Classifications

• • B—PERFORMING OPERATIONS; TRANSPORTING
  • B22—CASTING; POWDER METALLURGY
  • B22D—CASTING OF METALS; CASTING OF OTHER SUBSTANCES BY THE SAME PROCESSES OR DEVICES
  • B22D11/00—Continuous casting of metals, i.e. casting in indefinite lengths
  • B22D11/16—Controlling or regulating processes or operations

Definitions

• the invention relates to a control system for a plant in the basic material or processing industry or the like, e.g. for a metallurgical plant, for example for the production of strips of steel or non-ferrous metals, the control system using computer technology, based on input of prior knowledge, the condition of the plant and details of a manufacturing process running in the plant, e.g. a continuous casting process for strips, which recognizes itself automatically and gives instructions appropriate to the situation in order to achieve a safe production success.
• a control system for a plant in the basic material or processing industry or the like e.g. for a metallurgical plant, for example for the production of strips of steel or non-ferrous metals
• the control system using computer technology, based on input of prior knowledge, the condition of the plant and details of a manufacturing process running in the plant, e.g. a continuous casting process for strips, which recognizes itself automatically and gives instructions appropriate to the situation in order to achieve a safe production success.
• Expert systems as so-called intelligent systems, with which the quality of the manufactured products is also to be improved with regard to quality features that are difficult to control from a control point of view, are also known for systems in the basic material industry, e.g. from the article "Process optimization for maximum availability in continuous casting", published in the journal "Metallurgical Plant and Technologie International 5/1994". Expert systems of this type, which can certainly improve the success of production, do not, however, overcome the basic weaknesses of conventional control. These are particularly visible when it comes to processes that are more suitable (such as in the raw materials industry) due to the lack
• the task is solved by an actually intelligently designed control system, which, based on the input of prior knowledge, automatically gives appropriate instructions for a safe and as good as possible (optimal) process control. It is therefore a fully developed technical intelligence that, surprisingly, can already be implemented for the process control systems of large systems using the computing resources available today.
• control system is designed to optimize the situation-specific instructions step by step in terms of computing. This results in a further increase in the intelligent behavior, which leads to a quality of the process control which cannot be achieved by human operating personnel or at least not in the short time that can be achieved by computer technology. .
• the prior knowledge entered that is to say the process knowledge predefined by humans, preferably automatically, continuously improved by knowledge obtained internally computationally in the process during production, for example in different operating points, and this self-generated process knowledge in one, in particular, constantly updated data storage is adopted as new prior knowledge. So one becomes very advantageous constantly improved basis for further adaptation or optimization of the process.
• the knowledge gained is not only limited to more precise parameters etc., but also includes the principles of the algorithms used, etc. in particular.
• the control system has a basic functional system for the system components, which contains the instructions from the computational, e.g. knowledge gained from a process model, preferably an overall process model, is reliably implemented in the plant management.
• a safe basic function system which is preferably designed as a basic automation system that makes the system components safe to work on their own or in summary, with a static process model adapted to the situation, results in an implementation that is related to the safety of the process - Management of the conventional design of a control system is at least equivalent, in terms of the cost / benefit ratio and the reliably achievable process result, but is superior.
• the basic automation system is advantageously designed as an autonomous subsystem that guarantees a safe state of the system or the system components and the process state, for example as a hazard state relapse system, which, if necessary, recognized as safe instead of the computationally generated instructions , operating values stored in the data memory can be used. This enables safe, albeit suboptimal, operation of the system even in the event of a failure or malfunction of the intelligent computer part.
• the basic function system advantageously also has start-up and run-up routines that can be entered manually or automatically, as well as suboptimal normal operating routines in which individual, otherwise computationally determined, instructions can be replaced by constant, reliable specifications.
• Such a configuration of the basic function system is particularly advantageous for the commissioning phases, for operation with a sudden change in requirements, etc.
• For the, albeit sub-optimal, function of the intelligent computer part it is not always necessary for all model parts to be available in a specifically adapted form. Operation with an only partially elaborated and / or adapted overall process model is also advantageously possible.
• the process model itself in the form of an overall process model, is particularly modular and describes the behavior between the process input variables and the manipulated variables and the process output variables, for example quality parameters of the product produced.
• the modularity allows a particularly advantageous design and processing of the overall process model, since it is possible to start from individual, easily visible, partial models.
• the process model is advantageously based as far as possible on mathematical forms of description. Where such mathematical forms of description are not possible, use is made, for example, of linguistically formulated model parts, which can be implemented, for example, by fuzzy systems, neuro-fuzzy systems, expert systems or the like.
• self-learning systems for example neural networks, are used. This means that all production plants, regardless of how large they are designed or how they are designed, have the option of creating an overall process model.
• the process model is advantageously continuously adapted and further improved to the process on the basis of process data collected on the system and archived in a process database, this advantageously using adaptive methods, learning methods, e.g. through a backpropagation learning process or a selection process for different sub-models, such as neural networks or their parts. This results in a largely self-taught model that can be adapted or improved on-line or off-line.
• the adjustable process variables are optimized by the optimizer on the process model in such a way that the model output variables, which are in particular quality parameters of the product, are matched as well as possible with predetermined, e.g. the target,
• the off-line optimization can be carried out on a separate computing unit parallel to the model adaptation, as well as during breaks, for example at weekends or during repair downtimes on the computer, which outputs the reference values of the basic function system during operation.
• the optimization is advantageously carried out using known optimization methods, in particular using genetic algorithms.
• the selection of the optimization process depends on the situation and problem. It can be done by a specification, e.g. on the basis of an analysis of the course of the process, or by means of a computational selection from an optimization method collection. A simple "trial and error” procedure can be used for this, but in order to reduce the computational effort, it is advisable to support the "trial and error” procedure by means of convergence criteria, methods of pattern recognition in the course of error acceptance, etc.
• the respective start values for an optimization are advantageously determined on the basis of the suboptimal operating data archived in a process data memory. This reduces the optimization effort, since the optimization calculation starts with pre-optimized values when it uses intermediate values that are recognized as safe as start values.
• the overall system is improved at least in three stages.
• the lowest level is the continuous improvement of the existing process knowledge, stored in the data memory, for example in the form of suboptimal, safe operating points, which are continuously brought up to a better adapted level of knowledge, and which is then again assumed.
• the second stage essentially forms the model adaptation, which adapts the model behavior to the process behavior as well as possible.
• the third stage is a continuous improvement of the situation-specific instructions by the process optimizer, e.g. about evolutionary strategies, genetic algorithms etc. These strategies require a large computing time and preferably run off-line.
• the system improvement is advantageously also continuously supported by external simulation calculations, model tests, possibly also by tests on the production plant with new aids, etc.
• control system according to the invention is described below by way of example on the basis of a strip casting installation for steel. Further, also inventive details and advantages result from the drawing and the description of the drawing as well as from the subclaims.
• FIG. 5 essential components of the process model and their rough link structure
• FIG. 1, 1 denotes the casting rolls of a two-roll casting device
• the material, between the casting rolls 1, Liquid steel, for example, is poured in from the ladle 4 via the tundish 5 and a dip tube 6 and solidifies into a strip 3, which can be further deformed in a rolling mill symbolized by the circles 2 with arrows.
• the downstream rolling plant can also be simply replaced by conveyor rollers, a reel or the like if the rolling is not to take place immediately after the casting.
• the design of the overall system is carried out in a requirement-specific manner. It is also possible to design the system downstream of the casting device as a hot-cold rolling mill and is recommended at very high casting speeds, since the cold rolling section of the system can then also be sufficiently utilized.
• the casting roll device preferably has an electrodynamic system 8, 9, also shown only symbolically, and an induction heating system 10.
• the electrodynamic system part 8 advantageously serves to relieve the weight, the cast band 3, which is still very soft and therefore at risk of constriction, and the electrodynamic system part 9 guides the band 3, while the induction heating system 10 maintains a predetermined temperature profiles over the bandwidth is incumbent if, for example this is followed by direct post-forming in a rolling mill. This is particularly advantageous for crack-sensitive steels.
• the cast strip 3 is checked for cracks by a camera 73, it being possible advantageously to use the fact that the crack pattern in the scale is influenced by cracks in the base material.
• the formation of a measured variable is advantageously carried out by a neuro-fuzzy system.
• the casting roller device Due to the data transmissions 1, 11 and VI, which are symbolized by arrows, the casting roller device, in which the solidification shells of the steel formed on the two casting rollers 1 are not only combined, but also formed in a rolling manner, with the intelligent one Part of the control system connected.
• FIG. 2 shows the structure of the intelligent part of the control system. This essentially consists of the parts of process optimizer 15, model 20, model adaptation 16 and data storage device 17. These parts of the control system work together in such a way that the setpoint output 13 provides the best possible, situation-specific instructions via data line V for process control become. These instructions are then converted into setpoints for the basic automation. The task and function of the individual parts are described below.
• the model output variables are typical quality parameters of the product.
• the model description y, (" ⁇ -...,” ,, ...- v ,, - .., v ,, ...)
• the process behavior generally does not exactly understand why y, - and y, differ more or less from one another.
• the manipulated variables w and the non-influenceable manipulated variables v are transmitted via the data lines I and II.
• the model adaptation 16 has the task of improving the model so that the model behavior corresponds as well as possible to the process behavior. This can be done online, at least for model parts, by adapting or tracking these model parts on the basis of continuously acquired process data.
• the adaptation can also be carried out off-line at specific times. This is done on the basis of a number m of process states representing the process (“, *, v *, y *), which are stored in the data memory 17.
• the index k quantifies the respective process status.
• model parameters or the model structure are varied so that ⁇ becomes as small as possible.
• the process optimizer has the task of using an optimization method and the process model to find manipulated variables u t which lead to the best possible process behavior.
• the process optimizer works off-line at specific times, for example, which can be predetermined manually, as follows:
• the non-influenceable manipulated variables v for which the optimization is to take place - for example the current ones - are kept constant and supplied to the model via the data line II.
• the process optimizer is then connected to the model by means of switch 18. It gives control values u t on the Model.
• the initial values y i are determined via the model. These are compared with target output values y target l , and it becomes the error
• the error E should be minimized.
• the process optimizer varies the manipulated variables u t in an iterative loop, each of which contains the calculation of v and E and the new selection of u t , until the error cannot be reduced further or this optimization is terminated.
• genetic algorithms, hill-clibing methods, etc. can be used as optimization methods.
• the optimal manipulated variables u opll obtained in this way which are the result of the above-mentioned minimization, are then transferred as setpoints to the basic function system via the setpoint specification and the data line V.
• the main task of the data memory is to archive representative process states (w.-v ,,; - /). In doing so, he replaces old process data again and again with new ones, in order to enable a current, albeit selective, process description based on this data.
• the data memory then supplies the model adaptation, as described above. On the other hand, it also supplies start values w for the process optimizer.
• Start values are selected, for example, in such a way that the output values y belonging to these start values correspond as closely as possible to the target values y target l .
• model 20 and process optimizer 15 which uses, for example, genetic algorithms for, for example, evolutionary, model improvement, preferably works off-line because, because of the grain complexity of a plant control model with its many possible configurations, the computing time of an evolutionary optimization process becomes comparatively long. Even with good optimization strategies, which are selected, for example, on the basis of an analysis of the probable model behavior, many optimization processes can be calculated until a significant model improvement is achieved.
• Workstations are used as computers for process optimization and parameter adaptation, e.g. from the Sun company.
• computers working in parallel are advantageously used. This applies in particular if the model can be divided into groups of Mode11 modules, which can be partially optimized depending on one another.
• FIG. 3 shows more precise details of the program structure with which the optimization is terminated and the new setpoint output is started.
• 58 designates an error function to be selected in each case, into which the errors (setpoint deviations) determined flow.
• the error function fulfills the termination criteria of the optimization. If this is the case, further optimized control and regulating variables are output.
• start values continuously arrive from the data memory in the start value specification 59, from which in search steps in 60, not from the optimizer, but from the data memory, for example with the aid of fuzzy interpolation, control and regulating parameters for a sub-optional process control be won.
• Switching takes place after reaching the predetermined quality factor, which is adapted to the respective control system knowledge.
• the minimization which can never be absolute, is stopped when the specified quality factor is reached.
• the model is also advantageous if it is connected to the process, i.e. Switch 1 is closed, an alarm signal is also generated which signals that critical operating states have been reached.
• Switch 1 is closed, an alarm signal is also generated which signals that critical operating states have been reached.
• FIG. 4 which shows the structure of a model adaptation by means of an optimization algorithm
• data from the start value specification 61 arrive in a search step unit 62 and are passed on from there to the model 63 as model parameters.
• the model 63 forms, together with the data memory 64, a parameter improvement loop which compares the values formed and stored in a known manner in 65.
• the comparison values are fed to the error function 67, which forwards their values to the termination criteria unit 66. If the termination criteria are met, the model is no longer improved and the existing values are used. Otherwise the optimization is continued with further search steps and the intermediate values in the data memory.
• FIG. 5 which shows the essential partial models of the overall process model of the exemplary embodiment, 46 denotes this
• Input model in which the external influences, such as the influences from the quality of the material used, are summarized.
• the quality of the steel used results e.g. the liquidus value, the solidus value, and other variables that characterize the casting behavior.
• 47 denotes the tundish model in which e.g. the steel volume of the tundish, the dip tube position or the like, the stopper position and the steel outflow temperature.
• the input models 46 and 47 are combined in the sub-model 56, which reflects the status of the material supplied.
• Such partial models can advantageously be parallel to other partial models, such as the casting area model, the rolling area model or the like. be optimized.
• the input model 48 contains the influences that affect the solidification, e.g. the casting roll cooling, the infrared heating etc.
• the input model 49 contains the values that are necessary for the heat balance, such as the steel casting roll temperature difference, the influence of lubricant as a function of the amount of lubricant, the rate of crystal formation of the respective steel type and e.g. the roller surface condition.
• the input model 50 contains e.g. the influences of the casting level characteristics, such as the level of the casting level, the thickness of the slag layer and the radiation coefficient.
• the input models 48, 49 and 50 are combined to form a partial model 54, which represents the status of the casting area. This
• Model area summary is generally advantageous for production areas, since it simplifies and improves the overall model optimization.
• the sub-models are still partially dependent on one another, for example the input models 49 (input model heat balance) and 50 to a considerable extent (Initial model of the mold level characteristic). Secondary dependencies are not shown for simplification.
• the sub-model 51 contains all influences on the solidification front, i.e. to the area where the metal shells solidified on the two cooling rolls meet. Essentially, these influences are the forming work performed by the casting rolls, the vibration width of the casting rolls or the emerging strip, the side gap sealing influences and the degree of effort of the overall system, this is e.g. a fuzzy model.
• the partial model 52 reproduces the exit values, e.g. the quality of the strip, the outlet temperature and distribution, but also the tendency to stick and the condition of the scale formed.
• the input model 53 and the input model 74 which relate to the temperature profile transverse to the strip and to the surface condition of the strip, also go into the partial part 52.
• the rolling mill part models 54 are also included in this special process model, since the product formation after the exit from the roll stands is the decisive criterion.
• the partial models are combined to form the product training model 57, which summarizes the thickness profile of the strip formed, the strip thickness, a possible defect pattern, the grain structure of the strip, the surface structure, etc.
• the surface structure and in particular the grain structure of the strip can only be determined with a considerable time delay. It is therefore advantageous here to work with submodels based on neural networks for the qualitative and quantitative determination of influencing variables.
• the above illustration shows the particular advantage that results from the design of the model in module form, since in particular the parts of a complex overall process model can be processed in parallel. This is particularly advantageous for the commissioning period of a system, in which the input and partial models are adapted to the actual conditions, linked together, etc.
• 68 denotes the process data archive, 69 the model parameter memory part, 70 the part with the start values for the optimizer and 71 the memory part for the safe operating points.
• the respective model formation is also stored in 68.
• the basic automation which with its controls, controls, interlocks etc., is an indispensable part of the
• Control system because it Guaranteeing the safe functioning of the system even in the event of a malfunction of the model part of the control system operating according to the invention must perform a multitude of functions.
• 21 means in the exemplary embodiment the mass flow control via the individual speed controller, 22 the control of the tundish heating, 23 the mold level control, 24 the tundish outflow control and 25 the heating power of the infrared or above screen 7 for maintaining the operating temperature of the casting rolls.
• 26 means the regulation of the lubricant additions, for example in the form of loose casting powder or of casting powder paste applied to the casting rolls, 27 the cooling water quantity regulation, 28 possibly the roll oscillation regulation, 29 the electric drive regulation and 30 the roll gap setting.
• 31 means the roll speed control and 32 possibly the control of the roll torque, 33 the setting of the cleaning system, consisting for example of a brush and a scraper for the casting rolls and 34 the control of the electrodynamic System for balancing the belt weight and 35 the regulation of the vibration width of the cast belt.
• 36 means the regulation of the individual parts of an electrodynamic system for side gap sealing and 37 the regulation of the heating for the side walls of the space between the casting rolls.
• 38 means the temperature profile control of the induction heating system 10.
• 39 and indicated further control units relate to controls of the downstream deformation units, for example roll stands, the train between these roll stands etc.
• the timing control 45 acts on the above actuators, controllers etc. which coordinates the manipulated variable outputs etc. in time.
• auxiliary controls and the interlocks are summarized, for example 41 means the automatic start-up, 42 the automatic switch-off, 43 and 44 interlocks which, for example, prevent molten steel from flowing before the casting roller pair and the forming rollers are operational , etc.
• 41 means the automatic start-up
• 42 the automatic switch-off
• 43 and 44 interlocks which, for example, prevent molten steel from flowing before the casting roller pair and the forming rollers are operational , etc.
• the strip edge separation that may be required, for example by laser, for influencing the scale formation, for example by
• the casting and rolling process consists of a number of sub-processes, the formation and influences of which are decisive for the end product.
• the properties of the end product for example its thickness, its thickness profile and its surface formation, can be influenced and optimized by a number of adjustable process variables, such as, for example Casting roll gap, the casting roll profile, the casting level, etc., which in turn influence the position of the union zone of the solidified metal shells deposited on the casting rolls.
• an overall process model is advantageously created according to the invention, which describes the process behavior.
• the influencing variables with which the process is influenced can be adapted and optimized step by step in accordance with the process conditions.
• the intelligent, self-improving part of the control system is composed of three essential elements: the process model, the model adaptation and the process optimizer.
• the process model consists of subsystems (modules), which will be of different types depending on the process knowledge. Knowing the physical relationships, classic, physical-mathematical models can be created. If, on the other hand, one only has experience or estimates, then fuzzy or neuro-fuzzy systems are used. If one knows little or nothing about the process behavior, such as crack formation and surface formation, neural networks are used, at least initially, for process formation.
• the model describes the relationship between the process variables, such as the height of the mold level in the selected example, the condition values and the quality of the cast material, the setting values of the casting rolls, etc. and the quality parameters of the Band, for example the thickness, the profile and the surface training.
• model adaptation which is based on data from past process states.
• model adaptation provides the model parameters above. in such a way that the model behavior corresponds as closely as possible to that of the process.
• the models themselves are optimized in a changing manner, e.g. through genetic algorithms, combinatorial evolution etc. Appropriate optimization strategies are known, e.g. from Ulrieh Hoffmann, Hanns Hofmann “Introduction to Optimization", Verlag Chemie GmbH, 1971 Weinheim / Bergstrasse; H.P.
• control structure according to the invention with the procedure according to the invention described above leaves the previous structure of a control system.
• level I a basic automation
• level II a basic automation
• intelligent self-optimization it ensures better and better process results due to the process results already achieved.
• Individual feedback control loops can be omitted.
• Quality control sensors are only for checking the process results necessary.
• the control system according to the invention therefore only has two essential levels, of which the intelligent level requires no visualization except for programming. For control purposes, however, the elements of the basic automation can be visualized in a known manner.

Landscapes

• Engineering & Computer Science (AREA)
• Mechanical Engineering (AREA)
• Feedback Control In General (AREA)
• Testing And Monitoring For Control Systems (AREA)
• Organic Low-Molecular-Weight Compounds And Preparation Thereof (AREA)
• General Factory Administration (AREA)
• Control Of Metal Rolling (AREA)

Applications Claiming Priority (3)

Publications (2)

Family

ID=7756201

Family Applications (1)

Country Status (6)

Families Citing this family (60)

Family Cites Families (12)

• 1995
  • 1995-03-09 DE DE19508476A patent/DE19508476A1/de not_active Withdrawn
  • 1995-06-05 US US08/463,446 patent/US5727127A/en not_active Expired - Lifetime
• 1996
  • 1996-03-06 DE DE59603352T patent/DE59603352D1/de not_active Revoked
  • 1996-03-06 EP EP96905686A patent/EP0813701B1/fr not_active Revoked
  • 1996-03-06 CN CNB961929588A patent/CN1244032C/zh not_active Expired - Fee Related
  • 1996-03-06 AT AT96905686T patent/ATE185626T1/de not_active IP Right Cessation
  • 1996-03-06 WO PCT/DE1996/000397 patent/WO1996028772A1/fr not_active Application Discontinuation

Non-Patent Citations (1)

Also Published As

Similar Documents

Legal Events