Methods for predicting steel leakage, operating procedures for continuous casting machines, and devices for predicting steel leakage.

Abstract

The method for predicting steel leakage includes the following steps: inputting the dimensions of a casting sheet drawn from a mold in a continuous casting machine; detecting the temperature of the mold using multiple thermometers embedded in the mold; performing interpolation processing on the detected temperatures from the multiple thermometers based on the dimensions of the casting sheet; calculating the deviation from the normal operation without steel leakage based on the temperature calculated by the interpolation processing, using components in directions orthogonal to the influence coefficient vector obtained from principal component analysis; and predicting steel leakage based on the deviation.

Description

Technical Field

[0001] This invention relates to a method for predicting steel leakage, an operation method for a continuous casting machine, and a device for predicting steel leakage. Background Technology

[0002] Traditionally, the known continuous casting process involves pouring molten steel into a mold, cooling the molten steel using a mold with embedded water-cooling pipes to solidify its surface, pulling a semi-solidified sheet from the bottom of the mold using drawing rollers, and finally producing a fully solidified sheet through jet cooling. In continuous casting processes, there is an increasing demand for high-speed casting to improve productivity. However, increasing the casting speed can lead to a reduction in the thickness of the solidified shell at the bottom of the mold, or an uneven distribution of the solidified shell thickness. As a result, sometimes the solidified shell breaks as it exits the mold at thinner sections, causing a so-called breakout. When a breakout occurs, it results in a long downtime, significantly deteriorating productivity. Therefore, a breakout prediction method is desired that can accurately predict the occurrence of breakouts while performing high-speed casting.

[0003] As a method for predicting steel leakage, the following method is known: For countermeasures against steel leakage caused by the solidification shell being restricted by the mold, the steel leakage is predicted by detecting the change in temperature measured by a temperature measuring instrument such as a thermocouple embedded in a copper plate, thus preventing the solidification shell from being restricted by the mold.

[0004] For example, Patent Document 1 discloses a method for monitoring restricted steel leakage as follows: multiple temperature measuring instruments are arranged horizontally below the liquid surface of the mold of a continuous casting machine to form a temperature measuring column. Multiple layers of this temperature measuring column are arranged in the casting direction. Any two temperature measuring instruments in the upper layer and the lower layer of the multiple layers are arranged on the same vertical line. Their measured values ​​are transmitted to a computing device. If conditions 1 and 2 are met simultaneously, it is determined that restricted steel leakage has occurred.

[0005] Condition 1: In the upper and / or lower temperature measurement columns, the measured values ​​of adjacent temperature measuring devices rise and then fall.

[0006] Condition 2: The temperature readings of the lower temperature measuring device located on the vertical line are higher than those of the upper temperature measuring device.

[0007] Furthermore, Patent Document 2 discloses the following method for predicting steel leakage, including: a step of detecting the temperature of the mold using multiple thermometers whose sensitivity coefficients are calculated by embedding the mold in a continuous casting machine; a step of defining a detection temperature vector as a vector composed of the sensitivity coefficients of each of the multiple thermometers; a step of calculating the deviation degree as the component of the detection temperature vector in a direction orthogonal to the sensitivity coefficient vector; a step of assigning a first score to thermometers whose deviation degree exceeds a threshold; a step of defining different thermometer score vectors as the scores of different thermometers, using the first score as the score of different thermometers; a step of assigning a second score to the thermometer that becomes the center when scores are assigned to each thermometer and the thermometer adjacent to each thermometer in the different thermometer score vectors; and a step of detecting the occurrence of signs of steel leakage using the second score.

[0008] Existing technical documents

[0009] Patent documents

[0010] Patent Document 1: Japanese Patent Application Publication No. 2017-154155

[0011] Patent Document 2: Japanese Patent No. 5673100 Summary of the Invention

[0012] The problem that the invention aims to solve

[0013] However, in the restrictive method for monitoring leaked steel disclosed in Patent Document 1, the method is configured to calculate the temperature change based on time-series data of the detected temperature. Therefore, it is possible that even if the detected temperature changes due to factors other than signs of leaked steel, such as changes in casting speed, it could be falsely detected as a possible leak.

[0014] Furthermore, in the steel leakage prediction method disclosed in Patent Document 2, the temperature measurement value itself is defined as the detection temperature vector to calculate the deviation. Therefore, during unsteady operations such as changing the width of the casting sheet, the deviation increases due to changes in the casting width of the molten steel relative to the mold, which may lead to false detection of potential steel leakage.

[0015] The present invention was made in view of the above-mentioned problems, and its object is to provide a method for predicting steel leakage with high accuracy, an operation method for a continuous casting machine, and a device for predicting steel leakage.

[0016] Methods for solving problems

[0017] To address the aforementioned issues and achieve the objective, the method for predicting steel leakage of the present invention is characterized by comprising: a step of inputting the dimensions of a casting sheet drawn from a mold in a continuous casting machine; a step of detecting the temperature of the mold using a plurality of thermometers embedded in the mold; a step of performing interpolation processing on the detected temperatures of the plurality of thermometers based on the dimensions of the casting sheet; a step of calculating, based on the temperature calculated by performing the interpolation processing, a deviation degree relative to normal operation when steel leakage does not occur, using components in directions orthogonal to the influence coefficient vector obtained by principal component analysis; and a step of predicting steel leakage based on the deviation degree.

[0018] Furthermore, the method for predicting steel leakage of the present invention is characterized in that, in the above invention, in the step of performing the interpolation process, for the detected temperature of each of the plurality of thermometers, the temperature is calculated by performing interpolation process at the center point of each of the plurality of calculation units that are equally divided according to the size of the casting sheet.

[0019] Furthermore, the method for predicting steel leakage of the present invention is characterized in that, even if the size of the casting is changed, the number of calculation units remains constant.

[0020] Furthermore, the method for predicting steel leakage of the present invention is characterized in that, in the above invention, in the step of calculating the deviation, the average temperature of each of the plurality of calculation units located at the same distance from the upper end of the mold in the casting direction of the molten steel relative to the mold is calculated, the difference between the temperature of each of the plurality of calculation units and the average value is calculated, and the deviation is calculated using the influence coefficient vector based on the calculated difference.

[0021] Furthermore, the method for predicting steel leakage of the present invention is characterized in that, in the step of predicting steel leakage, when the time change rate of the deviation exceeds a preset first threshold, the steel leakage is predicted based on the adjacency of the calculation unit when the absolute value of the deviation exceeds a preset second threshold.

[0022] Furthermore, the method for predicting steel leakage of the present invention is characterized in that, in the above invention, the step of predicting the steel leakage includes: assigning a first score to the calculation unit whose deviation exceeds the second threshold; calculating a second score based on the adjacency of the calculation unit assigned the first score; and predicting the steel leakage based on the second score.

[0023] Furthermore, the method for predicting steel leakage of the present invention is characterized in that, in the above invention, the influence coefficient vector is a sensitivity coefficient vector with the sensitivity coefficient of each of the plurality of thermometers as a component.

[0024] Furthermore, the operating method of the continuous casting machine of the present invention is characterized in that, when a steel leakage is predicted based on the steel leakage prediction method of the above invention, the casting speed of injecting molten steel into the mold is reduced.

[0025] Furthermore, the leakage prediction device of the present invention is characterized by comprising: an input mechanism for inputting the dimensions of a casting sheet drawn from a mold in a continuous casting machine; a plurality of thermometers embedded in the mold and detecting the temperature of the mold; an interpolation processing execution mechanism for performing interpolation processing on the detected temperatures of the plurality of thermometers based on the dimensions of the casting sheet; a deviation calculation mechanism for calculating the deviation degree relative to normal operation when no leakage occurs, based on the temperature calculated by performing the interpolation processing and taking the components in the direction orthogonal to the influence coefficient vector obtained by principal component analysis as the deviation degree; and a leakage prediction mechanism for predicting leakage based on the deviation degree.

[0026] The effects of the invention

[0027] The steel leakage prediction method, continuous casting machine operation method, and steel leakage prediction device of the present invention can predict steel leakage with high accuracy. Attached Figure Description

[0028] Figure 1 This is a schematic diagram showing the general structure of a continuous casting machine according to an embodiment.

[0029] Figure 2 This is a perspective view showing the schematic structure of a mold with an embedded thermometer in a continuous casting machine according to an embodiment.

[0030] Figure 3 (a) is a diagram illustrating the condition of the molten steel and solidified shell inside the mold in the early signs of a steel leak. Figure 3 (b) is a diagram showing the condition of the fractured part of the solidified shell in the early signs of a leaking steel.

[0031] Figure 4 (a) is the temperature distribution of the mold at the instant sintering occurs. Figure 4 (b) is a graph showing the temperature distribution of the mold 10 seconds after the moment of sintering.

[0032] Figure 5 This is a flowchart illustrating an example of the steps in the method for predicting steel leakage according to an embodiment.

[0033] Figure 6 This is a graph showing the correlation between the thermometer's detection temperature and the normal temperature under conditions where no steel leakage has occurred.

[0034] Figure 7 This is a graph showing the correlation between the thermometer's detection temperature and the occurrence of signs such as sintering that can lead to steel leakage.

[0035] Figure 8 (a) is a graph showing the relationship between the temperature detected by the thermometer and the temperature obtained by interpolation in a case where the width of the casting sheet drawn from the lower end of the mold is large. Figure 8 (b) is a graph showing the relationship between the temperature detected by the thermometer and the temperature obtained by interpolation in a case where the width of the casting sheet drawn from the lower end of the mold is small.

[0036] Figure 9 It is a diagram showing the positional relationship between the thermometer and the calculation unit, which are located at the same distance from the top of the mold.

[0037] Figure 10 (a) is a graph showing the temporal variation of the absolute value of the deviation in cases where sintering occurs. Figure 10 (b) is a graph showing the temporal variation of the rate of change of deviation in cases where sintering occurs.

[0038] Figure 11 (a) is a graph showing the time-series changes in the absolute value of the deviation in cases where sintering did not occur. Figure 11 (b) is a graph showing the time-series change of the deviation rate in cases where sintering did not occur.

[0039] Figure 12 This is a diagram illustrating an example of a method for determining adjacency when the computational unit performing the interpolation process is a single-layer structure.

[0040] Figure 13 This diagram illustrates a method for determining whether the calculation units are configured in two layers, an upper layer and a lower layer, in the casting direction. The method assumes that the calculation units in the upper layer have scores for three adjacent units, and the calculation unit in the lower layer has a score for one of the three adjacent units in the upper layer.

[0041] Figure 14 This is a chart of time-series detection data of steel leakage cases predicted using the steel leakage prediction method according to an embodiment of the present invention. Detailed Implementation

[0042] The following describes the embodiments of the steel leakage prediction method, the operation method of the continuous casting machine, and the steel leakage prediction device of the present invention. It should be noted that the present invention is not limited to these embodiments.

[0043] Figure 1 This is a schematic diagram showing the general structure of the continuous casting machine 1 according to the embodiment. Figure 1As shown, the continuous casting machine 1 of this embodiment includes a tundish 3 into which molten steel 2 is injected, a copper mold 5 that cools the molten steel 2 injected from the tundish 3 via an immersion nozzle 4, multiple casting support rollers 7 that transport semi-solid casting sheets 6 drawn from the mold 5, and a determination unit 20 that determines signs of steel leakage based on the detected temperature of a thermometer 8 embedded in the mold 5. It should be noted that in this embodiment, a thermocouple is used as the thermometer 8, but the method is not limited to this.

[0044] Figure 2 A thermometer 8 is embedded in the continuous casting machine 1 shown in the embodiment. 1，1 ~8 m，n A three-dimensional diagram of the schematic structure of mold 5. (See figure) Figure 2 As shown, the mold 5 has a pair of long-side cooling plates 5a and a pair of short-side cooling plates 5b, which are formed into a generally square tube shape that runs through the vertical direction. Inside the long-side cooling plates 5a and the short-side cooling plates 5b, cooling water channels (not shown) are formed along the inner wall surface, and the molten steel 2 is cooled by circulating cooling water in these cooling water channels.

[0045] In addition, a thermometer 8 is embedded inside the long side cooling plate 5a of the mold 5 at a predetermined depth from the outer wall surface of the long side cooling plate 5a. 1，1 ~8 m，n It should be noted that in the following description, unless otherwise specified, the thermometer 8... 1，1 ~8 m，n In such cases, it is only recorded as thermometer 8. Figure 2 In the middle, the thermometer 8 1，1 ~8 m，n The structure is designed with three or more layers in the casting direction A, with the first layer consisting of a thermometer 8. 1，1 ~8 1，n 8 thermometers on the second layer 2，1 ~8 2，n Thermometer 8 on the nth floor m，1 ~8 m，n They are respectively embedded on the same plane. In this embodiment, the casting direction A refers to the direction in which molten steel 2 is injected from the tundish 3 through the immersion nozzle 4 relative to the mold 5, and is the same direction as the direction in which the casting sheet 6 is pulled out from the lower end of the mold 5.

[0046] It should be noted that, Figure 2The illustrated configuration of the thermometer 8 is merely one example to illustrate the present invention. The thermometer 8 can be arranged in at least one of the pair of long-side cooling plates 5a, at least one of the pair of short-side cooling plates 5b, or all of the pair of long-side cooling plates 5a and the pair of short-side cooling plates 5b of the mold 5. Preferably, the thermometer is arranged in all of the pair of long-side cooling plates 5a and the pair of short-side cooling plates 5b. Furthermore, the thermometer 8 can also be arranged in the mold 5 in multiple layers (more than three layers) or in a single layer along the casting direction A.

[0047] Next, the warning signs of steel leakage will be explained. Figure 3 (a) is a diagram illustrating the condition of the molten steel 2 and solidified shell 10 in the mold 5, which are early signs of steel leakage. Figure 3 (b) is a diagram showing the condition of the crack 11 in the solidified shell 10, which is a precursor to steel leakage.

[0048] like Figure 3 (a) and Figure 3 As shown in (b), in the precursory phenomenon of steel leakage, sintering occurs within the mold 5 for some reason, and the solidified shell 10 is confined by the mold 5. On the other hand, due to the presence of... Figure 3 (b) The casting sheet 6 is pulled from the lower end of the mold 5 in the same direction as casting direction A, so a crack 11 is formed in the solidified shell 10 directly below the sintering point. At the crack 11 of the solidified shell 10, the mold 5 comes into contact with the molten steel 2, and further sintering occurs. While repeating the above phenomenon, the crack 11 of the solidified shell 10 moves downward, and the solidified shell 10 above the crack 11 thickens. Then, when the crack 11 finally passes through the lower end of the mold 5, the molten steel 2 leaks out from the crack 11, resulting in a steel leak.

[0049] It should be noted that because the molten steel 2 comes into contact with the mold 5 at the fracture point 11, the temperature of the mold 5 locally rises. Therefore, for example, if using... Figure 3 Arrow B in (b) indicates the ground, with the rupture 11 moving downwards from the thermometer 8. m'，1 ~8 m'，n When the configuration position is passed, thermometer 8 m'，1 ~8 m'，n The detection temperature becomes high. Subsequently, because the solidified shell 10 above the fracture 11 is confined by the mold 5 and continues to cool, the thermometer 8... m'，1 ~8 m'，n The detection temperature decreases monotonically. On the other hand, since the fracture 11 propagates not only downwards but also laterally, so... Figure 3 As shown in (b), the rupture 11 expands in a V-shape. It should be noted that in thermometer 8... m'，1 ~8 m'，nIn the case where a solidified shell 10 cracks 11 at its lower part, due to the temperature gauge 8 m'，1 ~8 m'，n The location where the rupture 11 will not pass through will only be observed at thermometer 8. m'，1 ~8 m'，n The detection temperature decreased.

[0050] Figure 4 (a) is the temperature distribution of mold 5 at the instant of sintering. Figure 4 (b) is a graph showing the temperature distribution of mold 5 10 seconds after the moment of sintering. Based on... Figure 4 (a) and Figure 4 As shown in (b), the temperature distribution of mold 5 indicates that the high-temperature V-shaped part spreads downwards and laterally.

[0051] The temperature distribution variation in the mold 5 described above may also occur due to a decrease in the casting speed, changes in the liquid level, and changes in the width of the casting sheet 6. When the casting speed decreases or the liquid level changes, the mold temperature at a position equidistant from the upper end of the mold 5 changes synchronously. On the other hand, when the casting width of the molten steel 2 poured into the mold 5 is changed during operation—in other words, when the width of the casting sheet 6 pulled from the lower end of the mold 5 is changed—the temperature variation in the mold measured by the thermometers 8 located near both ends of the width of the casting sheet 6 increases.

[0052] Therefore, in the method for predicting steel leakage in the embodiment, by calculating the evaluation value of the non-linkage of the estimated temperature at multiple locations where interpolation processing was performed based on the width of the casting slab 6, and determining the rate of change of this evaluation value and the adjacency of temperature changes at the locations where the change occurred, the accuracy of steel leakage prediction is improved. The method for predicting steel leakage based on the above-described technical concept will be described in detail below.

[0053] Figure 5 This is a flowchart illustrating an example of the steps in a method for predicting steel leakage according to an embodiment. The steel leakage prediction method shown in this flowchart consists of... Figure 1 The determination unit 20 shown performs the operation. It should be noted that the determination unit 20 has at least the functions of the interpolation processing execution mechanism, the deviation calculation mechanism, and the leakage prediction mechanism described in this invention. Further details will be provided later. Figure 5 Details of each step in the process.

[0054] In the method for predicting steel leakage in the embodiment, the determination unit 20 calculates in advance the temperature 8 during normal operation (hereinafter also referred to as normal operation) when no steel leakage occurs. 1，1 ~8 m，nThe sensitivity coefficient is calculated (step S1). Here, in order to address issues such as casting width variations and thermometer malfunctions, as described later, the sensitivity coefficient is calculated using a temperature obtained by interpolation based on a normal temperature measured by the thermometer. It should be noted that since the surface condition of the mold 5 may change during operation, the sensitivity coefficient may change; therefore, it is preferable to update it at an appropriate time, such as during casting. Next, the determination unit 20 uses thermometer 8... 1，1 ~8 m，n Continuously monitor the temperature T of mold 5. 1，1 ~T m，n (Step S2). Next, the determination unit 20 checks the thermometer 8. 1，1 ~8 m，n The detected temperature is calculated by dividing the dimensions (e.g., width and thickness of the casting sheet 6) pulled from the mold 5 into equal parts by an input device (not shown) that is used by the operator as an input mechanism, such as a personal computer installed on the continuous casting machine 1. 1，1 ~12 k，p At the center point, perform temperature interpolation processing on mold 5 (step S3). Then, for the temperature T' of mold 5 obtained through interpolation processing... 1，1 ~T' k，p A mean bias is removed. That is, for the temperature T' of mold 5 obtained through interpolation... 1，1 ~T' k，p In the middle, calculation units 12 are located at positions equidistant from the upper end of the mold 5. 1，1 ~12 1，p Temperature T' 1，1 ~T' 1，p and calculation unit 12 2，1 ~12 2，p Temperature T' 2，1 ~T' 2，p T' k，1 ~T' k，p Calculate the average value for each. Then, calculate the average value for calculation unit 12. 1，1 ~12 1，p Temperature T' 1，1 ~T' 1，p Difference from the average and calculation unit 12 2，1 ~12 2，p Temperature T' 2，1 ~T' 2，p The difference from the average value (step S4). Next, the determination unit 20 calculates the deviation using a sensitivity coefficient based on the calculated difference from the average value (step S5).

[0055] Here, the sensitivity coefficient vector represents the temperature of thermometer 8 under normal operation.1，1 ~8 m，n The diagram shows the direction of the average temperature movement of the calculation unit obtained through the above interpolation process. The sensitivity coefficient vector is a vector with the sensitivity coefficient as a component, which is an influence coefficient. Furthermore, the component of the vector with the difference from the average value as a component, which is parallel to the direction of the sensitivity coefficient vector, is the component of the average movement, and the component in the direction orthogonal to the direction of the sensitivity coefficient vector is the component of the deviation from the average movement.

[0056] Next, if the calculated rate of change of deviation over time exceeds a threshold Y, the determination unit 20 determines whether a steel leakage is predicted based on the adjacent conditions of calculation units 12 whose absolute value of deviation exceeds a threshold X (step S6). It should be noted that the rate of change of deviation over time represents the proportion (degree) of change in the absolute value of deviation over a predetermined time period (per unit time). If it is determined that a steel leakage is not predicted (no in step S6), the determination unit 20 proceeds to step S2. On the other hand, if it is determined that a steel leakage is predicted (yes in step S6), the determination unit 20 automatically reduces the casting speed to a predetermined speed (step S7). Thus, if the determination unit 20 predicts a steel leakage, by sufficiently reducing the casting speed, a solidified shell 10 of sufficient thickness is formed within the mold 5 even at the location where sintering occurs, thus preventing steel leakage. Afterwards, the determination unit 20 returns to the processing routine after reducing the casting speed to the predetermined value.

[0057] Next, regarding the sensitivity coefficient used in the steel leakage prediction method in the implementation method, regarding the initial use of thermometer 8 1，1 ~8 m，n The situation regarding the detected temperature will be explained. Figure 6 This is a thermometer indicating the normal state when no steel leakage has occurred. 1，1 ~8 m，n A graph showing the correlation between the detection temperatures. Figure 7 This is a thermometer 8 indicating when signs such as sintering, which can lead to steel leakage, occur. 1，1 ~8 m，n A graph showing the correlation between the detected temperatures. It should be noted that, for simplicity... Figure 6 and Figure 7 This indicates two thermometers 8 located at the same distance from the top of the mold 5 in the casting direction A. i，j1 and thermometer 8 i，j2 The situation.

[0058] like Figure 6 As shown, the thermometer 8 is in normal working order. i，j1 and thermometer 8 i，j2 The detection temperature distribution is close to the dashed line (in Figure 6The example shown is a line diagonally 45 degrees to the right, and the dashed line indicates the direction of the sensitivity coefficient vector, which is a vector with the sensitivity coefficient as a component. Then, if a thermometer 8... i，j1 The detection temperature T i，j1 If the temperature rises, use a thermometer. i，j2 The detection temperature T i，j2 It also rises. On the other hand, if you use a thermometer... i，j1 The detection temperature T i，j1 If the temperature drops, use a thermometer. i，j2 The detection temperature T i，j2 It also declined.

[0059] As mentioned above, the thermometer 8 under normal conditions i，j1 and thermometer 8 i，j2 The reasons for this relevance are as follows. For example, the faster the casting speed of the continuous casting machine 1, the thinner the solidified shell 10 becomes due to the drawing of the casting sheet 6 during the period when the solidified shell 10 has not fully grown. As a result, the thermal resistance decreases, and the temperature of the molten steel 2 is more easily transferred to the thermometer 8. i，j1 and thermometer 8 i，j2 On the other hand, the slower the casting speed, the more the solidified shell 10 grows and is pulled out, resulting in a thicker solidified shell 10, greater thermal resistance, and difficulty in transferring the temperature of the molten steel 2 to the thermometer 8. i，j1 and thermometer 8 i，j2 These tendencies are present in all 8 thermometers. 1，1 ~8 m，n The CCP was established in Tongdi, therefore the normal temperature is 8. 1，1 ~8 m，n The detected temperatures are distributed in an almost elliptical shape within a range close to the dashed line, which indicates the direction of the sensitivity coefficient vector. However, due to the ease of temperature transfer of molten steel 2 according to thermometer 8... 1，1 ~8 m，n Different, so thermometer 8 1，1 ~8 m，n The sensitivity coefficient is generally not constant. Therefore, Figure 6 The slope of the sensitivity coefficient vector shown can be determined based on thermometer 8. 1，1 ~8 m，n Changes such as the setting position of mold 5 and construction deviations.

[0060] In addition, the thermometer is normally 8 i，j1 and thermometer 8 i，j2 In addition to the reasons mentioned above, the relevance can also be considered in terms of the flow and level changes of the molten steel 2 within the mold 5. However, the overall temperature change of the mold 5, which accompanies the increase or decrease in the casting speed, will affect the thermometer 8. 1，1 ~8 m，nThe sensitivity coefficient is largely contributed by this factor. Therefore, in order to incorporate more diverse phenomena of continuous casting processes into the sensitivity coefficient, the overall temperature change of mold 5 accompanying the increase or decrease of casting speed needs to be removed as an average bias.

[0061] As a method for removing the average bias, the following methods can be listed: Calculate using a thermometer 8 1，1 ~8 m，n The detected temperature T 1，1 ~T m，n The average of all T ave Take the detection temperature T 1，1 ~T m，n Each of them is related to the average value T. ave The difference. Other methods for removing the average bias include, for example, using a thermometer 8 located at the same distance from the top of the mold 5 in the casting direction A. i，1 ~8 i，n Perform: Determine the position of thermometer 8 located at the same distance. i，1 ~8 i，n The detected temperature T i，1 ~T i，n average value T i，ave Take the detection temperature T i，1 ~T i，n Each of them is related to the average value T. i，ave The difference.

[0062] Alternatively, principal component analysis can be used as a method to determine the sensitivity coefficient vector, where the sensitivity coefficient vector is the influence coefficient vector. Another method could be to experimentally determine the temperature changes of each thermometer as the overall temperature changes due to variations in liquid level, etc. 1，1 ~8 m，n The ease of temperature transfer in molten steel 2.

[0063] On the other hand, such as Figure 7 As shown, the thermometer 8 indicates the presence of signs such as sintering that may lead to steel leakage. i，j1 and 8 i，j2 The detection temperature distribution is along the dashed line (showing the direction of the sensitivity coefficient vector) Figure 7 The example shown shows the separation point of the line at a 45-degree right angle. This is because, in the event of sintering that could lead to steel leakage, the thermometer 8 is located near the crack 11 of the solidified shell 10. i，j1 Detection temperature T i，j1 It dropped, and later, it was located at thermometer 8. i，j1 Thermometers on both sides 8 i，j1+1 and thermometer 8 i，j1-1 Detection temperature Ti，j1+1 and detection temperature T i，j1-1 decline.

[0064] Based on the above investigation, it can be concluded that: It is possible to determine the thermometer's 8... 1，1 ~8 m，n Detection temperature T 1，1 ~T m，n The degree of deviation of the dashed line indicating the direction of the sensitivity coefficient vector determines the occurrence of steel leakage. In other words, it can be determined that: [the following text appears to be unrelated and possibly from a different source: "can determine the occurrence of steel leakage by measuring 8"] 1，1 ~8 m，n Detection temperature T 1，1 ~T m，n The deviation is calculated for the component in the vector of the composition, i.e., the temperature vector, in the direction orthogonal to the sensitivity coefficient vector, and the occurrence of steel leakage is determined based on this deviation.

[0065] For example, in Figure 6 and Figure 7 In the middle, the thermometer 8 was calculated. i，j1 and thermometer 8 i，j2 The detection temperature is the deviation component of the component in the temperature vector of the composition, which is in a direction orthogonal to the sensitivity coefficient vector. Then, the occurrence of steel leakage is determined based on this calculated deviation component. It should be noted that... Figure 6 and Figure 7 In the normal temperature distribution, the direction of the sensitivity coefficient vector is the same as the direction of the first principal component, and the direction orthogonal to the direction of the sensitivity coefficient vector is the same as the direction of the second principal component.

[0066] However, when the detection temperature T 1，1 ~T m，n When the steel leakage is intended to be predicted, if the casting width of the molten steel 2 injected into the mold 5 is changed during operation, or in other words, if the width of the casting sheet 6 pulled from the lower end of the mold 5 is changed, it is possible to incorrectly predict (falsely detect) that steel leakage has occurred even though there are no signs that would lead to steel leakage.

[0067] Figure 8 (a) shows a thermometer 8 in an example where the width (casting width) of the sheet 6 drawn from the lower end of the mold 5 is relatively large. m1，n1 ~8 m1，n1+18 Detection temperature T m1，n1 ~T m1，n1+18 The temperature T' obtained after interpolation m1，n1 ~T' m1，n1+18 A diagram showing the relationships between them. Figure 8 (b) shows a thermometer 8 in an example where the width (casting width) of the sheet 6 drawn from the lower end of the mold 5 is relatively narrow. m1，n1 ~8 m1，n1+18Detection temperature T m1，n1 ~T m1，n1+18 The temperature T' obtained after interpolation m1，n1 ~T' m1，n1+18 A diagram showing the relationships between them. It should be noted that in... Figure 8 (a) and Figure 8 In (b), thermometer 8 m1，n1 ~8 m1，n1+18 It is positioned at the same distance from the top of mold 5 in the casting direction A. Additionally, the temperature T' m1 , n1 ~T' m1，n1+18 The calculation unit 12 is formed by equally dividing the width of the casting 6. m1，n1 ~12 m1，n1+18 The center point of the thermometer 8 m1，n1 ~8 m1，n1+18 Detection temperature T m1，n1 ~T m1，n1+18 The estimated temperature of mold 5 is calculated by performing interpolation. It should be noted that the interpolation method will be explained later.

[0068] The casting width was changed during casting and from Figure 8 (a) Change to Figure 8 In the case of state (b), with regard to thermometer 8 m1，n1 ~8 m1，n1+18 Detection temperature T m1，n1 ~T m1，n1+18 In the case of only detecting temperature T m1，n1+3 and detection temperature T m1，n1+15 The temperature change was significant at one temperature, while no significant temperature changes were observed at other detection temperatures. Therefore, in Figure 8 (a) and Figure 8 In each of the examples shown in (b), when the detection temperature T m1，n1 ~T m1，n1+18 When used to predict steel leakage, it may deviate from the sensitivity coefficient vector and be mistakenly detected as a sign that will lead to steel leakage.

[0069] On the other hand, the casting width changes during casting and from Figure 8 (a) Change to Figure 8 In the case of (b), even if the size of the casting 6 is changed, the number of calculation units 12 (number of units) remains constant, and the temperature T' obtained by interpolation is taken into account. m1，n1 ~T' m1 , n1+18 At that time, the temperature T' m1，n1 ~T' m1，n1+18 The temperature change is relatively small. Therefore, in Figure 8 (a) and Figure 8In each of the examples shown in (b), the temperature T' is obtained by performing interpolation. m1，n1 ~T' m1，n1+18 This technology can be used to predict steel leakage, thereby reducing the risk of false detections that could lead to steel leakage.

[0070] In addition, Figure 8 (a) and Figure 8 In (b), the detection temperature T is measured respectively. m1，n1+7 Temperature T m1，n1+11 Temperature T m1，n1+12 and detection temperature T m1，n1+16 Thermometer 8 m1，n1+7 8 thermometers m1，n1+11 8 thermometers m1，n1+12 and thermometer 8 m1，n1+16 The temperature detection is faulty. Then, in the case of a thermometer 8 containing such a faulty temperature sensor, the temperature T is being measured... m1，n1 ~T m1，n1+18 When used to predict steel leakage, it may deviate from the sensitivity coefficient vector and falsely detect a sign of steel leakage. On the other hand, if the temperature T' obtained through interpolation is used… m1，n1 ~T' m1，n1+18 Even in the case of a thermometer 8 with a malfunctioning temperature sensor, the risk of misdetection as a sign of steel leakage can be reduced by using the estimated temperature of the mold 5 within the range of malfunction.

[0071] Next, the interpolation method will be explained. Figure 9 This indicates the position of thermometer 8, which is located at the same distance from the top of mold 5. i，1 ~8 i，j and calculation unit 12 i，1 ~12 i，j A diagram showing the positional relationships.

[0072] like Figure 9 As shown, computing unit 12 i，1 ~12 i，j This refers to the thermometer 8 located on the long side cooling plate 5a of the mold 5 at a position equidistant from the top of the mold 5. i，1 ~8 i，j The unit is formed by equally dividing a section of the long-side cooling plate 5a, which is equivalent to the width of the casting sheet 6 (the section sandwiched between a pair of short-side cooling plates 5b in the width direction of the mold 5), into units with a certain number of units. Then, the thermometer 8... i，1 ~8 i，j The detected temperature is linearly interpolated to calculate the value of calculation unit 12. i，1 ~12 i，jThe estimated temperature of the mold 5 (long-side cooling plate 5a) at the center point of each of the points. It should be noted that the number of calculation units 12 used for interpolation can be the same as or different from the number of thermometers 8 in the vertical and horizontal directions, but is not fixed depending on the variation of the casting width during casting.

[0073] The interpolation process described above can be applied to cases where principal component analysis (PCA) is used to obtain the sensitivity coefficient vector and to calculate the deviation. In this case, the temperature obtained through interpolation is used instead of the actual detection temperature for PCA. Since the same number of temperature vectors can be used even when the casting width changes, PCA can be performed with data for different widths. Therefore, it is not necessary to calculate different influence coefficients based on width; the influence coefficient vector can be determined by including data for different casting widths. Then, the deviation can also be calculated using the influence coefficient vector, which is calculated based on the temperature obtained through interpolation of the detection temperature. Therefore, leakage prediction for different casting widths can be performed based on a unified benchmark. Furthermore, when the casting width changes during casting, the risk of false detections related to signs that lead to leakage can be reduced.

[0074] Next, the determination of steel leakage prediction will be explained. Figure 10 (a) is a graph showing the temporal variation of the absolute value of the deviation in cases where sintering occurs. Figure 10 (b) is a graph showing the temporal variation of the rate of change of deviation in cases where sintering occurs. Figure 11 (a) is a graph showing the time-series changes in the absolute value of the deviation in cases where sintering did not occur. Figure 11 (b) is a graph showing the time-series change of the deviation rate in cases where sintering did not occur.

[0075] exist Figure 10 In (a), at some point during the operation, the absolute value of the deviation increases sharply. On the other hand, in Figure 11 In (a), during operation, the absolute value of the deviation is consistently large. Based on thermometer 8... 1，1 ~8 m，n If the sensitivity coefficient calculated from the detected temperature through interpolation deviates from the value pre-calculated based on factors such as changes in the surface shape of mold 5, such as... Figure 11 As shown in (a), even without anomalies such as sintering, the absolute value of the deviation can still consistently increase. Therefore, as Figure 10 (a) and Figure 11 As shown in (a), when a single threshold X is set for the absolute value of the deviation, it is difficult to determine whether sintering, which can be a precursor to steel leakage, has occurred.

[0076] Here, as a precursor to sintering failure, sintering suddenly occurs, and the cracks 11 in the solidified shell 10 propagate downwards and laterally towards the mold 5. Therefore, as Figure 10 As shown in (a), the absolute value of the deviation during sintering increases sharply at a certain point in the operation. Therefore, as Figure 10 As shown in (b), the rate of change of the deviation over time increases sharply. On the other hand, as... Figure 11 As shown in (a), even without anomalies such as sintering, if the absolute value of the deviation during operation is consistently large, such as Figure 11 As shown in (b), the rate of change of deviation over time does not increase dramatically. Therefore, as Figure 10 (b) and Figure 11 As shown in (b), by setting a single threshold Y for the time-varying rate of deviation, it becomes easy to determine whether sintering, which is a precursor to steel leakage, has occurred.

[0077] Next, a method for determining the adjacency of calculation units 12 that exceed the threshold X is explained when the absolute value of the deviation calculated from the sensitivity coefficient vector exceeds the threshold Y.

[0078] Figure 12 This illustrates the case where the computational unit 12 performing the interpolation process is a single-layer structure (computational unit 12). 1，1 ~12 1，p A diagram illustrating an example of a method for determining the adjacency of ( ). That is, in Figure 12 The diagram shows a calculation unit 12 located at the same distance from the upper end of the mold 5 in the casting direction A. 1，1 ~12 1，p Examples of methods for determining lateral adjacency. It should be noted that in... Figure 12 In the adjacentity determination method shown in this example, the condition that the time change rate of the deviation exceeds the threshold Y is taken as the premise.

[0079] In the adjacency determination method of this example, firstly, for calculation unit 12... 1，1 ~12 1，p In the calculation unit 12, where the absolute value of the deviation exceeds a preset threshold X as described above, 1 point is assigned as the first score, i.e., the score for different calculation units. On the other hand, for calculation unit 12... 1，1 ~12 1，pCalculation units 12 whose absolute deviation value does not exceed the threshold X are assigned a score of 0. Then, relative to the vector of these different calculation unit scores, the vector obtained by shifting the different calculation unit scores to the previous calculation unit 12 is designated as the forward shift vector, and the vector obtained by shifting the different calculation unit scores to the next calculation unit 12 is designated as the backward shift vector. Furthermore, the vector obtained by multiplying the elements of the forward and backward shift vectors is designated as the adjacent product vector. When the adjacent product vector determined in this way is calculated, if there are three adjacent calculation units 12 whose absolute deviation value exceeds the threshold X, the score of the central calculation unit 12 among the three adjacent calculation units 12 becomes 1 point, and the scores of the other calculation units 12 become 0 points; therefore, this score is designated as the second score.

[0080] use Figure 12 The example shown illustrates in detail that, Figure 12 First, due to computing unit 12 1，1 ~12 1，p Calculation unit 12 in 1，3 Calculation Unit 12 1，4 Calculation Unit 12 1，5 The absolute value of the deviation exceeds the set threshold X, so a signal is sent to calculation unit 12. 1，3 Calculation Unit 12 1，4 Calculation Unit 12 1，5 One point is assigned as the score for each calculation unit (the first score). On the other hand, for the other calculation units 12... 1，1 Calculation Unit 12 1，2 and calculation unit 12 1，6 ~12 1，p A score of 0 is assigned as the score for each calculation unit (first score). Then, the vector formed by arranging these different calculation unit scores (first scores) is (0, 0, 1, 1, 1, 0, ..., 0, 0, 0). The forward shift vector is (0, 1, 1, 1, 0, 0, ..., 0, 0, 0), and the backward shift vector is (0, 0, 0, 1, 1, 1, ..., 0, 0, 0). The adjacent product vector obtained by multiplying the elements of the forward and backward shift vectors is (0, 0, 0, 1, 0, 0, ..., 0, 0, 0). Therefore, it can be seen that when there are three adjacent calculation units 12 exceeding the threshold X, the three adjacent calculation units 12 exceeding the threshold X... 1，3 Calculation Unit 12 1，4 Calculation Unit 12 1，5 The central computing unit 12 1，4 The score (second score) becomes 1 point, except for the other calculation unit 12. 1，1 ~121，3 and calculation unit 12 1，5 ~12 1，p The score (second score) becomes 0.

[0081] Therefore, in use Figure 12 In the method for determining adjacency, if any element of the adjacent product vector becomes 1, it can be determined that a sintering or other precursor that could lead to steel leakage has occurred.

[0082] It should be noted that, in Figure 12 In this process, the vector obtained by shifting the scores of different calculation units to the previous calculation unit 12 is set as the forward shift vector, and the vector obtained by shifting the scores of different calculation units to the next calculation unit 12 is set as the backward shift vector. The adjacent product vector of three adjacent calculation units 12 is then calculated, but this is not a limitation. That is, depending on the set number of calculation units 12, the vector obtained by shifting the scores of different calculation units to one or more preceding calculation units 12 can be set as the forward shift vector, and the vector obtained by shifting the scores of different calculation units to one or more subsequent calculation units 12 can be set as the backward shift vector. It should be noted that, at this time, the number of shifts required to obtain the backward shift vector is the same as the number of shifts required to obtain the forward shift vector. Then, the vector obtained by multiplying the elements of the forward shift vector and the backward shift vector obtained in this way can be set as the adjacent product vector.

[0083] For example, the vector obtained by shifting the scores of different calculation units to the front three calculation units 12 is set as the forward shift vector, and the vector obtained by shifting the scores of different calculation units to the back three calculation units 12 is set as the backward shift vector. Then, the elements of the forward shift vector and the backward shift vector are multiplied together to calculate the adjacent product vector of the seven adjacent calculation units 12, and the second score is obtained. If any element of the adjacent product vector becomes 1, it is determined that a sintering or other precursor that could lead to steel leakage has occurred. Thus, since the occurrence of a precursor that could lead to steel leakage can be determined with higher accuracy, steel leakage can be predicted with high precision.

[0084] Furthermore, when the calculation unit 12 used for interpolation processing is configured in two or more layers in the casting direction A, the above-mentioned method for determining adjacency can also be extended.

[0085] Figure 13 This shows the calculation unit 12 in the casting direction A (longitudinal direction) with two layers: the upper layer and the lower layer (calculation unit 12). 1，1 ~12 1，p and calculation unit 12 2，1 ~12 2，p Configuration, in the upper-level computing unit 121，1 ~12 1，p Three adjacent numbers in the middle are scored, and the scores are calculated in the lower-level calculation unit 12. 2，1 ~12 2，p The computational unit 12 corresponds to one of the three adjacent units in the middle and upper layers. 2，i A diagram showing the determination method for satisfying the condition of adjacency when a score is obtained.

[0086] In this method, firstly, for the upper-level computing unit 12 1，1 ~12 1，p The upper-level computational unit 12 is determined by using scores from different computational units (first score) that show whether the absolute value of the deviation exceeds the threshold X. 1，1 ~12 1，p Based on the adjacency of the elements, calculate the adjacent product vector of the upper layer.

[0087] exist Figure 13 In the middle, it is the upper-level computing unit 12 1，1 ~12 1，p Middle Calculation Unit 12 1，3 Calculation Unit 12 1，4 Calculation Unit 12 1，5 In the case where the absolute value of the deviation exceeds the threshold X, the upper-level neighboring product vector is (0, 0, 0, 1, 0, 0, ..., 0, 0, 0). It should be noted that the method for calculating the upper-level neighboring product vector differs from the method used... Figure 12 The method for finding adjacent product vectors is the same, so detailed explanations are omitted here.

[0088] Next, regarding the lower-level computing unit 12 2，1 ~12 2，p The sum of all elements of the score vector, forward shift vector, and backward shift vector of different calculation units is taken. If any one of them has a score, then the calculation unit 12 is... 2，1 ~12 2，p The score is set to 1 point. Then, the vector formed by arranging these scores is set as the lower-level adjacent sum vector. Next, the vector formed by multiplying the upper-level adjacent product vector by each element of the lower-level adjacent sum vector is set as the upper and lower adjacent product vector. Finally, if any element of the upper and lower adjacent product vectors has a score of 1 (the second score), then the adjacency is determined to be valid.

[0089] exist Figure 13 In the example shown, it is the lower-level computing unit 12. 2，1 ~12 2，p Calculation unit 12 in 2，3If the absolute value of the deviation exceeds the threshold X, the lower-level adjacent sum vector is (0, 1, 1, 1, 0, 0, ..., 0, 0, 0). Then, since the upper and lower adjacent product vector is (0, 0, 0, 1, 0, 0, ..., 0, 0, 0), and there is an element that scores 1 point as the second score, it can be determined that the adjacency is valid.

[0090] By determining the adjacency, the location where sintering occurs in the mold 5 can be identified. Furthermore, by increasing the number of thermometers 8 in the casting direction A, even in the event of sintering that could lead to leakage, the longitudinal propagation of the fracture 11 in the casting direction A can be monitored based on the phenomenon of propagation in the casting direction A according to the adjacency determination.

[0091] Therefore, in use Figure 13 In the method for determining adjacency, if any element of the adjacent product vectors above and below becomes 1, it can be determined that a sintering or other precursor that could lead to steel leakage has occurred.

[0092] It should be noted that the calculation unit 12 was not considered in the above description of this embodiment. 1，1 ~12 k，p The mold 5 is positioned within the mold 5, but is equipped with temperature gauges 8 located on the long side cooling plate 5a and short side cooling plate 5b of the mold 5, as well as on the front and back sides of the mold 5. 1，1 ~8 m，n Interpolation is performed independently on each surface based on computational unit 12. 1，1 ~12 k，p The second fraction is calculated based on the adjacent relationships, enabling more precise discrimination. Furthermore, the number of adjacent vectors used to calculate the adjacent product vector and the adjacent sum vector is not limited to three and can be varied.

[0093] Furthermore, the phenomenon of steel leakage within the mold 5 in the continuous casting process not only propagates laterally but also manifests as a temperature change from the upstream side to the downstream side (from above to below the mold 5) in the casting direction A. That is, the crack 11 of the solidified shell 10 moves downwards while repeating the following phenomenon: due to some reason, the mold 5 comes into contact with the molten steel 2 and sintering occurs; the solidified shell 10 is restricted by the mold 5, further pulling the molten steel 2 from the lower part of the mold 5; therefore, at the crack 11 of the solidified shell 10 directly below the sintering, the mold 5 comes into contact with the molten steel 2 and further sintering occurs. Additionally, by taking the logical product of the adjacent sums and vectors of the calculation units 12 for the upper and lower layers, the adjacency (the occurrence of the same phenomenon at adjacent positions) of the upper and lower layers is determined. Therefore, it is not necessary for multiple thermometers 8 or multiple calculation units 12 to be arranged at the same distance from the upper end of the mold 5 in the casting direction A.

[0094] Figure 14This is a chart of time-series detection data of instances of steel leakage predicted using the steel leakage prediction method (the method of the present invention) according to an embodiment of the present invention. It should be noted that... Figure 14 In this context, time t1 is the instant at which the steel leakage is predicted using the leakage prediction method according to an embodiment of the present invention. Furthermore, in... Figure 14 In this process, time t2 is used to predict the moment of steel leakage using a conventional method for predicting steel leakage. It should be noted that this conventional method predicts leakage based on the fact that the temperature detected by the upper thermometer 8 in the two-layer thermometer structure is lower than the temperature detected by the lower thermometer 8 for a certain period of time. Furthermore, at time t2, by predicting the steel leakage, the casting speed is controlled to decrease to a predetermined value.

[0095] like Figure 14 As shown, by using the leakage prediction method of the present invention, leakage can be predicted at a faster time than conventional leakage prediction methods that calculate the temperature change from time-series data of the detected temperature.

[0096] Furthermore, Table 1 below shows the results of applying the leakage prediction method (the method of the present invention) according to the embodiments of the present invention to past leakage prediction cases. It should be noted that in Table 1 below, Examples 1 and 5 are examples where leakage occurred, and Examples 2 to 4 are examples where leakage did not occur. Moreover, in Table 1 below, "positive detection" means that in examples where leakage occurred, the occurrence of a sign that would lead to leakage was correctly detected, and thus the leakage was correctly predicted. In addition, in Table 1 below, "overdetection" means that in examples where leakage did not occur, excessive detection (false detection) would lead to a sign that would lead to leakage, and thus the leakage was incorrectly predicted. Furthermore, in Table 1 below, "no detection" means that in examples where leakage did not occur, the occurrence of a sign that would lead to leakage was not detected, and thus the leakage was not predicted.

[0097] [Table 1]

[0098] Previous methods The method of the present invention Case 1 Positive detection Positive detection Case 2 Over-detection Not detected Case 3 Over-detection Not detected Case 4 Over-detection Not detected Case 5 Positive detection Positive detection

[0099] As can be seen from Table 1 above, the steel leakage prediction method according to the embodiment of the present invention can correctly detect all past cases in which steel leakage has occurred, the occurrence of signs that will lead to steel leakage, and correctly predict the occurrence of steel leakage. Furthermore, for past cases in which steel leakage has not occurred, there is no over-detection (false detection) that occurred in conventional methods.

[0100] Industrial availability

[0101] This invention provides a method for predicting steel leakage with high precision, an operation method for a continuous casting machine, and a device for predicting steel leakage.

[0102] Explanation of reference numerals in the attached figures

[0103] 1. Continuous casting machine

[0104] 2. Molten steel

[0105] 3. Intermediate package

[0106] 4. Dipping nozzle

[0107] 5. Casting mold

[0108] 6 Castings

[0109] 7 Casting support rollers

[0110] 8. Thermometer

[0111] 10. Solidified shell

[0112] 11. Fractured section

[0113] 20 Judgment Department

Claims

1. A method for predicting steel leakage, characterized in that, have: The steps for inputting the dimensions of the sheet pulled from the mold in a continuous casting machine; The step of detecting the temperature of the mold using multiple thermometers embedded in the mold; The step of performing interpolation processing on the detected temperatures of the plurality of thermometers based on the size of the casting sheet; Based on the temperature calculated by performing the interpolation process, the component in the direction orthogonal to the influence coefficient vector obtained from the principal component analysis is used as the deviation from the normal operation when no steel leakage occurs. as well as The steps for predicting steel leakage based on the aforementioned deviation are as follows. In the step of performing the interpolation process For the temperature detected by each of the plurality of thermometers, interpolation is performed at the center point of each of the plurality of calculation units equally divided according to the size of the casting to calculate the temperature. Even if the size of the casting is changed, the number of calculation units remains constant. In the step of predicting the steel leakage... If the time rate of change of the deviation exceeds a preset first threshold, the calculation unit predicts steel leakage based on the adjacency of the absolute value of the deviation exceeding a preset second threshold. The influence coefficient vector is a sensitivity coefficient vector composed of the sensitivity coefficient of each of the plurality of thermometers. The sensitivity coefficient vector is a vector representing the direction of the average movement of the thermometers during normal operation, showing the temperature of the calculation unit obtained by the interpolation process.

2. The method for predicting steel leakage according to claim 1, characterized in that, In the step of calculating the deviation, The average temperature of each of the plurality of calculation units located at the same distance from the upper end of the mold in the casting direction of the molten steel relative to the mold is calculated. The difference between the temperature of each of the plurality of calculation units and the average temperature is calculated. Based on the calculated difference, the deviation is calculated using the influence coefficient vector.

3. The method for predicting steel leakage according to claim 1 or 2, characterized in that, The steps for predicting steel leakage include: The step of assigning a first score to the calculation unit whose deviation exceeds the second threshold; The step of calculating the second fraction based on the adjacency of the calculation units assigned the first fraction; and The steps for predicting steel leakage are based on the second score.

4. The operation method of a continuous casting machine, characterized in that, In the case where a steel leakage is predicted based on the steel leakage prediction method according to any one of claims 1 to 3, the casting speed of injecting molten steel into the mold is reduced.

5. A device for predicting steel leakage, characterized in that, have: Input mechanism: Input the dimensions of the casting sheet pulled from the mold in the continuous casting machine; Multiple thermometers are embedded in the mold to detect the temperature of the mold; An interpolation processing actuator performs interpolation processing on the detected temperatures of the plurality of thermometers based on the dimensions of the casting. The deviation calculation mechanism calculates the deviation relative to normal operation when no steel leakage occurs, based on the temperature calculated by performing the interpolation process and using the component in the direction orthogonal to the influence coefficient vector obtained from the principal component analysis as the component. as well as The steel leakage prediction mechanism predicts steel leakage based on the deviation. The interpolation processing mechanism calculates the temperature for each of the plurality of thermometers by performing interpolation processing at the center point of each of the plurality of calculation units equally divided according to the size of the casting. Even if the size of the casting changes, the number of calculation units remains constant. The steel leakage prediction mechanism, when the time change rate of the deviation exceeds a preset first threshold, predicts steel leakage based on the adjacency of the calculation unit when the absolute value of the deviation exceeds a preset second threshold. The influence coefficient vector is a sensitivity coefficient vector composed of the sensitivity coefficient of each of the plurality of thermometers. The sensitivity coefficient vector is a vector representing the direction of the average movement of the thermometers during normal operation, showing the temperature of the calculation unit obtained by the interpolation process.

Images