Arc furnace flat bath smelting stage temperature feedback control method and system

Abstract

This invention relates to the field of data processing technology, specifically to a temperature feedback control method and system for the smelting stage of an electric arc furnace flat-walled molten pool. The method periodically extracts the arc force index and vibration response index for each channel based on the current and vibration signals of each channel; stores these two data pairs in a historical data queue and updates it; for each channel, it determines whether the dispersion of the arc force index within the queue meets the validity condition; if so, it fits the vibration force transmission ratio based on all data pairs in the queue; it determines the reliability of the vibration signal based on the arc force index and vibration response index within the queue; it determines the real-time electrical energy thermal efficiency based on preset efficiency parameters, transmission ratio, and reliability; it corrects the real-time electrical power based on the real-time electrical energy thermal efficiency, and determines the net heating power of the molten steel by combining the chemical reaction heat power and heat loss power; it performs cumulative integration based on the net heating power of the molten steel to determine the predicted temperature of the molten steel, thereby improving the accuracy of the predicted temperature.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, specifically to a temperature feedback control method and system for the smelting stage of an electric arc furnace flat-bottomed pool. Background Technology

[0002] In electric arc furnace steelmaking, the flat-melt pool smelting stage is a crucial step determining the final temperature and quality of the molten steel. During this stage, the scrap steel in the furnace has essentially melted, forming a liquid molten pool. Foamed slag submerged arc operation is typically employed to improve thermal efficiency and protect the furnace lining. Precise control of the molten steel temperature is the core objective; steel must be tapped promptly upon reaching the target temperature to avoid energy waste or low-temperature steel accidents.

[0003] Currently, industrial applications primarily rely on energy integral models (also known as heat balance models) to estimate molten steel temperature. These models typically assume that the thermal conversion efficiency of electrical energy is constant. However, such models have the following drawbacks: First, they cannot detect the real-time dynamic drift in thermal efficiency caused by drastic changes in the state of the foamy slag. During the inefficient period of foamy slag collapse (bare arc), they still calculate based on high efficiency, which can easily lead to "false temperature rises" and result in overestimated predicted temperatures, seriously misleading steel tapping decisions. Second, although attempts have been made to introduce vibration signals to monitor slag conditions, they cannot effectively distinguish between effective vibrations caused by changes in arc force and ineffective vibrations caused by mechanical noise such as scrap steel collapse and molten pool boiling. This leads to frequent malfunctions in the control system under strong noise interference. These defects severely restrict the accuracy of temperature prediction and the reliability of process control. Summary of the Invention

[0004] To address the technical problem of inaccurate temperature predictions due to reliance on energy integral models, which severely misleads steel tapping decisions, this invention provides a temperature feedback control method and system for the flat-pool smelting stage of an electric arc furnace. The specific technical solution adopted is as follows: This invention proposes a temperature feedback control method for the smelting stage of an electric arc furnace flat-pool, the method comprising: In response to the smelting entering the flat melting pool stage, the actual temperature of the molten steel is obtained; The current signals and furnace wall vibration signals of each electrode channel of the electric arc furnace are collected synchronously. Based on the current signals and vibration signals of each channel, the arc force amplitude and vibration response amplitude of each channel are periodically extracted, and the arc force index and vibration response index are obtained after standardization. The two data pairs are stored in the historical data queue of each channel in chronological order and updated. For each channel, determine whether the dispersion of the arc force index within the queue meets the validity condition; if it does, fit the vibration force transmission ratio based on all data pairs within the queue; determine the reliability of the vibration signal based on the correlation between the arc force index and the vibration response index within the queue; determine the real-time electrical and thermal efficiency of the channel based on the preset efficiency parameters, transmission ratio, and reliability. The real-time electrical power is corrected based on the real-time electrical energy thermal efficiency, and the net heating power of the molten steel is determined by combining the preset chemical reaction heat power and the preset heat loss power. The predicted temperature of the molten steel is determined by accumulating the net heating power of the molten steel. The predicted temperature is then corrected based on the measured temperature. The smelting endpoint is controlled based on the predicted temperature, and corresponding optimized smelting operations are performed based on the real-time electrical energy thermal efficiency.

[0005] Furthermore, the current signal includes the instantaneous current value of each phase electrode channel acquired by the current transformer; the furnace wall vibration signal includes the instantaneous value of furnace wall vibration acceleration in each region acquired by the vibration acceleration sensor installed on the outer wall of the electric arc furnace shell and corresponding to the physical position of each phase electrode. The process of obtaining the arc force index and vibration response index includes: The instantaneous current values ​​at each acquisition moment in the current signal are squared to obtain the squared current value; based on the squared current values ​​at each acquisition moment, an equivalent arc force signal proportional to the Lorentz force of the arc is generated in chronological order. For each electrode channel, obtain the current square value sequence and vibration acceleration instantaneous value sequence within each cycle; and obtain the preset arc force reference amplitude and preset vibration response reference amplitude; Perform a Fast Fourier Transform (FFT) on the sequence of squared current values ​​within each period to obtain the first spectrum; perform a Fast Fourier Transform (FFT) on the sequence of instantaneous vibration acceleration values ​​within the same period to obtain the second spectrum; Read the complex number corresponding to the preset target frequency from the first spectrum, and take its modulus as the arc force amplitude; read the complex number corresponding to the same target frequency from the second spectrum, and take its modulus as the vibration response amplitude; Divide the arc force amplitude by the preset arc force reference amplitude to obtain the arc force index in each cycle channel; divide the vibration response amplitude by the preset vibration response reference amplitude to obtain the vibration response index in each cycle channel.

[0006] Furthermore, the historical data queue is maintained using a first-in-first-out queue data structure to store the arc force index and vibration response index within each electrode channel cycle, and the queue has a preset maximum length. The process of storing the data pairs (comprising the two data points) into the historical data queues of each channel in chronological order and updating them includes: For each electrode channel, the arc force index and vibration response index of the channel obtained in the current cycle are combined into a new data pair; the new data pair is added to the head of the historical data queue maintained for the channel; it is determined whether the current length of the queue maintained by the channel exceeds the preset maximum length. If the current length of the queue after addition exceeds the preset maximum length, the oldest data pair at the end of the queue is removed; if the number of data pairs in the queue maintained by the channel does not reach the preset maximum length, the channel adopts a preset safety efficiency value.

[0007] Furthermore, determining whether the dispersion of the arc force index within the queue satisfies the validity condition includes: For each electrode channel, calculate the standard deviation of all arc force indices in the queue maintained by the current cycle channel, and use it as a fluctuation index; If the fluctuation index is less than the preset fluctuation threshold, all data pairs in the queue of the current cycle channel are determined to be invalid, and subsequent transmission ratio and reliability calculations are skipped. The preset safety efficiency value is used as the real-time electrical and thermal efficiency of the current cycle channel.

[0008] Furthermore, the process of obtaining the vibration force transmission ratio includes: For each electrode channel, based on the data pairs of all arc force indices and vibration response indices in the queue maintained in the current cycle channel, linear regression is performed using the least squares method to obtain the regression slope of the vibration response index with respect to the arc force index. The regression slope is used as the vibrational force transmission ratio in the current cycle channel.

[0009] Furthermore, the process for determining the reliability of the vibration signal includes: For each electrode channel, based on all data pairs in the queue maintained in the current cycle channel, the arc force index sequence and vibration response index sequence for the current cycle channel are generated in chronological order. Calculate the Pearson correlation coefficient between the arc force index sequence and the vibration response index sequence, and use it as the correlation coefficient; The square of the correlation coefficient is used as the reliability of the vibration signal in the current cycle channel.

[0010] Furthermore, the preset efficiency parameters include a preset efficiency benchmark value and a preset maximum efficiency loss value; The real-time electrical energy thermal efficiency determination process includes: Obtain the credibility threshold; For each electrode channel, a preset S-shaped function is used to map the reliability of the vibration signal in the current cycle channel to a reliability weight between 0 and 1. Specifically, if the reliability of the vibration signal is much lower than the reliability threshold, the mapped reliability weight approaches 0; if the reliability of the vibration signal is much higher than the reliability threshold, the mapped reliability weight approaches 1. The product of the vibration force transmission ratio in the current cycle channel and the preset maximum efficiency loss value is calculated as the efficiency loss amount. The real-time electrical thermal efficiency in the current cycle channel is obtained by subtracting the product of the confidence weight and the efficiency loss from the preset efficiency benchmark value.

[0011] Furthermore, the process of determining the net heating power of the molten steel includes: Acquire the real-time electrical power of each electrode channel; and acquire the preset chemical reaction heat power and preset heat loss power; Multiply the real-time electrical power of each electrode channel by its corresponding real-time electrical energy thermal efficiency to obtain the effective electrical power of each channel. The total effective power is obtained by summing the effective power of all channels. The net heating power of molten steel is obtained by subtracting the preset heat loss power from the sum of the total effective electrical power and the preset chemical reaction heat power.

[0012] Furthermore, the process for determining the predicted temperature of the molten steel includes: Maintain a preset temperature prediction integrator, the status of which includes an initial temperature value and an accumulated heat energy value; The temperature prediction integrator uses the preset molten steel mass and preset specific heat capacity as parameters. It updates the accumulated heat energy value by continuously accumulating the product of the net heating power of the molten steel and time. The accumulated heat energy value is divided by the product of the preset estimated mass of molten steel and the preset specific heat capacity of molten steel to obtain the temperature rise value from the initial time. The temperature rise value is then added to the initial temperature value to obtain the predicted temperature of the molten steel at the current time. The initial temperature value is set to the first measured temperature of the molten steel when the system is started, and is updated to the new measured temperature value each time a new measured temperature of the molten steel is obtained. At the same time, the accumulated heat energy value is reset to zero.

[0013] A temperature feedback control system for the smelting stage of an electric arc furnace flat-bottomed pool includes a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of a temperature feedback control method for the smelting stage of an electric arc furnace flat-bottomed pool.

[0014] The present invention has the following beneficial effects: This invention synchronously acquires current and vibration signals, periodically extracts arc force and vibration response indices, and fits the vibration force transmission ratio based on a data queue composed of these two indices. Correlation analysis is then used to determine the reliability of the vibration signal. Based on this dynamic synthesis of the two-dimensional indices of vibration force transmission ratio and vibration signal reliability, the real-time electrothermal efficiency can accurately capture the instantaneous efficiency decrease caused by foam slag collapse, thereby eliminating invalid vibrations caused by noise in real time. This ensures that the temperature prediction trajectory truly reflects the actual heat absorption process of the molten steel, improving the accuracy of the endpoint temperature hit. Furthermore, by responding to the start of the flattening pool stage, real-time acquisition of molten steel data is achieved. By measuring the actual temperature and combining it with multi-source signals (i.e., current signal and vibration signal) to dynamically calculate the net heating power of molten steel, real-time temperature prediction is achieved. At the same time, the predicted temperature is dynamically corrected based on the measured temperature, which effectively offsets the errors caused by factors such as temperature measurement lag, thermal fluctuations in chemical reactions, and changes in heat loss. This reduces the deviation limit of molten steel temperature prediction and ensures the stability of molten steel composition. The power is dynamically corrected based on real-time electrical energy thermal efficiency, avoiding overheating or underheating caused by inaccurate efficiency estimation. This enables on-demand allocation of electrical energy and further improves arc stability and electrical energy utilization efficiency. Attached Figure Description

[0015] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a flowchart of a temperature feedback control method for the smelting stage of an electric arc furnace flat molten pool, provided in one embodiment of the present invention. Figure 2 An example diagram illustrating the process of obtaining arc force and vibration response indices according to an embodiment of the present invention; Figure 3 This is an example diagram illustrating the real-time electrical energy thermal efficiency determination process provided in one embodiment of the present invention. Detailed Implementation

[0017] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a temperature feedback control method and system for the smelting stage of an electric arc furnace flat-walled pool according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0018] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0019] The following description, in conjunction with the accompanying drawings, details the specific scheme of the temperature feedback control method and system for the smelting stage of an electric arc furnace flat-bottomed pool provided by the present invention.

[0020] Please see Figure 1 The diagram illustrates a flow chart of a temperature feedback control method for the smelting stage of an electric arc furnace flat-pool according to an embodiment of the present invention. The method includes: S101: In response to the smelting entering the flat melting pool stage, the actual temperature of the molten steel is obtained.

[0021] The flat molten pool stage refers to the process stage in which the solid scrap steel in the electric arc furnace has been basically melted away, forming a stable molten pool dominated by liquid steel, and the surface of the molten pool does not show obvious violent churning.

[0022] It should be noted that the specific method for determining whether smelting has entered the flat melting pool stage can be determined by referring to existing industrial experience, and will not be described in detail in this embodiment.

[0023] For example, the system can determine that the smelting has entered the flat melting pool stage and automatically activate the subsequent control process when the cumulative power consumption reaches a preset percentage (e.g., between 75% and 85%, indicating that the scrap steel has basically melted), the feeding operation has ended for more than a set time (usually 2 to 5 minutes), and the current and vibration signals in the furnace tend to stabilize.

[0024] It should be noted that the measured temperature of molten steel obtained in this invention is not a real-time continuous signal, but rather an intermittent measurement performed manually or by a bomb-type temperature measuring device at key process nodes. For example, the measured temperature of molten steel may be a one-time measurement performed by an operator using a temperature measuring gun or a bomb-type temperature measuring device.

[0025] S102: Synchronously acquire the current signal and furnace wall vibration signal of each electrode channel of the electric arc furnace; periodically extract the arc force amplitude and vibration response amplitude of each channel based on the current signal and vibration signal of each channel, and obtain the arc force index and vibration response index after standardization; store the data pairs composed of the two into the historical data queue of each channel in chronological order and update them.

[0026] It is understood that in this invention, each phase electrode channel and each electrode channel have the same meaning, namely, an independently established data acquisition and analysis circuit for each phase power supply electrode of the electric arc furnace (such as the A, B, and C phase electrodes of a three-phase AC furnace). Each channel uniquely corresponds to one phase electrode and synchronously acquires the current signal of the electrode and the vibration signal of its corresponding furnace wall area.

[0027] The corresponding furnace wall area refers to the physical area on the outer wall of the electric arc furnace shell that is adjacent to the insertion position of a certain phase electrode, i.e., the area where the vibration acceleration sensor is installed.

[0028] In this embodiment, the current signal includes the instantaneous current value of each phase electrode channel acquired by the current transformer.

[0029] In this embodiment, the furnace wall vibration signal includes the instantaneous values ​​of furnace wall vibration acceleration in each region, which are collected by vibration acceleration sensors installed on the outer wall of the electric arc furnace shell and corresponding to the physical positions of each phase electrode.

[0030] To ensure that the current signal and the vibration signal are strictly aligned on the time axis, all sensors (such as current transformers and vibration acceleration sensors) can be connected to the same data acquisition system and synchronous sampling can be triggered by a unified high-precision master clock. The system reads and stamps the same timestamp at each sampling moment to ensure that the instantaneous current value at each sampling moment can be accurately correlated with the instantaneous vibration acceleration value at the same moment.

[0031] The process of obtaining the arc force index and vibration response index is as follows: Figure 2 As shown, it includes: S102-1: Squaring the instantaneous current values ​​at each acquisition moment in the current signal to obtain the squared current value; generating an equivalent arc force signal proportional to the Lorentz force of the arc in chronological order based on the squared current values ​​at each acquisition moment.

[0032] It is important to understand that since the mechanical impact force (Lorentz force) of the electric arc on the molten pool is proportional to the square of the current flowing through the arc, it is necessary to perform point-to-point squaring on the collected instantaneous current value to convert the current signal (i.e., the instantaneous current value) into a linearly related mechanical force signal (i.e., the square of the current value). This step is a necessary prerequisite for establishing the correct correspondence between "electrodynamic force" and "vibration response" in the future.

[0033] The square value of the current is physically equivalent to the instantaneous intensity of the Lorentz force of the electric arc. The square value of the current reflects the magnitude of the "force" by which the electric arc impacts the molten pool and furnace wall at any given moment. Specifically, a larger square value at any given moment indicates a stronger mechanical impact force generated by the electric arc.

[0034] S102-2: For each electrode channel, obtain the current square value sequence and vibration acceleration instantaneous value sequence within each cycle; and obtain the preset arc force reference amplitude and the preset vibration response reference amplitude.

[0035] It is understandable that the sequence of current square values ​​can be periodically extracted from the equivalent signal of the electric arc force.

[0036] It should be noted that the specific value of the period is a configurable engineering parameter that can be determined based on the system response speed, and this embodiment does not impose a specific limitation. For example, the typical range of the period is between 0.1 seconds and 1.0 seconds. If a period contains hundreds or thousands of consecutive signal acquisition moments, then within a 0.2-second period, 200 consecutive instantaneous current values ​​and 200 consecutive instantaneous vibration values ​​will be acquired, thus forming a sequence of current square values ​​and a sequence of instantaneous vibration acceleration values ​​in chronological order.

[0037] It should be noted that the specific value of the preset arc force reference amplitude can be determined based on the transformer's nameplate parameters. It is typically calculated by multiplying the square of the rated short-circuit current (or maximum allowable operating current) on the transformer's secondary side by a safety factor greater than 1 (e.g., 1.2 to 1.5). This embodiment does not impose a specific limitation. For example, for an electric arc furnace with a rated short-circuit current of 50kA on its secondary side, the value can be... =3.25×10^9A 2 The preset arc force reference amplitude is taken as the value, where the safety factor is 1.3, which means that if the actual arc force index is close to 1, it means that the arc force intensity is close to 1.3 times the rated short circuit condition.

[0038] The preset vibration response reference amplitude defines a vibration level reference under ideal working conditions, which usually corresponds to the ideal state of "good foam slag coverage and stable electric arc combustion".

[0039] It should be noted that the specific value of the preset vibration response reference amplitude can be determined by statistical analysis of historical production data, and this embodiment does not impose a specific limitation. The specific determination process can be as follows: During the system debugging phase, engineers or operators identify a series of recognized ideal operating conditions, extract the characteristic amplitudes of the vibration signals within these periods, calculate their statistical average or median, and then multiply it by a factor (such as 3 to 5 times, based on empirically calibrated process parameters, which will not be elaborated further) as the vibration response reference amplitude.

[0040] For example, it was found that under good submerged arc conditions, the average amplitude of the 100Hz component of furnace wall vibration was approximately 5m / The vibration response reference amplitude can then be set to 3 × 5 = 15 m / s. 2 (Use 3 times the average as the benchmark).

[0041] S102-3: Perform a Fast Fourier Transform on the sequence of squared current values ​​within each period to obtain the first spectrum; perform a Fast Fourier Transform on the sequence of instantaneous vibration acceleration values ​​within the same period to obtain the second spectrum.

[0042] It should be noted that the specific process of performing a fast Fourier transform to obtain the spectrum is a well-known technique in the art, and will not be described in detail in this embodiment.

[0043] S102-4: Read the complex number corresponding to the preset target frequency from the first spectrum, and take its modulus as the arc force amplitude; read the complex number corresponding to the same target frequency from the second spectrum, and take its modulus as the vibration response amplitude.

[0044] It is important to understand that, since the instantaneous power and electromagnetic force pulsation frequency of the AC arc are always twice the power supply frequency, the energy of the mechanical excitation force generated by the arc and the vibration response it causes on the furnace wall is mainly concentrated at twice the power supply frequency. Selecting this frequency for analysis can capture the signal components driven by the arc itself to the greatest extent, while filtering out various mechanical background noises that are out of sync with the power grid (such as scrap steel collapse, boiling, etc.). Therefore, the preset target frequency is preferably the second harmonic frequency of the power grid frequency (for example, 100Hz for a 50Hz power grid and 120Hz for a 60Hz power grid).

[0045] In essence, the result of the Fast Fourier Transform (FFT) is a sequence of complex numbers, each corresponding to a specific frequency component. Therefore, the complex number corresponding to the preset target frequency is the one within this sequence that represents that frequency. Furthermore, a complex number contains a real part and an imaginary part, representing the amplitude of the sine wave at that frequency. Taking the complex modulus involves calculating the absolute value of this complex number.

[0046] S102-5: Divide the arc force amplitude by the preset arc force reference amplitude to obtain the arc force index in each cycle channel; divide the vibration response amplitude by the preset vibration response reference amplitude to obtain the vibration response index in each cycle channel.

[0047] The arc force index is the normalized Lorentz force intensity of the electric arc. A higher arc force index for a particular electrode channel within a given cycle indicates a stronger mechanical impact force generated by the arc in that electrode channel during that cycle.

[0048] The vibration response index is the normalized intensity of furnace wall vibration. If the vibration response index is for a specific electrode channel in a certain cycle, it indicates that the furnace wall vibration is stronger in that cycle and exceeds the level of the ideal submerged arc condition. This usually means that the foam slag layer is thinner or closer to disappearing (i.e., bare arc), the arc force is more likely to directly impact the furnace wall, or there is stronger mechanical interference in the furnace (such as the collapse of large pieces of scrap steel).

[0049] It should be noted that the historical data queue is maintained using a first-in-first-out queue data structure to store the arc force and vibration response indices for each electrode channel cycle, and the queue has a preset maximum length.

[0050] Each channel maintains a historical data queue.

[0051] For example, queue maintenance follows the principle of "new data in, old data out, maintaining a fixed capacity." Taking a queue with a maximum length of 5 as an example, assuming the current period is k, and the A-phase electrode calculates a new data pair (arc force index, vibration response index), we know the initial state is: the queue is empty []. Period k=1: the new data pair enters the queue and is placed at the head, the queue state is [(k=1)]; Period k=2: the new data pair enters the head, the existing data moves to the right, the queue state is [(k=2), (k=1)], and so on. Periods k=3, k=4, k=5 continue... The same operation is performed. When the period k=5, the new data pair enters the head of the queue. At this time, the queue is filled for the first time, reaching the maximum length of 5. The queue state is [(k=5), (k=4), (k=3), (k=2), (k=1)]. When the period k=6, the new data pair enters the head of the queue. Since the queue is full (maximum length = 5), the oldest data pair at the tail (rightmost) must be removed while inserting the new data pair. That is, the data pair corresponding to the period k=1. After the update, the queue state is [(k=6), (k=5), (k=4), (k=3), (k=2)].

[0052] In this embodiment, for each electrode channel, the arc force index and vibration response index of the channel obtained in the current cycle are combined into a new data pair; the new data pair is added to the head of the historical data queue maintained by the channel; it is determined whether the current length of the queue maintained by the channel exceeds the preset maximum length; if the current length of the queue after adding exceeds the preset maximum length, the oldest data pair located at the tail of the queue is removed; wherein, if the number of data pairs in the queue maintained by the channel does not reach the preset maximum length, a preset safety efficiency value is adopted for the channel.

[0053] It should be noted that the specific value of the preset maximum length can be determined based on process experience (such as typical time scales for changes in the state of foam slag), and this embodiment does not impose a specific limitation. For example, if the system cycle is 0.2 seconds, and we want to analyze the historical trend of the last 10 seconds, then the maximum length is 10 seconds ÷ 0.2 seconds = 50. Therefore, the typical range of the maximum length value can be 25 to 200, corresponding to data coverage of 5 to 40 seconds.

[0054] The preset safety efficiency value is a conservative estimate of the electrical energy thermal efficiency.

[0055] It should be noted that the specific value of the preset safety efficiency value can be determined based on engineering experience, and this embodiment does not impose a specific limitation. For example, in order to reflect the conservative principle of assuming good operating conditions when there is no conclusive evidence, the preset safety efficiency value is usually set to a relatively high value, such as 85%, that is, 85% of the electrical energy is converted into the thermal energy of the molten steel.

[0056] S103: For each channel, determine whether the dispersion of the arc force index in the queue meets the validity condition; if it does, fit the vibration force transmission ratio based on all data pairs in the queue; determine the reliability of the vibration signal based on the correlation between the arc force index and the vibration response index in the queue; determine the real-time electrical energy thermal efficiency of the channel based on the preset efficiency parameters, transmission ratio and reliability.

[0057] It is important to understand that during certain stages of smelting in a flat molten pool (such as the heat preservation period), the electrode current may be adjusted to be extremely stable by the control system, resulting in minimal fluctuations in the arc force index sequence. If subsequent linear regression and other analytical operations are forcibly performed at this time, the mathematical calculations may diverge due to the data variance approaching zero, or the algorithm may become overly sensitive to measurement noise, resulting in meaningless calculation results. Therefore, it is advisable to first determine the degree of dispersion of the arc force index sequence and only execute subsequent core analytical calculations when the data meets the validity conditions, thereby ensuring the numerical stability and reliability of the algorithm under all operating conditions.

[0058] In this embodiment, for each electrode channel, the standard deviation of all arc force indicators in the queue maintained by the current cycle channel is calculated as the fluctuation index. If the fluctuation index is less than the preset fluctuation threshold, all data pairs in the queue of the current cycle channel are determined to be invalid, and the subsequent transmission ratio and reliability calculations are skipped. The preset safety efficiency value is used as the real-time electrical and thermal efficiency of the current cycle channel.

[0059] The fluctuation index measures the degree to which the arc force index fluctuates around its average value over a period of time. Specifically, the smaller the fluctuation index of a channel in a given period (closer to zero), the more stable and less variable the mechanical force of the arc is within that period. This implies that the current is precisely controlled and the arc is in a "steady-state" operating phase (such as during the heat preservation period). In this case, the data in the queue is less suitable for statistical analysis, and forced calculations will result in meaningless output.

[0060] It should be noted that the specific value of the preset fluctuation threshold can be determined by offline analysis of historical data, and this embodiment does not impose a specific limitation. The specific method involves collecting a large amount of production data online and manually identifying generally accepted periods of "stable current and stable operating conditions" (such as the insulation period); calculating the standard deviation of the arc force index sequence within these periods and observing the distribution range of the standard deviation; and then, based on the upper limit of this distribution range, slightly increasing a safety margin (determined based on engineering experience, which will not be detailed in this embodiment), i.e., setting the fluctuation threshold. For example, if the maximum standard deviation of the stable period is 0.02 and the safety margin is 0.01, then the fluctuation threshold can be set to 0.03.

[0061] It is important to understand that, since the foamed slag layer acts as a crucial damping medium between the electric arc and the furnace wall, its thickness directly determines the efficiency of the transmission of the electric arc impact force to the furnace wall vibration: generally, the thicker the foamed slag layer, the greater the damping, and the smaller the vibration caused by a unit electric arc force; the thinner the foamed slag layer (or even a bare arc), the smaller the damping, and the more severe the vibration response. Therefore, by fitting the linear proportional coefficient between the electric arc force index and the vibration response index data pair, i.e., the vibration force transmission ratio, the damping characteristics of the current foamed slag layer can be accurately quantified, thereby directly diagnosing the physical state of the slag layer in the furnace (whether it is well covered or tends to be a bare arc), providing a core physical basis for subsequent real-time assessment of thermal efficiency loss.

[0062] In this embodiment, for each electrode channel, based on the data pairs of all arc force indexes and vibration response indexes in the queue maintained in the current cycle channel, linear regression is performed using the least squares method to obtain the regression slope of the vibration response index with respect to the arc force index; the regression slope is used as the vibration force transmission ratio in the current cycle channel.

[0063] It should be noted that the specific method of linear regression using the least squares method is a common technique, which will not be described in detail in this embodiment.

[0064] For example, first calculate the average value of all arc force indices and the average value of all vibration response indices in the historical data queue. Then calculate the deviation of each data point from its respective average value. Next, calculate the sum of the products of these deviations (covariance) and the sum of the squares of the arc force deviations (variance). Finally, divide the covariance by the variance to obtain the quotient, which is the optimal regression slope. The regression slope (i.e., the vibration force transmission ratio) is theoretically any real number. However, in the physical context of this invention, since the larger the arc force, the greater the mechanical impact transmitted to the furnace wall, the stronger the resulting vibration response should be (or at least unchanged). Therefore, the slope should be positive. The reasonable range of values ​​for the regression slope is usually [0, +∞), that is, zero or a positive number.

[0065] The vibration force transmission ratio reflects how many units the vibration response index changes on average for every unit change in the arc force index during the current cycle. It is a core indicator for quantifying the damping characteristics of the foamed slag layer. A higher vibration force transmission ratio for a particular channel during a given cycle indicates that changes in the arc force cause more significant vibration changes. This suggests that a thinner slag layer results in weaker damping, and the arc impact force is transmitted to the furnace wall more readily without attenuation. In this case, the operating condition is closer to that of a bare arc, and the thermal efficiency should be lower.

[0066] It is important to understand that, theoretically, the vibration force transmission ratio calculated through linear regression is a real number. However, based on physical facts (vibration response should not be negative) and engineering robustness considerations, the vibration force transmission ratio needs to be mapped to a reasonable physical range. Therefore, before using it to determine real-time electrical energy thermal efficiency, the vibration force transmission ratio needs to be limited. That is, if the vibration force transmission ratio is less than 0 (a very extreme case), the value is taken as 0; if the vibration force transmission ratio is greater than 1, the value is taken as 1; if 0 ≤ vibration force transmission ratio ≤ 1, the original value is retained.

[0067] It is important to understand that the furnace wall vibration signal contains not only the effective components generated by the electric arc, but also a large amount of random mechanical noise unrelated to the electric arc condition, such as scrap collapse and molten pool boiling. If all vibrations are attributed to the electric arc without distinction, it will lead to serious misjudgment of the slag layer condition (for example, misjudging a single collapse noise as a bare arc). Therefore, the confidence level of the current vibration signal change as originating from the electric arc can be quantified by determining the confidence level of the vibration signal. Only when the confidence level of the vibration signal is high can the slag condition information represented by the vibration force transmission ratio be accepted. If the confidence level of the vibration signal is low, it is determined that the vibration is mainly dominated by noise, thereby effectively shielding non-arc source interference and ensuring the reliability of subsequent thermal efficiency analysis conclusions.

[0068] In this embodiment, for each electrode channel, based on all data pairs in the queue maintained in the current cycle channel, an arc force index sequence and a vibration response index sequence are generated in the current cycle channel in chronological order; the Pearson correlation coefficient between the arc force index sequence and the vibration response index sequence is calculated as the correlation coefficient; and the square of the correlation coefficient is used as the reliability of the vibration signal in the current cycle channel.

[0069] It should be noted that the calculation method of Pearson correlation coefficient is a well-known technique in the art, and will not be described in detail in this embodiment.

[0070] It should be noted that since the Pearson correlation coefficient ranges from [-1, 1], its sign only indicates the direction of change (positive or negative correlation), while the amplitude (i.e. the strength of the correlation) is determined by the absolute value of the Pearson correlation coefficient. Therefore, in order to eliminate the influence of directionality, the squared value of the correlation coefficient is used as the reliability of the vibration signal.

[0071] It is important to understand that if the reliability of the vibration signal of a certain channel in a certain period is greater, it indicates that the change trend of the arc force and the vibration response is more synchronized and the direction is more consistent. That is, when the arc force increases, the vibration also increases, and when it decreases, the vibration decreases in time step, indicating that the vibration is more likely to be driven by the arc.

[0072] In this embodiment, the preset efficiency parameters include a preset efficiency benchmark value and a preset maximum efficiency loss value.

[0073] The preset efficiency benchmark is a theoretical or empirical upper limit of efficiency, used to represent the theoretically highest proportion of input electrical energy converted into the thermal energy of molten steel under ideal submerged arc conditions where the foam slag is perfectly covered and the electric arc is fully wrapped.

[0074] It should be noted that the specific value of the preset efficiency benchmark is an empirical constant determined based on process theory, equipment characteristics, and historical data, and this embodiment does not impose specific limitations. Specifically, it is determined by analyzing the heat balance data of the same furnace type under recognized favorable submerged arc conditions during a specific period, calculating the average thermal efficiency for that period, and then taking a slightly higher value (or the average value) as the efficiency benchmark. For example, based on electric arc furnace process experience and heat balance calculations, the efficiency benchmark is typically set between 0.85 and 0.95 (i.e., 85% to 95%).

[0075] The preset maximum efficiency loss value defines the maximum range in which efficiency may deteriorate, representing the total amount by which the efficiency value may decrease from the optimal state (i.e., the preset efficiency baseline value) to the worst state (i.e., a completely bare arc).

[0076] It should be noted that the specific value of the preset maximum efficiency loss is a constant set based on historical data analysis and engineering experience, and this embodiment does not impose specific limitations on it. Specifically, it is determined by analyzing historical data on severe bare arc periods, calculating the difference between the average efficiency and the efficiency benchmark value, and selecting a larger, representative value as the maximum efficiency loss. For example, based on industry experience, when a complete bare arc occurs, the thermal efficiency may decrease by 30% to 50% compared to the ideal value; therefore, the preset maximum efficiency loss value is typically set between 0.3 and 0.5.

[0077] To avoid calculating high efficiency under harsh operating conditions, before determining the real-time electrical energy thermal efficiency, it is first determined whether the vibration response index of the current cycle is abnormally high (e.g., greater than the preset abnormal threshold of 3.0). If the vibration response index is abnormally high and the reliability of the vibration signal is low (e.g., less than 0.1), it indicates that a severe mechanical disturbance (e.g., material collapse) has occurred without arc driving. In this case, the subsequent S-shaped weighted calculation formula is not executed, the real-time electrical energy thermal efficiency is directly set to the preset safe efficiency value, and the calculation of the real-time electrical energy thermal efficiency of the current cycle ends.

[0078] It should be noted that setting a preset safety efficiency value is intended to avoid numerical instability caused by forcibly performing regression analysis when the system is in steady state (with minimal current fluctuations), while also saving computational resources. Using a preset safety efficiency value in this case is a conservative and robust control strategy, avoiding unnecessary over-adjustment due to minor disturbances.

[0079] The process of determining real-time electrical thermal efficiency is as follows: Figure 3 As shown, it includes: S103-1: Obtain the credibility threshold.

[0080] It should be noted that the specific value of the credibility threshold can be determined based on historical production data, and this embodiment does not impose a specific limitation. For example, the credibility threshold can range from 0.3 to 0.5.

[0081] S103-2: For each electrode channel, a preset S-shaped function is used to map the reliability of the vibration signal in the current cycle channel to a reliability weight between 0 and 1; where, if the reliability of the vibration signal is much lower than the reliability threshold, the mapped reliability weight approaches 0; if the reliability of the vibration signal is much higher than the reliability threshold, the mapped reliability weight approaches 1.

[0082] Since a larger vibration signal in a certain electrode channel during the current cycle indicates a more synchronized and consistent trend between the arc force and vibration response, it reflects a higher degree of confidence in the operating condition information represented by the vibration force transmission ratio. This results in a higher confidence weight (closer to 1), indicating that the system is more certain that the current vibration change is caused by the arc force. Therefore, the efficiency loss calculated from the vibration force transmission ratio is more likely to be fully adopted, thus the calculated real-time electrical thermal efficiency more accurately reflects the efficiency decrease caused by the thinning of the slag layer. Therefore, the confidence weight can be represented by the following preset S-shaped function: in, The 'r' represents the confidence weight; exp() represents an exponential function with the natural constant as the base; r represents the preset sharpness coefficient. Indicates the reliability of the vibration signal; This indicates the credibility threshold.

[0083] It is important to understand that the preset sharpness coefficient controls the "steepness" of the S-curve. The larger the preset sharpness coefficient, the steeper the S-curve becomes near the confidence threshold. This means that as long as the confidence of the vibration signal is slightly higher or lower than the threshold, the confidence weight will quickly jump to close to 1 or 0, and the decision-making performance will be more sensitive.

[0084] It should be noted that the specific value of the preset sharpness coefficient can be determined through engineering experience, and this embodiment does not impose a specific limitation. For example, a commonly used sharpness coefficient that can produce obvious non-linear transitions is in the range of 5 to 15.

[0085] S103-3: Calculate the product of the vibration force transmission ratio in the current cycle channel and the preset maximum efficiency loss value, and use it as the efficiency loss amount.

[0086] S103-4: Subtract the product of the confidence weight and the efficiency loss from the preset efficiency benchmark value to obtain the real-time electrical energy thermal efficiency in the current cycle channel.

[0087] Since the efficiency loss represents the slag layer condition diagnosed in the current cycle (characterized by the transmission ratio), if the slag layer condition is fully accepted (i.e., the confidence weight is high), theoretically, the value that needs to be deducted from the efficiency baseline value. If the vibration force transmission ratio of a certain channel in a certain cycle is larger, it indicates a thinner slag layer and worse operating conditions, thus leading to a larger efficiency loss. This causes the real-time electrical thermal efficiency to tend to decrease. Furthermore, if the confidence weight is higher, it indicates a more reliable signal, and the efficiency loss calculated from the vibration force transmission ratio and the preset maximum efficiency loss value should be fully accepted. In this case, the real-time electrical thermal efficiency will decrease accordingly. Therefore, the real-time electrical thermal efficiency can be expressed by the following formula: in, Indicates real-time electrical energy thermal efficiency; This indicates the preset efficiency benchmark value; Indicates the credibility weight; This indicates the preset maximum efficiency loss value; clamp() represents the amplitude limiting function, which limits the vibration force transmission ratio to between 0 and 1; K represents the vibration force transmission ratio.

[0088] S104: Correct the real-time electrical power based on the real-time electrical energy thermal efficiency, and determine the net heating power of molten steel by combining the preset chemical reaction heat power and the preset heat loss power; determine the predicted temperature of molten steel by accumulating the net heating power of molten steel; and correct the predicted temperature based on the measured temperature.

[0089] In this embodiment, the real-time electrical power of each electrode channel is obtained; and the preset chemical reaction heat power and preset heat loss power are obtained; the real-time electrical power of each electrode channel is multiplied by its corresponding real-time electrical energy thermal efficiency to obtain the effective electrical power of each channel; the effective electrical power of all channels is summed to obtain the total effective electrical power; the sum of the total effective electrical power and the preset chemical reaction heat power is subtracted from the preset heat loss power to obtain the net heating power of the molten steel.

[0090] It should be noted that the instantaneous total active power of each phase electrode (such as phases A, B, and C of the electric arc furnace) is read in real time from the energy meter or power transmitter on the secondary side of the transformer supplying power to the electric arc furnace, and recorded as real-time power (unit: kilowatt or megawatt).

[0091] It is important to understand that since the real-time thermal efficiency is a discrete value calculated and updated at the end of each fixed cycle, while the real-time power is a continuously changing signal output by the energy meter, in order to ensure that the real-time thermal efficiency (one value per cycle) and the real-time power (a continuous value) can be matched in time, the cycle to which it belongs can be determined at any point in time. Within this entire cycle, the system treats the real-time thermal efficiency calculated at the end of the cycle as a constant and keeps it unchanged, and uses it to correct the real-time power at all times within the cycle.

[0092] The preset chemical reaction heat power is a power-based estimate of the heat released by chemical processes such as the carbon-oxygen reaction in the electric arc furnace molten pool.

[0093] It should be noted that the specific value of the preset chemical reaction heat power is based on empirical constants set for each process stage. A theoretical range of chemical reaction heat power can be estimated using the existing thermochemical equations for carbon-oxygen reactions, and this embodiment does not impose specific limitations. For example, during the oxidation period (large-scale oxygen blowing for decarburization), the chemical reaction is vigorous, and a higher constant can be used, such as a chemical reaction heat power of 8MW; during the reduction period or before tapping, the chemical reaction is weak, and the chemical reaction heat power can be set to 0.5MW.

[0094] It should be noted that the preset heat loss power is a parameter with adaptive learning capabilities. The initial value of the preset heat loss power can be determined based on historical data and used as the initial value input when initially calculating the net heating power of the molten steel. In subsequent calculations, the value of the heat loss power is not fixed, but will be dynamically adjusted according to the deviation between the measured temperature of the molten steel and the predicted temperature.

[0095] It should be noted that the initial value of the preset heat loss power can be determined based on historical data, and this embodiment does not impose a specific limitation. For example, by analyzing a large amount of historical production data, the average value of the input electrical power during the "no chemical reaction, constant temperature period" can be calculated and used as the initial value of the heat loss power. For example, for an electric arc furnace with a capacity of 100 tons, the typical initial value of the heat loss power may be in the range of 2MW to 5MW.

[0096] To continuously adjust the value of heat loss power, as an example, if a measured temperature of molten steel is obtained, the deviation between the measured temperature and the current predicted temperature is calculated; the time from the last measured temperature to the current measured temperature of molten steel is calculated; the deviation is divided by the time to obtain the average temperature deviation per unit time; the average temperature deviation per unit time is multiplied by a preset learning rate coefficient to obtain the correction amount of heat loss power; the current value of the preset heat loss power is subtracted from the correction amount to obtain the updated value of heat loss power.

[0097] The preset learning rate coefficient determines the magnitude of the correction to the heat loss power based on the temperature deviation each time.

[0098] It should be noted that the specific value of the preset learning rate coefficient is a dimensionless positive number, which can be determined through offline simulation and on-site debugging; this embodiment does not impose a specific limitation. The specific determination process involves using historical data to run the entire system's algorithm in a simulation environment; then setting different learning rate coefficient values ​​(e.g., 0.05, 0.1, 0.2, etc.) and observing the convergence trajectory of heat loss power during multiple temperature measurement calibrations; based on the convergence trajectory, selecting the value that allows for timely but not overly aggressive response to heat loss power, and using the learning rate coefficient that achieves stable convergence as the final set learning rate coefficient. For example, a typical range for the preset learning rate coefficient is between 0.05 and 0.3.

[0099] It is important to understand that if the measured temperature is higher than the predicted temperature, it means that the prediction is too low. This may be because the set heat loss power is too large, resulting in an underestimation of the net heating power of the molten steel. Therefore, it is necessary to reduce the value of the heat loss power. Thus, the correction amount is subtracted to obtain the updated value of the heat loss power.

[0100] It's important to understand that the net heating power of molten steel is the heat energy actually absorbed by the molten steel and used to raise its temperature per unit time. It is the net effective heat flow after deducting all losses (such as ineffective electrical power and furnace heat dissipation) from the total input energy (i.e., the sum of total effective electrical power and preset chemical reaction heat power). Therefore, if the net heating power of molten steel is higher at a certain moment, it indicates that the heating intensity is higher at that moment, and the temperature of the molten steel is rising more rapidly.

[0101] In this embodiment, a preset temperature prediction integrator is maintained, the state of which includes an initial temperature value and an accumulated heat energy value. The temperature prediction integrator uses a preset molten steel mass and a preset specific heat capacity as parameters, and updates the accumulated heat energy value by continuously accumulating the product of the net heating power of the molten steel and time. The accumulated heat energy value is then divided by the product of the preset estimated molten steel mass and the preset molten steel specific heat capacity to obtain the temperature rise value from the initial moment. The temperature rise value is then added to the initial temperature value to obtain the predicted temperature of the molten steel at the current moment. The initial temperature value is set to the first measured temperature of the molten steel when it is started, and is updated to the new measured temperature value each time a new measured temperature of the molten steel is obtained. At the same time, the accumulated heat energy value is reset to zero.

[0102] It should be noted that the specific value of the preset estimated mass of molten steel is a known process parameter directly calculated from the production process batching list. The calculation method is a common technical means, and this embodiment does not impose specific limitations. For example, for an electric arc furnace with a nominal capacity of 100 tons, considering a certain metal yield, its molten steel mass is usually set between 90 and 110 tons.

[0103] It should be noted that the specific value of the preset specific heat capacity of molten steel can be directly taken from publicly available materials physical properties handbooks, metallurgical textbooks, or industry standard data, and this embodiment does not impose specific limitations. For example, for conventional carbon steel molten steel, the typical range of specific heat capacity is 0.75 to 0.82 kJ / (kg·°C).

[0104] Since the core principle of the formula for calculating the predicted temperature is based on the application of the first law of thermodynamics (the law of conservation of energy) to the specific system of an electric arc furnace, and this principle is a publicly known technology, the specific calculation process will not be elaborated upon here. What needs to be emphasized is that this formula relies on its integrand... The revolutionary reconstruction of (i.e., net heating power of molten steel) means that the predicted temperature of molten steel can be expressed by the temperature prediction integral formula built into the temperature prediction integrator: in, Indicates the predicted temperature of molten steel; Indicates the initial temperature value; This indicates the estimated mass of the molten steel. Indicates the preset specific heat capacity of molten steel; Indicates the cumulative heat energy value; Indicates the first Net heating power of molten steel at any given moment; t represents the initial time; t represents the current time.

[0105] It should be noted that when the system obtains a measured temperature of molten steel, it immediately performs the following operations: directly overwrites the current predicted temperature value output by the temperature prediction integrator with the measured temperature, and simultaneously clears the accumulated heat energy value inside the integrator to zero.

[0106] S105: Control the smelting endpoint based on the predicted temperature and perform corresponding optimized smelting operations based on the real-time electrical energy thermal efficiency.

[0107] It is important to understand that, generally speaking, the smelting process has a preset target tapping temperature required by the process. The system will continuously compare the real-time predicted temperature of the molten steel output by the integrator with the target tapping temperature. If the system determines that the predicted temperature of the molten steel continuously and stably reaches or exceeds the target tapping temperature, it will automatically generate and issue a power outage command to stop supplying power to the electrodes, and at the same time trigger an audible and visual alarm to prompt the operator to prepare for tapping.

[0108] It should be noted that the specific value of the preset target tapping temperature can be determined according to the process design requirements, and this embodiment does not impose a specific limitation. For example, for ordinary carbon steel, the typical target tapping temperature range is between 1600°C and 1650°C.

[0109] It should be noted that the core of optimizing smelting operations based on real-time electrical energy thermal efficiency is proactive process intervention based on efficiency thresholds.

[0110] For example, the system continuously monitors the real-time electrothermal efficiency of each electrode channel. If the efficiency of any channel is detected to be continuously lower than the preset low efficiency alarm threshold (e.g., 0.6) for a certain period of time (e.g., 5 seconds), the system determines that the corresponding channel is in a state of continuous low efficiency bare arc or poor slag condition. Subsequently, the system automatically triggers optimization commands, such as increasing the carbon powder and oxygen blowing rate in the area where the channel is located to physically thicken the foam slag. If the real-time electrothermal efficiency still does not recover after a certain period of time (i.e., exceeds the low efficiency alarm threshold), an auxiliary command is initiated (e.g., adjusting the electrode position to forcibly shorten the arc). If the real-time electrothermal efficiency of the channel recovers to the normal range, the system automatically cancels the intervention.

[0111] A temperature feedback control system for the smelting stage of an electric arc furnace flat-bottomed pool includes a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of a temperature feedback control method for the smelting stage of an electric arc furnace flat-bottomed pool.

[0112] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0113] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A method for temperature feedback control during the smelting stage of an electric arc furnace flat-walled molten pool, characterized in that, The method includes: In response to the smelting entering the flat melting pool stage, the actual temperature of the molten steel is obtained; The current signals and furnace wall vibration signals of each electrode channel of the electric arc furnace are collected synchronously. Based on the current signals and vibration signals of each channel, the arc force amplitude and vibration response amplitude of each channel are periodically extracted, and the arc force index and vibration response index are obtained after standardization. The two data pairs are stored in the historical data queue of each channel in chronological order and updated. For each channel, determine whether the dispersion of the arc force index within the queue meets the validity condition; if it does, fit the vibration force transmission ratio based on all data pairs within the queue; determine the reliability of the vibration signal based on the correlation between the arc force index and the vibration response index within the queue; determine the real-time electrical and thermal efficiency of the channel based on the preset efficiency parameters, transmission ratio, and reliability. The real-time electrical power is corrected based on the real-time electrical energy thermal efficiency, and the net heating power of the molten steel is determined by combining the preset chemical reaction heat power and the preset heat loss power. The predicted temperature of the molten steel is determined by accumulating the net heating power of the molten steel. The predicted temperature is then corrected based on the measured temperature. The smelting endpoint is controlled based on the predicted temperature, and corresponding optimized smelting operations are performed based on the real-time electrical energy thermal efficiency.

2. The temperature feedback control method for the smelting stage of an electric arc furnace flat-walled molten pool according to claim 1, characterized in that, The current signal includes the instantaneous current value of each phase electrode channel acquired by the current transformer; the furnace wall vibration signal includes the instantaneous value of furnace wall vibration acceleration in each region acquired by the vibration acceleration sensor installed on the outer wall of the electric arc furnace shell and corresponding to the physical position of each phase electrode. The process of obtaining the arc force index and vibration response index includes: The instantaneous current values ​​at each acquisition moment in the current signal are squared to obtain the squared current value; based on the squared current values ​​at each acquisition moment, an equivalent arc force signal proportional to the Lorentz force of the arc is generated in chronological order. For each electrode channel, obtain the current square value sequence and vibration acceleration instantaneous value sequence within each cycle; and obtain the preset arc force reference amplitude and preset vibration response reference amplitude; Perform a Fast Fourier Transform (FFT) on the sequence of squared current values ​​within each period to obtain the first spectrum; perform a Fast Fourier Transform (FFT) on the sequence of instantaneous vibration acceleration values ​​within the same period to obtain the second spectrum; Read the complex number corresponding to the preset target frequency from the first spectrum, and take its modulus as the arc force amplitude; read the complex number corresponding to the same target frequency from the second spectrum, and take its modulus as the vibration response amplitude; Divide the arc force amplitude by the preset arc force reference amplitude to obtain the arc force index in each cycle channel; divide the vibration response amplitude by the preset vibration response reference amplitude to obtain the vibration response index in each cycle channel.

3. The temperature feedback control method for the smelting stage of an electric arc furnace flat-walled molten pool according to claim 1, characterized in that, The historical data queue is maintained using a first-in-first-out queue data structure to store the arc force index and vibration response index within each electrode channel cycle, and the queue has a preset maximum length. The process of storing the data pairs (comprising the two data points) into the historical data queues of each channel in chronological order and updating them includes: For each electrode channel, the arc force index and vibration response index of the channel obtained in the current cycle are combined into a new data pair; the new data pair is added to the head of the historical data queue maintained for the channel; it is determined whether the current length of the queue maintained by the channel exceeds the preset maximum length. If the current length of the queue after the addition exceeds the preset maximum length, the oldest data pair at the end of the queue is removed; if the number of data pairs in the queue maintained by the channel does not reach the preset maximum length, the channel adopts a preset safety efficiency value.

4. The temperature feedback control method for the smelting stage of an electric arc furnace flat-walled molten pool according to claim 3, characterized in that, The determination of whether the dispersion of the arc force index within the queue meets the validity condition includes: For each electrode channel, calculate the standard deviation of all arc force indices in the queue maintained by the current cycle channel, and use it as a fluctuation index; If the fluctuation index is less than the preset fluctuation threshold, all data pairs in the queue of the current cycle channel are determined to be invalid, and subsequent transmission ratio and reliability calculations are skipped. The preset safety efficiency value is used as the real-time electrical and thermal efficiency of the current cycle channel.

5. The temperature feedback control method for the smelting stage of an electric arc furnace flat-walled molten pool according to claim 4, characterized in that, The process of obtaining the vibration force transmission ratio includes: For each electrode channel, based on the data pairs of all arc force indices and vibration response indices in the queue maintained in the current cycle channel, linear regression is performed using the least squares method to obtain the regression slope of the vibration response index with respect to the arc force index. The regression slope is used as the vibrational force transmission ratio in the current cycle channel.

6. The temperature feedback control method for the smelting stage of an electric arc furnace flat-walled molten pool according to claim 4, characterized in that, The process for determining the reliability of the vibration signal includes: For each electrode channel, based on all data pairs in the queue maintained in the current cycle channel, the arc force index sequence and vibration response index sequence for the current cycle channel are generated in chronological order. Calculate the Pearson correlation coefficient between the arc force index sequence and the vibration response index sequence, and use it as the correlation coefficient; The square of the correlation coefficient is used as the reliability of the vibration signal in the current cycle channel.

7. The temperature feedback control method for the smelting stage of an electric arc furnace flat-walled molten pool according to claim 1, characterized in that, The preset efficiency parameters include a preset efficiency benchmark value and a preset maximum efficiency loss value; The real-time electrical energy thermal efficiency determination process includes: Obtain the credibility threshold; For each electrode channel, a preset S-shaped function is used to map the reliability of the vibration signal in the current cycle channel to a reliability weight between 0 and 1. Specifically, if the reliability of the vibration signal is much lower than the reliability threshold, the mapped reliability weight approaches 0; if the reliability of the vibration signal is much higher than the reliability threshold, the mapped reliability weight approaches 1. The product of the vibration force transmission ratio in the current cycle channel and the preset maximum efficiency loss value is calculated as the efficiency loss amount. The real-time electrical thermal efficiency in the current cycle channel is obtained by subtracting the product of the confidence weight and the efficiency loss from the preset efficiency benchmark value.

8. The temperature feedback control method for the smelting stage of an electric arc furnace flat-walled melting pool according to claim 7, characterized in that, The process for determining the net heating power of the molten steel includes: Acquire the real-time electrical power of each electrode channel; and acquire the preset chemical reaction heat power and preset heat loss power; Multiply the real-time electrical power of each electrode channel by its corresponding real-time electrical energy thermal efficiency to obtain the effective electrical power of each channel. The total effective power is obtained by summing the effective power of all channels. The net heating power of molten steel is obtained by subtracting the preset heat loss power from the sum of the total effective electrical power and the preset chemical reaction heat power.

9. The temperature feedback control method for the smelting stage of an electric arc furnace flat-walled molten pool according to claim 1, characterized in that, The process for determining the predicted temperature of molten steel includes: Maintain a preset temperature prediction integrator, the status of which includes an initial temperature value and an accumulated heat energy value; The temperature prediction integrator uses the preset molten steel mass and preset specific heat capacity as parameters. It updates the accumulated heat energy value by continuously accumulating the product of the net heating power of the molten steel and time. The accumulated heat energy value is divided by the product of the preset estimated mass of molten steel and the preset specific heat capacity of molten steel to obtain the temperature rise value from the initial time. The temperature rise value is then added to the initial temperature value to obtain the predicted temperature of the molten steel at the current time. The initial temperature value is set to the first measured temperature of the molten steel when the system is started, and is updated to the new measured temperature value each time a new measured temperature of the molten steel is obtained. At the same time, the accumulated heat energy value is reset to zero.

10. A temperature feedback control system for the smelting stage of an electric arc furnace flat-walled molten pool, characterized in that, The system includes a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that the processor executes the computer program to implement the steps of the method as described in any one of claims 1 to 9.

Images