TECHNICAL FIELD
[0001] The present invention relates to a method and device for controlling sizing mill
of pipes or tubes, and in particular, to a method and device for controlling a sizing
mill capable of suppressing defective thickness of a front or rear end (longitudinal
end) of a pipe or tube rolled by the sizing mill effectively.
BACKGROUND ART
[0002] Generally, in a step of finishing the outer diameter of a pipe or tube so as to have
a predetermined value, there is used a sizing mill (sizer, stretch reducer or the
like) comprising a plurality of stands each of which is provided with two or three
grooved rolls (hereinafter referred to as rolling rolls). A sizing mill applies a
tensile force in a pipe or tube axis direction of the pipe or tube being rolled by
giving a difference to the circumferential speeds of the rolling rolls between adjacent
stands to thereby control the thickness of a pipe or tube.
[0003] In a case of rolling a front or rear end of a pipe or tube, the tensile force in
the pipe or tube axis direction is not applied sufficiently comparing with the case
of rolling the intermediate portion of the pipe or tube, whereby there is caused a
phenomenon where the thickness of the front or rear end of the pipe or tube is thicker
than the thickness of the pipe or tube intermediate portion. Therefore, the front
or rear end of the pipe or tube is cut off as being a part of defective dimension,
which result in a lower yield.
[0004] In order to prevent such a lower yield due to an increase in the thickness of the
front or rear end of the pipe, there is considered a method of controlling the circumferential
speed (specifically, rotational speed) of rolling rolls when rolling the front or
rear end of the pipe or tube(reducing the rotational speed when rolling the front
or rear end of the pipe or tube). However, in order to activate such a method effectively,
it is important that the timing of starting control of the rotational speed of rolling
rolls must be accurate.
[0005] As shown in Fig. 1, assuming that, with respect to a timing when a front or rear
end of a pipe or tube 1 is detected by a pipe end detector 8 such as an HMD, a period
until the front or rear end of the pipe or tube reaches a first stand is T
0, and a period from the front or rear end of the pipe or tube leaves an (i-1)
th stand until it reaches an i
th stand (i ≥ 2) is T
i-1, after the period T
0 has passed from the timing the front or rear end of the pipe or tube 1 is detected
by the pipe end detector 8, a rotational speed control of a rolling roll 2 provided
to the first stand must be started, and after the period of T
0 + ΣT
j (j = 1 to i-1, i ≥ 2) has passed, the rotational speed control of a rolling roll
21 provided to the i
th stand must be started.
[0006] In order to grasp the period T
0 and the period T
i-1 accurately, a high precision sensor such as a load measuring device may be disposed
at each stand, but it requires enormous capital investment. The present invention
is based on the premise of using a method of predicting the time period T
0 and the time period T
i-1 (i ≥ 2) without using a high precision sensor.
[0007] However, due to various factors, there is caused an error between the predicted periods
of the period T
0 and the period T
i-1 (i ≥ 2) and the period that the front or rear end of the pipe or tube 1 actually
reaches (hereinafter, referred to as "prediction error"). Therefore, there is a problem
that an accurate control is impossible if the rotational speed control of the rolling
roll 21 is started only based on the estimated periods, whereby an increase in the
thickness of the front or rear end of the pipe or tube cannot be controlled effectively.
[0008] In
Japanese Patent No. 2541311, there is proposed a method in which a rolling torque is calculated from the driving
current and the rotational speed of a motor for driving rolling rolls, and a timing
when an front or rear end of the pipe or tube actually reaches each stand (a timing
when a front end bites into rolling rolls or rear end passes out of rolling rolls)
is detected from the fluctuating state of the calculated rolling torque, and the rotational
speed control start timing (the timing when the rotational speed control starts) of
the rolling rolls for the next rolling of the pipe or tube is corrected such that
the prediction error in each stand comes into a predetermined range.
[0009] In the method described in the above-mentioned publication, there is no analysis
for a factor causing a prediction error at the rotational speed control start timing
of rolling rolls which have been set with respect to each stand. Based on the prediction
error calculated for each stand, the rotational speed control start timing of rolling
rolls in the next rolling of the pipe or tube is corrected uniformly. Even if the
timing when the front or rear end of the pipe or tube actually reaches is detected
accurately, it is impossible to correct the rotational speed control start timing
of rolling rolls for the pipe or tube to be rolled next, in the state where the prediction
error varies at random.
[0010] In other words, a prediction error for rolling one pipe or tube is not always the
same for the next pipe or tube to be rolled. Accordingly, it is impossible to correct
the rotational speed control start timing of rolling rolls accurately for the next
pipe or tube to be rolled. As a result, it is impossible to suppress the defective
thickness of the front or rear end of the pipe or tube effectively.
[0011] Further, the pattern of the thickness fluctuation of a front or rear end of the pipe
or tube is not always the same, so defective thickness of a front or rear end of the
pipe or tube cannot be suppressed fundamentally only by grasping the timing when the
front or rear end of the pipe or tube reaches each stand accurately to thereby control
the rotational speed of the rolling rolls.
DISCLOSURE OF THE INVENTION
[0012] The present invention has been developed to solve such a problem in the conventional
art. An object of the present invention is to provide a method and device for controlling
a sizing mill, capable of suppressing defective thickness of a front or rear end of
a pipe or tube to be rolled by the sizing mill effectively.
[0013] In order to solve the above-described problem, the present invention provides a method
for controlling sizing mill of a pipe or tube, wherein a rotational speed control
start timing of rolling rolls set with respect to a predetermined stand of a sizing
mill is corrected based on a thickness measured value of a front or rear end of the
pipe or tube measured at an output side of the sizing mill.
[0014] According to such an invention, the rotational speed control start timing of rolling
rolls set with respect to a predetermined stand is corrected based on the thickness
measured values of the front or rear end of the pipe or tube measured at the output
side of the sizing mill, whereby it is expected to correct the rotational speed control
start timing of the rolling rolls appropriately according to the actual thickness
fluctuation of the front or rear end of the pipe or tube, so it is possible to effectively
suppress defective thickness of the front or rear end of the pipe or tube rolled by
the sizing mill.
[0015] It is preferable that the method for controlling sizing mill comprises: a first step
of calculating a prediction error between the rotational speed control start timing
of the rolling rolls set with respect to the predetermined stand and a timing when
the front or rear end of the pipe or tube actually reaches the predetermined stand;
and a second step of correcting the rotational speed control start timing of the rolling
rolls set with respect to the predetermined stand, based on the prediction error calculated
and the thickness measured value of the front or rear end of the pipe or tube measured
at the output side of the sizing mill.
[0016] According to such a preferable configuration, the rotational speed control start
timing of the rolling rolls is corrected by using not only the thickness of the front
or rear end of the pipe or tube actually measured but also a prediction error between
the rotational speed control start timing of the rolling rolls set with respect to
the predetermined stand and the timing when the front or rear end of the pipe or tube
actually reaches the predetermined stand. Therefore, it is expected to correct the
rotational speed control start timing of the rolling rolls appropriately and accurately,
so it is possible to suppress defective thickness of the front or rear end of the
pipe or tube rolled by the sizing mill, more effectively.
[0017] In order to solve the problem described above, the inventors of the present invention
intensively studied about factors causing a prediction error between the rotational
speed control start timing of the rolling rolls set with respect to each stand and
the timing when the front or rear end of the pipe or tube actually reached the stand.
As a result, it was anticipated that a prediction error for a predicted period T
0 until the front or rear end of the pipe or tube reaches a first stand, and a prediction
error for the predicted period ΣT
j (j = 1 to i-1, i ≥ 2) until the front or rear end of the pipe or tube reaches the
i
th stand from the first stand were caused due to different factors.
[0018] That is, a prediction error for the predicted period T
0 until the front or rear end of the pipe or tube reaches the first stand is caused
due to a difference between the predicted carried speed and the actual carried speed
of the pipe or tube due to the cross-sectional shape and a bend of the pipe or tube
as well as abrasion of the carrying conveyor. The prediction error for the predicted
period T
0 is a component included in common in the prediction error between the rotational
speed control start timing of the rolling rolls set with respect to each stand and
the timing when the front or rear end of the pipe or tube actually reaches the stand.
On the other hand, a prediction error for the predicted period ∑T
j (j = 1 to i-1, i ≥ 2) until the front or rear end of the pipe or tube reaches the
i
th stand from the first stand is caused by fluctuations in the rolling state of each
stand (for example, the extension rate of the pipe or tube in sizing mill is different
from predicted values due to fluctuation factors on the sizing mill side such as actual
rotational speed of rolling rolls being different from the set value and abrasion
of the rolling rolls and surface irregularity of the rolling rolls being not uniform,
as well as fluctuation factors on the pipe or tube side such as material, dimensions
and temperature of the pipe or tube).
[0019] Next, the inventors of the present invention calculated through experimentations
a prediction error between the rotational speed control start timing of the rolling
rolls set with respect to each stand and the timing when the front or rear end of
the pipe or tube actually reached the stand, and divided the calculated prediction
error into a prediction error for the predicted period T
0 until the front or rear end of the pipe or tube reaches the first stand (hereinafter,
referred to as "prediction error component of a factor outside the stand" as appropriate),
and a prediction error for the predicted period ΣT
j (j = 1 to i-1, i ≥ 2) until the front or rear end of the pipe or tube reaches the
i
th stand from the first stand (hereinafter, referred to as "prediction error component
of a factor inside the stand" as appropriate), whereby analyzed the dispersion states
of the both prediction error components. As a result, the inventors found that a tendency
shown in Fig. 2 was obtained. Note that in Fig. 2, the horizontal axis shows the sequence
of pipes or tubes rolled, and the vertical axis shows the ratio of each prediction
error component to the predicted period.
[0020] Hereinafter, Fig. 2 will be described in detail. First, for each rolled pipe or tube,
as shown in Fig. 3, the stand number was plotted in the horizontal axis X, and a prediction
error (prediction error between the rotational speed control start timing of the rolling
rolls set with respect to each stand and the timing when the front or rear end of
the pipe or tube actually reaches the stand) in each stand calculated as described
above was plotted in the vertical axis Y. Then, based on the plotted data (i, Yi),
a primary regression equation ofY in which X was a variable was calculated, and a
Y section of the primary regression equation (value at a point intersecting the vertical
axis) was defined as a prediction error component T
0' of the predicted period T
0 until a front or rear end of the pipe or tube reaches the first stand, and a value
obtained by subtracting the prediction error component T
0' from the Y coordinate Y
i' in the primary regression equation, where X = i, was defined as a prediction error
component ΣT'
j for the predicted period ΣT
j (j = 1 to i-1, i ≥2) until the front or rear end of the pipe or tube reached the
i
th stand from the first stand. In Fig. 2, "○" indicates a value obtained by dividing
the prediction error component T
0' calculated as described above by the predicted period T
0, and "□" indicates a value obtained by dividing the prediction error component ΣT
j' calculated as described above by the predicted period ΣT
j' respectively. Note that data shown in Fig. 2 indicates cases where pipes or tubes
of different materials were rolled (in a chance 1, A1 indicates that carbon steel
is rolled and A2 indicates that 2Cr steel is rolled, and in a chance 2, B1 indicates
that low alloy steel is rolled and B2 indicates that carbon steel is rolled) in two
manufacturing chances (chance 1 and chance 2) where the dates of rolling the pipes
or tubes were different (accordingly, various settings of the rolling mill are often
different).
[0021] As shown in Fig. 2, the dispersion tendency of the prediction error components T
0' for the predicted period T
0 (prediction error components of factors outside the stand) does not change a lot
even though the manufacturing chance and the material of the pipe or tube differ.
However, the dispersion tendency of the prediction error components ΣT
j' for the predicted period ΣT
j (prediction error components of factors inside the stand) changes when the material
of the pipe or tube differs. This is due to the fact that the generating factors of
the prediction error components of factors outside the stand and the prediction error
components of factors inside the stand are different as described above.
[0022] As described above, the dispersion tendencies in the prediction error components
of factors outside the stand and the prediction error components of factors inside
the stand are different since their generation factors are different. Therefore, if
the both prediction error components are divided and are provided separately (e.g.,
while being weighted differently) for correcting the rotational speed control start
timing of the rolling rolls, it is expected that the rotational speed control start
timing of the rolling rolls can be corrected appropriately even in a state where the
prediction error between the rotational speed control start timing of the rolling
rolls set with respect to each stand and the timing when a front or rear end of the
pipe or tube actually reaches the stand varies at random, due to the fluctuations
of the generation factors.
[0023] Accordingly, it is preferable that the second step includes the steps of: extracting
a first prediction error component until the front or rear end of the pipe or tube
reaches a first stand and a second prediction error component after the front or rear
end of the pipe or tube reaches the first stand, from the prediction error calculated;
applying a first weight to the first prediction error component extracted, and based
on the first prediction error component applied with the first weight, correcting
the rotational speed control start timing of the rolling rolls set with respect to
the predetermined stand; applying a second weight to the second prediction error component
extracted, and based on the second prediction error component applied with the second
weight, correcting the rotational speed control start timing of the rolling rolls
set with respect to the predetermined stand; and based on the thickness measured value
of the front or rear end of the pipe or tube measured at the output side of the sizing
mill, correcting the rotational speed control start timing of the rolling rolls set
with respect to the predetermined stand.
[0024] According to such a preferable configuration, a prediction error between the rotational
speed control start timing of the rolling rolls set with respect to a predetermined
stand and the timing when the front or rear end of the pipe or tube actually reaches
the predetermined stand is calculated, and from the prediction error calculated, a
first prediction error component (prediction error component of a factor outside the
stand) until the front or rear end of the pipe or tube reaches the first stand and
a second prediction error component (prediction error component of a factor inside
the stand) after the front or rear end of the pipe or tube reached the first stand
are extracted. Next, a first weight set to a value of 0 to 1 for example, is applied
to the prediction error component of the factor outside the stand, and based on the
prediction error component of the factor outside the stand applied with the first
weight, the rotational speed control start timing of the rolling rolls set with respect
to the predetermined stand is corrected. Further, a second weight set to a value of
0 to 1 for example, is applied to the prediction error component of the factor inside
the stand, and based on the prediction error component of the factor inside the stand
applied with the second weight, the rotational speed control start timing of the rolling
rolls set with respect to the predetermined stand is corrected.
[0025] In other words, the rotational speed control start timing of the rolling rolls is
corrected in such a manner that two prediction error components where the generation
factors are different (accordingly, tendency of variation may differ) are divided,
and each of the components is weighted separately (each of the prediction error components
may be weighted differently). Therefore, even in such a state that a prediction error
between the rotational speed control start timing of rolling rolls set with respect
to each stand and the timing when the front or rear end of the pipe or tube actually
reaches the stand varies at random according to the fluctuations of the generation
factors of the prediction error, the rotational speed control start timing of the
rolling rolls can be corrected appropriately, whereby it is possible to suppress defective
thickness of a front or rear end of the pipe or tube rolled by the sizing mill more
effectively.
[0026] Further, the present invention is also provided as a method for controlling sizing
mill of a pipe or tube, comprising the steps of: calculating a prediction error between
a rotational speed control start timing of rolling rolls set with respect to a predetermined
stand of a sizing mill and a timing when a front or rear end of the pipe or tube actually
reaches the predetermined stand; extracting a first prediction error component until
the front or rear end of the pipe or tube reaches a first stand and a second prediction
error component after the front or rear end of the pipe or tube reaches the first
stand, from the prediction error calculated; applying a first weight to the first
prediction error component extracted, and based on the first prediction error component
applied with the first weight, correcting the rotational speed control start timing
of the rolling rolls set with respect to the predetermined stand; and applying a second
weight to the second prediction error component extracted, and based on the second
prediction error component applied with the second weight, correcting the rotational
speed control start timing of the rolling rolls set with respect to the predetermined
stand.
[0027] Further, in order to solve the above-described problem, the present invention is
provided as a device for controlling sizing mill of a pipe or tube, comprising: a
thickness gauge provided at an output side of a sizing mill; a timing computing unit
for correcting a rotational speed control start timing of rolling rolls set with respect
to a predetermined stand, based on a thickness measured value of a front or rear end
of a pipe or tube measured by the thickness gauge; and a rolling controller for controlling
a rotational speed of rolling rolls provided to each stand, based on the rotational
speed control start timing corrected by the timing computing unit.
[0028] It is preferable that the device for controlling sizing mill further comprises a
detecting unit for detecting that a front or rear end of the pipe or tube reaches
the predetermined stand, wherein the timing computing unit executes a computation
including: a first step of calculating a prediction error between the rotational speed
control start timing of the rolling rolls set with respect to the predetermined stand
and a timing when the front or rear end of the pipe or tube actually reaches the predetermined
stand detected by the detecting unit; and a second step of correcting the rotational
speed control start timing of the rolling rolls set with respect to the predetermined
stand, based on the prediction error calculated and the thickness measured value of
the front or rear end of the pipe or tube measured by the thickness gauge.
[0029] Further, it is preferable that the second step includes the steps of: extracting
a first prediction error component until the front or rear end of the pipe or tube
reaches a first stand and a second prediction error component after the front or rear
end of the pipe or tube reaches the first stand, from the prediction error calculated;
applying a first weight to the first prediction error component extracted, and based
on the first prediction error component applied with the first weight, correcting
the rotational speed control start timing of the rolling rolls set with respect to
the predetermined stand; applying a second weight to the second prediction error component
extracted, and based on the second prediction error component applied with the second
weight, correcting the rotational speed control start timing of the rolling rolls
set with respect to the predetermined stand; and based on the thickness measured value
of the front or rear end of the pipe or tube measured by the thickness gauge, correcting
the rotational speed control start timing of the rolling rolls set with respect to
the predetermined stand.
[0030] Further, in order to solve the above-described problem, the present invention is
also provided as a device for controlling sizing mill of a pipe or tube, comprising:
a detecting unit for detecting that a front or rear end of the pipe or tube reaches
a predetermined stand of a sizing mill; a timing computing unit for correcting a rotational
speed control start timing of rolling rolls set with respect to a predetermined stand;
and a rolling controller for controlling a rotational speed of rolling rolls provided
to each stand based on the rotational speed control start timing corrected by the
timing computing unit, wherein the timing computing unit executes a computation including
the steps of: calculating a prediction error between the rotational speed control
start timing of the rolling rolls set with respect to the predetermined stand and
a timing when the front or rear end of the pipe or tube actually reaches the predetermined
stand detected by the detecting unit; extracting a first prediction error component
until the front or rear end of the pipe or tube reaches a first stand and a second
prediction error component after the front or rear end of the pipe or tube reaches
the first stand, from the prediction error calculated; applying a first weight to
the first prediction error component extracted, and based on the first prediction
error component applied with the first weight, correcting the rotational speed control
start timing of the rolling rolls set with respect to the predetermined stand; and
applying a second weight to the second prediction error component extracted, and based
on the second prediction error component applied with the second weight, correcting
the rotational speed control start timing of the rolling rolls set with respect to
the predetermined stand.
BRIEF DESCRIPTION OF THE DRAWINGS
[0031]
Fig. 1 is an illustration for describing a method of previously setting control start
timings of rotational speeds of rolling rolls provided to respective stands.
Fig. 2 shows exemplary results of analyzing dispersion states of prediction error
components of factor outside a stand and prediction error components of factor inside
the stand.
Fig. 3 is an illustration for describing the meaning of the prediction error components
of factor outside a stand and the prediction error components of factor inside the
stand, shown in Fig. 2.
Fig. 4 is a block diagram showing the schematic configuration of a sizing mill used
for performing a method for controlling sizing mill, according to an embodiment of
the present invention.
Figs. 5(a) and 5(b) are diagrams each showing an example of a thickness measured value
of a pipe or tube outputted from a thickness gauge shown in Fig. 4.
Fig. 6 is a flowchart showing a process flow in a timing computing unit shown in Fig.
4.
Figs. 7(a) and 7(b) show an exemplary result of evaluating a prediction error between
the rotational speed control start timing of rolling rolls corrected with respect
to a predetermined stand and the timing when a front or rear end of the pipe or tube
actually reached the predetermined stand, in a case where the method for controlling
sizing mill according to an embodiment of the present invention is applied.
Fig. 8 shows an example of the increased thickness ratio of a front or rear end in
a case where the method for controlling sizing mill according to an embodiment of
the present invention is applied.
BEST MODE FOR CARRYING OUT THE INVENTION
[0032] Hereinafter, an embodiment of the present invention will be described with reference
to the accompanying drawings.
First Embodiment
[0033] Fig. 4 is a block diagram showing the schematic configuration of a sizing mill used
for performing a method for controlling sizing mill, according to an embodiment of
the present invention. As shown in Fig. 4, a pipe or tube 1 to be rolled is carried
in an axial direction (direction shown by the outlined arrow in Fig. 4) by carrying
rolls (not shown), and is sized and rolled at each stand 2. In the vicinity of the
input side of the sizing mill in the carrying path of the pipe or tube 1, a pipe or
tube end detector 8 which consists of a photoelectric sensor and detects a front end
and rear end of the pipe or tube 1 by the operation of the photoelectric sensor. Further,
in the vicinity of the output side of the sizing mill in the carrying path of the
pipe or tube 1, a γ-ray thickness gauge 9 and a length gauge 10 consisting of a photoelectric
sensor or the like are disposed. An end detecting signal of the pipe or tube 1 outputted
from the pipe or tube end detector 8 is inputted into a rolling controller 7 and a
timing computing unit 6. Further, a thickness measured value of the pipe or tube 1
outputted from the thickness gauge 9 and a length measured value of the pipe or tube
1 outputted from the length gauge 10 are inputted into the timing computing unit 6.
[0034] A rolling roll 21 provided to each stand 2 is driven by a roll driving motor 3 via
a reduction gear 31. Among a plurality of stands 2, a roll driving motor 3 of a stand
2 in, for example, odd-number order counted from the first stand (stand disposed at
the most upstream side) is provided with a current detector 32 for detecting a drive
current of the roll driving motor 3 and a rotational speed detector 33 for detecting
the rotational speed (the present invention is not limited to this configuration,
and it is possible to adopt configurations in which the current detector 32 and the
rotational speed detector 33 are provided to another predetermined stand or to the
roll driving motors 3 of all stands). Detection signals of the current detector 32
and the rotational speed detector 33 are inputted into a motor drive controller 4
for drive-controlling the roll driving motor 3, respectively. In the motor drive controller
4, a rotational speed control starting signal of the rolling roll 21 has been inputted
from the rolling controller 7, and the motor drive controller 4 performs a rotational
speed control of the roll driving motor 3 based on the rotational speed control starting
signal. Further, the detection signals of the current detector 32 and the rotational
speed detector 33 are also inputted into a rolling torque computing unit 5 via the
motor drive controller 4.
[0035] The rolling torque computing unit 5 serves as a detecting unit for detecting that
a front or rear end of the pipe or tube reaches a predetermined stand in the present
invention. The rolling torque computing unit 5 calculates rolling torque based on
the detection signals of the driving current and the rotational speed inputted from
the motor drive controller 4, and outputs the calculated rolling torque signal to
the timing computing unit 6. To the timing computing unit 6, there are inputted the
calculated rolling torque signal, the end detection signal from the pipe or tube end
detector 8, the thickness measured value of the pipe or tube 1 outputted from the
thickness gauge 9, the length measured value of the pipe or tube 1 outputted from
the length gauge 10, and the rotational speed control starting signal of the rolling
roll 21 from the rolling controller 7. Based on these inputted signals, the timing
computing unit 6 calculates the correction amount of the rotational speed control
starting signal, and outputs the calculation result to the rolling controller 7 as
a correction signal.
[0036] To the rolling controller 7, the end detection signal from the pipe or tube end detector
8 and the correction signal from the timing computing unit 6 are inputted. Timing
is started at the time when the end detection signal from the pipe or tube end detector
8 is imputed, and when the timing result reaches a stored set value of the rotational
speed control start timing of the rolling rolls 21 of each stand 2, the rolling controller
7 outputs a rotational speed control starting signal to each motor drive controller
4 and to each timing computing unit 6. Each motor drive controller 4 lowers the rotational
speed of the roll driving motor 3 based on the rotational speed control starting signal
inputted. Note that a set value of the rotational speed control start timing is corrected
based on the correction signal inputted from the timing computing unit 6, and is stored
as a set value used for rolling the next pipe or tube 1.
[0037] Hereinafter, a method of computing the correction amount of the rotational speed
control starting signal (correction amount of the rotational speed control start timing)
in the timing computing unit 6, based on the rolling torque signal from the rolling
torque computing unit 5, the end detection signal from the pipe or tube end detector
8, the thickness measured value of the pipe or tube 1 outputted from the thickness
gauge 9, the length measured value of the pipe or tube 1 outputted from the length
gauge 10, and the rotational speed control starting signal of the rolling roll 21
from the rolling controller 7, will be described specifically with reference to Figs.
5 and 6 and Fig. 3 described above as appropriate. Note that the method for controlling
sizing mill, according to this embodiment, is configured so as to compute the correction
amount of the rotational speed control starting signal while taking into account both
of the correction amount based on the thickness measured value of a front or rear
end of the pipe or tube 1 (hereinafter, referred to as "correction amount based on
thickness result" as appropriate) and the correction amount based on a prediction
error between the rotational speed control start timing of the rolling rolls 21 and
the timing when the front or rear end of the pipe or tube 1 actually reaches (hereinafter,
referred to as "correction amount based on prediction error" as appropriate). Hereinafter,
the respective correction amounts will be described in sequence.
(1) Correction Amount Based on Thickness Result
[0038] First, the correction amount based on a thickness result will be described. Fig.
5 is a diagram showing an example of a thickness measured value of the pipe or tube
1 (average thickness in a pipe or tube circumferential direction) outputted from the
thickness gauge 9. The timing computing unit 6 first calculates an average thickness
tm in a length Lm, in the intermediate portion of the pipe or tube 1, shown by the
following equation (1), based on the thickness measured value of the pipe or tube
1 outputted from the thickness gauge 9 and the length measured value of the pipe or
tube 1 outputted from the length gauge 10:

[0039] Here, in the equation (1) above, L means the length of the pipe or tube 1 at the
output side of the sizing mill, Lct means the crop length at the front end of the
pipe or tube 1 defined in advance according to the type and dimensions of the pipe
or tube 1, Lt means the front end length of the product part of the pipe or tube 1
defined in advance, Lcb means the crop length at the rear end of the pipe or tube
1 defined in advance, and Lb means the rear end length of the product part of the
pipe or tube 1 defined in advance. Note that the front end length Lt and the rear
end length Lb of the product part are lengths in predetermined proportional to the
length (or target length) of the pipe or tube 1 at the output side of the sizing mill,
or are constant lengths irrespective of the length of the pipe or tube 1.
(1-1) When there is a thickness measured value exceeding the upper limit but no value
below the lower limit
[0040] Next, when, among thickness measured values of the pipe or tube 1 outputted from
the thickness gauge 9, there is a value exceeding the upper limit (= tm + tup) but
no value below the lower limit (= tm - tlo), within a part of the pipe or tube 1 corresponding
to the front end length Lt of the product part, the timing computing unit 6 calculates
the increased thickness length Lzt of the front end. Here, tup and tlo are values
which have been determined beforehand. Further, the increased thickness length Lzt
of the front end means the length from a portion where the thickness first increases
from the average thickness tm by tup viewed from the most inner side of the portion
corresponding to the front end length Lt of the product part of the pipe or tube 1,
to a portion coming inside by the crop length Lct from the tip of the pipe or tube
1, as shown in Fig. 5(a). Similarly, if, among thickness measured values of the pipe
or tube 1 outputted from the thickness gauge, there is a value exceeding the upper
limit (= tm + tup) but no value below the lower limit (= tm - tlo), within a portion
of the pipe or tube 1 corresponding to the rear end length Lb of the product part,
the timing computing unit 6 calculates the increased thickness length Lzb of the rear
end. The increased thickness length Lzb of the rear end means the length from a portion
where the thickness first increases from the average thickness tm by tup viewed from
the most inner side of the portion corresponding to the rear end length Lb of the
product part of the pipe or tube 1, to a portion coming inside by the crop length
Lcb from the rear end of the pipe or tube 1. The values of tup and tlo may take the
same values in the front end and the rear end of the pipe or tube 1, or may be different.
[0041] Then, as for the rotational speed control start timing of the rolling rolls 21 set
with respect to each stand 2 relating to the front end of the pipe or tube 1, the
timing computing unit 6 outputs to the rolling controller 7 a correction signal in
which ΔTt1 shown by the following equation (2) is set as the correction amount based
on the thickness result:

[0042] Here, in the equation (2), Kt means a constant (weight) set to a value of 0 to 1,
L0 means the length of the pipe or tube 1 at the input side of the sizing mill (it
is measurable by arranging a length gauge on the input side of the sizing mill or
by measuring the length in the step of the previous stage of the sizing mill), and
V0 means the speed of the pipe or tube 1 at the input side of the sizing mill (it
is measurable by arranging a speed meter on the input side of the sizing mill, or
by arranging two pipe or tube end detectors 8 described above and dividing the distance
between the pipe or tube end detectors 8 by the difference between the detection times).
[0043] Since L/L0 means the elongation percentage of the pipe or tube 1 (percentage of the
pipe or tube being elongated) by the sizing mill, a value, obtained by dividing the
increased thickness length Lzt of the front end by L/L0 (= Lzt·L0/L), corresponds
to the length of the increased thickness length of the front end at the input side
of the sizing mill. A value, obtained by dividing the length of the increased thickness
length of the front end at the input side of the sizing mill by the speed V0 of the
pipe or tube 1 at the input side of the sizing mill (= Lzt · L0/L/V0) means a time
period during which a portion corresponding to the increased thickness length of the
front end is generated. Accordingly, it is possible to suppress generation of the
portion corresponding to the increased thickness length of the front end by setting
ΔTt1 shown by the above-mentioned equation (2) as the correction amount, and rolling
the next pipe or tube 1 by uniformly adding the correction amount ΔTt1 to the rotational
speed control start timing of the rolling rolls 21 set with respect to each stand
2 (delaying the rotational speed control start timing by ΔTt1).
[0044] On the other hand, as for the rotational speed control start timing of the rolling
rolls 21 set with respect to each stand 2 relating to the rear end of the pipe or
tube 1, the timing computing unit 6 outputs to the rolling controller 7 a correction
signal in which ΔTb 1 shown by the following equation (3) is set as the correction
amount based on the thickness result:

[0045] Here, in the equation (3), Kb means a constant (weight) set to a value of 0 to 1.
[0046] Similar to the aforementioned case of the front end of the pipe or tube 1, Lzb ·
L0/L/V0 means a time period in which a portion correspond to the increased thickness
length Lzb of the rear end is generated. Accordingly, it is possible to suppress generation
of the portion corresponding to the increased thickness length of the rear end by
setting ΔTb1 shown by the above-mentioned equation (3) as the correction amount, and
rolling the next pipe or tube 1 by uniformly adding the correction amount ΔTb1 to
the rotational speed control start timing of the rolling rolls 21 set with respect
to each stand 2 (advancing the rotational speed control start timing by a period corresponding
to the absolute value of ΔTt1).
(1-2) When there is a thickness measured value below the lower limit
[0047] On the other hand, when there is a value below the lower limit (= tm - tlo) among
thickness measured values of the pipe or tube 1 outputted from the thickness gauge
9, in the portion of the pipe or tube 1 corresponding to the front end length Lt of
the product part, the timing computing unit 6 calculates the reduced thickness length
Lgt of the front end. Here, as shown in Fig. 5(b), the reduced thickness length Lgt
of the front end means the length from a portion where the thickness is first reduced
from the average thickness tm by tlo viewed from the most inner side of the portion
corresponding to the front end length Lt of the product part of the pipe or tube 1,
to a portion coming inside by the crop length Lct from the tip of the pipe or tube
1. Similarly, when there is a value below the lower limit (= tm - tlo) among thickness
measured values of the pipe or tube 1 outputted from the thickness gauge 9, in the
portion of the pipe or tube 1 corresponding to the rear end length Lb of the product
part, the timing computing unit 6 calculates the reduced thickness length Lgb of the
rear end. The reduced thickness length Lgb of the rear end means the length from a
portion where the thickness is first reduced from the average thickness tm by tlo
viewed from the most inner side of the portion corresponding to the rear end length
Lb of the product part of the pipe or tube 1, to a portion coming inside by the crop
length Lcb from the rear end of the pipe or tube 1.
[0048] Then, as for the rotational speed control start timing of the rolling rolls 21 set
with respect to each stand 2 relating to the front end of the pipe or tube 1, the
timing computing unit 6 outputs to the rolling controller 7 a correction signal in
which ΔTt1 shown by the following equation (4) is set as the correction amount based
on the thickness result:

[0049] On the other hand, as for the rotational speed control start timing of the rolling
rolls 21 set with respect to each stand 2 relating to the rear end of the pipe or
tube 1, the timing computing unit 6 outputs to the rolling controller 7 a correction
signal in which ΔTb1 shown by the following equation (5) is set as the correction
amount based on the thickness result:

(1-3) When there is no value exceeding the upper limit and no value below the lower
limit in thickness measured value
[0050] When there is no value exceeding the upper limit (= tm + tup) and no value below
the lower limit (= tm - tlo) in thickness measured values of the pipe or tube 1 outputted
from the thickness gauge 9 within the portion of the pipe or tube 1 corresponding
to the front end length Lt of the product part, the timing computing unit 6 ends the
computation since there is no need to correct the rotational speed control starting
signal relating to the correction amount based on the thickness result (that is, a
correction signal of the correction amount ΔTb1 = 0 based on the thickness result
is outputted to the rolling controller 7). Similarly, when there is no value exceeding
the upper limit (= tm + tup) and no value below the lower limit (= tm - tlo) in thickness
measured values of the pipe or tube 1 outputted from the thickness gauge 9 within
the portion of the pipe or tube 1 corresponding to the rear end length Lb of the product
part, the timing computing unit 6 ends the computation since there is no need to correct
the rotational speed control starting signal relating to the correction amount based
on the thickness result (that is, a correction signal of the correction amount ΔTb1
= 0 based on the thickness result is outputted to the rolling controller 7).
(2) Correction Amount Based on Prediction Error
[0051] Next, the correction amount based on a prediction error will be described. The timing
computing unit 6 starts timing at a timing where an end (front end or rear end) detection
signal is imputed from the pipe or tube end detector 8 to the timing computing unit
6 as a starting point, and based on the fluctuation state of a rolling torque signal
inputted from the rolling torque computing unit 5, detects the timing where the end
(front end or rear end) of the pipe or tube 1 actually reaches the predetermined stand
2 (stand in odd-number order in this embodiment) (that is, an elapsed time period
starting from the time when the end detecting signal is imputed, hereinafter referred
to as a "measured period" as appropriate). A specific method of detecting the timing
when the front or rear end of the pipe or tube 1 actually reaches the predetermined
stand 2, based on the fluctuation state of the rolling torque signal inputted, is
similar to that described in Patent Document 1 described above, so the detailed description
thereof is omitted in the present specification. On the other hand, the timing computing
unit 6 detects an elapsed time period starting from the time when the end detecting
signal is inputted from the pipe or tube end detector 8 to the timing computing unit
6, to the time when the rotational speed control starting signal of the rolling roll
21 is inputted from the rolling controller 7 (hereinafter, referred to as "predicted
period" as appropriate), and calculates the prediction error Y
j between the predicted period and the measured period.
[0052] In the timing computing unit 6, first, based on plural pieces of data (i, Y
i) plotted by assuming the horizontal axis X being the stand number and the vertical
axis Y being the prediction error between the predicted period and the measured period
(see Fig. 3), the correlation coefficient R between X and Y is calculated (S1 in Fig.
6), and the calculated correlation coefficient R is determined whether it is below
the predetermined value (S2 in Fig. 6).
[0053] Here, if the correlation coefficient R is below the predetermined value, it is determined
that the prediction error Y
i only includes a prediction error component of the factor outside the stand, and based
on (i, Y
i), a primary regression equation of Y, where X is a variable, is calculated (S3 in
Fig. 6). Then, a Y section of the primary regression equation is defined as a prediction
error T
0' between the predicted period and the measured period in the first stand (see S4
in Fig. 6, and Fig. 3). Next, for the stand in which the measured period has been
detected, the sum of squares Σ(Y
i - T
0')
2 of the difference between the prediction error Y
i and the prediction error T
0' and the sum of squares Σ(Y
i)
2 of the prediction error Y
i are compared (S5 in Fig. 6), and if Σ(Y
i - T
0')
2≥ Σ(Y
i)
2, the computation ends since there is no need to correct the rotational speed control
starting signal relating to the correction amount based on the prediction error. More
specifically, if the prediction error Y
i for the front end of the pipe or tube 1 satisfies the conditions described above,
the timing computing unit 6 outputs a correction signal where the correction amount
based on the prediction error is ΔTt2 = 0, to the rolling controller 7, as for the
rotational speed control start timing of the rolling rolls 21 set with respect to
each stand 2 relating to the front end of the pipe or tube 1. Similarly, if the prediction
error Y
i for the rear end of the pipe or tube 1 satisfies the conditions described above,
the timing computing unit 6 outputs a correction signal where the correction amount
based on the prediction error is ΔTb2 = 0, to the rolling controller 7, as for the
rotational speed control start timing of the rolling rolls 21 set with respect to
each stand 2 relating to the rear end of the pipe or tube 1.
[0054] On the other hand, if Σ(Y
i - T
0')
2 < Σ(Y
i)
2, the timing computing unit 6 multiplies the prediction error T
0' by a first weight (value of 0 to 1, e.g., 0.5), and outputs a correction signal
in which the prediction error T
0' multiplied by the first weight is set as the correction amount to the rolling controller
7 (S6 in Fig. 6). More specifically, if the prediction error Y
i for the front end of the pipe or tube 1 satisfies the conditions described above,
the timing computing unit 6 outputs a correction signal where the correction amount
based on the prediction error is ΔTt2 = -(first weight) T
0', to the rolling controller 7, as for the rotational speed control start timing of
the rolling rolls 21 set with respect to each stand 2 relating to the front end of
the pipe or tube 1. In the rolling controller 7, the correction amount ΔTt2 based
on the prediction error is added to the rotational speed control start timing of the
rolling rolls 21 of each stand 2 uniformly (subtracting the prediction error T
0' multiplied by the first weight), which is used for rolling the next pipe or tube
1 (S6 in Fig. 6). Similarly, if the prediction error Y
i for the rear end of the pipe or tube 1 satisfies the conditions described above,
the timing computing unit 6 outputs a correction signal where the correction amount
based on the prediction error is ΔTb2 = -(first weight). T
0', to the rolling controller 7, as for the rotational speed control start timing of
the rolling rolls 21 set with respect to each stand 2 relating to the rear end of
the pipe or tube 1.
[0055] On the other hand, if the correlation coefficient R is larger than the predetermined
value, it is determined that the prediction error Y
i includes both of the prediction error component of the factor outside the stand and
the prediction error component of the factor inside the stand, whereby the timing
computing unit 6 calculates a primary regression equation based on (i, Y
i) similar to that described above (S7 in Fig. 6), and sets the calculated Y section
of the of the primary regression equation as the prediction error T
0' between the predicted period and the measured period in the first stand (see S8
in Fig. 6 and Fig. 3). Next, the prediction error T
0' is multiplied by a first weight (value of 0 to 1, e.g., 0.5), and a correction signal
in which the prediction error T
0' multiplied by the first weight is set as the correction amount is outputted to the
rolling controller 7 (S9 in Fig. 6). In the rolling controller 7, the correction amount
is subtracted from the rotational speed control start timing of the rolling rolls
21 of each stand 2 uniformly (S9 in Fig. 6). In other words, through the processing
shown in S9 of Fig. 6, the prediction error component of the factor outside the stand
included in the prediction error Y
i is corrected. Further, in the timing computing unit 6, a prediction error Y
i' between the predicted period and the measured period in each stand (i
th stand) after the second stand is calculated based on the primary regression equation
(see S10 in Fig, 6 and Fig. 3), and a value obtained by subtracting T
0' from Y
i' (this value corresponds to a prediction error component ΣT
j' for a predicted period ΣT
j(j = 1 to i-1, i ≥ 2) until the front or rear end of the pipe or tube reaches i
th stand from the first stand) is multiplied by a second weight (values of 0 to 1, e.g.,
0.5), and a correction signal where the value multiplied by the second weight is set
as the correction amount is outputted to the rolling controller (S11 in Fig. 6). In
the rolling controller 7, the correction amount is further subtracted from the rotational
speed control start timing, for the next pipe or tube, of the rolling roll 21 in each
stand (i
th stand) (S11 in Fig. 6). In other words, through the processing shown in S11 of Fig.
6, the prediction error component of the factor inside the stand, included in the
prediction error Y
i, is corrected.
[0056] Namely, when the correlation coefficient R for the front end of the pipe or tube
1 is larger than the predetermined value, the timing computing unit 6 outputs a correction
signal where the correction amount based on the prediction error is ΔTt2 = -(first
weight)· T
0' - (second weight)· (Y
i' - T
0') is outputted to the rolling controller 7, as for the rotational speed control start
timing of the rolling rolls 21 set with respect to each stand 2 relating to the front
end of the pipe or tube 1. Similarly, when the correlation coefficient R for the rear
end of the pipe or tube 1 is larger than the predetermined value, the timing computing
unit 6 outputs a correction signal where the correction amount based on the prediction
error is ΔTb2 = -(first weight) T
0' - (second weight)·(Y
i' - T
0') is outputted to the rolling controller 7, as for the rotational speed control start
timing of the rolling rolls 21 set with respect to each stand 2 relating to the rear
end of the pipe or tube 1.
[0057] Note that this embodiment has described such a configuration that, based on the data
(i, Y
i) plotted on the premise that the horizontal axis X shows the stand number i and the
vertical axis Y shows the prediction error Y
i between the predicted period and the measured period, a primary regression equation
ofY where X is variable is calculated, and with the primary regression equation, a
prediction error component of a factor outside the stand and a prediction error component
of a factor inside the stand are separated. However, the present invention is not
limited to this configuration. It is possible to adopt such a configuration that an
N
th (integer of N > 1) regression equation ofY where X is variable is calculated based
on data (i, Y
i), and a prediction error component of a factor outside the stand and a prediction
error component of a factor inside the stand are separated. Further, as the first
weight and the second weight, the same values may be adopted in the front end and
the rear end of the pipe or tube 1, or different values may be adopted.
(3) Total Correction Amount
[0058] The method of controlling thicknesses of the front and rear end of the pipe or tube
according to this embodiment is so configured as to compute the correction amount
of the rotational speed control start timing of the rolling rolls 21 set at the current
rolling time with respect to each stand 2, taking into account both (1) correction
amount based on the thickness result and (2) the correction amount based on the prediction
error, as described above. That is, as for the rotational speed control start timing
of the rolling rolls 21 set with respect to each stand 2 relating to the front end
of the pipe or tube 1, ΔTt shown by the following equation (6) is considered as the
total correction amount, and is stored in the rotating controller 7:

[0059] Here, αt is a constant of 0 to 1, and βt is a constant of 1 - αt.
[0060] Further, the set value of the rotational speed control start timing of the rolling
rolls 21 set with respect to each stand 2 relating to the front end of the pipe or
tube 1 is corrected based on the correction amount ΔTt stored (correction amount ΔTt
is added), which is used as a set value for rotating the next pipe or tube 1.
[0061] Similarly, as for the rotational speed control start timing of the rolling rolls
21 set with respect to each stand 2 relating to the rear end of the pipe or tube 1,
ΔTb shown by the following equation (7) is considered as the total correction amount,
and is stored in the rotating controller 7:

[0062] Here, αb is a constant of 0 to 1, and βb is a constant of 1 - αb.
[0063] Further, the set value of the rotational speed control start timing of the rolling
rolls 21 set with respect to each stand 2 relating to the rear end of the pipe or
tube 1 is corrected based on the correction amount ΔTb stored (correction amount ΔTb
is added), which is used as a set value for rotating the next pipe or tube 1.
[0064] Although, in this embodiment, descriptions have been given in the order of (1) the
correction amount based on the thickness result and (2) the correction amount based
on the prediction error, it is not necessary to perform computation in this order.
A configuration in which either correction amount is computed first may be applied.
[0065] As described above, according to the method for controlling sizing mill of this embodiment,
the rotational speed control start timing of the rolling rolls 21 set with respect
to each stand 2 is corrected based on the thickness measured values of the front and
rear end of the pipe or tube 1 measured at the output side of the sizing mill. Therefore,
it is expected that the rotational speed control start timing of the rolling rolls
21 is corrected so as to be appropriate for the actual thickness fluctuations of the
front and rear end of the pipe or tube. Further, since the prediction error is divided
into two prediction error components of different factors (a prediction error component
of a factor outside the stand and a prediction error component of a factor inside
the stand) which are weighted respectively (it is possible to differ first and second
weights applied to the both prediction error components respectively) and are provided
to correct the rotational speed control start timing of the rolling rolls 21, it is
possible to correct the rotational speed control start timing of the rolling rolls
21 appropriately in a state where the prediction error between the rotational speed
control start timing of the rolling rolls 21 set with respect to each stand 2 and
the timing when a front or rear end of the pipe or tube 1 reaches each stand 2 varies
at random along with the fluctuation of the generation factors of the prediction error.
Accordingly, it is possible to suppress defective front and rear end thickness of
the pipe or tube 1 rolled by the sizing mill, effectively.
Second Embodiment
[0066] A method for controlling sizing mill, according to this embodiment, is configured
to use only the correction amount based on the thickness result in the first embodiment.
That is, as for the rotational speed control start timing of the rolling rolls 21
set with respect to each stand 2 relating to the front end of the pipe or tube 1,
ΔTt = ΔTt1 (that is, αt = 1, βt = 0 in the equation (6) above) is considered as the
total correction amount, and is stored in the rolling controller 7. Further, the set
value of the rotational speed control start timing of the rolling rolls 21 set with
respect to each stand 2 relating to the front end of the pipe or tube 1 is corrected
based on the stored correction amount ΔTt (correction amount ΔTt is added), which
is used as a set value for rolling the next pipe or tube 1. Similarly, as for the
rotational speed control start timing of the rolling rolls 21 set with respect to
each stand 2 relating to the rear end of the pipe or tube 1, ΔTb = ΔTb1 (that is,
αb = 1, βb = 0 in the equation (7) above) is considered as the total correction amount,
and is stored in the rolling controller 7. Further, the set value of the rotational
speed control start timing of the rolling rolls 21 set with respect to each stand
2 relating to the rear end of the pipe or tube 1 is corrected based on the stored
correction amount ΔTb (correction amount ΔTb is added), which is used as the set value
for rolling the next pipe or tube 1.
[0067] According to the method for controlling sizing mill of this embodiment, the rotational
speed control start timing of the rolling rolls 21 set with respect to each stand
2 is corrected based on the thickness measured values of the front and rear end of
the pipe or tube 1 measured at the output side of the sizing mill, so it is expected
that the rotational speed control start timing of the rolling rolls 21 can be corrected
so as to be appropriate for the actual thickness fluctuations of the front and rear
ends of the pipe or tube. Thereby, it is possible to suppress defective thicknesses
in the front and rear ends of the pipe or tube 1 rolled by the sizing mill, effectively.
Third Embodiment
[0068] A method for controlling sizing mill, according to this embodiment, is configured
to use a prediction error Y
i between the predicted period and the measured period in each stand 2 as it is. The
correction amounts based on the prediction errors are ΔTt2 and ΔTb2 (in this embodiment,
however, measured periods must be detected not only for the stands in odd-number order
but for all stands, different from the first embodiment). In other words, the correction
amount is computed based on the equations (6) and (7) above, where ΔTt2 = -Yi (prediction
error for the front end) and ΔTb2 = -Yi (prediction error for the rear end).
[0069] According to the method for controlling sizing mill of this embodiment, the rotational
speed control start timing of the rolling rolls 21 is corrected by using not only
the thickness measured values of the front and rear ends of the pipe or tube 1 measured
at the output side of the sizing mill but also the prediction errors between the rotation
speed control start timings of the rolling roll set with respect to each stand 2 and
the timing when the front or rear end of the pipe or tube 1 actually reach each stand
2. Therefore it is expected that the rotational speed control start timing of the
rolling rolls 21 can be corrected so as to be more appropriate than the case of the
second embodiment, whereby it is possible to suppress defective thicknesses in front
and rear ends of the pipe or tube 1 rolled by the sizing mill, effectively.
Fourth Embodiment
[0070] A method for controlling sizing mill, according to this embodiment, is configured
to only use the correction amount based on the prediction error in the first embodiment.
That is, as for the rotational speed control start timing of the rolling rolls 21
set with respect to each stand 2 relating to the front end of the pipe or tube 1,
ΔTt = ΔTt2 (that is, αt = 0, βt = 1 in the equation (6) above) is considered as the
total correction amount, and is stored in the rolling controller 7. Then, the set
value of the rotational speed control start timing of the rolling rolls 21 set with
respect to each stand 2 relating to the front end of the pipe or tube 1 is corrected
based on the stored correction amount ΔTt (the correction amount ΔTt is added), which
is used as the set value for rolling the next pipe or tube 1. Similarly, as for the
rotational speed control start timing of the rolling rolls 21 set with respect to
each stand 2 relating to the rear end of the pipe or tube 1, ΔTb = ΔTb2 (that is,
αb = 0, βb = 1 in the equation (7) above) is considered as the total correction amount,
and is stored in the rolling controller 7. Then, the set value of the rotational speed
control start timing of the rolling rolls 21 set with respect to each stand 2 relating
to the rear end of the pipe or tube 1 is corrected based on the stored correction
amount ΔTb (the correction amount ΔTb is added), which is used as the set value for
rolling the next pipe or tube 1.
[0071] According to the method for controlling sizing mill of this embodiment, since the
prediction error is divided into two prediction error components of different factors
(a prediction error component of a factor outside the stand and a prediction error
component of a factor inside the stand) which are weighted respectively (it is possible
to differ first and second weights applied to the both prediction error components
respectively) and are provided to correct the rotational speed control start timing
of the rolling rolls 21, it is possible to correct the rotational speed control start
timing of the rolling rolls 21 appropriately in a state where the prediction errors
between the rotational speed control start timings of the rolling roll 21 set to each
stand 2 and the timings when front and rear ends of the pipe or tube 1 reach each
stand 2 varies at random along with the fluctuations of the generation factors of
the prediction errors. Accordingly, it is possible to suppress defective thicknesses
in the front and rear ends of the pipe or tube 1 rolled by the sizing mill, effectively.
[0072] Figs. 7(a) and 7(b) show exemplary results of evaluating prediction error between
the rotational speed control start timing of the rolling rolls 21 corrected with respect
to a predetermined stand 2 and the timing when a front or rear end of the pipe or
tube 1 actually reaches the predetermined stand 2, in the case of applying the method
for controlling sizing mill (method for correcting rotational speed control start
timing of rolling roll 21) according to the fourth embodiment of the present invention.
Fig. 7(a) shows a prediction error in the case of applying the method according to
the fourth embodiment of the present invention, and Fig. 7(b) shows a prediction error
in the case of applying the conventional method (prediction error between the rotational
control start timing of the rolling rolls 21 which has been set beforehand with respect
to the predetermined stand 2 (including a case where correction is performed manually
by the operator) and the timing when the front or rear end of the pipe or tube 1 actually
reaches the predetermined stand 2). As shown in Fig. 7(a), it was found that according
to the method of this embodiment, the absolute value of the average value of prediction
errors becomes small and the dispersion also becomes small comparing with the case
of applying the conventional method (Fig. 7(b)), whereby it was possible to correct
the rotational speed control start timing of the rolling rolls 21, appropriately.
[0073] Thereby, as shown in Fig. 8, it is possible to suppress defective thickness (increased
thickness ratio of end) of the front end (except for the crop part shown in Fig. 8)
of the pipe or tube 1 rolled by the sizing mill more effectively comparing with the
conventional case. The increased thickness ratio of end shown in Fig. 8 is a value
indicated by (thickness at each portion of end - average thickness (tm))/average thickness
(tm) × 100 (%).
[0074] Further, the tolerance failure ratio of the thickness of the pipe or tube 1 after
being rolled was evaluated, in the case of applying the methods for controlling sizing
mill according to the first to fourth embodiments of the present invention and methods
of comparative examples. More specifically, 50 to 100 pipes or tubes for each manufacturing
chance were sized and rolled for three manufacturing chances in total under the following
conditions of (1) to (6), and the tolerance failure ratios were evaluated for the
front ends (portions corresponding to the crop lengths Lct and the front end lengths
Lt of the product parts) after being rolled, for each chance. Note that the tolerance
failure ratio means the ratio of the number of pipes or tubes in which the average
thicknesses of the front ends are out of the range of (tm - tlo) to (tm - tup) to
the total number of rolled pipes or tubes:
- (1) Pipe or tube dimensions at input side of sizing mill: outer diameter 100 mm, thickness
6.0 to 7.0 mm,
- (2) Pipe or tube dimensions at output side of sizing mill: outer diameter 30.0 mm,
thickness 5.0 to 6.0 mm,
- (3) Pipe or tube temperature at input side of sizing mill: 900 to 950°C,
- (4) Pipe or tube temperature at output side of sizing mill: 810 to 860°C,
- (5) The number of stands in sizing mill: 25 stands, and
- (6) Pipe or tube material: carbon steel.
[0075] Table 1 shows evaluation results. Note that Examples 1-1 and 1-2 in Table 1 show
a method for controlling sizing mill, corresponding to the first embodiment. Example
1-1 shows an example in which coefficients αt and βt in the equation (6) described
above are fixed (the first coefficients shown in Table 1 are αt = βt = 0.5, the second
coefficients shown in Table 1 are αt = βt = 0.5) in all three manufacturing chances.
Example 1-2 shows an example in which αt and βt are fixed (the first coefficients
shown in Table 1 are αt = 0.3 and βt = 0.7) in the same manufacturing chance, but
at the timing of the manufacturing chance being changed, different coefficients (the
second coefficients shown in Table 1 are αt = 0.7 and βt = 0.3) are used. Further,
Example 2, Example 3 and Example 4 are methods for controlling sizing and fixing,
corresponding to the second embodiment, the third embodiment and the fourth embodiment,
respectively. Comparative Example 1 is a method of using a prediction error between
the predicted period and the measured period in each stand as the correction mount
as it is, without performing thickness measurement. Comparative Example 2 is a method
in which correction is performed manually by the operator, without performing thickness
measurement.
Table 1
|
Correction based on thickness result |
Correction based on prediction error |
The first coefficient |
coefficient |
Tolerance failure ratio (%) |
αt |
βt |
αt |
βt |
Chance 1 |
Chance 2 |
Chance 3 |
Example 1-1 |
O |
O (Primary regression equation) |
0.5 |
0.5 |
0.5 |
0.5 |
3.5 |
1.5 |
1.2 |
Example 1-2 |
O |
O (Primary regression equation) |
0.3 |
0.7 |
0.7 |
0.3 |
3.8 |
1.0 |
0.7 |
Example 2 |
O |
× |
1.0 |
0.0 |
1.0 |
0.0 |
4.0 |
3.5 |
3.2 |
Example 3 |
O |
O (Prediction error in each stand) |
0.5 |
0.5 |
0.5 |
0.5 |
5.0 |
3.5 |
2.5 |
Example 4 |
× |
O (Primary regression equation) |
0.0 |
1.0 |
0.0 |
1.0 |
6.5 |
4.5 |
4.0 |
Comparative Example 1 |
× |
O (Prediction error in each stand) |
0.0 |
1.0 |
0.0 |
1.0 |
8.0 |
7.5 |
6.5 |
Comparative Example 2 |
× |
Δ (Manual) |
0.0 |
(1.0) |
0.0 |
(1.0) |
10.0 |
9.0 |
8.5 |
[0076] As shown in Table 1, the tolerance failure ratio was lowered in the method of Example
4, comparing with Comparative Examples 1 and 2. In particular, in the method of Comparative
Example 1, the prediction error between the predicted period and the measured period
in each stand was used as it was as the correction amount, so an influence of the
measurement error of the measured period directly affected, whereby the prediction
error was difficult to be solved. On the other hand, in the method of Example 4, the
correction amount was approximated by a primary regression equation, whereby it was
less likely to be affected by the measurement error of the measured period, so it
was considered that the tolerance failure ratio was lowered consequently. Further,
in the method of Example 3, correction based on the thickness result was added to
the method of Comparative Example 1, whereby it was possible to reduce the tolerance
failure ratio comparing with the methods of Comparative Examples 1 and 2. However,
since the prediction error in each stand was defined as the correction amount based
on the prediction error as it was, similar to the method of Comparative Example 1,
it was difficult to solve the prediction error, so the tolerance failure ratio was
also somewhat difficult to be solved, consequently. Further, as for the method of
Example 2, it was possible to reduce the tolerance failure ratio comparing with the
methods of Comparative Examples 1 and 2, since a correction based on the thickness
result was performed. However, since a correction based on the prediction error was
not performed, the tolerance failure ratio was less likely to be solved comparing
with the methods of Examples 1-1, 1-2 and 3. In the method of Example 1-1, since a
correction based on the thickness result and a correction based on the prediction
error based on a primary regression equation are performed, the tolerance failure
ratio was lowered and was possible to be solved quickly, comparing with not only Comparative
Examples 1 and 2 but also Examples 2 to 4. Further, in the methods of Example 1-2,
since different coefficients were used in the case of the same manufacturing chance
and in the case of the manufacturing chance being changed (at the timing where the
manufacturing chance is changed, the value of αt is increased such that the correction
amount based on the thickness result contributes more), it was possible to solve the
tolerance failure ratio more quickly than the method of Example 1-1. This is because
the correction amount based on the thickness result may depend more on the dimensions
of the pipe or tube than the manufacturing chances. That is, in the case where the
dimensions of the pipe or tube at respective chances are the same as in this example,
it is considered that the correction result in the prior manufacturing chance can
be utilized in a more effective manner by setting coefficients such that the correction
amount based on the thickness result contributes more at timing when the manufacturing
chance is changed.