JPH07286218A - Device for deciding melt down of scrap - Google Patents

Abstract

PURPOSE:To provide a device for deciding melt-down of scrap which is capable of preventing overflow of the scrap by erroneous decision and a failure of a measuring instrument, assuring the quality of stable operation and improving efficiency without requiring the skill and experience of skilled persons in spite of a decrease in manpower and has a function corresponding to a skilled operator. CONSTITUTION:This device is provided with an ITV camera 6 which monitors the in-furnace condition mounted in the cap section of an electric furnace, a microphone which is mounted at the outer peripheral part of the furnace body and collects the sound and noises generated from arcs and a making electric power managing device 10 which manages the electric power to be made to the electric furnace. The signals from at least :2 devices among these devices are inputted to a fuzzy inference arithmetic unit 15. The degree of progression of the melt-down of the scrap us determined by executing fuzzy inference in accordance with the membership function determined for the respective inputs in the fuzzy inference arithmetic unit 15 and is announced to the operator.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an electric furnace facility for melting scrap, and more particularly to determining scrap burn-through after initial charging of scrap and scrap burn-through after additional charging of scrap. Regarding the device.

[0002]

2. Description of the Related Art Scrap burn-through determination in the operation of an electric furnace depends on the intuition and speculation of a skilled operator who has many years of experience in operation, because it is difficult to observe the inside of the furnace currently in operation. ing.

The operator determines the proper voltage or current amount to be applied to the electrode from the amount of scrap loaded at the beginning of melting.
The power supplied to the scrap is adjusted according to the progress of melting during the middle period and the final period. That is, the operator comprehensively makes a scrap burn-through determination and switches operating conditions by referring to information such as voltage during energization, stability of current, sound generated from arcs generated during energization, and state of the electric furnace. ing.

FIG. 6 shows an example of the timing of charging scrap and the pattern of input power. In FIG. 6, the horizontal axis represents the operating time and the vertical axis represents the operating voltage.

After the initial charging of scrap, operation is performed at low voltage and small current for t1 hours (1) Boring period (I), t
Operate at high voltage and high current for 2 hours (2) Melt period (I), then add additional scrap and operate at low voltage and low current for t3 hours (3) Boring period ( I
I), operate at high voltage and large current for t4 hours (4)
After the melting period (II) and before scrap burn through, operate at low voltage and large current for t5 hours (5) After the heating phase, after scrap burn through determination (II), low for t6 + t7 hours Operate with voltage and large current (6) After the refining period, steel is tapped. During temperature measurement sampling during refining, the power is temporarily stopped.

Further, as disclosed in Japanese Patent Publication No. 55-17314, there is an example of a scrap burn-through judging device for judging scrap burn-through by acoustically analyzing only a sound generated from an arc. This is because the microphone is installed in contact with the furnace wall, the oscilloscope is connected through an electric circuit, the loudness of the furnace discharge is converted into a current, and this is detected as the detected current waveform of the oscilloscope. Is intended to be determined by utilizing the fact that it is large during the melting of scrap and becomes small when it melts down.

[0007]

[Problems to be Solved by the Invention] In the electric furnace industry, it is difficult to secure a labor force due to the severe working environment, and problems such as aging of workers and lack of manpower are becoming serious. Skilled operators need to understand the furnace condition,
Scrap burn-through is comprehensively determined based on information obtained from the senses of the eyes, ears, and the body, the amount of applied electricity, the stability of voltage and current, and the operating knowledge obtained from years of experience and intuition.

In recent years, a DC arc furnace has been popularized in which a large amount of electric power is supplied to rapidly and efficiently melt in order to increase the efficiency of melting in an arc furnace and improve productivity, and a large number of units are adopted. It's starting to happen. However, in such an arc furnace, if the melting situation is erroneously determined and the burn-through is determined earlier, the amount of scrap that can be charged into the furnace will be less than the specified amount because the scrap remains unmelted. There are problems such as scrap overflow, temperature measurement sampling device melting and remaining scrap colliding with damage, and a momentary delay in burnout judgment leads to wasteful power consumption and a large amount of heat loss. In operations that rely only on the guesswork of the experienced operators of
It is difficult to maintain the operating efficiency of the arc furnace at a high level and operate it.

In the acoustic device of Japanese Patent Publication No. 55-17314, the burn-through is determined by catching the point where the amplitude of the current conversion waveform of the detected sound changes from large to small, but not the absolute value. , We must capture the relative change in amplitude. Therefore, a good result has not yet been obtained because the amplitude of the current conversion waveform is regarded as an absolute value and it is difficult to make an electrical determination.

Therefore, an object of the present invention is to provide a function equivalent to that of a skilled operator for judging burn-through, to secure stable operation quality even if manpower is reduced, and to improve efficiency. It is an object of the present invention to provide a scrap burn-through determination device for performing scrap burn-through determination.

[0011]

According to the present invention, the above-mentioned object is a scrap burn-through determination device for melting scrap, which recognizes the unmelted portion of the scrap and the molten steel / slag portion. At least one of an image processing device for determination, a power supply management device that manages the power supplied to the scrap by applying a voltage from the power supply to the electrode, and a sound analysis device for sound generated from the arc of the electric furnace during energization This is achieved by a scrap burn-through determination device configured to include a fuzzy inference arithmetic unit that receives output signals from two devices and performs fuzzy inference to determine scrap burn-through.

The scrap burn-through determination device requires at least two signals out of the above-mentioned three output signals, and the combination thereof can be the following four cases. (1) Image processing device and input power management device (2) Image processing device and sound analysis device (3) Input power management device and sound analysis device (4) Image processing device, input power management device and sound analysis device

[0013]

In the scrap burn-through determination device, the image from the ITV camera at the upper part of the electric furnace is binarized by the detection means of the image processing device for the unmelted scrap portion and the molten steel / slag portion (dark portion ... Black: 0 , Bright part ... white:
By performing 1), the ratio of molten steel / slag portion on the screen is determined, and it can be estimated that the scrap is unmelted if the dark portion is large, and the scrap is unmelted if the bright portion is spread over almost the entire surface.

Then, the current, voltage, and
The electric power detection means determines how much electric power is applied to the scrap input amount, and the situation of scrap burn-through can be estimated from this data and the relationship between the scrap input amount and the electric power prepared in advance. .

Further, the amplitude of the generated sound gradually decreases from the large amplitude based on the detection signal of the size of the generated sound from the arc by the detection means of the acoustic analysis device for the arc generated sound, so that the scrap is generated. Can be presumed to have melted down.

In order to improve the estimation accuracy among the signal output of the ratio of "0" and "1" of the image binarization processing, the amount of input electric power with respect to the amount of scrap input, and the volume of sound generated from the arc. At least two output signals are input to and the fuzzy inference is performed in the inference part to obtain the degree of burn through.

In order to perform fuzzy inference, at least two or more signals among the above three output signals are required, and the combination thereof can be the following four cases. (1) Image processing device and input power management device (2) Image processing device and sound analysis device (3) Input power management device and sound analysis device (4) Image processing device, input power management device and sound analysis device

[0018]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of a scrap burn-through determination device according to the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram of an embodiment of the present invention.
The electric furnace is a pot-shaped furnace body 1 that holds scrap with a lid.
An upper electrode 2 which is connected to a power source and supplies electric power to scrap, an upper secondary conductor 11-a and a lower secondary conductor 11-b for flowing currents from the power source to the upper electrode 2 and the furnace bottom electrode 14, respectively. A DC reactor 12 installed between the upper electrode 2 and the thyristor transformer 13, a thyristor transformer 13 for converting an AC voltage into a DC voltage, and a furnace bottom electrode 14 installed at the bottom of the furnace body.

The heat-resistant ITV camera 6 for monitoring the inside of the furnace is attached to the furnace lid of the electric furnace, and the microphone 8 for collecting the sound generated from the arc is attached to the outer periphery of the furnace body, which is installed far away. The image processing device 7 and the acoustic analysis device 9 are connected.

The input power management device 10 for managing the power input to the electric furnace is connected to the thyristor transformer 13.

FIG. 2 shows a signal processing flow of the scrap burn-through determination device, and four combinations are possible.

In FIG. 2A, the output signals from the image processing device 7 and the input power management device 10 are input to the fuzzy inference operation device 15, and the fuzzy inference processing is performed by the software of the CPU of the built-in microcomputer. .

In FIG. 2B, output signals from the image processing device 7 and the acoustic analysis device 9 are converted into fuzzy inference operation device 15
And fuzzy inference processing is performed by the software of the CPU of the built-in microcomputer.

In FIG. 2C, the output signals from the acoustic analysis device 9 and the input power management device 10 are input to the fuzzy inference operation device 15, and fuzzy inference processing is performed by the software of the CPU of the built-in microcomputer. .

In FIG. 2 (4), the output signals from the image processing device 7, the acoustic analysis device 9 and the input power management device 10 are input to the fuzzy inference operation device 15 and the software of the CPU of the built-in microcomputer is used. Fuzzy reasoning is processed.

Next, fuzzy rules will be described.
Table 1 shows a fuzzy rule for burn-through determination used in fuzzy inference in a tabular form.

[0028]

[Table 1]

The knowledge for judging the state of scrap burn-through is, for example, "the amount of input electric power is large (B), and the ratio of one level during binarization is large (B), the degree of melting-through is large (B).
(Table 1, (1) Rule 1), and "If the input power amount is large (B) and the ratio of one level during binarization is medium (M), the degree of burn through is medium (M). "(Rule 2 of Table 1 (1)).

The same applies to (2), (3) and (4) in Table 1. In this way, what shows the knowledge for judgment about various conditions related to the state of scrap burn through,
It is a fuzzy rule table of Table 1.

Description will be given taking Table 1 (1) as an example. Table 1
In (1) of, the two columns on the left side show the condition groups related to the burn-through state of scrap (in fuzzy reasoning, the antecedent part), and the one column on the right side shows the degree of scrap burn-through judgment (fuzzy reasoning). Then, it is called the consequent part).

FIGS. 3 (1), (2), (3) and (4) represent membership functions used in fuzzy inference. The membership function represents the degree (fitness) of belonging to a certain phenomenon with a numerical value from "0" to "1". 3 (1), (A), (B), and (D) are membership functions used in fuzzy inference for determining scrap burn-through, and the vertical axis indicates the goodness of fit (grade). It is defined by the value of the section of ""-"1".

This will be described with reference to FIG. 3 (4) as an example. Figure 3
(4) (A) shows the membership function MF1, the horizontal axis represents the amount of electric power input per ton of scrap, and the unit is (kwh / ton).

FIG. 3 (4) (B) shows the membership function MF2, where the horizontal axis represents the total number of pixels in one screen when the image of the situation in the furnace is binarized. The ratio of white (1 level) is taken. This is a ITV camera with a special filter that allows light in a specific wavelength range to pass through, and when the captured image is binarized, the molten steel, arc, electrodes, and slag in the high temperature area are displayed in white (1 level). By applying the fact that the low temperature scrap and the furnace wall are displayed in black (0 level), the degree of scrap burn through is determined by the ratio of the number of white pixels in all the pixels of one screen. .

FIG. 3 (4) (C) shows the membership function MF3, and the horizontal axis shows the noise level of arc generation. When scrap remains in the furnace, the noise level is large due to scrap collapse and arc instability, and when completely scrapped, the arc is generated stably, so the noise level is lowered. However, the degree of progress of the scrap burn through is determined.

FIG. 4 shows a processing flow of fuzzy inference. According to the flow of FIG. 4, the fitness of each condition part is calculated for all fuzzy rules, and the membership function of each condition part (consequent part) of the fuzzy rule is weighted and corrected to determine the overall melt-through degree. If the logical sum function of the above-mentioned corrected membership function is taken and it is determined by the MIN-MAX centroid method obtained from the centroid position, if the result is above a certain level (for example 70%), the scrap melted down. It is determined that the scrap is burned out, and the guidance message “scrap burnout” is output on the CRT screen.

FIGS. 5 (1), (2), (3) and (4) show examples of the corrected OR function of the membership function.
A typical example will be described with reference to FIG.

In FIG. 5A, the membership function of FIG. 3D is calculated by weighting all the degrees that conform to the fuzzy rules of Table 1A, and then multiplied by the ratio. , And the degree of burn-through is obtained by the MIN-MAX centroid method.

As for the determination of the burn-through of scrap, as described above, the combination of the following four cases can be similarly performed. (1) Image processing device and input power management device (2) Image processing device and sound analysis device (3) Input power management device and sound analysis device (4) Image processing device, input power management device and sound analysis device

Table 1 (1), FIG. 3 (1), and (1) in FIG. 5 (1) show burn-through determination fuzzy rules, membership functions, and corrected membership in the case of the above combination (1). It represents the logical sum function of the functions.

In the cases of (2), (3), and (4) in Table 1, FIG. 3, and FIG. 5, as in the case of (1), the cases of (2), (3), and (4) of the above combinations are obtained. The logical sum function of the burn-through determination fuzzy rule, the membership function, and the corrected membership function is shown.

[0042]

According to the present invention, it is possible to judge scrap burn through without depending on the intuition and experience of a skilled person. As a result, it is possible to prevent scrap overflow due to erroneous scrap burn-through determination, and to prevent the temperature measurement sampling device from colliding with the unmelted scrap and damaging it, thus improving operating efficiency.

Further, it can be used as a device for determining the end of melting of the electric furnace operation, and is effective in automating the electric furnace operation.

[Brief description of drawings]

FIG. 1 is an equipment configuration diagram of a scrap burn-through determination device in an electric furnace equipment of the present invention.

FIG. 2 is a signal processing flow of the scrap burn-through determination device according to the present invention. FIG. 2 (1) shows a case where an image processing device and an input power management device are used, and FIG. 2 (2) shows an image processing device and acoustic analysis. 2 (3) shows the case of using the input power management device and the acoustic analysis device, and FIG. 2 (4) shows the case of using the image processing device, the input power management device, and the acoustic analysis device.

3 is a membership function of the present invention, and FIG.
3A is a membership function with respect to the input power amount, and FIG. 3B is a membership function with respect to the ratio of one level (white) when the image is binarized.
(C) is a membership function of noise level of arc generation, and FIG. 3 (D) is a membership function of scrap burn-through degree.

FIG. 4 is a processing flow of fuzzy inference according to the present invention.

FIG. 5 is a logical sum function of the corrected membership function of the present invention.

FIG. 6 is an example of a pattern of input power.

[Explanation of symbols]

 1 Furnace Body 2 Upper Electrode 3 Arc 4 Molten Steel 5 Scrap 6 Heat-Resistant ITV Camera 7 Image Processing Device 8 Sound Collecting Microphone 9 Acoustic Analysis Device 10 Input Power Management Device 11-a Upper Secondary Conductor 11-b Lower Secondary Conductor 12 DC Reactor 13 Thyristor transformer 14 Furnace bottom electrode 15 Fuzzy inference operation device 16 Temperature measurement sampling device

Claims (1)

[Claims] 1. A scrap burn-through determination apparatus for electric furnace equipment for melting scrap, which comprises an image processing apparatus for recognizing and determining the unmelted portion of the scrap and the molten steel / slag portion, and applying a voltage from a power source to the electrodes. Fuzzy inference is performed by inputting output signals from at least two of the input power management device that manages the power input to scrap and the acoustic analysis device for the sound generated from the arc of the electric furnace during energization. And a fuzzy inference calculation device for judging the burn-through of scrap, and a scrap burn-through judging device comprising: