Breakout prediction method and method for operating continuous casting machine

Abstract

A method of predicting a breakout of the present invention includes: a step of receiving input of a dimension of a cast strand to be pulled out from a mold included in a continuous casting machine; a step of detecting temperatures of the mold with a plurality of thermometers buried in the mold; a step of executing interpolation processing in accordance with a dimension of the cast strand on detected temperatures detected by the plurality of thermometers; a step of calculating temperature change amounts by comparison with temperatures before a first period; a step of acquiring a standard deviation of temperature change amounts of the thermometers in a second period; a step of standardizing temperature changes by dividing the temperature change amounts by the standard deviation; a step of calculating a component in a direction orthogonal to an influence coefficient vector obtained by principal component analysis, as a deviation degree from a normal operation time when no breakout occurs, based on a temperature change amount calculated by executing the interpolation processing; and a step of predicting the breakout based on the deviation degree.

Description

Field

[0001] The present invention relates to a method of predicting a breakout and a method of operating a continuous casting machine.Background

[0002] Patent Literatures 1 and 2 disclose a method of predicting a breakout by detecting a characteristic behavior of temperature by use of a temperature sensor arranged in a mold. Furthermore, Patent Literature 3 discloses a method of calculating a deviation degree from a mold temperature during normal operation in which no breakout occurs and predicting a breakout based on the calculated deviation degree.Citation ListPatent Literature

[0003] Patent Literature 1: JP S63-47545 B Patent Literature 2: JP 2005-296979 A Patent Literature 3: Japanese Patent No. 6950860 SummaryTechnical Problem

[0004] Patterns of breakouts are roughly classified into two patterns of a sticking breakout and a foreign object entrapment breakout. In the sticking breakout, a cast strand is stuck onto a mold, and a relatively large temperature fluctuation often occurs. The foreign object entrapment breakout occurs when casting proceeds with a foreign object such as powder being attached on the surface of the cast strand. The foreign object entrapment breakout thus has a feature in that a change width and a change magnitude of a mold temperature are extremely small.

[0005] In the methods disclosed in Patent Literatures 1 and 2, a temporal change and a spatial distribution of mold temperatures are detected using a characteristic temperature at the time when a breakout actually occurs. The sticking breakout and the foreign object entrapment breakout are detected by different methods. The sticking breakout having a relatively large temperature change is determined by the temperature relation between temperatures of different casting heights (normally high temperature in upper casting direction and, at abnormal time, high temperature in lower casting direction). Since the foreign object entrapment breakout is difficult to detect because a temperature change is small, the foreign object entrapment breakout is thus determined based on the magnitude of temperature fluctuation and downward propagation of a temperature change. However, since operation is not performed with a constant distribution of mold temperatures and the temperature changes depending on the situation inside the mold, erroneous detection often occurs, and detection has been often omitted.

[0006] Furthermore, in the method disclosed in Patent Literature 3, a breakout is predicted based on a deviation degree from a mold temperature at a normal operation time. Thus, when a cast strand is stuck onto a mold, and a relatively large temperature fluctuation occurs as in the sticking breakout, the breakout can be sufficiently detected. In contrast, in a foreign object entrapment breakout that occurs when casting proceeds with a foreign object such as powder being attached on the surface of the cast strand, a change width and a change magnitude of a mold temperature are very small, and the breakout cannot be sufficiently detected.

[0007] The present invention has been made in view of the above-described problems, and an object thereof is to provide a method of predicting a breakout and a method of operating a continuous casting machine capable of detecting, with sufficient detection accuracy, a foreign object entrapment breakout in which a change width of a mold temperature and temperature fluctuation of a mold is small.Solution to Problem

[0008] To solve the problems and achieve an object, (1) there is provided a method of predicting a breakout, including: a step of receiving input of a dimension of a cast strand to be pulled out from a mold included in a continuous casting machine; a step of detecting temperatures of the mold with a plurality of thermometers buried in the mold; a step of executing interpolation processing in accordance with a dimension of the cast strand on detected temperatures detected by the plurality of thermometers; a step of calculating temperature change amounts by comparison with temperatures before a first period; a step of acquiring a standard deviation of temperature change amounts of the thermometers in a second period; a step of standardizing temperature changes by dividing the temperature change amounts by the standard deviation; a step of calculating a component in a direction orthogonal to an influence coefficient vector obtained by principal component analysis, as a deviation degree from a normal operation time when no breakout occurs, based on a temperature change amount calculated by executing the interpolation processing; and a step of predicting the breakout based on the deviation degree. (2) In the method of predicting a breakout according to the above-described invention (1), in the step of executing interpolation processing, a temperature is calculated by executing interpolation processing on detected temperatures of the plurality of thermometers at a center point of each of a plurality of calculation cells obtained by equal division in accordance with the dimension of the cast strand. (3) In the method of predicting a breakout according to the above-described invention (2), a number of the calculation cells is kept constant even when the dimension of the cast strand is modified. (4) In the method of predicting a breakout according to the above-described invention (2) or (3), in the step of calculating the deviation degree, an average value of temperatures of the plurality of calculation cells at positions with a same distance from an upper end of the mold in a casting direction of molten steel toward the mold is determined, differences of the temperatures of the plurality of calculation cells from the average value are determined, and the deviation degrees are calculated by using the influence coefficient vector. (5) In the method of predicting a breakout according to the above-described invention (4), in the step of predicting the breakout, when some of calculated individual deviation degrees of the calculation cells exceed a preset first threshold or a calculated total deviation degree of the calculation cells exceeds a preset second threshold, the breakout is predicted. (6) In the method of predicting a breakout according to the above-described invention of any one of (1) to (5), the influence coefficient vector is a sensitivity coefficient vector having, as a component, a sensitivity coefficient of each of a plurality of the temperature change amounts. (7) A method of operating a continuous casting machine according to the present invention, includes decreasing a casting speed when a breakout is predicted based on the method of predicting a breakout according to any one invention of (1) to (6). Advantageous Effects of Invention

[0009] A method of predicting a breakout and a method of operating a continuous casting machine according to the present invention have an effect of detecting, with sufficient detection accuracy, a foreign object entrapment breakout having a small temperature change width and temperature fluctuation of a mold.Brief Description of Drawings

[0010] [FIG. 1] FIG. 1 is a schematic diagram illustrating a schematic configuration of a continuous casting machine according to an embodiment. [FIG. 2] FIG. 2 is a perspective view illustrating a schematic configuration of a mold, in which thermometers are buried, of the continuous casting machine according to the embodiment. [FIG. 3] FIG. 3(a) illustrates a situation of molten steel and a solidified shell inside the mold in a sign phenomenon of a breakout. FIG. 3(b) illustrates a situation of a fracture portion of the solidified shell in the sign phenomenon of a breakout. [FIG. 4] FIG. 4(a) illustrates a temperature distribution of the mold at the moment when sticking has occurred. FIG. 4(b) illustrates a temperature distribution of the mold 10 seconds after the moment when sticking has occurred. [FIG. 5] FIG. 5 illustrates a foreign object entrapment breakout. [FIG. 6] FIG. 6 is a flowchart illustrating an example of a procedure of a method of predicting a breakout according to the embodiment. [FIG. 7] FIG. 7 illustrates a correlation between detected temperatures of thermometers under normal operation when no breakout occurs. [FIG. 8] FIG. 8 illustrates a correlation between detected temperatures of thermometers at the time when a sign such as sticking that leads to a breakout occurs. [FIG. 9] FIG. 9(a) illustrates the relation between detected temperatures of thermometers and temperatures on which interpolation processing has been executed in a case of a wide width of a cast strand pulled out from the lower end of the mold. FIG. 9(b) illustrates the relation between detected temperatures of thermometers and temperatures on which interpolation processing has been executed in a case of a narrow width of a cast strand pulled out from the lower end of the mold. [FIG. 10] FIG. 10 illustrates the positional relation between thermometers at positions with the same distance from the upper end of the mold and calculation cells. [FIG. 11] FIG. 11(a) illustrates time-series changes of deviation degrees in a case where sticking (sticking breakout) is confirmed. FIG. 11(b) illustrates a change of a deviation degree at each of the calculation cells in a case where sticking (sticking breakout) is confirmed. [FIG. 12] FIG. 12(a) illustrates time-series changes of deviation degrees in a case where foreign object entrapment (foreign object entrapment breakout) is confirmed. FIG. 12(b) illustrates a change of a deviation degree at a position of each of the calculation cells in a case where foreign object entrapment (foreign object entrapment breakout) is confirmed. [FIG. 13] FIG. 13(a) illustrates a case where a result of slab inspection has no abnormality though total deviation degrees have peaks. FIG. 13(b) illustrates a deviation degree in each of the calculation cells in a case where a result of slab inspection has no abnormality though total deviation degrees have peaks. [FIG. 14] FIG. 14 illustrates a result of breakout prediction using a method of predicting a breakout according to the embodiment and a conventional method. Description of Embodiments

[0011] An embodiment of a method of predicting a breakout and a method of operating a continuous casting machine according to the present invention will be described below. Note that the present invention is not limited by the embodiment.

[0012] FIG. 1 is a schematic diagram illustrating a schematic configuration of a continuous casting machine 1 according to the embodiment. As illustrated in FIG. 1, the continuous casting machine 1 according to the embodiment includes a tundish 3, a mold 5, a plurality of strand support rolls 7, and a determination unit 20. Molten steel 2 is poured into the tundish 3. The mold 5 is made of copper, and cools the molten steel 2 poured from the tundish 3 via a submerged entry nozzle 4. The plurality of strand support rolls 7 conveys a semi-solidified cast strand 6 pulled out from the mold 5. The determination unit 20 determines a sign phenomenon of a breakout from a detected temperature of a thermometer 8 buried in the mold 5. Note that, although a thermocouple is used as the thermometer 8 in the embodiment, this is not a limitation.

[0013] FIG. 2 is a perspective view illustrating a schematic configuration of the mold 5, in which thermometers 8 1,1 to 8 m,n are buried, of the continuous casting machine 1 according to the embodiment. As illustrated in FIG. 2, the mold 5 includes a pair of long side cooling plates 5a and a pair of short side cooling plates 5b, and has a substantially rectangular cylindrical shape penetrating in the vertical direction. Cooling water channels (not illustrated) are formed along inner wall surfaces inside the long side cooling plates 5a and the short side cooling plates 5b. The molten steel 2 is cooled by distributing cooling water through the cooling water channels.

[0014] Furthermore, the thermometers 8 1,1 to 8 m,n are buried inside a long side cooling plate 5a of the mold 5 from an outer wall surface of the long side cooling plate 5a at a predetermined depth. Note that, in the following description, the thermometers 8 1,1 to 8 m,n are also simply referred to as thermometers 8 unless particularly distinguished. FIG. 2 illustrates a configuration of three or more stages of the thermometers 8 1,1 to 8 m,n in a casting direction A. The thermometers 8 1,1 to 8 m,n are divided into a first stage of thermometers 8 1,1 to 8 1,n , a second stage of thermometers 8 2,1 to 8 2,n , and an nth stage of thermometers 8 m,1 to 8 m,n , which are buried on the same plane. In the embodiment, in the casting direction A, the molten steel 2 is poured from the tundish 3 into the mold 5 via the submerged entry nozzle 4. The casting direction A is the same as a direction in which the cast strand 6 is pulled out from the lower end of the mold 5.

[0015] Note that the arrangement of the thermometers 8 in FIG. 2 is an example for describing the present invention. The thermometers 8 are only required to 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 among the pair of long side cooling plates 5a and the pair of short side cooling plates 5b of the mold 5. Among these options, the thermometers are preferably arranged in all of the pair of long side cooling plates 5a and the pair of short side cooling plates 5b. Furthermore, the thermometers 8 can be arranged in the mold 5 in a multi-stage array having more than three stages or a single-stage array in the casting direction A.

[0016] Next, the sign phenomenon of a breakout will be described. Breakouts targeted in the present proposal are roughly classified into two types. One type of breakout is called a sticking breakout, and the other type of breakout is called a foreign object entrapment breakout. First, for the sticking breakout, FIG. 3(a) illustrates a situation of the molten steel 2 and a solidified shell 10 inside the mold 5 in a sign phenomenon of a sticking breakout. FIG. 3(b) illustrates a situation of a fracture portion 11 of the solidified shell 10 in the sign phenomenon of a sticking breakout.

[0017] As illustrated in FIGS. 3(a) and 3(b), in the sign phenomenon of a sticking breakout, sticking occurs inside the mold 5 for some reason, and the solidified shell 10 is stuck onto the mold 5. In contrast, since the cast strand 6 is pulled out from the lower end of the mold 5 in the same direction as the casting direction A in FIG. 3(b), the fracture portion 11 of the solidified shell 10 is generated immediately below the sticking. In the fracture portion 11 of the solidified shell 10, the mold 5 and the molten steel 2 are in contact with each other, which causes further sticking. While the above-described phenomenon is repeated, the fracture portion 11 of the solidified shell 10 moves downward, and the solidified shell 10 above the fracture portion 11 becomes thicker. Then, finally, when the fracture portion 11 passes through the lower end of the mold 5, the molten steel 2 leaks from the fracture portion 11, and a sticking breakout occurs.

[0018] Note that, in the fracture portion 11, the molten steel 2 and the mold 5 are in contact with each other, so that the temperature of the mold 5 locally rises. Thus, for example, as indicated by an arrow B in FIG. 3(b), when the fracture portion 11 moving downward passes through the arrangement positions of thermometers 8 m',1 to 8 m',n , the detected temperatures of the thermometers 8 m',1 to 8 m',n increase. Thereafter, since the solidified shell 10 above the fracture portion 11 is stuck by the mold 5 and continues to be cooled, the detected temperatures of the thermometers 8 m',1 to 8 m',n monotonously decrease. In contrast, since the fracture portion 11 propagates not only in the downward direction but in the lateral direction, the fracture portion 11 enlarges to have a V shape as illustrated in FIG. 3(b). Note that, when the fracture portion 11 of the solidified shell 10 is generated below the thermometers 8 m',1 to 8 m',n , the fracture portion 11 does not pass through the positions of the thermometers 8 m',1 to 8 m',n , so that only a decrease in the detected temperatures of the thermometers 8 m',1 to 8 m',n is observed.

[0019] FIG. 4(a) illustrates a temperature distribution of the mold 5 at the moment when sticking has occurred. FIG. 4(b) illustrates a temperature distribution of the mold 5 10 seconds after the moment when sticking has occurred. From the temperature distributions of the mold 5 in FIGS. 4(a) and 4(b), it can be read that the V-shaped high temperature portion propagates in the downward direction and the lateral direction.

[0020] Next, the foreign object entrapment breakout will be described. FIG. 5 illustrates the foreign object entrapment breakout. Note that, in FIG. 5, a reference numeral 18 denotes a meniscus (molten-metal surface) of the molten steel 2. As illustrated in FIG. 5(a), mold powder 23 is used as a lubricant to prevent the above-described sticking between the mold 5 and the cast strand 6 during casting. After being supplied from an upper portion of the mold 5 onto the molten-metal surface, the mold powder 23 receives heat from the molten steel 2, comes into a molten state, and flows into a gap between the mold 5 and the cast strand 6. The mold powder 23 serves as a lubricant to prevent sticking. As illustrated in FIG. 5(b), however, the mold powder 23 supplied to the molten-metal surface of the molten steel 2 may be pulled into the mold 5 as an unmelted / massive (state called powder bear) foreign object 22. In the case, thermal resistance that is not originally assumed is generated in the gap between the mold 5 and the cast strand 6. As a result, as illustrated in FIGS. 5(c) and 5(d), the solidified shell 10 reaches the lower end of the mold 5 without being sufficiently grown at a site where the foreign object 22 is entrapped. In the case, as illustrated in FIG. 5(e), it is considered that, after the foreign object 22 passes through the lower end of the mold 5 with a thin local shell thickness, the solidified shell 10 cannot withstand molten steel static pressure, the solidified shell 10 fractures, and the molten steel 2 flows out to the outside. In the case, a temperature tendency that greatly differs from the temperature distribution inside the mold at the time when sticking occurs is exhibited. First, since the entrapped foreign object 22 is considered to be approximately 100 [mm] or less, the mold temperature changes with only approximately the size of the foreign object 22 when the foreign object 22 passes through the inside of the mold 5. That is, in a peripheral portion through which the foreign object 22 has passed, the temperature after the foreign object 22 has passed does not change from a temperature before the generation. Furthermore, a temperature change amount is relatively smaller than that in a case where sticking causes a fracture. Thus, in the breakout pattern, detection is generally difficult, and many erroneous detections occur.

[0021] Therefore, the method of predicting a breakout according to the embodiment improves accuracy of predicting a breakout by using a temperature change amount of a difference of a thermometer installed inside the mold 5 from that before a certain period (before first period) in order to accurately detect breakouts in the above-described two patterns and standardizing the temperature change amount with a standard deviation in a certain period (second period) of the temperature change amount. The method of predicting a breakout according to the embodiment based on the above-described technical idea will be described in detail below.

[0022] The method of predicting a breakout according to the embodiment includes a step of receiving input of the dimension of the cast strand 6 to be pulled out from the mold 5 included in the continuous casting machine 1. Furthermore, the method of predicting a breakout according to the embodiment includes a step of detecting the temperatures of the mold 5 with the plurality of thermometers 8 buried in the mold 5. Furthermore, the method of predicting a breakout according to the embodiment includes a step of executing interpolation processing in accordance with the dimension of the cast strand 6 on the detected temperatures detected by the plurality of thermometers 8. Furthermore, the method of predicting a breakout according to the embodiment includes a step of calculating temperature change amounts by performing comparison with temperatures before a certain period. Furthermore, the method of predicting a breakout according to the embodiment includes a step of acquiring the standard deviation of temperature change amounts for the thermometers 8 in the certain period. Furthermore, the method of predicting a breakout according to the embodiment includes a step of standardizing temperature changes by dividing the temperature change amounts by the standard deviation. Furthermore, the method of predicting a breakout according to the embodiment includes a step of calculating a component in a direction orthogonal to an influence coefficient vector obtained by principal component analysis, as a deviation degree from a normal operation time when no breakout occurs, based on the temperature change amounts calculated by executing the interpolation processing. Furthermore, the method of predicting a breakout according to the embodiment includes a step of predicting a breakout based on the deviation degree.

[0023] FIG. 6 is a flowchart illustrating an example of a procedure of the method of predicting a breakout according to the embodiment. The determination unit 20 in FIG. 1 executes the method of predicting a breakout in the flowchart. Note that the determination unit 20 has at least functions of a device of executing interpolation processing, a device of standardizing temperature change amounts, a device of calculating a deviation degree, and a device of predicting a breakout in the present invention. Furthermore, details of the steps in the flowchart in FIG. 6 will be appropriately described later.

[0024] In the method of predicting a breakout according to the embodiment, the determination unit 20 preliminarily calculates a sensitivity coefficient for standardized temperature change amounts for the thermometers 8 1,1 to 8 m,n under normal operation (hereinafter, also referred to as normal operation) when no breakout has occurred (Step S1). Here, the sensitivity coefficient is calculated by using temperature change amounts (standardized temperature change amounts) standardized for commonly processing casting conditions with different conditions of a casting speed, a steel type, and the mold powder 23 after the interpolation processing is performed to address casting at different widths and a failure of a thermometer to be described later. Note that, since the sensitivity coefficient may be changed by a change of the surface state of the mold 5 through operation, the sensitivity coefficient is preferably updated at an appropriate time between pieces of casting. Next, the determination unit 20 continuously detects temperatures T 1,1 to T m,n of the mold 5 by using the thermometers 8 1,1 to 8 m,n (Step S2). Next, the determination unit 20 executes, on the detected temperatures of the thermometers 8 1,1 to 8 m,n , interpolation processing on the temperatures of the mold 5 at center points of calculation cells 12 1,1 to 12 k,p obtained by equal division in accordance with the dimension of the cast strand 6 (e.g., width of cast strand 6 and thickness of cast strand 6) to be pulled out from the mold 5 (Step S3), the dimension of the cast strand 6 being input from an operator through an input apparatus (not illustrated) which is an input device such as a personal computer provided in the continuous casting machine 1. Next, the determination unit 20 performs mean bias removal on temperatures T' 1,1 to T' k,p of the mold 5 obtained by the interpolation processing. That is, for the temperatures T' 1,1 to T' k,p of the mold 5 obtained by the interpolation processing T' 1,1 to ΔT' k,p , the determination unit 20 determines average values for temperature change amounts ΔT' 1,1 to ΔT' 1,p of calculation cells 12 1,1 to 12 1,p and temperature change amounts ΔT' 2,1 to ΔT' 2,p and T' k,1 to T' k,p of calculation cells 12 2,1 to 12 2,p , the calculation cells 12 1,1 to 12 1,p being located at positions with the same distance from the upper end of the mold 5, and a standard deviation (of temperature change amounts) in a certain period for temperature change amounts ΔT' 1,1 to ΔT' k,p of calculation cells . Thereafter, the determination unit 20 determines standardized temperature change amounts obtained by dividing, by a standard deviation, the differences of the temperature change amounts ΔT' 1,1 to ΔT' 1,p of the calculation cells 12 1,1 to 12 1,p from the average value and the differences of the temperature change amounts ΔT' 2,1 to ΔT' 2,p of the calculation cells 12 2,1 to 12 2,p from the average value (Steps S4 to S6). Next, the determination unit 20 calculates deviation degrees by using sensitivity coefficients to the determined standardized temperature change amounts (Step S7).

[0025] Furthermore, a method using principal component analysis can be considered as one method of determining a sensitivity coefficient vector, which is an influence coefficient vector. Here, the sensitivity coefficient vector, which has as a component a sensitivity coefficient that is an influence coefficient, represents a direction indicating an average behavior of temperature change amounts of calculation cells obtained by the above-described interpolation processing for the thermometers 8 1,1 to 8 m,n at a normal operation time. Then, in a vector having, as a component, a difference from an average value, a component parallel to the direction of the sensitivity coefficient vector is a component of the average behavior, and a component in a direction orthogonal to the direction of the sensitivity coefficient vector is a component of a deviation degree from the average behavior. Furthermore, individual deviation degrees of calculation cells 12 indicate components of differences from the average values for individual calculation cells.

[0026] Next, when a calculated individual deviation degree (individual deviation degree) of the calculation cell 12 exceeds a threshold Y or a calculated total deviation degree (total deviation degree) of the calculation cells 12 exceeds a threshold X, the determination unit 20 determines breakout prediction (Step S8). When it is determined that a breakout is not predicted (No in Step S8), the determination unit 20 proceeds to Step S2.

[0027] In contrast, when it is determined that a breakout is predicted (Yes in Step S8), the determination unit 20 automatically decreases a casting speed to a predetermined speed (Step S9). As described above, when the determination unit 20 predicts a breakout, the casting speed is sufficiently decreased to form the solidified shell 10 having a sufficient thickness inside the mold 5 even at a location where sticking and foreign object entrapment have occurred. A breakout can thus be avoided. Thereafter, the determination unit 20 returns to a processing routine after lowering the casting speed to a predetermined value.

[0028] Next, a difference between a sensitivity coefficient used in the method of predicting a breakout according to the embodiment when a temperature is used and that when a temperature change amount is used will be described. In the sticking breakout, a temperature distribution generated in an extremely large range (approximately 100 to 500 [mm]) is obtained, and a temperature change amount itself is extremely large (20 to 50[°C]). Thus, when a deviation degree is determined from a sensitivity coefficient, a deviation degree is large and detection is easy. In the foreign object entrapment breakout, however, a mold temperature changes only in a width (100 [mm] or less) of the size of the foreign object 22, and a temperature change amount itself is also extremely small (20[°C] or less). Various temperature distributions are obtained also in a normal operation. If a small temperature change that occurs in foreign object entrapment is determined as abnormal, erroneous detection is permitted, which decreases superiority of this approach. In contrast, a large deviation degree can be set with respect to a rapid temperature change that does not occur under normal operation by determining a sensitivity coefficient by using a temperature change amount and obtaining a deviation degree. In a breakout to be detected this time, an abnormal state suddenly occurs without notice regardless of whether the breakout is a sticking breakout or a foreign object entrapment breakout. Expressing an abnormal temperature change amount with a deviation degree with a temperature change amount as a sensitivity coefficient is preferable to expressing an abnormal temperature distribution with a deviation degree with a temperature as a sensitivity coefficient.

[0029] Next, a standardized temperature change amount is used in calculation. A method of the standardization will be described. A change tends to be made overall when an operation condition changes, for example, when a casting speed increases or decreases, or when a molten steel temperature rises or falls. An overall temperature change accompanying a modification in an operation condition at the time of predicting a breakout may serve as a disturbance to deteriorate detection accuracy. Thus, it is necessary to exclude a mean value bias for use.

[0030] Examples of a method of removing a mean bias include a method of determining an average value T ave of all detected temperatures T 1,1 to T m,n detected by the thermometers 8 1,1 to 8 m,n and using the differences between the detected temperatures T 1,1 to T m,n and the average value T ave . Examples of another method of removing a mean bias include a method of determining an average value T i,ave of detected temperatures T i,1 to T i,n detected by the thermometers 8 i,1 to 8 i,n at positions with the same distance from the upper end of the mold 5 in the casting direction A and using the differences between the detected temperatures T i,1 to T i,n and the average value T i,ave for the thermometers 8 at the positions with the same distance.

[0031] Furthermore, in addition, the stability of a mold temperature may change depending on an operation condition such as the swing of the mold used for casting, casting speed, and steel type / mold powder 23 used. In addition, electromagnetic noise and the like may influence a thermometer used, which may be regarded as temperature fluctuation. When a breakout is predicted from a deviation degree by using temperature fluctuation or when the mold temperature becomes unstable due to a condition, erroneous detection is extremely highly likely to occur. It is thus necessary to handle the case. Processing of calculating a standard deviation of temperature change amounts in a certain period for the thermometers 8 and performing division from the temperature change amounts is performed as standardization. As a result, a standardized temperature change amount that enables certain threshold determination can be obtained while ignoring differences in the temperature change amounts, which are generated by differences in a condition in the processing.

[0032] Furthermore, examples of another method of determining a sensitivity coefficient vector, which is an influence coefficient vector, can include a method of experimentally determining how easily the temperature of the molten steel 2 is transmitted in the individual thermometers 8 1,1 to 8 m,n at the time when the overall temperature changes due to fluctuation of a molten-metal surface.

[0033] In contrast, as illustrated in FIG. 7, the detected temperatures of thermometers 8 i,j1 and 8 i,j2 at the occurrence of a sign of sticking or the like that leads to a breakout are distributed at positions away from a dashed line (line of 45 degrees in diagonally right direction in example in FIG. 7) indicating a direction of a sensitivity coefficient vector. This is because, when sticking that leads to a breakout occurs, a detected temperature T i,j1 of the thermometer 8 i,j1 close to the position of the fracture portion 11 of the solidified shell 10 decreases, and then a detected temperature T i,j1+1 and a detected temperature T i,j1-1 of a thermometer 8 i,j1+1 and a thermometer 8 i,j1-1 , which are located on both sides of the thermometer 8 i,j1 , decrease a little late.

[0034] From the above-described consideration, it can be seen that occurrence of a breakout can be determined by a degree with which temperature change amounts ΔT 1,1 to ΔT m,n of the thermometers 8 1,1 to 8 m,n deviate from the dashed line indicating a direction of a sensitivity coefficient vector. In other words, a component in a direction orthogonal to the sensitivity coefficient vector in a temperature change amount vector is calculated as a deviation degree. The temperature change amount vector is a vector having ΔT 1,1 to ΔT m,n calculated from the detected temperatures of the thermometers 8 1,1 to 8 m,n as components. Then, it can be seen that occurrence of a breakout can be determined based on the calculated deviation degree.

[0035] For example, in FIGS. 7 and 8, deviation degree components in temperature change amount vectors having, as components, temperature change amounts of the detected temperatures of the thermometers 8 i,j1 and 8 i,j2 are calculated. The deviation degree components are components in the direction orthogonal to the sensitivity coefficient vector. Then, occurrence of a breakout is determined based on the calculated deviation degree components. Note that, in FIGS. 7 and 8, the direction of the sensitivity coefficient vector is the same as a direction of a first principal component in a temperature change amount distribution under normal operation. The direction orthogonal to the direction of the sensitivity coefficient vector is the same as a direction of a second principal component in the temperature change amount distribution under normal operation.

[0036] If the detected temperatures T 1,1 to T m,n themselves are used to predict a breakout, however, erroneous detection may occur. That is, at an unsteady time when, for example, the casting width at the time when the molten steel 2 is poured into the mold 5, in other words, the width of the cast strand 6 pulled out from the lower end of the mold 5 is modified during operation, occurrence of a breakout may be erroneously predicted even though no sign that leads to a breakout is generated.

[0037] FIG. 9(a) illustrates the relation between detected temperatures T m1,n1 to T m1,n1+18 of thermometers 8 m1,n1 to 8 m1,n1+18 and temperatures T' m1,n1 to T' m1,n1+18 on which interpolation processing has been executed in a case of a wide width (casting width) of the cast strand 6 pulled out from the lower end of the mold 5. FIG. 9(b) illustrates the relation between the detected temperatures T m1,n1 to T m1,n1+18 of the thermometers 8 m1,n1 to 8 m1,n1+18 and the temperatures T' m1,n1 to T' m1,n1+18 on which interpolation processing has been executed in a case of a narrow width (casting width) of the cast strand 6 pulled out from the lower end of the mold 5. Note that, in FIGS. 9(a) and 9(b), the thermometers 8 m1,n1 to 8 m1,n1+18 are arranged at positions with the same distance from the upper end of the mold 5 in the casting direction A. Furthermore, the temperatures T' m1,n1 to T' m1,n1+18 are estimated temperatures of the mold 5 calculated by executing interpolation processing on the detected temperatures T m1,n1 to T m1,n1+18 of the thermometers 8 m1,n1 to 8 m1,n1+18 at center points of calculation cells 12 m1,n1 to 12 m1,n1+18 obtained by equal division in accordance with the width of the cast strand 6. Note that an approach of the interpolation processing will be described later.

[0038] When attention is paid to the detected temperatures T m1,n1 to T m1,n1+18 of the thermometers 8 m1,n1 to 8 m1,n1+18 in a case where the casting width is modified during casting and a change is made from a state in FIG. 9(a) to a state in FIG. 9(b), only detected temperatures T m1,n1+3 and T m1,n1+15 greatly change, and other detected temperatures do not prominently change. Thus, in the cases in FIGS. 9(a) and 9(b), if the detected temperatures T m1,n1 to T m1,n1+18 themselves are used to predict a breakout, a deviation from the sensitivity coefficient vector may occur to lead to erroneous detection that a sign that leads to a breakout is generated.

[0039] In contrast, when attention is paid to the temperatures T' m1,n1 to T' m1,n1+18 on which interpolation processing has been executed while keeping the number of calculation cells 12 (cell number) constant even when the dimension of the cast strand 6 is modified in a case where the casting width is modified during casting and a change is made from a state in FIG. 9(a) to a state in FIG. 9(b), the changes of the temperatures T' m1,n1 to T' m1,n1+18 are small. Thus, in the cases in FIGS. 9(a) and 9(b), a risk of erroneously detecting occurrence of a sign that leads to a breakout can be decreased by using the temperatures T' m1,n1 to T' m1,n1+18 on which interpolation processing has been executed for predicting the breakout.

[0040] Furthermore, in FIGS. 9(a) and 9(b), thermometers 8 m1,n1+7 , 8 m1,n1+11 , 8 m1,n1+12 , and 8 m1,n1+16 , which respectively detect detected temperatures T m1,n1+7 , T m1,n1+11 , T m1,n1+12 , and T m1,n1+16 , defectively detect the temperatures. Then, also when a thermometer 8 that defectively detects a temperature as described above is included, if the detected temperatures T m1,n1 to T m1,n1+18 themselves are used to predict a breakout, a deviation from the sensitivity coefficient vector may occur to lead to erroneous detection that a sign of occurrence of a breakout is generated. In contrast, in the temperatures T' m1,n1 to T' m1,n1+18 on which interpolation processing has been executed, even when a thermometer 8 that defectively detects a temperature is included, a risk of erroneously detecting occurrence of a sign that leads to a breakout can be decreased by using an estimated temperature of the mold 5 in a section in which a temperature is defectively detected.

[0041] Next, the approach of the interpolation processing will be described. FIG. 10 illustrates the positional relation between thermometers 8 i,1 to 8 i,j at positions with the same distance from the upper end of the mold 5 and calculation cells 12 i,1 to 12 i,j .

[0042] As illustrated in FIG. 10, the calculation cells 12 i,1 to 12 i,j are obtained by equally dividing, by a certain cell number, a section corresponding to the width of the cast strand 6 (section sandwiched by pair of short side cooling plates 5b in width direction of mold 5) of the long side cooling plate 5a with respect to the thermometers 8 i,1 to 8 i,j at the positions with the same distance from the upper end of the mold 5 in the long side cooling plate 5a of the mold 5. Then, detected temperatures detected by the thermometers 8 i,1 to 8 i,j are linearly interpolated, and estimated temperatures of the mold 5 (long side cooling plate 5a) at positions of the center points of the calculation cells 12 i,1 to 12 i,j are calculated. Note that, although the cell number of the calculation cells 12 for interpolation processing may be the same as or different from the number of the thermometers 8 in a vertical direction and a horizontal direction, the cell number is constant without depending on fluctuation in the casting width during casting.

[0043] The above-described interpolation processing can be applied to a case where a sensitivity coefficient vector is determined by using principal component analysis and a case where a deviation degree is calculated. In this case, the principal component analysis is performed by using temperatures on which interpolation processing has been performed instead of actual detected temperatures. Even when the cast strand width is modified, the same number of temperature vectors can be used, so that principal component analysis can be performed including data of different widths. This eliminates the need for determining different influence coefficients for each width. An influence coefficient vector can be determined including data of different cast strand widths. Then, the deviation degree can also be calculated by using the influence coefficient vector calculated based on temperatures obtained by performing interpolation processing on the detected temperatures. Therefore, a breakout can be predicted in different cast strand widths based on a standardized reference. Moreover, a risk of erroneously detecting occurrence of a sign that leads to a breakout can be reduced also in a case where the cast strand width is modified during casting.

[0044] Note that a temperature change amount is derived from the difference between a temperature detection value before a certain time and a value of this time. The temperature change amount can be simply derived by substituting the difference between detected temperatures of two thermometers (e.g., difference between 8 1,2 and 8 2,2 in mold thermometer arrangement diagram of FIG. 2) at the same location in the width direction and different locations in the casting direction.

[0045] Next, determination of breakout prediction will be described. FIG. 11(a) illustrates time-series changes of total deviation degrees in a case where sticking (sticking breakout) is confirmed. FIG. 11(b) illustrates a change of an individual deviation degree at each of the calculation cells 12 in a case where sticking (sticking breakout) is confirmed.

[0046] Furthermore, FIG. 12(a) illustrates time-series changes of total deviation degrees in a case where foreign object entrapment (foreign object entrapment breakout) is confirmed. FIG. 12(b) illustrates a change of an individual deviation degree at a position of each of the calculation cells 12 in a case where foreign object entrapment (foreign object entrapment breakout) is confirmed.

[0047] Furthermore, FIG. 13(a) illustrates a case where a result of slab inspection has no abnormality though total deviation degrees have peaks. FIG. 13(b) illustrates an individual deviation degree at each of the calculation cells 12 in a case where the result of slab inspection has no abnormality. Note that FIGS. 13(a) and 13(b) are considered to illustrate a case where noise is added to measured values of the thermometers 8 due to generation of electromagnetic noise near the thermometers.

[0048] Note that two graphs in each of FIGS. 11(a), 12(a), and 13(a) illustrate total deviation degrees calculated for an upper portion and a lower portion of the mold 5. Furthermore, in each of FIGS. 11, 12, and 13, an abnormality threshold is 150 of a total deviation degree and 20 of an individual deviation degree. The abnormality threshold is indicated by a dashed line in each figure.

[0049] In FIG. 11(a), a deviation degree rapidly rises at a certain time during operation. As described above, in sticking, temperature fluctuation occurs in a relatively large range. The large temperature fluctuation width indicates that a total deviation degree has a value clearly different from that of normal operation. Furthermore, as illustrated in FIG. 11(b), some calculation cells 12 of the calculation cells 12 have a large value of deviation degree.

[0050] Furthermore, as illustrated in FIG. 12(a), it can be seen that a total deviation degree in a case where foreign object entrapment is confirmed has a smaller peak than in FIG. 11(a). Also in FIG. 12(b), it can be seen that only two calculation cells 12 have peaks in deviation degrees.

[0051] Furthermore, as illustrated in FIG. 13(a), a total deviation degree has a peak. In contrast, as illustrated in FIG. 13(b), all the calculation cells 12 have low individual deviation degrees. Temperatures of almost all the thermometers 8 simultaneously change. It is thus estimated that the condition is caused by electromagnetic noise generated near and around the thermometers 8 and actual temperatures do not change.

[0052] Here, in order to detect sticking and foreign object entrapment, which are signs that lead to a breakout, and in order not to detect a case having no abnormality, a breakout is predicted under the following two conditions. In the first condition, when a total deviation degree exceeds a preset threshold X (second threshold), an abnormality is determined. Note that, even in this case, at least one or more points need to be equal to or more than the preset threshold Y. The second condition is whether or not there is, in a deviation degree, even one point having a value equal to or more than a preset threshold Z (first threshold).

[0053] FIG. 14 illustrates a result of breakout prediction using the method of predicting a breakout according to the embodiment and a conventional method. As illustrated in FIG. 14, in a conventional method (conventional example) based on Patent Literatures 1 and 2, a breakout is predicted by the behavior of a temperature buried inside a mold. The breakout can be detected only at a detection rate of approximately 50[%]. There are many erroneous detections. In contrast, as illustrated in FIG. 14, in a method (invention example) of predicting a breakout according to the embodiment, an erroneous detection rate was successfully and greatly lowered. Furthermore, even though an abnormality is not detected in the conventional method (conventional example), the abnormality was sometimes successfully detected in a case where a deviation degree is determined by using the method (invention example) of predicting a breakout according to the embodiment. Furthermore, in the method (invention example) of predicting a breakout according to the embodiment, an abnormality that has not led to a breakout was often found in an actual cast strand. The method (invention example) of predicting a breakout according to the embodiment enables prediction of a breakout that cannot be detected by the conventional method (comparative example).Industrial Applicability

[0054] The present invention can provide a method of predicting a breakout and a method of operating a continuous casting machine capable of detecting, with sufficient detection accuracy, a foreign object entrapment breakout having a small temperature change width and temperature fluctuation of a mold.Reference Signs List

[0055] 1CONTINUOUS CASTING MACHINE 2MOLTEN STEEL 3TUNDISH 4SUBMERGED ENTRY NOZZLE 5MOLD 6CAST STRAND 7STRAND SUPPORT ROLL 8THERMOMETER 10SOLIDIFIED SHELL 11FRACTURE PORTION 12CALCULATION CELL 18MENISCUS 20DETERMINATION UNIT 22FOREIGN OBJECT 23MOLD POWDER

Examples

Embodiment Construction

[0011]An embodiment of a method of predicting a breakout and a method of operating a continuous casting machine according to the present invention will be described below. Note that the present invention is not limited by the embodiment.

[0012]FIG. 1 is a schematic diagram illustrating a schematic configuration of a continuous casting machine 1 according to the embodiment. As illustrated in FIG. 1, the continuous casting machine 1 according to the embodiment includes a tundish 3, a mold 5, a plurality of strand support rolls 7, and a determination unit 20. Molten steel 2 is poured into the tundish 3. The mold 5 is made of copper, and cools the molten steel 2 poured from the tundish 3 via a submerged entry nozzle 4. The plurality of strand support rolls 7 conveys a semi-solidified cast strand 6 pulled out from the mold 5. The determination unit 20 determines a sign phenomenon of a breakout from a detected temperature of a thermometer 8 buried in the mold 5. Note that, although a therm...

Field

[0001] The present invention relates to a method of predicting a breakout and a method of operating a continuous casting machine.Background

[0002] Patent Literatures 1 and 2 disclose a method of predicting a breakout by detecting a characteristic behavior of temperature by use of a temperature sensor arranged in a mold. Furthermore, Patent Literature 3 discloses a method of calculating a deviation degree from a mold temperature during normal operation in which no breakout occurs and predicting a breakout based on the calculated deviation degree.Citation ListPatent Literature

[0003] Patent Literature 1: JP S63-47545 B Patent Literature 2: JP 2005-296979 A Patent Literature 3: Japanese Patent No. 6950860 SummaryTechnical Problem

[0004] Patterns of breakouts are roughly classified into two patterns of a sticking breakout and a foreign object entrapment breakout. In the sticking breakout, a cast strand is stuck onto a mold, and a relatively large temperature fluctuation often occurs. The foreign object entrapment breakout occurs when casting proceeds with a foreign object such as powder being attached on the surface of the cast strand. The foreign object entrapment breakout thus has a feature in that a change width and a change magnitude of a mold temperature are extremely small.

[0005] In the methods disclosed in Patent Literatures 1 and 2, a temporal change and a spatial distribution of mold temperatures are detected using a characteristic temperature at the time when a breakout actually occurs. The sticking breakout and the foreign object entrapment breakout are detected by different methods. The sticking breakout having a relatively large temperature change is determined by the temperature relation between temperatures of different casting heights (normally high temperature in upper casting direction and, at abnormal time, high temperature in lower casting direction). Since the foreign object entrapment breakout is difficult to detect because a temperature change is small, the foreign object entrapment breakout is thus determined based on the magnitude of temperature fluctuation and downward propagation of a temperature change. However, since operation is not performed with a constant distribution of mold temperatures and the temperature changes depending on the situation inside the mold, erroneous detection often occurs, and detection has been often omitted.

[0006] Furthermore, in the method disclosed in Patent Literature 3, a breakout is predicted based on a deviation degree from a mold temperature at a normal operation time. Thus, when a cast strand is stuck onto a mold, and a relatively large temperature fluctuation occurs as in the sticking breakout, the breakout can be sufficiently detected. In contrast, in a foreign object entrapment breakout that occurs when casting proceeds with a foreign object such as powder being attached on the surface of the cast strand, a change width and a change magnitude of a mold temperature are very small, and the breakout cannot be sufficiently detected.

[0007] The present invention has been made in view of the above-described problems, and an object thereof is to provide a method of predicting a breakout and a method of operating a continuous casting machine capable of detecting, with sufficient detection accuracy, a foreign object entrapment breakout in which a change width of a mold temperature and temperature fluctuation of a mold is small.Solution to Problem

[0008] To solve the problems and achieve an object, (1) there is provided a method of predicting a breakout, including: a step of receiving input of a dimension of a cast strand to be pulled out from a mold included in a continuous casting machine; a step of detecting temperatures of the mold with a plurality of thermometers buried in the mold; a step of executing interpolation processing in accordance with a dimension of the cast strand on detected temperatures detected by the plurality of thermometers; a step of calculating temperature change amounts by comparison with temperatures before a first period; a step of acquiring a standard deviation of temperature change amounts of the thermometers in a second period; a step of standardizing temperature changes by dividing the temperature change amounts by the standard deviation; a step of calculating a component in a direction orthogonal to an influence coefficient vector obtained by principal component analysis, as a deviation degree from a normal operation time when no breakout occurs, based on a temperature change amount calculated by executing the interpolation processing; and a step of predicting the breakout based on the deviation degree. (2) In the method of predicting a breakout according to the above-described invention (1), in the step of executing interpolation processing, a temperature is calculated by executing interpolation processing on detected temperatures of the plurality of thermometers at a center point of each of a plurality of calculation cells obtained by equal division in accordance with the dimension of the cast strand. (3) In the method of predicting a breakout according to the above-described invention (2), a number of the calculation cells is kept constant even when the dimension of the cast strand is modified. (4) In the method of predicting a breakout according to the above-described invention (2) or (3), in the step of calculating the deviation degree, an average value of temperatures of the plurality of calculation cells at positions with a same distance from an upper end of the mold in a casting direction of molten steel toward the mold is determined, differences of the temperatures of the plurality of calculation cells from the average value are determined, and the deviation degrees are calculated by using the influence coefficient vector. (5) In the method of predicting a breakout according to the above-described invention (4), in the step of predicting the breakout, when some of calculated individual deviation degrees of the calculation cells exceed a preset first threshold or a calculated total deviation degree of the calculation cells exceeds a preset second threshold, the breakout is predicted. (6) In the method of predicting a breakout according to the above-described invention of any one of (1) to (5), the influence coefficient vector is a sensitivity coefficient vector having, as a component, a sensitivity coefficient of each of a plurality of the temperature change amounts. (7) A method of operating a continuous casting machine according to the present invention, includes decreasing a casting speed when a breakout is predicted based on the method of predicting a breakout according to any one invention of (1) to (6). Advantageous Effects of Invention

[0009] A method of predicting a breakout and a method of operating a continuous casting machine according to the present invention have an effect of detecting, with sufficient detection accuracy, a foreign object entrapment breakout having a small temperature change width and temperature fluctuation of a mold.Brief Description of Drawings

[0010] [FIG. 1] FIG. 1 is a schematic diagram illustrating a schematic configuration of a continuous casting machine according to an embodiment. [FIG. 2] FIG. 2 is a perspective view illustrating a schematic configuration of a mold, in which thermometers are buried, of the continuous casting machine according to the embodiment. [FIG. 3] FIG. 3(a) illustrates a situation of molten steel and a solidified shell inside the mold in a sign phenomenon of a breakout. FIG. 3(b) illustrates a situation of a fracture portion of the solidified shell in the sign phenomenon of a breakout. [FIG. 4] FIG. 4(a) illustrates a temperature distribution of the mold at the moment when sticking has occurred. FIG. 4(b) illustrates a temperature distribution of the mold 10 seconds after the moment when sticking has occurred. [FIG. 5] FIG. 5 illustrates a foreign object entrapment breakout. [FIG. 6] FIG. 6 is a flowchart illustrating an example of a procedure of a method of predicting a breakout according to the embodiment. [FIG. 7] FIG. 7 illustrates a correlation between detected temperatures of thermometers under normal operation when no breakout occurs. [FIG. 8] FIG. 8 illustrates a correlation between detected temperatures of thermometers at the time when a sign such as sticking that leads to a breakout occurs. [FIG. 9] FIG. 9(a) illustrates the relation between detected temperatures of thermometers and temperatures on which interpolation processing has been executed in a case of a wide width of a cast strand pulled out from the lower end of the mold. FIG. 9(b) illustrates the relation between detected temperatures of thermometers and temperatures on which interpolation processing has been executed in a case of a narrow width of a cast strand pulled out from the lower end of the mold. [FIG. 10] FIG. 10 illustrates the positional relation between thermometers at positions with the same distance from the upper end of the mold and calculation cells. [FIG. 11] FIG. 11(a) illustrates time-series changes of deviation degrees in a case where sticking (sticking breakout) is confirmed. FIG. 11(b) illustrates a change of a deviation degree at each of the calculation cells in a case where sticking (sticking breakout) is confirmed. [FIG. 12] FIG. 12(a) illustrates time-series changes of deviation degrees in a case where foreign object entrapment (foreign object entrapment breakout) is confirmed. FIG. 12(b) illustrates a change of a deviation degree at a position of each of the calculation cells in a case where foreign object entrapment (foreign object entrapment breakout) is confirmed. [FIG. 13] FIG. 13(a) illustrates a case where a result of slab inspection has no abnormality though total deviation degrees have peaks. FIG. 13(b) illustrates a deviation degree in each of the calculation cells in a case where a result of slab inspection has no abnormality though total deviation degrees have peaks. [FIG. 14] FIG. 14 illustrates a result of breakout prediction using a method of predicting a breakout according to the embodiment and a conventional method. Description of Embodiments

[0011] An embodiment of a method of predicting a breakout and a method of operating a continuous casting machine according to the present invention will be described below. Note that the present invention is not limited by the embodiment.

[0012] FIG. 1 is a schematic diagram illustrating a schematic configuration of a continuous casting machine 1 according to the embodiment. As illustrated in FIG. 1, the continuous casting machine 1 according to the embodiment includes a tundish 3, a mold 5, a plurality of strand support rolls 7, and a determination unit 20. Molten steel 2 is poured into the tundish 3. The mold 5 is made of copper, and cools the molten steel 2 poured from the tundish 3 via a submerged entry nozzle 4. The plurality of strand support rolls 7 conveys a semi-solidified cast strand 6 pulled out from the mold 5. The determination unit 20 determines a sign phenomenon of a breakout from a detected temperature of a thermometer 8 buried in the mold 5. Note that, although a thermocouple is used as the thermometer 8 in the embodiment, this is not a limitation.

[0013] FIG. 2 is a perspective view illustrating a schematic configuration of the mold 5, in which thermometers 8 1,1 to 8 m,n are buried, of the continuous casting machine 1 according to the embodiment. As illustrated in FIG. 2, the mold 5 includes a pair of long side cooling plates 5a and a pair of short side cooling plates 5b, and has a substantially rectangular cylindrical shape penetrating in the vertical direction. Cooling water channels (not illustrated) are formed along inner wall surfaces inside the long side cooling plates 5a and the short side cooling plates 5b. The molten steel 2 is cooled by distributing cooling water through the cooling water channels.

[0014] Furthermore, the thermometers 8 1,1 to 8 m,n are buried inside a long side cooling plate 5a of the mold 5 from an outer wall surface of the long side cooling plate 5a at a predetermined depth. Note that, in the following description, the thermometers 8 1,1 to 8 m,n are also simply referred to as thermometers 8 unless particularly distinguished. FIG. 2 illustrates a configuration of three or more stages of the thermometers 8 1,1 to 8 m,n in a casting direction A. The thermometers 8 1,1 to 8 m,n are divided into a first stage of thermometers 8 1,1 to 8 1,n , a second stage of thermometers 8 2,1 to 8 2,n , and an nth stage of thermometers 8 m,1 to 8 m,n , which are buried on the same plane. In the embodiment, in the casting direction A, the molten steel 2 is poured from the tundish 3 into the mold 5 via the submerged entry nozzle 4. The casting direction A is the same as a direction in which the cast strand 6 is pulled out from the lower end of the mold 5.

[0015] Note that the arrangement of the thermometers 8 in FIG. 2 is an example for describing the present invention. The thermometers 8 are only required to 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 among the pair of long side cooling plates 5a and the pair of short side cooling plates 5b of the mold 5. Among these options, the thermometers are preferably arranged in all of the pair of long side cooling plates 5a and the pair of short side cooling plates 5b. Furthermore, the thermometers 8 can be arranged in the mold 5 in a multi-stage array having more than three stages or a single-stage array in the casting direction A.

[0016] Next, the sign phenomenon of a breakout will be described. Breakouts targeted in the present proposal are roughly classified into two types. One type of breakout is called a sticking breakout, and the other type of breakout is called a foreign object entrapment breakout. First, for the sticking breakout, FIG. 3(a) illustrates a situation of the molten steel 2 and a solidified shell 10 inside the mold 5 in a sign phenomenon of a sticking breakout. FIG. 3(b) illustrates a situation of a fracture portion 11 of the solidified shell 10 in the sign phenomenon of a sticking breakout.

[0017] As illustrated in FIGS. 3(a) and 3(b), in the sign phenomenon of a sticking breakout, sticking occurs inside the mold 5 for some reason, and the solidified shell 10 is stuck onto the mold 5. In contrast, since the cast strand 6 is pulled out from the lower end of the mold 5 in the same direction as the casting direction A in FIG. 3(b), the fracture portion 11 of the solidified shell 10 is generated immediately below the sticking. In the fracture portion 11 of the solidified shell 10, the mold 5 and the molten steel 2 are in contact with each other, which causes further sticking. While the above-described phenomenon is repeated, the fracture portion 11 of the solidified shell 10 moves downward, and the solidified shell 10 above the fracture portion 11 becomes thicker. Then, finally, when the fracture portion 11 passes through the lower end of the mold 5, the molten steel 2 leaks from the fracture portion 11, and a sticking breakout occurs.

[0018] Note that, in the fracture portion 11, the molten steel 2 and the mold 5 are in contact with each other, so that the temperature of the mold 5 locally rises. Thus, for example, as indicated by an arrow B in FIG. 3(b), when the fracture portion 11 moving downward passes through the arrangement positions of thermometers 8 m',1 to 8 m',n , the detected temperatures of the thermometers 8 m',1 to 8 m',n increase. Thereafter, since the solidified shell 10 above the fracture portion 11 is stuck by the mold 5 and continues to be cooled, the detected temperatures of the thermometers 8 m',1 to 8 m',n monotonously decrease. In contrast, since the fracture portion 11 propagates not only in the downward direction but in the lateral direction, the fracture portion 11 enlarges to have a V shape as illustrated in FIG. 3(b). Note that, when the fracture portion 11 of the solidified shell 10 is generated below the thermometers 8 m',1 to 8 m',n , the fracture portion 11 does not pass through the positions of the thermometers 8 m',1 to 8 m',n , so that only a decrease in the detected temperatures of the thermometers 8 m',1 to 8 m',n is observed.

[0019] FIG. 4(a) illustrates a temperature distribution of the mold 5 at the moment when sticking has occurred. FIG. 4(b) illustrates a temperature distribution of the mold 5 10 seconds after the moment when sticking has occurred. From the temperature distributions of the mold 5 in FIGS. 4(a) and 4(b), it can be read that the V-shaped high temperature portion propagates in the downward direction and the lateral direction.

[0020] Next, the foreign object entrapment breakout will be described. FIG. 5 illustrates the foreign object entrapment breakout. Note that, in FIG. 5, a reference numeral 18 denotes a meniscus (molten-metal surface) of the molten steel 2. As illustrated in FIG. 5(a), mold powder 23 is used as a lubricant to prevent the above-described sticking between the mold 5 and the cast strand 6 during casting. After being supplied from an upper portion of the mold 5 onto the molten-metal surface, the mold powder 23 receives heat from the molten steel 2, comes into a molten state, and flows into a gap between the mold 5 and the cast strand 6. The mold powder 23 serves as a lubricant to prevent sticking. As illustrated in FIG. 5(b), however, the mold powder 23 supplied to the molten-metal surface of the molten steel 2 may be pulled into the mold 5 as an unmelted / massive (state called powder bear) foreign object 22. In the case, thermal resistance that is not originally assumed is generated in the gap between the mold 5 and the cast strand 6. As a result, as illustrated in FIGS. 5(c) and 5(d), the solidified shell 10 reaches the lower end of the mold 5 without being sufficiently grown at a site where the foreign object 22 is entrapped. In the case, as illustrated in FIG. 5(e), it is considered that, after the foreign object 22 passes through the lower end of the mold 5 with a thin local shell thickness, the solidified shell 10 cannot withstand molten steel static pressure, the solidified shell 10 fractures, and the molten steel 2 flows out to the outside. In the case, a temperature tendency that greatly differs from the temperature distribution inside the mold at the time when sticking occurs is exhibited. First, since the entrapped foreign object 22 is considered to be approximately 100 [mm] or less, the mold temperature changes with only approximately the size of the foreign object 22 when the foreign object 22 passes through the inside of the mold 5. That is, in a peripheral portion through which the foreign object 22 has passed, the temperature after the foreign object 22 has passed does not change from a temperature before the generation. Furthermore, a temperature change amount is relatively smaller than that in a case where sticking causes a fracture. Thus, in the breakout pattern, detection is generally difficult, and many erroneous detections occur.

[0021] Therefore, the method of predicting a breakout according to the embodiment improves accuracy of predicting a breakout by using a temperature change amount of a difference of a thermometer installed inside the mold 5 from that before a certain period (before first period) in order to accurately detect breakouts in the above-described two patterns and standardizing the temperature change amount with a standard deviation in a certain period (second period) of the temperature change amount. The method of predicting a breakout according to the embodiment based on the above-described technical idea will be described in detail below.

[0022] The method of predicting a breakout according to the embodiment includes a step of receiving input of the dimension of the cast strand 6 to be pulled out from the mold 5 included in the continuous casting machine 1. Furthermore, the method of predicting a breakout according to the embodiment includes a step of detecting the temperatures of the mold 5 with the plurality of thermometers 8 buried in the mold 5. Furthermore, the method of predicting a breakout according to the embodiment includes a step of executing interpolation processing in accordance with the dimension of the cast strand 6 on the detected temperatures detected by the plurality of thermometers 8. Furthermore, the method of predicting a breakout according to the embodiment includes a step of calculating temperature change amounts by performing comparison with temperatures before a certain period. Furthermore, the method of predicting a breakout according to the embodiment includes a step of acquiring the standard deviation of temperature change amounts for the thermometers 8 in the certain period. Furthermore, the method of predicting a breakout according to the embodiment includes a step of standardizing temperature changes by dividing the temperature change amounts by the standard deviation. Furthermore, the method of predicting a breakout according to the embodiment includes a step of calculating a component in a direction orthogonal to an influence coefficient vector obtained by principal component analysis, as a deviation degree from a normal operation time when no breakout occurs, based on the temperature change amounts calculated by executing the interpolation processing. Furthermore, the method of predicting a breakout according to the embodiment includes a step of predicting a breakout based on the deviation degree.

[0023] FIG. 6 is a flowchart illustrating an example of a procedure of the method of predicting a breakout according to the embodiment. The determination unit 20 in FIG. 1 executes the method of predicting a breakout in the flowchart. Note that the determination unit 20 has at least functions of a device of executing interpolation processing, a device of standardizing temperature change amounts, a device of calculating a deviation degree, and a device of predicting a breakout in the present invention. Furthermore, details of the steps in the flowchart in FIG. 6 will be appropriately described later.

[0024] In the method of predicting a breakout according to the embodiment, the determination unit 20 preliminarily calculates a sensitivity coefficient for standardized temperature change amounts for the thermometers 8 1,1 to 8 m,n under normal operation (hereinafter, also referred to as normal operation) when no breakout has occurred (Step S1). Here, the sensitivity coefficient is calculated by using temperature change amounts (standardized temperature change amounts) standardized for commonly processing casting conditions with different conditions of a casting speed, a steel type, and the mold powder 23 after the interpolation processing is performed to address casting at different widths and a failure of a thermometer to be described later. Note that, since the sensitivity coefficient may be changed by a change of the surface state of the mold 5 through operation, the sensitivity coefficient is preferably updated at an appropriate time between pieces of casting. Next, the determination unit 20 continuously detects temperatures T 1,1 to T m,n of the mold 5 by using the thermometers 8 1,1 to 8 m,n (Step S2). Next, the determination unit 20 executes, on the detected temperatures of the thermometers 8 1,1 to 8 m,n , interpolation processing on the temperatures of the mold 5 at center points of calculation cells 12 1,1 to 12 k,p obtained by equal division in accordance with the dimension of the cast strand 6 (e.g., width of cast strand 6 and thickness of cast strand 6) to be pulled out from the mold 5 (Step S3), the dimension of the cast strand 6 being input from an operator through an input apparatus (not illustrated) which is an input device such as a personal computer provided in the continuous casting machine 1. Next, the determination unit 20 performs mean bias removal on temperatures T' 1,1 to T' k,p of the mold 5 obtained by the interpolation processing. That is, for the temperatures T' 1,1 to T' k,p of the mold 5 obtained by the interpolation processing T' 1,1 to ΔT' k,p , the determination unit 20 determines average values for temperature change amounts ΔT' 1,1 to ΔT' 1,p of calculation cells 12 1,1 to 12 1,p and temperature change amounts ΔT' 2,1 to ΔT' 2,p and T' k,1 to T' k,p of calculation cells 12 2,1 to 12 2,p , the calculation cells 12 1,1 to 12 1,p being located at positions with the same distance from the upper end of the mold 5, and a standard deviation (of temperature change amounts) in a certain period for temperature change amounts ΔT' 1,1 to ΔT' k,p of calculation cells . Thereafter, the determination unit 20 determines standardized temperature change amounts obtained by dividing, by a standard deviation, the differences of the temperature change amounts ΔT' 1,1 to ΔT' 1,p of the calculation cells 12 1,1 to 12 1,p from the average value and the differences of the temperature change amounts ΔT' 2,1 to ΔT' 2,p of the calculation cells 12 2,1 to 12 2,p from the average value (Steps S4 to S6). Next, the determination unit 20 calculates deviation degrees by using sensitivity coefficients to the determined standardized temperature change amounts (Step S7).

[0025] Furthermore, a method using principal component analysis can be considered as one method of determining a sensitivity coefficient vector, which is an influence coefficient vector. Here, the sensitivity coefficient vector, which has as a component a sensitivity coefficient that is an influence coefficient, represents a direction indicating an average behavior of temperature change amounts of calculation cells obtained by the above-described interpolation processing for the thermometers 8 1,1 to 8 m,n at a normal operation time. Then, in a vector having, as a component, a difference from an average value, a component parallel to the direction of the sensitivity coefficient vector is a component of the average behavior, and a component in a direction orthogonal to the direction of the sensitivity coefficient vector is a component of a deviation degree from the average behavior. Furthermore, individual deviation degrees of calculation cells 12 indicate components of differences from the average values for individual calculation cells.

[0026] Next, when a calculated individual deviation degree (individual deviation degree) of the calculation cell 12 exceeds a threshold Y or a calculated total deviation degree (total deviation degree) of the calculation cells 12 exceeds a threshold X, the determination unit 20 determines breakout prediction (Step S8). When it is determined that a breakout is not predicted (No in Step S8), the determination unit 20 proceeds to Step S2.

[0027] In contrast, when it is determined that a breakout is predicted (Yes in Step S8), the determination unit 20 automatically decreases a casting speed to a predetermined speed (Step S9). As described above, when the determination unit 20 predicts a breakout, the casting speed is sufficiently decreased to form the solidified shell 10 having a sufficient thickness inside the mold 5 even at a location where sticking and foreign object entrapment have occurred. A breakout can thus be avoided. Thereafter, the determination unit 20 returns to a processing routine after lowering the casting speed to a predetermined value.

[0028] Next, a difference between a sensitivity coefficient used in the method of predicting a breakout according to the embodiment when a temperature is used and that when a temperature change amount is used will be described. In the sticking breakout, a temperature distribution generated in an extremely large range (approximately 100 to 500 [mm]) is obtained, and a temperature change amount itself is extremely large (20 to 50[°C]). Thus, when a deviation degree is determined from a sensitivity coefficient, a deviation degree is large and detection is easy. In the foreign object entrapment breakout, however, a mold temperature changes only in a width (100 [mm] or less) of the size of the foreign object 22, and a temperature change amount itself is also extremely small (20[°C] or less). Various temperature distributions are obtained also in a normal operation. If a small temperature change that occurs in foreign object entrapment is determined as abnormal, erroneous detection is permitted, which decreases superiority of this approach. In contrast, a large deviation degree can be set with respect to a rapid temperature change that does not occur under normal operation by determining a sensitivity coefficient by using a temperature change amount and obtaining a deviation degree. In a breakout to be detected this time, an abnormal state suddenly occurs without notice regardless of whether the breakout is a sticking breakout or a foreign object entrapment breakout. Expressing an abnormal temperature change amount with a deviation degree with a temperature change amount as a sensitivity coefficient is preferable to expressing an abnormal temperature distribution with a deviation degree with a temperature as a sensitivity coefficient.

[0029] Next, a standardized temperature change amount is used in calculation. A method of the standardization will be described. A change tends to be made overall when an operation condition changes, for example, when a casting speed increases or decreases, or when a molten steel temperature rises or falls. An overall temperature change accompanying a modification in an operation condition at the time of predicting a breakout may serve as a disturbance to deteriorate detection accuracy. Thus, it is necessary to exclude a mean value bias for use.

[0030] Examples of a method of removing a mean bias include a method of determining an average value T ave of all detected temperatures T 1,1 to T m,n detected by the thermometers 8 1,1 to 8 m,n and using the differences between the detected temperatures T 1,1 to T m,n and the average value T ave . Examples of another method of removing a mean bias include a method of determining an average value T i,ave of detected temperatures T i,1 to T i,n detected by the thermometers 8 i,1 to 8 i,n at positions with the same distance from the upper end of the mold 5 in the casting direction A and using the differences between the detected temperatures T i,1 to T i,n and the average value T i,ave for the thermometers 8 at the positions with the same distance.

[0031] Furthermore, in addition, the stability of a mold temperature may change depending on an operation condition such as the swing of the mold used for casting, casting speed, and steel type / mold powder 23 used. In addition, electromagnetic noise and the like may influence a thermometer used, which may be regarded as temperature fluctuation. When a breakout is predicted from a deviation degree by using temperature fluctuation or when the mold temperature becomes unstable due to a condition, erroneous detection is extremely highly likely to occur. It is thus necessary to handle the case. Processing of calculating a standard deviation of temperature change amounts in a certain period for the thermometers 8 and performing division from the temperature change amounts is performed as standardization. As a result, a standardized temperature change amount that enables certain threshold determination can be obtained while ignoring differences in the temperature change amounts, which are generated by differences in a condition in the processing.

[0032] Furthermore, examples of another method of determining a sensitivity coefficient vector, which is an influence coefficient vector, can include a method of experimentally determining how easily the temperature of the molten steel 2 is transmitted in the individual thermometers 8 1,1 to 8 m,n at the time when the overall temperature changes due to fluctuation of a molten-metal surface.

[0033] In contrast, as illustrated in FIG. 7, the detected temperatures of thermometers 8 i,j1 and 8 i,j2 at the occurrence of a sign of sticking or the like that leads to a breakout are distributed at positions away from a dashed line (line of 45 degrees in diagonally right direction in example in FIG. 7) indicating a direction of a sensitivity coefficient vector. This is because, when sticking that leads to a breakout occurs, a detected temperature T i,j1 of the thermometer 8 i,j1 close to the position of the fracture portion 11 of the solidified shell 10 decreases, and then a detected temperature T i,j1+1 and a detected temperature T i,j1-1 of a thermometer 8 i,j1+1 and a thermometer 8 i,j1-1 , which are located on both sides of the thermometer 8 i,j1 , decrease a little late.

[0034] From the above-described consideration, it can be seen that occurrence of a breakout can be determined by a degree with which temperature change amounts ΔT 1,1 to ΔT m,n of the thermometers 8 1,1 to 8 m,n deviate from the dashed line indicating a direction of a sensitivity coefficient vector. In other words, a component in a direction orthogonal to the sensitivity coefficient vector in a temperature change amount vector is calculated as a deviation degree. The temperature change amount vector is a vector having ΔT 1,1 to ΔT m,n calculated from the detected temperatures of the thermometers 8 1,1 to 8 m,n as components. Then, it can be seen that occurrence of a breakout can be determined based on the calculated deviation degree.

[0035] For example, in FIGS. 7 and 8, deviation degree components in temperature change amount vectors having, as components, temperature change amounts of the detected temperatures of the thermometers 8 i,j1 and 8 i,j2 are calculated. The deviation degree components are components in the direction orthogonal to the sensitivity coefficient vector. Then, occurrence of a breakout is determined based on the calculated deviation degree components. Note that, in FIGS. 7 and 8, the direction of the sensitivity coefficient vector is the same as a direction of a first principal component in a temperature change amount distribution under normal operation. The direction orthogonal to the direction of the sensitivity coefficient vector is the same as a direction of a second principal component in the temperature change amount distribution under normal operation.

[0036] If the detected temperatures T 1,1 to T m,n themselves are used to predict a breakout, however, erroneous detection may occur. That is, at an unsteady time when, for example, the casting width at the time when the molten steel 2 is poured into the mold 5, in other words, the width of the cast strand 6 pulled out from the lower end of the mold 5 is modified during operation, occurrence of a breakout may be erroneously predicted even though no sign that leads to a breakout is generated.

[0037] FIG. 9(a) illustrates the relation between detected temperatures T m1,n1 to T m1,n1+18 of thermometers 8 m1,n1 to 8 m1,n1+18 and temperatures T' m1,n1 to T' m1,n1+18 on which interpolation processing has been executed in a case of a wide width (casting width) of the cast strand 6 pulled out from the lower end of the mold 5. FIG. 9(b) illustrates the relation between the detected temperatures T m1,n1 to T m1,n1+18 of the thermometers 8 m1,n1 to 8 m1,n1+18 and the temperatures T' m1,n1 to T' m1,n1+18 on which interpolation processing has been executed in a case of a narrow width (casting width) of the cast strand 6 pulled out from the lower end of the mold 5. Note that, in FIGS. 9(a) and 9(b), the thermometers 8 m1,n1 to 8 m1,n1+18 are arranged at positions with the same distance from the upper end of the mold 5 in the casting direction A. Furthermore, the temperatures T' m1,n1 to T' m1,n1+18 are estimated temperatures of the mold 5 calculated by executing interpolation processing on the detected temperatures T m1,n1 to T m1,n1+18 of the thermometers 8 m1,n1 to 8 m1,n1+18 at center points of calculation cells 12 m1,n1 to 12 m1,n1+18 obtained by equal division in accordance with the width of the cast strand 6. Note that an approach of the interpolation processing will be described later.

[0038] When attention is paid to the detected temperatures T m1,n1 to T m1,n1+18 of the thermometers 8 m1,n1 to 8 m1,n1+18 in a case where the casting width is modified during casting and a change is made from a state in FIG. 9(a) to a state in FIG. 9(b), only detected temperatures T m1,n1+3 and T m1,n1+15 greatly change, and other detected temperatures do not prominently change. Thus, in the cases in FIGS. 9(a) and 9(b), if the detected temperatures T m1,n1 to T m1,n1+18 themselves are used to predict a breakout, a deviation from the sensitivity coefficient vector may occur to lead to erroneous detection that a sign that leads to a breakout is generated.

[0039] In contrast, when attention is paid to the temperatures T' m1,n1 to T' m1,n1+18 on which interpolation processing has been executed while keeping the number of calculation cells 12 (cell number) constant even when the dimension of the cast strand 6 is modified in a case where the casting width is modified during casting and a change is made from a state in FIG. 9(a) to a state in FIG. 9(b), the changes of the temperatures T' m1,n1 to T' m1,n1+18 are small. Thus, in the cases in FIGS. 9(a) and 9(b), a risk of erroneously detecting occurrence of a sign that leads to a breakout can be decreased by using the temperatures T' m1,n1 to T' m1,n1+18 on which interpolation processing has been executed for predicting the breakout.

[0040] Furthermore, in FIGS. 9(a) and 9(b), thermometers 8 m1,n1+7 , 8 m1,n1+11 , 8 m1,n1+12 , and 8 m1,n1+16 , which respectively detect detected temperatures T m1,n1+7 , T m1,n1+11 , T m1,n1+12 , and T m1,n1+16 , defectively detect the temperatures. Then, also when a thermometer 8 that defectively detects a temperature as described above is included, if the detected temperatures T m1,n1 to T m1,n1+18 themselves are used to predict a breakout, a deviation from the sensitivity coefficient vector may occur to lead to erroneous detection that a sign of occurrence of a breakout is generated. In contrast, in the temperatures T' m1,n1 to T' m1,n1+18 on which interpolation processing has been executed, even when a thermometer 8 that defectively detects a temperature is included, a risk of erroneously detecting occurrence of a sign that leads to a breakout can be decreased by using an estimated temperature of the mold 5 in a section in which a temperature is defectively detected.

[0041] Next, the approach of the interpolation processing will be described. FIG. 10 illustrates the positional relation between thermometers 8 i,1 to 8 i,j at positions with the same distance from the upper end of the mold 5 and calculation cells 12 i,1 to 12 i,j .

[0042] As illustrated in FIG. 10, the calculation cells 12 i,1 to 12 i,j are obtained by equally dividing, by a certain cell number, a section corresponding to the width of the cast strand 6 (section sandwiched by pair of short side cooling plates 5b in width direction of mold 5) of the long side cooling plate 5a with respect to the thermometers 8 i,1 to 8 i,j at the positions with the same distance from the upper end of the mold 5 in the long side cooling plate 5a of the mold 5. Then, detected temperatures detected by the thermometers 8 i,1 to 8 i,j are linearly interpolated, and estimated temperatures of the mold 5 (long side cooling plate 5a) at positions of the center points of the calculation cells 12 i,1 to 12 i,j are calculated. Note that, although the cell number of the calculation cells 12 for interpolation processing may be the same as or different from the number of the thermometers 8 in a vertical direction and a horizontal direction, the cell number is constant without depending on fluctuation in the casting width during casting.

[0043] The above-described interpolation processing can be applied to a case where a sensitivity coefficient vector is determined by using principal component analysis and a case where a deviation degree is calculated. In this case, the principal component analysis is performed by using temperatures on which interpolation processing has been performed instead of actual detected temperatures. Even when the cast strand width is modified, the same number of temperature vectors can be used, so that principal component analysis can be performed including data of different widths. This eliminates the need for determining different influence coefficients for each width. An influence coefficient vector can be determined including data of different cast strand widths. Then, the deviation degree can also be calculated by using the influence coefficient vector calculated based on temperatures obtained by performing interpolation processing on the detected temperatures. Therefore, a breakout can be predicted in different cast strand widths based on a standardized reference. Moreover, a risk of erroneously detecting occurrence of a sign that leads to a breakout can be reduced also in a case where the cast strand width is modified during casting.

[0044] Note that a temperature change amount is derived from the difference between a temperature detection value before a certain time and a value of this time. The temperature change amount can be simply derived by substituting the difference between detected temperatures of two thermometers (e.g., difference between 8 1,2 and 8 2,2 in mold thermometer arrangement diagram of FIG. 2) at the same location in the width direction and different locations in the casting direction.

[0045] Next, determination of breakout prediction will be described. FIG. 11(a) illustrates time-series changes of total deviation degrees in a case where sticking (sticking breakout) is confirmed. FIG. 11(b) illustrates a change of an individual deviation degree at each of the calculation cells 12 in a case where sticking (sticking breakout) is confirmed.

[0046] Furthermore, FIG. 12(a) illustrates time-series changes of total deviation degrees in a case where foreign object entrapment (foreign object entrapment breakout) is confirmed. FIG. 12(b) illustrates a change of an individual deviation degree at a position of each of the calculation cells 12 in a case where foreign object entrapment (foreign object entrapment breakout) is confirmed.

[0047] Furthermore, FIG. 13(a) illustrates a case where a result of slab inspection has no abnormality though total deviation degrees have peaks. FIG. 13(b) illustrates an individual deviation degree at each of the calculation cells 12 in a case where the result of slab inspection has no abnormality. Note that FIGS. 13(a) and 13(b) are considered to illustrate a case where noise is added to measured values of the thermometers 8 due to generation of electromagnetic noise near the thermometers.

[0048] Note that two graphs in each of FIGS. 11(a), 12(a), and 13(a) illustrate total deviation degrees calculated for an upper portion and a lower portion of the mold 5. Furthermore, in each of FIGS. 11, 12, and 13, an abnormality threshold is 150 of a total deviation degree and 20 of an individual deviation degree. The abnormality threshold is indicated by a dashed line in each figure.

[0049] In FIG. 11(a), a deviation degree rapidly rises at a certain time during operation. As described above, in sticking, temperature fluctuation occurs in a relatively large range. The large temperature fluctuation width indicates that a total deviation degree has a value clearly different from that of normal operation. Furthermore, as illustrated in FIG. 11(b), some calculation cells 12 of the calculation cells 12 have a large value of deviation degree.

[0050] Furthermore, as illustrated in FIG. 12(a), it can be seen that a total deviation degree in a case where foreign object entrapment is confirmed has a smaller peak than in FIG. 11(a). Also in FIG. 12(b), it can be seen that only two calculation cells 12 have peaks in deviation degrees.

[0051] Furthermore, as illustrated in FIG. 13(a), a total deviation degree has a peak. In contrast, as illustrated in FIG. 13(b), all the calculation cells 12 have low individual deviation degrees. Temperatures of almost all the thermometers 8 simultaneously change. It is thus estimated that the condition is caused by electromagnetic noise generated near and around the thermometers 8 and actual temperatures do not change.

[0052] Here, in order to detect sticking and foreign object entrapment, which are signs that lead to a breakout, and in order not to detect a case having no abnormality, a breakout is predicted under the following two conditions. In the first condition, when a total deviation degree exceeds a preset threshold X (second threshold), an abnormality is determined. Note that, even in this case, at least one or more points need to be equal to or more than the preset threshold Y. The second condition is whether or not there is, in a deviation degree, even one point having a value equal to or more than a preset threshold Z (first threshold).

[0053] FIG. 14 illustrates a result of breakout prediction using the method of predicting a breakout according to the embodiment and a conventional method. As illustrated in FIG. 14, in a conventional method (conventional example) based on Patent Literatures 1 and 2, a breakout is predicted by the behavior of a temperature buried inside a mold. The breakout can be detected only at a detection rate of approximately 50[%]. There are many erroneous detections. In contrast, as illustrated in FIG. 14, in a method (invention example) of predicting a breakout according to the embodiment, an erroneous detection rate was successfully and greatly lowered. Furthermore, even though an abnormality is not detected in the conventional method (conventional example), the abnormality was sometimes successfully detected in a case where a deviation degree is determined by using the method (invention example) of predicting a breakout according to the embodiment. Furthermore, in the method (invention example) of predicting a breakout according to the embodiment, an abnormality that has not led to a breakout was often found in an actual cast strand. The method (invention example) of predicting a breakout according to the embodiment enables prediction of a breakout that cannot be detected by the conventional method (comparative example).Industrial Applicability

[0054] The present invention can provide a method of predicting a breakout and a method of operating a continuous casting machine capable of detecting, with sufficient detection accuracy, a foreign object entrapment breakout having a small temperature change width and temperature fluctuation of a mold.Reference Signs List

[0055] 1CONTINUOUS CASTING MACHINE 2MOLTEN STEEL 3TUNDISH 4SUBMERGED ENTRY NOZZLE 5MOLD 6CAST STRAND 7STRAND SUPPORT ROLL 8THERMOMETER 10SOLIDIFIED SHELL 11FRACTURE PORTION 12CALCULATION CELL 18MENISCUS 20DETERMINATION UNIT 22FOREIGN OBJECT 23MOLD POWDER

Examples

Embodiment Construction

[0011]An embodiment of a method of predicting a breakout and a method of operating a continuous casting machine according to the present invention will be described below. Note that the present invention is not limited by the embodiment.

[0012]FIG. 1 is a schematic diagram illustrating a schematic configuration of a continuous casting machine 1 according to the embodiment. As illustrated in FIG. 1, the continuous casting machine 1 according to the embodiment includes a tundish 3, a mold 5, a plurality of strand support rolls 7, and a determination unit 20. Molten steel 2 is poured into the tundish 3. The mold 5 is made of copper, and cools the molten steel 2 poured from the tundish 3 via a submerged entry nozzle 4. The plurality of strand support rolls 7 conveys a semi-solidified cast strand 6 pulled out from the mold 5. The determination unit 20 determines a sign phenomenon of a breakout from a detected temperature of a thermometer 8 buried in the mold 5. Note that, although a therm...

Claims

1. A method of predicting a breakout, comprising: a step of receiving input of a dimension of a cast strand to be pulled out from a mold included in a continuous casting machine; a step of detecting temperatures of the mold with a plurality of thermometers buried in the mold; a step of executing interpolation processing in accordance with a dimension of the cast strand on detected temperatures detected by the plurality of thermometers; a step of calculating temperature change amounts by comparison with temperatures before a first period; a step of acquiring a standard deviation of temperature change amounts of the thermometers in a second period; a step of standardizing temperature changes by dividing the temperature change amounts by the standard deviation; a step of calculating a component in a direction orthogonal to an influence coefficient vector obtained by principal component analysis, as a deviation degree from a normal operation time when no breakout occurs, based on a temperature change amount calculated by executing the interpolation processing; and a step of predicting the breakout based on the deviation degree.

2. The method of predicting a breakout, according to claim 1, wherein, in the step of executing interpolation processing, a temperature is calculated by executing interpolation processing on detected temperatures of the plurality of thermometers at a center point of each of a plurality of calculation cells obtained by equal division in accordance with the dimension of the cast strand.

3. The method of predicting a breakout, according to claim 2, wherein a number of the calculation cells is kept constant even when the dimension of the cast strand is modified.

4. The method of predicting a breakout, according to claim 2, wherein, in the step of calculating the deviation degree, an average value of temperatures of the plurality of calculation cells at positions with a same distance from an upper end of the mold in a casting direction of molten steel toward the mold is determined, differences of the temperatures of the plurality of calculation cells from the average value are determined, and the deviation degrees are calculated by using the influence coefficient vector.

5. The method of predicting a breakout, according to claim 3, wherein, in the step of calculating the deviation degree, an average value of temperatures of the plurality of calculation cells at positions with a same distance from an upper end of the mold in a casting direction of molten steel toward the mold is determined, differences of the temperatures of the plurality of calculation cells from the average value are determined, and the deviation degrees are calculated by using the influence coefficient vector.

6. The method of predicting a breakout, according to claim 4, wherein, in the step of predicting the breakout, when some of calculated individual deviation degrees of the calculation cells exceed a preset first threshold or a calculated total deviation degree of the calculation cells exceeds a preset second threshold, the breakout is predicted.

7. The method of predicting a breakout, according to claim 5, wherein, in the step of predicting the breakout, when some of calculated individual deviation degrees of the calculation cells exceed a preset first threshold or a calculated total deviation degree of the calculation cells exceeds a preset second threshold, the breakout is predicted.

8. The method of predicting a breakout, according to any one of claims 1 to 7, wherein the influence coefficient vector is a sensitivity coefficient vector having, as a component, a sensitivity coefficient of each of a plurality of the temperature change amounts.

9. A method of operating a continuous casting machine, comprising decreasing a casting speed when a breakout is predicted based on the method of predicting a breakout according to any one of claims 1 to 7.

10. A method of operating a continuous casting machine, comprising decreasing a casting speed when a breakout is predicted based on the method of predicting a breakout according to claim 8.

Images