JPS63168262A - Estimating method for surface flaw in cast slab - Google Patents

Abstract

PURPOSE:To directly observe molten surface in a mold and the behavior of lubricating agent by detecting the mold surface variation arranging photo conductor at the corresponding part of meniscus and estimating the development of surface flaw as compared with the allowable variation. CONSTITUTION:An optical fiber 6a is used as the photo conductor and the obtd. observing picture is inputted to a PGB treating device 21a and in succession, treated to be binarized by a binarization device 21b. The boundary line between the molten steel 12 and the lubricating agent is operated by a boundary line arithmetic unit 21c through the treated signal, to find the molten surface. In a molten surface variation calculating device 22, the condition of variation for the molten surface near the meniscus is quantitatively detected. The allowable variation for eliminating the development of surface defect is inputted to a comparator, to compare with the inputted detected molten surface variation through the picture treating device 21 and when the detected molten surface variation exceeds the allowable variation, the probability for developing the surface defect is high. In this case, alarm is given by a warning device 27 or in some case, warning signal is inputted to a process control device 24, to control changing and supplying of the lubricating agent.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a method for rapidly and accurately estimating surface defects of a slab in continuous casting, particularly in a mold, in an online state.

[Conventional technology]

As is well known, a mold for continuous casting is composed of a combination of long sides and short sides. Molten steel is injected into the mold to form a solidified shell of a predetermined thickness on the surface in contact with the inner surface of the mold. Slabs are manufactured continuously from below. Therefore, the molten steel begins its initial solidification within the mold. This initial solidification state has a significant effect on the occurrence of surface defects and other quality of the slab.

The mold is provided with vibrations that reciprocate in the casting direction in order to efficiently flow a lubricant such as powder between the inner surface of the mold and the solidified shell and to prevent seizure between the solidified shell and the inner surface of the mold. If the type of lubricant mentioned above, its supply status, vibration conditions, etc. are not properly controlled according to the operating conditions, pinholes may be generated in the solidified shell, or the lubricant, slag, or other In extreme cases, various defects may occur, including cracks on the surface.

Such defects remain as surface defects on the manufactured slab and cause sludge, sliver flaws, cracks, etc. during subsequent rolling steps. Moreover, in extreme cases, it may induce a breakout and lead to a situation where the operation becomes impossible.

Particularly in recent years, active efforts have been made to increase casting speeds and to directly connect continuous casting and rolling processes (hereinafter referred to as direct rolling), so the occurrence of the defects described above is a problem that is difficult to implement. But it is a big obstacle.

However, conventionally there is no way to directly detect the conditions inside the mold. For example, by observing the surface properties of the slab after it has been manufactured, it is possible to determine the suitability of past casting conditions, in other words, the suitability of the selected lubricant. It was common to judge the supply state or the vibration conditions such as the frequency and amplitude of the mold.

On the other hand, for defects such as seizure and cracking of the solidified shell, for example, as shown in Japanese Patent Application Laid-Open No. 57-115, a plurality of temperature detection ends are buried in the casting direction of the wall surface of the mold, and these temperature detection ends are used. A method of estimating the occurrence of casting defects by comparing measured temperature values with each other or with a reference value, or a method of measuring temperature using multiple temperature detection terminals in the casting direction as shown in Japanese Patent Application Laid-Open No. 61-200453. Methods have been proposed to perform Fourier transform on the values, recognize patterns of temperature changes within the mold, and estimate the type and location of casting defects from this, and some have been put into practical use.

[Problem that the invention seeks to solve]

As mentioned above, in the conventional method, there is no way to directly detect the situation inside the mold, especially the behavior of the molten metal level near the meniscus and the lubricant, and it is necessary to estimate the past casting condition by looking back from the manufactured slab. It was common to make judgments. For this reason, during casting, for example, fluctuations in the level of the molten metal or waves occurring in the molten metal surface itself, insufficient supply of lubricant, uneven inflow, etc., could not be directly detected online. Therefore, it has not been possible to implement appropriate controls corresponding to these, and the reality has been to rely on empirical operations based on past operational results.

In addition, in all of the above conventional methods for detecting various casting defects in the mold, the temperature of the buried portion is measured with a temperature detection end buried in the wall of the mold, and the surface defects of the slab are estimated from the measured temperature value. It was an indirect method. For this reason, even though no defects have occurred, it has often been mistakenly recognized as a surface defect, and there have been cases in which casting defects have been overlooked. Such erroneous recognition and oversight of casting defects have a great effect on direct rolling, and there is a major problem in that it cannot be carried out efficiently.

The present invention aims to fundamentally solve the above-mentioned problems.

[Means for solving problems]

In order to solve the above-mentioned problems, the present invention attaches a light guide connected to an image processing device to a through hole drilled in a portion corresponding to a meniscus on a long side or a short side of a continuous casting mold. The amount of fluctuation in the melt level is detected from the observation image obtained by directly observing the melt level during casting through a conductor, and
The detected amount of fluid level fluctuation is compared with the allowable amount of fluctuation under the operating conditions, which is set based on the correlation between the amount of fluid level fluctuation determined in advance and the occurrence of mold inner surface defects, and the detected amount of fluid level fluctuation is determined by This is a method for estimating surface defects in slabs in continuous casting, which is characterized by estimating the occurrence of surface defects when an allowable variation amount is exceeded.

[For production]

It has been learned from past operational experience that when the vibration conditions of the mold are set to a small amplitude and a high frequency, the depth of the oscillation mark is reduced, its shape improves, and the surface quality of the slab improves. There is. However, it was conventionally thought that this was due to the following reasons. That is, to explain the case where the mold is vibrating sinusoidally as shown in Fig. 3, the amplitude of this vibration is A, the frequency is f,
If the casting speed is Vc and the time is t, the displacement X of the mold is given by the following equation (1).

x = Asin (2πft) ”... (1) The velocity Vm of mold 1 is calculated by differentiating equation (1) with respect to time t, and Vm = dx
/dt = 2 πf Acos (2zft) −・
”・(2) Here, if the descending speed Vm of the mold 1 is greater than the casting speed Vc, the slab will be pushed downward from the mold. It is said that the smaller the negative strip time (hereinafter referred to as TN time), the better the surface and surface layer quality of the slab.This TN time is determined by the following equation (3), and the amplitude A is small and the frequency f
It can be seen that the larger the value, the smaller the TN time.

TN = (1/πf)・cos-' (Vc/2π
fA) (3) It is also known that in order to reduce the TN time, non-sinusoidal vibration may be used instead of the sinusoidal vibration.

However, such reasoning is based on the above-mentioned molten metal surface (in the present invention, the contact area between the mold inner surface 1a and molten steel shown in FIG. .Also, this continuous surface of the hot water surface is called the hot water surface level 15) is stationary with respect to the absolute coordinates.
That is, it is assumed that the molten metal level within the mold 1 does not fluctuate even if the mold 1 vibrates. It is common knowledge among those skilled in the art that in the conventional mold 1, the inner surface 1a of the mold and the molten steel completely slip due to the action of the lubricant, and that even if the mold 1 moves up and down, the molten metal level remains constant. Based on this basic understanding, we carefully observed the interior of the mold from above the mold and found that ripples were occurring on the surface of the mold, even though the level was stable. Due to this, the present inventors had doubts about the conventional basic understanding, and conducted various experimental studies on a method for detecting fluctuations in the hot water level near the meniscus during actual operation, and succeeded in detecting it.

First, a method of detecting the amount of fluctuation in the molten metal level during continuous casting in the present invention will be explained.

FIG. 1 is a cross-sectional structural diagram of the vicinity of a mold for explaining the basic configuration for actually measuring the amount of variation in the level of hot water, and FIG. 2 is a perspective view of the mold. In the figure, 1 is a mold, which is composed of a long side 2 and a short side 3.

In this example, a through hole 5 penetrating to the inner surface of the meniscus portion 4 of the long side 2 is bored. A light guide 6 is attached to the through hole 5 with a heat-resistant glass 7 interposed therebetween. The light guide 6 can detect the brightness and/or wavelength of the light inside the mold, for example, an optical fiber, a combination of a spectrometer and a photomultiplier, a COD element combining a filter and a lens, an MO3 type element, an ITV camera, etc. It is possible to use However, in the experience of the present inventors, an optical fiber that allows the diameter of the through hole 5 to be made small and also allows for a large observation field of view is the most excellent.

The heat-resistant glass 7 prevents leakage of the molten steel 112 and protects the light guide 6 from the heat of the molten steel 12.

The light guide 6 is connected to an image processing device 21 whose functions will be described below. 9 is a vibration generator that applies vibration to the mold l; 10 is a vibration control device that controls the frequency, amplitude, etc. of the vibration generator 9. Further, 11 is a nozzle for injecting molten steel 12 into the mold 1, 3 is a slab, and 14 is a solidified shell formed on the surface layer of the slab 13.

Next, an example of a processing mechanism in the image processing device 21 will be explained.

If, for example, an optical fiber 6a is used as the light guide 6, an observed image obtained by the optical fiber 6a is input to the RGB processing device 21a. Optical fiber 6a in this example
Assuming that the number of pixels is 30,000, each detection signal from each pixel input to the RGB processing device 21a is binarized using a preset reference value in the binarization processing device 21b following the RGB processing device 21a. It is well known. For example, the brightness emitted by the temperature of molten steel can be expressed in terms of well-known lux, etc., and that value can be used as a reference value. That is, the lubricant 17 supplied into the mold 1 covers the surface of the melt m412 as shown in FIG.
As it melts, it flows between the molten steel 12 and the mold inner surface 1a and exerts a lubricating function. This molten lubricant 1
The temperature of No. 7 is about 1400″C, which is about 152″C of the molten steel temperature.
It is slightly lower than 0°C. However, lubricant 1
7 has CaO and 5iOz as its main components, and has transparency like glass, so when directly observed with an optical fiber 6a, its brightness is lower than that of molten steel 12. Therefore, if the signal of the observed image is binarized using the brightness of the molten steel 12 as a reference, for example, parts with a brightness higher than the brightness of the molten steel 12 will be displayed white, and conversely, parts with a brightness lower than the brightness of the molten steel 12 will be displayed black for discrimination. I can do it.

On the other hand, since the thickness of the lubricant is thin, there are cases in which the surface of the lubricant is substantially at the level of the hot water level without causing any actual harm. There is a significant brightness difference between the molten lubricant 17 and the undissolved lubricant 17a. Therefore, in such a case, the brightness of the molten lubricant may be used as the reference value. In addition, the input observation image is divided into an arbitrary number of blocks, and - vertical differential and horizontal differential operations are performed on the image between each block.
For example, the calculation result is binarized based on whether it is greater than or less than the luminance difference between the lubricant and molten steel as a reference value, or it is output only when the luminance corresponding to the boundary between molten steel and lubricant is input. Further, it is possible to adopt a method of performing binarization processing based on, for example, a red wavelength of the molten steel 12. The processing in the binarization processing device 21b is a general term for such methods.

The signal thus processed by the binarization processing device 21b is used to calculate the boundary line between the molten fj 112 and the lubricant 17 in the boundary line calculation device 21c, and the level of the molten metal is determined.

As shown in FIG. 4, when the hot water level 15 within the observation field Y of the optical fiber 68 is determined continuously or at an arbitrary period, the amount of displacement of this hot water level 15 at preset parts is calculated one after another. By doing so, it is possible to detect the amount of variation in the hot water level at the relevant location. FIG. 6 shows an example of a specific calculation method, where the abscissa represents the distance l from the inner surface of the mold 1a to a predetermined interval or a set point, and the molten metal level 15 is represented as the ordinate (h). The hot water surface level 1 is expressed by the time h for a predetermined abscissa (setting position).
By calculating the difference Δh in the hot water level 15a after a lapse of unit time from 5, the amount of fluctuation in the hot water level per unit time is obtained. In FIG. 1, reference numeral 22 denotes a hot water level fluctuation amount calculation device for calculating this hot water level fluctuation amount.

The liquid level fluctuation calculation device 22 performs the above-mentioned calculations, for example, at intervals of 1 to 3 decorations from the mold inner wall surface, or for each specific region such as a region 3 cylinders and 10 mm from the mold inner wall surface. This makes it possible to quantitatively detect fluctuations in the hot water level near the meniscus. The amount of variation in the level of the molten metal in the mold detected in this way will be defined and used as the amount of variation in the molten metal level detected hereinafter.

As a result of repeatedly detecting the fluctuations in the molten metal level during casting from the observation images using the aforementioned light guide 6 and conducting repeated research, the present inventors discovered that the molten metal level is not static, which is completely different from the conventional basic understanding. First, we obtained new knowledge that it is synchronized with the vibration of the mold. That is, (1) the hot water level is not stationary, and with time =
This level fluctuation is not only due to the balance between the amount of molten steel injected into the mold and the withdrawal of slabs, but also due to surface waves due to mold vibration. The relative fluctuation (the movement of the molten metal surface obtained by the light guide 6 attached to the mold 1) has the same period and opposite phase as shown in FIG. 7(b). (3) The molten metal surface relative to the mold The absolute fluctuation (the movement of the hot water surface when mold vibration is canceled out from the relative movement and the earth is taken as an absolute reference) has the same period and same phase as shown in Figure 7 (C), and the amplitude of the hot water surface is 70-80% of the mold amplitude.

Next, based on the above findings, the present inventors investigated whether there was any correlation between defects that actually occurred on the surface of the slab (hereinafter referred to as surface defects) and the amount of fluctuation in the melt level.

Under the casting conditions shown in Table 1, the short side 3 of the mold long side 2
Optical fibers 6a were attached to the meniscus-equivalent portions of the joints (4 locations at each corner where slab surface defects are likely to occur), and the amount of variation in the level of the molten metal in the 3-barrel portion was detected from the mold inner surface 1a. Simultaneously with the detection of the amount of variation in the level, the operational data at that time, such as casting speed, level, vibration speed of the mold, etc., is inputted from the data acquisition device 25 shown in FIG. and simultaneously recorded on the recording device 26. As a result, when a normal slab with no surface defects was obtained, the detected amount of fluctuation in the molten metal level was approximately 70% of the mold amplitude on average, although there was some variation.

Table 1 On the other hand, it was also confirmed that when a surface defect occurs, the amount of detected melt level fluctuation tends to increase compared to when it is normal. In other words, if the detected level fluctuation is large, it means that the level at the meniscus is large (it does not fluctuate), which means that the initial solidification is stable.Therefore, the initial solidification at the meniscus is stable. It was confirmed that this method is more advantageous for surface defects.

FIG. 8 shows an example of the results of a concrete investigation of the correlation between the above-mentioned melt level fluctuation amount and surface defects under the above-mentioned operating conditions. The occurrence of surface defects is expressed as an index of the number of surface defects such as pinholes and inclusions per meter of slab (number of defects/m) (hereinafter referred to as surface defect index).
In addition, the amount of fluid level fluctuation is the ratio of the mold amplitude to the amount of detected fluid level fluctuation, that is, the absolute amount of fluctuation described above is expressed as an index (hereinafter referred to as the amount of fluid level fluctuation index), and is shown on the vertical and horizontal axes, respectively. .

As is clear from FIG. 8, when the hot water level variation index becomes 5 or more, the surface defect index rapidly increases, indicating that surface defects are likely to occur.

By determining the correlation between the amount of fluctuation in the melt level and the occurrence of surface defects corresponding to each operating condition based on past operating results, the allowable amount of fluctuation to eliminate the occurrence of surface defects within the mold can be set. Is possible.

Once this allowable variation amount is set, for example, the comparator 2 in FIG.
3, the value is inputted and compared with the detected amount of water level fluctuation inputted via the image processing device 21.

As is clear from the above explanation, if the detected amount of variation in the melt level exceeds the allowable amount of variation, the probability of surface defects occurring is high.

Therefore, in that case, the alarm device 27 may issue an alarm, and depending on the case, the generated signal may be input to the process control device 24, and the process control device 24 may, for example, change the type of lubricant, control the supply amount, or vibrate the lubricant. It is also possible to issue a command to the control device 10 to automatically take corresponding measures such as changing the mold vibration conditions.

The detected hot water level fluctuation amount is detected by the image processing device 21 or by the hot water level fluctuation calculation device 22 before being input to the comparator 23, as described above, as a hot water level fluctuation index that incorporates the mold amplitude at that time, or as a relative fluctuation amount. It is also possible to express it as
It is sufficient to understand the amount of variation that shows the most significant correlation with the occurrence of surface defects.

〔Example〕

The present invention was carried out in a curved continuous casting machine with a monthly production capacity of 160,000 tons.

In this example, the apparatus shown in Fig. 1 was used, and 13 mm through-holes were made in the meniscus-equivalent part of the part where the long side of the mold joins with the short side (4 places at each corner where slab surface defects are likely to occur). This through hole has 50,000 pixels and a diameter of 12 mo.
A + optical fiber was attached.

The observation field of the optical fiber was 50°. Silicon glass 10 mm thick was fitted onto the inner surface of the through-hole mold to prevent leakage of molten steel and protect the optical fiber. The recording device 26 also records the casting speed.

The level of the hot water, the vibration speed of the mold, etc. were recorded from the data acquisition device 25 via the hot water level variation calculation device 22.

The casting conditions in this example are as shown in Table 1.

The correlation between the amount of fluid level fluctuation and the occurrence of surface defects under the above casting conditions is as shown in FIG. 8. Therefore, in this example, the allowable amount of fluctuation is also expressed as a fluid surface fluctuation index, and its value is set to 5. On the other hand, the observed image from the optical fiber 6a is processed by the image processing device 21.
The above-mentioned process was carried out, and the amount of variation in the level of the molten metal at a location 3rm away from the inner surface of the mold was detected.

The operation continued as described above, and the accuracy of estimating the occurrence of surface defects and the occurrence of surface defects in the manufactured slabs were investigated.

Figure 9 compares the accuracy of estimating the occurrence of surface defects with the method of the present invention and the conventional method that indirectly estimates the occurrence of surface defects using the temperature measurement value of the detection end embedded in the mold. . In this example, a notice of surface defects is issued during casting, and when the part becomes a slab, it is investigated whether or not surface defects actually occur. I judged it by how warm it was. As is clear from Part 9, the behavior of the meniscus, which has an important influence on the occurrence of surface defects, is directly observed, and the amount of fluctuation in the melt level is detected based on the signal from the observed image to estimate the occurrence of surface defects. By implementing the present invention, it was possible to significantly reduce false alarms compared to conventional methods.

Next, FIG. 1O shows an example of the results of investigating the incidence of defects in products when direct rolling is performed, in comparison with the conventional method. As can be seen from FIG. 10, the accuracy of detecting surface defects by implementing the present invention was higher than that of the conventional method, and the occurrence of defects in the final product could be reduced.

In this example, the optical fibers 6a were installed at four locations at the corners of each long side where surface defects are likely to occur, but the installation locations may be set as appropriate depending on the type of steel to be cast and casting conditions. For example, it may be attached to the short side 3 without any problem.

〔Effect of the invention〕

By implementing the present invention, it becomes possible to directly observe the behavior of the molten metal level and lubricant in the mold. In addition, by detecting the amount of fluctuation in the melt level based on the direct observation results, and by correlating the amount of fluctuation in the mold surface with surface defects based on past operational results, it is possible to check the occurrence status of surface defects in the mold in an online state. This makes it possible to estimate accurately and quickly. As a result, precise operational actions can be taken according to the situation, and normal slabs without surface defects can be manufactured. In addition, since the number of misrecognitions has been drastically reduced, the implementation rate of the above-mentioned direct rolling will be significantly improved.

[Brief explanation of the drawing]

Fig. 1 is a cross-sectional structural diagram of the vicinity of the mold to explain the basic configuration of the present invention, Fig. 2 is a perspective view of the mold, Fig. 3 is a diagram illustrating an example of the displacement state of the mold, and Fig. 4 Figure 5 is a partial cross-sectional view of the vicinity of the meniscus; Figure 6 is a diagram showing an example of a specific method for calculating the amount of fluctuation in the level; Figure 7 The figures are diagrams comparing the displacement status of the mold and the corresponding amount of fluctuation in the melt level. Figure 7(a) shows the displacement status of the mold due to an approximate sine wave, and Figure 7(b) shows the displacement status of the mold due to the approximate sine wave. Figure 7 (C) shows the relative variation in the surface, and Figure 7 (C) shows the absolute variation in the melt level. Figure 8 shows an example of the results of investigating the relationship between the variation in the melt level and the occurrence of surface defects. Figure 10 is a comparison of the accuracy of estimating the occurrence of surface defects using the conventional method and the method of the present invention, which indirectly estimate the occurrence of surface defects using the temperature value of the detection end embedded in the mold. FIG. 2 is a diagram illustrating an example of the results of investigating the incidence of defects in products during rolling in comparison with a conventional method. 1... Mold, 1a... Mold inner surface, 2... Long side, 3
... short side, 4 ... meniscus equivalent part, 5 ... through hole, 6 ... light guide, 6a ... optical fiber, 7 ...
- Heat-resistant glass, 9... Vibration generator, lO... Vibration control device, 11... Nozzle, 13... Slab, 14.
... Solidified shell, 15.15a ... Hot water level, 16...
・Meniscus, 17.17a...Lubricant, 21...
Image processing device, 21a...RGB processing device, 21b.
...Binarization processing device, 21c...Boundary line calculation device, 2
2... Hot water level fluctuation amount calculation device, 23... Comparator, 24
... process control device, 25 ... data collection device,
26... Recording device, 27... Alarm device. Agent Patent attorney Masamitsu Aki Sawa and 1 other person 71'3 Figure 7F4 off view time C3eC)

Claims (1)

[Claims] (1) A light guide connected to an image processing device is attached to a through hole drilled in the meniscus-equivalent part of the long side or short side of the continuous casting mold, and the molten metal level during casting is monitored via this light guide. The amount of fluid level fluctuation is detected from the observation image obtained by direct observation of A method for estimating surface defects in a slab in continuous casting, comprising comparing the amount of variation with an allowable amount of variation under the operating conditions, and estimating the occurrence of a surface defect when the detected amount of variation in the melt level exceeds the amount of variation. .