Continuous casting tundish nozzle blockage prediction method

Abstract

The invention relates to a continuous casting tundish nozzle blockage prediction method. The method comprises the following steps of 1, collecting data; 2, intelligent algorithm formula inference; 3, periodically calculating and predicting the steel passing amount; 4, the actual steel passing amount is periodically calculated; 5, predicting the water gap blocking degree; and step 6, when the blockage degree reaches a threshold value (0.65), notifying an automatic device for replacing the nozzle to complete the automatic replacement of the tundish nozzle. According to the technical scheme, a method for judging whether the tundish nozzle is blocked or not completely depending on artificial experience is abandoned, the state of the nozzle can be judged timely and accurately, casting blanks with abnormal quality produced due to nozzle blocking are effectively reduced, the molten steel yield is increased, and meanwhile the continuous casting production organization is more stable.

Description

Technical Field

[0001] The invention belongs to the technical field of steelmaking continuous casting, and in particular relates to a method for predicting the blockage of a water nozzle of a continuous casting tundish. Background Art

[0002] At present, whether the nozzle of the continuous casting tundish is blocked mainly depends on manual judgment, and the timing of replacement depends entirely on the experience of the operator, mainly based on: 1. Observe the flow: By observing the flow of molten steel from the nozzle during the continuous casting process, it is preliminarily judged whether the nozzle is blocked. If the molten steel flows poorly or intermittently, it may be a sign of nozzle blockage. 2. Check temperature changes: The blockage of the nozzle may cause the accumulation of molten steel at the nozzle during the continuous casting process, thereby increasing the temperature around the nozzle. By measuring the temperature changes around the nozzle, if an abnormal increase is found locally, it may indicate that the nozzle is blocked. 3. Observe sparks or bubbles: Use specific equipment (such as infrared thermal imager) for observation. If abnormal sparks or bubbles are observed around the nozzle, it may be due to the blockage that prevents the molten steel from flowing out smoothly. 4. Check the outside of the nozzle: Carefully check the outer surface of the nozzle. If impurities, solids or other foreign matter are found attached to the nozzle, it may be a sign of blockage. 5. Check pressure changes: By monitoring the changes in nozzle pressure during the continuous casting process, if the pressure is found to increase or fluctuate greatly, it may be caused by increased resistance caused by blockage.

[0003] These judgment methods that completely rely on manual experience often fail to detect that the nozzle is blocked in time, resulting in abnormal casting in a serious blockage state, which seriously reduces the quality of the casting. Therefore, a new solution is urgently needed to solve the above technical problems. Summary of the invention

[0004] The present invention aims at the problems existing in the prior art and provides a method for predicting the blockage of the water inlet of a continuous casting tundish. The present invention collects the data of the stopper position information, casting speed, and the lower mouth width of the crystallizer 30 minutes after the water inlet is replaced, the casting is started, and the tundish is replaced. Through the least squares intelligent algorithm, the linear relationship between the steel flow rate and the stopper position is regressed, thereby predicting the degree of water inlet blockage in the actual production process. When the blockage degree reaches the threshold, the water inlet automatic device is notified to complete the automatic replacement of the water inlet of the tundish. Therefore, compared with the prior art, this method has the characteristics of intelligence and automation, which greatly improves the quality of the ingot and reduces the labor intensity of the workers on the production line.

[0005] In order to achieve the above object, the technical solution of the present invention is as follows: a method for predicting the blockage of a continuous casting tundish nozzle, the method comprising the following steps:

[0006] Step 1: Collect data;

[0007] Step 2: Intelligent algorithm formula inference;

[0008] Step 3: Periodic calculation and forecast of steel throughput;

[0009] Step 4: Calculate the actual steel throughput in cycles;

[0010] Step 5: Predict the degree of nozzle blockage;

[0011] Step 6: When the blockage degree reaches a threshold, the water inlet automatic device is notified to complete the automatic replacement of the tundish water inlet.

[0012] The present invention collects data on the on-site stopper position information, casting speed, and lower mouth width of the crystallizer 30 minutes after replacing the water inlet, starting casting, and changing the tundish, and regresses the linear relationship between the steel throughput and the stopper position through the least squares intelligent algorithm, thereby predicting the degree of water inlet blockage in the actual production process. When the blockage degree reaches a threshold, the water inlet automatic replacement device is notified to complete the automatic replacement of the tundish water inlet.

[0013] As an improvement of the present invention, the data collection described in step 1 is specifically performed as follows:

[0014] Step 1: Collect data on the tundish nozzle within 30 minutes after the continuous casting nozzle is replaced, pouring is started, and the tundish is replaced as sample data for the prediction model, mainly including the actual position information of the stopper rod, the pulling speed, and the data on the lower mouth width of the crystallizer. The pulling speed and the lower mouth width of the crystallizer are mainly used to estimate the actual steel throughput.

[0015] As an improvement of the present invention, the intelligent algorithm formula inference described in step 2 is specifically performed as follows:

[0016] Step 2: Since the actual position of the stopper rod and the actual steel throughput have a linear relationship, the intelligent algorithm adopts the least squares linear regression algorithm, taking the actual position data of the stopper rod in 30 minutes as the independent variable x, and the actual steel throughput data in 30 minutes as the dependent variable y, and finds a linear curve or quadratic curve algorithm formula through the least squares algorithm. The specific algorithm is as follows;

[0017] Step 2-1: First, calculate the linear curve formula: y = β0 + β1x, where β0 is the intercept and β1 is the slope;

[0018]

[0019]

[0020] Among them, xi and yi represent the variable value of the i-th stopper position and the variable value of the steel flow, respectively. and represent the mean values ​​of stopper position and steel flow rate respectively;

[0021] Step 2-2: Calculate β0 and β1 to determine the linear curve algorithm equation;

[0022] Step 2-3: Calculate the coefficient of determination R2, R2 = 1-(residual sum of squares / total sum of squares);

[0023] Step 2-4: If R2<0.8, optimize the algorithm formula and calculate according to the quadratic curve formula:

[0024] y=β2*xi2+β1*xi+β0

[0025] SSR=Σ(yi-(β2*xi2+β1*xi+β0))2

[0026] Among them, xi and yi represent the variable value of the i-th stopper position and the variable value of the steel flow, respectively.

[0027] Minimize the objective function SSR, and find the optimal solution by taking its partial derivative and setting the derivative to zero. Take the partial derivatives of β2, β1, and β0 respectively, and solve the following system of equations:

[0028]

[0029]

[0030]

[0031] Get the optimal coefficient values ​​β2, β1, and β0.

[0032] As an improvement of the present invention, the periodic calculation and prediction of the steel throughput in step 3 is specifically performed as follows: using the formula determined in step 2, the steel throughput is periodically predicted by the actual position of the stopper rod,

[0033] Throughputpre=β0+β1*X

[0034] X: actual stopper position data

[0035] As an improvement of the present invention, the actual steel throughput is calculated periodically in step 4, and the specific steps are as follows:

[0036] Throughputact = pulling speed * lower opening width of crystallizer.

[0037] As an improvement of the present invention, the specific steps of predicting the blockage degree of the water outlet in step 5 are as follows:

[0038] Predicted water outlet blockage degree = (Throughputpre-Throughputact) / Throughputpre.

[0039] As an improvement of the present invention, when the blockage degree reaches a threshold value, the water inlet automatic device is notified in step 6 to complete the automatic replacement of the tundish water inlet. The specific steps are as follows:

[0040] When the blockage degree reaches the threshold value (0.65), the water inlet automatic device is notified to complete the automatic replacement of the tundish water inlet.

[0041] Compared with the prior art, the present invention has the following advantages: 1) This technical solution abandons the method of completely relying on manual experience to judge whether the water inlet of the tundish is blocked, and can timely and accurately judge the state of the water inlet, thereby effectively reducing the production of ingots with abnormal quality due to water inlet blockage, improving the molten steel yield and making the continuous casting production organization more stable; 2) This method has a high degree of automation, which can greatly reduce the labor intensity of production line workers. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 This is a flow chart for predicting the blockage of the water inlet of the continuous casting tundish;

[0043] Figure 2 This is a first-order curve fitting diagram of the stopper rod position and the amount of steel passing through. DETAILED DESCRIPTION

[0044] In order to deepen the understanding of the present invention, the present embodiment is described in detail below with reference to the accompanying drawings.

[0045] Example 1: See Figure 1 , a method for predicting the blockage of the water inlet of a continuous casting tundish, wherein the step 1 collects data: since there is no blockage in the water inlet of the tundish within 30 minutes after the continuous casting tundish is replaced, the pouring is started, and the tundish is replaced, the data of this period of time is collected as the sample data of the prediction model. As shown in Table 1 below, the main process data of 7 streams of a continuous casting plant 30 minutes after the tundish is replaced mainly include the data of casting length, mold bottom mouth width, casting speed, and stopper rod position information, among which the casting speed and mold bottom mouth width data are mainly used to estimate the actual steel throughput.

[0046]

[0047]

[0048]

[0049]

[0050] Table 1 Main process data of a continuous casting plant 30 minutes after the 7th strand changed the middle package

[0051] Step 2, intelligent algorithm formula inference: Since the actual position of the stopper rod and the actual steel flow have a linear relationship, the intelligent algorithm uses the least squares linear regression algorithm. Specifically, the actual position data of the stopper rod in 30 minutes is taken as the independent variable x, and the actual steel flow data in 30 minutes is taken as the dependent variable y. The least squares algorithm is used to fit a linear curve algorithm formula, as follows: Figure 2 shown.

[0052] The formula of the linear curve is: y = 0.1347x-9.6804, where the slope is 0.1347, the intercept is -9.6804, and R 2 =0.9809.

[0053] Step 3, periodic calculation and prediction of steel throughput: using the above fitted formula, the steel throughput is predicted based on the actual position of the stopper rod in a 1-minute period.

[0054] If the actual position of the stopper rod is 114.05 mm at a certain time, the predicted steel flow rate is calculated as:

[0055] Throughput pre =0.1347*114.05-9.6804=5.682 tons / minute

[0056] Step 4: Calculate the actual steel throughput in a cycle

[0057] For example, if the casting speed is 1.5m / min and the lower opening width of the crystallizer is 1300mm, the actual steel throughput is calculated as: act =1.5*1.3=1.95 tons / minute

[0058] Step 5: Predict the degree of sprue blockage

[0059] Predicted water outlet blockage degree = (Throughput pre -Throughput act ) / Throughput pre =(5.682-1.95) / 5.682

[0060] =65.68%

[0061] Step 6, notifying the water inlet automatic device. Since the predicted water inlet blockage degree reaches 0.6568, which exceeds the threshold value (0.65), the water inlet automatic device is notified to complete the automatic replacement of the tundish water inlet.

[0062] It should be noted that the above embodiments are not intended to limit the protection scope of the present invention, and equivalent changes or substitutions made on the basis of the above technical solutions all fall within the protection scope of the claims of the present invention.

Claims

1. A method for predicting the blockage of a continuous casting tundish nozzle, characterized in that: The method comprises the following steps: Step 1: Collect data; Step 2: Intelligent algorithm formula inference; Step 3: Periodic calculation and forecast of steel throughput; Step 4: Calculate the actual steel throughput in cycles; Step 5: Predict the degree of nozzle blockage; Step 6: When the blockage degree reaches a threshold, the water inlet automatic device is notified to complete the automatic replacement of the tundish water inlet.

2. The method for predicting the blockage of the water nozzle of the continuous casting tundish according to claim 1, characterized in that: The specific steps for collecting data in step 1 are as follows: Step 1: Collect data on the tundish nozzle within 30 minutes after the continuous casting nozzle is replaced, pouring is started, and the tundish is replaced as sample data for the prediction model, mainly including the actual position information of the stopper rod, the pulling speed, and the data on the lower mouth width of the crystallizer. The pulling speed and the lower mouth width of the crystallizer are mainly used to estimate the actual steel throughput.

3. The method for predicting the blockage of the water nozzle of the continuous casting tundish according to claim 2, characterized in that: The specific steps of the intelligent algorithm formula inference in step 2 are as follows: Step 2: Since the actual position of the stopper rod and the actual steel throughput have a linear relationship, the intelligent algorithm adopts the least squares linear regression algorithm, taking the actual position data of the stopper rod in 30 minutes as the independent variable x, and the actual steel throughput data in 30 minutes as the dependent variable y, and finds a linear curve or quadratic curve algorithm formula through the least squares algorithm. The specific algorithm is as follows; Step 2-1: First, calculate the linear equation: y = β0 + β1x, where β0 is the intercept and β1 is the slope. Among them, xi and yi represent the variable value of the i-th stopper position and the variable value of the steel flow, respectively. and represent the mean values ​​of stopper position and steel flow rate respectively; Step 2-2: Calculate β0 and β1 to determine the linear curve algorithm equation; Step 2-3: Calculate the coefficient of determination R2, R2 = 1-(residual sum of squares / total sum of squares); Step 2-4: If R2<0.8, optimize the algorithm formula and calculate according to the quadratic curve formula: y=β2*xi2+β1*xi+β0 SSR=Σ(yi-(β2*xi2+β1*xi+β0))2, Among them, xi and yi represent the variable value of the i-th stopper position and the variable value of the steel flow, respectively. Minimize the objective function SSR, and find the optimal solution by taking its partial derivative and setting the derivative to zero. Take the partial derivatives of β2, β1, and β0 respectively, and solve the following system of equations: Get the optimal coefficient values ​​β2, β1, and β0.

4. The method for predicting the blockage of the water nozzle of the continuous casting tundish according to claim 3, characterized in that: The cycle calculation in step 3 predicts the steel throughput; the specific steps are as follows: Using the formula determined in step 2, the steel throughput is predicted periodically based on the actual stopper position. Throughputpre=β0+β1*X X: Actual stopper position data.

5. The method for predicting the blockage of the water nozzle of the continuous casting tundish according to claim 4, characterized in that: The actual steel throughput is calculated in the cycle in step 4. The specific steps are as follows: Throughputact = pulling speed * lower opening width of crystallizer.

6. The method for predicting the blockage of the water nozzle of the continuous casting tundish according to claim 5, characterized in that: In step 5, the degree of nozzle blockage is predicted. The specific steps are as follows: Predicted water outlet blockage degree = (Throughputpre-Throughputact) / Throughputpre.

7. The method for predicting the blockage of the water nozzle of the continuous casting tundish according to claim 6, characterized in that: The specific steps in step 6 are as follows: when the blockage degree reaches the threshold value (0.65), the water inlet automatic device is notified to complete the automatic replacement of the tundish water inlet.

Images