A method for controlling the consumption of an lf refining electrode

Abstract

The application discloses a LF refining electrode consumption control method, S1: real-time collection of cumulative heating time T in the LF refining process t , smelting sound decibel L t and molten steel heating rate V t ; S2: input of the parameters into an electrode consumption rate prediction model; S3: comparison of real-time calculated Y real with a preset warning threshold Y limit ; S4: if Y>Y limit , dominant factors are determined according to each item value in the model formula, and process adjustment is carried out according to the determination result. The method realizes real-time monitoring and early warning of the electrode consumption by real-time collection of the heating time, the smelting sound decibel and the heating rate, establishment of the electrode consumption rate prediction model, and further optimization of the power supply system and reduction of the electrode consumption. The application realizes real-time visualization and early warning of the electrode consumption state, solves the problem of lagging of manual experience control, and effectively reduces the electrode consumption cost of the LF refining process.

Description

Technical Field

[0001] This invention relates to an LF refining method, and more particularly to an LF refining electrode consumption control method. Background Technology

[0002] The LF (Ladle Furnace) refining furnace is an important ladle refining device in modern steelmaking processes. Graphite electrodes, as its core conductive consumable, account for a significant proportion of the refining cost. Electrode consumption is mainly affected by factors such as arc stability, slag submersion arc effect, and electrode surface oxidation.

[0003] Traditional electrode consumption control methods mostly focus on improving the quality of the electrode itself or simple voltage and current adjustments. For example, the arc brightness is judged by manual observation, or the power supply curve is set based on experience. However, in actual production, the refining process is a complex and dynamically changing system: as the heating time increases, the electrode surface temperature rises, and the rate of lateral oxidation consumption increases non-linearly; the foaming state of the slag directly affects the arc combustion sound (noise), and excessive noise usually indicates poor arc sublimation, leading to increased sublimation at the electrode tip; changes in the heating rate reflect the arc energy utilization efficiency, and an abnormally low heating rate is often accompanied by the risk of electrode breakage or ineffective dry burning.

[0004] Application number CN201820224192.8 provides a low-electrode-consumption LF furnace cover. The core control concept is to reduce the oxidizing atmosphere of the high-temperature part of the LF furnace electrode by optimizing the LF furnace cover structure, thereby reducing electrode consumption by more than 15%. However, it does not address adjusting process control in a timely manner according to changes in conditions to reduce electrode consumption. Application number CN202511691040.X provides a method for creating foam slag in LF refining using waste graphite electrodes. This method involves grinding waste graphite electrodes to obtain graphite electrode powder. Utilizing the high carbon content, high conductivity, and low sulfur characteristics of graphite electrodes, it achieves submerged arc heating, reducing power consumption and shortening refining time. This method also does not systematically optimize and control process parameters.

[0005] Currently, there is a lack of a method that can comprehensively analyze various parameters of LF refining in real time, quantitatively predict the electrode consumption rate, and perform closed-loop control accordingly. Therefore, developing an electrode consumption control method based on a multi-parameter fusion prediction model is of significant practical importance for reducing steelmaking costs. Summary of the Invention

[0006] The technical problem to be solved by the present invention is to provide a method for quantitatively predicting and controlling the consumption of LF refining electrodes.

[0007] To solve the above-mentioned technical problems, the technical solution adopted by the present invention includes the following steps: S1: Real-time acquisition of the cumulative heating time T during the LF refining process. t , Smelting sound decibels L t and the heating rate V of molten steel t ; S2: Input the above parameters into the electrode consumption rate prediction model. The model formula is shown in the following formula (1). (1) Among them, Y real To predict the electrode consumption rate exponent; T t The cumulative heating time is expressed in minutes (L). t The sound level is measured in decibels (dB) for smelting; L0 is the reference noise level (dB); V t , where is the heating rate of molten steel, °C / min; a, b, and c are correction coefficients; K1, K2, and K3 are weighting coefficients; S3: Calculate Y in real time real Compared with the preset warning threshold Y limit Compare; S4: If Y > Y limit Then, the dominant factor is determined according to the values ​​of each item in the model formula (1), and the process is adjusted according to the judgment result.

[0008] Furthermore, in step S2, the values ​​of the correction coefficients a, b, and c are obtained based on experience, while K1, K2, and K3 are obtained by fitting historical data using a multiple linear regression algorithm.

[0009] Furthermore, the correction coefficients a, b, and c are set to 0.03, b, and 4.0 respectively.

[0010] Furthermore, in step S4, if the dominant factor for determination is the cumulative heating time, then the electrode is sprayed with cooling, or the electrode is sprayed with cooling and then powered off for heat preservation.

[0011] Furthermore, in step S4, if the dominant factor for determination is the decibel level of the smelting sound, then the amount of foaming agent sprayed and / or the secondary voltage level is increased.

[0012] Furthermore, in step S4, if the dominant factor is the heating rate of the molten steel, then the input power and / or the intensity of argon blowing and stirring should be increased.

[0013] Furthermore, in step S3, the preset warning threshold Y limit Take 2.5 to 3.0.

[0014] The beneficial effects of adopting the above technical solution are as follows: This invention establishes an electrode consumption rate prediction model by real-time acquisition of heating time, smelting sound decibels and heating rate, realizing real-time monitoring and early warning of electrode consumption, thereby optimizing the power supply system and reducing electrode consumption; This invention realizes real-time visualization and early warning of electrode consumption status, solves the problem of lagging traditional manual experience control, and effectively reduces the electrode consumption cost of LF refining process.

[0015] This invention achieves multi-dimensional accurate prediction: It breaks through the limitations of single-parameter control by integrating three parameters: T (representing physical time), L (representing slag state), and V (representing energy efficiency). The model established can more comprehensively and accurately reflect the actual state of electrode consumption.

[0016] This invention enables real-time closed-loop control: it can calculate the electrode consumption rate in real time and compare it with a threshold, realizing the transformation from "post-event statistical consumption" to "process control consumption", effectively reducing ineffective consumption.

[0017] This invention enables targeted intervention: the model can not only predict the rate of consumption, but also identify the specific reasons for high consumption (such as poor arc burial or low thermal efficiency) by analyzing the contribution rate of each factor, thereby guiding the equipment to implement precise intervention measures and avoiding blind operation. Detailed Implementation

[0018] The present invention will now be described in further detail with reference to specific embodiments.

[0019] This LF refining electrode consumption control method includes the following steps: S1: During the LF refining process, acoustic sensors placed above the LF furnace cover or inside the transformer room collect smelting noise signals in real time. After filtering, the smelting sound level in decibels (L) is obtained. t The sampling frequency for smelting sound decibels should not be lower than 10Hz; the temperature of molten steel should be obtained through thermocouples or temperature measuring guns, and the heating rate V of molten steel should be calculated. t The cumulative heating time T for the current furnace cycle is recorded by the control system. t .

[0020] S2: Input the above parameters into the electrode consumption rate prediction model. The model formula is shown in the following formula (1). (1) Among them, Y real To predict the electrode consumption rate exponent; T t The cumulative heating time is expressed in minutes (L). t The sound level during smelting is expressed in decibels (dB); L0 is the reference noise level, taken as 80dB; V tK1 represents the steel heating rate, in °C / min; a is the time correction coefficient, empirically set at 0.03; b is the noise correction coefficient, empirically set at 0.12; c is the heating rate correction coefficient, empirically set at 4.0; K1 is the time weight coefficient, obtained by fitting historical data using a multiple linear regression algorithm; K2 is the noise weight coefficient, obtained by fitting historical data using a multiple linear regression algorithm; K3 is the heating rate weight coefficient, obtained by fitting historical data using a multiple linear regression algorithm. The weight coefficients K1, K2, and K3 are based on machine learning algorithms using neural networks or support vector machines (SVM), and are periodically updated using the latest production data through model self-learning.

[0021] In the formula (1): , is a time-dependent factor function, which exhibits an exponential or polynomial relationship that increases with heating time, characterizing the intensification of oxidation on the electrode side. , is a sound influence factor function, which shows an increasing relationship with the increase of decibels, characterizing the end sublimation caused by poor arc burial effect; , is the heating rate influence factor function, which shows an increasing relationship as the heating rate decreases, characterizing the ineffective losses caused by low thermal efficiency. S3: Calculate Y in real time real Compared with the preset warning threshold Y limit Comparison with the preset warning threshold Y limit Take 2.5 to 3.0.

[0022] If Y real ≤Y limit The system is determined to be in the normal power consumption range, and the current power supply system is maintained. If Y real >Y limit If the system is determined to be in a high-consumption abnormal range, an alarm is triggered and step S4 intervention strategy is executed.

[0023] S4: If Y real >Y limit Then, based on the values ​​of each component in the model formula (1), the dominant factor is determined. The determination process is as follows: calculate the time influence factor function, sound influence factor function, and heating rate influence factor function respectively, and compare them with the predicted electrode consumption rate index Y. real The ratio, that is, calculate separately , The factor function corresponding to the largest ratio is the dominant factor.

[0024] S5: Adjust the process based on the results of the identified dominant factors: If the dominant factor in the determination is the cumulative heating time, then the heating time T t The largest contribution rate is due to overheating and oxidation caused by prolonged heating. In this case, the electrode spray cooling device is activated to enhance cooling. Electrode spray cooling + power-off heat preservation can also be performed. The water volume control range of the spray cooling is 5-15 L / min. The power-off heat preservation time is 3-8 minutes.

[0025] If the dominant factor in the determination is the decibel level of the smelting sound, then the sound level is in decibels (L). t The largest contribution rate is due to poor arc submersion. To achieve short arc submersion, increase the amount of foaming agent sprayed and / or reduce the secondary voltage level. The amount of foaming agent added is increased to 0.8-1.5 kg / t to ensure arc submersion. The secondary voltage is reduced to 200-500 V, and different levels can be adjusted according to different equipment.

[0026] If the dominant factor in the determination is the heating rate of the molten steel, then the heating rate V t The largest contribution rate is due to low thermal efficiency, so the input power of the LF furnace should be increased and / or the argon blowing and stirring intensity should be increased; the LF power should be increased to 10-20MW, and the argon blowing and stirring intensity should be increased to .

[0027] In the following embodiments, K1 is 1.0; K2 is 1.0; K3 is 1.0, and the above formula (1) can be simplified to: .

[0028] Example 1: Taking the 120-ton LF refining furnace of a steel plant for smelting Q345B steel as an example for illustration.

[0029] S1: Heating to the 18th minute, cumulative heating time T t =18min; Acoustic sensor detection, smelting sound decibels L t =93dB, baseline value 80dB; at this time, the steel heating rate .

[0030] S2: Substitute the above parameters into model formula (1) to calculate Y. real =3.39. Analyzing the contributions of each factor, the sound decibel level accounts for 45.98%.

[0031] S3: Preset warning threshold Y limit Take 2.8, Y real >Y limit The system was determined to be in a high-consumption abnormal range.

[0032] S4: The dominant factor is determined to be the decibel level of smelting noise, caused by poor arc submersion; therefore, a foaming agent should be added to the ladle. Meanwhile, the secondary voltage is controlled at 500V.

[0033] Effect verification: 3 minutes after adjustment, L t The value dropped to 83 Db, and Y dropped to 2.19. The model predicted that electrode consumption decreased by 21.7% during this period.

[0034] Example 2: Taking a 100-ton LF refining furnace in a steel plant as an example to illustrate the process of smelting 45# steel.

[0035] S1: Heating to the 22nd minute, cumulative heating time T t =22min; Smelting sound decibels L t =82Db, which is within the normal range; at this time, the steel heating rate is... The value is significantly lower than expected.

[0036] S2: Substitute the above parameters into model formula (1) to calculate Y. real =3.51. Analysis revealed that the heating rate term contributed 63.28%.

[0037] S3: Preset warning threshold Y limit Take 2.5, Y real >Y limit The system was determined to be in a high-consumption abnormal range.

[0038] S4: The dominant factor is determined to be the steel heating rate, and the dominant cause is low thermal efficiency. Therefore, it is determined that the arc power is not effectively absorbed by the molten steel, and the bottom-blown argon flow rate should be increased to... To improve molten pool stirring and increase input power to 10MW.

[0039] Effect verification: After 5 minutes, V t The temperature was raised to 3.5℃ / min, and Y dropped to 2.43, thus avoiding prolonged dry burning of the electrode in an inefficient state.

[0040] Example 3: Taking the long-cycle smelting of high-grade pipeline steel in a 150-ton LF refining furnace of a steel plant as an example.

[0041] S1: Heating to the 40th minute, cumulative heating time T t =40min; Smelting sound decibels L t =85dB; Steel heating rate .

[0042] S2: Substitute the above parameters into model formula (1) to calculate Y. real =4.02. Analysis revealed that the time term contributed 51.91%, which is because as the heating time increases, the electrode surface temperature becomes extremely high, and the lateral oxidation consumption increases exponentially.

[0043] S3: Preset warning threshold Y limit Take 3.0, Yreal >Y limit The system was determined to be in a high-consumption abnormal range.

[0044] S4: The dominant factor is determined to be cumulative heating time, and the dominant cause is continuous high-temperature oxidation. Therefore, the electrode spray cooling system is activated, with a water flow rate of [missing information]. Simultaneously, a power-off and heat preservation operation is performed for 8 minutes to reduce electrode heat accumulation.

[0045] Effect verification: After power was restored, the electrode surface temperature was controlled, and the average consumption rate decreased by 25.3% in the following 20 minutes.

[0046] Example 4: Taking the 80-ton LF refining furnace of a steel plant for smelting spring steel as an example.

[0047] S1: Heating to the 25th minute, cumulative heating time T t =25min; Smelting sound decibels (L) t =90dB, which is too high; molten steel heating rate The value is too low.

[0048] S2: Substitute the above parameters into model formula (1) to calculate Y. real =4.23. Analysis revealed that the heating rate contributed 42.95%, making it the dominant factor.

[0049] S3: Preset warning threshold Y limit Take 2.8, Y real >Y limit The system was determined to be in a high-consumption abnormal range.

[0050] S4: The dominant factor is determined to be the steel heating rate, and the dominant cause is low thermal efficiency. Therefore, it is determined that the arc power is not effectively absorbed by the molten steel, and the bottom-blown argon flow rate should be increased to... To improve molten pool stirring and increase input power to 20MW.

[0051] Effect verification: After the slag melts, L t Dropped to 80dB, V t Rise to Y quickly fell back to 2.22.

[0052] Example 5: Taking the smelting of gear steel in a 60-ton LF refining furnace of a steel plant as an example.

[0053] S1: Heating for 8 minutes, cumulative heating time T t =8min; Smelting sound decibels L t =92dB; Steel heating rate .

[0054] S2: This furnace is a small-capacity furnace with a fast production pace and short processing time per furnace. Small furnaces are more sensitive to submerged arc. Substituting the above parameters into model formula (1), Y is calculated. real =2.57; of which the sound decibels accounted for 56%.

[0055] S3: Preset warning threshold Y limit Take 2.5, Y real >Y limit The system was determined to be in a high-consumption abnormal range.

[0056] S4: If the dominant factor is determined to be the decibel level of smelting sound, primarily driven by noise, then a foaming agent should be added to the ladle. At the same time, the secondary voltage is controlled at 200V.

[0057] Performance verification: The model responded quickly. After adjustment, Lt dropped to 84Db and Y dropped to 1.61 within 3 minutes. The model predicted that the electrode consumption during this period was reduced by 37%.

[0058] Example 6: Taking the smelting of cord steel in a 150-ton LF refining furnace of a steel plant as an example.

[0059] S1: Heating to the 35th minute, cumulative heating time T t =35min; Smelting sound decibels (L) t =84dB; Steel heating rate .

[0060] S2: Substitute the above parameters into model formula (1) to calculate Y. real =3.41. Analysis revealed that the time term contributed 52.45%, indicating a long heating time and extremely high electrode surface temperature.

[0061] S3: Preset warning threshold Y limit Take 3.0, Y real >Y limit The system was determined to be in a high-consumption abnormal range.

[0062] S4: The dominant factor is determined to be cumulative heating time, and the dominant cause is continuous high-temperature oxidation. Therefore, the electrode spray cooling system is activated, with a water flow rate of [missing information]. Simultaneously, a power-off and heat preservation operation is performed for 3 minutes to reduce the heat accumulation on the electrodes.

[0063] Effect verification: After power was restored, the electrode surface temperature was controlled, and the average consumption rate decreased by 12% in the following 30 minutes.

[0064] Example 7: Taking the smelting of high-speed rail steel in a 120-ton LF refining furnace of a steel plant as an example.

[0065] S1: 30 minutes of heating, total heating time T t =30min; Smelting sound decibels (L) t =82dB; Steel heating rate .

[0066] S2: Substitute the above parameters into model formula (1) to calculate Y. real =2.79.

[0067] S3: Preset warning threshold Y limit Take 3.0, Y real <Y limit The system is determined to be in the normal consumption range.

[0068] Results verification: The process control was normal and no adjustments were required, ensuring the applicability of the model throughout the entire lifecycle.

[0069] Example 8: Taking a 100-ton LF refining furnace in a steel plant as an example to illustrate the process of smelting industrial pure iron.

[0070] S1: Heating for 35 minutes, cumulative heating time T t =35min; Smelting sound decibels (L) t =80dB; Steel heating rate .

[0071] S2: Substitute the above parameters into model formula (1) to calculate Y. real =2.93.

[0072] S3: Preset warning threshold Y limit Take 3.0, Y real <Y limit The system is determined to be in the normal consumption range.

[0073] Results verification: The process control was normal and no adjustments were required, ensuring the applicability of the model throughout the entire lifecycle.

[0074] As can be seen from the above eight embodiments, this control method can flexibly monitor electrode consumption in LF refining and implement early warning and response measures. Compared with the traditional fixed-ratio process, this method solves the problem of lagging traditional manual experience-based control by providing real-time visualization and early warning of electrode consumption status, effectively reducing the electrode consumption cost of the LF refining process.

Claims

1. A method for controlling the consumption of LF refining electrodes, characterized in that, Includes the following steps: S1: Real-time acquisition of the cumulative heating time T during the LF refining process. t , Smelting sound decibels L t and the heating rate V of molten steel t ; S2: Input the above parameters into the electrode consumption rate prediction model. The model formula is shown in the following formula (1). （1） Among them, Y real To predict the electrode consumption rate exponent; T t The cumulative heating time is expressed in minutes (L). t The sound level is measured in decibels (dB) for smelting; L0 is the reference noise level (dB); V t , where is the heating rate of molten steel, °C / min; a, b, and c are correction coefficients; K1, K2, and K3 are weighting coefficients; S3: Calculate Y in real time real Compared with the preset warning threshold Y limit Compare; S4: If Y > Y limit Then, the dominant factor is determined according to the values ​​of each item in the model formula (1), and the process is adjusted according to the judgment result.

2. The method for controlling the consumption of LF refining electrodes according to claim 1, characterized in that: In step S2, the values ​​of the correction coefficients a, b, and c are obtained based on experience, while K1, K2, and K3 are obtained by fitting historical data using a multiple linear regression algorithm.

3. The method for controlling the consumption of LF refining electrodes according to claim 2, characterized in that: The correction coefficients a, b, and c are set to 0.03, 0.12, and 4.0, respectively.

4. The method for controlling the consumption of LF refining electrodes according to claim 1, characterized in that: In step S4, if the dominant factor is the cumulative heating time, then the electrode is sprayed with cooling, or the electrode is sprayed with cooling and then the power is cut off to keep it warm.

5. The method for controlling the consumption of LF refining electrodes according to claim 1, characterized in that: In step S4, if the dominant factor is the decibel level of the smelting sound, then increase the amount of foaming agent sprayed and / or decrease the secondary voltage level.

6. The method for controlling the consumption of LF refining electrodes according to claim 1, characterized in that: In step S4, if the dominant factor is the heating rate of molten steel, then the input power and / or the intensity of argon blowing and stirring should be increased.

7. A method for controlling the consumption of LF refining electrodes according to any one of claims 1-6, characterized in that: In step S3, the preset warning threshold Y limit Take 2.5 to 3.0.

Images