Continuous casting tundish plasma heating temperature control method

Abstract

The invention relates to a continuous casting tundish plasma heating temperature control method, which comprises the following steps of: firstly, analyzing temperature influence factor data of a procedure before continuous casting by adopting a K-Means clustering algorithm, setting a temperature interval, and dividing predicted refining finishing temperature into three types of H-high, M-medium and L-low; according to the technical scheme, the heating starting time is determined in advance according to prediction, energy is input into the molten steel before the temperature is obviously reduced, large fluctuation of the temperature is prevented, and the temperature trend of the molten steel in the tundish is kept gentle, so that the product quality is improved, and the use requirements of users are met.

Description

Technical Field

[0001] The invention relates to a control method, in particular to a continuous casting tundish plasma heating temperature control method, belonging to the technical field of steelmaking continuous casting. Background Art

[0002] In the continuous casting process of steel, qualified molten steel is carried by a ladle and transported to the continuous casting machine. The continuous casting machine has a special rotary device for receiving the ladle, which has two stations to realize the alternating continuous supply of the ladle. The specific continuous casting process is that the molten steel enters the tundish from the ladle, and then enters the crystallizer from the tundish and gradually solidifies, which is the casting process. The tundish plays the role of connecting the ladle and buffering, realizing the continuous production of multiple ladle continuous casting.

[0003] In order to obtain stable production quality and prevent production interruptions, it is necessary to control the temperature of the molten steel in the tundish within an appropriate range. The temperature of the molten steel in the tundish is affected by many factors, including: the number of times the ladle is used, the time the ladle waits for steel tapping, the steelmaking tapping temperature, the duration of tapping, the amount of material added to tapping, the time from tapping to refining, the refining start temperature, the refining end temperature, the refining duration, the amount of refining material added, the time from the end of refining to the start of continuous casting, the ladle casting time, the ladle continuous casting sequence, the tundish capacity, etc. In actual production, these variables vary greatly and affect each other, resulting in fluctuations in the temperature of the tundish. Under current technical conditions, the temperature fluctuation range of the molten steel in the tundish is generally 10 to 50°C.

[0004] Through real-time monitoring, the temperature trend of the tundish steel liquid in the continuous casting process is a continuous change in the shape of a "peak" over time. The temperature trend of each ladle casting process presents a "peak". The temperature of each ladle in the middle of the casting is the highest, that is, the "peak", and the temperature of the opening and final casting is the lowest, that is, the "peak valley". There are also differences in temperature between each ladle. The temperature trend of all ladles in the entire casting is a continuous and uneven "peak". The variation range of these "peaks" is large, often beyond the ideal control range.

[0005] Chinese patent application CN202210010603.4 proposes a temperature control method for a chemical reaction system based on an artificial intelligence mechanism, specifically a reaction temperature control method and system based on deep reinforcement learning. The invention is particularly based on the framework of a flexible action evaluation algorithm, and can update the weights of the strategy network inside the temperature control model through deep reinforcement learning. At present, there is no public model method for continuous casting tundish heating temperature control. Summary of the invention

[0006] The present invention is aimed at the problems existing in the prior art, and provides a method for controlling the temperature of the continuous casting tundish by plasma heating. The technical solution uses the tundish plasma heating to control the temperature of the molten steel during the continuous casting process, and predictably regulates and compensates the temperature of the molten steel, that is, by heating to increase or make up for the "peaks and valleys" of the temperature trend; at the same time, it also predictably reduces the high-temperature "peak" of the previous process, so that the temperature fluctuation amplitude is reduced, or the temperature values ​​are completely consistent, so that the process and quality of the continuous casting production are stabilized. In order to reduce the temperature fluctuation amplitude of the molten steel in the tundish, or make it consistent, on the one hand, it is necessary to heat the low-temperature section, that is, to raise the "peak and valley"; on the other hand, it is also necessary to control the highest value of the temperature within a reasonable height, that is, to reduce the "peak".

[0007] The factors affecting the temperature of molten steel in the tundish include: the number of times the ladle is used, the time the ladle waits for tapping, the tapping temperature of steelmaking, the duration of tapping, the amount of material added to tapping, the time from the end of tapping to refining, the refining start temperature, the refining end temperature, the refining duration, the amount of refining material added, the time from the end of refining to the start of continuous casting, the ladle casting time, the ladle continuous casting sequence, the tundish capacity, etc. These factors also affect each other, that is, there is a serious uncertainty and collinearity.

[0008] In order to achieve the above object, the technical solution of the present invention is as follows: a temperature control method based on plasma heating of a continuous casting tundish, the method comprising the following steps:

[0009] Step 1: A ladle with a specific number carrying molten steel arrives at the refining station, and the computer collects relevant information of the ladle, including: "Number of times the ladle is used", "Time the ladle waits for steel tapping", "Steelmaking and steel tapping temperature", "Steel tapping duration", "Amount of steel tapping materials added", "Time from steel tapping completion to refining", "Refining start temperature" and other data.

[0010] Step 2: Based on the data collected in step 1, the K-Means clustering algorithm is used to analyze the temperature of the molten steel and predict the "refining end temperature" after the general "standard refining operation mode" is completed.

[0011] Step 3: According to the actual steel composition, the temperature range is pre-set, and the refining end temperature is divided into three categories: "H-high", "M-medium", and "L-low". If the "refining end temperature" of the batch is expected to be within the "H-high" range in step 2, then add an appropriate amount of cooling material to the molten steel before the end of refining, and extend the refining time. This is the "modified refining operation mode".

[0012] Step 4: After the "correction of the refining operation mode", the temperature of the molten steel is made to reach the "M-medium" range at the end of refining, and then the ladle and the molten steel it carries are sent to the continuous casting process.

[0013] Step 5: If the "refining end temperature" of the ladle is estimated to be within the "M-medium" or "L-low" range in step 2, after the "standard refining operation mode" is completed, the ladle and the molten steel it carries are sent to the continuous casting process.

[0014] Step 6: In the continuous casting process, for the ladle with a specific number carrying molten steel that arrives, the computer collects relevant information of the ladle, including: "number of times the ladle is used", "time the ladle waits for steel tapping", "steelmaking and steel tapping temperature", "steel tapping duration", "amount of steel tapping material added", "time from steel tapping completion to refining", "refining start temperature", "refining end temperature", "refining duration", "amount of refined material added", "time from refining completion to continuous casting start", "ladle casting time", "ladle continuous casting sequence", "intermediate ladle capacity" and other data,

[0015] Step 7: In the continuous casting process, the SVR support vector regression algorithm is used to analyze the data of factors affecting the temperature collected in the above step 6 to predict the "peak" and "peak valley" of the temperature of the molten steel in the tundish. When the expected "peak valley" value is lower than the threshold, the molten steel in the tundish is heated after the ladle is poured.

[0016] Step 8: In the continuous casting process, during each ladle casting process, without heating, the temperature trend of the molten steel in the intermediate ladle is displayed as a peak. The first half of the peak, i.e., the rising stage of the trend, corresponds to the first half of the ladle casting, and the second half of the peak, i.e., the falling stage of the trend, corresponds to the second half of the ladle casting. The temperature is lowest at the start and end of casting. Here, the BP neural network algorithm is used to analyze the data of factors affecting the temperature collected in the above step 6 to determine the low point of temperature, the rate of temperature rise, and the rate of temperature drop. In the first half, use a smaller heating power and a shorter heating time, and stop heating in advance; in the second half, use a larger heating power and a longer heating time, and keep heating until the end of casting.

[0017] Step 9: When training the BP neural network, optimize the parameters of the neural network and determine that the number of hidden layers is preferably 3. The training effect is better when the number of hidden nodes in the hidden layer is 15, 11, and 13.

[0018] Compared with the prior art, the present invention has the following advantages: the technical solution determines the start time of heating in advance according to the prediction, inputs energy into the molten steel before the temperature drops significantly, prevents large temperature fluctuations, and keeps the temperature trend of the molten steel in the tundish gentle, thereby improving product quality and meeting the user's use requirements. The comprehensive effect is that the solidification structure of the continuous casting slab is uniformly and effectively controlled. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1Tundish molten steel temperature prediction and analysis flow chart,

[0020] Figure 2 Temperature trend of molten steel in tundish under normal mode,

[0021] Figure 3 Temperature trend of molten steel in tundish under temperature control mode. DETAILED DESCRIPTION

[0022] In order to deepen the understanding of the present invention, the present embodiment is described in detail below with reference to the accompanying drawings.

[0023] Embodiment 1: A temperature control method based on plasma heating of a continuous casting tundish, the method comprising the following steps:

[0024] Step 1: A ladle with a specific number carrying molten steel arrives at the refining station, and the computer collects relevant information of the ladle, including: "Number of times the ladle is used", "Time the ladle waits for steel tapping", "Steelmaking and steel tapping temperature", "Steel tapping duration", "Amount of steel tapping materials added", "Time from steel tapping completion to refining", "Refining start temperature" and other data.

[0025] Step 2: Based on the data collected in step 1, the K-Means clustering algorithm is used to analyze the temperature of the molten steel and predict the "refining end temperature" after the general "standard refining operation mode" is completed.

[0026] Step 3: According to the actual steel composition, the temperature range is pre-set, and the refining end temperature is divided into three categories: "H-high", "M-medium", and "L-low". If the "refining end temperature" of the batch is expected to be within the "H-high" range in step 2, then add an appropriate amount of cooling material to the molten steel before the end of refining, and extend the refining time. This is the "modified refining operation mode".

[0027] Step 4: After the "correction of the refining operation mode", the temperature of the molten steel is made to reach the "M-medium" range at the end of refining, and then the ladle and the molten steel it carries are sent to the continuous casting process.

[0028] Step 5: If the "refining end temperature" of the ladle is estimated to be within the "M-medium" or "L-low" range in step 2, after the "standard refining operation mode" is completed, the ladle and the molten steel it carries are sent to the continuous casting process.

[0029] Step 6: In the continuous casting process, for the ladle with a specific number carrying molten steel that arrives, the computer collects relevant information of the ladle, including: "number of times the ladle is used", "time the ladle waits for steel tapping", "steelmaking and steel tapping temperature", "steel tapping duration", "amount of steel tapping material added", "time from steel tapping completion to refining", "refining start temperature", "refining end temperature", "refining duration", "amount of refined material added", "time from refining completion to continuous casting start", "ladle casting time", "ladle continuous casting sequence", "intermediate ladle capacity" and other data,

[0030] Step 7: In the continuous casting process, the SVR support vector regression algorithm is used to analyze the data of factors affecting the temperature collected in the above step 6 to predict the "peak" and "peak valley" of the temperature of the molten steel in the tundish. When the expected "peak valley" value is lower than the threshold, the molten steel in the tundish is heated after the ladle is poured.

[0031] Step 8: In the continuous casting process, during each ladle casting process, without heating, the temperature trend of the molten steel in the intermediate ladle is displayed as a peak. The first half of the peak, i.e., the rising stage of the trend, corresponds to the first half of the ladle casting, and the second half of the peak, i.e., the falling stage of the trend, corresponds to the second half of the ladle casting. The temperature is lowest at the start and end of casting. Here, the BP neural network algorithm is used to analyze the data of factors affecting the temperature collected in the above step 6 to determine the low point of temperature, the rate of temperature rise, and the rate of temperature drop. In the first half, use a smaller heating power and a shorter heating time, and stop heating in advance; in the second half, use a larger heating power and a longer heating time, and keep heating until the end of casting.

[0032] Step 9: When training the BP neural network, optimize the parameters of the neural network and determine that the number of hidden layers is preferably 3. The training effect is better when the number of hidden nodes in the hidden layer is 15, 11, and 13.

[0033] Embodiment 2: As attached Figure 2 The figure shows a record of the temperature trend of the tundish steel liquid under normal conditions without temperature model control, with continuous pouring of 5 ladles, represented by ① to ⑤. The temperature of the molten steel in the tundish fluctuates greatly, with the maximum and minimum temperature difference between furnaces reaching more than 25°C, and the temperature difference within the same furnace reaching about 15°C.

[0034] For attachment Figure 2 The results shown are achieved using the method of the present invention.

[0035] First, for the furnace situation represented by ①, the temperature is high. Under the control of the model, the following data are collected: the number of times the ladle is used, the time the ladle waits for steel tapping, the steelmaking and steel tapping temperature, the duration of steel tapping, the amount of steel tapping materials added, the time from steel tapping to refining, the refining start temperature, the refining end temperature, the duration of refining, the amount of refining materials added, etc. Using the K-Means clustering algorithm, the collected data on temperature influencing factors are analyzed, and it can be predicted that the temperature of the molten steel at the end of refining is "H-high".

[0036] Furthermore, the amount of coolant added is calculated based on the furnace capacity, heat capacity of molten steel and other values, and then added to the ladle. Then the calculation in the previous step is repeated. When the molten steel temperature at the end of refining is expected to be "M-Medium", the refining process is terminated and the ladle is transported to the continuous casting process. If the calculation is still "H-High" after adding the coolant, continue with the previous step until it is predicted to reach "M-Medium".

[0037] Furthermore, when the ladle arrives at the continuous casting station and starts pouring, the following data are collected: the number of times the ladle is used, the time the ladle waits for tapping, the steelmaking tapping temperature, the tapping duration, the amount of tapping material added, the time from tapping completion to refining, the refining start temperature, the refining end temperature, the refining duration, the amount of refining material added, the time from the end of refining to the start of continuous casting, the ladle casting time, the ladle continuous casting sequence, the tundish capacity, etc. The collected temperature influencing factor data are analyzed using the SVR support vector regression algorithm.

[0038] Furthermore, it is predicted whether the temperature of the molten steel in the tundish is lower than the "peak-to-valley" value of the previous bulge. If so, the molten steel in the tundish is heated after the ladle is poured; if not, wait for the right time.

[0039] Furthermore, during the ladle casting process, the trained BP neural network is used to determine the temperature change trend of the tundish steel liquid. When the temperature is expected to be lower than the set threshold (25°C in this embodiment), heating begins. This situation generally occurs in the second half of a ladle casting.

[0040] As a result, the "valley" of the temperature trend of the tundish steel liquid is eliminated, and the minimum temperature is kept above the threshold. The improved temperature trend is shown in the attached figure. Figure 3 As shown. Figure 2 In contrast, Figure 3 One of the effects is to reduce the maximum temperature from 45℃ to 30℃; the second effect is to increase the minimum temperature from 20℃ to 25℃. The comprehensive effect is to reduce the temperature fluctuation range from 25℃ to 5℃.

[0041] According to the effect of the temperature model, a stable casting speed range is ensured, and the surface and internal quality of the continuous casting slab are good.

[0042] The above is a schematic description of the present invention and its implementation methods, which is not restrictive. The drawings are only one of the implementation methods of the present invention, and the actual application is not limited thereto. Therefore, if a person skilled in the art is inspired by it and does not deviate from the purpose of the invention, he or she can design a structure and an implementation method similar to the technical solution without creativity, which should fall within the protection scope of the present invention.

Claims

1. A method for controlling the temperature of a continuous casting tundish by plasma heating, characterized in that: The method comprises the following steps: Step 1: A ladle with a specific number carries the molten steel to the refining station. Step 2: Based on the data collected in step 1, the K-Means clustering algorithm is used to analyze the temperature of the molten steel and predict the "refining end temperature" after the general "standard refining operation mode" is completed. Step 3: According to the actual steel grade composition, the temperature range is pre-set, and the refining end temperature is divided into three categories: "H-high", "M-medium", and "L-low". If the "refining end temperature" of the batch is expected to be within the "H-high" range in step 2, then add an appropriate amount of cooling material to the molten steel before the end of refining, and extend the refining time. This is the "modified refining operation mode". Step 4: After "correcting the refining operation mode", the temperature of the molten steel is made to reach the "M-medium" range at the end of refining, and then the ladle and the molten steel it carries are sent to the continuous casting process. Step 5: If the "refining end temperature" of the ladle is estimated to be within the "M-medium" or "L-low" range in step 2, after the "standard refining operation mode" is completed, the ladle and the molten steel it carries are sent to the continuous casting process. Step 6: In the continuous casting process, for the ladle with a specific number carrying molten steel that arrives, the computer collects the relevant information of the ladle. Step 7: In the continuous casting process, the SVR support vector regression algorithm is used to analyze the data of factors affecting the temperature collected in the above step 6 to predict the "peak" and "peak valley" of the temperature of the molten steel in the tundish. When the expected "peak valley" value is lower than the threshold, the molten steel in the tundish is heated after the ladle is poured. Step 8: In the continuous casting process, during each ladle casting process, without heating, the temperature trend of the molten steel in the intermediate ladle is displayed as a peak. The first half of the peak, i.e., the rising stage of the trend, corresponds to the first half of the ladle casting, and the second half of the peak, i.e., the falling stage of the trend, corresponds to the second half of the ladle casting. The temperature is lowest during the start and end of the casting. Here, the BP neural network algorithm is used to analyze the data of factors affecting the temperature collected in the above step 6 to determine the low point of the temperature, the rate of temperature rise, and the rate of temperature drop. In the first half, a smaller heating power and a shorter heating time are used, and the heating is stopped in advance; in the second half, a larger heating power and a longer heating time are used, and the heating is maintained until the end of the casting. Step 9: When training the BP neural network, the parameters of the neural network are optimized and it is determined that the number of hidden layers is preferably 3, and the training effect is better when the number of hidden nodes in the hidden layer is 15, 11, and 13.

2. The method for controlling the temperature of a continuous casting tundish by plasma heating according to claim 1, characterized in that: Step 1: A ladle with a specific number carrying molten steel arrives at the refining station, and the computer collects relevant information of the ladle, including: "number of times the ladle is used", "time the ladle waits for steel tapping", "steelmaking and steel tapping temperature", "steel tapping duration", "amount of steel tapping material added", "time from steel tapping completion to refining", and "refining start temperature" data.

3. The method for controlling the temperature of a continuous casting tundish by plasma heating according to claim 1, characterized in that: Step 6: In the continuous casting process, for the ladle carrying molten steel that has arrived with a specific number, the computer collects relevant information of the ladle, including: "number of times the ladle is used", "the time the ladle waits for steel tapping", "steelmaking and steel tapping temperature", "steel tapping duration", "amount of steel tapping material added", "time from steel tapping completion to refining", "refining start temperature", "refining end temperature", "refining duration", "amount of refined material added", "time from refining completion to continuous casting start", "ladle casting time", "ladle continuous casting sequence", and "intermediate ladle capacity" data.

Images