JP2003290894A - Method for detecting fluidized state in nozzle for transporting molten metal and this instrument - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a method for detecting the fluidized state in a nozzle for transporting molten metal by which the detection of mixing, etc., of the air into the nozzle for transporting the molten metal and the abnormal fluidized state of the molten metal in the nozzle, or in the case of providing a swirling blade, whether the swirl stream of the molten metal in the nozzle exists or not, etc., are directly judged from nozzle vibration. <P>SOLUTION: To the nozzle vibration having periodicity, a frequency analyzing processing is applied to the vibration of the nozzle for transporting the molten metal and also, with a feature in a frequency spectral distribution obtained from the result of this processing, the judgement of the abnormal fluidized state of the molten metal in the nozzle or whether the swirl stream of the molten metal exists or not, in the case of providing the swirling blade in the nozzle is performed. To the nozzle vibration having no periodicity, the analysis is performed with a phase-shift plane orbit figure. <P>COPYRIGHT: (C)2004,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for detecting a flow state in a molten metal transfer nozzle in which molten metal such as molten steel flows and an apparatus therefor.

[0002]

2. Description of the Related Art In a continuous casting apparatus for molten steel, molten steel flowing out of a tundish flows through a molten steel transfer nozzle and is discharged into a mold. This molten steel transfer nozzle has a three-layer sliding nozzle and a dipping nozzle. It is configured by vertically connecting in series. In continuous casting of molten steel using a continuous casting apparatus having this molten steel transfer nozzle, the prevention of defects in the cast has become an important issue with the demand for higher quality of the cast cast. The cause of this defect is that air is mixed in through the gap between the three-layer sliding nozzles to oxidize the molten steel and deteriorate the quality. In this case, it is considered that the flow of air becomes abnormal in the three-layer sliding nozzle and the immersion nozzle due to the inclusion of air. Further, due to uneven flow in the immersion nozzle, the flow of molten steel in the immersion nozzle becomes abnormal, and the molten steel discharged in the mold becomes unevenly flowed, resulting in quality deterioration. That is, when the molten steel discharged into the mold becomes a non-uniform flow, the molding powder is caught in the molten steel or inclusions such as air bubbles are deeply penetrated,
Since it is easy to be taken into the slab, it causes deterioration of quality. Generally, if the molten steel is filled and flowing in the immersion nozzle, the occurrence of uneven flow of the discharged molten steel is suppressed, but if the molten steel does not fill due to clogging of the immersion nozzle, etc. If the flow of molten steel in the immersion nozzle becomes abnormal, such as when a low-pressure air space occurs or uneven flow occurs in the immersion nozzle, it causes uneven flow of the discharged molten steel. Therefore, in such a case, it is necessary to promptly stop the injection of the molten steel into the immersion nozzle and take action such as removing the cause. Need to detect.
Therefore, conventionally, the presence or absence of molten steel in the immersion nozzle is detected by using electromagnetic force, or the change in magnetic permeability due to mixing of slag is detected electromagnetically to indirectly detect the immersion nozzle. I try to know the abnormal flow state of the molten steel inside.

Further, as a method of stabilizing the discharge flow so as to prevent uneven flow when the molten steel is discharged from the immersion nozzle into the mold, the method disclosed in JP-A-2000-237852 is used.
A method for making molten steel in a submerged nozzle into a swirling flow by devising a swirl vane as described in Japanese Patent No. 3120, has been devised and used. According to this method, by swirling the molten steel flow in the immersion nozzle, the speed of the discharge flow can be reduced even if the molten steel is injected into the nozzle at a high speed, and a stable discharge flow can be obtained. , Is effective. However, in this method, if the swirl vane provided inside the immersion nozzle is damaged or if slag or the like is clogged in the swirl vane, the swirl flow disappears and a stable discharge flow cannot be obtained. Will end up. Therefore, when the swirling flow disappears, it is necessary to immediately stop the injection of molten steel and take action such as removing the cause.
It is necessary to detect the presence or absence of swirling flow in the immersion nozzle. Therefore, conventionally, also in this case, the presence of the swirl flow is indirectly detected by electromagnetically detecting the change in the magnetic permeability in the immersion nozzle to detect whether slag is mixed in the immersion nozzle. I try to check the existence.

[0004]

However, in the method of indirectly checking the presence or absence of an abnormal flow state or swirl flow of molten steel in the immersion nozzle, it is possible to accurately determine whether or not there is an abnormal flow state or swirl flow. It is impossible to know, and in order to prevent the occurrence of defects in the slab, a method of directly detecting these has been required. Further, as described above, when air is mixed from the gaps of the three-layer sliding nozzle, the molten steel is oxidized and the quality is deteriorated. However, the abnormal state of the flow in the three-layer sliding nozzle or the immersion nozzle due to the mixing of air is caused. There was no way to detect it, and there was a need for that method. The present invention has been made in order to meet such a demand, and in a molten steel transfer nozzle including a three-layer sliding nozzle, a dipping nozzle, etc., mixing of air into the three-layer sliding nozzle and dipping A method for detecting an abnormal flow state of molten steel in the nozzle, or the presence or absence of swirling flow of molten steel in the immersion nozzle equipped with swirl vanes is directly determined from the flow characteristics of molten steel in the molten steel transfer nozzle. The present invention is intended to provide a method for detecting a flow state in a molten steel transfer nozzle and an apparatus therefor.

[0005]

[Means for Solving the Problems] The vibration generated from a molten steel transfer nozzle including a three-layer sliding nozzle and a dipping nozzle does not seem to be periodic when air is mixed from the three-layer sliding nozzle. is there. On the other hand, it is known that the vibration has a periodicity when the drift in the immersion nozzle that causes the drift of the molten steel discharged in the mold is caused. Therefore, the present invention has been made paying attention to this point, and with respect to the nozzle vibration having periodicity, the vibration analysis of the molten steel transfer nozzle, particularly the immersion nozzle, is performed along with the frequency analysis processing. The characteristics of the frequency spectrum distribution obtained from the results determine the abnormal flow state of molten steel in the immersion nozzle and the presence or absence of a swirl flow when the swirl vane is provided in the immersion nozzle. The present invention also considers shortening the time required for this detection and determination and improving accuracy. In addition, we have made unique efforts to eliminate nozzle vibration that does not have periodicity.
Since these points apply not only to molten steel but also to molten metal in general, molten metal in general is targeted in the following description.

Next, the method for detecting the flow state in the molten metal transfer nozzle of the present invention will be described. First, a method for detecting a flow state in the molten metal transfer nozzle, which mainly targets vibration of the molten metal transfer nozzle without periodicity, will be described. This method detects the nozzle vibration generated from the molten metal transfer nozzle, converts it into a nozzle vibration signal, draws a phase-shift plane trajectory diagram of this nozzle vibration signal, and captures the characteristics of the pattern of this phase-shift trajectory diagram. It is characterized by determining the flow state in the molten metal transfer nozzle.

In the above-described method for detecting the flow state in the molten metal transfer nozzle, a vibration detection sensor attached to the molten metal transfer nozzle or a microphone installed in the vicinity of the molten metal transfer nozzle is used to detect nozzle vibration. That is practical.

Next, a method for detecting the flow state in the molten metal transfer nozzle will be described, which is intended for periodic vibration of the molten metal transfer nozzle. This method detects the nozzle vibration generated by the flow of the molten metal in the molten metal transfer nozzle and converts it into a nozzle vibration signal, performs frequency analysis processing of this nozzle vibration signal, and obtains the frequency obtained from this frequency analysis processing. The feature is that the flow state of the molten metal in the molten metal transfer nozzle is determined by capturing the characteristics of the spectral distribution.

As a method embodying the above-mentioned flow state detecting method in the molten metal transfer nozzle, in the frequency spectrum distribution obtained from the frequency analysis processing, with respect to the average value of all peaks in all bands, If the average value of the peaks is equal to or higher than a certain ratio, there is a method of determining that a low pressure air space is generated in the flow of the molten metal in the molten metal transfer nozzle.

With this method, it is recommended that the fixed ratio be 250%.

Further, as a method embodying the above-mentioned flow state detecting method in the molten metal transfer nozzle, if the highest peak in the frequency spectrum distribution occurs at a specific frequency determined by the size of the molten metal transfer nozzle. There is a method of determining that the flow of the molten metal in the molten metal transfer nozzle is uneven.

Alternatively, as a method embodying the above-mentioned flow state detecting method in the molten metal transfer nozzle in the case where the molten metal transfer nozzle is provided with swirl vanes for making the flowing molten metal a swirl flow , In the frequency spectrum distribution, the value of the next highest peak is less than a certain ratio among the remaining peaks excluding the peaks due to the harmonics whose frequency is an integral multiple of this peak frequency. If so, there is a method of determining that a swirling flow exists.

In this method, it is recommended that the fixed ratio be 30%.

In each of the above methods, a predetermined number of data blocks formed by sampling nozzle vibration signals at a predetermined unit time to form a data string are continuously combined in the order of formation to form a set data block. Each time a new data block is formed, it is inserted at the beginning of the set data block, each subsequent data block is sequentially shifted backward, and the last data block is discarded, and the set At the end of the data block, a dummy data block composed of a predetermined number of data whose data values are all constant values is added to form a frame, and this frame is analyzed by using FFT. It is recommended to perform frequency analysis processing from the viewpoint of improving the processing speed and accuracy of detection. Here, FFT refers to fast Fourier transform.

In this method, it is convenient in actual processing to set the constant value of the dummy data block to 0.

Alternatively, in each of the above methods, a data block formed by sampling the nozzle vibration signal and forming a data string of a predetermined number of data obtained every predetermined unit time,
A set number of data blocks are consecutively combined in the order of formation to form a set data block. At this time, each time a new data block is formed, it is inserted at the beginning of the set data block, and each subsequent data block is sequentially moved backward. Shift,
While discarding the last data block and adding a first dummy data block formed of a predetermined number of data whose data values are all constant values to the end of the aggregate data block to form a first frame, at the same time, A second dummy data block composed of a predetermined number of data whose data values are all constant values is added to the end of the same data block inserted at the beginning of the aggregate data block, and the number of data is equal to the first frame. A second frame is formed, and further, using FFT, a first frequency analysis process for analyzing the first frame and a second frequency analysis process for analyzing the second frame are performed, and a geometric mean of the analysis process results of both is performed. It is recommended that the frequency analysis processing of the nozzle vibration signal be performed by obtaining the value from the viewpoint of further improving the detection processing speed and accuracy. That.

In this method, it is convenient in actual processing to set the constant values of the first dummy data block and the second dummy data block to 0.

Next, the flow state detecting device in the molten metal transfer nozzle of the present invention will be described. This apparatus includes nozzle vibration detection means for detecting nozzle vibration generated by the flow of molten metal in a molten metal transfer nozzle and converting it into a nozzle vibration signal, and frequency analysis processing means for performing frequency analysis processing of the nozzle vibration signal. It is characterized in that it is a determination means for determining the flow state of the molten metal in the molten metal transfer nozzle by capturing the characteristics of the frequency spectrum distribution obtained from the frequency analysis processing.

As a concrete example of the above-mentioned molten metal transfer nozzle in-flow state detecting device, the determining means is such that the determining means has all peaks in a specific band with respect to an average value of all peaks in all bands in the frequency spectrum distribution. There is a device that determines that a low-pressure air space is generated in the flow of the molten metal in the molten metal transfer nozzle if the average value of is more than a certain ratio.

With this apparatus, it is recommended that the fixed ratio be 250%.

Further, as a specific example of the above-mentioned molten metal transfer nozzle flow state detection device, the determination means has the highest peak in the frequency spectrum distribution at a specific frequency determined by the size of the molten metal transfer nozzle. If so, there is an apparatus that determines that the flow of the molten metal in the molten metal transfer nozzle is uneven.

Alternatively, as a specific example of the above-mentioned flow state detecting device in the molten metal transfer nozzle in the case where a swirl vane is provided for swirling the molten metal flowing in the molten metal transfer nozzle, In the frequency spectrum distribution, the determination means determines that the value of the highest peak is the next highest peak value among the remaining peaks excluding the peaks due to the harmonics having a frequency that is an integral multiple of the frequency of this peak. There is a device that determines that a swirl flow exists if the ratio is less than a certain ratio.

With this apparatus, it is recommended that the fixed ratio be 30%.

In each of the above devices, it is recommended to use a vibration detection sensor attached to the molten metal transfer nozzle to detect nozzle vibration.

Alternatively, in each of the above devices, a microphone installed near the molten metal transfer nozzle may be used to detect nozzle vibration.

In each of the above-mentioned devices, the frequency analysis processing unit combines a predetermined number of data blocks formed by sampling the nozzle vibration signal and forming a data string of a predetermined data number obtained every predetermined unit time in the order of formation. To form an aggregate data block, and each time a new data block is formed, this is inserted at the beginning of the aggregate data block, and the subsequent data blocks are sequentially shifted backward to the end data block. , A dummy data block composed of a predetermined number of data whose data values are all constant values is added to the end of the aggregate data block to form a frame, and this frame is analyzed using FFT. It is recommended to do so from the viewpoint of improving the detection processing speed and accuracy.

In this device, it is convenient in actual processing to set the constant value of the dummy data block to 0.

Alternatively, in each of the above-mentioned devices, the frequency analysis processing means continuously determines a data block formed by a predetermined number of data strings obtained by sampling the nozzle vibration signal every predetermined unit time in the order of formation. The number of data blocks is combined to form a set data block. At this time, each time a new data block is formed, it is inserted at the beginning of the set data block, and subsequent data blocks are sequentially shifted backward, and the last The data block is discarded, and at the end of the aggregate data block, a first dummy data block composed of a predetermined number of data whose data values are all constant values is added to form a first frame, and at the same time, the aggregate data block is formed. At the end of the same data block that was inserted at the beginning of the By adding a second dummy data block constituting a second number of data is equal to the first frame
A frame is formed, and further, using FFT, a first frequency analysis process for analyzing the first frame and a second frequency analysis process for analyzing the second frame are performed, and a geometric mean of the analysis process results of both is obtained. This is recommended from the viewpoint of further improving the detection processing speed and accuracy.

In this apparatus, it is convenient in actual processing to set the constant values of the first dummy data block and the second dummy data block to 0.

[0030]

Embodiments of the present invention will now be described in detail with reference to the drawings. In the embodiments of the present invention, molten steel will be described as the molten metal. Further, as the molten metal transfer nozzle, a molten steel transfer nozzle provided between the tundish and the mold, which is configured by connecting a three-layer sliding nozzle and a dipping nozzle in the vertical direction, will be described. In each of the examples described below, the nozzle vibration generated from the molten steel transfer nozzle is checked to check whether air is mixed in the three-layer sliding nozzle,
The abnormal flow state of molten steel in the immersion nozzle, or the presence or absence of swirling flow of molten steel in the immersion nozzle equipped with swirl vanes,
The present invention relates to a determination device. As described above, the nozzle vibration generated from the molten steel transfer nozzle does not seem to show periodicity when the air is mixed from the three-layer sliding nozzle. On the other hand, it is known that there is periodicity when the drift in the immersion nozzle that causes the drift in the molten steel discharged in the mold is caused. Therefore, first, the main part of the molten steel transfer nozzle is configured by checking the nozzle vibration with periodicity as a check target, and the flow state in the molten steel transfer nozzle with respect to the immersion nozzle equipped with a swirl vane inside is configured. The detector will be described, and then the molten steel transfer nozzle flow state detector for an immersion nozzle having no swirl vane inside will be described.
Then, after that, the case where the nozzle vibration having no periodicity is set as the check target will be described.

First, the flow state detecting device in the molten steel transfer nozzle of the first embodiment for the immersion nozzle having the swirl vane inside will be described. As described above, in continuous casting of molten steel, using a dipping nozzle provided with swirl vanes inside,
The swirling flow of the molten steel is generated in the inside to stabilize the discharging flow of the molten steel discharged from the immersion nozzle into the mold. However, the apparatus of the first embodiment is necessary for stabilizing the discharging flow of the molten steel. The presence or absence of swirling flow of molten steel in the immersion nozzle is detected. As described above, this detection is performed by checking the vibration frequency component of the immersion nozzle vibration by utilizing the fact that the vibration frequency component, which is the characteristic of the immersion nozzle vibration, differs depending on the presence or absence of the swirling flow of molten steel. . This vibration frequency component is checked by performing frequency analysis processing by checking the frequency spectrum obtained by Fourier transforming the nozzle vibration signal of the immersion nozzle, which is real-time data. As the actual processing work of this frequency analysis processing, in order to perform the processing work efficiently and easily, the nozzle vibration signal is converted into digital data by sampling, and this digital data is processed by the discrete Fourier transform (hereinafter Fourier transform). The FFT (Fast Fourier Transform) is used for this Fourier transform.

FIG. 1 is an explanatory view showing a state in which the apparatus of the first embodiment is used. In FIG. 1, 1 is an immersion nozzle, 2 is a swirl blade, 3a is a vibration sensor, 4 is a molten steel injection direction, 5 is a swirling direction of molten steel flow, 6 is a discharging direction of molten steel, 7 is a mold, 8 is molten steel, and 9 is a processing device. The immersion nozzle 1 has a swirl vane 2 inside. Among them, the apparatus of the first embodiment is composed of a vibration sensor 3a attached to the outer surface of the immersion nozzle 1 and a processing device 9 for receiving an output signal of the vibration sensor 3a as an input signal. In FIG. 1, the tundish and the three-layer sliding nozzle are not shown. The vibration sensor 3a has a function of detecting the nozzle vibration of the immersion nozzle 1 and converting it into a nozzle vibration signal which is an electric signal. The vibration sensor
In general, it has excellent heat resistance and signal transmission characteristics of a detection signal, and has an advantage that it can be easily fixed to the immersion nozzle by using an adhesive or the like. A microphone 3b may be used instead of the vibration sensor 3a. When the microphone 3b is used, there is an advantage that the sensitivity is low and high frequency, but it is necessary to detect the nozzle vibration of the immersion nozzle 1 as a sound. Therefore, it is necessary to install it in the vicinity of the immersion nozzle 1 without adhering it. The processing device 9 includes an amplifier 9a, a computing unit 9b, and a display 9c. The amplifier 9a has a function of amplifying the nozzle vibration signal output from the vibration sensor 3a. Arithmetic unit 9b
Is equipped with a microcomputer and a program used for it. It is converted into digital data by sampling nozzle vibration signals, FFT processing of this digital data, and swirling based on the processing result, as described below. It has functions such as determining the presence or absence of flow. Further, the display 9c has a function of displaying the result of the FFT processing and the determination result.

Next, the operation of the apparatus of the first embodiment will be described. FIG. 2 is a flowchart showing the operation of the device of the first embodiment. First, the vibration sensor 3a converts the vibration of the immersion nozzle 1 into a nozzle vibration signal which is an electric signal (S1). Next, an FF that converts the nozzle vibration signal into digital data and frequency analyzes the data using an FFT.
T processing is performed (S2). This digital data conversion and FF
There are the following problems in performing the T process.

In order to accurately detect the presence or absence of the swirling flow of molten steel in the immersion nozzle, it is necessary to secure the necessary frequency resolution in the FFT process. For that purpose, the nozzle vibration signal is digitized and sufficient. Amount of data is needed. Therefore, considering this point, assuming that the frequency resolution in FFT processing is Δf (Hz), the sampling frequency is fs (Hz), and the number of digital data is N (pieces), Δf = fs / N (1) Therefore, N = fs / Δf (2) Further, the time T (seconds) required to obtain N pieces of data is T = N / fs (3). Then, when a swirl flow exists in the immersion nozzle, the upper limit of the frequency included in the nozzle vibration signal of the immersion nozzle is about 10 Hz, so it is necessary and sufficient to set the sampling frequency fs to 20 Hz according to the sampling theorem. Is. Therefore, in order to accurately detect the presence or absence of the swirling flow of molten steel, it is considered necessary to secure 0.02 Hz as the frequency resolution Δf.
From equation (2), the number of required data N = 1000,
For convenience of processing, it is convenient to use the power of 2 as the power of the FFT processing. Therefore, if N = 1024, the time required to obtain the data is T = 51.2 from the above equation (3). Seconds. In other words, in order to secure 0.02 Hz as the frequency resolution Δf, it takes 51.2 seconds to collect necessary data, and if time such as FFT processing is added, it is possible to determine whether or not a swirling flow exists. , Will take even more time.

However, when the molten steel is injected into the immersion nozzle 1 equipped with the swirl vanes 2 and a swirl flow of the molten steel is generated in the immersion nozzle 1 and the swirl flow disappears during the supply of the molten steel to the mold 7, the If the time difference between the disappearance of the swirling flow and the time when the disappearance of the swirling flow is detected by the device is large, no matter how accurately the disappearance of the swirling flow is detected, the state where the swirling flow disappears continues during that period. As a result, there is an extremely high risk of causing defects that lead to defective cast pieces. Therefore, in order to improve the responsiveness, the time required to detect the disappearance of the swirling flow needs to be as short as possible, but as can be seen from the above calculation, in order to shorten the processing / judgment time required to obtain a judgment. When attempting to shorten the data acquisition time, the number of acquired data decreases, and it becomes difficult to maintain the required frequency resolution.

Therefore, as a method for solving this problem,
The method shown in FIG. 3 is adopted. in this way,
First, the nozzle vibration signal is sampled and converted into digital data, and one data block 11 is formed by using this data string obtained within a predetermined unit time. For example, if the sampling frequency is 20 Hz and the predetermined unit time is 1.6 seconds, 32 data can be obtained within the predetermined unit time.
One data block 11 is formed of 32 data strings of data. By changing the predetermined unit time, the number of data forming the data block 11 changes. In this case, the number of data forming the data block 11 can be set to one by setting the predetermined unit time to 50 msec. A predetermined number of the data blocks 11 are continuously combined in the order of formation to form a collective data block 12. Although the number is four in FIG. 3, the number is not limited to this and may be set to an optimum number according to each case. As shown in FIG. 3, this aggregated data block 12 is added to the beginning of the aggregated data block 12 formed immediately before each new data block 11 is formed, that is, every time a predetermined unit time elapses. The newly formed data block 11 is inserted and the last data block 11 is discarded, and a new aggregate data block 1
Form 2. That is, the data block 11 used once
Will be used to form the aggregate data block 12 from the next time onward, and the data block 11 will be reused. By reusing the data, the data amount is secured and the processing accuracy is improved. This is because. In the above example, the aggregate data block 12 is formed by combining four data blocks 11 each consisting of 32 pieces of data.
It consists of 28 data.

A dummy data block 13 composed of a predetermined number of data whose data values are all constant is added to the end of the aggregate data block 12 formed in this way to form a frame 14. The reason for adding such a dummy data block 13 is to increase the above-mentioned frequency resolution while saving the data acquisition time by securing a predetermined amount of data. The reason why all the data values used for the configuration of the dummy data block 13 are constant values is
This is because the dummy data block 13 is used as data of a signal having only a direct current component when viewed as real-time data, and the addition of the dummy data block 13 does not adversely affect the FFT processing result. The specific data value of the dummy data block 13 is 0 from the viewpoint of practical processing, but it may be a value other than 0. The number of pieces of data forming the dummy data block 13 is set so that the number of pieces of data forming the frame 14 after the dummy data block 13 is added is a power of 2 for convenience of FFT processing. In the above example, assuming that the number of data forming the frame 14 is 1024, the aggregate data block 12 is formed of 128 data, so the number of data forming the dummy data block 13 is the difference between the two. There are 896. Figure 4
3 shows the configuration of the frame 14 in the above example. The frame 14 thus formed is then F
FT processing is performed. That is, in the first embodiment, as can be seen from FIG. 3, the frame 14 is formed every time a new data block 11 is formed, that is, every time a predetermined unit time elapses, and the FFT process of the frame 14 is performed. .

Digital data of nozzle vibration signal and FF
When the T process is performed as described above, as described above, in the above example, as shown in FIG.
Hz, the predetermined unit time is 1.6 seconds, the number of data blocks 11 is 32, the number of data blocks 12 is 4, and the number of dummy data blocks 13 is
Since the number of data pieces in the frame 14 is 896 and 1024, and the frame 14 is FFT processed, 0.02 Hz can be secured as the frequency resolution Δf, and the required frequency analysis accuracy can be secured. It can be a good device. In addition, since the number of data in the data block 11 is 32 and the data that composes one data block 11 is collected in a predetermined unit time of 1.6 seconds, practically sufficient processing / determination is required. Time can be realized, and a device with good responsiveness can be obtained. Further, if the number of data samples of the actually measured nozzle vibration signals forming the aggregated data block 12 is too small with respect to the total number of data items forming the frame 14, the sampling interval can be secured even if the frequency resolution can be secured. Although it may be shorter than the cycle of vibration due to the swirling flow, the aggregate data block 12 is formed of 128 data, and it can be said that there is no such fear. In the above example, as shown in Fig. 5, the sampling frequency is 20H.
z, the number of data in the frame 14 is 1024, the number of data in the dummy data block 13 remains 896, the predetermined unit time is 6.4 seconds, and the number of data in the data block 11 is 128.
Alternatively, the number of data blocks in the aggregate data block 12 may be one. In this case, the frequency resolution etc.
Although it is similar to the above example, it takes a predetermined unit time of 6.4 seconds to collect the data that constitutes one data block 11, and therefore, there remains a problem in responsiveness.

In FIG. 2, when the nozzle vibration signal is converted into digital data and the FFT process (S2) is completed as described above, a frequency spectrum distribution is created (S3).
6 and 7 show this frequency spectrum distribution. FIG. 6 shows an example in which a swirl flow exists, and FIG. 7 shows an example in which no swirl flow exists. In the figure,
21 represents a spectrum. As can be seen from FIGS. 6 and 7, when a swirl flow exists in the immersion nozzle, a sharp peak with a significantly high vibration intensity appears in the curve of the frequency spectrum distribution at a specific frequency, whereas the swirl flow disappears. If it does not exist, such a sharp peak of vibration intensity does not appear, and the curve of the frequency spectrum distribution draws a flat curve as a whole. It is considered that the frequency spectrum distribution shows such a tendency for the following reason. When swirl flow exists in the immersion nozzle, the molten steel swirls at a specific angular velocity in the immersion nozzle, so vibration proportional to this specific angular velocity occurs in the immersion nozzle, and vibration at a specific frequency is significantly large. Become. As the vibration frequency proportional to the specific angular velocity, generally only the fundamental wave or this fundamental wave and its harmonics are generated. On the other hand, when the swirling flow in the immersion nozzle disappears and does not exist, the vibration proportional to the specific angular velocity does not occur, and the vibration of the specific frequency does not significantly increase.

Therefore, the presence or absence of the swirling flow can be determined by whether or not the curve of the frequency spectrum distribution obtained by the FFT process has a sharp peak with a remarkably large vibration intensity. As a specific method for making this determination, it is possible to make the determination by the following method as a result of actually analyzing a large number of nozzle vibration signals in the case where a swirl flow exists and the case where no swirling flow exists. I came to the conclusion.
This method, in the curve of the frequency spectrum distribution,
If the value of the second highest peak is less than a certain ratio among the remaining peaks excluding the peak due to the harmonics of the frequency of this peak, which is an integer multiple of the frequency of this highest peak This is a method of determining that a flow exists. As a comparison target, except for the peak due to the harmonic wave which is an integral multiple of the highest peak frequency, as mentioned above, the vibration frequency proportional to the specific angular velocity is not only the fundamental wave but also its harmonic wave. This is because it is necessary to remove the higher harmonics of the fundamental wave, because the noise is also generally generated.

This method will be specifically described with reference to the drawings in the case where only the fundamental wave and no harmonics are generated.
The flowchart of FIG. 2 and FIGS. 6 and 7 show this case. In FIGS. 6 and 7, the value of the maximum peak 32 (P1) in the curve of the frequency spectrum distribution
And the value (P2) of the second largest peak 33 among the remaining peaks is extracted (S4), and it is checked whether or not this P2 is below a certain ratio with respect to P1 (S5), If it is a fixed ratio or less, for example, 30% or less, it is a method of determining that a swirling flow exists (S6, S7). Conversely, if it exceeds a certain ratio, it is determined that the swirling flow has disappeared (S6, S8). Setting this fixed ratio to 30% is considered to be a reasonable value because it was obtained from actual analysis results, but this fixed ratio is not limited to 30%, and the optimum ratio according to each case You can set it to.

According to the apparatus of the first embodiment, since the vibration of the submerged nozzle is subjected to the frequency analysis processing to determine the presence or absence of the swirl flow in the submerged nozzle, the presence or absence of the swirl flow is characteristic of the swirl flow. Can be directly determined from. Also,
Since the nozzle vibration signal is sampled and converted into digital data and processed using the FFT, the processing work can be performed efficiently and easily. Further, the data block 11 that has been used once is reused for the formation of the aggregate data block 12 from the next time onward, and by reusing the data to secure the data amount, the processing accuracy can be improved. Further, a dummy data block 13 consisting of data of a constant value is added to the aggregate data block 12 to secure a predetermined data amount to form a frame 14, which is practically used while securing a necessary frequency resolution and the like. Therefore, it is possible to realize a sufficient processing / determination required time, and it is possible to provide a fluidized state detection device in the molten steel transfer nozzle having excellent accuracy and responsiveness.

In the first embodiment, as described above, the frame 14 is formed every time a predetermined unit time elapses, and the FFT processing of this frame 14 is performed. That is, the presence or absence of the swirling flow in the immersion nozzle is determined every time a predetermined unit time elapses. The new data used for the determination is only the head data of the aggregate data block 12, that is, the head data of the frame 14. Yes, the data after the head data is the data used for the determination before the previous time. The reason why the data used for the determination before the last time is used again is to improve the accuracy such as the frequency resolution as described above.
Strictly speaking, the frequency spectrum obtained by FFT processing the frame 14 including these data is not only the result of analysis of the situation of the nozzle vibration in the latest predetermined unit time but also the situation before that. However, the state of the latest nozzle vibration only during the predetermined unit time is not accurately analyzed and shown. For example, when the curve of the frequency spectrum distribution of FIG. 9 is obtained from the latest FFT processing result in the predetermined unit time,
The characteristic of the actual situation of the nozzle vibration in the latest predetermined unit time is that there is a peak 42 at the frequency f2, and the presence of the peak 41 at the maximum frequency f1 is before the latest predetermined unit time. In some cases, the result may reflect the situation. In this case, the method of the first embodiment cannot remove the presence of the peak 41 at the frequency f1, which is the influence of the situation before the predetermined unit time. In other words, the determination according to the latest situation occurring in the immersion nozzle has not been made.
Therefore, the flow state detecting device in the molten steel transfer nozzle of the second embodiment will be described next, which incorporates a device capable of removing the influence of the latest situation before the predetermined unit time as much as possible.

The configuration of the apparatus of the second embodiment is exactly the same as that of the first embodiment. The operation of the apparatus of the second embodiment is also shown in FIG.
The operation is exactly the same as that of the apparatus of the first embodiment shown in the flowchart of FIG. However, in the second embodiment, compared with the first embodiment, the nozzle vibration signal is converted into digital data and the FFT is performed.
We are working to improve accuracy and responsiveness by incorporating further innovations in processing. As the digital data of the nozzle vibration signal and the FFT processing, the method shown in FIG. 8 is adopted in the second embodiment. In this method, the method of forming the first frame 24 is the same as that of the frame 1 in the first embodiment except that the dummy data block is referred to as the first dummy block.
This is exactly the same as the forming method of No. 4. The difference is that, in addition to the first frame 24, a second frame 26 is formed, FFT processing is performed on the first frame 24 and the second frame 26, respectively, and then the processing results of both are geometrically averaged to obtain a frequency spectrum. is there.

The second frame 26 is the first frame 2
The same data block 21 used at the beginning of 4
And the same number of data as the first frame 24. As shown in FIG. 8, the method of forming the second frame 26 is performed by adding a new data block 2
1 is formed, that is, every time a predetermined unit of time elapses, the new data block 21 is used to form the first frame 24 and at the same time, the same data block 2 is formed.
At the end of 1, a second dummy data block 25 composed of a predetermined number of data whose data values are all constant values is added,
The second frame 26 is formed. The number of pieces of data forming the dummy data block 25 to be added is set so that the number of pieces of data forming the frame 26 after the addition of the dummy data block 25 is a power of 2 as in the first embodiment. However, as described above, the number of data in the second frame 26 needs to be the same as the number of data in the first frame 24. In addition, all the data values forming the second dummy data block 25 are constant values, like the first dummy data block 23. Here, as an example, similar to the first embodiment, the sampling frequency is 20 Hz and the predetermined unit time is 1.
6 seconds, the number of data of the data block 21 is 32, the number of data block of the set data block 22 is 4, the number of data of the first dummy data block 23 is 896, and the number of data of the first frame 24 is 1024. 1 frame 24
, The number of data in the second frame 26 becomes
Since there are 1024 frames, which is the same as the frame 24, the second frame 2
The number of data of the second dummy data block 25 added to 6 is 992, which is obtained by subtracting 32, which is the number of data forming one data block 21, from 1024. In addition, the data value forming the second dummy data block 25 is set to the first value.
Although it is set to 0 as in the embodiment, it may be set to a value other than 0.

As shown in FIG. 8, the first frame 24 and the second frame 26 thus formed are subjected to FFT processing, respectively, and then geometric mean calculation processing is performed to perform geometric mean of the processing results of both. A frequency spectrum is obtained from this result. The geometric mean is calculated by FF of the first frame 24.
The T processing result is F1, and the FFT processing result of the second frame 26 is F2. If the result of the geometric mean calculation process is F3,
F3 is obtained by F3 = (F1 × F2) ^1/2 (4). Based on the frequency spectrum obtained from F3, the same judgment as in the first embodiment is performed.

The operation and effect of using the result of the geometric mean calculation process described above will be described with reference to FIGS. 9, 10 and 11. Individual spectra are omitted in these figures. The curve of the frequency spectrum distribution based on the FFT processing result of the first frame 24 is based on the head data of the aggregate data block 22 in the latest predetermined unit time, that is, the head data of the frame 24 and the data after the head data. Is. Therefore, for example, the curve of the frequency spectrum distribution based on the FFT processing result of the first frame 24 shows that the characteristic of the actual situation of the nozzle vibration in the latest predetermined unit time is the frequency as shown in FIG. 9 exemplified above. It is assumed that the presence of the peak 42 at f2 and the presence of the peak 41 at the frequency f1 that is the maximum peak reflect the latest situation before the predetermined unit time. Then, the curve of the frequency spectrum distribution based on the FFT processing result of the second frame 26 formed based on only the head data of the aggregate data block 22 in the latest predetermined unit time, that is, the head data of the frame 24 is As shown in 10, there is a peak 43 at the frequency f2, and only the latest situation in the predetermined unit time is reflected. Therefore, the curve of the frequency spectrum distribution obtained from the result of the geometric mean of the FFT processing result of the first frame 24 and the FFT processing result of the second frame 26 is, as shown in FIG. 11, the peak 42 of the frequency f2 in FIG. Is highlighted as peak 44 and is present in FIG.
The peak 41 of the frequency f1 at is present in a reduced size. That is, the influence of the latest situation before the predetermined unit time can be removed as much as possible.

According to the apparatus of the second embodiment, the first frame 24 composed of the latest data of the predetermined unit time and the data before the first frame 24 and the first frame 24 composed of the data of the latest predetermined unit time only. After forming the two frames 26 and performing FFT processing on both of them, the presence or absence of the swirling flow is determined using the frequency spectrum obtained by geometrically averaging the processing results of both of them, so that the latest before the predetermined unit time can be determined. It is possible to remove the influence of the situation as much as possible, and it is possible to carry out the judgment according to the latest situation occurring in the immersion nozzle while maintaining the accuracy such as frequency resolution.

Next, the molten steel transfer nozzle in-flow state detecting device of the third embodiment for an immersion nozzle having no swirl vane inside will be described. FIG. 12 is an explanatory diagram showing a state in which the device of the third embodiment is used. In FIG. 12, the tundish and the 3-layer sliding nozzle are not shown. The configuration of this device is exactly the same as that of the first or second embodiment except that the immersion nozzle 10 having no swirl vane inside is used instead of the immersion nozzle having the swirl vane inside. . Therefore, the microphone 3b can be used instead of the vibration sensor 3a, and the vibration of the immersion nozzle 10 is converted into a nozzle vibration signal which is an electric signal to be digital data, and the data is FFT processed. Is. Further, not only the method of using the result of FFT processing of one frame, but also the method of using the result of geometric mean calculation processing of the FFT processing result of the first frame and the FFT processing result of the second frame can be used. Is. What is different from the first or second embodiment is the determination method based on the frequency spectrum distribution obtained as a result of these.

Then, this determination method will be described below. FIG. 13, FIG. 14, and FIG. 15 are the third embodiment,
It is an example of a curve of a frequency spectrum distribution obtained from the result of FFT processing of one frame. Individual spectra are omitted in these figures. Figure 1
3 is an example of the curve of the frequency spectrum distribution in the normal case, in which the vibration intensity is almost constant over the entire frequency band, and there is no portion where the vibration intensity is extremely large. Such vibration is referred to as white noise, which is considered to indicate that state when particular flow of the molten steel in the dipping nozzle is normal, because specific vibration with particularly high vibration intensity does not occur. . Therefore, as a determination method, the vibration intensity is almost constant over the entire frequency band,
When there is no portion where the vibration intensity is extremely large, a determination method for determining normal can be adopted.

FIG. 14 is an example of a curve of the frequency spectrum distribution when there is a low-pressure air space in the flow of molten steel in the immersion nozzle. The vibration intensity increases in the frequency band from frequency f11 to f12, but this identification No particular characteristic is found in the shape of the curve of the frequency spectrum distribution in the frequency band of. This state is supposed to be because the inside of the immersion nozzle 10 should be completely filled with molten steel. However, when a low-pressure air region such as a vacuum region occurs in a part of the immersion nozzle, the molten steel passes through this low-pressure air region. Since the phenomenon of falling straight like a waterfall occurs on the upper surface of the molten steel filled area below the low pressure air area, the vibration intensity at the specific frequency band increases due to the vibration that occurs at the upper surface of the molten steel filled area at that time. It is thought to be because. As a result of actually analyzing a large number of nozzle vibration signals in such a state, this state shows that the average value of all the peaks in all bands in the frequency spectrum distribution is the average value of all the peaks in a specific band. Was found to be above a certain ratio. Therefore, as the determination method, in the frequency spectrum distribution, the average value of all the peaks in a specific band should be a certain ratio or more, for example, 250% or more, with respect to the average value of all the peaks in all the bands. For example, it is possible to employ a determination method for determining that a low pressure air space is generated in the flow of molten steel in the immersion nozzle. This fixed ratio is not limited to 250%, and may be set to an optimum ratio according to each case.

FIG. 15 is an example of a curve of the frequency spectrum distribution when there is a drift in the flow of molten steel in the immersion nozzle, and the highest peak is present at the natural frequency f21 determined by the size of the immersion nozzle. . This is because when the above-mentioned low-pressure air space is generated and the space in this low-pressure air space becomes even larger, the molten steel does not fall straight onto the upper surface of the molten steel-filled area below the low-pressure air space, but obliquely to the inner wall surface of the immersion nozzle Since the phenomenon of falling while colliding occurs,
It is considered that when the inner wall of the immersion nozzle is obliquely hit, vibration of a specific frequency determined by the size of the immersion nozzle occurs. Therefore, as a determination method, in the frequency spectrum distribution, if the highest peak occurs at a specific frequency determined by the size of the immersion nozzle, it means that the flow of molten steel in the immersion nozzle has a drift. A judgment method for judgment can be adopted.

FIG. 16 is a flow chart showing the operation of the apparatus of the third embodiment. S11 of this flowchart
Up to S13 are exactly the same as S1 to S3 in the flowchart of the first embodiment of FIG. In addition, in the device of the third embodiment, the presence or absence of drift in the above description is first determined (S14 to S16), and then the presence or absence of the low pressure airspace is determined (S17 to S21). , And is determined to be normal (S22).

In the third embodiment, the result of FFT processing of one frame is used.
The result of the geometric mean calculation processing of the FT processing result and the FFT processing result of the second frame may be used.

As can be seen from the above description, according to the apparatus of the third embodiment, the presence or absence of the low pressure air space of the flow of the molten steel in the immersion nozzle is analyzed by frequency analysis of the vibration of the immersion nozzle,
Since the presence or absence of uneven flow of the molten steel flow in the immersion nozzle is determined, it is possible to provide a molten steel transfer nozzle flow state detection method or apparatus that can be directly determined from the characteristics of the flow in the immersion nozzle.

Next, a flow state detecting device in the molten steel transfer nozzle according to the fourth embodiment will be described, which is intended for vibrations generated from the molten steel transfer nozzle and having no periodicity. FIG. 17 is an explanatory diagram showing a state in which the device of the fourth embodiment is used. In FIG. 17, 51 is a tundish, 52 is a porous refractory, 53 is a three-layer sliding nozzle, 54 is a three-layer sliding nozzle drive cylinder, 55 is a dipping nozzle, 56 is a mold, 57 is a microphone, and 58 is a processing device. And 59 is molten steel. The microphone 57 is installed in the vicinity of the three-layer sliding nozzle 53, and mainly detects a vibration sound generated by the three-layer sliding nozzle 53 and converts it into a vibration signal. Instead of the microphone 57, a vibration detection sensor attached to the three-layer sliding nozzle 53 may be used. Processor 5
8 receives the vibration signal output from the microphone 57 and performs processing. Therefore, the processing device 58 is equipped with a microcomputer having a memory and a necessary program.

In FIG. 17, the three-layer sliding nozzle 53 is a three-layer sliding nozzle drive cylinder 54.
Thus, in order to adjust the amount of molten steel 59 supplied from the tundish 51 to the mold 56, the middle part of the three layers slides. Therefore, the middle part of the three layers vibrates due to this slide, and a vibration sound is generated by this vibration. In the three-layer sliding nozzle 53, for example, air may be mixed into the three-layer sliding nozzle 53 from its slide surface, but in that case, the vibration sound may change. Therefore, by capturing the nozzle vibration sound generated by the three-layer sliding nozzle 53 with the microphone 57, converting it into a nozzle vibration signal, and drawing a phase shift plane trajectory diagram of this vibration signal, that is, a trajectory, the processing device 58 Three
It is checked whether or not air is mixed in the layer sliding nozzle 53.

This trajectory is generated by removing noise in the nozzle vibration signal from the direct current component and commercial power supply frequency band using a filter, then sampling this nozzle vibration signal, and based on this sampling data, The axis is the signal level of the nozzle vibration signal, and the vertical axis is the time derivative of this nozzle vibration signal. This trajectory draws a clear trajectory when the nozzle vibration signal has periodicity, that is, a periodic phenomenon, but clears when the nozzle vibration signal does not have a periodicity, that is, a roughly periodic phenomenon. No trace is drawn. Here, in the case of drawing a clear trajectory, that is, in the case of a periodic phenomenon, it is possible to analyze by the method using the frequency analysis described above, but in the case of not drawing a clear trajectory, that is,
It has been found that in the case of the roughly periodic phenomenon, it is difficult to analyze by the method using the frequency analysis described above. The above-mentioned check for the presence or absence of the inclusion of air needs to be targeted for this roughly periodic phenomenon.

Therefore, as a method for checking the presence or absence of air inclusion, the trajectory in the case of air inclusion is stored in the processing device in advance as a reference trajectory and the nozzle vibration signal drawn by the microphone 57 is drawn. Compare the inspection trajectory with this reference trajectory. As a method of comparison, each of the above trajectories is drawn on a lattice plane where the horizontal axis and the vertical axis are orthogonal,
The trajectory is represented as a pattern composed of each lattice through which the trajectory passes. FIG. 18 shows an example of the trajectory pattern drawn in this way. Then, in each lattice of the reference trajectory and the inspected trajectory, the distance between the center coordinates of the lattices that are closest to each other is obtained, and this total is defined as the deviation degree. If the reference trajectory and the inspected trajectory pass through the same grid, the distance at that grid is zero. Then, when the deviation degree is within a certain range, it is determined that the inspected trajectory matches the reference trajectory.

According to the apparatus of the fourth embodiment, even if the nozzle vibration generated from the molten steel transfer nozzle shows an approximately periodic phenomenon, it is possible to analyze.

In the above description of the embodiment, as the molten steel transfer nozzle provided between the tundish and the mold,
Although the description has been made for the three-layer sliding nozzle and the dipping nozzle, a tundish nozzle 62 using a stopper 63 as shown in FIG. 19 can also be targeted. In this case, clogging is likely to occur near the entrance of the tundish nozzle 62, and it is conceivable that the presence or absence of clogging will be the center of detection of the flow state. In the figure, 61 is a tundish, 64 is a stopper holder, 65 is a stopper drive cylinder, 66 is molten steel, and 67 is
Is a microphone, and 68 is a processor. A vibration detection sensor may be used instead of the microphone 67.

Further, in addition to the molten steel transfer nozzle provided between the tundish and the mold, the molten steel transfer nozzle between the ladle (ladle) and the tundish can be targeted. FIG. 20 is directed to the long nozzle 72 of the ladle 71 provided between the ladle (ladle) 71 and the tundish 73. In this case, since a drift in the nozzle is likely to occur, this drift It can be considered that the presence or absence of the flow is the center of detection of the flow state. Alternatively, when the molten steel 74 is completely filled in the long nozzle 72, an unstable factor such as an excessively high flow velocity of the molten steel in the long nozzle 72 may occur. It can be the target of detection. In addition, in FIG.
4 is molten steel, 75 is a microphone, and 76 is a processing device. Also in this case, a vibration detection sensor may be used instead of the microphone 75. Further, as shown in FIG. 21, a dipping pipe 83 between a ladle (ladle) 81 and a tundish 84 can also be targeted. In this case, air entrapment easily occurs. It can be considered to be the center of detection of the flow state.
In the figure, 82 is a short nozzle, 85 is molten steel, 86
Is a microphone, and 87 is a processor.

Although molten steel is used as the molten metal in the above description of the embodiments, the present invention is not limited to this, and the present invention can be applied to other molten metals such as molten aluminum.

[0064]

According to the first or second aspect of the present invention, it is possible to provide a method for detecting a flow state in a molten metal transfer nozzle that can be analyzed even when the nozzle vibration shows a roughly periodic phenomenon.

According to the third or thirteenth aspect of the present invention, since the abnormal flow state in the molten metal transfer nozzle is determined by performing frequency analysis processing on the vibration of the molten metal transfer nozzle, the abnormal flow state is transferred to the molten metal transfer nozzle. It is possible to provide a method for detecting a flow state in a molten metal transfer nozzle or an apparatus therefor, which can be directly determined from the characteristics of the flow in the nozzle.

According to the present invention, the vibration of the molten metal transfer nozzle is subjected to frequency analysis processing to determine whether or not there is a low pressure air space for the flow of the molten metal in the molten metal transfer nozzle. It is possible to provide a method for detecting a flow state in a molten metal transfer nozzle or an apparatus therefor capable of directly determining the occurrence of a low pressure air region from the characteristics of the flow in the molten metal transfer nozzle.

According to the sixth or sixteenth aspect of the present invention, the vibration of the molten metal transfer nozzle is subjected to frequency analysis processing to determine whether or not there is a drift of the molten metal flow in the molten metal transfer nozzle. It is possible to provide a method for detecting a flow state in a molten metal transfer nozzle or an apparatus therefor capable of directly determining the presence or absence of a drift in the molten metal transfer nozzle from the characteristics of the flow in the molten metal transfer nozzle.

According to the invention described in claim 7, 8, 17 or 18, since the presence or absence of the swirl flow in the molten metal transfer nozzle is determined by performing the frequency analysis processing on the vibration of the molten metal transfer nozzle, the presence of the swirl flow is detected. It is possible to provide a method for detecting a flow state in a molten metal transfer nozzle or an apparatus therefor, which can directly determine the presence or absence of a swirl flow based on the characteristics of the swirl flow.

According to the invention of claim 9, 10, 21 or 22, the nozzle vibration signal is sampled and converted into digital data, and is processed using the FFT, so that the processing work can be performed efficiently and easily. You can Further, the data block that has been used once is reused for the formation of the aggregate data block from the next time onward, and the processing precision can be improved by reusing the data to secure the data amount. In addition, a dummy data block consisting of fixed-value data is added to the aggregate data block to secure a predetermined amount of data to form a frame. It is possible to provide a method for detecting a flow state in a molten metal transfer nozzle or an apparatus therefor, which can realize processing / determination required time and is excellent in accuracy and responsiveness.

According to the eleventh, twelfth, twenty-third, or twenty-fourth aspect of the present invention, the first frame composed of the latest data of the predetermined unit time and before that and the data of the latest data of the predetermined unit time only. After forming the second frame configured and performing FFT processing on both of them, the abnormal flow state in the molten metal transfer nozzle is determined using the frequency spectrum obtained by the geometric mean of the processing results of both. , The influence of the latest situation before the predetermined unit time can be removed as much as possible, and the determination in accordance with the latest situation occurring in the molten metal transfer nozzle can be performed while maintaining the accuracy such as frequency resolution. A method for detecting a flow state in a metal transfer nozzle or an apparatus therefor can be provided.

According to the nineteenth aspect of the invention, since the vibration sensor which is easily fixed to the molten metal transfer nozzle and has excellent heat resistance and signal transmission characteristics of the detection signal is used for detecting the nozzle vibration, the sensor is used. It is possible to provide a method or apparatus for detecting a fluidized state in a molten metal transfer nozzle, which is easily fixed to the molten metal transfer nozzle and has excellent heat resistance and signal transmission characteristics of a detection signal.

According to the twentieth aspect of the present invention, since the microphone is used to detect the nozzle vibration, it is possible to provide a method for detecting a flow state in the molten metal transfer nozzle with high sensitivity at a low frequency or an apparatus therefor.

[Brief description of drawings]

FIG. 1 is an explanatory view showing a state in which a molten steel transfer nozzle in-flow state detecting device of a first embodiment is used.

FIG. 2 is a flowchart showing the operation of the molten steel transfer nozzle in-flow state detecting device of the first embodiment.

FIG. 3 is an explanatory diagram of digitizing nozzle vibration signals and FFT processing in the first embodiment.

FIG. 4 is an explanatory diagram showing a configuration example of a frame in the first embodiment.

FIG. 5 is an explanatory diagram showing another configuration example of the frame in the first embodiment.

FIG. 6 is a graph showing an example of frequency spectrum distribution when a swirling flow is present in the first embodiment.

FIG. 7 is a graph showing an example of a frequency spectrum distribution when a swirling flow does not exist in the first embodiment.

FIG. 8 is an explanatory diagram of digitizing nozzle vibration signals and FFT processing in the second embodiment.

FIG. 9 is a graph showing an example of frequency spectrum distribution based on the FFT processing result of the first frame in the second embodiment.

FIG. 10 is a graph showing an example of frequency spectrum distribution based on the FFT processing result of the second frame in the second embodiment.

FIG. 11 is a graph showing an example of a frequency spectrum distribution based on a result of geometric mean calculation processing of the FFT processing result of the first frame and the FFT processing result of the second frame in the second embodiment.

FIG. 12 is an explanatory view showing a state in which the molten steel transfer nozzle inflow state detecting device of the third embodiment is used.

FIG. 13 is a graph showing an example of a frequency spectrum distribution when the flow of molten steel in the immersion nozzle is normal in the third embodiment.

FIG. 14 is a graph showing an example of frequency spectrum distribution in the case where a low pressure air region exists in the flow of molten steel in the immersion nozzle in the third embodiment.

FIG. 15 is a graph showing an example of a frequency spectrum distribution when a drift exists in the flow of molten steel in the immersion nozzle in the third embodiment.

FIG. 16 is a flowchart showing the operation of the molten steel transfer nozzle inflow state detecting device of the third embodiment.

FIG. 17 is an explanatory view showing a state in which the molten steel transfer nozzle inflow state detecting device of the fourth embodiment is used.

FIG. 18 is a graph showing an example of a trajectory pattern of the fourth embodiment.

FIG. 19 is an explanatory view showing a state in which the flow state detecting device is used for a tundish nozzle using a stopper.

FIG. 20 is an explanatory view showing a state in which the flow state detecting device is used for the long nozzle of the ladle.

FIG. 21 is an explanatory view showing a state in which the flow state detecting device is used for the ladle dipping tube.

[Explanation of symbols]

1 immersion nozzle 2 swirl vanes 3a Vibration sensor 3b microphone 4 Molten Steel Injection Direction 5 Turning direction of molten steel 6 Discharge direction of molten steel 7 mold 8 Molten steel 9 processing equipment 9a amplifier 9b Operation unit 9c display 10 Immersion nozzle 11 data blocks 12 aggregate data blocks 13 Dummy data block 14 frames 21 data blocks 22 aggregate data block 23 First dummy data block 24 First Frame 25 Second dummy data block 26 Second Frame 31 spectrum 32 maximum peak 33 second peak 41 Peak at frequency f1 42 Peak at frequency f2 43 Peak at frequency f2 44 Peak at frequency f2 51 tundish 52 Porous refractory 53 3-layer sliding nozzle 54 3-layer sliding nozzle drive cylinder 55 Immersion nozzle 56 mold 57 microphone 58 processor 59 Molten Steel 61 Tundish 62 tundish nozzle 63 stopper 64 stopper holder 65 Stopper drive cylinder 66 molten steel 67 microphone 68 Processor 71 Ladle 72 Long nozzle 73 Tundish 74 Molten Steel 75 microphone 76 Processor 81 ladle 82 short nozzle 83 Immersion tube 84 Tundish 85 Molten Steel 86 microphone 87 Processor

   ─────────────────────────────────────────────────── ─── Continued front page    F-term (reference) 2G047 AA01 BA04 BC04 CA07 EA08                       EA11 GD02 GG09 GG12 GG24                       GG28                 4E004 FB10 MB20

Claims (24)

[Claims] 1. A nozzle vibration generated from a molten metal transfer nozzle is detected and converted into a nozzle vibration signal, and a phase shift surface trajectory diagram of this nozzle vibration signal is drawn. A method for detecting a flow state in a molten metal transfer nozzle, characterized by determining a flow state in the molten metal transfer nozzle. 2. The melt according to claim 1, wherein a vibration detection sensor attached to the molten metal transfer nozzle or a microphone installed in the vicinity of the molten metal transfer nozzle is used for detecting the nozzle vibration. A method for detecting a flow state in a metal transfer nozzle. 3. The nozzle vibration generated by the flow of the molten metal in the molten metal transfer nozzle is detected and converted into a nozzle vibration signal, and the nozzle vibration signal is subjected to frequency analysis processing, and is obtained from the frequency analysis processing. A method for detecting a flow state in a molten metal transfer nozzle, characterized in that the flow state of the molten metal in the molten metal transfer nozzle is determined by capturing the characteristics of the frequency spectrum distribution. 4. In the frequency spectrum distribution, if the average value of all peaks in a specific band is a certain ratio or more with respect to the average value of all peaks in all bands, melting in the molten metal transfer nozzle is performed. The method for detecting a flow state in a molten metal transfer nozzle according to claim 3, wherein it is determined that a low pressure air space is generated in the flow of metal. 5. The method for detecting a flow state in a molten metal transfer nozzle according to claim 4, wherein the fixed ratio is 250%. 6. In the frequency spectrum distribution, if the highest peak occurs at a specific frequency determined by the size of the molten metal transfer nozzle, a drift of the molten metal flow in the molten metal transfer nozzle occurs. The molten metal transfer nozzle flow state detecting method according to claim 3, wherein 7. A swirl vane is provided in the molten metal transfer nozzle for swirling the flowing molten metal, and a swirl vane is provided for the highest peak value in the frequency spectrum distribution. Among the remaining peaks excluding the peaks due to the harmonics having a frequency that is an integral multiple of the frequency, if the value of the next highest peak is less than a certain ratio, it is determined that the swirling flow is present. Item 4. A method for detecting a flow state in a molten metal transfer nozzle according to Item 3. 8. The method for detecting a flow state in a molten metal transfer nozzle according to claim 7, wherein the fixed ratio is 30%. 9. A set data block is formed by continuously combining a predetermined number of data blocks formed by a predetermined number of data rows obtained by sampling the nozzle vibration signal every predetermined unit time, in order of formation. At that time, each time the new data block is formed, it is inserted at the beginning of the aggregate data block, each subsequent data block is sequentially shifted backward, and the last data block is discarded, At the end of the aggregate data block, a dummy data block composed of a predetermined number of data having all constant data values is added to form a frame, and the frame is analyzed using an FFT, whereby the nozzle The molten metal transfer nozzle according to any one of claims 3 to 8, wherein the frequency analysis processing of the vibration signal is performed. The inner flow state detection method. 10. The method for detecting a flow state in a molten metal transfer nozzle according to claim 9, wherein the constant value of the dummy data block is set to 0. 11. A set data block is formed by continuously combining a predetermined number of data blocks formed by a predetermined number of data sequences obtained by sampling the nozzle vibration signal every predetermined unit time, in order of formation. At that time, each time the new data block is formed, it is inserted at the beginning of the aggregate data block, each subsequent data block is sequentially shifted backward, and the last data block is discarded, At the end of the aggregate data block, a first dummy data block composed of a predetermined number of data whose data values are all constant values is added to form a first frame, and at the same time, it is inserted at the beginning of the aggregate data block. At the end of the same data block as the above, the second data composed of a predetermined number of data, all of which have constant values. By adding over data blocks, the number of data to form a second frame equal to the first frame, further, by using the FFT,
By performing a first frequency analysis process of analyzing the first frame and a second frequency analysis process of analyzing the second frame, and obtaining a geometric mean of the analysis process results of both, the frequency of the nozzle vibration signal is calculated. The method for detecting a flow state in a molten metal transfer nozzle according to any one of claims 3 to 8, wherein an analysis process is performed. 12. The method for detecting a flow state in a molten metal transfer nozzle according to claim 11, wherein the constant values of the first dummy data block and the second dummy data block are set to 0. 13. Nozzle vibration detection means for detecting nozzle vibration generated by the flow of molten metal in a molten metal transfer nozzle and converting it into a nozzle vibration signal, and frequency analysis processing means for performing frequency analysis processing of the nozzle vibration signal. And a determination means for determining the flow state of the molten metal in the molten metal transfer nozzle by capturing the characteristics of the frequency spectrum distribution obtained from the frequency analysis processing, and the flow in the molten metal transfer nozzle. State detection device. 14. The molten metal, if the average value of all peaks in a specific band is a certain ratio or more with respect to the average value of all peaks in all bands in the frequency spectrum distribution. 14. The molten metal transfer nozzle inflow state detecting device according to claim 13, wherein it is determined that a low pressure air space is generated in the flow of the molten metal in the transfer nozzle. 15. The molten metal transfer nozzle in-flow state detecting device according to claim 14, wherein the fixed ratio is 250%. 16. The molten metal in the molten metal transfer nozzle is determined by the determination means if the highest peak in the frequency spectrum distribution occurs at a specific frequency determined by the size of the molten metal transfer nozzle. 14. The molten metal transfer nozzle inflow state detecting device according to claim 13, wherein it is determined that the flow is unbalanced. 17. A swirl vane for turning a molten metal flowing in the molten metal transfer nozzle into a swirling flow is provided, and the judging means is characterized in that in the frequency spectrum distribution,
With respect to the value of the highest peak, among the remaining peaks excluding the peak due to the harmonics having a frequency that is an integral multiple of the frequency of this peak, if the value of the next highest peak is below a certain ratio, then the turning The molten metal transfer nozzle inflow state detecting device according to claim 13, wherein it is determined that a flow exists. 18. The apparatus for detecting a flow state in a molten metal transfer nozzle according to claim 17, wherein the fixed ratio is 30%. 19. The flow state detection in the molten metal transfer nozzle according to claim 13, wherein a vibration detection sensor attached to the molten metal transfer nozzle is used to detect the nozzle vibration. apparatus. 20. The molten metal transfer nozzle in-flow state detecting device according to claim 13, wherein a microphone installed near the molten metal transfer nozzle is used for detecting the nozzle vibration. . 21. The frequency analysis processing unit continuously combines a predetermined number of data blocks formed by a predetermined number of data sequences obtained by sampling the nozzle vibration signal every predetermined unit time in a forming order. Forming an aggregate data block, at which time each time the new data block is formed, it is inserted at the beginning of the aggregate data block, and each subsequent data block is sequentially shifted backward, While discarding the data block, at the end of the aggregate data block,
21. A frame is formed by adding a dummy data block composed of a predetermined number of data whose data values are all constant values, and the frame is analyzed using FFT. The apparatus for detecting a fluidized state in the molten metal transfer nozzle according to 1. 22. The molten metal transfer nozzle inflow state detecting device according to claim 21, wherein the constant value of the dummy data block is set to 0. 23. The frequency analysis processing means continuously combines a predetermined number of data blocks formed by a predetermined number of data sequences obtained by sampling the nozzle vibration signal every predetermined unit time in a forming order. Forming an aggregate data block, at which time each time the new data block is formed, it is inserted at the beginning of the aggregate data block, and each subsequent data block is sequentially shifted backward, While discarding the data block, at the end of the aggregate data block,
A first dummy data block composed of a predetermined number of data whose data values are all constant is added to form a first frame, and at the same time, the end of the same data block that is inserted at the beginning of the aggregate data block. In addition, a second dummy data block composed of a predetermined number of data whose data values are all constant values is added to form a second frame in which the number of data is equal to the first frame, and further, using FFT, A first frequency analysis process for analyzing the first frame and a second frequency analysis process for analyzing the second frame
The molten metal transfer nozzle inflow state detecting device according to any one of claims 13 to 20, wherein frequency analysis processing is performed and a geometric mean of both analysis processing results is obtained. 24. The molten metal transfer nozzle inflow state detecting device according to claim 23, wherein the constant values of the first dummy data block and the second dummy data block are set to 0.