EP2566633A2

EP2566633A2 - Operating method for a production line with prediction of the command speed

Info

Publication number: EP2566633A2
Application number: EP11710447A
Authority: EP
Inventors: Klaus Weinzierl
Original assignee: Siemens AG
Current assignee: Primetals Technologies Germany GmbH
Priority date: 2010-05-06
Filing date: 2011-03-09
Publication date: 2013-03-13
Anticipated expiration: 2031-03-09
Also published as: US20130054003A1; CN102939173B; EP2566633B1; WO2011138067A2; RU2545872C2; CN102939173A; BR112012028373A2; EP2386365A1; WO2011138067A3; US9630227B2; RU2012152449A

Abstract

At the latest at an instant, at which a first belt point (12) of a belt (2) is still situated in front of a production line (1), in each case an actual value (G) and a setpoint value (G*) are known to a control computer (8) for the first and a number of second and third belt points (12, 13, 13') of the belt (2). For each belt point (12, 13, 13'), the actual value (G) is characteristic of the actual energy content which the respective belt point (12, 13, 13') has at a location (xE) in front of the production line (1). For each belt point (12, 13, 13'), the setpoint value (G*) is characteristic of the setpoint energy content which the respective belt point (12, 13, 13') has at a location (xA) behind the production line (1). The second belt points (13) run after the first belt point (12) into the production line (1), and the third belt points (13') run in in front of the first belt point (12). Before the first belt point (12) runs into the production line (1), the control computer (8) determines in each case one command variable (L*) for the first belt point (12) and at least part of the second belt points (13) using a respective determining regulation. Using the respective command variable (L*), the control computer (8) determines in each case one command speed (vL) and operates the production line (1) at the respective command speed (vL) at the instant when the respective belt point (12, 13) runs into the production line (1). For the respective command variable (L*), its determining regulation involves the actual and setpoint values (G, G*) of the belt point (12, 13) which runs in each case into the production line (1) at this instant and the actual and setpoint values (G, G*) of at least one belt point (12, 13, 13') which has already entered the production line (1) at this instant.

Description

description

Operating method for a finishing train with prediction of the guide speed

The present invention relates to an operating method for a finishing train for rolling a strip,

- wherein a control computer for the finishing train at the latest at a time at which a first band point of the band is still in front of the finishing train, for the first

Band point an actual size and a nominal size are known,

- where the actual size for the actual energy content of the first

Band point and nominal value are characteristic of the desired energy content of the first band point,

- the actual value is bezo ^¬ gen to a location in front of the finishing line and the target size is based on a location downstream of the finishing ^¬ road,

- wherein the control computer determines a leading variable prior to the arrival of the first band point in the finishing train for the first band point on the basis of a determination rule,

wherein the control computer determines a Leitge ^¬ speed based on the Leitgröße and operates the finishing train at the time of entry of the first band point in the finishing train at the conduction velocity,

- in the determination of the lead the

Actual size and the target size of eintre ^¬ border in the finishing line band point received.

The present invention further relates to a computer program comprising machine code which can be processed directly by a control computer for a finishing train for rolling a strip and whose execution by the control computer causes the control computer to operate the finishing line according to such an operating method.

The present invention further relates to a control computer for a finishing train for rolling a belt, wherein the control computer is designed such that it operates the finished road ^¬ according to such a method of operation.

The present invention further relates to a finishing line for rolling a belt, which is equipped with such a control computer.

As a rule, a hot strip mill consists of at least one finishing train and one cooling line downstream of the finishing train. Optionally, the finishing train may be preceded by a roughing train - alternatively or in addition to the cooling section - or the finishing train may be preceded by a casting device. The finishing train has a number of rolling stands. The number of rolling stands can be determined as needed. As a rule, there are several rolling stands, for example four to seven rolling stands. In individual cases, however, only a single rolling stand can be present. For each roll stand - regardless of their number - a setpoint decrease is specified for every rolling pass to be performed. If several rolling stands are available, generally further input and / or outlet-side desired trains are specified. If only a single roll stand is present, one and / or Auslaufseifiger Sollzug can be specified. However, this is not mandatory.

One of the targets to be met in a hot strip mill is the final rolling temperature, that is, the temperature at which the strip exits the finishing train. Al ^¬ tively to finish rolling temperature can also be used the energy content of the tape at this location variable describing others, such as the enthalpy. The target size should be kept as far as possible over the entire length of the tape. The target size may alternatively be constant or vary along the length of the band. To achieve the target usually the Leitge ^¬ speed of the finishing train is adjusted accordingly. The conduction velocity is a speed from the - gege ^¬ appropriate in conjunction with the einzu- in the finishing train Vice border pass reductions and desired features - within the

Finishing occurring strip and Walzenumfangsgeschwindig ^¬ speeds are clearly determined. For example, it may be a fictitious speed of the tape head or the speed of the first rolling stand of the finishing train. The guide speed can be defined, for example, as a function as a location of the tape head.

As a further actuators Zwischenge ^¬ Ruest cooling equipment and / or the finishing train superiors arranged induction furnace can optionally be present. These actuators act - as well as cooling devices of the cooling section - only locally on the tape. In the context of the present invention, however, the presence of these other actuators is of minor importance. Decisive, it depends on the conduction velocity (or cha for the conduction velocity ^¬ istic size, for example, the mass flow) and de ^¬ ren investigation.

As already mentioned, the finishing train is usually arranged downstream of a cooling section. In the cooling section, the strip is cooled in a defined manner to a reel temperature (or enthalpy). The speed with which the belt passes through the cooling ^¬ stretch is determined by the Leitgesch ^¬ sets. The adjustment of the cooling curves required for the individual strip points is achieved by tracking the strip points and by controlling the control valves of the cooling devices of the cooling section, which adjust the flow of coolant, in a time-correct manner. The control valves, in practice considerable tarry ^¬ delay times that are often in the order of several seconds. In order to be able to control the control valves in good time, it is therefore necessary to be on time knowing in advance when a certain band point is within the control of a particular cooling device. To calculate exactly when a particular strip point enters this sphere of influence, and when it comes out of him, it is necessary not only to know the current value of the Leitge ^¬ speed, but also the future course of conduction velocity, at least in the context of the tarry ^¬ time of the control valves. In addition, the cycle time as such, ie the time required by the respective band point for passing through the cooling section, also influences the reel temperature. The cycle time is - of course - influenced by the course of the conduction velocity. In the prior art it is known to determine the Leitgeschwindigkeits- course in a simplified manner. For example, an initial value is specified with which the tape head is to pass through the finishing train. Furthermore, a Accelerat ^¬ nigungsrampe is specified, over which the tape is accelerated to a speed Endge- when the tape head is leaked from the finishing train. In practice, this procedure proves to be unsuitable for maintaining a predetermined desired final rolling temperature (or a corresponding temperature profile) with high accuracy.

In the prior art, the (actual) is further known to detect and final rolling temperature track the Leitgeschwindig ^¬ ness in the sense of a minimization of the deviation of the actual final rolling temperature of the predetermined Sollendwalztempe-. This tracking can be done by means of a classical or - as described for example in DE 103 21 791 AI - by means of a model predictive control. Regardless of the type of control (classic or model predictive), the control intervention, ie changing the conduction velocity, is carried out simultaneously to determine the

Conduction velocity. Any prediction is limited - analogous to the unregulated procedure - to the specification of a future expected acceleration ramp. Whether on reason the set and actual sizes of the next control step is actually assumed the predicted conduction velocity is not certain. Furthermore, the prediction systematically extends to a single control step.

Although in practice this procedure usually proves to be suitable for maintaining a predetermined desired final rolling temperature (or a corresponding course) with high accuracy. However, with this procedure it is not possible to predict in which direction and by which value the guide speed will actually vary in the next control step. The possible prediction is more of a guess than a real determination.

In addition, the prediction, even if it were correct or at least approximately correct, would in principle be limited to a single control step in the teaching of DE 103 21 791 A1. This would be completely insufficient for a timely tracking of the control signals for the actuators of the cooling section or of the intermediate structure cooling devices of the finishing train. Varying the conduction velocity results therefore that the set of the actuators of the cooling line ^¬ brought coolant quantities not been to the strip points ^¬ be introduced for which the coolant amounts previously calculated. Therefore, deviations of the temperature (or of the energy content) of the band points at the end of the cooling ^¬ stretch (for example, on a reel) of desired setpoints arise. The exact compliance with the final rolling temperature is therefore "bought" in the prior art with an increased fluctuation, for example, of the coiler temperature.

In the earlier, not yet published on the filing date of the present invention the European patent application 09 171 068.1 (filed 23/09/2009) describes a model predictive regulation-which a Kühlstre ^¬ blocks regulates a finishing mill and jointly by means of a prognosis. Here also the mass flow is predicted. This approach requires coolant quantities output by actuators of the cooling section, to determine the mass flow. Furthermore, the mass flow is always readjusted immediately. Therefore, this approach does not solve the problem of reliably determining a conduction velocity history in advance.

The object of the present invention is to provide opportunities to be able to reliably realistically determine the Leit ^¬ size not only for this band point, but also for incoming after this band point in the finishing line tape points before entering a band point in the finishing mill.

The object is achieved by an operating method for a finishing train with the features of claim 1. Advantageous embodiments of the operating method are the subject of the dependent claims 2 to 14.

According to the invention, it is provided

- That a control computer for the finishing train at the latest at a time at which a first band point of the band is still in front of the finishing line, for the first band point, a number of second band points and a number of third band points of the band one each

Actual size and a nominal size are known,

- that for each band point the respective actual size for the

Is the energy content of the respective band point and the jewei ^{¬ term} setpoint for the target energy content of the respective band point are characteristic,

- That for each band point the respective actual size is related to a location in front of the finishing train and the respective

Target size is related to a location behind the finishing train,

- that the second band points after the first band point and the third band points before the first band point enter the finishing line,

- That the control computer before entering the first band point in the finishing line for the first band point and at least a portion of the second band points based on a Determines a reference variable for each specific band point

- That the control computer determined on the basis of the determined for the respective band point guide each a Leitgeschwindig- speed and operates the finishing train at the time of running ^{¬ of} the respective band point in the finishing mill with the respective conduction velocity and

- that for the respective command variable in their Ermittlungsvor ^¬ writing the actual value and the desired size of the entering at that time in each case into the finishing train band point as well as the actual quantity and the target quantity received at least one already occurred at this time in the finishing train band point.

For example, it may be provided

- that the control computer each of the guide sizes based on a

Variety of individual wire sizes determined

- that each individual size is related in each case to one of the band points, the actual and nominal size of which go into the determination of the respective reference variable,

- That the control computer for each band point whose Einzelleitgröße determined such that a respective expectation ^¬ size matches the corresponding target size, and

- that the respective expected size for an expected

Energy content is characteristic, the respective band ^¬ point at the place behind the finishing train, to which the je ^¬ Weils corresponding nominal size would assume if the control computer, the finishing train throughout the passage of the respective band point through the finishing line with one with the Single wire size correspon ^¬ dierende leading speed would operate.

To determine the respective command variable based on the respective plurality of Einzelleitgrößen the control computer, for example, carry a weighted or unweighted Mittelwertbil ^¬ dung. Alternatively it can be provided that the control computer for each band point, for which he determined its Leitgröße,

- based on the actual quantities used in determining the benchmark for the particular band point, an effective

Actual size and based on the setpoint values, which are included in the determination of the guide variable for the respective band point, an effective setpoint,

- determines an expected size, which is characteristic of an expected energy content which the respective band would assume point at the location behind the finishing train, to which the ef ^¬ fective target size is received if the control computer the finishing train of the respective during the entire run Bandpunktes would run through the finishing mill with a corresponding with the Leitgröße for the respective band point leading speed, and

- Determines the Leitgröße such that the expected size at the location behind the finishing train to which the effective target size is related, the effective target size. Again, the control computer to determine the effective actual size and the effective target size make a weighted or unweighted averaging.

Likewise, it may alternatively be provided

- That the control computer to determine the guide the

Initially, the guide values are provisional values

- That the control computer for the first band point and at least a part of the second and third band points determines a je¬ ^¬ expectation variable,

- That each expectation is characteristic of an expected energy content, the respective band point at the place behind the finishing train, to which the kor ^¬ respondierende target size would assume if the control computer, the finishing train throughout the passage of the respective band point through the Fertigstra ^¬ SSE would operate with conduction velocities corresponding to the scheduled command variables, and - That the control computer varies the applied guide values, so that a target function is optimized, in which the Beträ ^¬ ge the differences of the expectation variables of the corresponding target values received.

In the latter alternative is preferably vorgese ^¬ hen that additionally, a penalty term enters into the objective function will be punished by means of which changes in conduction velocity.

Regardless of which of the three alternatives mentioned above is taken, the operating method according to the invention is still very compute-intensive. To reduce the computational effort is preferably provided

that the control computer prepares in advance a data field in which the control computer stores the expected variable resulting for the respective possible actual variable at the respective possible guide speed for a multiplicity of possible guide speeds and possible actual variables, and

- That the control computer determines the parameters for the band points using the data field.

The operating method, as described so far, is already working quite well. It can thereby be further improved that the control computer

- At least for a part of the band points a respective He ^¬ maintenance size determined, which is characteristic of an expected energy ^¬ content, for the respective band point at the place behind the finishing train to which the ^¬ Weil's corresponding target size is related to Reason of the guide speeds with which the control computer operates the finishing train during the entire passage of the respective band point through the finishing train, is expected - after the passage of the respective band point through the

Finished road accepts a measure that for an actual energy content of each band point at the place is characteristic behind the finishing train, to which the corresponding nominal value is related, and

- automatically adapted a model of the prefabricated street (1) on the basis of a comparison of the expected energy content with the actual energy content, and

- the model of the finishing train, characterized adapted that it adds in the use of the data field to the actual variables an offset, scaled, the conduction velocities with a scaling factor and / or adds an offset to them and / or to the ermittel ^¬ th using the data field expectation quantities a Offset added.

In a preferred embodiment of the present invention, the actual size and the desired size of the points already entered the finishing line are received for each master in their determination only if these tape points at the time for which the respective guide is determined, not from the finishing train have leaked. In particular, the actual and nominal values of all band points, which are located at the time when the determined band point enters the finishing train in the finishing train, can be included in the determination of the reference variable for a specific band point.

The operating method, as described so far, is already working quite well. It can thereby be further improved that the control computer at least for a part of the band points

- determining a respective expected size that is characteristic of a he ^¬ waiting energy content that is for each strip point at the location behind the finishing train, to which the respective corresponding target size is related, due to the conduction velocities with which the control computer the finishing train during the entire pass of the respective strip point through the finishing train is expected,

- After the passage of the respective band point through the

- Based on a comparison of the expected energy content with the actual energy content automatically nachzuführen at least a portion of the already determined benchmarks.

When the control computer compares the expected energy content with the actual energy content and tracks the control values, it is possible for the computer to perform the comparison for all the tape points one after the other. However, it is sufficient to perform the comparison for a part of the band points, for example every third or every tenth band point.

Of course, when the control computer tracks the guiding parameters, the control computer takes into account the changed course of the parameter when determining expectation variables.

It is possible that the control computer carries out the tracking for all already determined control variables. Preferably, however, it is provided that the control computer automatically readjusts only those command variables based on the comparison, which were determined for strip points which have a minimum distance at the time of performing After ^¬ from the entrance of the finishing train. This procedure is particularly advantageous if the control computer or other control means uses the determined command variables for determining at least a further manipulated variable and the further controlling variable ^¬ SSE delayed by a dead time and acts locally on the tape. This procedure is optimal if the minimum distance is determined in such a way that a time difference corresponding to the minimum distance is at least as long as the dead time.

In addition to tracking already determined command variables can control computer - of course - the determination rule for not yet determined command variables as such adap ^¬ animals. The adaptation result, depending on the circumstances of the individual case ^¬, dessel- already in identifying additional command variables ben Bandes or only in the determination of guide values for subsequent bands are considered.

The two last-mentioned procedures - keyword "tracking already determined parameters" on the one hand and "adapting the determination rule" on the other hand, for example, be coupled to each other such that the control computer comprises a model of the finishing train, is determined by means of which temperature for a band point on the outlet side of the finishing train is expected if the jewei ^¬ lige band point upstream of the finishing train has a given temperature and the finishing train passes through, while the finishing train at a given conduction ^{speed is ¬} driven. In this case, the model can be adapted immediately. This corresponds to the adaptation of the investigation rule. Is then re-determined for at least one of the already ermit ^¬ telten command variables the guide variable using the adapted model of the finishing train. This ent ^¬ speaks from the approach of tracking the already determined control variables. Optionally, a smooth transition from the originally determined control variables to the newly determined control parameters can take place.

The operating method according to the invention already provides then represents a significant advance over the prior art, when the prediction horizon is relatively small, amounts to ^¬ play, three to five points on the strip. Its full superiority but showing operation method of the invention particularly when the first tape point, and the part of the second strip points for which their respective Leitgrö ^¬ SSE was determined prior to entry of the first tape point in the finished ^¬ road, a prediction horizon entspre ^¬ Chen, at least as large as the dead time with which the further manipulated variable acts on the band. This applies in particular in cooperation with the tracking of the already determined control variables, if the tracking is also tuned to the said dead time. In a preferred embodiment of the present invention, it is further provided that the control computer chains together the determined guiding variables or the corresponding guiding speeds by a spline, so that a guiding-speed course resulting from the linking is continuous and differentiable. The resulting advantage is a smoother and more uniform operation of the finishing train. This is especially true if the is erge ^¬ bende Leitgrößenverlauf is not just differentiable special countries is continuously differentiable.

The control computer preferably executes the determination of the control variables online or in real time as part of a precalculation.

The inventive object is also achieved by a Compu ^¬ terprogramm of the type mentioned. The computer program is designed in this case such that the control computer executes an operating method with all steps of an operating method according to the invention.

The object is further achieved by a control computer for a finishing train for rolling a belt, which is formed from ^¬ such that it performs such operation during operation.

The object is further achieved by a finishing train for rolling a belt, which is equipped with such a control computer.

Further advantages and details will become apparent from the following description of exemplary embodiments in conjunction with the drawings. In a schematic representation:

1 shows schematically a hot strip mill,

2 shows a flow chart,

Figures 3 to 6 exemplify various states of a Fer ^¬ tigstraße, FIG 7 an example of a snapshot of the finished ^¬ road,

FIGS. 8 to 11 are flow charts,

12 shows a model of the finishing train,

13 shows a flow chart,

14 shows a timing diagram and

FIG. 15 is a flow chart. FIG.

According to FIG 1, a hot strip mill comprising at least one Fer ^¬ tigstraße 1. In the finishing mill 1, a tape 2 is to be rolled. The band 2 is usually a metal band, in ^¬ example, a steel strip. Alternatively (to steel), the band may be made of copper, brass, aluminum or another metal.

The finishing train 1 has for rolling the belt 2 a Walzge ^¬ setup 3 or - as shown in Figure 1 - several rolling stands 3 on. Three such rolling mills 3 are shown in FIG. 1. The actual number of rolling mills 3 can be three, as shown. Alternatively, it may be different from three, in particular larger. In general, the number is at ^¬ at roll stands 3, four to eight, in particular 5 to 7. Further, illustrated only the work rolls of the roll stands 3 in FIG 1 (2-high). In general, the rolling stands 3 include in addition to the work rolls back-up rolls (4-high), sometimes in addition also intermediate rolls (6-high).

The finishing train 1 may comprise a heating device 4, for example an induction furnace. If the heating device 4 is present, it is usually located at the entrance of the finishing train 1. Alternatively or additionally - as with intermediate stand cooling devices - 3 heaters may be present between the rolling stands. The heating device 4, if it is present, in the context of the present invention as part of the finishing train 1 ^¬ see. As an alternative or in addition to the heating device 4, the finishing train 1 may have interstitial cooling devices 5. If the inter-frame cooling devices 5 present are each inter-frame cooling device 5 of two of the rolling stands 3 eingabelt. They are, if they are present, part of the finishing train 1. Each inter-frame cooling device 5 has at least one control valve 5 'and at least one spray nozzle 5 ".

The finishing train 1 can continue nachge ^¬ a cooling track 6 nachge ^¬ assigns. If the cooling section 6 is present, it has cooling devices 7. Each cooling device 7 has at least one control valve 1 'and at least one spray nozzle 7 ".

Both by means of the intermediate stand cooling devices 5 and by means of the cooling devices 7, the strip 2 is cooled with a liquid cooling medium (usually water with or without admixtures). The difference between the intermediate stand cooling devices 5 and the cooling devices 7 of the finishing train 6 is that the cooling devices 7 are arranged behind the last rolling stand 3 of the finishing train 1, the intermediate stand cooling devices 5 between each two of the rolling stands 3.

According to FIG. 1, the finishing train 1 is furthermore equipped with a control computer 8. The control computer 8 serves At the very least the control of the finishing train 1, that is, the Walzgerüs ^¬ te 3 and - if present - of the heater 4 and the intermediate stand cooling means 5. If desired, also control other devices, the control computer 8, for example, the cooling section 6 and the cooling means 7 Alternatively, the cooling section 6 can be controlled by another control device 8 '.

The operation of the control computer 8 is determined by a Compu ^¬ terprogramm 9, the control computer 8 - is fed - beispiels- example via a mobile data carrier 10th The mobile data carrier 10 can be designed as required, for example as a CD-ROM, as a USB memory stick or as an SD memory card. On disk 10 is the computer pro stored program 9 in machine-readable form, for example in electronic form.

The computer program 9 comprises machine code 11 with which the control computer 8 is programmed and which can be processed directly by the control computer 8. The execution of the machine code 11 by the control computer 8 causes the control ^¬ calculator 8 operates the finishing train 1 according to an operating method, which will be explained in more detail below. The programs mieren with the computer program 9 thus effects a entspre ^¬ sponding embodiment of the control computer. 8

In the context of the operating method, the control computer 8 according to FIG. 2 must have an actual size G and a number of second band points 13 of the band 2 in a step S1 for a first band point 12 of the band 2, a number of second band points 13 of the band 2 Target size G * be known, and at the latest at a time when the first band point 12 is still in front of the finishing train 1.

It will be apparent from the explanations that follow that the control computer 8 does not have to know all the actual variables G and the nominal values G * for the first band point 12, the second band points 13 and the third band points 13 'at the same time. However, it will also be apparent that the knowledge must be completed before the first band point 12 enters the finishing train 1.

The second band points 13 are all located behind the first band point 12, that is, they run into the finishing line 1 after the first band point 12. The third band points 13 'enter the finishing train 1 before the first band point 12. The Figu ^¬ ren 3 to 6 show respective embodiments. The actual size G of each band point 12, 13, 13 'is for the

Energy content characteristic of the respective band point 12, 13, 13 'at a location xE before the finishing train 1 has. The actual size G is thus at the location xE in front of the finishing train 1 based. The location xE can be determined as needed. In particular, according to FIG. 1, it may be a location which is located immediately before the first device 4, 3 of the finishing train 1, by means of which - directly or indirectly - the temperature of the belt 2 is influenced. It is still possible that a temperature measuring device is arranged at this location. In general, however, the temperature measuring device 14 is arranged upstream of the location Xe. The target quantity G * each band point 12, 13, 13 'is characteristic of the energy content of the respective strip point 12, 13, 13' behind the finishing train 1 is to aufwei ^¬ sen in a place xA. The desired values G * are therefore related to the location xA behind the finishing train 1. The location xA can be determined as required, analogously to the location xE in front of the finishing train 1. In ^¬ example, it may be the location of a temperature measuring device 15, which downstream of the finishing train 1, the cooling section 6, however, upstream. The type of the actual size G and the target size G * can be determined as needed. As a rule, these are the corresponding temperatures. Alternatively, an enthalpy in particular comes into question. For good order's sake it is noted that the term "place" always refers below to a location that is relative to the finishing line 1 stationary. The term "strip-point" be ^¬ takes place in contrast, always on a point be ^¬ züglich of the band 2 is stationary. Distances between the points on the strip 12, 13, 13 'from each other are determined in the present invention is not limited by their geometric distances, since these distances change due to the rolling of the strip 2 in the Fer ^¬ tigstraße. 1 Rather, the distances are defined by the mass which is located between the band points 12, 13, 13 '.

The band points 12, 13, 13 'may, based on the mass of the band 2 located between them, be equidistant. AI ternatively, the band can points 12, 13, 13 'thus be defined, that - is detected in time-equidistant steps in each case a measured value for the actual value G - for example by means of Temperaturmessein ^¬ direction fourteenth The time interval between two successive band points 12, 13, 13 'is generally between 100 ms and 500 ms, typically between 150 ms and 300 ms. For example, it may be 200 ms. In a step S2, the control computer 8 determines - even ^¬ understood prior to entry of the first tape point 12 in the finishing mill - for the first strip point 12 based on a determination rule a master value L *. In a step S3, the control computer 8 is also determined at least for a part of the second strip points 13 by means of a Ermittlungsvor ^¬ writing a respective master value L *. The control computer 8 also executes step S3 before entering the first strip point 12 into the finishing train 1. The steps S2 and S3 of FIG. 2 usually form one unit in practice. The separate illustration in FIG 2 serves single ^¬ Lich of better illustrate the present invention.

Preferably, the control computer 8 determines in the context of step S3 for all second band points 13, which - starting from the first band point 12 - are within a predetermined Prä ^¬ diktionshorizontes H, whose master size L *. If, as part of step S3, the guide variable L * is determined for a specific second band point 13, the respective other band points 13, which lie between the first band point 12 and the determined second band point 13, will generally also be the other Guide size L * determined.

The command variables L determined * are each for charac- table with which conduction velocity vL, the control computer 8, the finishing mill 1 operates when the strip point 12, 13, for which the respective command variable L was determined *, enters the Fer ^¬ tigstraße. 1 The guide speed vL can be For example, be the speed at which the belt 2 enters the finishing train 1. Alternatively, it may be the speed at which the tape 2 runs from the finishing ^¬ road first Other sizes - for example, a determination of the mass flow or a roller speed or a roller peripheral speed - are conceivable. Is decisive ^¬ DEND that by conduction velocity vL - given ^¬ appropriate in combination with reductions per pass and desired features - are uniquely determined every band and Walzenumfangsge- occurring in the finishing train 1 speeds.

In a step S4, the control computer 8 determines, if necessary, the corresponding guide speeds vL on the basis of the control variables L *. In a step S5, the control computer 8 operates the finishing train 1 in accordance with the guide speeds vL determined in step S4. The control computer 8, the conduction velocity vL thus always such a that the finishing train 1 is operated with the Leitge ^¬ speed vL at any time, with the guidance parameter L * of the currently entering the finishing train 1 Volume point 12, corresponds. 13

The determination rule for determining the guiding variables L * is in each case specific to the respective band point 12, 13. The determined value of the reference variable L * for a particular band point 12, 13 can not therefore be readily determined by the value of the reference variable L * for another band point 12, 13 CLOSED ^¬ sen. In particular, go into the determination procedure for the guide variable L * for a particular strip point 12, 13, first the actual value G, and the target quantity of G * ^¬ correspond to Tape point 12,. 13 In addition, the actual quantities G and the set values G * of at least one further band point 12, 13, 13 ', which at the time at which the considered band point 12, 13 enters the finishing train 1, already enter the finishing train 1 in the respective determination rule is turned ^¬ occur. This situation will be explained below in conjunction with FIG. 7. FIG. 7 shows by way of example a snapshot of the finishing train 1 while the belt 2 is being rolled in the finishing train 1. The band points 12, 13 are referred to in connection with the explanations to FIG. 7 as band points Pi (i = 1, 2, 3,...).

Suppose, are as shown by FIG 7 is currently the band points P5 to P30 in the finishing mill 1. The strip points PI to P4 have in this case, the finished ^¬ road 1 already left again, so are from the finished road one already resigned. The band points P31 to P35 are still in front of the finishing train 1. The band point P31 occurs in this case next in the finishing train 1 a. After the band point P31, the band points P32, P33, P34 and P35 enter the finishing train 1 in succession. The actual and target variables G, G * are known up to and including the band point P35.

In the situation illustrated in FIG. 7, the determination of the guiding variable L * for the strip point P4 must already have been completed for a long time, since the strip point P4 has not only already entered the finishing train 1, but has even already left the finishing train 1 again , In the development of the guide variable Determined ^¬ L *, with which the finishing mill 1 was operated at the time point at which the strip point P4 has occurred in the finishing mill 1, according to the invention is considered ^¬

- The actual size G and the target size G * for the band point P4 and

- The actual size and the target size G, G * for at least one of the band points PI, P2 and P3.

Assuming that the prediction horizon H is equivalent to four band ^¬ points that determine the command variable L * need for the strip point P4 have been completed in the finishing train 1 one clock before the date of Eintre ^¬ least the band point PI. Analogously, the determination of the reference variable L * for the band point P7 is discussed

- The actual and the target size G, G * for the band point P7 and - the actual and the target size G, G * for at least one of

Band points PI to P6.

This determination must have been completed at the latest at the time of the occurrence of the band point P3.

The band point P30 is the band point that has just entered the finishing train 1. In the determination of the control variable L *, which must have been completed by the time of the occurrence of the band point P26 at the latest, have arrived

- The actual and the target size G, G * for the band point P30 and

- The actual and the target size G, G * for at least one of the band points PI to P29.

As a rule, it is sufficient for the determination of the guiding quantity L * for the band point P30 to take into account the actual and nominal quantities G, G * of the band points P5 to P30, ie those band points which, according to the representation of FIG Finishing line 1 are located.

In an analogous manner, the guiding variables L * are determined for the band points P31 to P35. In the representation of FIG. 7, the band point P31 corresponds to the first band point 12, the band points P32 to P35 to the second band points 13. The determination of the guide quantities L * for these band points P31 to P35 must each be at the latest at the time of entry of the band point P27 to P31 in FIG completed the finishing train 1. The band ^¬ points PI to P30 corresponding to the third band points 13 '.

In the determination of the guide variable L * for the band point P31 ge ^¬ hen - The actual and the target size G, G * for the band point P31 and

- The actual and target variables G, G * for at least one of

Band points PI to P30, preferably for at least one of the band points P6 to P30.

The latter is particularly true because the band points PI to P5 at the time when the band point P31 enters the Fer ^¬ tigstraße 1, have already exited the finishing train 1 again.

In an analogous manner, the guiding variables L * for the band points P32 to P35 can also be determined. For example, for the band point P35 go into the determination for the guide size L *

- The actual and the target size G, G * for the band point P35 and

- The actual and target variables G, G * for at least one of

Band points PI to P34. The actual and nominal sizes G, G * for the strip points PI to P9 can hereby be disregarded, as the band points already quit PI to P9 to the date on which the band point P35 enters the Fer ^¬ tigstraße one back from the finishing train 1 are.

For the other band points P32, P33 and P34 analogue versions apply.

In a preferred embodiment of the present invention for each entering the finishing train 1 band point 12, 13 - for example, for the band point P31 of FIG 7 - the lead L * thus on the basis of the actual and nominal values G, G * those band points 12, 13, 13 'determines that are located to the ^¬ sem time just in the finishing train 1 are therefore not yet withdrawn from the finishing line first

In the finishing train 1 are usually at the same time a plurality of band points 12, 13, 13 '. Typical numbers values are between 10 and 200, for example between 50 and 100. It is possible to take into account only a few band points 12, 13, 13 'of the band points 12, 13, 13' which are currently located in the finishing train 1 at a particular time - For example, every second or every fourth band point 12, 13, 13 '. This approach leads to a redu ^¬ ed computational effort and yet gives acceptable resulting ^¬ nisse. Preferably, however, the actual and nominal values G, G * of all band points 12, 13, 13 'are taken into account for the determination of the guide variable L * for a specific band point 12, 13, which coincide at the time of entry of that band point

12, 13, the guide variable L * is determined already in the finished ^¬ road 1 in the finishing train. 1 The illustration shown in FIG. 7 is of course purely exemplary. For example, the number of the Fer ^¬ tigstraße 1 located (third) band points 13 'purely exemplary. Also, the number of (second) band points 13 whose lead variable L * is predicted (here the band points P32 to P35) is purely exemplary. Also, the prediction horizon H is purely exemplary. In particular, in practical An ^¬ applications the prediction horizon H several seconds Betra ^¬ gene that is, at a timing of, for example 200 ms per measurement recording of the actual value G corresponding to five times the number of strip points 12, 13. In some cases even a prediction horizon H of up to one minute and more possible, which ent ^¬ speaking at a timing of 200 ms of tape point to strip point ei ^¬ nem prediction horizon H 300 tape points and more.

It is possible that the control computer 8 in step Sl of FIG 2, the actual and target variables G, G * for all band points 12,

13, 13 'of the (entire) band 2 are known. In this case, it is possible that the control computer 8 passes through the steps S2 and S3 only once and in the steps S2 and S3 - so ^¬ say in one stroke - the guiding variables L * for all band points 12, 13, 13 'of the band 2 determined. In this case the control computer 8 executes the determination of the control variables L * on the basis of a prediction online.

Alternatively, it is possible that the control computer 8. Although Sl of Figure 2 are within the step, the actual and desired sizes G, G * for all points on the strip 12, 13, 13 'of the entire tape 2 be ^¬ known, he in steps S2 and S3 of FIG 2 but always only for some of the band points 12, 13, 13 'whose Leit ^¬ sizes L * determined. In this case, the steps S2 and S3, as indicated by dashed lines in FIG 2, involved in a loop. In this case, the control computer 8 executes the determination of the control variables L * in real time with the control of the finishing train 1. In this case, the control computer 8 determines the guiding variables L * by the prediction horizon H, so to speak.

Is likewise indicated by broken lines in FIG 2, it is so ^¬ even possible that the step Sl is integrated in the loop. Also in this case, the control computer 8 executes the determination of the control variables L * in real time.

In the event that also the step Sl is involved in the loop, the control computer 8 in a given passage of the loop only the actual and target variables G, G * of band points 12, 13 known, which not yet in the finishing mill. 1 occurred. The actual and target variables G, G * of the band points 13 ', which have already entered the finishing train 1, are known to the control computer 8 in this case, however, on the basis of earlier loop passes. In this case, it is only necessary for the control computer 8 to "remember" the "old" actual and target variables G, G *.

For determining the guiding variables L * for a specific band point 12, 13 - ie for implementing steps S2 and

S3 of FIG 2 - are different approaches possible. The different alternatives are listed below in succession Connection with the 8, 9 and 10 explained in more detail. After Be ^¬ may hereby FIG 7 is to be used.

In a first possible embodiment of the steps S2 and S3 of FIG 2 8 selects the control computer shown in FIG 8, in a step Sil first one of the strip points 12, 13, ^¬ sen actual and desired size G, G *, the control computer 8 already be ^¬ known are. For example, the control computer 8 selects the band point P31 of FIG. 7.

In a step S12, the control computer 8 13 determines all band ^¬ points 12, 13 ', their actual and nominal sizes G, G * in the determination of the guide variable L * for the strip point 12, 13 is ^¬ hen, the control computer 8 in Step Sil has selected. For example, the control computer 8 - see FIG. 7 - determine the band points P6 to P31 for the band point P31. In an analogous manner, the control computer would determine the band points P7 to P32 for the band point P32 in step S12, for the band point P33 the band points P8 to P33, etc.

In a step S13, the control computer 8 selects one of the band points 12, 13, 13 'determined in step S12. In a step S14, the control computer 8 determines a single-wire size 1 * for the tape point 12, 13, 13 'selected in step S13-for example for the tape point P6. * Go 'only the actual value G and the target size ^¬ G * of the selected band in step S13, point 12, 13, 13 in the determination of the Einzelleitgröße. 1 The respective individual wire size 1 * is therefore related to this one tape point 12, 13, 13 '.

The Einzelleitgröße 1 * determines a corresponding Leit ^¬ speed vL. The control computer 8 assumes that the band point 12, 13, 13 'considered in step S14 passes through the finishing line 1 and the finishing line 1 through the finishing line 1 during the entire passage of the considered band point 12, 13, 13' Time of entry into the finishing train 1 until the time of departure from the tigstraße 1 - is constantly driven ^¬ with this guide speed vL, which is determined by the corresponding Einzelleitgröße 1 *. In this case, an energy ^¬ content is for the considered strip point 12, 13, 13 'at the location xA, to the desired size G * of the considered band point 12, 13, 13' is related expected. The control computer 8 determines this expected energy content. The determination of the expected energy content can be determined by the control computer 8, for example by means of a finishing road model. Suitable finishing road models are known as such. They are used, for example, to determine the expected final rolling temperature, see the already mentioned DE 103 21 791 AI.

The expected energy content is characterized by a corresponding expectation quantity GE. The expectation variable GE can alternatively be the temperature or the enthalpy, analogous to the actual and set values G, G *. The control computer 8 determines the Einzelleitgröße 1 * for the considered band point 12, 13, 13 'in step S14 such that the expected size GE with the target size G * for the considered band point 12, 13, 13' matches.

In a step S15, the control computer 8 checks whether it has already performed step S14 for all band points 12, 13, 13 'to be used. If this is not the case, the control computer 8 returns to step S13. In the renewed execution of step S13, the control computer 8 naturally selects another, previously not yet considered, band point 12, 13, 13 ', which enters into the determination of the sought-after control variable L *, for example the band point P7.

If the control computer 8 determines in step S15 that it has already determined all the required individual line sizes 1 *, the control computer 8 proceeds to a step S16. In step S16, the control computer 8 determines based on all ^¬ A zelleitgrößen 1 * he has determined in the context of the repeated working off of the step S14, the guide variable for L * the selected band in step Sil point 12, 13. Example ^¬ example, the control computer 8 are the weighted or unweighted average of the Einzelleitgrößen 1 *. In a step S17, the control computer 8 checks whether it has the

Steps S16 to Sil already for all strip point 12, 13 out ^¬ leads has whose command variables L * should be calculated. If this is not the case, the control computer 8 returns to step Sil. There, the control computer 8 selects of course another, not yet considered strip point 12, 13. Otherwise, the process of FIG 8 is completed ^¬ det.

The procedure of FIG. 8 is implemented slightly differently in practice than explained above. For the A ^¬ zelleitgröße 1 * for a particular strip point 12, 13, 13 '- for example, the strip point P28 of FIG 7 - is in the determination of the guide variable L * of many points on the strip 12, 13, 13' a, for example - based on FIG 7 - in the determination of the band points P28, P29, ... P53. Of course, it is possible and even preferred to determine the respective Einzelleitgröße 1 * only once and then save, so that they must be retrieved for later uses only from the memory.

As an alternative to the procedure of FIG. 8, according to FIG. 9, it is possible to replace steps S13 to S16 of FIG. 8 by steps S21 to S23. The steps Sil, S12 and S17 of FIG. 8 are adopted in the procedure of FIG. 9 from FIG.

In step S21, the control computer 8 determines on the basis of

Actual variables G of the band points 12, 13, 13 'determined in step S12 are an effective actual variable G'. In an analogous manner, the control computer 8 determines an effective setpoint G '* in step S22 on the basis of the setpoint values G * of the band points 12, 13, 13' determined in step S12. For example, the control computation ^¬ ner 8 in Steps S21 and S22 may be a weighted or unsaturated weighted averaging. Regardless of which approach is taken, the Vorgehenswei ^¬ sen of the steps S21 and S22 should, however, correspond to each other.

In step S23, the control computer 8 determines the reference variable L * for the band point 12, 13 selected in step S11.

The master variable L * determined in step S23 corresponds to a corresponding master speed vL. If the im

Step selected Sil point 12, 13 at the location xE, to which the actual size G of the selected in step Sil band point 12, 13, the effective actual size G 'would have and the control computer 8, the finishing line 1 during the entire run of the in step Sil selected band point 12, 13 would operate at this conduction velocity vL, would be expected for this band point 12, 13 at location xA, to which the target size G * of the selected in step Sil point 12, 13, an actual energy content, the is characterized by an expectation GE. The STEU ^¬ errechner 8 determines the master value L * in step S23 such that the expected size GE determined coincides with the effective target size G '*. The determination of the expectation ^¬ size GE may - be effected by means of an appropriate, known finishing train model - analogous to the procedure of step S14 of FIG. 8

As an alternative to the procedures of FIGS. 8 and 9, it is possible for the guiding variables L * according to FIG. 10 to be determined as follows:

According to FIG 10 sets the control computer 8 in a step S31, the command variables L *, which it is to determine, - that the Leitgrö ^¬ SEN L * for the first swath dot 12 and for at least a portion of the second strip points 13 - first as a preliminary ^¬ te at . In step S32 the control unit 8 in step S31 for the considered strip points 12 determines, 13 a respective He ^¬ maintenance size GE. The determined at step S32 expectation ^¬ sizes GE are provided for the expected energy content of the respective corresponding band point 12, 13 charakteris ^¬ table that is expected for each strip point 12, 13 when the respective strip point 12, 13, the finishing mill 1 ent ^¬ speaking the scheduled course of the guide speed vL - as defined by the sequence of the guiding variables L * - goes through. The expected energy contents GE are each related to the location xA, to which the desired quantities G * for the band points 12, 13 are related.

In a step S33, the control computer 8 forms a target function Z. At least the amounts of the differences of the expectation variables GE from the corresponding desired values G * enter into the objective function Z. For example, the Zielfunkti ^¬ on Z contain a sum, for example, each summand is the square of the difference of an expectation GE of the corresponding desired size G * according to the representation in FIG.

It is possible to use the above-described objective function Z as described so far. Preferably, however, further variables enter into the objective function Z. In particular, a penalty term ^¬ can enter into the objective function Z additionally be punished by means of which changes in conduction velocity vL. For example, the objective function Z can thus have the following form:

'j

In the two sums, different indices i, j were used because the indices i and j run over different ranges. ± and ßj are - in principle arbitrary, non-negative - weighting factors. In a step S34, the control computer 8 varies the applied guiding variables L * with the aim of optimizing the target function Z according to the embodiment above. With a correspondingly different design of the objective function Z, maximizing would also be possible.

The procedures of FIGS. 8 and 9 are applicable regardless of whether only a few control variables L * are determined in a single execution of steps S2 and S3 of FIG. 2 or whether the control variables L * are valid for all band points 12, 13, 13 'of the band 2 be determined in advance. The approach of FIG 10, however, provides only meaningful results when the prediction horizon H covers the entire band 2 or - if the band 2 is long enough - is sufficiently large. In particular, the prediction horizon H should be as large in the procedure of FIG 10 in case of a long tape 2, that it corresponds to at least the effective Fertigstra ^¬ ßenlänge, is preferably at least twice as large. The effective length of the finishing train is determined by the maximum number of simultaneously located arrival in the finishing train 1 Volume ^¬ points 12, 13, 13 determines'.

Both in the procedure of FIG 8 as well as part of the procedure of FIG 9 as well as part of the procedure of FIG 10 expectancy sizes GE must be ermit ^¬ telt. The determination of the expected quantities GE takes place - from the point of view - by means of a model of the finishing train 1, which models the thermal processes (heat conduction and heat transfer, possibly also phase transformations and microstructure) in the finishing train 1. Such models are known per se, see DE 103 21 791 AI.

It is possible to use such a model as such also in steps S14, S23 and S32. Preferred, JE but that the control computer 8 as shown by FIG 11 in a step S41 in advance - He ^¬ submit to the command variables L * that is in front of the - creates a data field. The control computer 8 deposits in the data field in a step S42 for a multiplicity of possible guide speeds vL and possible actual variables G, which expectation variable GE results for the respective possible actual variable G and the respective possible guide speed vL. Because in this case, the control computer 8 in the context of appropriately designed

Steps S2 and S3 of FIG 2 (or the steps S14, S23 and S32) determine the guiding variables L * for the band points 12, 13 using the data field. In the procedure ge ^¬ Mäss FIG 8, the control unit 8 determines the Einzelleitgrößen 1 * using the data field so that the use of the data field is of an indirect nature. In the procedure according to FIG. 9, the respective control variable L * is determined directly. In the procedure according to FIG. 10, the data field is used to determine the respectively resulting expected quantities GE.

By using the data field, a considerable ^acceleration can be achieved. For even the data field must indeed be determined within the framework of a preliminary calculation, that is, when the hot strip 2 for rolling in the finishing train 1 is already ready. So the data field is not determined offline ^¬ to. Rather, the data field must be determined online, al ^¬ so after the tape data have been given to the control computer 8. Therefore, only a few seconds are available for the determination of the data field. Nevertheless, a significant acceleration occurs. Because part of the data field Müs ^¬ sen relatively few values by means of the model of the finished ^¬ street 1 are completely calculated, for example, for every 10 possible actual values G and the 10 possible guiding principle speeds vL so that the model calculation for a total of 100 values are performed got to. However, this is still considerably faster than later in the context of steps S14, S23, S32 always for each individual band point 12, 13, 13 'with ^{¬ means of} the model of the finishing train 1 whose expected size GE to determine.

The nature of the integration of the data field in the procedures of FIG 8 and 9 is immediately apparent, since the actual size G the Control computer 8 is known and the relationship between the possible conduction velocity vL and the expected variable GE is unambiguous (the larger the given actual variable G is the conduction velocity vL, the greater the expected energy content of the corresponding band point 12, 13,

13 '). However, the data field can also be used in conjunction with the procedure of FIG. Because it can be formed in the first and usually already very good approximation for a particular band point 12, 13, 13 ', the mean of all Leitgrößen G * or all conduction velocities vL, with the finishing train 1 during the passage of the respective band ^¬ point 12, 13, 13 'is operated by the finishing train 1. This mean value can be regarded as the effective guide speed vL. The data field can thus be evaluated at this point to determine the expected size of GE for the ent ^¬ speaking strip point 12, 13, 13 '.

The data field can be designed as needed. Example ^¬, it may be a pure checkpoint field with examples play, 5, 8, 10, ... reference points for each dimension. Between individual linear interpolation points may be (for example by means of splines) INTERPO ^¬ lines linearly or not in this case. Alternatively, the data field may be formed, for example, as a neural network.

If the actual value G based on a measured variable is detected at ^¬ game by means of the temperature measuring device 14, it is possible to process the measured values directly. As a rule, the location xE is located in front of the prefabricated street 1, to which the actual quantities G are related, but behind the temperature measuring device 14. It is therefore necessary to divide the measured quantities into the actual quantities G (which refer to the location xE). convert. This is relatively easy, since only an air gap must be calculated. Input values for the air gap is measured by means of the Tempe ^¬ raturmesseinrichtung 14 temperature value and the time for each strip point 12, 13, 13 'is obtained until the corresponding strip point 12, 13, 13' the location xE before reached the finishing train 1. The time is given for each band point 12, 13, 13 'by the conduction velocities of the upstream band points 12, 13, 13'. This creates a feedback problem. To solve this problem, a preliminary course of the guide speed vL is initially set. Assuming that this set course applies, the actual quantities G are determined, which are related to the location xE before the finishing train 1. With the now determined actual variables G, the course of the guide speed vL is determined. The course of the conduction velocity vL determined is in turn herangezo ^¬ gen to determine the actual values G new. In practice it turns out that the approach converges very fast. As a rule, only a few iterations are required - for example, three to five iterations - in order to achieve sufficiently stable results.

In the context of the previous explanations of the present invention, it was assumed that the finishing train 1 has neither an input-side heating device 4 nor inter-frame cooling devices 5. If the heating device 4 and / or the inter-frame cooling devices 5 are present, the operating method according to the invention can be adapted accordingly. The necessary adjustments will be explained below in connection with a single inter-frame cooling device 5. However, the corresponding embodiments are also readily applicable to embodiments of the finishing train 1, which more than one inter-frame cooling device 5 and / or an input-side heater 4, wherein the heater 4 may be provided alternatively or in addition to the inter-frame cooling devices 5 , Thus, assume the finishing mill 1, a single Zvi ^¬ rule scaffold-cooling device 5, for example between the second and the third roll stand 3 according to the depicting ^¬ lung of FIG 1. In this case, the model of the finished Road 1 - this is immediately and readily apparent - are divided into three submodels, which are designated in FIG 12 as Teilmo ^¬ model TMl, partial model TM2 and partial model TM3. The partial model TMI corresponds in its approach a model of a ^¬ ner finishing mill 1, as previously thought, so ei ^¬ nem model of a finishing mill 1 without intermediate stand cooling facilities. It models the behavior of the belt 2 in the finishing train 1 to before the inter-frame cooling device 5. The sub-model TMl receives as input variables the actual size G of a band point 12, 13, 13 'and the leading speed vL and the corresponding Leitgeschwindigkeitsverlauf. The partial model TMI provides as an output an expectation TE size that corresponds to an expected energy content with which the corresponding strip point 12, 13, 13 enters' into the Zwischenge ^¬ Ruest cooling device. 5 The partial model TMI is two ^¬ dimensional, because it has two input variables, namely the actual variable G and the conduction velocity vL. The partial model TM2 models the inter-frame cooling device 5 as such. It receives as input variables supplied by the part ^¬ model TMI expected size TE, the Leitgeschwin ^¬ speed vL with which the band point in question 12, 13, 13 ', the intermediate stand cooling device 5 passes and - as such given - refrigerant quantity M, at which the tape 2 Per

Time unit is applied. The quantity M of cooling liquid per unit time is preferably defined as a function of the material of the tape 2 ^¬ amount which has the Zwischengerüst- cooling device 5 already happened. Alternatively, the amount M of cooling fluid per unit of time may be defined, for example, as a function of the respective band point 12, 13, 13 ', which is just entering the inter-frame cooling device 5. The partial model TM2 thus has three input variables, in contrast to a model of a finishing train 1 without interstand cooling devices. Setting up a corresponding three-dimensional data field for the three-dimensional part Model TM2 may still be possible depending on the computing power available. However, the partial model TM2 is preferably split into two submodels ΤΜ2 ', TM2 ", which are multiplicatively linked to one another, because with sufficient accuracy, a three-dimensional function f, which predetermines an expected variable TA behind the inter-frame cooling device 5 as a function of the expected variable TE the inter-frame cooling device 5, the guide speed vL and the amount M of cooling fluid per unit time, are shown as the product of a two-dimensional function g and a one-dimensional function h, the function g is here of the expected value TE, which is supplied by the submodel TM1, The function h depends only on the quantity M of cooling liquid per unit of time, so it can be used

TA = f (TE, vL, M) = g (TE, vL) ^■ h (M)

Denote this

- TA expectation size for the energy content of the look ^¬ th band point 12, 13, 13 'downstream of the intermediate stand cooling device 5,

TE is the expected size for the energy content of the considered band point 12, 13, 13 'in front of the inter-frame cooling device 5,

- vL the conduction velocity and

M is the amount of cooling fluid applied to the belt 2 per unit of time.

The submodel TM3 has the same structure as the submodel TM1. It models the part of the finishing train 1 which is arranged behind the intermediate stand cooling device 5.

The submodels TM1 to TM3 are connected to each other and concatenated with each other, so that the output variables of the one submodel TM1, TM2 input variables of the next mo- dells TM2, TM3. By concatenating the sub-models TM1 to TM3 the dimensionality of Mo ^¬ dellierungsproblems can already be substantially reduced with each other, namely on the viewing a three-dimensional and two-dimensional two- problems. By splitting the dreidimensiona ^¬ len problem - Keyword partial model TM2 - in a eindimensi ^{¬-dimensional} and two-dimensional function, the complexity can be reduced even further. In particular, by this Redu ^¬ cation of the complexity of the three-dimensional problem, the real-time and online capability is maintained even if the inter-frame cooling devices 5 and / or the heater 4 are present.

When the intermediate stand cooling devices 5 and / or the heating device 4 are available, so they can, provided that the course of the set M is added to the cooling liquid per unit of time, the command variables L * calculated ^¬ the. In a second step, the quantity M can then be varied for each intermediate stand cooling device 5 in order to obtain the expected energy contents of the strip points 12, 13, 13 'as far as possible from the corresponding desired energy contents the band ^¬ points 12, 13, 13 'approach. The determination of the correct quantities M takes place completely analogously to the determination of the correct quantities of cooling liquid for the cooling devices 7 of the cooling section 6.

It is possible for the control computer 8 to control the finishing train 1 without detecting a measured quantity GM which is characteristic of the actual energy content of the band points 12, 13, 13 'behind the finishing train 1. In a preferred embodiment of the present invention, however, takes the control computer 8 - in this case, of course, after the passage of the respective band points 12, 13, 13 'through the finishing line 1 - according to FIG 13 in a step S51 for the corresponding band points 12, 13, 13th 'in each case a corre ^¬ sponding measure GM against. For example, the control computer 8 can ^¬ a corresponding measured temperature value against take, which was detected by the temperature measuring device 15.

Furthermore, the control computer 8 according to FIG 13 in a step S52 for at least a portion of the band points 12, 13, 13 '- preferably for all band points 12, 13, 13' - each have an expectation GE '. In general, the control computer 8 ^¬ determined for each strip point 12, 13, 13 'whose expectation ^¬ size GE', while the respective strip point 12, 13, 13 'the finishing train 1 passes. However, it is alternatively possible for the control computer 8 to determine the corresponding expectation variable GE 'before the respective band point 12, 13, 13' passes through the finishing train 1. Any such identified expectations ^¬ tung size GE 'is characteristic of the energy content of the band for the respective point 12, 13, 13' is at the location xA, to which the desired quantities are G * based expected. The control computer 8 determines the expected quantities GE 'using the Leitgeschwindigkeitsverlaufs, with the jewei ^¬ lige band point 12, 13, 13' actually passes through the finishing train 1.

In the event that the model of the finishing train 1 - regardless of the exact nature of the model of the finishing train 1 - error ^{¬ is} free, the determined in step S52 actual energy contents of the band points 12, 13, 13 'exactly the actual ^¬ union energy content , which are determined by the corresponding measured quantities GM. In many cases, however, the model of finishing train 1 is faulty. The reasons for this can be manifold. For example, the modeling can be easily attached or it may have a sys ^¬ atic error, the model, such as the heat transfer be wrong modeled. In a step S53, the control computer 8 therefore compares the energy content according to the measured variable GM and the energy content according to the corresponding expected variable GE 'with each other. Depending on the ^comparison of step S53, in a step S54 the control computer 8 automatically carries out at least a part of those control points. large L *, which the control computer 8 has already determined at the time of the comparison.

The tracking of the control variables L * in the context of step S54, of course, refers only to those control variables L *, which are indeed already determined at this time, but are still pending execution. The step S54 is thus carried out only for guiding variables L * which were determined for band points 12, 13 which have not yet entered the finishing train 1 at the time of the tracking.

It is possible to track all tracked variables L * immediately in their entirety. However, it is preferred to make a smoother transition. For example, the first tracked guide variable L * can be tracked by 10% of its variation, the second tracked guide variable by 20% of its amendments ^¬ tion, the third tracked guide variable L * by 30% of their Än ^¬ alteration etc .. Alternatively, or in addition to the presence of the step

554, it is possible that the control computer 8 in one step

555 adapted on the basis of the comparison, the determination rule for determining the guiding variables L * as such. This ensures that in the future determined command variables L * that are not nor ^¬ averages at the time of the comparison of the step S53, are determined in an improved manner. The adapting the determination procedure can in particular comprise adapting the model of the finishing train 1 and here insbeson ^¬ particular the heat transfer model.

In particular, if the expectation quantities GE, GE 'are determined by means of the above-mentioned data field, it is possible to carry out the adaptation of the model of the finishing line 1 for the strip 2, which is currently passing through the finishing line 1, in a simplified manner. Because in this case the adaptation can be carried out, for example, in that an offset is added to the actual variables G, before being used as input ^¬ size of the data field. Alternative or additional In addition, the guide speed vL can be scaled by a factor and / or an offset added to it before it is used as input to the data field. Alternatively or additionally, an offset can be added to the expected variable GE, GE 'determined using the data field in each case. In particular, the real-time capability of the operating method according to the invention is maintained in this ver ^¬ simplified type of adaptation of the model of the finishing line first

Within the scope of step S54, it is possible to track all the control variables L * which have already been determined at this time, but have not yet been executed, that is to say, for example, the master variable L * for the (first) band point entering the finishing train 1 next 12. Preferably, the STEU ^¬ errechner 8 only those command variables L resulting from the comparison of step S53 selbsttä ^¬ tig However * after that were determined for (second) strip points 13, which was a minimum distance at the time of tracking of the entrance of the finishing train 1 MIN (see FIG. 14).

Because, as shown in FIG. 14, the operating method according to the invention has a prediction horizon H with regard to the control variable course. The prediction horizon H is determined by the second band point 13, whose master size L * is already determined and which has the greatest distance from the finishing line 1 from the second band points 13 whose guiding variables L * have already been determined. It may be useful if the control computer 8 self based on the comparison act only those command variables L * readjusts that were determined for second strip points 13 which have at the time of the post ^¬ passing from the entrance of the finishing mill 1 the minimum distance MIN. This will be illustrated below in connection with FIG.

According to the illustration of FIG. 7 the strip points PI to P4 have already left the finishing train 1,

- are the band points P5, P6, P7 ... P30 in the Fer ^¬ tigstraße 1,

- The band point P31 next enters the finishing train 1 and

extends the prediction horizon H, starting from

Band P31, to band P35. Based on the actual temperature, for example, the band point P2 in front of the finishing train 1 and on the Leitgeschwindigkeitsverlip, with the band point P2 the finishing train 1 durchlau ^¬ FEN, determines the control computer 8, the temperature for the band point P2 at the output of the finishing train 1 (ie at location xA) is expected. This corresponds to the step S52 of FIG

FIG. 13. The control computer 8 furthermore receives from the temperature measuring device 15 the actual temperature which is measured for the belt point P2. This corresponds to the step S51 of FIG. 13. It is assumed that the comparison of the step S53 gives a deviation. Despite the deviation of the control computer 8 leaves - for example - the already determined Leitgrö ^¬ Shen L * for the band points P31 to P34 unchanged. Based on the comparison of step S53 in step S54, it only traces the reference variable L * of the band point P35. The guiding variables L * for subsequent band points P36, P37 associated with this

Time have not yet determined, determines the control ^¬ computer 8 based on a determination rule, which he

Step S55 adapted based on the comparison of step S53 adap ^¬ .

It may be permissible in individual cases to also change the guiding variables L * of the band points P31 to P34 as well. In this case, however, the change of the corresponding control variables L * does not take place on the basis of the comparison of step S53, but on the basis of a higher-level control intervention which is specified to the control computer 8 by another control device - for example the control device 8 '- or by an operator. As already mentioned, the finishing train 1 is usually downstream of a cooling section 6. The cooling section 6 has Kühlein ^¬ directions 7. Each cooling device has (at least) a control valve 1 'and a number of spray nozzles 7 "assigned to the respective control valve 1'

Control valve 1 'is set, how much cooling liquid is delivered locally to the belt 2. The control valves 1 'rea ^¬ yaw relatively sluggish. Calculated from the time at which a control valve 1 'is controlled with a changed manipulated variable S up to the time at which the changed control affects the band 2, there is a dead time T, which is often ^¬ times in the second range. Dead times of two to five seconds are quite common. Furthermore, the course of the guide speed vL also influences the passage time of the belt points 12, 13, 13 'through the cooling path 6. It is therefore necessary that the control device 8', which performs the control of the cooling devices 7 of the cooling section 6, not only the instantaneous value the guiding speed vL knows, but also their future course. Because only then the controller 8 'of the cooling section 6 can predict response to future upcoming changes in Leitge ^¬ speed vL time. The control device 8 'of the cooling section 6 must therefore use the control variable L * - and also ^¬ future upcoming control variables L * - to determine the set large S for the control valves 1' when the correct amounts of coolant to the "right" places of This applies, of course, in an analogous form even if the control of the cooling section 6 is carried out by the control computer 8.

In the case where inter-frame cooling devices 5 are present, 5 analog dead times occur in the inter-frame cooling devices. Here again, therefore, the course of the control variable in the determination of the manipulated variables S for the inter-frame cooling devices 5 should also be used in order to be able to react in good time to future changes in the guide speed vL. Preferably, therefore, the prediction horizon H according to FIG 14 is at least as the dead time T is preferably larger than the dead time T explained above. Preferably, the prediction horizon H is even greater than the dead time T. If, for example - see FIG. 7 - the dead time T corresponds to the band points P31 to P33, the prediction horizon H should exceed more than two band points extend, for example, according to the illustration of FIG 7 over four band points.

For essentially the same reasons supported the minimum distance should least MIN within which the tracking of the SEN Leitgrö ^¬ L * is suppressed at least as large as the dead time to be T, 7 for example, be in accordance with FIG three strip points.

The guiding variables L * are determined point by point for the individual band points 12, 13. In order to determine a continuous conduction velocity course, step S4 is designed in the form of a step S61 according to FIG. In step S61, the control computer 8 chains the determined control variables L * to each other by a spline, so that the chaining results in a control variable course that is continuous and differentiable. Corresponding to this, the so-defined Leitgeschwindigkeitsverlauf is continuous and differen ^¬ zierbar. Alternatively to step S61, a step S62 could be present. In step S62, the control computer 8 determines the corresponding point-specific guide speeds vL on the basis of the guide variables L * determined on a point-by-point basis. In this case, the control computer 8 chains the corresponding guide speeds vL by a spline to each other, so that the chaining results in a continuous and differentiable conduction velocity course.

Steps S61 and S62 are alternative to each other. Although they are therefore both shown in FIG 15, but both drawn only ge ^¬ dashed. The operating method for the Fer ^¬ tigstraße 1 described above provides - first - conduction velocities v L, until the last strip point 13 of the tape 2 in the finishing train

1 is running. However, the conduction velocity vL must be finiert de-, as long as at least one band point 12, 13 is in the finishing train 1, so even if no further strip points 12, 13 more einlau ^¬ fen into the finishing train. 1 It is easily possible to expand the inventive ago ^¬ go as accordingly. It is only necessary within the control computer 8 in addition to the band points 12, 13, 13 'for the physically existing band

2 virtual band points to be considered, which are attached to the erstge ^¬ called band points. For these virtuel ^¬ len strip points a corresponding command variable L * telt ermit-. The virtual band points, however, is neither one

Actual quantity G is still assigned a desired value G *, so that the virtual band points themselves do not contribute to the determination of the corresponding guiding quantities L *.

As part of the explanation of the present invention, the Leitgröße L * was further explained in each case in conjunction with the band points 12, 13, which run into the finishing train 1 at certain times. However, this is not to be understood as meaning that the corresponding guiding variables L * are permanently assigned to the corresponding band points 12, 13. For the corresponding guide variable L * global effect on all the tape 2. It is crucial, therefore, only the assignment of the respective command variable L * at a given time, said time is defined by the fact that the corresponding band ^¬ point 12, 13 at this time in the finishing train 1 on ^¬ runs.

The present invention has many advantages. Insbeson ^¬ wider it possible one or Leitgrößen- Leitgeschwin- digkeitsverlauf predict that will later actually met during operation of the finishing line first Associated with this results in improved accuracy in maintaining the nominal energy content at the outlet of the finishing train 1 and In addition, an improved - even considerably improved ^¬ serte - accuracy with which the cooling section 6 controlled ^¬ who can. Thus, it is possible, for example, both a final ^¬ Walztemperatur (at the outlet of the finishing train 1) and a reel temperature (at the outlet of the cooling section 6) comply with high accuracy.

The above description is only for explanation of the present invention. The scope of the present invention, however, is intended to be determined solely by the appended claims.

Claims

claims

1. Operating method for a finishing train (1) for rolling a strip (2),

- wherein a control computer (8) for the finishing train (1) later ^¬ at a time at which a first band point (12) of the band (2) is still before the finishing train (1) be ^¬ , for the first band point ( 12), a number of second band points (13) and a number of third band points (13 ') of the band (2) each have an actual size (G) and a nominal size (G *) are known,

- For each band point (12, 13, 13 '), the respective Ist ^¬ size (G) for the actual energy content of the respective band ^¬ point (12, 13, 13') and the respective target size (G *) for the target Energy content of the respective band point (12, 13,

13 ') are characteristic,

- For each band point (12, 13, 13 '), the respective

Actual size (G) is related to a location (xE) in front of the finishing train (1) and the respective nominal value (G *) is related to a location (xA) behind the finishing train (1),

- wherein the second band points (13) after the first band point (12) and the third band points (13 ') before the first band ^¬ point (12) run into the finishing train (1),

- wherein the control computer (8) before the arrival of the first band point (12) in the finishing line (1) for the first

Band point (12) and at least a portion of the second Band ^¬ points (1) based on a specific for each band point (12, 13) specific determination rule in each case a Leit ^¬ size (L *) determined

- wherein the control computer (8) based on the determined for the respective band point (12, 13) Leitgröße (L *) each determines a conduction velocity (vL) and the finishing train (1) at the time of entry of the respective band point (12, 13) in operates the finishing train (1) at the respective master speed (vL) and

Eintre- wherein for the respective command variable (L *) in whose valuation ^¬ lung provision, the actual variable (G) and the target quantity (G *) of the at that time in each case into the finishing train (1) - border point (12, 13) and the actual size (G) and the target size (G *) at least one at this time already in the finishing line (1) occurred band point (12, 13) received.

2. Operating method according to claim 1,

characterized ,

- that the control computer (8) determines each of the guide variables (L *) on the basis of a plurality of individual line sizes (1 *),

- that each Einzelleitgröße (1 *) in each case on one of the band ^¬ points (12, 13, 13 ') is related, the actual and nominal size (G, G *) received in the determination of the respective guide variable (L *),

- That the control computer (8) for each band point (12, 13, 13 ') whose Einzelleitgröße (1 *) determined such that ei ^¬ ne respective expected size (GE) with the corresponding ^¬ the target size (G *) matches, and

- that the respective expected size (GE) is characteristic of an expected ^¬ th energy content which the respective strip point (12, 13, 13 ') at the location (x A) downstream of the finishing ^¬ road (1) to which the respective corresponding target size (G *), would assume if the control computer (8) the finishing train (1) during the entire passage of the respective band point (12, 13, 13 ') through the finished street (1) with one with the Einzelleitgröße ( 1 *) would operate corres ^¬ ponding conduction velocity (vL).

3. Operating method according to claim 1,

characterized ,

in that the control computer (8) determines, for each band point (12, 13) for which it determines its master variable (L *),

- Based on the actual quantities (G), which are included in the determination of the guide variable (L *) for the respective band point (12, 13), an effective actual size (G ') and on the basis of the target values (G *), which in the determination of the command variable (L *) for each ^¬ weiligen strip point (12, 13) received, determines an effective target size ^¬ (G '*), - an expectation size (GE) is determined, which is characteristic of an expected ^¬ th energy content which the respective strip point (12, 13) at the location (x A) behind the finishing train (1) on which the effective target size (G '*) if the control computer (8) assumed the finishing line

(1) during the entire passage of the respective band ^¬ point (12, 13) through the finishing train (1) with a Leitgröße (L *) for the respective band point (12, 13) corresponding Leitgeschwindigkeit (vL) would operate , and

- The Leitgröße (L *) determined such that the expected ^¬ size at the location (xA) behind the finishing train (1) to which the effective target size (G '*) based, the effective target size (G' *) ,

4. Operating method according to claim 1,

characterized ,

- That the control computer (8) for determining the control variables

(L *) initially sets the guiding values (L *) as provisional values,

- that the control computer (8) for the first band point (12) and at least a part of the second and third band points (13, 13 ') determines a respective expectation variable (GE),

- Each expectation variable (GE) is characteristic for an expected energy content, the respective band point

(12, 13, 13 ') at the location (x A) behind the finishing train (1) to which the respective corresponding target quantity (G *) is bezo ^¬ gene would assume, when the control computer (8) the finishing train (1) during the entire passage of the respective band point (12, 13, 13 ') would operate through the finishing train (1) with guide speeds (vL), which correspond to the set guide values (L *), and

- That the control computer (8) the applied guide variables (L *) varies, so that an objective function (Z) is optimized, in which the amounts of the differences of the expectation quantities (GE) received by the corresponding target values (G *).

5. Operating method according to claim 4,

characterized ,

that in the target function (Z) additionally enters a Strafterm, by means of which changes in the conduction velocity (vL) are punished.

6. Operating method according to one of claims 2 to 5, d a d u c h e c e n e c e n e,

- That the control computer (8) in advance creates a data field in which the control computer (8) for a variety of possible

Conduction velocities (VL) deposited and possible actual values (G) which for each possible actual value (G) at the time jewei ^¬ possible conduction velocity (VL) resulting expectations ^¬ tung size (GE), and

- That the control computer (8) the guiding variables (L *) for the

Band points (12, 13) using the data field ermit ^¬ determined.

7. Operating method according to claim 6,

characterized ,

that the control computer (8)

- determines at least for a part of the band points (12, 13, 13 ') a respective expectation variable (GE') which is characteristic of an expected energy content which for the respective band point (12, 13, 13 ') at the location (xA ) behind the finishing train (1), to which the respective corresponding target size (G *) is related, due to the Leitgeschwindig ^¬ speeds (vL), with which the control computer (8) the finishing train (1) during the entire passage of the respective band point (12, 13, 13 ') through the finishing train (1) ^{¬ is driven} , is expected

- After the passage of the respective band point (12, 13, 13 ') through the finishing train (1) receives a measured variable (GM), which for an actual energy content of the respective band point (12, 13, 13') on the location (xA) downstream of the finishing ^¬ road (1), is related to the target, the corresponding size (G *) is characteristic, - Based on a comparison of the expected energy content with the actual energy content automatically a model of Fertigstraße ^¬ (1) adapted and

adapted the model of the finishing train (1) by adding an offset to the actual quantities (G) when using the data field, scaling the guide speeds (vL) with a scaling factor and / or adding an offset to them and / or the using the data field determined expectation quantities (GE, GE ') adds an offset.

8. Operating method according to one of the above claims,

characterized ,

that for each reference variable (L *) in their determination the actual size (G) and the target size (G *) of the already in the finishing train

(1) occurred band points (12, 13, 13 ') only if these band points (12, 13, 13') at the time for which the respective guide variable (L *) is determined, not yet from the finishing train ( 1) have leaked.

9. Operating method according to one of the above claims,

characterized ,

that the control computer (8) at least for a part of the band points (12, 13, 13 ')

'Determined, which is characteristic of egg ^¬ NEN expected energy content that is for each strip point (12, 13, 13 a respective expected size (GE)') at the location (x A) behind the finishing train (1), on each - Correspondie ^¬ ing setpoint size (G *) is based, due to the Leitgeschwindigkeiten (vL), with which the control computer (8) the

Finishing train (1) during the entire run of the belt jewei ^¬ time point (12, 13, 13 ') by the finishing mill (1) operates, it is expected

- After the passage of the respective band point (12, 13, 13 ') through the finishing train (1) receives a measured variable (GM), for an actual energy content of the respective band ^¬ point (12, 13, 13') on the Place (xA) behind the finished Street (1), to which the corresponding target size (G *) is related, is characteristic, and

- Based on a comparison of the expected energy content with the actual energy content automatically nachzuführen at least a portion of the already determined guide values (L *).

10. Operating method according to claim 9,

characterized ,

that the control computer (8) on the basis of the comparison automatically only those guiding variables (L *) tracks that were determined for band points (12, 13), which have a minimum distance (MIN) at the time of tracking from the entrance of the finishing train (1).

11. Operating method according to claim 10,

characterized ,

that the control computer (8) or another control device (8 ') uses the determined control variables (L *) to determine at least one further control variable (S),

- that the further control value (S) by a dead time (T) deferrers ^¬ siege and locally on the belt (2) and acts

- That the minimum distance (MIN) is determined such that ei ^¬ ne with the minimum distance (MIN) corresponding Zeitdif ^¬ ference is at least as large as the dead time (T).

12. Operating method according to one of the above claims, d a d u c h e c e n e c e n e,

- That the first band point (12) and the part of the second

Strip points (13) for which the respective command variable (L *) prior to entry of the first tape point (12) was determined in the Fertigstra ^¬ SSE (1), a prediction horizon (H) speak ent ^¬, the at least as large as the dead time (T) is.

13. Operating method according to one of the above claims, characterized in that:

in that the control computer (8) links the determined guide variables (L *) or the corresponding guide speeds (vL) to one another by a spline, so that a conduction velocity profile resulting from the concatenation is continuous and differentiable.

14. Operating method according to one of the above claims, d a d u c h e c e n e c e n e,

in that the control computer (8) carries out the determination of the control variables (L *) as part of a precalculation online or in real time.

15. Computer program comprising machine code (11) which can be processed directly by a control computer (8) for a finishing train (1) for rolling a belt (2) and whose execution by the control computer (8) causes the control computer (8 ) operates the finishing train (1) according to an operating method with all the steps of an operating method according to one of the above claims.

16. control computer for a finishing train (1) for rolling a belt (2),

characterized ,

in that the control computer is designed such that it operates the finishing train (1) according to an operating method with all the steps of an operating method according to one of claims 1 to 14.

17. Finishing line for rolling a strip (2),

characterized ,

that the finishing train is equipped with a control computer (8) according to ^¬ demanding sixteenth