (S5) of a dimensionless key figure optimally suited to the manufacturing process is carried out by training a machine learning model. [7] The method of claim 6, wherein training the machine learning model comprises determining a respective model performance for each of the plurality of dimensionless metrics, wherein determining the respective model performance is preceded by minimizing a respective metric-dependent loss function. [8] Method according to one of the preceding claims, wherein the process parameter comprises a power, a frequency, a temperature, a processing time and / or a feed rate of a processing machine. [9] Computer program with program code for carrying out at least parts of a method according to one of claims 1 to 8 when the computer program is executed on a computer. [10] Computer-readable data carrier with program code of a computer program for carrying out at least parts of a method according to one of claims 1 to 8 when the computer program is executed on a computer. [11] Processing machine for processing a component in a manufacturing process, wherein at least one process parameter of the processing machine can be adjusted by the method according to one of claims 1 to 8. [12] Device (100) for adjusting a process parameter of a manufacturing process to optimize a component quality of a component that is processed by the manufacturing process; the device (100) comprising an evaluation and computing device that is designed to carry out the following steps: - Providing quality data of the component, which is determined by a quality inspection of the component quality; - Extracting at least one dimensionless metric from the quality data; - Assessment of component quality based on at least one dimensionless key figure; and - Setting the process parameter based on the evaluation.

Images