JPS6133447B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a shape detection method in strip rolling such as hot rolling.

In strip rolling, it goes without saying that shape defects (hereinafter referred to as "shape") caused by differences in elongation rates in the width direction are important control and management items in terms of operation and quality. Research and development of shape detection methods has been carried out on many lines, and some lines have put them into practical use, and systems that automatically control the shape have also been developed. For example, in cold rolling lines, etc., the shape is elastically deformed due to the tension applied to the strip, becoming latent and not appearing on the surface. Therefore, shape detection is possible by measuring the tension distribution in the width direction in some way. It has been put into practical use.

However, in the case of a hot rolling line, etc., unlike the above case, the shape becomes latent and is observed as a wave that appears on the surface. For this reason, attempts have been made to measure the passing position of the strip using a rod light source, laser light source, water column resistance method, etc., and to grasp the size of the irregularly shaped wave from the temporal movement thereof.

However, in reality, in most cases, large disturbances such as flapping are added to the actual measurement data, as seen before the strip wraps around the coiler in a hot rolling line.

In this way, during actual rolling, the movement of the strip is observed to include disturbances other than the signals caused by waves with defective shapes, but in the past, processing for these disturbances was insufficient, and there were problems as a shape detection method. It was effective only in cases where there was a small number of cases.

SUMMARY OF THE INVENTION An object of the present invention is to provide a method for detecting a shape by extracting only a signal due to a shape defect from a measurement signal even in a situation where there is a disturbance such as fluttering of a plate. In the strip shape detection method of the present invention, during strip rolling, the passing position of the strip is measured multiple times over at least one period of the plate wave at a plurality of observation points arranged in the width direction, and is memorized. Abnormal values are removed from the measurement signal for each point, and the difference between the maximum and minimum values (fluctuation range) for each observation point is determined from the remaining measurement values, and the smallest fluctuation among the fluctuation widths for each observation point is calculated. The method is characterized in that the width is determined, and the wave height value of the plate wave at each observation point is determined from the difference between the fluctuation width of each observation point and the minimum fluctuation width.

A detailed explanation will be given below based on embodiments of the present invention. Figure 1 shows the overall device configuration.
1 is the strip whose shape is to be detected, 2 (subscript _-1 ,
- ₂ ... is a device that measures the passing position of the strip, and multiple units are installed in the width direction (or scanning may be performed to perform multiple samplings). The passing position of the strip at each point (passing position in the height direction; hereinafter referred to as displacement as it can be considered as displacement from a fixed position) is detected. 3 is a displacement signal detected by the detector, 4 is an interface for inputting the detection signal into a microcomputer, 5 is a microcomputer that performs signal processing, and 6 is a signal obtained as a result of signal processing by the microcomputer. This is the shape signal that is generated and becomes the output signal.

FIG. 2 shows a practical example of a detector signal, in which the horizontal axis represents time, the vertical axis represents the magnitude of displacement, and 6-1 to 6-7 represent the device 2- ₁
~ _2--7 represents the detection signal. The detection signal Hi
(t) is the wave hi(t) due to shape defects at each point, as well as the fluctuation Ni of displacement due to strip flapping, etc.
(t) and measurement noise εi(t), and is generally expressed by equation (1).

Hi(t)=Ni(t)+hi(t)+εi(t) (1) Here the desired signal is hi(t) and Hi
(t) as a shape signal produces erroneous results.

Figure 3 shows an example of erroneous results that occur when Hi(t) is used as a shape signal. However, the antiphase curve 8-1
~ When receiving a disturbance (dotted line) as shown in 8-3, a large wave is generated on the other DS (drive side) as a displacement as shown by curves 9-1 to 9-3, and this signal is used as the shape. This shows the opposite result of DS kananobi. Note that C is the center. From the above, in order to correctly detect the shape, a method of extracting only the shape signal from the displacement signal is essential.

The main object of the present invention is to provide a method for extracting only a shape signal by removing components due to plate vibration disturbance and measurement noise from a signal observed including disturbance due to unknown vibrations. According to the inventors' study based on actual data and physical considerations, if the number of measurement points in the width direction exceeds a certain level, there is at least one point where no irregular shaped wave occurs (hereinafter referred to as a reference point). It is clear that the fluctuation width is minimum at the reference point in the time interval of one wave period or more.The fluctuation of the strip can be considered to be almost uniform in the width direction of the plate.It is clear that the shape of the plate does not change within a fairly short observation period. We have found that this is guaranteed with sufficient accuracy for practical use. Therefore, if such characteristics are used, the shape signal can be obtained by finding the displacement fluctuation range in the time interval described above at each point in the width direction, and then subtracting the minimum fluctuation width from it.

In the embodiment of the present invention, there are seven measurement points in the width direction, and since fluctuations due to unknown shape defects are observed several times per second, the measurement period is set to 1 second. Displacement measurements were performed 200 times within this time, and abnormal values were removed to determine the variation range of displacement. FIG. 4 shows the example (results). The vertical axis is displacement x, and the horizontal axis is probability P.
(x).

A displacement signal is input to the microcomputer 5 for each displacement measurement, and the distribution of displacement for one second is calculated by calculating the probability of taking a value larger than that displacement value (the number of measurements of the number of displacement values larger than that value for each value 200 It is calculated as a ratio of The maximum displacement value x _nax of normal data from the measured data is P in Figure 4 a.
With x such that (x)=0.1, and the maximum displacement value x _n
Find _ax with x such that P(x) = 0.9 of b, and x
Measured data larger than _nax and measured data smaller than x _nio are removed as abnormal values. Through this processing, disturbances such as instantaneous fluctuations are removed. From this, the displacement fluctuation range Δx during that time is determined by Δx=x _nax −x _nio (2). The removal of abnormal values does not need to be limited to the above method.

If the fluctuation range of displacement at each point is found as Δx(i), (i is the measurement point number), then the minimum fluctuation range Δx _o
This is the fluctuation range due to disturbance. Therefore, the shape wave height value S(i) at each point is determined by S(i)=Δx(i)−Δx _o (3). Here, the shape wave height value has the meaning of the average wave height value of the wave with a defective shape within the shape detection time.

FIG. 5 shows an example of the results obtained by performing the above shape detection method. The horizontal axis represents the measurement point position in the width direction of the strip, and the vertical axis represents the displacement value of the strip at each point and the obtained shape peak value.
Polygonal lines 11 and 14 represent the maximum and minimum values of displacement, polygonal lines 12 and 13 represent displacements x _nax and x _nio , and polygonal line 15 represents the finally determined shape peak value.

The shape wave height value is used for monitoring as a shape signal value for shape control or for management purposes.
In the above embodiment, the shape signal (shape peak value) is extracted using a microcomputer.
The present invention can also be implemented with other signal processing devices, and is not limited to microcomputers.

According to the present invention, it is now possible to measure the shape of a plate even when there are disturbances such as fluttering of the strip, such as on a hot rolling line, and the shape peak value can now be determined at regular intervals. In addition, this method determines the variation range of displacement and determines the shape based on this, so the actual shape can be measured without affecting the warp or inclination of the plate. With this shape detection method, the shape can be measured during rolling even in hot rolling lines, etc., making it possible to detect the shape, thereby automating operations,
This greatly contributes to improving quality.

[Brief explanation of the drawing]

FIG. 1 is an overall device configuration diagram of an embodiment of the present invention;
Figure 2 is a waveform diagram showing a practical example of a detector signal, Figure 3 is a waveform diagram showing an example of erroneous results when Hi(t) is used as a shape signal, and Figure 4 is an abnormal value. A graph showing an example of a removal method, and FIG. 5 is a graph showing an example of a shape detection result. 1... Strip, 2... Passing position measuring device, 3...
displacement signal, 4...interface, 5...microcomputer, 6...output signal.

Claims (1)

[Claims] 1. In strip rolling, the passing position of the strip is measured multiple times over at least one period of the plate wave at multiple observation points arranged in the width direction and stored, and abnormal values are detected from the measurement signal at each observation point. The difference between the maximum value and the minimum value (fluctuation range) for each observation point is determined from the remaining measured values, and the minimum fluctuation range is determined from among the fluctuation ranges for each observation point. A method for detecting the shape of a strip, characterized in that the peak value of a plate wave at each observation point is determined from the difference between the width and the minimum fluctuation width.