JP7632366B2 - Method for generating prediction model for coiled steel sheet, method for cooling coiled steel sheet, and method for manufacturing coiled steel sheet - Google Patents

Description

Embodiments of the present invention relate to a method for generating a predictive model for coiled steel sheets, a method for cooling coiled steel sheets, a method for manufacturing coiled steel sheets, and cooling equipment for coiled steel sheets.

When manufacturing coiled steel sheet, hot rolling or cold rolling is carried out. For example, when manufacturing hot-rolled steel sheet, a slab is first heated to a predetermined temperature in a heating furnace, and the heated slab is rolled in a rough rolling mill to form a rough bar. This rough bar is then passed through a continuous hot finishing rolling mill consisting of multiple rolling stands to form a hot-rolled steel sheet of a predetermined thickness.

The hot-rolled steel sheet that leaves the continuous hot finishing rolling mill is cooled with cooling water supplied from above and below the cooling device installed on the run-out table, and then wound up by a winding machine to become a coiled hot-rolled steel sheet. It is then stored in this state.

On the other hand, in the case of cold-rolled steel sheets, the hot-rolled steel sheets coming out of the continuous hot finishing mill go through various processes such as pickling, annealing, and cold rolling to become cold-rolled steel sheets. The cold-rolled steel sheets are then wound up on a winding machine to become coiled cold-rolled steel sheets. Such coiled cold-rolled steel sheets are also stored in this state.

In this way, various steel sheets are wound up into coils for storage. Incidentally, the temperature of various steel sheets in this wound state is, for example, 500-650°C in the case of coiled hot-rolled steel sheets, and below this temperature in the case of coiled cold-rolled steel sheets. However, since they cannot be shipped or passed on to the next process in this temperature state, they are cooled to room temperature in the coil storage area where these coiled steel sheets are kept (hereinafter, such storage areas will be referred to as "cooling beds" as appropriate) before being transported and shipped.

A large amount of coiled steel plates are placed on the cooling bed and cooled. This means that other coiled steel plates at similarly high temperatures are placed around the hot coiled steel plates. In this case, the temperature around each coiled steel plate is therefore high. This reduces the cooling efficiency of the coiled steel plates, and it may take, for example, 3 to 5 days to complete the cooling of the coiled steel plates.

As a result, the area required to store the coiled steel sheets as a cooling bed must be large. In addition, because it takes time to cool, the time it takes to ship the coiled steel sheets increases. Furthermore, the time it takes to cool leads to problems such as a lack of space to store the coils in the cooling bed and an increase in coil inventory.

In order to shorten the cooling time of coiled steel sheets, various methods have been proposed for spraying cooling water onto the coiled steel sheets. However, when cooling water is sprayed onto the coiled steel sheets when the surface temperature of the coiled steel sheets is below 100°C, such as in the case of coiled cold-rolled steel sheets, the cooling water does not evaporate easily. As a result, the surface of the coiled steel sheets becomes wet with water, and if left as is, rust may form on the surface.

If rust occurs on the coiled steel plate, it will impair its appearance and the coiled steel plate will not be able to be shipped as a product. Even if cooling water is not sprayed, condensation may form on the coiled steel plate, especially during the cold seasons from winter to early spring, which may lead to rust.

The following Patent Documents 1 to 3 disclose various methods for cooling steel sheets. For example, the technology described in Patent Document 1 proposes a water-cooling method for hot-rolled coils that measures the temperature and humidity around the coil and controls the amount of cooling water sprayed to cool the coil, in order to prevent the surface of the hot-rolled steel sheet that constitutes the coil from getting wet.

Meanwhile, the technology described in Patent Document 2 also proposes a method of cooling hot-rolled coils in which cooling nozzles are placed on both end faces of each coil in the width direction, and cooling is performed from both sides using the respective cooling nozzles.

Furthermore, as in the technology described in Patent Document 3, a method for cooling a coiled hot-rolled steel sheet has also been proposed in which cooling is performed by intermittently spraying cooling mist onto both widthwise ends of the coiled steel sheet after it has been wound by a winding machine installed in a hot rolling line.

Japanese Unexamined Patent Publication No. 57-134207 Japanese Patent Application Publication No. 5-177240 JP 2013-188753 A

However, the water-cooling method for hot-rolled coils described in Patent Document 1 has the problem that cooling water is sprayed onto the coils from the ceiling of the coil storage area, which is far away from the coils, resulting in low coil cooling capacity and inefficient coil cooling.

In addition, the water-cooling method for hot-rolled coils described in Patent Document 2 uses only one cooling nozzle on each of the widthwise end faces of a single hot-rolled coil, so the minute droplets coming out of the cooling nozzles (hereinafter, such "minute droplets" will be referred to as "mist" as appropriate) do not spread sufficiently to both widthwise end faces of the hot-rolled coil. In other words, since the cooling area is small, it is difficult to cool the entire hot-rolled coil. As a result, the water-cooling method for hot-rolled coils described in Patent Document 2 has the problem of reduced coil cooling efficiency.

Here, in order to improve the cooling capacity for the coiled steel plate, it is important to increase the mist spray flow rate per unit area. However, if the spray flow rate is increased, the mist becomes difficult to evaporate due to factors such as an increase in humidity around the coiled steel plate and a decrease in the temperature of the coil surface.

As a result, the mist that does not evaporate is likely to adhere to the coil. In particular, when the surface temperature of the coiled steel sheet is below 100°C, the mist that adheres to the coil is unlikely to evaporate, and the coil surface is likely to remain wet, which may induce rust.

To prevent the coil surface from getting wet during water cooling, it is believed that using mist with a small particle size is effective. This is because mist with a smaller particle size requires less heat to evaporate, making it easier to evaporate.

On the other hand, if the particle size of the mist is reduced, it is possible that the mist will be dispersed by the influence of the wind around the coiled steel sheet, and the mist will be more likely to miss the coiled steel sheet that is the target for cooling. In addition, the spray distance from the mist nozzle to the coil will be shortened, which raises the problem that the sprayed mist may not even reach the coiled steel sheet in the first place.

The cooling method for coiled hot-rolled steel sheet described in Patent Document 3 involves constantly spraying cooling mist from a mist nozzle onto the coiled steel sheet, and also using an oscillating blower to entrain the cooling mist in an air current and intermittently blow it onto the widthwise ends of the coiled steel sheet.

However, even with this method, there is a possibility that the coiled hot-rolled steel sheet may not be cooled properly due to the problems mentioned above, and it may even lead to an increase in the humidity in the surrounding area other than the coiled steel sheet being cooled. In particular, condensation occurs in the coil once cooling is complete using the above-mentioned method for cooling the coiled hot-rolled steel sheet, increasing the risk of rust occurring in the coil.

Therefore, in order to cool the coiled steel plate without causing rust, it is necessary to properly manage the temperature, humidity, etc., and it is also necessary to control the cooling device in consideration of cooling the coiled steel plate more quickly. To perform this control, more appropriate control can be achieved if there is a predictive model that can predict, for example, the time when the coiled steel plate will be cooled and the state in which condensation will occur.

The present invention has been made with a focus on the above-mentioned points, and aims to provide a method for generating a predictive model for coiled steel plate, a method for cooling coiled steel plate, a method for manufacturing coiled steel plate, and cooling equipment for coiled steel plate that can simultaneously suppress wetting of the surface of the coiled steel plate and improve the cooling capacity of the coiled steel plate, prevent rust from forming on the coiled steel plate, and enable the coiled steel plate to be cooled with high efficiency.

A method for generating a predictive model for a coiled steel plate in an embodiment of the present invention includes the steps of acquiring component information of the coiled steel plate, operational parameters of the coiled steel plate, and performance data in the manufacture of the coiled steel plate, and generating by machine learning a predictive model that predicts the occurrence of condensation on the coiled steel plate on a cooling bed that is used when controlling at least the cooling mist generator and the cooling fan, which are arranged at intervals such that the cooling mist sprayed from the cooling mist generator can be evaporated to draw in cooled air and blow it to the coiled steel plate, using the component information and operational parameters as input data and the performance data as training data.

In the method for generating a predictive model for coiled steel plate in an embodiment of the present invention, the operational parameters are the dimensions and temperature of the coiled steel plate, the flow rate of the cooling medium used during cooling, and the flow rate of the cooling medium around the coiled steel plate.

In addition, in the method for generating a predictive model for coiled steel plate in an embodiment of the present invention, information obtained by measurement using a measuring device that detects the state of both or either of the coiled steel plate and the cooling bed is acquired as training data.

A method for cooling a coiled steel sheet in an embodiment of the present invention includes the steps of: acquiring component information of the coiled steel sheet, operation parameters of the coiled steel sheet, and performance data in the manufacture of the coiled steel sheet in order to predict the cooling completion time of the coiled steel sheet on the cooling bed or the occurrence of condensation; setting control contents for the cooling process of the coiled steel sheet using a prediction model generated using the component information and operation parameters as input data and the performance data as teaching data; and controlling the cooling equipment based on the set control contents. The cooling equipment is composed of at least a cooling mist generator and cooling fans arranged at intervals sufficient to allow the cooling mist sprayed from the cooling mist generator to evaporate and to prevent the cooling mist from diffusing and not being sucked in. The step of controlling the cooling equipment further includes the steps of spraying cooling mist from the cooling mist generator based on the set control contents, and sucking in air cooled by the evaporation of the cooling mist sprayed by the cooling mist generator and blowing it to the coiled steel sheet placed on the cooling bed by the cooling fan.

Furthermore, in the method for cooling a coiled steel plate in an embodiment of the present invention, after the step of the cooling fan blowing air onto the coiled steel plate, a step of exhausting the air around the coiled steel plate, including the gas blown onto the coiled steel plate by the cooling fan, outside the cooling bed based on the control content is provided.

The coiled steel sheet manufacturing method according to an embodiment of the present invention includes a step of acquiring component information of the coiled steel sheet, operating parameters of the coiled steel sheet, and performance data in the manufacture of the coiled steel sheet in order to predict the cooling completion time of the coiled steel sheet on the cooling bed or the occurrence of condensation, a step of setting the control content of the cooling process of the coiled steel sheet using a prediction model generated using the component information and operating parameters as input data and the performance data as training data, and a step of controlling the cooling equipment based on the set control content.

The coiled steel sheet cooling method, coiled steel sheet manufacturing method, and coiled steel sheet cooling equipment according to the embodiments of the present invention make it possible to simultaneously suppress wetting of the surface of the coiled steel sheet and improve the cooling capacity of the coiled steel sheet, preventing rust from forming on the coiled steel sheet and enabling the coiled steel sheet to be cooled with high efficiency.

FIG. 2 is a perspective view showing an entire building in which the coiled steel plate is cooled in the embodiment of the present invention. FIG. 2 is an explanatory front view of a building in which the coiled steel plate is cooled in an embodiment of the present invention. 2 is a block diagram showing an internal configuration of a control device according to an embodiment of the present invention. FIG. 4 is a flowchart showing a flow of machine learning performed by a control device in an embodiment of the present invention to generate a model for predicting the cooling completion time of a coiled steel sheet or the occurrence of condensation. 5 is a flowchart showing a flow in which a control device controls cooling equipment using a prediction model generated when predicting the cooling completion time of a coiled steel sheet or the occurrence of condensation in an embodiment of the present invention.

Below, an embodiment of the present invention will be described in detail with reference to the drawings. Note that the contents described below are not limited to the coiled steel sheet to be cooled, and may be either a hot-rolled steel sheet or a cold-rolled steel sheet. Therefore, in the following description, no particular distinction will be made and the term "coiled steel sheet" will be used.

The method of generating a predictive model for coiled steel sheets and the method of cooling coiled steel sheets in the embodiments of the present invention described below are used when coiled steel sheets S are formed by winding them up on a winding machine attached to or installed in hot rolling lines, cold rolling lines, and other pickling or annealing equipment, and are placed on a cooling bed to be cooled.

Fig. 1 is a perspective view showing an entire building B in which a coiled steel sheet S is cooled in an embodiment of the present invention. In Fig. 1, in order to clarify the inside of the building B, the building B itself is shown by a dashed line _.

On the other hand, FIG. 2 is an explanatory diagram showing the front view of the building B in which the coiled steel plate S is cooled in an embodiment of the present invention. Note that in FIG. 2, unlike FIG. 1, the building B is shown in solid lines. Also, the flow of air sent from a cooling fan, which will be described later, toward the coiled steel plate S is shown in dashed lines.

In addition, in the embodiment of the present invention, building B is formed in a rectangular shape as shown in Figures 1 and 2, but the structure of building B may be any type and is not limited to the structure shown in Figure 1, etc.

As shown in FIG. 1, the building B is composed of a cooling bed CF on which the coiled steel sheet S to be cooled is placed, walls surrounding the cooling bed CF on all four sides, and a ceiling R located opposite the cooling bed CF. Note that in FIG. 1 and other figures, structures that are not necessary for explaining the method for cooling the coiled steel sheet S in the embodiment of the present invention are omitted.

In addition, in FIG. 1 etc., the entire floor surface of building B is used as a cooling bed CF for cooling the coiled steel plate S. However, the extent to which an area of building B is used as a cooling bed CF can be set arbitrarily. Furthermore, building B itself does not have to be a building dedicated to cooling the coiled steel plate S.

Coiled steel plates S are placed on the cooling bed CF. This is placed inside the building B in order to cool the coiled steel plates S to the desired temperature. In FIG. 1, three rows of five coiled steel plates S are placed on the cooling bed CF. However, the number of coiled steel plates S placed on the cooling bed CF can be set arbitrarily depending on the operation.

When multiple coiled steel sheets S are arranged between cooling equipment C, such as the cooling mist generator 1 and cooling fan 2 described below, it is preferable to place the coiled steel sheets S, for example, 1 to 2 m apart.

The cooling bed CF shown in Figure 1 is shown placed so that the direction of the rotation axis when winding the coiled steel sheet S, i.e., the central axis of the coil, is in the left-right direction in the drawing. On the other hand, the cooling bed CF shown in Figure 2 is shown so that the central axis is in a direction perpendicular to the left-right direction in the drawing. In other words, the coiled steel sheet S may be placed in either direction on the cooling bed CF.

For the purpose of drawing the drawings, the size of each coiled steel sheet S placed on the cooling bed CF is the same, but the size of the coiled steel sheet S to be cooled may be any size. As described above, the coiled steel sheet S may be a coiled hot-rolled steel sheet or a coiled cold-rolled steel sheet.

In FIG. 1, cooling equipment C for cooling the coiled steel plate S is provided at a position parallel to the central axis of the coiled steel plate S, i.e., on both sides of the row in which the coiled steel plate S is arranged. The cooling equipment C is composed of at least a cooling mist generator 1 and a cooling fan 2.

First, the cooling mist generator 1 generates cooling mist, which is the cooling medium used to cool the coiled steel sheet S. Any liquid may be used as the cooling mist, but water is preferably used because of its workability in the cooling bed CF, cost, ease of availability, etc.

The method for generating the cooling mist in the cooling mist generator 1 can be either a one-fluid method or a two-fluid method. For example, when using a two-fluid method, the volume ratio of the gas that makes up the cooling mist to the cooling liquid (i.e., the "air-water ratio") can be set in the range of 10 to 1000, taking into account certain conditions.

The cooling fan 2 sucks in air (cooling medium) that has been cooled by the evaporation of the cooling mist sprayed from the cooling mist generator 1, as described below, and blows it onto the coiled steel plate S placed on the cooling bed CF. Therefore, the cooling fan 2 is positioned in front of the cooling mist generator 1, i.e., downwind, so that the wind from the cooling fan 2 is sent to the coiled steel plate S.

The cooling fan 2 is installed at a distance of, for example, about 1 m from the coiled steel plate S. If the cooling fan 2 and the coiled steel plate S are placed too close to each other, there is a risk that the cooling fan 2 will be damaged by the heat of the coiled steel plate S that is being cooled.

On the other hand, the distance between the cooling fan 2 and the cooling mist generator 1 is preferably large enough to allow the cooling mist to evaporate and avoid the cooling mist diffusing and not being sucked in by the cooling fan 2, approximately 1 to 2 m being a guideline.

Furthermore, the cooling mist sprayed from the cooling mist generator 1 is designed to evaporate before being sucked into the cooling fan 2. In other words, the cooling fan 2 blows a gas consisting of air cooled by the evaporation of the cooling mist and the surrounding air or gas onto the coiled steel plate S, which is the object to be cooled.

This is because if the cooling mist sprayed from the cooling mist generator 1 is sent to the coiled steel plate S via the cooling fan 2 without evaporating, the cooling mist adhering to the coiled steel plate S increases the risk of rust occurring, especially at low temperatures.

Here, heat removal from the coiled steel plate S, i.e., cooling, is achieved by two types of heat transfer: radiation and convection. Of these, radiation is expressed as the product of the emissivity and the difference between the fourth power of the absolute temperature of the surface of the coiled steel plate S and the fourth power of the absolute temperature of the surroundings. On the other hand, convection is expressed as the product of the convection heat transfer coefficient and the difference between the surface temperature of the coiled steel plate S and the surrounding air temperature. Therefore, the effect of radiation is greater when cooling high-temperature objects, and the effect of convection is greater at low temperatures.

In addition, when looking at the time required to cool the coiled steel sheet S, there is a tendency for the time taken to complete cooling, i.e., to be longer at low temperatures, particularly from below 100°C, to reach the desired temperature. Therefore, it is important to improve the cooling capacity by convection, which has a large impact on cooling at low temperatures as described above.

Of the above mentioned conditions, improving the convection heat transfer coefficient and lowering the surrounding air temperature are effective in improving the cooling capacity of convection. Furthermore, installing a cooling fan is effective in improving the convection heat transfer coefficient, while spraying cooling mist is effective in lowering the surrounding air temperature.

In this way, by increasing the convection heat transfer coefficient of the cooling fan, the evaporation rate of the condensed water can be increased, so that even if condensation occurs on the coiled steel plate S, the occurrence of rust can be suppressed.

On the other hand, experiments have shown that when the cooling mist evaporates, the latent heat of vaporization causes the air temperature to drop by around 5°C. In this way, spraying cooling mist from a cooling mist generator contributes to improving the cooling capacity of convection, especially at low temperatures.

From the above viewpoints, the coiled steel sheet cooling method in the embodiment of the present invention uses at least a cooling mist generator 1 and a cooling fan 2 as the cooling equipment C. Furthermore, based on the cooling concept described above, the cooling mist sprayed by the cooling mist generator 1 is not sent directly to the coiled steel sheet S, but is evaporated once and the cooled air is sent to the coiled steel sheet S via the cooling fan 2.

By using this cooling method, the occurrence of rust caused by the cooling mist adhering to the coiled steel plate S can be reduced, and even if condensation occurs on the coiled steel plate S, the surface of the coiled steel plate S can be prevented from becoming wet because air is blown from the cooling fan 2. In addition, because the air at a lower temperature is blown to the coiled steel plate S via the cooling fan 2, the coiled steel plate S can be cooled more effectively.

Therefore, the flow rate of the cooling mist sprayed from the cooling mist generator 1 is limited to the amount at which the cooling mist can evaporate before reaching the cooling fan 2. If the flow rate of the cooling mist is too high, the relative humidity around the coiled steel plate S increases, and when the relative humidity reaches 100%, the cooling mist sprayed from the cooling mist generator 1 cannot evaporate completely.

On the other hand, if the flow rate of the cooling mist is small, the cooling capacity will be insufficient because the air will not be cooled sufficiently by the latent heat of vaporization caused by the evaporation of the cooling mist. Therefore, it is necessary to adjust the balance between the air flow rate of the cooling fan 2 and the amount of cooling mist sprayed by the cooling mist generator 1 to control the relative humidity in the cooling bed CF.

In this manner, in the coiled steel sheet cooling method according to the embodiment of the present invention, the cooling mist sprayed from the cooling mist generating device 1 evaporates, and the cooled air is sent to the coiled steel sheet S via the cooling fan 2 to cool the coiled steel sheet S. Therefore, it is desirable that the particle diameter of the cooling mist, i.e., the diameter of the droplets sprayed from the cooling mist generating device 1, is 40 μm or less. If the particle diameter of the cooling mist is 40 μm or less, the cooling mist evaporates in approximately 1 second.

On the other hand, the larger the particle diameter of the cooling mist, the longer it takes for the cooling mist to evaporate. Therefore, if the particle diameter of the cooling mist is larger than 40 μm, the efficiency of evaporation of the cooling mist is poor, and the cooling of the air by the heat of vaporization is insufficient. If the cooling mist is sucked into the cooling fan 2 without evaporating, there is a high possibility that the cooling mist will adhere to the coiled steel plate S without evaporating, increasing the risk of rust.

As described above, it is desirable for the particle size of the cooling mist to be as large as that mentioned above, but if the ease of evaporation is taken into too much consideration, the amount of cooling mist sprayed will be small, which will inevitably result in a small amount of evaporation. Therefore, in such cases, it is desirable to spray a large amount of cooling mist with a small particle size.

Furthermore, in the method for cooling the coiled steel sheet S according to the embodiment of the present invention, the cooling mist generator 1 intermittently sprays the cooling mist and stops spraying the cooling mist, and the spraying is controlled to be repeated alternately for a certain period of time. The spray interval can then be set according to the target cooling time for the coiled steel sheet S.

In this way, by alternately repeating the spraying and stopping of spraying of cooling mist for a certain period of time, it is possible to suppress an increase in humidity around the coil-shaped steel sheet S compared to when spraying is continuous. More preferably, since the risk of condensation increases as the temperature of the coil-shaped steel sheet S decreases, it is possible to more easily suppress rust by setting the spray interval of cooling mist to be longer.

In addition, when the coiled steel sheet S has cooled sufficiently and cooling is complete, it is preferable to stop spraying the cooling mist from the cooling mist generator 1 from the standpoint of running costs and the risk of condensation.

Here, when the cooling mist evaporates, the humidity increases, particularly around the coiled steel sheet S. If the humidity increases, condensation will form, especially when the coiled steel sheet S is at a low temperature, making it more likely to rust. For this reason, an increase in humidity is undesirable.

Since warm air rises, it is possible to prevent an increase in humidity by installing an exhaust fan on the ceiling R of the cooling floor CF and actively exhausting the air from the cooling floor CF, for example, outside the building B.

For example, if the temperature of the cooling bed CF is 40°C, the relative humidity is 90%, and the dew point is 38°C, the coiled steel plate S will condense at or below 38°C. Therefore, when cooling causes the surface temperature of the coiled steel plate S to drop by 2°C relative to the temperature around the coiled steel plate S, or when the relative humidity reaches 100%, condensation will form on the surface of the coiled steel plate S, so it is necessary to exhaust the air around the cooling bed CF to reduce the humidity.

Therefore, in the cooling equipment C for the coiled steel plate S in the embodiment of the present invention, roof fans RF are provided on the ceiling R of the building B. That is, as shown in Figures 1 and 2, for example, three roof fans RF are provided in the building B. The number of roof fans RF to be provided can be freely set. Also, as shown in Figure 1, the roof fans RF are installed in a row on the ceiling R, but the position of the roof fans RF can also be freely set.

In Figure 2, the dashed lines show the movement of air when the cooling equipment C, consisting of the cooling mist generator 1, the cooling fan 2, and the roof fan RF, is in operation. As shown in Figure 2, the cooling mist generator 1 and the cooling fan 2 are arranged on both sides of the coiled steel plate S placed in a row.

The cooling mist sprayed from the cooling mist generator 1 evaporates between the cooling mist generator 1 and the cooling fan 2, thereby cooling the surrounding air. The cooling fan 2 sucks in this cooled air and blows it out together with the surrounding gas toward the coiled steel plate S. The gas sent from the cooling fan 2 toward the coiled steel plate S is exhausted from the roof fan RF to the outside of the building B, including the air around the cooling bed CF, as shown by the dashed arrows in Figure 2.

Furthermore, the cooling mist generator 1 and cooling fan 2 described above can be controlled based on information obtained by grasping the conditions inside building B using a measuring device 3. Figures 1 and 2 show a thermometer 3a and a video camera 3b installed on the wall of building B as an example of measuring device 3. A separate hygrometer may also be provided.

If the temperature and humidity can be determined using such a measuring device 3, the dew point can be estimated. This makes it possible to determine the risk of condensation occurring on the coiled steel sheet S and control the cooling equipment C, such as the cooling mist generator 1 and cooling fan 2. In addition, based on the information on temperature and other information determined by the measuring device 3, it is also possible to remove the coiled steel sheet S from the cooling bed CF before condensation occurs.

Since many coiled steel plates S are placed on the cooling bed CF, it is preferable to use, for example, a radiation thermometer as the thermometer 3a used for temperature measurement. Also, a two-dimensional radiation thermometer may be used to measure the temperatures of multiple coiled steel plates S with one device. The measurement position is preferably a surface perpendicular to the central axis of the coiled steel plate S, i.e., the surface formed by its width ends.

This is because the surface of the coiled steel sheet S that is made up of the width ends has poor thermal conductivity because the steel sheets are stacked in the radial direction. In addition, the outermost and innermost circumferences may be loosely wound, which may further worsen thermal conductivity. When thermal conductivity is poor in this way, the temperature is not easily transmitted from the inside, making it necessary to underestimate the temperature and making accurate measurement difficult. Therefore, the surface made up of the width ends is measured, which allows the temperature of the coiled steel sheet S to be measured reliably without such problems.

In addition, a video camera 3b is also provided as the measuring device 3. By providing such a video camera 3b, it is possible to directly observe the coiled steel plate S, and therefore it is also possible to directly determine whether condensation has occurred. Furthermore, as will be described later, information from the video camera 3b can also be used when generating a model for predicting the occurrence of condensation on the coiled steel plate S.

Here, a video camera 3b is used as a measuring device to observe whether condensation has occurred, but any type of imaging device can be used as long as it can capture an image. In addition, the timing of the measurement process by the measuring device 3 can be set arbitrarily, such as constant measurement or measurement at predetermined intervals.

In this way, the cooling equipment C in this embodiment of the present invention is composed of the above-mentioned cooling mist generating device 1, cooling fan 2, measuring device 3, and roof fan RF, and also includes a control device 4 for controlling the cooling equipment C.

Figure 3 is a block diagram showing the internal configuration of the control device 4 in an embodiment of the present invention. The control device 4 includes an information acquisition unit 41, a learning unit 42, a storage unit 43, a cooling condition calculation unit 44, and a cooling condition output unit 45. The control device 4 controls the operation of each device constituting the cooling equipment C described above, generates a predictive model for reference during control by machine learning, and also updates the generated predictive model.

Examples of predictive models include a predictive model used to predict the cooling completion time of the coiled steel plate and a predictive model used to predict the occurrence of condensation on the coiled steel plate.

In other words, the coiled steel sheet cooling completion time prediction model is a model that predicts the time it takes for the coiled steel sheet S to reach the cooling completion temperature after being cooled by the cooling equipment C on the cooling bed CF. On the other hand, the coiled steel sheet condensation occurrence prediction model is a model that predicts whether or not condensation will occur on the coiled steel sheet S during the process of cooling the coiled steel sheet S on the cooling bed CF.

Note that the control device 4 shown in Figure 3 only shows functions related to the method of generating a predictive model for coiled steel sheets and the method of cooling the coiled steel sheets, which will be described below, and various other functions are not shown.

Therefore, the control device 4 may have a configuration in which, for example, a CPU (Central Processing Unit), a ROM (Read Only Memory), a RAM (Random Access Memory), and an input/output interface (not shown in FIG. 3) are connected via a bus. In addition, the input/output interface may be connected to the above-mentioned components, and may also be connected to components such as a display unit, a communication control unit, or an input unit.

The control device 4 is also connected to a host computer 5 and a database 6. The host computer 5 is a computer that connects the control device 4 that controls the cooling process of the coiled steel plate S in the building B equipped with the cooling bed CF described above, as well as the control devices used in each manufacturing process of the coiled steel plate S.

On the other hand, the database 6 stores various information related to the manufacturing of the coiled steel sheet S. The database 6 can be accessed by the host computer 5 as well as the control devices in each manufacturing process. Therefore, in FIG. 3, it is drawn outside the control device 4.

Therefore, for example, information from the host computer 5 is sent to the database 6 and stored, and the host computer 5 can access the database 6 to obtain necessary information. In addition, as described below, information on the set cooling conditions from the control device 4 and the cooling equipment C and information on values measured by the measuring device 3 are sent and stored.

The information acquisition unit 41 acquires information, for example, on the temperature measured by the measuring device 3 and the state of condensation. The information acquisition unit 41 also acquires various information on the coiled steel plate S to be cooled from the host computer 5. Specifically, the information acquisition unit 41 acquires component information of the coiled steel plate S, operating parameters of the coiled steel plate S, and performance data on the manufacture of the coiled steel plate S, as described below.

The learning unit 42 generates a prediction model for the cooling completion time of the coiled steel plate S and a prediction model for the occurrence of condensation on the coiled steel plate S by performing machine learning. That is, the component information and operational parameters acquired by the information acquisition unit 41 are used as input data, and a prediction model for predicting the state of the coiled steel plate S on the cooling bed is generated by machine learning using the performance data as teaching data.

Here, "composition information" refers to the components contained in the coiled steel sheet S, and is information that determines the transformation temperature of the coiled steel sheet S and the phase fraction before and after the transformation. Therefore, this composition information is essential for predicting the transformation of the coiled steel sheet S.

In general steel, the transformation is generally complete before coiling. However, for example, high-tensile steel sheets contain a large amount of hardening elements such as manganese (Mn), and depending on the temperature conditions before coiling, the transformation may not have progressed sufficiently. In such cases, the temperature history of the coiled steel sheet S after coiling changes due to the influence of the transformation. For this reason, this information is necessary for predicting the time to complete cooling and whether or not condensation will occur.

In addition, since the component information differs for each coiled steel sheet S, each piece of information is acquired from the host computer 5 by the information acquisition unit 41.

"Operation parameters" are values used in the cooling process of the coiled steel sheet S on the cooling bed CF. However, while it is certainly preferable to use various values that are set based on information on the target cooling completion time as described below, it is also possible to use values obtained by performing calculation processing in the cooling condition calculation unit 44, for example.

The information used as "operation parameters" is, for example, the "width of the coiled steel plate S," the "outer diameter of the coiled steel plate S," the "flow rate of the mist in the cooling equipment C (cooling mist generating device 1)," the "wind speed of the gas blown by the cooling fan 2," the "wind speed around the coiled steel plate S on the cooling bed CF," and the "temperature value measured for the coiled steel plate S."

The above-mentioned operating parameters may also include other parameters such as "cooling bed temperature" and "cooling bed humidity." Furthermore, by taking into account the wind speed of the gas around the coiled steel plate S, the prediction accuracy of the low-temperature side, where the contribution of convection heat transfer is greater, is improved.

Of these, the "width" and "outer diameter" of the coiled steel plate S are information acquired from the host computer 5 for each cooling target. The "flow rate of the mist in the cooling equipment C" and the "wind speed of the gas blown by the cooling fan 2" are values used when the control device 4 controls the drive of the cooling mist generator 1 and the cooling fan 2 when cooling the coiled steel plate S.

Here, the reason why information on the "wind speed around the coiled steel plate S of the cooling bed CF" is also collected is that, as described above, convective heat transfer is a more effective condition, particularly when the surface of the coiled steel plate S is at a low temperature, and this is a necessary parameter for accurate prediction. From this perspective, therefore, by further adding the temperature and humidity of the cooling bed CF and the exhaust flow rate of the roof fan RF, as described above, more accurate prediction is possible.

Note that "around the coiled steel plate S" here means that it is preferable to measure closer to the coiled steel plate S, although this depends on the heat resistance of the measuring device. Specifically, the measurement is performed at a position several hundred mm away from the coiled steel plate S, more realistically about 500 mm away.

The temperature of the coiled steel plate S is measured using the thermometer 3a shown in Figures 1 and 2. It is preferable that the temperature is measured on the surface formed by the widthwise end of the coiled steel plate S. The temperature of the coiled steel plate S is used to grasp the temperature history of the coiled steel plate S during the cooling process, and since it changes depending on the position where the coiled steel plate S is placed and the arrangement of the surrounding coiled steel plates S, it is preferable to obtain the temperature for as many coiled steel plates S as possible.

In addition to the information described above, the input data can also include various types of information stored in the database 6. This is because it is possible that the accuracy of predicting the cooling completion time can be improved by using operational performance data that can directly or indirectly affect the temperature history of the coiled steel sheet S as input data. Examples of the various types of information include environmental information such as the outside temperature, humidity, and weather at the location of the cooling bed CF.

On the other hand, the actual data in the manufacture of the coiled steel sheet S used as training data is, for example, information actually obtained by the measuring device 3 while the process of cooling the coiled steel sheet S has been carried out.

For example, the actual data used when generating a prediction model for the cooling completion time of a coiled steel sheet S is information on the temperature measured by the thermometer 3a and information on the cooling completion time calculated from that information.

That is, the temperature during the cooling process is measured by the thermometer 3a for each coiled steel plate S, or for each area set in the cooling bed CF on which the coiled steel plate S is placed. And, because the time when the coiled steel plate S to be cooled is brought into the building B is stored in the database 6, it is possible to know the cooling time until the temperature reaches the point where cooling is determined to be complete.

When generating a prediction model for predicting the occurrence of condensation, for example, information on the surface condition of the coiled steel sheet S captured by the video camera 3b becomes actual data. In other words, the occurrence of condensation can be determined from the information based on the video camera 3b, and other information linked to the information based on the video camera 3b at that time becomes training data.

The learning unit 42 uses the component information and operational parameters as input data and the performance data as training data to generate a prediction model through machine learning. Specifically, the learning unit 42 uses a machine learning algorithm to generate a prediction model for the cooling completion time or the occurrence of condensation.

As machine learning algorithms, for example, neural networks, decision tree learning, random forests, support vector regression, etc. can be adopted. Furthermore, the prediction of the cooling completion time and the occurrence of condensation may not be made into separate prediction models, but may be configured as an ensemble model that combines these prediction models.

Furthermore, the various performance data stored in the database 6 can be used not only to construct the learned cooling completion time and condensation occurrence prediction model described above, but also as training data. These performance data may be data measured by various measuring devices 3 as described above.

The timing at which the learning unit 42 performs machine learning can be set arbitrarily, for example, at any time or at a specified interval. In addition, the learning unit 42 may perform machine learning by providing the teaching data read by the host computer 5 from the database 6 to the cooling completion time prediction model and the condensation occurrence prediction model.

The cooling completion time prediction model and the condensation occurrence prediction model generated (constructed) by the learning unit 42 are stored, for example, in the memory unit 43 and are used to control the cooling of the coiled steel plate S.

Here, the memory unit 43 is composed of, for example, a semiconductor or a magnetic disk, and stores the above-mentioned cooling completion time prediction model and condensation occurrence prediction model, as well as programs used by the cooling condition calculation unit 44 described below when setting them.

As shown in FIG. 3, the control device 4 in the embodiment of the present invention is described on the assumption that it has a storage unit 43 inside. However, the storage unit 43 may not be provided inside the control device 4, and for example, the database 6 may be configured to perform that function, and the configuration method can be freely selected.

The cooling condition calculation unit 44 sets the control contents for driving each cooling equipment C when cooling the coiled steel plate S placed on the cooling floor CF in the building B. That is, based on information on the target cooling completion time received from the upper computer 5, for example, it sets the input data required to drive and control each part constituting the cooling equipment C, such as the cooling mist generator 1.

Specifically, for example, for the cooling mist generator 1, it is the amount of cooling mist sprayed (flow rate). For the cooling fan 2 and roof fan RF, the control content is set to achieve a specified wind speed.

Furthermore, the cooling process of the coiled steel plate S is performed for each cooling area set on the cooling floor CF of the building B. Therefore, when the cooling condition calculation unit 44 executes such a process, the control contents are set for each of these areas. In this case, in many cases, multiple coiled steel plates S will exist in the cooling area, and in such a case, the control contents of the cooling equipment C are set, for example, based on the coiled steel plate S that is most susceptible to rusting, so as to prevent rust from occurring in such coiled steel plate S.

The cooling condition calculation unit 44 then applies each of the set conditions to a prediction model for the cooling completion time obtained from the memory unit 43 and a prediction model for determining whether condensation will occur, and calculates predicted values for the cooling completion time and the presence or absence of condensation.

As a result, it is judged whether the cooling completion time was achieved in line with the target, and whether condensation occurred, whether the coiled steel sheet S could be cooled in a manner that prevented condensation from occurring.

If the desired result is not obtained as a result of the judgment, the control content is reviewed and the calculation is repeated using each prediction model again so that the desired result is obtained. On the other hand, if the desired result is obtained, the cooling condition calculation unit 44 sends a control signal to each cooling device C via the cooling condition output unit 45 and drives and controls the cooling device C based on the set input data.

The cooling condition output unit 45 transmits the input data received from the cooling condition calculation unit 44 to the cooling equipment C, and also transmits it to the database 6, where it is stored as teaching data. The cooling equipment C is then driven based on the control signal set and output by the control device 4, and performs the cooling process on the coiled steel plate S placed on the cooling bed CF. Furthermore, information on the control content may be transmitted from the cooling equipment C to the database 6.

The thus constructed cooling completion time prediction model, the method for generating the condensation occurrence prediction model, and the cooling method for coiled steel sheet S using these cooling completion time prediction model and condensation occurrence prediction model make it possible to predict and control with high accuracy the temperature history and condensation occurrence of the coiled steel sheet S placed on the cooling bed while being cooled.

In addition, because the cooling equipment C is controlled in this manner, it becomes easier to grasp the time when the cooling of the coiled steel plate S to be cooled will be completed, which makes it easier to predict the schedule for removing the coiled steel plate S from building B, which also contributes to the management of shipments.

In addition, the input data stored in the database 6 may be used to perform machine learning by the learning unit 42 described above, thereby updating the prediction model for the cooling completion time and the prediction model for the occurrence of condensation. By repeating such processing, the accuracy of each prediction model is improved, and the cooling method for the coiled steel plate S can be executed more precisely.

Furthermore, when controlling the cooling equipment C, if the input data used by the cooling condition calculation unit 44 to set the control content is, for example, an abnormal value, there is a high possibility that the result of applying the predicted data will also be an abnormal value, and in such a case, the administrator can be notified of this via the display unit described above, etc.

[Action]
Next, the flow of generating a prediction model for the cooling completion time and a prediction model for the occurrence of condensation in the control device 4, and the flow of control of the cooling equipment C using the generated prediction models will be explained with reference to Figures 4 and 5.

First, the flow of generating a prediction model for the cooling completion time and the occurrence of condensation is shown. Figure 4 is a flowchart showing the flow of machine learning performed by the control device 4 in an embodiment of the present invention to generate a model for predicting the cooling completion time of the coiled steel plate S or the occurrence of condensation.

First, the control device 4 acquires input data (ST1). The input data is the same whether generating a prediction model for the cooling completion time or generating a prediction model for the occurrence of condensation, and as described above, is the component information and operational parameters of the coiled steel plate S to be cooled.

Meanwhile, training data is also acquired (ST2). The training data is actual data, but in the case of a prediction model for the cooling completion time, it is information on the temperature measured by the thermometer 3a and information on the cooling completion time calculated from that information. On the other hand, in the case of a prediction model for the presence or absence of condensation, it is information on the surface condition of the coiled steel sheet S photographed by the video camera 3b.

The learning unit 42 acquires, for example, the input data and teacher data obtained by the host computer 5 accessing the database 6 by receiving them via the information acquisition unit 41. Alternatively, the learning unit 42 may acquire them by accessing the storage unit 43 _.

Then, the learning unit 42 performs machine learning using the obtained input data and teacher data (ST3). As for the algorithm used in the machine learning, any algorithm may be used as described above.

The prediction model generated (constructed) as a result of the machine learning is then sent to the memory unit 43 and stored therein (ST4). With the above steps, a prediction model for the cooling completion time and a prediction model for the presence or absence of condensation are generated.

Note that here, the predictive model generated by the learning unit 42 through machine learning is stored as is in the storage unit 43, but for example, the predictive model generated by the learning unit 42 may be evaluated as to whether it is appropriate. Depending on the results of the evaluation, the predictive model may be reconstructed.

Next, we will explain the flow of the cooling method for the coiled steel plate S using each of the generated prediction models. Figure 5 is a flowchart showing the flow of the control device 4 controlling the cooling equipment C using the prediction models generated when predicting the cooling completion time of the coiled steel plate S or the occurrence of condensation in an embodiment of the present invention.

When executing the cooling process for the coiled steel plate S to be cooled, the control device 4 acquires information on the target cooling completion time from the host computer 5 (ST11). Then, it sets input data for control so that the cooling process is completed at the received target cooling completion time (ST12).

Specifically, as described above, for example, the flow rate of the cooling mist in the cooling mist generator 1 and the air volume in the cooling fan 2 and the roof fan RF are set by the cooling condition calculation unit 44. Alternatively, it is also possible to adopt the conditions used in each prediction model stored in the memory unit 43 as input data.

The cooling condition calculation unit 44 then applies the input data set in this manner to a prediction model of the cooling completion time obtained from the memory unit 43, or a prediction model regarding the occurrence or non-occurrence of condensation (ST13).

By applying the input data to the prediction model in this way, it is possible to predict the time until cooling is complete, and also to predict whether condensation will occur on the coiled steel plate S under the cooling conditions.

The cooling condition calculation unit 44 checks the results of applying the input data to each prediction model. In other words, it checks against the information on the target cooling completion time to determine whether the cooling process of the target coiled steel sheet S will be completed by the target cooling completion time and whether condensation will occur (ST14).

If the desired results cannot be obtained (NO in ST14), the input data is changed (ST15) and the process returns to step ST12 to set the input data and apply it to the prediction model.

On the other hand, if the desired result is obtained (YES in ST14), the cooling condition calculation unit 44 sends the input data to the cooling condition output unit 45, which then sends the input data to the cooling equipment C via the cooling condition output unit 45 (ST16). This causes the cooling equipment C, such as the cooling mist generator 1, to be driven and controlled based on the input data.

In addition, the cooling condition output unit 45 transmits the input data to the database 6 as teacher data and stores it (ST17). Note that the input data to be used as teacher data can also be transmitted to the storage unit 43, for example, and stored there, instead of being transmitted to the database 6.

The control device 4 executes the above-mentioned series of processes, and the coiled steel sheet S is cooled by the cooling equipment C using a cooling method that uses a prediction model for the cooling completion time and a prediction model for the occurrence of condensation.

Next, we will explain the flow of generating the prediction model for the cooling completion time and the prediction model for the occurrence or non-occurrence of condensation, as well as an example of a method for cooling a coiled steel sheet S using cooling equipment C such as a cooling mist generator 1.

In this embodiment, as shown in Figures 1 and 2, multiple coiled steel plates S were placed at 1 m intervals on the cooling bed CF. In addition, a cooling mist generator 1 and a cooling fan 2 were placed as cooling equipment C at both ends of the row of coiled steel plates S. The positional relationship between the two is such that the cooling mist generator is placed behind the cooling fan 2.

The coiled steel plate S was cooled using this cooling equipment C. The starting cooling temperature of the coiled steel plate S was 600°C, the width of the coiled steel plate S was 1000 mm, and the outer diameter was 1950 mm. As shown in the figure, five coiled steel plates S were arranged in one row, the temperature of the cooling bed CF was 30°C, and the relative humidity was 80%. Water was used as the raw material for the cooling mist, and its temperature was 30°C. The cooling of the coiled steel plate S was stopped when the coil surface temperature reached 50°C.

Furthermore, as input data for generating the prediction model, data for the width of the coiled steel plate S of "800 to 1200 mm," the outer diameter of "1800 to 2100 mm," and the steel plate components of "C: 0.01 to 0.1%, Si: 0.1 to 1.0%, Mn: 0.5 to 2.0%" were used. The same data set was also used as input data for predicting condensation.

Then, in predicting the cooling completion time of the coiled steel sheet S, the temperature of the steel sheet before winding was used as input data. That is, the temperature of the coiled steel sheet S was measured by measuring the temperature of the center of the width direction of the steel sheet with a spot-type radiation thermometer before winding, and the temperature of the top surface of the steel sheet was measured.

A prediction model for the cooling completion time was generated using the components contained in the coiled steel plate S and the operating parameters of the cooling equipment C as input data and temperature information from the thermometer 3a, which is the measuring device 3, as teacher data. In addition, using the same input data and information on the presence or absence of condensation captured by the video camera 3b as teacher data, a prediction model for the occurrence or absence of condensation on the coiled steel plate S was generated.

The cooling completion time and condensation occurrence of the coiled steel plate S were predicted using a prediction model for the cooling completion time and a prediction model for the occurrence of condensation. Furthermore, the parameters of the cooling equipment C and the exhaust flow rates of the cooling fan 2 and roof fan RF were adjusted so that the predicted cooling completion time was within the target range and condensation did not occur, and the operation of the cooling equipment C was controlled based on the control content.

As a result, the cooling of the coiled steel plate S, which was the object to be cooled, was completed within the target cooling completion time, while condensation was avoided, and rust was prevented. By using the above-mentioned method for generating a predictive model and the cooling method using the predictive model, the cooling time of the coiled steel plate S was ultimately optimized, and the lack of space for the cooling bed was resolved.

The embodiments of the present invention described so far are merely examples of the present invention. Various modifications or improvements can be made to each of these embodiments, and such modifications or improvements can also be included in the present invention. Each of these embodiments and their variations are included in the scope and gist of the invention, and are included in the scope of the invention and its equivalents as set forth in the claims.

1 Cooling mist generator 2 Cooling fan 3 Measuring device 3a Thermometer 3b Video camera 4 Control device B Building C Cooling equipment CF Cooling floor RF Roof fan S Coiled steel plate

Claims (11)

Acquiring component information of a coiled steel sheet, operation parameters of the coiled steel sheet, and performance data in the manufacture of the coiled steel sheet;
a step of generating, by machine learning, a prediction model for predicting the occurrence of condensation on the coiled steel sheet on a cooling bed used when controlling at least the cooling mist generator and the cooling fan, the cooling fan being arranged at an interval sufficient to allow the cooling mist sprayed from the cooling mist generator to be evaporated, the cooling mist being sucked in and blown to the coiled steel sheet;
A method for generating a prediction model for a coiled steel sheet, comprising: The method for generating a prediction model for coiled steel plate according to claim 1, characterized in that the operational parameters are the dimensions and temperature of the coiled steel plate, the flow rate of the cooling medium used during cooling, and the flow rate of the cooling medium around the coiled steel plate. A measuring device is provided for detecting the state of both the coiled steel plate and the cooling bed, or one of the two,
The method for generating a prediction model for a coiled steel sheet according to claim 1 or 2, characterized in that information obtained by measurement with the measuring device is acquired as the teaching data. The method for generating a prediction model for a coiled steel sheet according to any one of claims 1 to 3, characterized in that the prediction model for the coiled steel sheet is a model for predicting the cooling completion time of the coiled steel sheet. The method for generating a predictive model for coiled steel sheet according to any one of claims 1 to 4, characterized in that the machine learning uses at least one machine learning algorithm from among neural network, decision tree learning, random forest, and support vector regression. A step of setting control contents of a cooling process of the coiled steel sheet using a prediction model for predicting a cooling completion time of the coiled steel sheet on a cooling bed or occurrence of condensation, which is generated by the method for generating a prediction model related to the coiled steel sheet according to any one of claims 1 to 5;
and controlling the cooling equipment based on the set control content.
The cooling equipment is composed of at least a cooling mist generating device and cooling fans arranged at intervals to allow the cooling mist sprayed from the cooling mist generating device to evaporate and to prevent the cooling mist from diffusing and not being sucked in ,
The step of controlling the cooling facility further comprises:
A step of spraying cooling mist from a cooling mist generating device based on the set control content;
a cooling fan sucking in air cooled by evaporation of the cooling mist sprayed by the cooling mist generator and blowing the air onto the coil-shaped steel plate placed on a cooling bed;
A method for cooling a coiled steel plate, comprising: After the step of blowing air onto the coiled steel sheet by the cooling fan,
The method for cooling a coiled steel plate according to claim 6, further comprising a step of exhausting the air around the coiled steel plate, including the gas blown to the coiled steel plate by the cooling fan, outside the cooling bed based on the control content. The method for cooling a coiled steel sheet according to claim 6 or 7, characterized in that the cooling mist generating device is disposed behind the direction of air blown by the cooling fan to the coiled steel sheet. The method for cooling a coiled steel sheet according to any one of claims 6 to 8, characterized in that the cooling mist has a particle diameter of 40 μm or less. A method for manufacturing a coiled steel sheet, comprising the method for cooling a coiled steel sheet according to any one of claims 6 to 9. The method for manufacturing coiled steel sheet according to claim 10, characterized in that hot-rolled steel sheet or cold-rolled steel sheet is manufactured.

Images