Method and system for adapting a manufacturing process

Abstract

The invention relates to a method (100) system (1) for adapting a manufacturing process (P), in particular a hot rolling process, and to the use of a method (100) for periodically operating a plant (10). In this case, i) a product property (E *) of the process product (2) is specified (S1). Furthermore, ii) an operating starting point (X) of a plant (10), in particular a hot-rolling plant, which is designed to carry out a manufacturing process (P) for producing the process product (2), is specified (S2), and iii) the process product (2) is produced in such a way that the manufacturing process (P) is carried out starting from the specified operating starting point (X) by means of the plant (10). Iv) determining (S4) an actual product property (E) of the produced process product (2) and v) determining (S5) an optimized operating point (X *) of the plant (10) by means of an optimizer (30) on the basis of the predefined product property (E *) and the determined actual product property (E).

Description

Technical Field

[0001] The invention relates to a method and a system for adapting a manufacturing process, in particular a hot rolling process. Background Art

[0002] In order to put a plant, such as a hot rolling mill, into operation, it is usually necessary to correctly set various parameters of the plant in order to be able to obtain products with the desired properties during the subsequent operation of the plant.

[0003] For this purpose, it is known to create a model of the system and to calculate the model under predefined target values ​​representing the desired product properties. Assumed external influences, for example in the form of parameterized ambient conditions, and model parameters (so-called start-up parameters, IBS parameters for short, which control the model behavior or parameterized control curves) are used as input variables for the model. Such a model can output manipulated variables with which the system can be operated. If necessary, at least individual model parameters can be optimized by operating the system several times with the initially calculated manipulated variables and comparing the properties of the products produced therewith with the desired properties.

[0004] For example, a method for optimizing a model of a device for the raw materials industry is known from WO 2018 / 192798 A1. The control device determines the desired rated operation (setting calculation) of the device using the model and the pre-given rated parameters for the output product. The model is parameterized using a large number of characteristic parameters (= model parameters) of the device to be modeled. After a large number of output products are produced, the actual size of the output product is compared with the expected size of the output product, and the first model parameter is re-determined based on this, and the model is re-parameterized accordingly. The expected parameter is determined here based on the actual operation of the device with the help of the model.

[0005] However, it is problematic in this case that a system which has been put into operation based on selected model input variables is often subject to changing external influences and / or the same manipulated variables lead to different product properties over a longer period of time, for example due to aging / wear of the system.

[0006] Furthermore, there are also manufacturing processes which are too complex for sufficiently accurate modeling or whose modeling is extremely time-consuming or computationally intensive anyway. If a sufficiently accurate model cannot be found or used, the IBS parameters cannot be optimized automatically by an algorithm either. In such cases, the IBS parameters must be determined manually. However, this requires a lot of experience from the corresponding personnel and is also very time-consuming.

[0007] In order to control the thickness quality in the rolling process in an optimized manner, it is known from EP 1 488 863 A2 to provide an online process model, which also performs an adaptation of model parameters by identifying variable process parameters. By means of an adaptation unit that communicates with the online process model, the adaptation of the control parameters is performed based on the adapted model parameters, and the adapted control parameters are applied online for implementing and adapting control functions and / or loops of a model-based multi-parameter control unit.

[0008] DE 195 08 476 A1 discloses a control system for a plant in the raw material or processing industry, which is designed by computer technology based on input a priori knowledge to automatically identify the state of the plant and the details of the production process carried out in the plant, and outputs instructions suitable for the situation, such as setpoints, for achieving a safe production result. The instructions suitable for the situation are optimized in a predetermined optimization routine, wherein the input a priori knowledge is continuously improved by the knowledge acquired internally in the process model during production using computer technology. Summary of the invention

[0009] Against this background, it is an object of the present invention to specify an improved method and an improved system with which even complex production processes can be automatically adapted and / or the properties of process products can be kept constant over a long period of time.

[0010] These objects are achieved by a method and a system for adapting a production process according to the independent claims.

[0011] Preferred embodiments are the subject matter of the dependent claims and the following description.

[0012] In the case of a method for adapting a manufacturing process, in particular a hot rolling process, in particular automatically, according to a first aspect of the invention, i) a product characteristic of a process product is predefined. Furthermore, ii) an operating start point of a plant, in particular a hot rolling plant, which is set up to carry out a manufacturing process for producing the process product is predefined, and iii) the process product is produced by carrying out the manufacturing process with the aid of the plant starting from the predefined operating start point. Iv) the actual product characteristic of the produced process product is determined, and v) based on the predefined product characteristic and the determined actual product characteristic, an optimized operating point of the plant is determined with the aid of an optimizer.

[0013] A manufacturing process in the sense of the invention is preferably a technical process in which a product is prepared, that is, produced. The product need not necessarily be a "finished" (final) product, but can also be a semi-finished product. In this context, a manufacturing process can also be part of a larger production process for the final product, such as hot rolling of a slab, in particular width forming in a roughing mill using a vertical stand.

[0014] An optimizer in the sense of the present invention is preferably a device for solving an optimization problem, in particular iteratively. For example, the optimizer can be set up to determine one or more values ​​of process parameters with which the process provides the desired result. Suitably, the optimizer uses a predefined optimization strategy for solving the optimization problem. For example, the selected process parameters can be changed according to the optimization strategy and, starting from the process results obtained therefrom, the changes in the process parameters are adapted so that the process results converge to the desired result.

[0015] Preferably, the optimizer is based on an optimization algorithm. The optimization algorithm may embody an optimization strategy.

[0016] In the sense of the present invention, executing a manufacturing process with the aid of a device starting from a predetermined operating start preferably means operating the device at a predetermined operating start or at least at an operating point determined based on the predetermined operating start. However, it may happen that, despite corresponding setting of operating parameters, the device still does not operate at the desired operating point, for example, the predetermined operating start. However, such execution of a manufacturing process can also be advantageously understood as execution with the aid of a device starting from a predetermined operating start.

[0017] One aspect of the invention is based on the following concept: an optimizer is used to determine an optimal operating point in conjunction with the (real) operation of a plant, at which the plant produces a process product with predetermined product properties. In particular, the optimization strategy on which the optimizer is based can be implemented by operating the plant. In this regard, the plant can be iteratively operated by the optimizer to determine the optimal operating point (or at least an operating point very close to the optimal operating point). The optimizer expediently predetermines boundary conditions, such as changes in operating parameters of the plant or corresponding operating parameter values, at which the plant should be operated. By analyzing the manufacturing process carried out by the operation of the plant, in particular of the produced process product, the optimizer can propose optimized operating parameter values, which may enable the production of process products with product properties that are closer to the predetermined product properties. It is not necessary for the optimizer to evaluate a model or a so-called (mathematically expressed) target function that characterizes the manufacturing process. The implementation of the optimization strategy by the (real) operation of the plant or the evaluation of the manufacturing process associated therewith, in particular over a longer period of time, also advantageously allows the consideration of changing operating conditions. For example, changes in external influences and / or changes in the state of the plant, for example due to aging, can be taken into account.

[0018] In order to adapt the manufacturing process, the optimization algorithm can be divided into individual calculation steps, for example. At the end of each calculation step, an evaluation of the manufacturing process is preferably performed using the process parameter values ​​predefined by the optimization algorithm, i.e. at a predefined operating start point. This evaluation is preferably performed by operating the system from a predefined operating start point, which is characterized by the predefined process parameter values. In this regard, the objective function is not called for a specific parameter composition, but an evaluation of the actual physical process can be performed for exactly this parameter composition. The result of the evaluation can then be used as a basis for calculating updated process parameter values, i.e. an optimized operating point, in the next calculation step.

[0019] The evaluation of the manufacturing process is preferably performed by determining the actual product properties of the process product produced by the manufacturing process. In particular, the determined actual product properties can be compared with predetermined product properties. The comparison result can be the quality of the actual product properties of the produced process product. Based on this quality, the optimizer can determine optimized process parameter values, i.e., optimized operating points for the system.

[0020] By means of this method for adapting the production process, the effort for starting up the system - and thus the costs - can be significantly reduced compared to manual methods. At the same time, a consistently good quality of the process products produced by the system can be ensured. In particular, when a new product range is produced or the system status changes, the optimal operating point of the system can be easily adapted.

[0021] Preferred embodiments of the invention and their developments are described below. Unless this is explicitly excluded, these embodiments can be combined with each other and with the aspects of the invention described below as desired.

[0022] In a preferred embodiment, the process product is produced by means of the plant instead of evaluating the mathematical target function by means of an optimizer. This allows the production process to be adapted by means of conventional optimizers, in particular conventional optimization algorithms. For example, plants for which no model or mathematical target function can be created can thus also be started by means of conventional optimizers.

[0023] According to the invention, the method is preferably repeated continuously with a predetermined product characteristic, i.e., while maintaining the predetermined product characteristic. In particular, the optimal operating point can be constantly re-determined or updated during normal operation of the system. This makes it possible to compensate for changes in operating conditions, such as changes in external influences or changes in system characteristics, such as due to aging or wear, if necessary, and thus achieve a consistently good quality of the process product.

[0024] The respectively determined optimized operating point is expediently specified as the starting point of operation, in particular in the corresponding subsequent repetition. In this regard, the optimizer can check, starting from the previously determined optimized operating point, whether operation of the system is possible at an operating point that is more suitable in view of the predetermined product characteristics. In this way, a particularly permanent convergence of the actual product characteristics to the predetermined product characteristics can be ensured.

[0025] Alternatively, the respectively determined optimized operating point can be used as the basis for predetermining the operating start point, in particular in the case of corresponding subsequent repetitions. In other words, the operating start point can be predetermined by determining the operating start point starting from the determined optimized operating point, for example by deriving the operating start point from the determined optimized operating point. For example, it is conceivable that one or more operating points in the vicinity of the determined optimized operating point are predetermined as the operating start point. Here, the operating points in the vicinity of the determined optimized operating point can be determined taking into account predetermined rules, that is, for example, by a predetermined selection or determination process, for example, according to a predetermined optimization strategy. As a result, the convergence of the actual product characteristics to the predetermined product characteristics can be accelerated in some cases.

[0026] However, it is also conceivable to randomly select the operation start point starting from the determined optimized operating point. In principle, the predetermined operation start point does not have to be located near the determined optimized operating point.

[0027] In another preferred embodiment, a plurality of process products are produced by carrying out a manufacturing process at different operating points with the device starting from a predetermined operating start. In other words, the behavior of the device—and thus the corresponding manufacturing process—can be tested at a plurality of operating points before the optimizer determines an optimized operating point. For example, the manufacturing process can be carried out with the device at different operating points around a predetermined operating start. In particular, so-called gradients for different process parameters can be determined in this way, based on which the optimizer ultimately determines an optimized operating point.

[0028] Starting from a predetermined operating start, for example from a previously determined optimized operating point, different operating points can be selected by changing at least one operating parameter of the system. It should be noted that when a manufacturing process is carried out multiple times at the same operating point of the system, the actual product properties are never exactly the same. Such dispersion, also known as natural noise of the process, is generally unavoidable, since each system component is only "precise" within a certain tolerance range. Therefore, a standard deviation can be determined as a measure of the dispersion of the product properties. Therefore, in order to determine the operating point, the operating parameters are preferably changed so much that the product properties change by more than the standard deviation of the product properties. In this context, the operating points in the vicinity of the predetermined operating start are preferably operating points that are obtained from the predetermined operating start by changing the operating parameters of the system by a maximum of 10%, preferably a maximum of 5%, in particular a maximum of 2%, as long as the product property change associated therewith is greater than the standard deviation of the product property. The other operating parameters are preferably kept constant here.

[0029] The actual product properties are expediently determined for each produced process product. Accordingly, it is preferred to determine an optimized operating point based on a predetermined product property and a plurality of determined actual product properties. In this regard, the optimizer can check which of the different operating points—that is, with which process parameter changes—the best result is achieved in view of the predetermined product property. The corresponding operating point can be provided as the most promising candidate for achieving the predetermined product property as an optimized operating point, for example as a new starting point for optimizing the process.

[0030] In order to enable efficient operation of the method, in another preferred embodiment, the dispersion of the actual product characteristics of the process product obtained when the manufacturing process is performed multiple times in the same operating point of the device is determined. For example, the standard deviation can be determined as a measure of the dispersion of the product characteristics. The determined dispersion measure is preferably used as a basis for determining different operating points. In other words, the distances between the individual operating points, in particular the minimum distance, and / or the distances from a predetermined operating starting point, such as the last determined optimized operating point, in particular the minimum distance, can be specified based on the determined dispersion. For example, the intensity of the change of at least one operating parameter of the device can depend on the intensity of the determined dispersion. It can be ensured that the difference between the determined actual product characteristics and the predetermined product characteristics originates from the change of the operating point and not just from "noise".

[0031] Preferably, different operating points around a predetermined operating start point are determined by respectively changing one of a plurality of operating parameters. Preferably, this corresponds to a change in a process parameter. Each change in an operating parameter can represent a gradient formation in one direction in the corresponding parameter space. The gradients determined in this way for different operating parameters are preferably linearly independent in the operating parameter space and, in a further development, are even orthogonal to one another. Based on the respective determined actual product characteristics corresponding to the changes or the formed gradients, the optimizer can decide in which direction in the parameter space the search for the optimal operating point should continue.

[0032] In another preferred embodiment, a plurality of process products are produced by carrying out the manufacturing process several times at the same operating point with the aid of the system. The product properties corresponding to the operating point can thus be determined more accurately. In particular, a standard deviation can be determined, which serves as a measure of the possible dispersion of the product properties of the respectively produced process products.

[0033] Expediently, the actual product property is determined for each produced process product, and the optimized operating point is determined based on the predetermined product property and the location parameter of a plurality of determined actual product properties. In this case, the location parameter can preferably be understood as a mathematical summary of a plurality of determined actual product properties, in particular a characteristic number of a sample represented by a plurality of determined actual product properties. Such a location parameter can be, for example, the arithmetic mean, median, mode, etc. of a plurality of determined actual product properties.

[0034] For example, a production process can be carried out at least three times, preferably at least four times, and in particular at least five times by means of the device at the same operating point. It has been shown that with this number of production processes per operating point, i.e. the produced process products, changes in the actual product properties can be reliably detected and accordingly taken into account, for example averaged. At the same time, with this number of production processes per operating point, the resulting expenditures can still be fully understood.

[0035] In another preferred embodiment, if the difference between the predetermined product characteristic and the determined actual product characteristic, in particular the target value representing the actual product characteristic and the predetermined target value representing the predetermined product characteristic, reaches or does not exceed a threshold value, the optimizer determines the predetermined operation starting point as the optimized operation point. In other words, if the determined actual product characteristic substantially matches the predetermined product characteristic, the optimizer does not propose a new operation point as the operation starting point. In this regard, a limit can be introduced, which can prevent continuous switching of the operation point when the process products are manufactured successively in the region of the optimal operation point. Therefore, the constant quality of the process products can be ensured during normal operation of the device.

[0036] It is advantageous here if, depending on the difference between the predefined product property and the determined actual product property, the number of times the manufacturing process is carried out at the same operating point by means of the device is increased in order to determine the actual product property at this operating point as accurately as possible, preferably until a threshold value is reached or not exceeded. The number of repetitions is expediently selected such that the standard deviation of the product property is smaller than the difference. The threshold value can accordingly be selected as the value to which the standard deviation of the product property converges when a high number of repetitions of the manufacturing process is repeated.

[0037] In a further preferred embodiment, the optimizer is designed to detect whether there are more than one optimal operating point for producing a process product with predetermined product properties. For example, it is conceivable that by operating the rolling mill stand at different operating points, a specific material quality of the longitudinal flat-rolled product can be achieved. In this case, at least one of the determined operating points that is equivalent with respect to the predetermined product properties is expediently output as an optimized operating point. In principle, the detection can be performed online during the production process or offline using recorded data of the production process.

[0038] If the predefined operating start point is within the limits of the performance of the device, it may happen that the device does not operate at the desired operating point despite the setting of operating parameters corresponding to the operating start point. Therefore, in another preferred embodiment, it is checked whether the operation of the device meets at least one predefined criterion when executing the manufacturing process. In particular, it can be checked whether the manufacturing process is executed at or at least in the vicinity of the predefined operating start point. For this purpose, recorded data of the manufacturing process, such as manipulated variables of the device, can be evaluated.

[0039] If it is determined that the manufacturing process is not carried out at the desired operating point by means of the device, for example not at or near a predetermined operating start point, an optimized operating point can be determined in various ways. It is therefore preferred to determine an optimized operating point of the device based on the result of the check.

[0040] For example, it is conceivable to repeat the production of the process product again at the predefined operation start point, hoping that the operation of the plant now meets the predefined criteria. Alternatively, a new operation start point that is closer to the operation point at which the operation of the plant meets the criteria can be predefined. If the optimizer predefined the operation start point, for example, starting from an operation point, for example a previously determined optimized operation point, and indicates that the criteria are not met at the predefined operation start point, the optimizer can predefined an operation point between the predefined operation start point and the previously determined optimized operation point as the new operation start point.

[0041] However, even if the criterion is not met, it is conceivable that the determined actual product characteristic is also accepted and used for determining the (next) optimized operating point. For example, the determined actual product characteristic can be compared with the prediction of the optimizer. If the actual product characteristic is closer to the predetermined product characteristic than the product characteristic predicted for a predetermined operating start, the (next) optimized operating point is preferably determined based on the actually determined product characteristic.

[0042] In another preferred embodiment, the optimized operating point is determined by the optimizer with the aid of a trust region method and / or a line search method. In particular, the optimized operating point can be determined with the aid of a Sequential Quadratic Programming (SQP) variant of the trust region method. Preferably, the optimizer determines the "correction" of the operating point starting from a predetermined operating start in the sense of the trust region method, in particular using the SQP variant, by solving a quadratic minimization problem under constraints. The "usability" of this correction is then measured, expediently, by producing a process product with the aid of the plant at the optimized operating point and, for example, comparing the corresponding actual product characteristics with the predetermined product characteristics. Based on the results of this "measurement", the constraints of the quadratic minimization problem can be changed so that a higher level of correction can be made to the (new) operating start in further iterations of the method.

[0043] In this regard, the optimizer preferably has an optimization algorithm, in particular an SQP variant of a trust region method or a reinforcement learning-based optimization algorithm. However, other optimization algorithms are also conceivable in principle. In order to be able to adapt to changing operating conditions, such as changing external influences or aging of the system, the optimization algorithm preferably has no termination criteria. Even if the determined actual product properties correspond essentially to the predefined product properties, the optimization of the operating point or the method for that matter is preferably not terminated. This is because at a later point in time, the changed operating conditions may make it unnecessary to operate the system at another (optimal) operating point in order to be able to continue to produce a process product with the predefined product properties.

[0044] In this respect, the method is suitable, in particular in this embodiment, but also in all other embodiments described above or in combinations thereof, for regular, in particular permanent, operation of the installation.

[0045] According to a second aspect of the invention, the method according to the first aspect of the invention is used for regular operation of a plant, in particular a hot rolling mill. In normal operation, the plant can basically operate continuously, while the optimizer iteratively determines the optimized operating point, and the plant is operated at the respectively determined optimized operating point after each iteration. The process can be fully automated, so that the expenditure and thus also the costs not only for startup but also in normal operation of the plant can be significantly reduced.

[0046] According to a third aspect of the invention, a system for adapting a manufacturing process, in particular a hot rolling process, in particular automatically, comprises a device for producing a process product. The device is preferably a hot rolling plant, in particular a vertical rolling mill stand for rolling longitudinal flat products with a lower horizontal rolling mill stand for longitudinal flat products. The process product can be produced expediently by carrying out the manufacturing process with the aid of the device near a predetermined start of operation. The system further comprises a device for determining the actual product characteristics of the produced process product and an optimizer, which is set up to determine an optimized operating point for the device based on the predetermined product characteristics of the process product and the determined actual product characteristics. Such a system allows the manufacturing process to be evaluated under boundary conditions predetermined by the optimizer, in particular selected in a test manner, in order to find an optimal value of an operating parameter or at least a value close to the optimal value. The system furthermore makes it possible to take into account operating conditions that change during normal operation, such as changes in external influences, changes in the state of the device, for example due to aging, and / or changes in the material being processed, for example rolled.

[0047] The plant may be an industrial plant, such as a hot rolling plant, which may include, for example, a casting machine for producing metal strips, in particular slabs, from molten metal, a vertical rolling mill stand (which is sometimes also referred to as an edge former) and a horizontal rolling mill stand for reducing the thickness of the slab.

[0048] The device for determining the actual product property can have a sensor device which can detect the process product, for example a slab, in a sensory manner and determine the actual product property from the generated sensor data.

[0049] The optimizer can be constructed in hardware and / or software technology. The optimizer can have, in particular, a processing unit preferably connected to a storage and / or bus system data or signal. For example, the optimizer can have a microprocessor unit (CPU) or a module of such a microprocessor unit and / or one or more programs or program modules. The optimizer can be constructed to execute commands implemented as programs stored in a storage system, detect input signals from a data bus and / or send output signals to a data bus. The storage system can include one or more, in particular different storage media, in particular optical, magnetic, solid-state and / or other non-volatile media. The program can have the characteristic that the program at least partially embodies or can execute the method described here, so that the optimizer can at least perform the individual steps of such a method, and thus in particular can determine the optimized operating point of the device. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] The above-mentioned characteristics, features and advantages of the present invention and the manner in which they are achieved will become clearer and more clearly understandable in conjunction with the following description of embodiments, which are explained in more detail in conjunction with the accompanying drawings. Here, at least partially schematically:

[0051] Figure 1 An example of a conventional method for actuating a device is shown;

[0052] Figure 2 An example of a system for adapting a manufacturing process is shown; and

[0053] Figure 3 An example of a method for adapting a manufacturing process is shown.

[0054] Whenever this is expedient, the same reference symbols are used in the figures for identical or mutually corresponding elements of the invention. DETAILED DESCRIPTION

[0055] Figure 1 An example of a conventional method for starting a device is shown, the device being set up to perform a manufacturing process P. Such a manufacturing process P is usually associated with operating conditions B and operating parameters of the device, which can take different operating parameter values ​​a. Depending on the operating conditions B and the operating parameter values ​​a selected, for example, by a user, the product characteristics E of the process product produced by the manufacturing process P change.

[0056] In order to be able to produce a process product with a predetermined product property E* using the system or a production process P that can be performed using the system, the “correct” operating parameter value a must therefore be found. This process is called startup.

[0057] For this purpose, a system or a production process P can be mapped with a model M. In addition to the operating conditions B and the desired predefined product properties E*, which can be characterized by so-called target values, model parameters Z can also be input as input variables into the model M. The model parameters Z, which are sometimes also referred to as start-up parameters (IBS parameters), can be parameters that are intrinsic to the model M, but can also be used, alternatively or additionally, to set the sequence of calculations and / or to influence the algorithm for evaluating the model M.

[0058] The evaluation of the model M in method step V1 provides operating parameter values ​​a using these operating conditions B, the predefined product properties E* and the model parameters Z. In addition, such a model M can optionally also be used to determine a suitable sensitivity for regulating the system or production process P (not shown).

[0059] However, it has been shown in practice that the desired, predefined product characteristics E* cannot usually be achieved exactly by operating the system with the operating parameter values ​​a determined in this way. This is due in particular to the generally incomplete detection of all operating conditions B and the inherent limitations of the model M in describing reality (the model M can never describe reality 100% accurately).

[0060] Therefore, an attempt can be made in a further method step V2 to optimize the model parameters Z. The optimization takes place afterwards, ie after the production process P has been carried out by means of the system at the operating point defined by the operating parameter values ​​a provided by the model M. In this context, offline optimization is also referred to.

[0061] For this offline optimization, all relevant data D relating to the manufacturing process P are recorded while the manufacturing process P is being carried out with the aid of the device. With these data D, the model M can be evaluated once again—however, now the operating parameter values ​​a are not calculated, but the product characteristics E' to be expected. Comparing this product characteristic E' to be expected with the product characteristic E actually achieved by the manufacturing process P provides a quality G for the product characteristics that can be achieved by the manufacturing process P carried out with the operating parameter values ​​a determined by the model.

[0062] If the relationship between the model parameters Z and the quality G for the achievable product properties, i.e. the achievable target values, is known, the model parameters Z can be optimized (offline) using conventional optimization methods using the data D. The model parameters Z optimized in this way ultimately allow the system to be started up with the corresponding (optimized) operating parameter values ​​a.

[0063] However, the described method cannot be applied without a model M that can describe the manufacturing process P sufficiently well with reasonable effort. An example for such a manufacturing process P is the production of strips using a hot rolling mill. In this case, a typical predetermined product property E* is the rectangular shape of the strip end at a predetermined width at the end of the rolling process. In practice, the strip end shape is influenced by hydraulic adjustment of the upset in front of the roughing stand of the hot rolling mill. The goal at start-up is therefore to find the correct setting for the upset.

[0064] However, due to the material flows associated with rolling, the relationship between the upsetting process and the strip end shape obtained after rolling is extremely complex. Therefore, extremely high computing power is required, if it is possible at all, for simulating the forming of the strip end at a specific upsetting machine setting, for example by means of finite element calculations. However, due to calculation time and accuracy reasons, such a model M is not suitable for use at the start-up of the plant and is not suitable at all for continuous control of the upsetting process.

[0065] Therefore, in practice, parameterized curves are predefined as a function of the control parameters of the upset. At startup, the curve parameters must be manually adapted to the operating conditions of the system. This requires a lot of experience from the corresponding startup personnel and a high expenditure of time until the parameters for the entire production spectrum of the system are set.

[0066] Figure 2 A system 1 for adapting a manufacturing process is shown, which can be carried out with the aid of a device 10 and with which a process product 2 can be produced. The device 10 is currently designed as a hot rolling plant, which is set up for hot rolling a metal strip, which in this example is the process product 2. The hot rolling plant comprises a casting machine 11 for preferably continuously casting a metal strip, a separating device 12 for splitting the cast metal strip, a vertical rolling stand 13 for reducing the width of the split metal strip, a roughing stand 14 for a first thickness reduction of the split metal strip, and a finishing stand 15 for a second thickness reduction of the split metal strip. The hot rolling plant can have further components (not shown), such as one or more furnaces for (re)heating the cast or split metal strip, a descaling station for removing oxide layers, further separating devices for splitting the metal strip downstream of the roughing stand 14 and / or the finishing stand 15, and one or more coilers for winding the split metal strip into coils. A conveying device for conveying the process product 2 along the conveying direction T is not shown.

[0067] The system 1 further comprises a device 20 for determining an actual product property E of the process product 2 and an optimizer 30 for determining an optimized operating point based on the determined actual product property E and a predetermined product property E*.

[0068] The device 20 preferably has a sensor device, with which the technological product 2 can be detected in a sensory manner. For example, the device 20 can have one or more sensors in the form of a camera, an ultrasonic (distance) sensor or the like.

[0069] The actual product characteristic E of the process product 2 can be determined based on the sensor data generated when the process product 2 is detected in a sensory manner. In the variant shown, a device 20 can be set up for this purpose. Alternatively, however, the evaluation of the sensor data with respect to the actual product characteristic E of the process product 2 can also be carried out elsewhere, for example by the optimizer 30. In any case, the actually determined product characteristic E is provided at the optimizer 30.

[0070] In the example shown, the device 20 is arranged between the roughing stand 14 and the finishing stand 15 with respect to the conveying direction T. However, other arrangements of the device 20 are also conceivable, for example an arrangement downstream of the finishing stand 15 .

[0071] The optimizer 30 can be constructed in hardware and / or software technology. The optimizer 30 can be, for example, a computing device, such as a microprocessor and / or a storage device connected to the microprocessor signals or data. The optimizer 30 can have a dedicated hardware architecture designed with a view to determining an optimized operating point. The optimizer 30 can be constructed, for example, as an ASIC. Alternatively or additionally, the optimizer 30 can have, for example, one or more software modules that embody an optimization algorithm.

[0072] The optimizer 30 is expediently connected to the device 10 via an interface in order to bring about the operation of the device 10 at the (newly) determined optimized operating point, in particular in order to be able to send a correspondingly optimized operating parameter value a to the device 10 .

[0073] The vertical rolling stand 13 is designed to shape the side faces, sometimes also referred to as longitudinal sides, of the divided metal strip according to a control curve, so that the side faces do not extend in a straight line over the length of the divided metal strip. In other words, the vertical rolling stand 13 can influence the width of the divided metal strip. This predetermined extension of the side faces is intended to serve as compensation for the material flow produced during rolling in the roughing stand 14 and, if necessary, also in the finishing stand 15, so that a rectangular shape of the divided metal strip can be achieved after the roughing stand 14 or the finishing stand 15.

[0074] The control curve is suitably characterized by one or more operating parameters, the values ​​of which can be predetermined by the optimizer 30. Therefore, in the present example, the optimizer 30 is connected to the vertical rolling mill stand 13 by signals or data. After detecting each segmented metal strip by the device 20, in particular measuring the strip width in the start and end regions of each metal strip, the optimizer 30 can change at least one of the operating parameters according to the correspondingly determined actual product characteristics E, and thus change the control curve. The resulting shape of the segmented metal strip can be detected again by the device 20, and the optimizer 30 can once again change one of the operating parameters. For example, a substantially square shape of the segmented metal strip is thus formed after twenty such traversals.

[0075] Figure 3 An example of a method 100 for adapting a manufacturing process P is shown.

[0076] In method step S1, a product characteristic E* of a process product 2 that can be produced using the manufacturing process P is predefined. In a further method step S3, the system 10 is operated at a predefined operating start point X, which is set up for producing the process product 2 by carrying out the manufacturing process P. The operating start point X was previously predefined for this purpose in method step S2.

[0077] In a further method step S4, an actual product property E of the process product 2 produced by the manufacturing process P is determined. For this purpose, a device 20 can be provided for determining the actual product property E. For example, in method step S4, the produced process product 2 is scanned in a sensory manner and the actual product property E is derived from the sensor data generated thereby.

[0078] The determined actual product characteristic E is provided to the optimizer 30. In a further method step S5, the optimizer 30 determines an optimized operating point X* based on the predetermined product characteristic E* and the determined actual product characteristic E, for example based on the quality of the determined actual product characteristic E with respect to the predetermined product characteristic E*.

[0079] When determining the optimized operating point X*, the optimizer 30 expediently uses an optimization algorithm 31, such as a trust region method or a line search method. Starting from the current operating starting point X of the device 10, which is characterized by a specific combination of values ​​of the operating parameters A of the device 10, the optimization algorithm 31 can change the individual operating parameters A. The change is expediently made according to the optimization strategy on which the optimization algorithm 31 is based. For example, the optimization algorithm 31 can be set up to find the minimum value in the potential scenario from a set of all operating points that reflect the difference ΔE, E* between the actual product characteristic E and the predetermined product characteristic E*. In the present example, this is indicated in a one-dimensional manner by the gradient determination when changing the operating parameter A.

[0080] The optimized operating point X* determined in method step S5 is expediently predefined as the operating starting point X in method step S2 in a further iteration of method 100. That is, method 100 can be repeated starting from method step S2.

[0081] Unlike conventional optimization methods, no termination criteria are provided for the method 100. That is, the system 10 can also be operated in normal operation with the aid of the method 100. This makes it possible for the system 10 to be automatically transferred to the changed operating point even in the case of changed operating conditions that require a changed operating point for producing a process product 2 having a predetermined product characteristic E*.

[0082] However, in order to prevent the newly predetermined operating starting point X from fluctuating continuously in a region C around the optimal operating point X', a limit can be predetermined at which the plant 10 produces a process product 2 having a predetermined product characteristic E* or at least a product characteristic E* that is very close to the predetermined product characteristic. For example, if the difference between the predetermined product characteristic E* and the determined actual product characteristic E reaches or does not exceed a threshold value, the predetermined operating starting point X can be output as an optimized operating point X*. Thus, once the region C, e.g. Figure 3By removing the "bottom" of the minimum value shown in , a "re-adjustment" of the operating point of the device 10 can be avoided.

[0083] Although the present invention has been illustrated and described in more detail by means of preferred embodiments, the present invention is limited to the disclosed examples and other variants may be derived therefrom by a person skilled in the art without departing from the scope of protection of the present invention.

[0084] Reference numerals list

[0085] 1 System

[0086] 2 Craft products

[0087] 10. Equipment

[0088] 11 Casting Machine

[0089] 12 Separation device

[0090] 13 Vertical rolling mill stand

[0091] 14 Roughing mill stand

[0092] 15 Finishing mill stand

[0093] 20 Device for determining actual product characteristics

[0094] 30 Optimizer

[0095] 31 Optimization Algorithm

[0096] 100 Methods

[0097] S1-S5 Method Steps

[0098] V1 and V2 Method Steps

[0099] A. Operating parameters

[0100] a Operation parameter value

[0101] B. Operating conditions

[0102] E Product Features

[0103] E* Predetermined product characteristics

[0104] E' Expected product characteristics

[0105] Z model parameters

[0106] M Model

[0107] Manufacturing Process

[0108] D Data

[0109] G Quality

[0110] T Conveying direction

[0111] X Start point

[0112] X* Optimized operating point

[0113] X' Optimal operating point

[0114] C Zone

[0115] ΔE,E* difference.

Claims

1. A method (100) for adapting a manufacturing process (P), in particular a hot rolling process, comprising the following steps: - predetermining (S1) a product property (E*) of a process product (2) characterized by a target value; - predetermining (S2) an operating start point (X) of a plant (10), in particular a hot rolling plant, which is set up to carry out a manufacturing process (P) for producing the process product (2); - producing (S3) the process product (2) by carrying out the manufacturing process (P) with the aid of the device (10) starting from a predetermined operating starting point (X); - determining (S4) actual product characteristics (E) of the produced process product (2); - determining (S5) an optimized operating point (X*) of the device (10) based on the predetermined product characteristics (E*) and the determined actual product characteristics (E) by means of an optimizer (30) for iteratively solving the optimization problem using an optimization strategy; - repeating the method (100) with the predefined product characteristics (E*), wherein the respectively determined optimized operating point (X*) is predefined as the operating start point (X) during the repetition or is used as the basis for predefining the operating start point (X).

2. The method (100) according to claim 1, wherein - producing a plurality of process products (2) by carrying out the production process (P) at different operating points starting from a predetermined operating start (X) with the aid of the device (10), - for each produced process product (2), determining said actual product characteristics (E), and - determining the optimized operating point (X*) based on the predetermined product property (E*) and a plurality of determined actual product properties (E).

3. A method (100) according to claim 2, wherein the dispersion of actual product characteristics (E) of the process product (2) obtained when the manufacturing process (P) is performed multiple times in the same operating point of the device (2) is determined, and a measure of the determined dispersion is used as the basis for determining different operating points.

4. The method (100) according to claim 2 or 3, wherein different operating points around the predetermined operating starting point (X) are determined by respectively changing one of a plurality of operating parameters (A).

5. The method (100) according to any one of the preceding claims, wherein - producing a plurality of process products (2) by carrying out the manufacturing process (P) a plurality of times at the same operating point with the aid of the device (10), and - for each produced process product (2), determining said actual product characteristics (E), and - determining the optimized operating point (X*) based on the predetermined product property (E*) and a plurality of determined position parameters of the actual product property (E).

6. The method (100) according to claim 5, wherein the manufacturing process (P) is performed at least three times, preferably at least four times, in particular at least five times at the same operating point with the aid of the device (10).

7. A method (100) according to any of the preceding claims, wherein if the difference (ΔE, E*) between the predetermined product characteristic (E*) and the determined actual product characteristic (E) reaches or does not exceed a threshold value, the optimizer (30) determines the predetermined operating starting point (X) as the optimized operating point (X*).

8. A method (100) according to claim 7, wherein the number of times the manufacturing process (P) is performed at the same operating point with the aid of the device (10) is increased according to the difference (ΔE, E*) between the predetermined product characteristics (E*) and the determined actual product characteristics (E).

9. The method (100) according to any one of the preceding claims, wherein the optimized operating point (X*) is determined by the optimizer (30) by means of a trust region method and / or a line search method.

10. A method (100) according to any one of the preceding claims, wherein it is checked whether the operation of the device (10) meets at least one predetermined criterion when executing the manufacturing process, and an optimized operating point (X*) of the device (10) is determined based on the result of the check.

11. The method (100) according to any one of the preceding claims, wherein the optimizer (30) comprises an optimization algorithm (31), and wherein the optimization algorithm (31) has no stopping criterion.

12. A system (1) for adapting a manufacturing process (P), in particular a hot rolling process, comprising - a plant (10), in particular a hot rolling plant, for producing a process product (2) by carrying out a manufacturing process (P) with the aid of the plant (10) starting from a predetermined operating start (X); - means (20) for determining the actual product characteristics (E) of the produced process product (2); and - An optimizer (30) for iteratively solving an optimization problem using an optimization strategy, the optimizer being configured to determine an optimized operating point (X*) for the device (10) based on predetermined product characteristics (E*) of the process product (2) characterized by target values ​​and determined actual product characteristics (E), so that the determined optimized operating point (X*) can be predetermined as an operating starting point (X) when the manufacturing process (P) is re-executed or can be used as a basis for predetermining the operating starting point (X).

Images