EP4117836A1 - Verfahren zur reduzierung von prozessstörungen bei der herstellung eines walzproduktes - Google Patents
Verfahren zur reduzierung von prozessstörungen bei der herstellung eines walzproduktesInfo
- Publication number
- EP4117836A1 EP4117836A1 EP21711817.3A EP21711817A EP4117836A1 EP 4117836 A1 EP4117836 A1 EP 4117836A1 EP 21711817 A EP21711817 A EP 21711817A EP 4117836 A1 EP4117836 A1 EP 4117836A1
- Authority
- EP
- European Patent Office
- Prior art keywords
- rolled product
- rolling
- production
- target
- rolling train
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B23/00—Testing or monitoring of control systems or parts thereof
- G05B23/02—Electric testing or monitoring
- G05B23/0205—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults
- G05B23/0218—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterised by the fault detection method dealing with either existing or incipient faults
- G05B23/0243—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterised by the fault detection method dealing with either existing or incipient faults model based detection method, e.g. first-principles knowledge model
- G05B23/0254—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterised by the fault detection method dealing with either existing or incipient faults model based detection method, e.g. first-principles knowledge model based on a quantitative model, e.g. mathematical relationships between inputs and outputs; functions: observer, Kalman filter, residual calculation, Neural Networks
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B21—MECHANICAL METAL-WORKING WITHOUT ESSENTIALLY REMOVING MATERIAL; PUNCHING METAL
- B21B—ROLLING OF METAL
- B21B37/00—Control devices or methods specially adapted for metal-rolling mills or the work produced thereby
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B2219/00—Program-control systems
- G05B2219/30—Nc systems
- G05B2219/33—Director till display
- G05B2219/33296—ANN for diagnostic, monitoring
Definitions
- the invention relates to a method for reducing process disturbances in the manufacture of a rolled product.
- Process disruptions in the manufacture of a rolled product can have a number of causes.
- a process disruption is initially any deviation of an actual value from a target value. Not every process disturbance that occurs has a negative effect on the output of the rolling train. Yield is understood here in the sense of the invention as the quotient of the amount of rolling stock produced in a time interval corresponding to the specifications and the maximum amount to be produced in the rolling train in this time interval
- That Output according to the invention can therefore be reduced by qualitative deviations in the rolling stock or deviations in relation to the production output or throughput of the rolling train, with qualitative deviations likewise being described by deviations in the actual values from the target values for the rolled product.
- Rolling trains usually consist of a number of units, for example furnaces, roll stands or cooling sections, through which the rolled product passes during manufacture without interrupting the manufacturing process.
- a large number of target specifications for a control and regulation of the overall process must be given to a rolling train or the units.
- These target specifications of the units for the manufacturing process are based on the target specifications for the rolled product and the empirical knowledge of the Plant operator or process models for individual units with which the target specifications of the rolled product can be adhered to.
- the setpoints and / or the permissible control range of the unit must then be adapted on the basis of empirical values and / or process models. In this case, however, the adjustments are usually only related to the unit if thickness deviations are adjusted or optimized by adjusting the pass schedules or piercing temperatures. Interactions going beyond this are mostly not considered.
- process models are known which describe and predict relationships between target specifications or actual values and the effects on the rolled product for individual process variables and / or individual units. Process models are, for example, temperature models, microstructure models or also deformation models in the roll gap. These are known from the prior art.
- Some of these are able to determine target specifications across all units, independently of a system control, or even Calculate actual values in advance.
- the previous models only take into account a small number of input variables and (only) track a relevant actual value of the rolled product.
- the specialist is able to plan target specifications for the units in advance, analyze process errors afterwards and adapt individual target specifications.
- these models can also influence the process within a control or regulation system across all units. The more target specifications and process parameters a previously known process model takes into account, the less it can be integrated into a known control or regulation concept of a rolling mill, since previously known linear process models with increasing complexity require disproportionately high computing capacity.
- Process disruptions that arise from a complex interaction of deviations from target specifications in individual units of the rolling train can therefore only be analyzed and remedied by a specialist using the known models or based on experience after the production of the rolled product has been completed. This creates a certain time delay between the error detection and the adjustment of the target specifications. This can significantly reduce the output of a rolling train.
- the object of the invention is therefore to improve the control or regulation of a rolling train in such a way that these aggregate interactions between the setpoints, actual values and deviations between the values are taken into account in the control or regulation of the rolling train and thereby process disruptions are reduced.
- the invention is achieved with a method having the features according to claim 1.
- a process disruption affecting the output of the rolling train is caused by at least a deviation of an actual value from a target specification of the rolled product itself and / or a target specification of an aggregate of the rolling train during the manufacture of the Rolled product justified, whereby these deviating actual values are associated with more than 4 actual values or target values.
- Connected in the sense of the invention means that a fluctuation, in particular a deviation from a target specification, of this value leads directly or indirectly to a fluctuation of at least 4 further actual values and / or is caused.
- the rolled product is manufactured with the rolling train by at least one heating device, at least one rolling stand, at least one cooling section, with a higher-level control of the rolling train controlling or regulating the production on the basis of setpoints.
- the output, the actual values and the associated target specifications for the manufacture and the rolled product, in particular the dimensions of the units and / or process measurement values, are recorded at different points in the rolling train and the output, the actual values and the target specifications are stored for further use.
- the output, the actual values and the target specifications are available on a data memory for further evaluation and retrieval by the higher-level control or an external evaluation system.
- a model in particular a model connected to the control of the rolling train, determines a relationship for determining a malfunction indicator.
- the malfunction index is a in the sense of the invention Key figure for a combination of related setpoints, actual values and output that lead to a process disruption.
- a model in the sense of the invention is any mathematical relationship or algorithm with which such a characteristic number can be determined from this data. From the prior art, for example, statistical evaluation methods, in particular regression analyzes, are known as a mathematical method for this.
- the model takes into account the interactions between the target specifications and actual values over the entire production of the rolled product in the rolling train by creating relationships between stored target specifications and / or actual values with the individual process disruption and / or the setpoints and / or actual values without process disruption.
- the higher-level control continuously determines the malfunction indicator during the manufacture of the rolled product, in particular online, on the basis of actual values during the manufacture of a rolled product.
- the ongoing production of the rolled product is interrupted or the production of a subsequent rolled product is not started if the malfunction indicator reaches a previously determined critical value range. Any range of values of the malfunction index that leads to a deviation of an actual value of the rolled product from its target value can initially be viewed as critical.
- the higher-level control can, as an alternative to interrupting production, specify changed target specifications for the units of the rolling train for further production, the changed target specifications using further process models connected to the higher-level control of the rolling train, in particular temperature - and / or structural models, can be determined and / or optimized.
- the output of the rolling train is reduced, for example, by cobble formation of the rolled product.
- a cobble is understood to mean any incorrect rolling which leads to an interruption in the rolling process of a single rolled product or a series of identical rolled products. This can be, for example, the rise of a rolled product in the cooling section or between two roll stands.
- Cobble arise in particular from deviations in the process variables of rolling force, rolling temperature, degrees of deformation and / or the course of recrystallization of the rolled product in the course of the process.
- target specifications and / or actual values are preferably taken into account when determining the malfunction index, which influence the Cobble formation of a metallic strip. As a result, the number of target specifications and / or actual values to be taken into account can be reduced.
- the output of the rolling train is reduced, for example, by a geometric deviation of the rolled product.
- Geometry deviations arise, for example, from roll wear, uneven rolling forces and / or an uneven temperature distribution on the surface of the rolled product.
- Preference is given to avoiding the reduction in output Target specifications and / or actual values are taken into account when determining the malfunction index, which influence the geometric deviation of a metallic strip.
- the number of target specifications and / or actual values to be taken into account can be reduced.
- the output of the rolling train is reduced, for example, by a deviation in the product quality, preferably mechanical properties, surface quality and / or dimensions.
- target specifications and / or actual values are preferably taken into account when determining the malfunction index, which influence a deviation in product quality, preferably mechanical properties, surface quality and / or dimensions.
- the roll age, the temperature control in the rolling train and / or the actual chemical analysis of the rolled product in relation to the material can have an influence on the product quality. As a result, the number of target specifications and / or actual values to be taken into account can be reduced.
- the output of the rolling train is reduced, for example, through a reduction in production output, in particular through unplanned machine downtimes.
- the occurrence of unplanned downtimes can, for example, be influenced by the process parameters of cooling water quality, backup roll age, maintenance intervals of the units or parts of units, and / or the power consumption of drives.
- target specifications and / or actual values are preferably taken into account when determining the malfunction index, which influence a reduction in production output, in particular due to equipment downtimes.
- the number of target specifications and / or actual values to be taken into account can be reduced.
- Preference is given to more than 5, preferably more than 25, even more preferably more than 100 actual values in the entire rolling train when determining the Incident code used.
- the more actual values and target values are linked, the more interactions can be taken into account. This also reduces the risk of receiving different key figures due to target specifications and / or actual values that are not taken into account with the same input variables.
- the malfunction index is preferably determined in the form of a probability index.
- the same actual values of the production can lead to a reduction of the output if there are not recorded disturbance-relevant influencing parameters.
- This is preferably represented by a malfunction indicator in the form of a probability.
- the model preferably uses a self-learning algorithm, particularly preferably a KI algorithm, to determine the malfunction index.
- a self-learning algorithm particularly preferably a KI algorithm
- the model for determining the relationship recalculates, preferably cyclically, more preferably once per rolling campaign, the relationship on the basis of the actual values, target specifications and / or the output and uses the new relationship to calculate new malfunction indicators.
- a rolling campaign can be a series of different rolled products or a defined number of similar rolled products. The number should be chosen so that there are enough data records.
- at least one warning message is generated and sent to an operator by means of the control of the rolling train transmitted. This allows the quality of the relationship to be determined and monitored on a regular basis. Ideally, this is done by executing the model for determining the relationship on the process computer of the higher-level controller. As a result, such a value can be output and displayed quickly and easily within the control system of the rolling train.
- the self-learning behavior of the algorithm increases the accuracy of the determination of the malfunction indicator with each rolled product manufactured.
- intervention by a person skilled in the art to rectify a process disturbance is not necessary or is significantly less
- the self-learning algorithm for determining the characteristic number is preferably trained with existing data records from a rolling mill. This means that a functioning model can be set up before the rolling mill is put into operation. For this purpose, for example, data from units that are not currently producing in the network can be combined with one another.
- the determination of changed target process parameters for avoiding or leaving the critical value range of the malfunction indicator is preferably carried out by a self-learning algorithm, preferably a KI algorithm.
- a self-learning algorithm preferably a KI algorithm.
- the self-learning part then increases the accuracy of the target specification.
- the self-learning algorithm for determining the changed target specifications is preferably trained with existing data records from a rolling mill. This means that a functioning model can be set up before the rolling mill is put into operation. For this purpose, for example, data from aggregates can be combined that are not currently producing in a network.
- the invention is achieved by a method having the features of claim 12.
- Process disruptions in the rolling train of a casting and rolling plant with a superordinate control are reduced by a method according to one of claims 1 to 11 and the actual values and the target values of the units before the first tapping of the rolled product in a rolling stand are taken into account in the method.
- the casting and solidification process has a significant influence on the quality of the rolled product.
- a large number of process parameters are also available here in addition to those of the rolling train that were previously not taken into account, or only in a simplified manner, in the control or regulation of a rolling train.
- direct physical modeling of individual processes, such as solidification with simultaneous structure formation has not yet been possible.
- the effect-based modeling according to the invention without a mandatory physical connection is a possibility of avoiding process disturbances.
- the changed target specifications are preferably specified by the control for the units before the first piercing in a roll stand.
- the ongoing casting process can usually only be interrupted for ongoing production with difficulty. It is mostly possible to discharge a cast product with these deviations from a casting and rolling plant. This ejected cast products can, for example, be re-introduced after a review and approval. This will make that Bringing out the casting and rolling plant versus a failure of the finished one
- the malfunction indicator is determined by the model for each rolled product on the basis of planned actual values.
- the influence of the respective individual actual value is weighted in relation to the malfunction index and a necessary changed target specification.
- Target specifications through maintenance and / or replacement of whole or parts of the units involved in the production then take place before the start of the production of the rolled product.
- the risk of reducing the output due to the standstill of a unit can be determined and technical measures can be taken to react. This includes the replacement of parts of the units or even entire units of the rolling train that unexpectedly fail with the previous control method.
- rolled products can also be removed from the production sequence and replaced by others with non-critical malfunction indicators.
- the object of the invention is achieved by a method with the features of claim 13.
- the malfunction indicator determined by the model is determined for each rolled product on the basis of planned actual values and In the case of a critical malfunction index, the influence of the respective individual actual value is weighted in relation to the malfunction index and a necessary changed target specification.
- An optimization algorithm minimizes the malfunction indicator in the area of the system-technical possible target specifications by adapting the Target specifications.
- a minimization routine before the start of production can reduce the risk of a reduction in output, especially when assuming the probability of malfunctions as a malfunction indicator.
- the description of the invention is accompanied by three figures.
- wear data from rolls which are recorded cyclically, can be assigned to a series of rolled products in the same way.
- the key figure output a is initially defined as described above and, if necessary, can be weighted, for example, by quality factors f q and / or monetary value factors f m.
- quality factors f q and / or monetary value factors f m As a result, the economic success of the overall production can be better compared and evaluated with different materials of the rolled product. This results in the double assignment of the output a to the measured values MW and the target values SV (formula 1.1).
- an output probability a w or, reciprocally to a or a w, a malfunction indicator s and malfunction probability s w can be formed from the fixed output a.
- Measured values are, for example, rolling forces in the finishing train 8, a temperature profile of the strip / rolls across the width, a center line deviation, roll bending, roll age, furnace dwell times / temperatures and / or a material of the rolled product.
- Target specifications also include, for example, the target dimensions of the rolled product, the coil weight, the surface quality, the material properties and / or a structural composition of the rolled product.
- Measured values from a soft sensor can be, for example, the core temperature of a slab calculated on the basis of the surface temperature.
- This data so present is stored with the above-described assignment in a database with which, for example, a neural network 13 is trained.
- Training means that the existing data is shared and the neural network 13 with one half in a first step Establishes connection 15 and the prediction quality of the connection for the malfunction indicator is checked on the basis of the second half of the data.
- the relationship determined in this way between the malfunction indicator or malfunction probability and the actual values and target specifications is then used to generate a current production based on the actual values and target specifications measured during the manufacture of the rolled product
- the higher-level control 12 of the rolling train (2-11) begins to vary the setpoint values SV during ongoing production using process models, taking into account the technical possibilities of the units.
- the procedure takes place here in iteration steps.
- new target specifications are determined on the basis of a predetermined variance v and within the possibilities of the units and the target specifications for the rolled product in relation to the existing ones.
- these are used to calculate new actual values for the rolling train using process models.
- this actual data record is then evaluated with the aid of the relationship 15 with regard to its probability of malfunction. If the failure probability is within the desired range, production is continued with these target specifications.
- a termination criterion for example a number of iteration steps, should interrupt the iteration and then interrupt the ongoing production.
- the model 13 for determining the relationship 15 for the malfunction indicator also checks the quality of the relationship 15 cyclically. For this purpose, analogous to the training of the neural network, the data records stored during the production of a series of rolled products are combined in groups. With this new group of data sets, the model establishes a new relationship for the malfunction indicator. This relationship and the values determined for the malfunction index are compared with the previous relationship and the previous values for the malfunction index and evaluated. If a defined deviation is exceeded, a message is output to the operator. Furthermore, the previous connection can also be automatically replaced by the new connection.
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Mathematical Physics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Control Of Metal Rolling (AREA)
Abstract
Description
Claims
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| DE102020202955 | 2020-03-09 | ||
| PCT/EP2021/055912 WO2021180719A1 (de) | 2020-03-09 | 2021-03-09 | Verfahren zur reduzierung von prozessstörungen bei der herstellung eines walzproduktes |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| EP4117836A1 true EP4117836A1 (de) | 2023-01-18 |
Family
ID=74873727
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| EP21711817.3A Pending EP4117836A1 (de) | 2020-03-09 | 2021-03-09 | Verfahren zur reduzierung von prozessstörungen bei der herstellung eines walzproduktes |
Country Status (2)
| Country | Link |
|---|---|
| EP (1) | EP4117836A1 (de) |
| WO (1) | WO2021180719A1 (de) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118976785B (zh) * | 2024-08-27 | 2025-09-02 | 宝武集团鄂城钢铁有限公司 | 一种轧机轧制过程中出现推床故障的应急方法 |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP3293594A1 (de) * | 2016-09-13 | 2018-03-14 | Primetals Technologies Germany GmbH | Verwendung umfassender künstlicher intelligenz bei anlagen der grundstoffindustrie |
-
2021
- 2021-03-09 EP EP21711817.3A patent/EP4117836A1/de active Pending
- 2021-03-09 WO PCT/EP2021/055912 patent/WO2021180719A1/de not_active Ceased
Also Published As
| Publication number | Publication date |
|---|---|
| WO2021180719A1 (de) | 2021-09-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| EP0813701B1 (de) | Leitsystem für eine anlage der grundstoff- oder der verarbeitenden industrie o.ä. | |
| EP0813700B1 (de) | Intelligentes rechner-leitsystem | |
| DE4338615B4 (de) | Verfahren und Vorrichtung zur Führung eines Prozesses in einem geregelten System | |
| DE112019001512T5 (de) | EINSPRITZGIEßMASCHINENSYSTEM | |
| EP3358431A1 (de) | Ganzheitliche planung von produktions- und/oder wartungsplänen | |
| DE4008510A1 (de) | Regeleinheit mit optimal-entscheidungsmitteln | |
| EP0756219B1 (de) | Verfahren zur Überwachung von Produkteigenschaften und Verfahren zur Regelung eines Herstellungsprozesses | |
| EP4150419B1 (de) | System und verfahren zur steuerung einer aus mehreren anlagenteilen bestehenden produktionsanlage, insbesondere einer produktionsanlage zur erzeugung von industriegütern wie metallischem halbzeug | |
| DE4416317B4 (de) | Verfahren und Regeleinrichtung zur Regelung eines materialverarbeitenden Prozesses | |
| EP4146414B1 (de) | Verfahren zum steuern oder regeln der temperatur eines stahlbandes bei der warmumformung in einer warmbandstrasse | |
| EP3791971A1 (de) | Kaltwalzen eines walzguts in einer walzstrasse mit mehreren walzgerüsten | |
| DE102019214548A1 (de) | Steuervorrichtung und steuerverfahren | |
| WO2021180719A1 (de) | Verfahren zur reduzierung von prozessstörungen bei der herstellung eines walzproduktes | |
| EP0837825B1 (de) | Verfahren zur regelung einer bahnspannung | |
| EP4045986A1 (de) | Verfahren für eine modellbasierte bestimmung von modellparametern | |
| DE102004005011B4 (de) | Regelverfahren und Regler für ein Walzgerüst | |
| WO2022106707A1 (de) | VERFAHREN ZUR EINSTELLUNG DER EIGENSCHAFTEN EINES WARMBANDES MIT EINER BESTIMMTEN CHEMISCHEN ZUSAMMENSETZUNG IN EINER WARMBANDSTRAßE | |
| EP4045982A1 (de) | Kontrollverfahren für die kontrolle wenigstens eines teils eines produktionsprozesses einer folienextrusionsanlage | |
| DE3425129A1 (de) | Verfahren zur walzenkuehlung an einem kaltwalzwerk | |
| WO2023011937A1 (de) | Verfahren zum ermitteln eines steuerparameters zum steuern einer walzanlage | |
| WO2021110300A1 (de) | Verfahren und computersystem zur vorhersage einer schrumpfung eines gegossenen metallproduktes | |
| WO2025114222A1 (de) | Verfahren zum betreiben eines walzwerks sowie walzwerk | |
| DE4325074C2 (de) | Verfahren zur Regelung der Walzgutgeschwindigkeit in einer mehrgerüstigen kontinuierlichen Warmwalzstraße zur Gewährleistung minimaler Längszugkräfte | |
| DE102022204937A1 (de) | Anlagensteuersystem, steuerverfahren und programm für anlagen | |
| WO2025228641A1 (de) | Sollwert-ermittlung bei der metallbandproduktion |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: UNKNOWN |
|
| STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: THE INTERNATIONAL PUBLICATION HAS BEEN MADE |
|
| PUAI | Public reference made under article 153(3) epc to a published international application that has entered the european phase |
Free format text: ORIGINAL CODE: 0009012 |
|
| STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: REQUEST FOR EXAMINATION WAS MADE |
|
| 17P | Request for examination filed |
Effective date: 20221010 |
|
| AK | Designated contracting states |
Kind code of ref document: A1 Designated state(s): AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO RS SE SI SK SM TR |
|
| DAV | Request for validation of the european patent (deleted) | ||
| DAX | Request for extension of the european patent (deleted) | ||
| P01 | Opt-out of the competence of the unified patent court (upc) registered |
Effective date: 20230707 |
|
| RAP3 | Party data changed (applicant data changed or rights of an application transferred) |
Owner name: SMS GROUP GMBH |
|
| STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: EXAMINATION IS IN PROGRESS |
|
| 17Q | First examination report despatched |
Effective date: 20250910 |