JPH0513731B2 - - Google Patents

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention provides a rolling mill roll eccentricity control method for preventing plate thickness variation due to roll eccentricity in plate thickness control of a hot or cold hydraulic rolling mill. Regarding.

[Conventional technology]

Roll eccentricity control of a rolling mill detects the amount of roll eccentricity by rotating the roll in a kiss-roll state off-line during non-rolling, but the detected amount of roll eccentricity is calculated from the mill constant and plasticity coefficient set at the time of detection. is a fixed value.

[Problem that the invention seeks to solve]

However, the mill constant and plasticity coefficient of the plate during actual rolling change depending on rolling conditions such as rolling force, plate width, and rolling material. Therefore, the amplitude value of the roll eccentricity control signal and the amplitude of the actual roll eccentricity vary. The value may differ greatly, and this error causes a control error. As a means for correcting such a control error, no particular online correction is performed other than adjusting the gain of the output value at the time of initial adjustment. Furthermore, it has the disadvantage that it is powerless against changes in characteristics over time, and that re-adjusting it periodically is labor-intensive and tends to be complicated.

The present invention has been made in view of the above points, and the roll eccentricity control signal is provided with an output coefficient for amplitude adjustment, and the amount corresponding to the control error is calculated from the roll eccentricity control signal and the rolling force detection signal using a Fourier series. The amplitude of the roll eccentricity control signal is determined by detecting the polarity from the rolling force detection signal and the cylinder displacement signal, and by constantly correcting the output coefficient online using these values. It is an object of the present invention to provide a more effective method for controlling roll eccentricity of a rolling mill by correcting the amount of roll eccentricity.

[Means and actions for solving problems]

During roll eccentricity control during rolling, the amplitude ratio of the roll eccentricity control signal and the rolling force detection signal is determined, and on the other hand, a coefficient is determined from the cross-correlation value between the rolling force detection signal and the cylinder displacement signal to determine whether the roll eccentricity control is excessive or insufficient. are calculated and the amplitude value of the roll eccentricity is always corrected online using these values.

〔Example〕

An embodiment of the present invention will be described below with reference to the drawings. In Figure 1, 1 is a work roll, 2a
2b is the upper back-up roll, 2b is the lower back-up roll, 3 is the material to be rolled, and 4 is the rolling force detector. Reference numerals 5 and 6 indicate rotational position detectors (pulse generators) for detecting the rotational angles of the backup rolls 2a and 2b. Reference numeral 7 denotes a reduction cylinder, which has a reduction position detector 8 attached to its movable part. The output of this reduction position detector 8 is fed back to the cylinder 7 through computing units 9, 10 and a servo valve 11, thereby forming a position control system. Further, the output of the rolling force detector 4 passes through a computing unit 12 and a converter 13 and enters a computing unit 9, forming a so-called BISRA type automatic plate thickness control system. Further, the output of the pressure reduction force detector 4 is passed through the calculator 14 to the sampler 1.
5, A/D (analog-digital) converter 16
The digital computer 17 is then entered. Then, the sampler 1 is
At the same time as timing 5 is taken, the counter 18 is counted and the rotation angle of the upper roll 2a is detected.
This counter 18 is reset by the output of the digital computer 17. Furthermore, the output of this pulse generator 5 is input to a digital computer 17, and in synchronization with this pulse, an A/D converter 16 is input.
The output is read into the digital computer 17. 19 is a pulse oscillator, and 20 is a counter. The counter 20 is set and reset by one pulse per revolution emitted from the pulse generator 5, thereby detecting the time for one rotation of the upper roll 2a. The output of this counter 20 is input to the digital computer 17, and is used by the digital computer 17 to calculate the rotational frequency and compensate for delays in the control system. Further, 21 is a counter that counts the output pulses of the pulse generator 6 provided on the lower roll 2b side and detects the rotation angle of the lower roll 2b, and 22 is an analog value for measuring the roll eccentricity compensation amount created by the digital computer 17. D/A (digital-analog) that is converted into and input to the arithmetic unit 12
It is a converter. A switch 23 is interposed between the D/A converter 22 and the arithmetic unit 12. This switch 23 turns the roll eccentricity control on and off, and is operated from the console 24. This console 24 can also send signals to the digital computer 17. In addition, the D/A converter 22
The roll eccentricity control signal that is the output of the roll eccentricity control signal, the roll-down force detection signal that is the output of the roll-down force detector 4, and the cylinder displacement signal that is the output of the roll-down position detector 8 are input to the output coefficient correction calculator 25. This calculator 25 constantly calculates a correction value for the output coefficient during roll eccentricity control, and outputs the result to the digital computer 17. Further, 26 and 27 are counters, the counter 26 counts the pulses from the pulse generator 5, and the counter 26 counts the pulses from the pulse generator 6.
It is used to detect the difference in diameter between the upper and lower rolls 2a and 2b by continuously counting several to several tens of revolutions each.

Next, the operation of the present invention configured as described above will be explained.

First, the output coefficient α, which is a feature of the present application, will be described. This is achieved by introducing an output coefficient α between the roll eccentricity detection signal R(t) and the roll eccentricity control signal S(t).

S(t)=α·R(t) (1) The role of this output coefficient α is to correct the amplitude of the roll eccentricity control signal. Now, the relationship between the roll eccentricity ΔR and the rolling force fluctuation ΔF is as follows: the mill constant of the rolling mill is M,
If the plastic modulus of the rolled material is Q, it can be found as ΔF=MQ/M+QΔR...(2).

Therefore, the amount of roll eccentricity is determined by detecting the rolling force fluctuation ΔF and performing analytical calculations on the results. Here, if the values of M and Q can be grasped accurately, the amplitude of the roll eccentricity can be determined correctly; however,
In reality, the value changes due to differences in rolling conditions such as rolling material, rolling force, and strip width, and these differences become an error factor in roll eccentricity detection calculations.

Next, a specific method will be described. First, the switch 23 is turned off using the console 24.
In this way, one rotation of the back-up roll is approximated by a first-order Fourier series for the roll eccentricity control signal S(t) detected offline.

If the rotational angular frequency of the back-up roll is ω, then S(t)≒a _s1 sin(ωt+ _s1 ) ……(3) Here, a _s1 =√ _s1 ² + _s1 ² (amplitude of first-order component) _s1 = tan ^-1 B _s1 /A _s1 (Phase of first-order component) However, A _s1 = 2/T∫ ^T ₀ a _s1 sin (ωt + _s1 ) sinωt dt B _s1 = 2/T∫ ^T ₀ a _s1 sin (ωt + _s1 ) cotωt dt Similarly, the rolling force detection signal F(t) output from the rolling force detector 4 is expanded into a first-order Fourier series.

F(t)≒a _f1 sin(ωt+ _f1 ) ...(4) Here, a _f1 = √ _f1 ² + _f2 ² (amplitude of first-order component) _f1 = tan ^-1 B _f1 /A _f1 (first-order component However, A _f1 = 2/T∫ ^T ₀ a _f1 sin (ωt + _f1 ) sinωt dt B _f1 = 2/T∫ ^T ₀ a _f1 sin (ωt + _f1 ) cosωt dt Therefore, the amplitude of the control signal is , |S(t)|=√ _s1 ² + _s1 ² ...(5) Moreover, the amplitude of the rolling pressure fluctuation is |F(t)|=√ _f1 ² + _f1 ² ...(6).

Here, the amplitude ratio of the control signal and the rolling force detection signal is ｜F(t)｜／｜S(t)｜ …(7) think of.

This value indicates the error rate of roll eccentricity control. Furthermore, the polarity of this control error, that is, its excess or deficiency, is determined by taking a cross-correlation between the rolling force detection signal F(t) and the cylinder position displacement signal H(t) of the rolling cylinder operated by the roll eccentricity control signal, etc. Discrimination is made by

Rolling force variation ΔF for one rotation of back up roll
(t) is proportional to the difference between the roll eccentricity R'(t) and the cylinder position displacement, so it can be expressed as ΔF(t)=K{R'(t)−H(t)}...(8) Ru.

Here, the ratio between the roll eccentricity detection signal R(t) and the actual roll eccentricity R'(t) is assumed to be β.

R'(t)=β・R(t)...(9) On the other hand, the cylinder position displacement for one rotation of the backup roll is H(t)≒S(t)=αR(t).

Therefore, if we find the cross-correlation between ΔF(t) and H(t), we get
From formulas (8) and (9), ∫ ^T ₀ Δ(t)・H(t)dt =K・(β−α)・α∫ ^T ₀ {R(t)} ² dt ……(10) This formula It can be seen from the above that roll eccentricity control can be performed reliably by making the value of α match the value of β. Therefore, by checking the sign of equation (10), it is determined whether the roll eccentricity control output is excessive or insufficient. If equation (10) has a positive value, R'(t)>S(t) because β>α, indicating that the roll eccentricity control signal is smaller than the actual amount of roll eccentricity. Therefore, it is necessary to increase the value of α. Conversely, if equation (10) has a negative value, β<α, so R′(t)
<S(t), indicating that the roll eccentricity control signal is larger than the actual roll eccentricity. Therefore, it is necessary to reduce the value of α.

From the above results, as a correction term for the output coefficient α, ε×｜F(t)｜／｜S(t)｜……(11) will be introduced.

Here, ε is a polar variable whose value in equation (10) is +1 when it has a positive value and -1 when it has a negative value.

Therefore, the output coefficient correction formula α _o = α _o-1 + ε× | F _o-1 (t) | / | S _o-1 (t) |...
(12) α _o ; Current output coefficient value α _o-1 ; Previous output coefficient value | F _o-1 (t) | ; Previous rolling force detection signal amplitude value for one rotation of the back-up roll | S _{o- 1} (t) | ; Roll eccentricity control signal amplitude value for one rotation of the previous back-up roll. Then, the correction shown in equation (12) above is performed online by the output coefficient correction calculator 25. That is, the roll eccentricity control signal that is the output of the D/A converter 22, the rolling force detection signal that is the output of the rolling force detector 4, and the cylinder displacement signal that is the output of the rolling position detector 8 are corrected by the output coefficient. This is done by inputting the data into the computer 25. This calculator 25 constantly calculates a correction value for the output coefficient during roll eccentricity control, and outputs the result to the digital computer 17. In this way, during roll eccentricity control, the amplitude correction amount of the roll eccentricity control signal is calculated and corrected using the roll eccentricity control signal, rolling force detection signal, and cylinder position displacement signal. Even if the amplitude of the quantity is different from the actual one, it can be automatically corrected.

〔Effect of the invention〕

As detailed above, according to the present invention, even if the amplitude of the roll eccentricity detected off-line is different from the actual one, it is automatically corrected, thereby realizing highly accurate control and making it even more effective. It is possible to provide a method for controlling roll eccentricity of a rolling mill.

[Brief explanation of the drawing]

FIG. 1 is a diagram showing the configuration of an embodiment of the present invention. 1... Upper and lower work rolls, 2a... Upper back up roll, 2b... Lower back up roll, 4... Down force detector, 5, 6... Pulse generator, 8... Down position detector ,9,10,1
2, 14... Arithmetic unit, 13... Converter, 15...
Sampler, 16...A/D converter.

Claims (1)

[Claims] 1 Regarding the roll eccentricity control device of a hydraulic rolling mill, the off-line roll eccentricity signal component is extracted for each back-up roll arranged oppositely, and the extracted eccentricity amount component is correlated to the roll rotation angle and calculated. In a roll eccentricity control method for a rolling mill, the stored contents are synchronized with and synthesized with each back-up roll eccentricity during rolling, and the output is used as a roll eccentricity control signal. An output coefficient for amplitude adjustment is provided, and the amount corresponding to the control error is detected from the roll eccentricity control signal and the rolling force detection signal by performing calculations using a Fourier series, and the rolling force detection signal and the cylinder displacement signal are By detecting the polarity from and constantly correcting the output coefficient online with those values,
A roll eccentricity control method for a rolling mill, comprising correcting the amplitude of a roll eccentricity control signal.