JP2020016345A - Melting state determination device - Google Patents

Abstract

To provide a melting state determination device capable, in determining a melting state of a metal material in an arc furnace based on measurement results of a sound generated in the furnace and a current or a voltage on the primary side of a furnace transformer, of determining the melting state after estimating a corresponding relationship between the measurement results and the melting state of the metal material.SOLUTION: A melting state determination device 1 of this invention comprises: data acquiring means 20 that acquires determination data consisting of at least one of sound frequency data obtained by frequency analyzing intensity of a sound generated in a furnace of an arc furnace and electric frequency data obtained by frequency analyzing a current value or a voltage value on the primary side of a furnace transformer that supplies an electric source to the arc furnace; determination means 11 that determines a melting state of a metal material from determination data acquired by the data acquiring means 20 based on a machine learning result concerning correlation between the melting state of the metal material in the arc furnace and the determination data; and notifying means 12 that notifies the determination result by the determination means 11.SELECTED DRAWING: Figure 2

Description

  The present invention relates to a melting state determining device, and more particularly, to a melting state determining device for determining a melting state of a metal material in a furnace in an arc furnace.

  In an arc furnace in which a metal material such as metal scrap is melted by arc discharge, it is important to determine the melting state of the metal material. Usually, the metal material is charged and melted in the furnace a plurality of times, and when the melting of the already charged metal material has progressed to some extent, the next metal material is recharged. Then, when all of the charged metal materials are melted down, the next step such as oxidation refining is performed. At this time, it is necessary to accurately determine a point in time at which the already charged metal material is melted and the reloading becomes possible, and a point at which melting occurs and the transition to the next step becomes possible. If the point at which it is determined that additional melting or melting is possible is too slow or too early with respect to the actual progress of melting, the operation efficiency of the arc furnace may be affected.

  However, it is difficult to determine the melting state of the metal material by directly observing the inside of the arc furnace, and by observing a phenomenon that has a correlation with the melting state of the metal material from outside the arc furnace, it is possible to observe the melting state of the metal material. Is used. For example, as described in Patent Literature 1, the sound generated inside the arc furnace is used to determine the completion of melting. 2. Description of the Related Art Conventionally, an operator may hear a sound generated in a furnace and determine the completion of melting based on a change in the sound. Further, in Patent Document 1, the frequency of a sound signal of an in-furnace generated sound is analyzed, and the intensity of a signal component in a region centered on an even multiple of the fundamental frequency is reduced on each of a low frequency side and a high frequency side of the region. It is determined that the dissolution is complete when the intensity is higher than a predetermined amount for a predetermined time or more compared to the intensity of the signal component in the region. Also, in Patent Document 2, a primary current / voltage of a furnace transformer of an arc furnace is detected to detect a harmonic current component / voltage component of an even multiple of the fundamental frequency, and the harmonic current component / voltage is detected. Dissolution completion is determined when the value of the voltage component falls below a predetermined value for a certain period of time.

JP 2013-170748 A JP 2012-207904 A

  As described in Patent Literature 1 and Patent Literature 2, melting of the metal material in the arc furnace is completed based on the result of the frequency analysis of the sound generated in the furnace and the intensity of the harmonic component of the primary current / voltage. If the determination is made, it is possible to timely determine whether or not it is possible to perform additional loading and determine whether or not the metal material has melted off, based on the high correlation between the measurement results and the melting state of the metal material. Then, a predetermined threshold value is set for each measurement result, and when the state of being equal to or higher than the threshold value is maintained for a predetermined time, the completion of melting is determined. Unlike the case of listening to the generated sound to judge the completion of melting, etc., the determination of the completion of melting is made quantitatively without depending on the experience and skill of the worker, and the efficiency of the operation of the arc furnace is promoted. be able to. However, the accuracy of the determination can be improved even when the completion of melting is determined by comparing the frequency analysis result of the in-furnace sound and the evaluation data such as the intensity of the harmonic components of the primary current / voltage with a threshold value. Greatly depends on whether the set threshold value is appropriate.

  The correlation between the melting state of the metal material in the arc furnace and the evaluation data depends on many factors such as the configuration of the individual arc furnace, the type and amount of the metal material charged, and the evaluation data and the metal material. Depending on the correlation between the dissolution states, it may not be possible to determine whether or not the dissolution has been completed according to the actual state at a predetermined threshold value. In addition, the melting state determined to be complete based on how much the melting of the metal material has progressed, and whether or not to determine whether or not the metal material can be additionally mounted or melted down, also depends on many factors as described above. I do. As described above, it is difficult to accurately determine the melting state unless an appropriate threshold is used according to the state of the arc furnace or the metal material, or the desired melting state. Appropriately estimate the correspondence between the evaluation data and the melting state of the metal material in the furnace according to the state of the individual arc furnace and metal material, and the desired melting state, regardless of the predetermined threshold. If it is possible, the state of the metal material in the furnace can be more accurately estimated, and it can be effectively used for the operation of the arc furnace, including the determination of the completion of melting.

  The problem to be solved by the present invention is to determine the melting state of the metal material in the arc furnace based on the sound generated in the furnace and the measurement result of the current or voltage on the primary side of the furnace transformer. An object of the present invention is to provide a melting state determination device capable of performing a determination after estimating a correspondence relationship between a result and a melting state of a metal material.

  In order to solve the above problems, a melting state determination apparatus according to the present invention includes sound frequency data obtained by frequency-analyzing the intensity of an in-furnace sound of an arc furnace, and a primary transformer for a furnace for supplying power to the arc furnace. Data acquisition means for acquiring evaluation data composed of at least one of electric frequency data obtained by frequency analysis of a current value or a voltage value on the side, and machine learning relating to a correlation between the melting state of the metal material in the arc furnace and the evaluation data. Based on the result of the above, from the evaluation data obtained by the data obtaining means, a determining means for determining the dissolution state of the metal material, and a notifying means for notifying the result of the determination by the determining means, is there.

  Here, in the machine learning, from the waveform of the evaluation data, a feature amount having a correlation with the dissolution state of the metal material is calculated, and the determination unit is configured based on the feature amount calculated by the machine learning. The state of dissolution of the metal material may be determined.

  Alternatively, in the machine learning, for a parameter included in the evaluation data, a threshold value indicating the completion of melting of the metal material is determined, and the determination unit determines the threshold value of the parameter in the evaluation data acquired by the data acquisition unit. It is preferable to determine whether or not the melting of the metal material is completed by comparing a value with the threshold value determined by the machine learning.

  In this case, the evaluation data includes the sound frequency data, and in the machine learning, the threshold is determined for the sound frequency data with respect to the height of a peak including a frequency that is an even multiple of the power supply frequency. Then, the determining means, in the sound frequency data obtained by the data obtaining means, the state where the height of the peak is equal to or more than the threshold, if the state has continued for a predetermined time or more, the melting of the metal material. It is good to judge that it was completed.

  In addition, the evaluation data includes the electric frequency data, and in the machine learning, for the electric frequency data, for the intensity of a harmonic component of a frequency that is an even multiple of the power supply frequency, determines the threshold value, In the electric frequency data acquired by the data acquisition unit, the determination unit determines that the intensity of the harmonic component is equal to or less than the threshold, and when the state has continued for a predetermined time or more, the melting of the metal material is completed. Should be determined.

  In the melting state determination device according to the present invention, the sound frequency data obtained by frequency-analyzing the intensity of the sound generated in the furnace of the arc furnace, and the electric frequency data obtained by frequency-analyzing the current value or the voltage value of the primary side of the furnace transformer. The dissolution state of the metal material is determined based on at least one of the evaluation data. These evaluation data have a high correlation with the melting state of the metal material in the arc furnace, and can be suitably used for judging the melting state as reflecting the melting state sensitively.

  Further, by learning the correlation between the evaluation data and the dissolution state of the metal material by machine learning, the correspondence relationship between the acquired evaluation data and the dissolution state of the metal material can be adjusted according to the actual situation. It can be used for estimation and determination of the dissolved state. As a result, it becomes possible to determine the dissolution state of the metal material with higher accuracy than when interpreting the evaluation data based on a preset threshold or the like for the evaluation data, such as a fixed threshold. Further, the correspondence between the evaluation data and the melting state of the metal material and the melting state desired to be determined may vary depending on the individual operation mode such as the configuration and operating conditions of the arc furnace, the type and amount of the metal material, and the like. In such a case, the correspondence between the evaluation data and the dissolution state of the metal material can be appropriately estimated according to the individual operation mode or the like. For example, a threshold value for determining whether or not the dissolution has been completed can be appropriately set.

  Here, in the machine learning, a characteristic amount having a correlation with the melting state of the metal material is calculated from the waveform of the evaluation data, and the determination unit determines the melting state of the metal material based on the characteristic amount calculated by the machine learning. In the case of the determination, a characteristic amount having a correlation with the melting state of the metal material is obtained by machine learning in the waveform of the evaluation data, and can be used for determining the melting state. As a result, waveforms can be obtained without having to set parameters that allow humans to clearly recognize and set thresholds in the evaluation data, such as the height of a specific peak or the signal intensity at a specific frequency. It is possible to accurately determine the dissolution state of the metal material based on its own characteristics. In addition, since it is possible to find a characteristic amount that well reflects the melting state of the metal material according to the individual operation mode, it is possible to find a feature value that is more suitable for the individual operation mode than when a parameter to be focused on is determined in advance. Thus, the dissolution state of the metal material can be accurately determined.

  Alternatively, in the machine learning, for a parameter included in the evaluation data, a threshold value indicating the completion of melting of the metal material is determined, and the determining unit determines the parameter value in the evaluation data acquired by the data acquiring unit and the value determined by the machine learning. When it is determined whether or not the melting of the metal material is completed by comparing with the threshold value, a correlation with the melting state of the metal material, such as a specific peak height and a signal intensity at a specific frequency, is required. By paying attention to parameters for which is known, the information included in the evaluation data can be clearly and simply associated with the dissolution state of the metal material and used for determination of the dissolution state. In addition, a threshold value is provided for the parameter, and the determination of the dissolution state is performed by an alternative method of determining whether or not the dissolution is completed, so that the determination by the determination unit and the notification by the notification unit can be simply performed. . At this time, instead of determining a specific value of the threshold value in advance, by using machine learning to determine the value, a threshold value that enables an accurate determination of melting completion in accordance with an individual operation mode. Can be obtained.

  In this case, the evaluation data includes sound frequency data, and in machine learning, a threshold is determined for the sound frequency data with respect to the peak height including a frequency that is an even multiple of the power supply frequency. According to the configuration in which it is determined that the melting of the metal material has been completed when the state in which the peak height is equal to or higher than the threshold in the sound frequency data acquired by the data acquisition means continues for a predetermined time or more. In the frequency data, the height of the peak including a frequency that is an even multiple of the power supply frequency has a high correlation with the melting state of the metal material, and as the melting proceeds, the height of the peak increases. Therefore, the threshold value at the peak height is determined by machine learning and used for determining the completion of dissolution, thereby making it possible to accurately determine the completion of dissolution. In particular, since the change in the peak height appears at a relatively late stage in the time near the completion of dissolution, it is suitable for judging the completion of dissolution at a time when the dissolution has progressed particularly well.

  Further, the evaluation data includes electric frequency data, and in machine learning, a threshold value is determined for the electric frequency data with respect to the intensity of a harmonic component of an even multiple of the power supply frequency. According to the configuration in which the melting of the metal material is determined to be completed when the state in which the intensity of the harmonic component is equal to or less than the threshold continues for a predetermined time or more in the electric frequency data acquired by the primary power supply, The intensity of the harmonic component on the side has a high correlation with the dissolution state of the metal material, and the intensity decreases as the dissolution proceeds. Therefore, by determining the threshold value of the intensity of the harmonic component by machine learning and using the threshold value for determining the completion of melting, it is possible to accurately determine the completion of melting. In particular, since the change in the intensity of the harmonic component appears relatively early in the time near the completion of dissolution, it is suitable for judging the completion of dissolution with a margin when the dissolution has progressed to some extent. Further, by combining the determination based on the intensity change of the harmonic component in the electric frequency data and the determination based on the peak height in the sound frequency data, the operating conditions of the arc furnace and the progress of melting to be determined are determined. The progress of the dissolution of the metal material can be accurately determined depending on the degree of the melting.

It is a mimetic diagram showing the composition of the arc furnace provided with the melting state judging device concerning one embodiment of the present invention. It is a block diagram showing the composition of the above-mentioned melting state judging device. It is an example of sound frequency data, (a) shows data at the beginning of dissolution, and (b) shows data at the end of dissolution. It is an example of electric frequency data, and shows the intensity change of the harmonic component of each order. It is a mimetic diagram showing the composition of the arc furnace provided with the melting state judging device concerning a modification.

  Hereinafter, a dissolution state determination device according to an embodiment of the present invention will be described with reference to the drawings. The melting state determining apparatus 1 according to one embodiment of the present invention is used together with the arc furnace 5 to determine the melting state of the metal material M in the arc furnace 5. FIG. 1 shows an example of an arc furnace 5 provided with a melting state determination device 1. FIG. 2 shows a configuration of the dissolution state determination device 1.

[Configuration of arc furnace]
First, the configuration of the arc furnace 5 including the melting state determination device 1 will be briefly described. In the arc furnace 5, the metal material M such as metal scrap is stored in the furnace body 50, and the metal material M is melted by performing arc discharge between the electrode 52 attached to the lid 51 and the metal material M. I do. The electrode 52 is held by a lifting device (not shown), and the height position thereof can be changed.

  A power supply circuit for supplying electric power to the arc furnace 5 is provided with a furnace transformer 53 having a tap changer, and a primary circuit 54 of the furnace transformer 53 is connected to a commercial power supply. In the primary circuit 54, a vacuum circuit breaker (VCB) 56 is provided between the furnace transformer 53 and a commercial power supply. The secondary circuit 55 of the furnace transformer 53 reaches the electrode 52 and supplies power to the electrode 52. In the illustrated embodiment, a three-phase alternating current is supplied to the electrode 52. Although not shown, the secondary circuit 55 is provided with a current transformer for the instrument and a transformer for the instrument, and is based on the current value and the voltage value of the secondary circuit 55 measured through them. By the feedback control, the height position of the electrode 52 is adjusted, and the output of the furnace transformer 53 is adjusted.

  Further, the current transformer 22 for the instrument is provided in the primary circuit 54. The instrument current transformer 22 constitutes the data acquisition means 20 of the melting state determination device 1, and inputs an electric signal corresponding to a current value flowing through the primary circuit 54 to the arithmetic device 10. The signal input from the current transformer for instrument 22 is subjected to calculation by the electric frequency analysis means 23 included in the calculation device 10.

  The sound level meter 21 is installed near the furnace body 50. The sound level meter 21 also constitutes the data acquisition means 20 of the melting state determination device 1, detects a sound generated inside the arc furnace 5, and sends an electric signal corresponding to the detected sound intensity to the arithmetic device 10. input. The signal input from the sound level meter 21 is subjected to calculation by a sound frequency analysis unit 24 included in the calculation device 10. Since the sound level meter 21 is installed near the furnace body 50, it is housed in a casing (not shown) as a heat-resistant measure and a dust-proof measure.

  The vacuum circuit breaker 56 provided in the primary circuit 54 of the furnace transformer 53 is ON / OFF controlled by a control signal from the arithmetic unit 10. While power is supplied to the electrode 52 and the metal material M is melted, the vacuum circuit breaker 56 is turned on. On the other hand, as described later, when it is determined in the arithmetic unit 10 that the melting of the metal material M is completed, the vacuum circuit breaker 56 is switched to the OFF state, and the power supply to the electrode 52 via the furnace transformer 53 is performed. Supply is shut off. Alternatively, when it is determined that the melting of the metal material M is completed, the determination result may be simply notified to the operator by the notification panel 60 described below. In this case, the operator performs ON / OFF control of the vacuum circuit breaker 56, and the operator may use the notified determination result as an index of the switching timing.

[Configuration of dissolution state determination device]
Next, the configuration of the melting state determination device 1 will be described. The melting state determination device 1 includes a computing device 10 in addition to the instrument current transformer 22 of the primary circuit 54 and the sound level meter 21 installed near the furnace body 50 described above.

  The arithmetic unit 10 is composed of a computer and is installed around the arc furnace 5. The arithmetic device 10 has, as arithmetic functions, a determination unit 11, a notification unit 12, a storage unit 13, and a learning unit 14, in addition to the electric frequency analysis unit 23 and the sound frequency analysis unit 24. The arithmetic unit 10 may also serve as a control device for controlling the output of the furnace transformer 53 and the height position of the electrode 52 based on the current value and the voltage value of the secondary circuit 55 in the arc furnace 5. .

  The sound level meter 21, the sound frequency analyzing means 24, the current transformer 22 for the instrument, and the electric frequency analyzing means 23 constitute a data acquiring means 20 in the melting state determination device 1. As described above, the signal corresponding to the intensity of the in-furnace sound detected by the sound level meter 21 is subjected to frequency analysis by the sound frequency analysis means 24 by Fourier transform, Laplace transform, or the like. Then, sound frequency data indicating an intensity distribution for each frequency is input to the determination unit 11 as evaluation data. Further, the signal corresponding to the current value of the primary circuit 54 input from the current transformer 22 is subjected to frequency analysis by the electric frequency analysis means 23. Then, electrical frequency data consisting of data of the intensity distribution itself for each frequency or data obtained by extracting a change in the intensity of a harmonic component having a frequency that is an even multiple of the fundamental frequency (power frequency of a commercial power supply) from the intensity distribution, The evaluation data is input to the determination unit 11.

  The judging means 11 converts the evaluation data inputted from the data acquiring means 20, that is, the sound frequency data inputted from the sound frequency analyzing means 24, and the electric frequency data inputted from the electric frequency analyzing means 23 into the electric arc furnace 5. Of the metal material M, that is, how much or in what form the metal material M is dissolved is determined. At the time of the determination, the evaluation data actually input from the data acquisition means 20 is interpreted based on a model relating to the correlation between the evaluation data and the dissolution state of the metal material M given by the learning means 14, and the evaluation data is obtained. Estimate the corresponding dissolution state. For example, by comparing the value of a parameter included in the evaluation data, such as the height of a peak of a specific frequency in the sound frequency data, with a threshold value given by the learning unit 14, the value of the metal material M can be compared. Determine whether dissolution is complete.

  The notifying unit 12 notifies the result of the determination regarding the melting state of the metal material M by the determining unit 11 to the outside of the melting state determination device 1. In the illustrated embodiment, the notifying unit 12 outputs the result of the determination to, for example, a notifying panel 60 provided as a display screen of the arithmetic unit 10, and notifies the worker of the melting state of the metal material M. For example, the fact that melting has been completed and the metal material M can be reloaded, or that all the metal material M has melted down is notified by an image or text. The determination result may be notified remotely using a communication facility such as the Internet. Further, when the judging means 11 judges that the melting is completed, the notifying means 12 outputs a control signal as a notice to the vacuum circuit breaker 56 provided in the primary circuit 54 of the arc furnace 5, and turns ON. The vacuum circuit breaker 56 in the state is turned off, and the power supply to the electrode 52 is stopped.

  The storage unit 13 stores data including a correspondence relationship between the waveform of the evaluation data, that is, the waveforms of the sound frequency data and the electric frequency data, and the melting state of the metal material M in the arc furnace 5. In other words, the waveform of the evaluation data is stored in combination with the melting state of the metal material M when the waveform is observed. In addition to the waveform of the evaluation data and the melting state of the metal material M, information on the operation mode when the evaluation data is obtained may be stored in association with the waveform and the melting state of the evaluation data.

  In this specification, the “operation mode” encompasses various conditions related to the operation when melting a certain metal material M using a certain arc furnace 5. As specific items of various conditions, conditions relating to the operation of the arc furnace 5 (the configuration of the arc furnace 5, the current value and the voltage value in the secondary side circuit 55, the scheduled electric energy, the height position of the electrode 52, and the like). , The type and amount (bulk) of the metal material M, the conditions for charging the metal material M (whether at the time of initial loading or at the time of reloading, etc.), the desired degree of dissolution progression (complete dissolution Whether it must be completed or whether some unmelted part may remain), the time required to move to the next process such as reloading, the surrounding environment of the arc furnace 5 (external noise level, etc.) Can be mentioned.

  Examples of the data stored in the storage unit 13 include data obtained in the past in the arc furnace 5 provided with the melting state determination device 1 or data obtained in another arc furnace in addition thereto. . While the operation of the arc furnace 5 is repeated from a state in which a certain amount of initial data is stored in the storage means 13, the evaluation data newly acquired by the data acquisition means 20 and the judgment means 11 using the evaluation data. The correspondence relationship with the dissolution state determined in (1) may be further added to the initial data and stored in the storage unit 13.

  The learning unit 14 performs machine learning on the correlation between the evaluation data and the melting state of the metal material M in the arc furnace 5 based on the data stored in the storage unit 13. That is, in the arc furnace 5 to be operated, what kind of evaluation data is observed and what kind of melting state the metal material M in the furnace can be determined to be in is analyzed. I do. For example, the value of a parameter included in the evaluation data, which can be extracted from the evaluation data, such as the height of a specific peak or the signal intensity of a specific frequency in the evaluation data, is compared with the value in the actually measured evaluation data. A threshold for determining whether or not the melting of the metal material M is completed is determined as a specific numerical value. The details of the machine learning will be described later. The learning unit 14 inputs the result of the machine learning to the determination unit 11. The determination means 11 interprets the evaluation data actually input from the data acquisition means 20 using the result of the machine learning as described above.

  In the embodiment described here, the arithmetic unit 10 including the electric frequency analysis unit 23, the sound frequency analysis unit 24, the determination unit 11, and the notification unit 12 further includes a storage unit 13 and a learning unit 14 as a part of the function. However, the storage unit 13 and the learning unit 14 may be configured to be provided in an external device such as a computer independent of the arithmetic device 10. And an external device may be installed in a place remote from the arc furnace 5 such as outside the factory. In particular, when the newly acquired evaluation data is not added to the initial data and stored in the storage unit 13, the operation mode is not frequently changed, and the machine learning result input to the determination unit 11 is frequently updated. When there is no need to update, when the load of the machine learning calculation in the learning means 14 is large, the storage means 13 and the learning means 14 may be provided in an external device.

[Metal material melting method]
An example of a method of melting a metal material M such as a metal scrap using the arc furnace 5 and the melting state determination device 1 having the above-described configurations will be briefly described. Before the melting is started, a model of the correlation to be applied to the operation with respect to the evaluation data and the melting state, for example, a threshold value to be applied to a certain parameter is set in the learning means 14 in accordance with the planned operation mode. It is determined by machine learning and input to the determination means 11.

  When the melting is started, first, the operator first mounts the metal material M on the arc furnace 5. Then, power is supplied to the electrode 52 to promote dissolution of the metal material M. During this time, the dissolution state determination device 1 acquires the evaluation data by the data acquisition unit 20. Then, the obtained evaluation data is input to the judging means 11, and the evaluation data is interpreted according to the correlation inputted in advance from the learning means 14 to judge the melting state of the metal material M in the arc furnace 5. The acquisition of the evaluation data by the data acquisition unit 20 and the determination of the dissolution state by the determination unit 11 are repeatedly performed throughout the dissolution process. In the meantime, if it is determined by the determining means 11 that the initially loaded metal material M is sufficiently melted and the reloading is possible, a signal indicating that fact is output from the notifying means 12. Specifically, the operator is notified by the notification panel that reloading is possible, the vacuum circuit breaker 56 of the arc furnace 5 is turned off, and power supply to the electrode 52 is interrupted.

  Next, the worker performs reloading of the metal material M. After that, the melting of the metal material M, the acquisition of the evaluation data by the data acquiring unit 20, and the judgment of the melting state by the judging unit 11 are performed in the same manner as in the initial mounting. The reloading and melting can be repeated a plurality of times (for example, three times). The determining unit 11 determines that the melting state of the metal material M is in a state in which the reloading is possible, and the power supply to the electrode 52 is performed. After the suspension, the next reloading may be performed.

  When the metal material M added last is melted, when the judging means 11 judges that the melt-down has occurred and the molten state has reached a melting state capable of proceeding to the next step such as oxidizing refining, the signal from the notifying means 12 is sent. , The notification panel notifies that the process can be shifted to the next step, the vacuum circuit breaker 56 of the arc furnace 5 is turned off, and the power supply to the electrode 52 is terminated.

  The model relating to the correlation between the evaluation data and the melting state, which is input from the learning unit 14 to the determination unit 11 and used for the determination by the determination unit 11, is a case where it is determined that the reloading is possible and a case where it is determined that the burnout has occurred. And may be different. Further, when it is determined that the reloading is possible, it may be different between the initial reloading and the plurality of reloading times. For example, when the determination of whether the reloading is possible or the burnout is performed in comparison with a threshold value of a certain parameter, the degree of progress of the required melting is different between the determination of the reloading possibility and the determination of the burnout. By reflecting the threshold value, different threshold values can be applied to the determination of whether the reloading is possible and the determination of the burnout.

[Machine learning in learning means]
As described above, in the melting state determination device 1 according to the present embodiment, in the determination of the melting state of the metal material M by the determination unit 11, a learning unit is used as a basis for interpreting the evaluation data actually acquired. A model relating to the correlation between the evaluation data and the melting state of the metal material M, which is obtained by machine learning in 14, is used. If a model relating to the correlation between the evaluation data used as a basis for determining the melting state of the metal material M and the melting state of the metal material M is set without performing machine learning, the relationship between the two in the actual arc furnace 5 is assumed. It can be difficult to set up a model that accurately reflects gender. Further, when the form of influence of the melting state of the metal material M on the sound generated in the furnace or the behavior of the current value of the primary circuit 54 changes due to factors such as a change in the operation mode, the melting state based on the evaluation data. May not be able to be determined accurately. The mode of operation can vary from one arc furnace 5 to another, and from one operation to a single arc furnace 5.

  On the other hand, as described above, when the correlation between the evaluation data and the dissolution state of the metal material M is analyzed using machine learning, the evaluation data is determined based on the results actually observed between the two. After estimating the correspondence between the and the dissolved state, the determination of the dissolved state can be performed using the model of the correspondence as a basis. Therefore, it is possible to accurately grasp the correlation between the evaluation data and the dissolution state of the metal material M and determine the dissolution state based on the correlation according to the actual situation. In addition, even if the correspondence between the melting state and the evaluation data changes depending on the operation mode, by deriving a model of the correspondence corresponding to the operation mode by machine learning, the evaluation data and the dissolution state in the operation mode can be obtained. It is possible to accurately determine the dissolution state in accordance with the actual correspondence between them.

  A specific machine learning method is one that can execute learning relating to the correlation between the evaluation data and the metal material M from the data including the correspondence between the evaluation data stored in the storage unit 13 and the melting state of the metal material M. However, there is no particular limitation. Although any of supervised learning, semi-supervised learning, and unsupervised learning may be used, supervised learning or semi-supervised learning can be suitably used. The algorithm used for machine learning is not particularly limited, and linear regression, regression type such as support vector machine, clustering type such as k-nearest neighbor method, tree type such as decision tree, deep learning and self-organizing map, etc. Neural network type algorithms can be exemplified. In an example described later, it is preferable to use deep learning in a form in which a parameter to be noted is not specified.

  Hereinafter, analysis of the correlation between the evaluation data and the dissolution state of the metal material M by machine learning will be described. The correlation analysis can be broadly classified into a case where a parameter to be focused on is specified and a case where it is not specified.

(1) When designating parameters to be focused on Sound frequency data and electric frequency data constituting evaluation data include various parameters (amount that can be extracted as numerical values from waveforms). The dissolution state of M can be reflected in those parameters. The parameters included in the evaluation data include signal strength, peak height, peak width, noise level, and the like. Further, a difference or a ratio of a value in a different area on data, or a ratio of a value between different parameters can be cited.

  Here, a description will be given of an embodiment in which machine learning is performed by focusing on what parameters, analyzing the correlation between the evaluation data and the dissolution state of the metal material M, and specifying whether to create a model. That is, after the worker or the designer of the melting state determination device 1 specifies the parameter, the learning unit 14 determines how the melting state of the metal material M changes when the value of the specified parameter changes. It is analyzed by machine learning whether it can be associated with such a change. For example, for a parameter whose value increases or decreases as the dissolution progresses, the threshold value at which the dissolution proceeds to a desired degree and the dissolution can be determined to be completed is set to what specific value by machine learning. decide. In this way, by associating the evaluation data with the dissolution state of the metal material M by focusing on the known parameters, unknown parameters are calculated as feature amounts by machine learning as described later. The interpretation of the evaluation data can be performed clearly and simply as compared with the form.

  Then, for the known parameters, the determination of the melting state of the metal material M is performed by comparing the threshold value calculated by the learning means 14 with the actually measured value as an alternative determination of whether or not the melting is completed. Thus, the determination by the determination unit 11 and the notification by the notification unit 12 can be simply performed. However, the threshold value is set only in one stage, and by comparing with the threshold value, the melting state of the metal material M is not limited to the alternative method of determining whether or not the melting is completed. It may be set in multiple stages, and the degree of progress of dissolution may be determined in multiple stages. Alternatively, without setting a specific threshold value, the value of the parameter may be continuously associated with the melting state of the metal material M, such as the degree of progress of the melting or the non-uniformity of the melting.

  The parameter of interest can be set based on past results and the like. Then, for the parameter of interest selected by the worker or the designer, a specific threshold may be determined by machine learning in the learning unit 14 based on the data stored in the storage unit 13. Hereinafter, a specific example will be described in which a parameter to be focused on is known, and the learning unit 14 performs machine learning by designating the parameter to be focused on.

(1-1) Form Focusing on the Height of Harmonic Peaks in Sound Frequency Data It is known that as the melting of the metal material M proceeds in the arc furnace 5, the sound generated inside the furnace changes. I have. While the melting is not progressing much, intermittent sound such as thunder is generated in response to discrete arc discharge of the non-uniformly deposited metal material M. It is known that as dissolution progresses, the sound changes to a continuous sound. Such a change in sound is often reflected in sound frequency data obtained by frequency-analyzing the sound generated in the furnace.

  As shown in FIG. 3 (a), in the initial stage of the dissolution, in a region including an even multiple of the fundamental frequency (for example, 4 times; 200 Hz), no clear frequency dependence is seen in the signal intensity, and a wide frequency region is used. Sound is observed with similar intensity. On the other hand, as shown in FIG. 3B, at the end of dissolution, while the signal intensity is increased in a part of the frequency region centered on the even multiple of the fundamental frequency, the signal of the surrounding frequency is increased. The intensity becomes weaker, and a clear peak structure centered on a frequency that is an even multiple of the fundamental frequency is observed.

  Therefore, the height of a peak (harmonic peak) including a frequency that is an even multiple of the fundamental frequency may be employed as a parameter of interest and used by the determination unit 11 to determine the melting state. For example, a threshold (T in FIG. 3) is provided for the height of the harmonic peak, and when the state where the height of the actually observed harmonic peak is equal to or higher than the threshold continues for a predetermined time or more, the metal material It can be determined that the dissolution of M has been completed. In the evaluation of the height of the harmonic peak, the height of the harmonic peak is read from the sound frequency data, and when the state in which the height is equal to or higher than the threshold T has continued for a predetermined time or more, it may be determined that the melting is completed. . The height of the harmonic peak is, for example, the intensity of the peak of the harmonic peak and the signal intensity in the low frequency side and high frequency side regions (frequency regions corresponding to the region Q and the region R in FIG. 3) of the harmonic peak. Can be read as the difference from the average value. For example, in FIG. 3A in the early stage of dissolution, the height of the harmonic peak is 10 dB or less, and in FIG. 3B in the last stage of dissolution, it is about 40 dB.

  Alternatively, instead of reading the height of the harmonic peak and comparing it with the threshold, it may be determined whether or not the signal component of each frequency is included in a plurality of regions set using the threshold. For example, in FIG. 3, three regions (regions P, Q, and R) are set based on two thresholds T1 and T2 (T1> T2). A region Q centered on an even multiple of the fundamental frequency in a region P defined by a threshold T1, at least a part of a signal component enters the region P, and a region Q defined by a threshold T2 on a lower frequency side than the region P; It may be determined that the melting is completed when the state where all the signal components of the frequency enter the region R defined by the threshold value T1 on the high frequency side for a certain period of time or longer. In FIG. 3A at the initial stage of dissolution, no signal is contained in the region P, and a part of the signal is not contained in the region Q. On the other hand, in FIG. 3B at the end of dissolution, a part of the signal is in the region P and is also in the regions Q and R.

  Here, the threshold used for the determination, that is, the threshold T of the height of the harmonic peak, or the thresholds T1 and T2 defining the regions P, Q, and R are determined by machine learning in the learning means 14. At this time, the storage means 13 associates the sound frequency data with the melting state of the metal material M when the sound frequency data was observed, that is, information on whether melting has been completed or not completed. A large number of data are stored. The data is assumed to have been obtained in the arc furnace 5 being operated and / or another arc furnace having the same configuration. Further, data acquired by an arc furnace having another configuration may be added.

  The learning unit 14 performs machine learning based on the data stored in the storage unit 13 and calculates a threshold value related to the height of the harmonic peak to be used for determining whether or not the melting is completed in the arc furnace 5. . That is, what threshold value is set is determined by comparing the threshold value with the actually measured value to accurately determine whether or not the dissolution is completed. Further, the length of the predetermined time for judging the completion of melting with the state where the height of the harmonic peak is equal to or higher than the threshold value or the state where the signal intensities satisfy the predetermined relationship with the regions P, Q, R is maintained. Similarly, it can be determined by machine learning. For example, it is only necessary to determine how long the comparison with the threshold value should be performed so that the influence of accidental noise can be sufficiently suppressed to make an accurate determination.

  Further, the data stored in the storage means 13 in association with the sound frequency data is stored not only in the dissolution state of the metal material M when the sound frequency data is observed, but also when the sound frequency data is observed. If the information on the mode is included, when the threshold value is determined by the learning means 14, it is possible to accurately determine the melting state of each individual melting process scheduled to be operated in consideration of the individual operation mode. It is possible to set an appropriate threshold. The operation mode items included in the data stored in the storage means 13 include, as described above, the conditions related to the operation of the arc furnace 5 (the configuration of the arc furnace 5, the current value and the voltage value in the secondary circuit 55, Expected power amount, height position of the electrode 52, etc.), type and amount (bulk) of the metal material M, conditions for charging the metal material M (whether at the time of initial mounting or additional mounting, etc.), desired The degree of progress of melting (whether the melting must be completely completed or whether some undissolved parts may remain), the time required to move to the next process such as reloading, arc The surrounding environment (external noise level and the like) of the furnace 5 can be exemplified. For example, when the amount of the metal material M is large or when it is required that the melting is completely completed, the threshold value of the height of the harmonic peak is set high, and the external noise level is high. In this case, the threshold and the predetermined time can be set by considering factors such as setting a predetermined time for performing the comparison with the threshold to be long in machine learning.

  In the case where the items of the operation mode to be considered in determining the threshold value are various, if an operator or a designer or the like determines in advance the threshold value corresponding to the individual operation mode composed of a combination of those items, If given, it is necessary for a human to analyze how each item affects the threshold, and to set a threshold appropriately reflecting the contents of each item. Is often difficult. For example, assuming that, among the operation modes, the threshold must be determined in consideration of the three items of the scheduled electric energy of the arc furnace 5 and the type and amount of the metal material M, each of the three items is set. The threshold value must be set appropriately in consideration of how it affects the magnitude of the threshold value to be performed and how the effects of the three items interact. As the number of items that define the operation mode increases and the change of the mode in each item becomes diversified, the number of thresholds that must be prepared in advance also increases according to the number of combinations of items. However, by determining the threshold value in consideration of each item by using machine learning in the learning unit 14, the threshold value reflecting the influence of each item is appropriately adjusted for each individual operation mode including a combination of a plurality of items. Can be determined. It is easy to cope with an increase in the number of items and various changes. Furthermore, even if data obtained in an operation to which the same operation mode as the scheduled operation is applied does not exist in the storage unit 13, a threshold suitable for the scheduled operation mode is set to a machine threshold. It can be calculated by learning and used for the determination of the dissolution state by the determination means 11. Further, in the case where the operation mode is unintentionally changed during the melting step, machine learning in the learning means 13 is performed again by using the evaluation data already acquired in the melting step, and the threshold value is set. Can be changed on the way.

(1-2) Form Focusing on the Intensity of Harmonic Components in Electric Frequency Data As the melting of the metal material M progresses in the arc furnace 5, the harmonics generated from the arc furnace 5 may decrease. Are known. The harmonics generated from the arc furnace 5 appear in the current and voltage of the primary circuit 54 of the power supply circuit for supplying power to the electrodes 52 as harmonic components having an even multiple of the fundamental frequency. Accordingly, by observing the time change of the intensity of the harmonic corresponding to an even multiple (for example, twice) of the fundamental frequency in the current and the voltage of the primary circuit 54, the melting state of the metal material M in the arc furnace 5 is observed. Can be determined.

  FIG. 4 shows a temporal change in the intensity of the harmonic component of each order in the current of the primary circuit 54 measured by the current transformer 22 for the instrument. The figure shows measurement results for a plurality of even-order harmonic components having different orders (how many times the frequency is the fundamental frequency). According to FIG. 4, the intensity of the harmonic component does not significantly change until the end of melting in the melting process at the time of initial loading and at the time of two additional loadings. However, at the end of dissolution, the intensity of the harmonic component gradually decreases.

  Therefore, the intensity of the harmonic component may be adopted as the parameter of interest and used for the determination of the melting state by the determination unit 11. For example, a threshold is set for the intensity of a harmonic component of a certain order, and the intensity of the actually observed harmonic component of the order is equal to or less than the set threshold, when the state continues for a predetermined time or longer, the metal material It can be determined that the dissolution of M has been completed. In the example of FIG. 4, when the threshold value for the second harmonic is set to 0.15 kV in the melting at the time of initial loading, the intensity of the second harmonic component exceeds the threshold value at the beginning of the melting. , The intensity of the second harmonic gradually decreases to fall below the threshold value. After that, the intensity of the second harmonic continues to decrease, but it is determined that the melting at the time of initial loading has been completed and that additional loading is possible when the intensity has fallen below the threshold for a predetermined time or more. Can be.

  Here, the threshold of the intensity of the harmonic component used for the determination is determined by machine learning by the learning means 14. At this time, the storage unit 13 associates the electric frequency data with the melting state of the metal material M when the electric frequency data was observed, that is, information on whether the melting is completed or not completed. A large number of data are stored.

  The learning unit 14 performs machine learning based on the data stored in the storage unit 13 and calculates a threshold value related to the intensity of the harmonic component, which is to be used for determining whether or not the melting in the arc furnace 5 is completed. That is, as a threshold value used to determine whether or not the dissolution is completed by comparison with the actually measured value, a threshold value that can accurately perform the determination is determined. Similarly, the length of the predetermined time for judging completion of dissolution when the intensity of the harmonic component is equal to or less than the threshold value can be determined by machine learning.

  Further, similarly to the case of the sound frequency data, the information to be stored in the storage means 13 in association with the electric frequency data includes not only the dissolved state of the metal material M when the electric frequency data is observed but also the information. If the information on the operation mode when the electric frequency data is observed is included, when the learning means 14 determines the threshold value, the individual operation mode is determined for each individual melting process scheduled for operation. In consideration of the above, it is possible to set a threshold suitable for accurate determination of the dissolution state. In addition, as for the form of storage in the storage unit 13 and the form of machine learning in the learning unit 14, the same form as described above in the case of using sound frequency data is preferably used in the case of using electric frequency data. Can be applied.

  In the embodiment described here, the current value of the primary circuit 54 obtained by the current transformer 22 for the instrument is subjected to frequency analysis, and the result is used as electric frequency data. The harmonic component of the voltage of the primary circuit 54, that is, the voltage component of an even multiple of the fundamental frequency, is used to determine the dissolution state of the material M, as well as the harmonic component of the current. It has a high correlation with the dissolution state, and exhibits a behavior in which the strength decreases as the dissolution proceeds. Therefore, an instrument transformer is installed in the primary circuit 54 instead of the instrument current transformer 22, the voltage value of the primary circuit 54 is measured and the frequency is analyzed, and the result is used as electric frequency data. Can also.

(1-3) A form in which a plurality of parameters are set as candidates In the two specific examples described above, attention is paid to the height of the harmonic peak in the sound frequency data and the intensity of the harmonic component in the electric frequency data. The melting state of the metal material M is determined using a threshold value set by machine learning in the learning unit 14 for each value.

  However, both the height of the harmonic peak in the sound frequency data and the intensity of the harmonic component in the electric frequency data can be used simultaneously to determine the dissolution of the metal material M. At this time, the determination of which one to use for the determination and the determination of the form of each contribution to the determination result when both are used for the determination can be made by machine learning in the learning means 14.

  In this case, the storage unit 13 stores both information that associates sound frequency data with the melting state of the metal material M and information that associates electrical frequency data with the melting state of the metal material M. Preferably, information in which both the sound frequency data and the electric frequency data acquired in the same melting step are associated with the melting state of the common metal material M may be stored. Also in this embodiment, information including not only the dissolution state of the metal material M but also the operation mode when each sound frequency data and electric frequency data are observed may be stored.

  The learning unit 14 performs machine learning based on the data stored in the storage unit 13 and determines the types of parameters to be used for determining whether melting in the arc furnace 5 is completed. In this case, it is determined which of the height of the harmonic peak of the sound frequency data and the intensity of the harmonic component of the electric frequency data is used for the determination, or whether both are used. Further, when it is determined that a plurality of parameters are to be used, in the determination of the dissolution state, it is determined in what kind of contribution form the determination based on each parameter is used together. After determining the parameters to be adopted, the learning means 14 further determines a threshold value to be applied to each parameter and a predetermined time for comparing the threshold value.

  In making these determinations, which of the candidate parameters can be used to accurately determine the melting state of the metal material M, that is, which parameter is sensitive to the change in the target melting state, and It is only necessary to consider whether to reflect accurately. For example, a decrease in the intensity of the harmonic component in the electric frequency data reflects the end of dissolution of the metal material M relatively early, while the growth of the harmonic peak in the sound frequency data indicates that the metal material M Observed after dissolution has progressed relatively well. Therefore, when it is desired to notify relatively early that the melting is about to end, for example, when the bulk of the metal material M to be mounted next is small, the lowering of the harmonic component in the electric frequency data is determined. It should be used for. On the other hand, in the case where the bulk of the metal material M to be mounted next is large, or when it is desired to notify until the melting is completely completed or immediately before the melting is advanced, the harmonic peak in the sound frequency data is Growth may be used for the determination.

  In this way, the parameters to be adopted can be selected according to the planned operation mode. For a plurality of parameters, when the determination based on the threshold value is used in combination, as a contribution form of the determination for each parameter, for example, when the measured values of all the adopted parameters exceed the threshold value, a predetermined melting state (For example, completion of dissolution) may be determined, or when a measured value of at least one parameter exceeds a threshold, it may be determined that a predetermined dissolution state has been reached. . Alternatively, weighting can be performed among a plurality of parameters. As a specific example of the weighting, for a certain parameter A, a state satisfying the threshold may be maintained for a predetermined time Ta, but for another parameter B, a state satisfying the threshold may be longer than a predetermined time Ta. There may be a form in which the duration is different, for example, it is necessary to last for Tb. In addition, it is necessary to always exceed a threshold value for a certain parameter A, but for another parameter B and parameter C, only one of them needs to exceed the threshold value. Is possible.

  In the specific example shown here, only the height of the high-frequency peak of the sound frequency data and the intensity of the high-frequency component of the electric frequency data are handled as candidate parameters. However, other parameters appearing in the sound frequency data and the electric frequency data may be adopted as candidates for the parameters used for determining the melting state of the metal material M. Such waveform features include the intensity and width of various peaks, noise, signal intensity at a given frequency, or the difference or ratio of measured values at different points on the data (eg, 2 Ratios of intensity at two different frequencies), ratios of values between different parameters (eg, ratio of peak intensity to noise level), and the like. In addition to the parameters appearing in the sound frequency data and the electric frequency data, other parameters having a correlation with the dissolution state of the metal material M, such as the amount of input power, may be used for the determination.

(2) When a parameter to be focused is not specified In the above, in the evaluation data including the sound frequency data and / or the electrical frequency data, what parameter is focused on and the determination of the melting state of the metal material M is performed. After designating by a designer, a designer, or the like, the learning unit 14 causes machine learning to make a specific correspondence between the parameter and the melting state of the metal material M, such as determination of a threshold value. However, it is also possible to cause the learning means 14 to perform machine learning without designating a parameter to be focused on by an operator or a designer, and to extract a correlation between the evaluation data and the melting state of the metal material M. In other words, in the evaluation data, the viewpoint itself, which is to be focused on and what should be used to determine the correspondence between the metal material M and the melted state and determine the melted state, is obtained by machine learning in the learning means 14. Can be done.

  Also in this case, the storage unit 13 stores data in which the evaluation data is associated with the melting state of the metal material M when the evaluation data is observed. It is preferable that the stored data further includes information on the operation mode when each evaluation data is observed.

  The learning unit 14 performs machine learning on the correlation between the evaluation data and the dissolution state of the metal material M based on the data stored in the storage unit 13. At this time, the learning unit 14 calculates a characteristic amount having a correlation with the melting state of the metal material M in the waveform of the evaluation data. For example, a feature common to the waveform of the evaluation data when the melting of the metal material M is completed and / or a feature common to the waveform of the evaluation data when the melting of the metal material M is not completed are found and extracted. I do.

  For the calculation of the feature amount, for example, deep learning can be used. In the deep neural network, evaluation data labeled in a dissolved state of the metal material M is input to an input layer. Data input to the input layer is transmitted to an intermediate layer including a plurality of hidden layers. Then, the feature amount in the waveform of the evaluation data is calculated by the intermediate layer, and the feature amount is output from the output layer.

  The feature amount calculated by the learning unit 14 as a result of the machine learning is input to the determination unit 11. Then, the judging means 11 interprets the evaluation data actually obtained by the data obtaining means 20 based on the characteristic amount, and judges the melting state of the metal material M in the arc furnace 5. For example, the learning unit 14 may apply a feature common to the waveform of the evaluation data when the melting of the metal material M is completed and / or a feature common to the waveform of the evaluation data when the melting of the metal material M is not completed. When the calculation is performed as a feature amount, the determination unit 11 interprets the waveform of the evaluation data acquired by the data acquisition unit 20 based on the feature amount output by the learning unit 14 and calculates the waveform of the evaluation data. Is determined to be associated with a state in which the melting of the metal material M is completed or in a state in which the melting of the metal material M is not completed.

  In this way, instead of specifying the parameter to be focused on, the feature amount having a correlation with the dissolution state of the metal material M is calculated from the waveform of the evaluation data by machine learning, so that the dissolution state of the metal material M is calculated. Features that reflect well can be found. Then, the association between the evaluation data and the dissolution state of the metal material M is performed with higher accuracy than when the parameter to be focused on is specified as described above, and the dissolution state can be accurately determined. In particular, even when the dissolution state of the metal material M appears in the waveform of the evaluation data as a characteristic that cannot be grasped as a clear parameter, such as a specific peak height or a specific frequency signal intensity, Such features can be extracted and used as a basis for interpretation of the evaluation data in the determination means 11. That is, similarities and differences between waveforms that cannot be expressed as clear parameters are extracted from the evaluation data stored in the storage unit 13, and are compared with the evaluation data actually acquired by the data acquisition unit 20. By comparing the waveforms themselves as image patterns between them, similarities and differences between the waveforms can be compared to judge the dissolution state. In some cases, a feature that a human cannot recognize from the evaluation data can be found by machine learning. In such a case, it is possible to determine the dissolution state with particularly high accuracy.

  Further, the information stored in the storage means 13 includes, in addition to the evaluation data and the dissolution state of the metal material M when the evaluation data is observed, information on the operation mode when the evaluation data is observed. In this case, in addition to the dissolution state, it is possible to discover by machine learning by the learning means 14 how many items relating to the operation mode and the evaluation data have a correlation. . Although it is difficult for humans to analyze how each of the effects of various items appears as characteristics on the waveform of the evaluation data, and how those effects interact on the waveform, By storing a large number of data in the storage unit 13 and causing the learning unit 14 to execute machine learning, it is possible to associate the effect of each item with the characteristic on the waveform and to determine the interaction in the waveform of the effect. The analysis can be performed quantitatively. As a result, the correlation between the waveform of the evaluation data and the melting state depends on the individual operation mode, even if the number of items that define the operation mode increases and the mode of each item changes in various ways. , A melting state of the metal material M can be determined with high accuracy.

[Modification]
In the embodiment described so far, as shown in FIG. 1, a three-phase alternating current is supplied from a primary side circuit 54 via a furnace transformer 53 to an electrode 52 provided on a lid 51 of the arc furnace 5. I have. However, as shown in a modified form in FIG. 5, electrodes 52 and 52 'may be provided on the lid 51 and the bottom of the furnace body 50, respectively, and a direct current may flow between the electrodes 52 and 52'. In this case, the AC output from the furnace transformer 53 is converted into DC by a power converter 53 ′ including a rectifier circuit and the like, and one of the negative side and the positive side of the DC output is covered with a cover 51. What is necessary is just to connect the other electrode 52 to the electrode 52 'of the furnace bottom. Also in this case, the frequency analysis of the current or voltage on the primary side of the furnace transformer 53 can be used as electric frequency data.

  In FIGS. 1 and 5, the current transformer 22 for the instrument (or the transformer for the instrument) is installed in the primary circuit 54, and the signal corresponding to the current value (or the voltage value) of the primary circuit 54 is converted into an electric signal. It is input to the frequency analysis means 23. The electric frequency analysis means 23 analyzes the frequency of the signal and comprises data of the intensity distribution itself for each frequency or data obtained by extracting a change in the intensity of a harmonic component having an even multiple of the fundamental frequency from the intensity distribution. The electric frequency data is input to the determination means 11 as evaluation data. However, when the determining means 11 exclusively uses the information on the intensity change of the harmonic component having an even multiple of the fundamental frequency as the electric frequency data, the current transformer 22 for the instrument (or the transformer for the instrument) and The electric frequency analysis means 23 may be omitted, and a harmonic meter for measuring the intensity of the harmonic component of the primary current (or voltage) may be provided in the primary circuit 54 instead. In this case, an electric signal corresponding to the intensity of the harmonic component output by the harmonic meter may be directly input to the determination unit 11 and used as evaluation data.

  As described above, the embodiments of the present invention have been described in detail. However, the present invention is not limited to the above embodiments, and various modifications can be made without departing from the gist of the present invention.

DESCRIPTION OF SYMBOLS 1 Melting state determination apparatus 10 Computing apparatus 20 Data acquisition means 21 Sound level meter 22 Current transformer for instrument (Primary side current transformer)
Reference Signs List 5 arc furnace 50 furnace body 51 lid 52 electrode 53 furnace transformer 54 primary circuit 55 secondary circuit 56 vacuum circuit breaker M metal material

Claims (5)

From sound frequency data obtained by frequency-analyzing the intensity of the sound generated inside the arc furnace, and electric frequency data obtained by frequency-analyzing a current value or a voltage value on a primary side of a furnace transformer for supplying power to the arc furnace. Data acquisition means for acquiring evaluation data
Judgment for judging the melting state of the metal material from the evaluation data acquired by the data acquiring means based on a result of machine learning on a correlation between the melting state of the metal material in the arc furnace and the evaluation data. Means,
A dissolving state determining apparatus, comprising: a notifying unit configured to notify a result of the determination by the determining unit. In the machine learning, from the waveform of the evaluation data, to calculate a feature amount having a correlation with the melting state of the metal material,
2. The apparatus according to claim 1, wherein the determination unit determines a melting state of the metal material based on the feature amount calculated by the machine learning. 3. In the machine learning, for the parameters included in the evaluation data, determine a threshold indicating the completion of melting of the metal material,
The determination unit determines whether or not the melting of the metal material is completed by comparing the value of the parameter in the evaluation data acquired by the data acquisition unit with the threshold value determined by the machine learning. The dissolution state determination device according to claim 1, wherein: The evaluation data includes the sound frequency data,
In the machine learning, for the sound frequency data, for the height of a peak including a frequency of an even multiple of the power supply frequency, determine the threshold,
In the sound frequency data acquired by the data acquisition unit, the determination unit is configured to complete the melting of the metal material when the state where the height of the peak is equal to or greater than the threshold continues for a predetermined time or more. The dissolution state determination device according to claim 3, wherein the determination is made. The evaluation data includes the electric frequency data,
In the machine learning, for the electric frequency data, for the intensity of the harmonic component of a frequency of an even multiple of the power supply frequency, determine the threshold,
In the electric frequency data acquired by the data acquisition unit, the determination unit is configured to complete the melting of the metal material when a state in which the intensity of the harmonic component is equal to or less than the threshold is maintained for a predetermined time or more. The dissolution state determination device according to claim 3, wherein the determination is made.

Images