A method and system for automatic control of power of an electric furnace for melting iron

Abstract

This invention belongs to the field of metallurgical technology, specifically relating to an automatic power control method and system for electric furnaces used in iron melting. The method includes: collecting multi-dimensional electric furnace data; constructing state vectors from the multi-dimensional state data; obtaining a state transition index by the distance between state vectors; constructing a temperature-degraded state vector based on the state vectors; obtaining a power adjustment factor based on the temperature difference between the temperature-degraded state vector and the temperature vectors, and the state transition index; adjusting PID parameters based on the power adjustment factor; and controlling the electric furnace power through the adjustment parameters. This invention calculates the adjustment factor based on temperature differences, state similarity, and stage changes to dynamically adjust the PID parameters, achieving adaptive power control.

Description

Technical Field

[0001] This invention relates to the field of metallurgical technology. More specifically, this invention relates to an automatic power control method and system for electric furnaces used in molten iron. Background Technology

[0002] In the production of metal pipe connectors, the electric furnace is the core equipment for melting the furnace charge and heating the molten iron. The control level of the electric furnace power directly determines the temperature uniformity and stability of the molten iron, thus affecting the performance of the final product. Therefore, the control of the electric furnace power is crucial in the production of metal pipe connectors.

[0003] In related technologies, for example, Chinese patent document CN101968318B discloses an automatic control method for the power of an electric furnace. This method includes: setting a power value for the electric furnace and calculating a resistance setting value corresponding to each of the three electrodes based on the set power value; detecting the current insertion depth of the three electrodes into the electric furnace and detecting the current current and voltage values ​​corresponding to each of the three electrodes; calculating the current resistance value of each electrode based on the detected current current and voltage values; searching a preset expert reasoning table based on the current resistance value and the resistance setting value to determine the adjustment range value of the three electrodes; and adjusting the insertion depth of the three electrodes into the electric furnace based on the current insertion depth and the adjustment range value to ensure that the electric furnace reaches the set power value.

[0004] However, the pre-set expert inference table is static; its rules and parameters remain fixed after being set. The electric furnace smelting process includes multiple dynamically changing stages such as preheating, melting, heating, and refining, each with different system characteristics and control requirements. A fixed set of control rules is insufficient to adapt to the nonlinear and time-varying characteristics of the entire smelting process. Therefore, this method has limited adaptability in dealing with complex operating conditions and process stage transitions, making it difficult to achieve optimal control throughout the entire process. Summary of the Invention

[0005] To address the technical problem that the fixed expert reasoning tables mentioned above are difficult to adapt to the dynamic changes in the entire smelting process, thereby affecting the iron smelting process, the present invention provides solutions in the following aspects.

[0006] In a first aspect, the present invention provides an automatic power control method for an electric arc furnace for molten iron, comprising: constructing a state vector using multidimensional electric arc furnace data; clustering historical state vectors using a clustering algorithm to obtain cluster centers for each category of historical state vectors; taking any given time as the target time and the time preceding the target time as the reference time; obtaining a state transition index based on the difference between the state distance at the target time and the state distance at the reference time, and the difference between the reference distance and the state distance at the target time; wherein the state distance is the distance between the cluster center closest to the state vector and the state vector; the reference distance is the distance between the cluster center closest to the state vector at the reference time and the state vector at the target time; obtaining a power adjustment factor based on the similarity between the temperature-removed state vector at the target time and the temperature-removed reference vector, the difference in temperature parameters between the state vector at the target time and historical state vectors, and the state transition index; wherein the temperature-removed state vector is the vector obtained after removing the temperature parameters from the state vector; the temperature-removed reference vector is the temperature-removed state vector of all state vectors corresponding to the cluster center closest to the state vector at the target time; and adjusting the basic PID parameters using the power adjustment factor to obtain the adjusted PID parameters.

[0007] This invention collects multi-dimensional data such as temperature, current, and charge weight, and uses a clustering algorithm to dynamically divide the smelting process into stages, enabling real-time sensing of whether the electric furnace is in different stages such as preheating, melting, or heating. It generates a dynamic power adjustment factor by comprehensively considering three dimensions: the difference between real-time and normal temperatures, the similarity of non-temperature states, and the state transition index representing stage changes. This power adjustment factor is then used to adjust the PID control parameters in real time. This overcomes the shortcomings of existing technologies where fixed parameters cannot adapt to the nonlinearity and time-varying nature of the electric furnace smelting process. It effectively avoids control lag, temperature overshoot, or oscillation caused by stage switching or changes in operating conditions, thus achieving more precise and efficient power control throughout the smelting process.

[0008] Preferably, the multidimensional electric furnace data includes: temperature, current, and weight of the furnace charge.

[0009] In this invention, the state vector includes parameters such as temperature, current, and charge weight. Compared with existing technologies that mainly rely on single electrical parameters such as resistance, voltage, and current, the introduction of charge weight as an indicator allows the electric furnace power control system to more accurately grasp the melting state and energy demand of the material inside the furnace. This enables the constructed state vector to more comprehensively and realistically reflect the physical process of electric furnace heating, providing a more reliable data foundation for subsequent dynamic stage division and precise control.

[0010] Preferably, the clustering algorithm is the K-means clustering algorithm.

[0011] This invention employs a clustering algorithm to cluster historical data, objectively identifying typical operating stages with significant differences during the smelting process and forming cluster centers. This avoids uncertainties caused by subjective factors in stage division, improving the robustness and accuracy of the entire electric furnace power control system.

[0012] Preferably, the state transition index satisfies the following relationship: In the formula, For electric furnace at all times State transition index, For a moment The Euclidean distance between the state vector and its nearest cluster center For a moment The Euclidean distance between the state vector and its nearest cluster center For a moment State vector and time The Euclidean distance between the nearest cluster centers of the state vectors. To prevent division by zero parameters.

[0013] This invention obtains a state transition index by quantifying the change in the distance between the current state vector and the previous state vector and their respective nearest cluster centers. This state transition index can sensitively capture changes in the electric furnace's operating state. It provides a clear early warning signal for the electric furnace power control system, enabling the system to adjust its control strategy quickly and proactively, thus achieving a rapid transition between the molten iron stage and the furnace.

[0014] Preferably, the similarity between the target time-de-temperature state vector and the de-temperature reference vector is a cosine similarity.

[0015] Preferably, the power adjustment factor satisfies the following relationship: In the formula, For electric furnace at all times The power adjustment factor, For electric furnace at all times State transition index, For a moment Temperature parameters of the state vector For a moment The state vector of the th Temperature parameters of a reference vector, For a moment temperature state vector and time The state vector of the first Cosine similarity between temperature reference vectors For a moment The number of reference vectors for the state vector.

[0016] Preferably, the power adjustment factor satisfies the following relationship: In the formula, For electric furnace at all times The power adjustment factor, For electric furnace at all times State transition index, For a moment Temperature parameters of the state vector For a moment The state vector of the th Temperature parameters of a reference vector, For a moment temperature state vector and time The state vector of the first Cosine similarity between temperature reference vectors For a moment The number of reference vectors for the state vector. To correct the degree adjustment parameters, It is an exponential function with the natural constant as the base.

[0017] Preferably, the basic PID parameters are obtained by means of obtaining the basic PID parameters used in the electric furnace iron melting process through the Ziegler-Nichols experimental method.

[0018] Preferably, the adjusted PID parameters satisfy the following relationship: ; ; In the formula, , and They are time points Adjustment of use Parameters, Adjustment Parameters and adjustments parameter, , and They are respectively Parameter base value Parameter base values ​​and Base value of parameter For electric furnace at all times The power adjustment factor, , , They are respectively , , The sensitivity coefficient of parameter adjustment.

[0019] This invention provides a basis for adaptive updating of PID parameters by dynamically adjusting the basic parameters using dynamically calculated power adjustment factors and sensitivity coefficients of each PID parameter. This makes the magnitude of parameter adjustment directly related to the degree to which the electric furnace heating system deviates from its normal state and the severity of stage transitions, achieving adaptive adjustment of control intensity and improving the timeliness of electric furnace control.

[0020] Secondly, the present invention provides an automatic power control system for an electric furnace used for molten iron, comprising a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned automatic power control method for an electric furnace used for molten iron is implemented.

[0021] By adopting the above technical solution, a computer program for the automatic power control method of an electric furnace used for molten iron is generated and stored in a memory so that it can be loaded and executed by a processor. A terminal device can then be made based on the memory and the processor for convenient use.

[0022] The beneficial effects of this invention are as follows: This invention achieves dynamic stage division of the smelting process through a clustering algorithm and constructs a power adjustment factor. The power adjustment factor integrates the temperature difference between real-time and historical normal temperatures, the similarity between real-time and historical normal states, and the stage transition index of the smelting process, enabling a comprehensive and dynamic assessment of the electric furnace's real-time status. Applying the power adjustment factor to the parameters of the PID algorithm allows power control to intelligently adapt to the time-varying characteristics of the smelting process, overcoming the shortcomings of traditional PID algorithms and static expert inference tables in adapting to the dynamic changes in the iron-melting process, and achieving efficient control of the electric furnace iron-melting process. Attached Figure Description

[0023] Figure 1 This is a flowchart illustrating an automatic power control method for an electric furnace used for melting iron according to the present invention.

[0024] Figure 2 This is a schematic diagram illustrating a comparison between the power control process of the present invention and the power control process of existing algorithms. Detailed Implementation

[0025] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0026] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0027] This invention discloses an automatic power control method for electric furnaces used in molten iron, referring to... Figure 1 This includes steps S1-S4:

[0028] S1. Collect multi-dimensional electric furnace data, construct a state vector through the multi-dimensional state data, and obtain the state transition index through the distance between the state vectors.

[0029] It should be noted that the electric furnace iron melting process has multiple stages, such as preheating, melting, heating and refining. The power requirements of different stages of the iron melting process are different. Therefore, when controlling the power at different stages, it is necessary to adapt to the characteristics of each stage. Therefore, this invention constructs a state vector through multi-dimensional electric furnace data and obtains the state transition index by the distance between the state vectors.

[0030] Specifically, parameters such as temperature, current, and weight of the furnace charge are collected by temperature sensors, current sensors, and the furnace charge bearing system. Each parameter is normalized, and the normalized values ​​of temperature, current, and weight of the furnace charge at each moment are used to construct a state vector for each moment. Historical state vectors at each moment during normal historical iron melting are collected, and clustering algorithms are used to cluster the historical state vectors into K classes, obtaining the cluster center for each class.

[0031] For example, parameters from 100 batches of normal historical iron melting processes are collected to construct state vectors for each historical moment.

[0032] In one embodiment, the clustering algorithm is the K-means clustering algorithm, wherein the K value used in the K-means clustering algorithm is 4, and the implementer can determine the value of K according to the actual situation.

[0033] Furthermore, during the real-time power control of the electric furnace, the data on the electric furnace's power output at various times is collected. Parameter construction time The state vector, obtaining the time step The Euclidean distance between the state vector and each cluster center; measured at distance time... The nearest cluster center and time of the state vector The Euclidean distance between the state vectors is at time 1 State distance; distance time The nearest cluster center and time of the state vector The Euclidean distance between the state vectors is at time 1 The reference distance is used to obtain the state transition index based on the difference between the state distances at adjacent time points and the difference between the reference distance and the state distance.

[0034] Specifically, the state transition index satisfies the following relation:

[0035] ;

[0036] In the formula, For electric furnace at all times State transition index, For a moment The Euclidean distance between the state vector and its nearest cluster center For a moment The Euclidean distance between the state vector and its nearest cluster center For a moment State vector and time The Euclidean distance between the nearest cluster centers of the state vectors. To prevent division by zero in this example, .

[0037] in, The electric furnace at any time The state vector and at time t The difference between the state vectors at two different times indicates that the larger the value, the greater the difference between the state vectors at two different times, and the more likely the molten iron state of the electric furnace will change. Therefore, the electric furnace at time [time missing]... The larger the state transition exponent, the smaller the value, indicating a smaller difference between the state vectors at two moments, and the more likely the molten iron state of the electric furnace is to remain stable. Therefore, the electric furnace at time [time value missing]... The smaller the state transition index, the better.

[0038] When the electric furnace transitions between different iron-melting stages, the nearest cluster centers of the state vectors at adjacent time points will change, resulting in different nearest cluster centers for the state vectors at different time points. The distance between the state vector and the nearest cluster center of the state vector at different times is different, therefore, by... right An upward correction is performed when the nearest cluster center of the state vector changes. If it is greater than 0, then ,at this time right This serves as an upward correction, resulting in a larger state transition exponent for the electric furnace during the molten iron stage transition; when the nearest cluster center of the state vector remains unchanged, If it is 0, then The value is 1, which is incorrect at this point. Make corrections.

[0039] S2. Construct a temperature-reduced state vector based on the state vector, and obtain the power adjustment factor based on the temperature difference between the temperature-reduced state vector and the temperature vector, as well as the state transition exponent.

[0040] It should be noted that when the electric furnace transitions from one stage to another, insufficient heating or overheating can easily occur if the power regulation response is not timely. Therefore, different power control strategies are required for different stages of the electric furnace iron melting process. To avoid this problem during the electric furnace iron melting process, this invention constructs a temperature-reduced state vector based on the state vector, and obtains the power regulation factor based on the temperature difference between the temperature-reduced state vector, the temperature vector, and the state transition exponent.

[0041] Specifically, the historical state vector is obtained, the temperature parameter is removed from the historical state vector to form the historical de-temperature state vector, and the temperature parameter is removed from the real-time state vector to obtain the real-time de-temperature state vector.

[0042] For example, if the historical state vector is (0.1, 0.2, 0.15), where 0.1 is the temperature parameter in the historical state vector, then the historical temperature state vector is (0.2, 0.15).

[0043] Furthermore, all state vectors corresponding to the nearest cluster center of the real-time state vector are used as reference vectors, and the de-temperature vectors corresponding to the reference vectors are used as de-temperature reference vectors. The similarity between the real-time de-temperature state vector and the de-temperature reference vector is obtained, and the power adjustment factor is obtained based on the similarity between the real-time de-temperature state vector and the de-temperature reference vector, as well as the difference between the temperature parameters of the real-time state vector and the temperature parameters of the historical state vector.

[0044] In one embodiment, the power regulation factor satisfies the following relationship:

[0045] ;

[0046] In the formula, For electric furnace at all times The power adjustment factor, For electric furnace at all times State transition index, For a moment Temperature parameters of the state vector For a moment The state vector of the th Temperature parameters of a reference vector, For a moment temperature state vector and time The state vector of the first Cosine similarity between temperature reference vectors For a moment The number of reference vectors for the state vector. In order to correct the degree adjustment parameters, in this embodiment The value is 1.5, and implementers can choose according to the actual situation. The value, It is an exponential function with the natural constant as the base.

[0047] in, This value represents the difference between the real-time temperature of the electric furnace and its historical normal temperature. A larger value indicates that the furnace's molten iron state deviates more from the historical normal temperature, and therefore the furnace power should be adjusted to bring the molten iron state back to the historical normal temperature. The larger the power adjustment factor, the closer the furnace's molten iron state is to the historical normal temperature; thus, the less adjustment of the furnace power is needed, and the more stable the furnace becomes at any given time. The smaller the power regulation factor, the better.

[0048] The larger the value, the more likely it is to indicate the time. The more similar the de-temperature state vector is to the de-temperature reference vector, the more similar the real-time parameters of the electric furnace are to the parameters under historical normal conditions, indicating that the electric furnace is at the same temperature at time... The closer the heating state of the electric furnace is to its historical normal heating state; The smaller the value, the more accurate the time. The greater the difference between the de-temperature state vector of the electric furnace and the de-temperature state vector of the reference vector, the better the electric furnace is at time [time value missing]. The greater the difference between the electric furnace's heating state and its historical normal heating state, the more likely it is to cause problems. When there is a difference between the current electric furnace temperature and the historical normal temperature, continuing to heat the furnace under the historical normal heating conditions may delay the furnace temperature from returning to the historical normal temperature. In this case, it is necessary to adjust the furnace heating conditions to allow the furnace temperature to return to the historical normal temperature more quickly. Therefore, by... right Make an upward correction when When it is larger, The smaller the value, the greater the upward correction, resulting in a larger power regulation factor, which allows the furnace temperature to return to the historical normal temperature more quickly when it deviates from the historical normal temperature. The smaller, The larger the value, the smaller the upward correction. In this case, excessive correction is not needed, and the electric furnace temperature can quickly return to the historical normal temperature when it deviates from the historical normal temperature. In the formula, 1 is added and 2 is divided to simplify the calculation. Normalization was performed, and parameters were adjusted by modifying the degree of correction. adjust The range of values ​​for makes It plays a better corrective role; because and It has been normalized, therefore The range is between 0 and 1. The range is between 0 and 1. According to the principle of gamma transformation, at this time... The smaller right The greater the upward correction; The larger right The smaller the upward correction.

[0049] State transition index The larger the value, the more likely the electric furnace iron melting process is to be transitioning between different stages. In order to facilitate the transition of the electric furnace iron melting process, the power adjustment factor should be larger. The smaller the value, the less likely the electric furnace iron melting process is to be in a transitional phase between different stages. The power adjustment factor should be as small as possible to maintain the stability of the electric furnace iron melting process.

[0050] In another embodiment, the power adjustment factor satisfies the following relationship:

[0051] ;

[0052] In the formula, For electric furnace at all times The power adjustment factor, For electric furnace at all times State transition index, For a moment Temperature parameters of the state vector For a moment The state vector of the th Temperature parameters of a reference vector, For a moment temperature state vector and time The state vector of the first Cosine similarity between temperature reference vectors For a moment The number of reference vectors for the state vector.

[0053] In this embodiment, linear correction is used to... The corrections are made to make the power regulation factor change more stably, thus making the control of the electric furnace power more stable.

[0054] S3. Adjust the PID parameters according to the power adjustment factor.

[0055] It should be noted that the Proportional-Integral-Differential (PID) algorithm is a commonly used control algorithm. Power control of electric furnaces can be achieved through the PID algorithm. However, when using the PID algorithm for power control, only fixed PID parameters are used, which cannot adapt to the changes in power demand during the electric furnace molten iron process. Existing technologies for electric furnace power control using the PID algorithm are difficult to meet the needs of the electric furnace molten iron process. Therefore, this invention adjusts the PID parameters according to the power adjustment factor to control the electric furnace power.

[0056] In one embodiment, the method used in the electric furnace iron melting process is obtained through the Ziegler-Nichols experimental method. Parameter base value Parameter base values ​​and The baseline parameter values ​​are used to adjust the PID parameters using a power adjustment factor to obtain the adjusted values. Parameters, Adjustment Parameters and adjustments parameter.

[0057] Specifically, adjustment Parameters, Adjustment Parameters and adjustments The parameters satisfy the following relationship:

[0058] ;

[0059] ;

[0060] ;

[0061] In the formula, , and They are time points Adjustment of use Parameters, Adjustment Parameters and adjustments parameter, , and They are respectively Parameter base value Parameter base values ​​and Base value of parameter For electric furnace at all times The power adjustment factor, , , They are respectively , , The parameter adjustment sensitivity coefficient, in this embodiment For 1, 0.8 The sensitivity coefficient is set to 0.5, and implementers can adjust it according to the actual situation.

[0062] in, The larger the value, the greater the difference between the electric furnace temperature and the historical normal state, or the period of transition. A larger PID parameter should be used to control the electric furnace power in order to smoothly switch the molten iron state of the electric furnace. The smaller the value, the more likely the electric furnace is to be in a stable state. Therefore, a smaller PID parameter should be used to control the furnace power to avoid unstable power fluctuations; hence, a power regulation factor is used. Adjust the base values ​​of the PID parameters.

[0063] S4, by adjustment Parameters, Adjustment Parameters and adjustments The parameters are used to control the power of the electric furnace.

[0064] Specifically, during the electric furnace iron melting process, the adjustments at each moment are obtained. Parameters, Adjustment Parameters and adjustments Parameters are used to obtain the target power and real-time power at each time step, and adjustments are made at each time step. Parameters, Adjustment Parameters and adjustments The parameters are controlled by a PID algorithm to achieve the target power for the electric furnace.

[0065] Figure 2 The diagram shows a comparison between the power control process of this invention and the power control process of existing algorithms. In the diagram, the curve obtained by this invention through adjusting the PID parameters for power control is closer to the target power, and the power control method of this invention can respond to changes more quickly when the target power changes.

[0066] This invention also discloses an automatic power control system for an electric furnace used for molten iron, including a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, an automatic power control method for an electric furnace used for molten iron according to the present invention is implemented.

[0067] The system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces, the settings and functions of which are known in the art and will not be described in detail here.

Claims

1. An automatic power control method for electric furnaces used in molten iron, characterized in that, the power of the electric furnace is controlled by a PID algorithm. Adjusting PID parameters through power adjustment factors includes: A state vector is constructed using multidimensional electric furnace data. The historical state vectors are then clustered using a clustering algorithm to obtain the cluster center for each category of historical state vectors. Take any given moment as the target moment and the moment before the target moment as the reference moment; The state transition index is obtained based on the difference between the state distance at the target time and the state distance at the reference time, and the difference between the state distance at the reference time and the state distance at the target time, including: In the formula, For electric furnace at all times State transition index, For a moment The Euclidean distance between the state vector and its nearest cluster center For a moment The Euclidean distance between the state vector and its nearest cluster center For a moment State vector and time The Euclidean distance between the nearest cluster centers of the state vectors. To prevent division by zero, the state distance is the distance between the nearest cluster center to the state vector and the state vector itself; the reference distance is the distance between the nearest cluster center to the state vector at the reference time and the state vector at the target time. The power adjustment factor is obtained based on the similarity between the target time-de-temperature state vector and the de-temperature reference vector, the difference in temperature parameters between the target time state vector and the historical state vector, and the state transition exponent, including: In the formula, For electric furnace at all times The power adjustment factor, For a moment Temperature parameters of the state vector For a moment The state vector of the th Temperature parameters of a reference vector, For a moment temperature state vector and time The state vector of the first Cosine similarity between temperature reference vectors For a moment The number of reference vectors for the state vector; the de-temperature state vector is the vector obtained by removing the temperature parameter from the state vector; the de-temperature reference vector is the de-temperature state vector of all state vectors corresponding to the cluster center closest to the state vector at the target time. The adjusted PID parameters are obtained by adjusting the basic PID parameters through the power adjustment factor.

2. The method for automatic power control of an electric furnace for molten iron according to claim 1, characterized in that, The multidimensional electric furnace data includes: temperature, current, and weight of the furnace charge.

3. The method for automatic power control of an electric furnace for molten iron according to claim 1, characterized in that, The clustering algorithm is the K-means clustering algorithm.

4. The method for automatic power control of an electric furnace for molten iron according to claim 1, characterized in that, The similarity between the target time-de-temperature state vector and the de-temperature reference vector is the cosine similarity.

5. The method for automatic power control of an electric furnace for molten iron according to claim 1, characterized in that, The power adjustment factor satisfies the following relationship: ； In the formula, For electric furnace at all times The power adjustment factor, For electric furnace at all times State transition index, For a moment Temperature parameters of the state vector For a moment The state vector of the th Temperature parameters of a reference vector, For a moment temperature state vector and time The state vector of the first Cosine similarity between temperature reference vectors For a moment The number of reference vectors for the state vector. To correct the degree adjustment parameters, It is an exponential function with the natural constant as the base.

6. The method for automatic power control of an electric furnace for molten iron according to claim 1, characterized in that, The method for obtaining the basic PID parameters includes obtaining the basic PID parameters used in the electric furnace iron melting process through the Ziegler-Nichols experimental method.

7. The method for automatic power control of an electric furnace for molten iron according to claim 1, characterized in that, The adjusted PID parameters satisfy the following relationship: ； ； ； In the formula, , and They are time points Adjustment of use Parameters, Adjustment Parameters and adjustments parameter, , and They are respectively Parameter base value Parameter base values ​​and Base value of parameter For electric furnace at all times The power adjustment factor, , , They are respectively , , The sensitivity coefficient of parameter adjustment.

8. An automatic power control system for an electric furnace used for molten iron, characterized in that, include: A processor and a memory, the memory storing computer program instructions that, when executed by the processor, implement an automatic power control method for an electric furnace for molten iron according to any one of claims 1-7.

Images