Method and system for computer-implemented design and / or (further) development and / or layout and / or service life optimisation of machine elements of a metallurgical plant

EP4630958A1Pending Publication Date: 2025-10-15SMS GROUP GMBH
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
EP2023821933
Authority / Receiving Office
EP · EP
Patent Type
Applications
Current Assignee / Owner
Priority Date
2023-09-14
Filing Date
2023-12-06
Publication Date
2025-10-15

AI Technical Summary

Technical Problem

Designing and optimizing machine elements in metallurgical plants, such as rolling mills, is hindered by the lack of information about wear-induced functional reduction due to insufficient knowledge of operational loads and load history, leading to suboptimal maintenance and performance.

Method used

A computer-implemented method and system that calculates and optimizes machine element design characteristics using operational and production data, incorporating measurement data to adapt design parameters dynamically, allowing for iterative improvements and extended service life through feedback loops and machine learning techniques.

Benefits of technology

This approach enhances the precision of load prediction and service life estimation, enabling targeted maintenance and optimization of machine elements, improving performance and extending their service life while reducing operational costs.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 00000024_0000
    Figure 00000024_0000
  • Figure 00000025_0000
    Figure 00000025_0000
  • Figure 00000026_0000
    Figure 00000026_0000
Patent Text Reader

Abstract

The invention relates to a method for the computer-implemented design and / or (further) development and / or layout and / or service life optimisation of machine elements of a metallurgical plant, comprising the following method steps: calculating and / or providing at least one first data set comprising structural characteristic data of the machine element based on an assumed load and service life expectation of the machine element in operational use; detecting the actual load and / or load history of the machine element in the metallurgical plant and / or from the production history of the machine element in the form of application-related measurement data and information from operational use and / or production of the machine element; and returning and using the measurement data to calculate at least one second changed data set with structural characteristic data, which are optimised taking into consideration the installation situation of the machine element and / or the application-related measurement data.
Need to check novelty before this filing date? Find Prior Art

Description

[0001] Method and system for computer-implemented design and / or (further) development and / or layout and / or lifetime optimization of machine elements of a metallurgical plant

[0002] The invention relates to a method and system for computer-implemented design and / or (further) development and / or layout and / or service life optimization of machine elements of a metallurgical plant.

[0003] Machine elements in rolling mill drive trains, in particular, are subject to wear. Their wear is due to numerous continuous and discontinuous, regular and sporadic, avoidable and unavoidable phenomena associated with the rolling process. Examples include maximum torques during strip threading and unthreading, alternating gripping and slipping of the strip, and also the continuous loading resulting from the deformation of the rolled stock. When a critical level of wear is finally reached, the machine element must be replaced. For some machine elements, such as rollers or rolls, or, for example, gear spindles or gear spindle sleeves of roll drives, repair or overhaul can extend the service life of the machine element while maintaining or improving performance.

[0004] When designing machine components, the designer usually has no information about the causes of the continuous reduction in functionality due to wear. Often, the exact load on the machine component during operational use, and in particular the load history and current condition of the component, are unknown or only insufficiently known, making targeted improvement or further development of the product or machine component impossible.

[0005] The invention is therefore based on the object of providing a method and a system for the computer-implemented design and / or (further) development and / or design and / or service life optimization of machine elements of a metallurgical plant, which takes this disadvantage into account.

[0006] The problem is solved by a method according to patent claim 1 and by a system having the features of patent claim 15.

[0007] Advantageous embodiments of the invention emerge from the subclaims.

[0008] According to a first aspect of the invention, a method for the computer-implemented design and / or configuration and / or (further) development and / or service life optimization of machine elements of a metallurgical plant is provided, which method comprises calculating and / or providing at least one first data set, comprising design characteristics of the machine element based on an assumed load and service life expectancy of the machine element in operational use, recording the actual load and / or production history of the machine element during manufacture and the load history of the machine element in the metallurgical plant in the form of application-related measurement data and information from the manufacturing process and the operational use of the machine element, and feeding back and using the measurement data to calculate at least one second modified data set with design characteristics,which are optimized taking into account the installation situation of the machine element and / or the production- and / or application-related measurement data. According to a first aspect, the method according to the invention comprises an improvement in the static design of the machine element. In particular, the method according to the invention provides more and improved target values ​​for the production of the machine element, which are taken into account in the design and, in particular, make the comparison of permissible and actual stress more precise.

[0009] The additional recording of actual values ​​during production enables the calculation or determination of correlations between these values ​​and the load during operational use. The influence of variations in the actual values ​​of the machine element is thus immediately recognizable, so that the method according to the invention provides additional certainty in the design and / or layout of the machine element.

[0010] The term design or layout in the sense of the present invention is to be understood as meaning that the method includes both a design of the component in the narrowest sense, ie with regard to material selection, dimensioning, surface design, heat treatment, etc., as well as a complete redesign of the component, for example in the sense of contour optimization or the like.

[0011] A further aspect of the invention encompasses the dynamic design of the component or machine element in the sense that operational data is recorded and evaluated and compared with the assumptions of the original, static design and, in particular, with the manufacturing data. This enables, in particular, the possibility of adjusting process variables and limit values ​​during operational use of the component or machine element, as well as the possibility of adapting the manufacturing process, based on the determination of the actual service life and the currently remaining service life. One aspect of the invention relates to automatic component optimization or automatic / automated further development of the machine element.

[0012] The method provided according to the invention can be summarized as providing, from one aspect, a quasi-dynamic design and / or service life optimization and / or (further) development and / or design of machine elements of a metallurgical plant, with feedback and appropriate use of information and measurement data relating to the machine element from production and / or operational use. Using the measurement data from operational use and / or the data from the manufacture of the machine element, the design and other characteristic data of the machine element are preferably adapted and optimized automatically. This can also be done repeatedly and iteratively, for example, with respect to a specific machine element.

[0013] The method according to the invention relates both to the computer-implemented design and / or redesign of the machine element in the sense of new production and to the service life optimization of the machine element in the sense of repair, maintenance or restoration.

[0014] Computer-implemented in the sense of the present invention means that the method is carried out at least partially automatically with the aid of computer hardware and software.

[0015] The manufacture of machine elements subject to high mechanical stress typically comprises numerous steps up to the production of the raw part, e.g. melting, casting, forming, heat treatment, and finally to the production of the finished machine element, e.g. by machining with a geometrically defined cutting edge, machining with a geometrically undefined cutting edge, welding, heat treatment. For all steps, the specified parameters (target values), which often characterize the quality of the machine element (e.g. geometry and surface finish), and the actually achieved values ​​(actual values) can be documented in detail, and the usually extensive, often continuously occurring measured values ​​(process parameters such as forces, moments or feed rates) can be recorded and stored.

[0016] These data contain information or can be converted into information using suitable methods that show a direct influence of various manufacturing-related variables on the load and properties of the machine element during subsequent operation and reveal interactions and correlations that were previously unknown and undiscovered.

[0017] The precise geometry and surface shape, for example, influence the loads encountered during operation. The actual material composition and heat treatment condition influence, for example, the permissible load or load capacity.

[0018] The data, preferably after plausibility checks and validation and conversion into relevant information, can be used to evaluate the performance of the machine element, taking into account the structural design on the one hand and the assumed operating load on the other, and to quantitatively determine important related static parameters such as the expected service life. This improves the accuracy and quality of load prediction during design compared to the standard approach, which usually only considers a few target values ​​from the manufacturing process.

[0019] It is also possible, for example, within the framework of a parametric study, to determine how a variation in key production process variables affects the load variables during operation and the resulting static characteristic values ​​of the design. In this context, static characteristic values ​​are those calculated using load variables based on assumptions or specified by plant limits or process constraints, and which are not compared and compared with the actual load during operation. Once the static characteristic values ​​have been calculated, they no longer change.

[0020] In contrast, dynamic parameters take into account the actual loads during operation, which deviate from the forecast or which can only be imprecisely represented by the assumptions made during design, and possibly also the load history, which leads, for example, to only a remaining service life remaining instead of the calculated service life.

[0021] In the sense of dynamic design and / or service life optimization (e.g. determination of safety factors within the framework of the mathematical strength verification), the subsequent operation of the machine element within a metallurgical plant is also recorded as comprehensively and seamlessly as possible according to the invention.

[0022] The recording and evaluation of the measurement data using suitable methods and models and the comparison with the parameters and measurement data from the production of the machine element allow the conversion of the static parameters from the production phase into dynamic and current parameters that are valid for the respective operation and thus the specific application.

[0023] For example, actual service lives and current remaining service lives can be calculated, but also limit values ​​regarding the load on the machine elements that must be observed during production can be adapted to the actual condition of the machine element and the actual course of the process and can also be exceeded in a defined manner if the associated and always known reduction in safety and / or service life is acceptable.

[0024] Conversely, particularly important, expensive or difficult to procure or replace machine elements can be relieved to a certain extent to the detriment of other machine elements if precise information is available on the currently occurring and currently permissible load.

[0025] The method according to the invention preferably comprises changing the boundary conditions assumed for calculating and / or providing the design parameters, both during production and during operation, in the event of a detected deviation between the assumed boundary conditions and information obtained from the measured data and / or information regarding the actual parameters and loads during the manufacturing process and operation, as well as calculating and / or specifying optimized design parameters taking the actual load into account. These optimized parameters were also referred to above as dynamic parameters or characteristic values.

[0026] It is therefore possible to adjust important process variables so that the load on the machine element corresponds to the prediction determined during the design process, or so that it is deliberately higher or lower if, for example, this relieves the load on machine elements or allows process-related or product-related goals to be achieved. In this case, precise knowledge of the actual values ​​or specified or calculated characteristic data based on assumed load, as well as comparison with the measured values ​​from operational use and, above all, the deviation between target values ​​and actual values, is of great importance. The individual process steps are preferably carried out in the order in which they are listed and can be repeated if necessary in order to achieve iterative optimization of the machine element.

[0027] Preferably, the measurement data are plausibility checked and / or validated on the basis of physical models and / or artificial intelligence.

[0028] The method may further comprise the step of contour optimization of the machine element based on the recorded actual load and / or load history of the machine element during operational use.

[0029] The method may further additionally or alternatively comprise one or more of the steps of optimizing the machine element with regard to the surface, the material used, and a heat treatment of the machine element.

[0030] In a variant of the method according to the invention, an automated review of given assumptions and / or derivation of new assumptions for the design of the machine element and / or the establishment of new relationships between the variables relevant for the design of the machine element is provided.

[0031] In one variant, the method according to the invention can comprise the method step of deliberately varying the load on the machine element during operational use to expand the database of measured data. For example, a deliberate exceedance of limit values / load limits of the machine element during operational use can be provided in order to achieve process optimization with regard to the component, process, and product (rolled product). The plausibility check and / or validation of the measured data and / or the contour optimization method step and / or the further optimization steps are preferably carried out using methods based on machine learning methods.

[0032] The calculation and / or specification of modified design parameters can also be carried out using procedures based on machine learning methods.

[0033] The method according to the invention includes the possibility of making changes to the machine element by recording comprehensive data and converting it into relevant information, thereby increasing its performance and expanding its functionality. Linking data from production and operation with data characterizing the current state of the machine element can allow not only a more precise design of the machine elements but also an optimization of the structural design.

[0034] For example, if contour optimization is carried out based on a measurement, the stresses acting on the machine element can be reduced. An example of such a measurement is the determination of wear-related material removal in the area of ​​a gear spindle sleeve that is in direct contact with a roll neck. As a result of the measurement and subsequent evaluation and assessment of the underlying mechanism, the geometry of a functional surface of the gear spindle sleeve can be changed, for example. The complexity of the process increases if, instead, several or even numerous measurements are evaluated together using suitable methods, and several or even numerous measures for machine element optimization are derived from them using suitable methods.Using suitable methods and special calculation procedures based on physical models and / or artificial intelligence, the influence of process variables and system settings on the machine element load during operation, which is calculated as part of the design of the machine element, can be verified and specified. For this purpose, in addition to the variables such as rolling force or temperature that are directly measured or determined using process models, for example, a rolling mill, a pass schedule model, or the profile, contour, and flatness model, additional variables are preferably measured or determined that allow direct conclusions to be drawn about the load on the machine element. This can be, for example, a current measurement of the geometry and / or surface or a local measurement of forces or stresses.With additional measurements, which, for example, allow for direct information about the wear condition or an indirect calculation of the remaining service life, the relationship between the above-mentioned influencing factors and the associated load, on the one hand, and the condition of the machine element, safety, or service life, on the other, can also be determined using methods and calculation procedures based on physical models and / or artificial intelligence. These two types of measurements also improve the quality of the design and enable the verification of important assumptions, such as the relationship between load variables and wear of the machine element, and the development of new correlations.

[0035] The relationship between the average or maximum torque transmitted by a machine element, such as a gear spindle sleeve, on the one hand, and the wear of the inner surfaces of the gear spindle sleeve that are in contact with the journal of the work roll, on the other hand, is a relationship that can initially be estimated using assumptions. This wear can lead to failure and thus the end of the service life of the gear spindle sleeve. The method according to the invention makes it possible to improve these assumptions by comparing the recorded and / or measured data. The relationship between characteristic quantities and parameters during the manufacture of the machine element and its subsequent loading during operation makes it possible, according to the invention, to improve previously made assumptions or to provide new relationships or assumptions.

[0036] The relationship between characteristic quantities and parameters during the manufacture of the machine element and its subsequent load during operation, which is often not known to the same extent as the relationship between characteristic quantities and parameters during operation and the load during operation, can also be analyzed using methods and calculation procedures based on physical models and / or artificial intelligence and verified analogously to the procedure described above.

[0037] Examples of such artificial intelligences include neural networks, adaptive algorithms, evolutionary algorithms, genetic algorithms, and the like. State-of-the-art examples include: Bayesian belief networks, decision trees, hidden Markov models, case-based reasoning, k-nearest neighbors, self-organizing maps, instance-based learning, support vector machines, artificial neural networks (ANNs), recurrent neural networks (RNNs), deep neural networks (DNNs), or convolutional neural networks (CNNs). Any conceivable combination of such approaches can also be used.

[0038] The comprehensive data situation and the evaluation of the data using, for example, analytical and numerical methods allow process optimization with regard to the machine element, but to the same extent also with regard to the process, e.g. stability, yield, use of alloying elements and energy, use of input materials, and product, e.g. properties, quality, reproducibility. This means that plant limits and process-related limit values ​​can be adjusted if the currently permissible load on the machine element is known. This can give the plant operator more flexibility in production or improve product quality. This can be the case, for example, in a hot rolling mill if larger (or smaller) pass reductions and / or higher or lower temperatures are used, thereby saving alloying elements or achieving better product properties.

[0039] The information measured directly or obtained through evaluation can be used according to the invention to further develop a component. Furthermore, component optimization can also be carried out automatically by initiating the necessary changes to the component using suitable rules and algorithms or highly complex dynamic and iterative models. While changes to the process to change the load can be made immediately, changes to the component to reduce the occurring load or increase the permissible load must be made by or at the manufacturer of the component and system. However, they can be initiated automatically, and the ordering process and the actual development process can also be automated if the infrastructure and suitable development tools are available.Finally, the method according to the invention can be used to design and manufacture customer-specific components based on actual loads. These take into account the actual driving behavior of the customer's operational application and thus the actual load on the components. They represent customized solutions that are difficult to implement by other suppliers of mechanical components.

[0040] Preferably, the characteristics or parameters of the production are selected from a group of data comprising material characteristics and / or material composition, surface quality, as well as the geometric dimensions and contours of the machine element.

[0041] Preferably, the measurement data from the operational use of the machine element are selected from a group of data comprising the tribological characteristics of the machine element in the specific installation situation, temperatures, torques, speeds and static loads.

[0042] Preferably, the measurement data are fed back at least partially from a process control and guidance system of a steelworks plant or a continuous casting plant or an annealing furnace or a hot-dip galvanizing line or a rolling mill or a strip treatment plant or a forging plant or a casting-rolling plant into a device for the computer-aided design of machine elements and used to optimize the characteristics or parameters of the machine element.

[0043] Preferably, the machine element is selected from a group of machine elements comprising the rolls of a rolling mill, roller table rollers of a rolling mill, linear guides of roller tables, rolling stands, and edging stands as well as drive shafts and gear parts and toothed spindles of rolling stands

[0044] The invention further comprises a system for the design and / or layout and / or service life optimization and / or further development of machine elements of a metallurgical plant, comprising at least one device for the computer-aided design of machine elements and means for measuring, recording, and storing operating data from the operational use of the machine element, wherein the system is configured to carry out the method described above. The invention is explained below using an exemplary embodiment and with reference to the accompanying drawings.

[0045] They show:

[0046] Figure 1 is a perspective view of a finishing train of a rolling mill as a metallurgical plant,

[0047] Figure 2 is a perspective view of the drive side of a rolling stand with parts of the drive as machine elements to be designed according to the method,

[0048] Figure 3a shows a cross section through a gear spindle sleeve as a machine element and

[0049] Figure 3b shows a perspective longitudinal section through the gear spindle sleeve.

[0050] The invention is described below based on a further development of gear spindle sleeves 1 as machine elements of drive trains for rolling stands F1-Fn of a finishing train 2 of a hot rolling mill for steel. The finishing train 2 comprises, in a known manner, a plurality of individual rolling stands F1-FN, which are part of the hot rolling mill as a metallurgical plant.

[0051] Such a finishing train 2 can, for example, comprise up to eight rolling stands F1-Fn, each of which is driven by drives 3 via geared spindles 4 as drive shafts. The geared spindle shafts 4 are each laterally connected to a gear 5 via geared spindle sleeves 1 and are part of the drive train for the rolling stands F1-Fn. On the drive side of the rolling stands F1-FN, an upper work roll 6 and a lower work roll 7 are connected to the respective geared spindle shaft 4 via a work roll journal 8 with a geared spindle sleeve 1, wherein the geared spindle shaft 4 engages with a corresponding gearing of the geared spindle sleeve 1 via a circumferential toothing 9. On the side of the geared spindle sleeve 1 opposite the toothing 9, the sleeve is provided with a cross-sectional profile complementary to the work roll journal 8.

[0052] Process variables relevant to the load on the gear spindle sleeve 1 and changing over the course of the finishing train 2 include the speed n and the torque M, which can change, for example, from the rolling stands F1 to F7 as follows:

[0053] Roll stand F1 F2 F3 F4 F5 F6 F7 n [rpm] 21 46 81 114 148 179 196

[0054] M [kNm] 2430 1460 770 340 250 150 70

[0055] Due to the reduction in thickness of the rolled stock, the speed of the work rolls increases significantly from an upstream roll stand F1 to a downstream roll stand F7, as can be seen from the table above. Accordingly, the torque M decreases.

[0056] The load on the gear spindle sleeve 1 during operational use also depends on the installation situation, for example, depending on whether the gear spindle sleeve 1 is connected to the gear spindle shaft 4 on the stand side or the gearbox side, or whether the shaft is arranged in the drive train of an upper work roll 6 or a lower work roll 7. The corresponding data, which can be provided in the form of measured data, and the information on which stand F1 to F7 of the finishing train 2 the respective gear spindle sleeve 1 is arranged in, are important for optimizing the service life of the gear spindle sleeve. Using simple sensors, the position and direction of rotation of the various gear spindle sleeves 1 in the finishing train 2 can be determined.An evaluation of the measurement data supplied by the sensors also provides information as to whether the respective gear spindle sleeve 1 is always used in the same position or whether it is installed and operated at a different position within the rolling stand F1-Fn or in another rolling stand F1-Fn.

[0057] On the one hand, key load data for the gear spindle sleeve, such as rpm, speed, rolling force, and rolling torque, can be estimated. On the other hand, rpm and speed can also be measured using additional sensors. Alternatively or additionally, the rolling force and torque values ​​can be obtained from the automation systems or a higher-level process control system of the rolling mill or from corresponding databases.

[0058] The measured or otherwise determined values ​​are assigned to each individual gear spindle sleeve 1 for the entire operating period. This creates an individual load history for each gear spindle sleeve 1 as a load collective with all available and relevant measurement data as a function of time.

[0059] The load collective is used as the basis for the design.

[0060] Gear spindle sleeves 1 that are always used in the same location must be able to withstand the load occurring there permanently or for a defined period of time. Gear spindle sleeves 1 that are exposed to high forces and moments can be designed differently than those operated at low loads. Gear spindle sleeves 1 that are operated at high speeds can be designed differently than slow-running gear spindle sleeves. For gear spindle sleeves 1 that are used in both positions within at least one stand, it is taken into account that the stress is of a changing rather than pulsating nature due to the change in direction of rotation and that the permissible load is therefore reduced without changing the design.

[0061] According to the invention, the most important design criteria, guidelines, and algorithms are stored in appropriate software based on machine learning methods. This software is trained accordingly so that the individual design of the machine element, in particular the gear spindle sleeve 1, is automated and runs without intervention by a designer.

[0062] The load collective can be used in a next and further process step as a basis for an individual further development of the gear spindle sleeve 1 that is optimally tailored to the respective application.

[0063] Gear spindle sleeves 1 that are always used in the same location must be able to withstand the loads occurring there permanently or for a defined period of time. Gear spindle sleeves 1 that are exposed to high forces and moments are dimensioned larger using this process. These can be equipped with a further developed and thus improved geometry of the gearing 9 and / or the roll neck mount.

[0064] Gear spindle sleeves 1 that operate at high speeds can have different structural designs and be manufactured from different materials and using different production methods than slow-running gear spindle sleeves 1. According to the invention, gear spindle sleeves 1 that are used in various rolling stands F1-Fn are specifically optimized for their heterogeneous and complex load scenario.

[0065] For gear spindle sleeves 1 that are used only at one position within a rolling stand F1-Fn, a pulsating load can be assumed instead of a fluctuating one. Most importantly, an asymmetrical tooth shape can be selected for the gearing 9. This results in a restriction to one direction of rotation, but higher torques can then be transmitted in this direction. If higher torques are not required, a smaller and more economical size can be selected.

[0066] Particularly for gear spindle sleeves 1 that are used in various rolling stands F1-Fn, but also for all other gear spindle sleeves, the data-supported load analysis may reveal that a technically optimal or economically favorable design or construction is difficult to achieve due to the loading situation or the loading history. The software can then make alternative suggestions for modified operation or implement them automatically. This may mean, on the one hand, that the rolling stand in which the gear spindle sleeve 1 is currently located is relieved of the load with regard to the critical parameters, or, on the other hand, that the gear spindle sleeve 1 is used in a specific rolling stand F1-Fn but not in another.

[0067] Furthermore, the process may include a computer-aided analysis of all load histories and all measures, derivation and manual or automated application of design principles, possibly taking customer specifications into account as additional boundary conditions.

[0068] 1 gear spindle sleeve

[0069] 2 finishing line

[0070] 3 drives

[0071] 4 toothed spindles

[0072] 5 gearboxes

[0073] 6 upper working roll

[0074] 7 lower working roll

[0075] 8 work roll necks

[0076] 9 Gearing

[0077] F1-Fn rolling stands n speed

[0078] M torque

Claims

Patent claims 1. Method for the computer-implemented design and / or (further) development and / or layout and / or lifetime optimization of machine elements of a metallurgical plant, comprising the following process steps: Calculating and / or providing at least one first data set comprising design characteristics of the machine element, based on an assumed load and service life expectancy of the machine element in an operational application Recording the actual load and / or the load history in the metallurgical plant and / or the manufacturing history of the machine element in the form of application-related measurement data and / or information from the operational use and / or manufacturing of the machine element, and Trace back and use the measurement data to calculate at least a second modified data set with design characteristics for a modified component, which are optimized taking into account the installation situation of the machine element and / or the application-related measurement data.

2. Method according to claim 1, comprising changing the assumed load and / or expected lifetime in case of a detected deviation between the assumed load and an actual load derived from the measurement data and / or information and calculating and / or specifying optimized design characteristics taking the actual load into account. Method according to one of claims 1 or 2, characterized in that the method steps are carried out in the order in which they are listed and, if necessary, repeated iteratively. Method according to one of claims 1 to 3, characterized in that the measured data are checked for plausibility and / or validated. Method according to one of claims 1 to 4, further comprising the method step of contour optimizing the machine element based on the recorded actual load and / or load history of the machine element during operational use. Method according to one of claims 1 to 5, further comprising one or more of the steps of optimizing the machine element with regard to the surface, the material used, and a heat treatment.Method according to one of claims 1 to 6, further comprising the automated verification of given assumptions and / or derivation of new assumptions for the design of the machine element and / or the establishment of new relationships between the variables relevant to the design of the machine element. Method according to one of claims 1 to 7, further comprising the method step of varying the load on the machine element during operational use to expand the database of the measured data. Method according to one of claims 1 to 8, characterized in that a plausibility check and / or validation of the measured data and / or the method step of contour optimization and / or the further optimization steps are carried out using methods based on machine learning methods. Method according to one of claims 1 to 9, characterized in that the calculation and / or specification of modified design characteristics is carried out using methods based on machine learning methods. Method according to one of claims 1 to 10, characterized in that the characteristics are selected from a group of data comprising material characteristics and / or material composition, surface quality, as well as the geometric dimensions and contours of the machine element.Method according to one of claims 1 to 11, characterized in that the measurement data are selected from a group of data comprising the tribological parameters of the machine element in the specific installation situation, temperatures, torques, speeds, and static loads. Method according to one of claims 1 to 12, characterized in that the measurement data are fed back, at least in part, from a process control and guidance system of a steelworks, a continuous casting plant, an annealing furnace, a hot-dip galvanizing line, a rolling mill or a strip treatment plant, a forging plant, or a casting-rolling plant to a device for the computer-aided design of machine elements. Method according to one of claims 1 to 13, characterized in that the machine element is selected from a group of machine elements comprising rolls of a rolling mill, roller table rollers of a rolling mill, linear guides of roller tables, rolling stands, and edging stands as well as drive shafts and gear parts and toothed spindles of rolling stands. System for the design and / or layout and / or service life optimization of machine elements of a metallurgical plant comprising at least one device for computer-aided Construction of machine elements and means for measuring, recording and storing operating data from the operational use of the machine element, wherein the system is adapted to carry out the method according to one of claims 1 to 14.