Method for controlling a system for rolling metal strips

Abstract

The invention relates to a method for controlling a plant for rolling metal strips with one or more rolling stands arranged in a rolling mill to achieve wear-optimized operation of the rolling stands, comprising reading and collecting and / or summarizing state variables of at least one rolling stand and / or the metal strip from a control system for process automation of the plant or directly from a field level of the plant, wherein the state variables comprise physical measured values ​​and / or values ​​derived from measured values ​​in the form of actual values ​​of individual rolls of at least one rolling stand and / or the metal strip during operation of the rolling mill and / or predicted state variables of individual rolls of a rolling stand and / or the metal strip from at least one process model of the process automation, comprising an evaluation of the collected state variables in a rule-based manner or in the form of mass data-based pattern recognition.wherein the reading, collecting and / or summarizing and evaluating of the state variables is computer-implemented and carried out automatically, and the evaluation includes the detection of the current wear state of individual rolls, the calculation of a lifetime prediction for individual rolls and at least one automatically generated roll replacement forecast and / or adapted roll program planning based on the lifetime prediction.

Description

[0001] The invention relates to a method for controlling a plant for rolling metal strips with one or more rolling stands arranged in a rolling mill to achieve a wear-optimized operation of the rolling stands.

[0002] The service life of rolls in rolling stands of rolling mills used to produce metal flat products is fundamentally critical for the operation of the rolling mill. From time to time, work rolls in rolling stands must be replaced, renewed, or repaired. These are typically machined in a roll shop, where the worn roll contour is restored to its original condition by deposition welding and turning or grinding. Remachining of the rolls in the roll shop is usually carried out using CNC-controlled machining centers.

[0003] Rolls are replaced from time to time after visual or measured assessment of the roll wear, which requires a temporary shutdown of the rolling mill. Therefore, for an optimized rolling process, the longest possible service life of each roll is desirable to reduce the number of roll changes.

[0004] Automated methods for detecting the wear condition of individual rolls are known from the prior art in order to initiate appropriate maintenance and repair measures. US 2002 / 0116980 A1, for example, discloses a method for inspecting a rolling stand. This method comprises non-destructive monitoring of individual rolls of the rolling stand. In this method, the amplitude of a voltage signal is monitored, and the wear condition of the individual roll is deduced from the change in amplitude. Based on the detected signal, any defect in the roll or roll is classified by comparison with stored patterns of voltage curves. The method comprises calculating a maximum threshold value for the detected voltage signal and determining a repair measure depending on a difference between the threshold value and the detected voltage value.The method further comprises the recording of historical data relating to the individual rollers, whereby the historical data as well as the wear indication of the respective roller are coupled with a CNC control of the processing machines for the restoration of the roller.

[0005] The linking of a roll workshop with data from the rolling mill for controlling machine tools in the roll workshop is also known, for example, from JP H08 10810 A. JP 2001300609 A discloses a method for controlling a rolling mill, which also provides for a computer-assisted linking of the roll workshop's maintenance plan with the rolling program currently running on the rolling mill.

[0006] The invention is based on the object of providing a link between a rolling mill and a roll maintenance system such that the operator receives a lead time for commissioning and decommissioning and / or repairing rolls, enabling service life-optimized operation of the rolling mill. Overall, the invention aims to increase the available rolling time and the process stability of the rolling mill.

[0007] The object is achieved by providing a method having the features of claim 1. Advantageous embodiments of the invention emerge from the subclaims.

[0008] One aspect of the invention relates to methods for controlling a plant for rolling metal strips with one or more rolling stands arranged in a rolling mill to achieve a wear-optimised operation of the rolling stands, - comprising the reading out and collecting and / or summarising of process values, in particular in the form of state variables of at least one rolling stand and / or the metal strip from a control system for process automation of the plant or directly from a field level of the plant, - wherein the process values ​​comprise physical measured values ​​and / or values ​​derived from measured values ​​in the form of actual values ​​of individual rolls of at least one rolling stand and / or of the metal strip during operation of the rolling mill and / or predicted state variables of individual rolls of a rolling stand and / or of the metal strip from at least one process model of the process automation, - comprehensive evaluation of the collected process values ​​based on rules or in the form of mass data-based pattern recognition, - wherein the reading, collecting and / or summarizing and evaluating of the process values ​​is carried out in a computer-implemented and automated manner, and the evaluation comprises the detection of the current wear condition of individual rolls, the calculation of a service life prediction concerning individual rolls and at least one automatically generated roll change forecast and / or adapted rolling program planning based on the service life prediction.

[0009] One aspect of the invention relates to the individual observation of the respective rolls of the rolling stands and the evaluation of the process values ​​or state variables collected for the rolls of the rolling stands in order to achieve maximum utilization of the respective rolls before a roll change, to be able to detect process-critical roll wear in a timely manner, for example, to avoid malfunctions such as strip breaks due to the roll condition, and to be able to carry out a sensible combination of roll changes. The latter can be achieved, for example, by appropriate interventions in the process automation.

[0010] Rolls within the meaning of the invention are work rolls, intermediate rolls, and backup rolls of individual rolling stands. The rolling mill according to the invention can, for example, be a hot rolling mill with one or more roughing stands and one or more finishing stands.

[0011] The control system for process automation is preferably a multi-level control system in which a pre-calculation of the working variables of the rolling process is carried out in a Level-2 automation in the form of setting values ​​for a Level-1 automation.

[0012] The method according to the invention comprises reading out, collecting and / or summarizing process values, in particular physical state variables of at least one rolling stand and / or the metal strip from the control system for process automation of the plant or directly from a field level of the plant, ie from a level 1 automation or by measuring data from the process control level within which the direct control interventions in the individual units of the rolling plant take place.

[0013] The collected process values ​​can be evaluated based on fixed rules and / or through mass data-based pattern recognition. The latter is useful and advantageous for identifying system malfunction patterns that can be traced back to the maintenance status of individual rollers.

[0014] Preferably, process values ​​or state variables from a process model for roller wear, from historical data such as process disturbances, and from measured values ​​or sensors at the process level are combined or summarized. The correspondingly prepared state variables or process values ​​can be summarized in a database and evaluated and / or classified, for example, using AI and / or rules.

[0015] For the evaluation of the process values, the current wear condition of individual rolls is determined according to the invention and a service life forecast for individual rolls as well as a roll change forecast based on the service life forecast are created.

[0016] Process values ​​within the meaning of the invention are both current and historical physical state variables of individual rolls of the rolling mill, individual rolling stands and / or other units of the rolling mill as well as of the rolling stock, preferably with a temporal and / or spatial reference, namely both measured and derived and / or estimated process values.

[0017] The method according to the invention comprises the individual consideration of individual rolls and rolling stands.

[0018] Based on the roll change forecast, it is possible to optimize rolling programs or to operate the rolling mill in a wear-optimized manner in order to reduce the number of roll changes.

[0019] Roll change predictions are usefully based on both future and current events, such as process disturbances or state variables. Process disturbances can be caused, for example, by strip breaks. Events can also be process anomalies, such as instabilities in lead, rolling force, strip speed, or the like, which do not directly lead to a disturbance. Events can also be transitions between rolling programs.

[0020] The process values ​​are preferably selected from a group of process values ​​comprising roll identifiers and / or roll positions; identifiers of chocks and / or their position; installation position and / or position and / or type of rolls; identifiers and / or position of bearings; roll geometry; material and / or surface quality of the rolls; roll speed; roll angle; roll rise; setting position of the rolls; roll lead; actual service life of the rolls; strip tension; rolling force; strip speed; roll temperature; mass flow of the strip; thickness of the strip, in each case before and / or after a roll stand; bearing temperatures of the rolls; pressure, temperature, volume flow and composition of the rolling emulsion.

[0021] The predicted process values ​​are preferably selected from a group of process values ​​comprising the temperature of the rolls and / or a temperature profile of the rolls and / or the contour of the rolls resulting from wear.

[0022] The method preferably further comprises reading out historical data of process disturbances of the rolling mill and / or data from the control system, which are read out, collected and evaluated, wherein these data are correlated with wear and repair data of rolls from a roll repair to identify disturbance patterns.

[0023] The detection of disturbance patterns can be achieved using machine learning methods.

[0024] The evaluation of the process values ​​can be carried out using machine learning methods and / or depending on rule-based limit values ​​for the state variables.

[0025] The method preferably further comprises a roll change indicator for an operator in a process control station or an HMI (human-machine interface) and / or an automatic initiation of a forced stop of the rolling mill, ie without operator intervention, when a roll wear condition that is critical for process stability is reached.

[0026] In an advantageous variant of the method according to the invention, it is provided that values ​​of the service life prediction and / or current process values ​​of the rolling mill are fed back into the process automation of the rolling mill and the fed back values ​​are used as reference variables for calculating set values ​​in a Level 2 automation of the mill or are incorporated into the process automation in order to directly adapt the load and / or mass flow distribution in the rolling mill in order to adapt its operation with regard to the expected wear limits of certain individual rolls.

[0027] The fed-back values ​​of the lifetime prediction can be used as reference variables for the automatic control of the speed of the rolling mill and / or the roll rise of individual rolling stands and / or the load distribution between individual rolling stands.

[0028] In a particularly advantageous variant of the method according to the invention, it is provided that an automated repair and / or provision request is generated and transmitted to a roll workshop from the service life prediction and / or the roll change forecast.

[0029] Furthermore, interventions in the process automation are preferably initiated automatically and / or maintenance and / or repair requirements and / or maintenance and / or repair instructions and / or maintenance and / or repair recommendations are issued.

[0030] Machine learning methods or methods based on artificial intelligence include methods selected from a group comprising physical-mathematical models, neural networks, decision gates, if-then queries, self-learning algorithms, statistical models as well as state queries and adaptive models.

[0031] Current events and disturbances in the operation of the rolling mill are preferably assessed and / or classified by artificial intelligence

[0032] The evaluation of the collected process values ​​can be carried out using various evaluation algorithms, whereby a first evaluation algorithm can provide the following results: - Maximum utilization of the rollers before roller changes, - Timely detection of process-critical roll wear to avoid disruptions, such as strip breaks due to the roll condition, - A sensible combination of roll changes. For example, the rolls of one rolling stand may be at their wear limit, but due to the appropriately selected roll program, they can continue to be used until other rolls of this or other rolling stands reach their wear limit.

[0033] Input data of such a first evaluation algorithm can be: - The advance of the rolling stands, - The tensile and rolling force behavior of the rolling stands, - the temperature of the rolls, which can be derived from process models, for example, - The total running time of the rollers concerned, - The status of rolling emulsions, - the flatness data in the ongoing process

[0034] A second evaluation algorithm can: - select and suggest suitable rolls (material, grinding, diameter) for the upcoming rolling programs, - check whether already selected rolls are suitable (material, grinding, diameter), also in combination with each other with regard to a single rolling stand (for example, determine whether the diameter combination is suitable with regard to the running time of the selected rolling program), - check which rolls coming from a roll workshop have already had damage and whether the surface restoration after the last removal of the roll in question is suitable for the upcoming requirement, - provide information for the rolling programs to be selected as to whether the individual rolls have the corresponding roll contour or calculate the relevant contour for the work roll pairing, for example for CVC (Continuous Variable Crown) ground rolls, - regroup individual rolls for upcoming or ongoing rolling programs to optimize roll utilization, - regroup current or upcoming rolling programs for better use of the rolls, - Plan or suggest roll changes and maintenance to avoid unnecessary additional roll changes or to automatically and sensibly combine activities planned on other parts of the system.

[0035] Input data of the second evaluation algorithm can be: - Status of the roller emulsion, - rollers available in a roller warehouse, - The storage time of the available rollers.

[0036] The method preferably comprises summarizing state variables of the individual rollers in a database, which may, for example, include the following data: - Running time or mileage of individual rollers since the final grinding, - Identification feature of the individual rollers, - Diameter change due to grinding, - wear and tear, - thermal stress, - Contour of the finishing touches during installation and removal, - Roughness when removing the rollers (based on the change in lead due to the mileage), - process disruptions, a) Downtimes (min / max / average rolling force with respective duration) b) Strip breaks / strip paths (classified by maximum rolling force, duration until plant stop, area in which the break occurred, installation location of the roll) - Type and identification of the chocks used - Bearing temperatures and / or lubricant status of the roller bearings, - Times of installation / removal, grinding and assembly of the rollers.

[0037] The invention is explained below using a process sequence shown schematically in the drawing figures.

[0038] They show: Fig. 1a the first, upper part of a block diagram of the process flow according to the invention and Fig. 1b the second, lower part of the block diagram.

[0039] The process sequence shown in the illustration is shown as a block diagram which, for reasons of clarity, extends over two pages of the drawing, with the interfaces between the parts of the diagram being indicated.

[0040] In the block diagram, the process control of a rolling mill comprising one or more rolling stands is designated by reference numeral 1 at the top level. The process control transfers process values ​​in the form of physical state variables to a process model 11, which determines the wear of rolls of individual rolling stands, to a module 12 for recording process disturbances, and to an interface 13 for outputting process values ​​or state variables in the form of direct measured values ​​from the field level of the rolling process and / or derived state variables.

[0041] Information regarding the position, arrangement, diameter, roll profile, running performance, and thermal load of the rolls is fed into the process model 11 for wear of individual rolls. The module for outputting state variables 13 records the rolling force, system pressures, roll setting position, strip tension before and after individual roll stands, strip speed, strip thickness, measured roll eccentricity, bearing temperatures, oil flow rates, and other state variables of the rolling mill's components, particularly the individual roll stands.

[0042] The output data from the process model 11, the module for recording faults 12 and the module for outputting state variables 13 are summarized in a module for process value processing 14. The process values ​​and measurement data are summarized there in time- and / or length- and / or volume-based segments related to the rolling stock or the metal strip. In addition to the aforementioned process values ​​and measurement data, the module for process processing 14 can also contain other data such as the time until a planned or unscheduled plant stoppage, the rolling stand in which a fault may have occurred, the rolling material, standstill locations, alarm messages, etc. The process values ​​summarized in this way are stored in a database 20. The database 20 also contains data from a rolling program 30 and data from a rolling workshop 40.

[0043] Data concerning planned coils 31 and data concerning planned roll changes 32 are transferred from the rolling program 30 into the database 20.

[0044] The data transferred from the roll workshop 40 to the database 20 comprises measured values ​​41 from the grinding process or repair process of the roll workshop 40.

[0045] The data stored in the database is transferred to a first and second evaluation unit 51, 52, with the first evaluation unit 51 performing AI-based evaluation of mass data and detecting fault patterns. In the second evaluation unit 52, which is arranged in series and parallel with the first evaluation unit 51, the process values ​​are evaluated based on rules. The evaluated process values ​​can be fed back to the process control system 1, where they can directly trigger control interventions at the field level.

[0046] Furthermore, based on the evaluation of the data, various measures are initiated which, on the one hand, provide feedback to the rolling program 30, on the other hand, serve for analysis, and are finally fed back into a Level 2 process automation system 60 for calculating new setting values ​​and indirectly intervene in the process control system 1 via the process automation system 60. The measures include suggestions 53 for rescheduling the rolling program, a further evaluation 54 of the relationship between the condition of individual rolls and a process disturbance, and model-based operator guidance 55 which, for example, indicates a required roll change or shows the operator the remaining running time of the rolling program until a required roll change. This serves, for example, to prevent strip breakage and maintain maximum process quality.The model-based operator guidance 55 suggests appropriate measures for flatness, for example in the case of decreasing roll diameter or changing settling values.

[0047] The evaluation 54 of the relationship between the rolls' state variables and process disturbances can be used, for example, to generate feedback to the roll workshop or purchasing department. An analysis of process disturbances is designated 56, and the review of appropriate repair measures, which may be performed automatically, is designated 57.

[0048] Immediate interventions in the process automation 60 are carried out via a model-based automatic replanning 61 of the rolling program, which is model-based, an automatic adjustment 62 of the rolling strategy, which is also model-based and causes a forced stop of the rolling mill 63.

[0049] Furthermore, an automatic generation of a provision request 42 to the roll shop 40 is provided. The automatic adjustment of the rolling strategy includes automatic load distribution between individual roll stands and / or an automatic reduction of the rolling speeds. The forced stop 63 of the rolling mill can be provided, for example, if the operator ignores a displayed wheat change. List of reference symbols 1 Process control 11 Process model roll wear 12 Module for recording faults 13 Module for outputting process values 14 Module for process value processing 20 Database 30 rolling program 31 Dates of planned roll changes 40 roller workshop 41 Data from the loop process 42 Provisioning request 51 first evaluation unit - AI based 52 second evaluation unit - rule-based 53 Proposal for replanning 54 Evaluation of the relationship between wheat rotation process and disturbance 55 model-based operator guidance 56 Analysis of process disturbances 57 Examination of repair measures 60 Process automation 61 automatic rescheduling of the rolling program 62 automatic adjustment of the rolling strategy 63 Compulsory detention QUOTES CONTAINED IN THE DESCRIPTION

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

[0000] US 2002 / 0116980 A1

[0004] JP H08 10810 A

[0005] JP 2001300609 A

[0005]

Claims

[1] Method for controlling a plant for rolling metal strips with one or more rolling stands arranged in a rolling mill to achieve a wear-optimised operation of the rolling stands, - comprising the reading out and collecting and / or summarising of process values, in particular in the form of state variables of at least one rolling stand and / or the metal strip from a control system for process automation of the plant or directly from a field level of the plant, - wherein the process values comprise physical measured values and / or values derived from measured values in the form of actual values of individual rolls of at least one rolling stand and / or of the metal strip during operation of the rolling mill and / or predicted state variables of individual rolls of a rolling stand and / or of the metal strip from at least one process model of the process automation, - comprehensive evaluation of the collected state variables based on rules or in the form of mass data-based pattern recognition, - wherein the reading, collecting and / or summarizing and evaluating of the process values is carried out in a computer-implemented and automated manner, and the evaluation comprises the detection of the current wear condition of individual rolls, the calculation of a service life prediction concerning individual rolls and at least one automatically generated roll change forecast and / or adapted rolling program planning based on the service life prediction. [2] Method according to claim 1, characterized by that the roll change forecast includes both future and current events. [3] Method according to one of claims 1 or 2, characterized bythat the process values are selected from a group of process values comprising roll identifiers and / or roll positions; identifiers of chocks and / or their position; installation position and / or position and / or type of rolls; identifiers and / or position of bearings; roll geometry; material and / or surface quality of the rolls; roll speed; roll angle; roll rise; setting position of the rolls; roll lead; actual service life of the rolls; strip tension; rolling force; strip speed; strip temperature; mass flow of the strip; thickness of the strip, in each case before and / or after a roll stand; bearing temperatures of the rolls; measured or calculated flatness values of the strip; pressure, temperature, volume flow and composition of the rolling emulsion. [4] Method according to one of claims 1 to 3, characterized bythat the predicted process values are selected from a group of process values comprising the temperature of the rolls and / or a temperature profile of the rolls and / or the contour of the rolls resulting from wear. [5] Method according to one of claims 1 to 4, characterized by that historical data of process disturbances of the rolling mill and / or data from the control system are read out, collected and evaluated, whereby this data is correlated with wear and repair data of rolls from a roll repair in order to identify fault patterns. [6] Method according to one of claims 1 to 5, characterized by that the detection of disturbance patterns is carried out using machine learning methods [7] Method according to one of claims 1 to 6, characterized by that the evaluation of the process values in their development (trend) and / or absolute change is carried out depending on rule-based limit values. [8] Method according to one of claims 1 to 7, comprising a roll change indication for an operator in a process control station and / or the initiation of a forced stop of the rolling mill without operator intervention when a wear condition of a roll that is critical for the process stability is reached. [9] Method according to one of claims 1 to 8, characterized by that a feedback of values from the service life prediction and / or current process values of the plant is provided into the process automation of the rolling mill and that the returned values are used as reference variables for calculating setting values in a Level 2 automation of the plant or are incorporated into the process automation in order to directly adapt the load and / or mass flow distribution in the rolling mill. [10] Method according to claim 9, characterized bythat the values fed back from the service life prediction are used as reference variables for the automatic control of the speed of the rolling mill and / or the roll rise of individual rolling stands and / or the load distribution between the individual rolling stands. [11] Method according to one of claims 1 to 10, characterized by that an automated repair and / or provision request is generated and transmitted to a roll workshop from the service life prediction and / or the roll change forecast. [12] Method according to one of claims 1 to 11, characterized by that current events and malfunctions in the operation of the rolling mill are assessed and classified by artificial intelligence and automatically initiate operator interventions in the process automation and / or issue maintenance and / or repair requirements and / or instructions and / or recommendations. [13] Method according to one of claims 6, 11 or 12, characterized by that methods of machine learning or artificial intelligence include methods selected from a group comprising physical-mathematical models, neural networks, decision gates, if-then queries, self-learning algorithms, statistical models, status queries, adaptive models.

Images