A method, system, and medium for processing operational data of a refining electric furnace steelmaking

Abstract

The application relates to the technical field of electric furnace steelmaking process control, and discloses a running data processing method and system for refining electric furnace steelmaking and a medium. The method comprises the following steps: acquiring electric resistance characteristic data and smelting stage information of a scrap steel smelting process, analyzing the influence of component diversity on the electric resistance characteristics, and establishing an initial power factor adjustment model; extracting a key feature vector, predicting a power mutation risk through a neural network, and determining a potential energy waste point; constructing a dynamic adjustment sequence, optimizing parameters of the dynamic adjustment sequence through a gradient descent algorithm, and generating a smooth transition control strategy; extracting adaptive adjustment parameters, fusing an electric resistance characteristic change trend, and determining a final power factor value; evaluating equipment loss, generating a power input sequence, acquiring a stable smelting efficiency index, and updating the model based on feedback data. Through the cooperation of multiple algorithms, the application realizes self-adaptive and accurate adjustment of smelting power, effectively suppresses power mutation, reduces energy consumption and equipment loss, and improves smelting efficiency and stability.

Description

Technical Field

[0001] This application relates to the field of electric arc furnace steelmaking process control technology, and in particular to a method, system and medium for processing operating data in refining electric arc furnace steelmaking. Background Technology

[0002] With the widespread use of scrap steel resources, refining electric furnaces (such as electric arc furnaces) play a crucial role in the steelmaking process. However, the complexity and diversity of scrap steel composition during smelting lead to frequent fluctuations in its electrical resistance characteristics. These fluctuations make adjusting the furnace power more difficult, directly impacting smelting efficiency, energy consumption, and the stability and lifespan of the equipment. The smelting process in an electric arc furnace typically includes preheating, melting, and refining stages, each with varying power requirements. The changes in scrap steel composition make it impossible to accurately predict the furnace's power demands, resulting in frequent power surges.

[0003] Sudden power fluctuations not only lead to energy waste but can also cause equipment overload or damage, and even interrupt the smelting process. Meanwhile, resistance fluctuations during smelting can cause instability in the electric furnace's power input, increasing energy consumption and reducing production efficiency. Therefore, effectively monitoring resistance fluctuations, anticipating sudden power fluctuations, and adjusting power input in a timely manner have become significant technical challenges for improving smelting efficiency and equipment stability.

[0004] In past research and technological applications, traditional power regulation methods relied on human experience and simple control algorithms, which often failed to adequately cope with the complex changes in the smelting process. With the rapid development of computer technology and machine learning technology, more and more intelligent algorithms are being introduced into the control systems of refining electric furnaces, aiming to improve the accuracy and flexibility of power regulation.

[0005] Currently, intelligent control technologies for refining electric arc furnaces mainly include power regulation based on real-time monitoring data, energy consumption optimization, and equipment loss assessment. By using machine learning algorithms such as Support Vector Machines (SVM) to analyze the impact of scrap steel composition on resistance fluctuations, the intelligence level of power factor adjustment is gradually improved. However, existing technical solutions still face many challenges, such as how to effectively predict the risk of power surges, how to smoothly transition through dynamic adjustment of the power sequence, and how to assess equipment loss levels and energy efficiency in real time. Summary of the Invention

[0006] This application provides a method, system, and medium for processing operating data in electric arc furnace steelmaking, which aims to precisely adjust the power input of the electric arc furnace during scrap steel smelting through intelligent algorithms, so as to cope with the resistance fluctuation problem caused by the diversity of scrap steel composition, reduce the risk of power sudden change, and improve smelting efficiency and equipment stability.

[0007] In a first aspect, this application provides a method for processing operating data in electric arc furnace steelmaking, the method comprising:

[0008] Step S1: Obtain resistance characteristic data and smelting stage information during the scrap steel smelting process, use support vector machine algorithm to analyze the impact of composition diversity on resistance characteristics, and establish an initial power factor adjustment model.

[0009] Step S2: Extract key feature vectors representing the dynamic changes in resistance during stage switching from the initial power factor adjustment model. If the key feature vectors exceed the preset vector threshold, predict the risk of power mutation through a neural network algorithm and determine the potential energy waste points in the smelting process.

[0010] Step S3: Construct a dynamic adjustment sequence based on potential energy waste points, and use the gradient descent algorithm to optimize the parameters of the dynamic adjustment sequence to generate a smooth transition control strategy;

[0011] Step S4: Extract adaptive adjustment parameters from the smooth transition control strategy. If the adaptive adjustment parameters match the information of the current melting stage, then integrate the real-time monitored resistance characteristic change trend to determine the final power factor value for the current melting stage.

[0012] Step S5: Based on the equipment loss monitoring data obtained during operation of the final power factor value, the support vector machine algorithm is used to evaluate the equipment loss level. If the loss level is lower than the preset loss threshold, an optimized power input sequence is generated.

[0013] Step S6: Calibrate the resistance characteristic fluctuations using the optimized power input sequence to obtain a stable smelting efficiency index;

[0014] Step S7: Extract feedback data on the frequency and magnitude of power mutations from the smelting efficiency index. If the feedback data indicates a reduction in power mutations, update the initial power factor adjustment model.

[0015] Secondly, this application provides an operating data processing system for refining electric arc furnace steelmaking, the system comprising:

[0016] The model generation unit is used to acquire resistivity data and smelting stage information during the scrap steel smelting process, and uses the support vector machine algorithm to analyze the impact of compositional diversity on resistivity characteristics and establish an initial power factor adjustment model.

[0017] The risk prediction unit is used to extract key feature vectors representing the dynamic changes in resistance during stage switching from the initial power factor adjustment model. If the key feature vectors exceed the preset vector threshold, the power change risk is predicted by a neural network algorithm, and potential energy waste points in the smelting process are identified.

[0018] The strategy generation unit is used to construct a dynamic adjustment sequence based on potential energy waste points, and to optimize the parameters of the dynamic adjustment sequence using a gradient descent algorithm to generate a smooth transition control strategy.

[0019] The power determination unit is used to extract adaptive adjustment parameters from the smooth transition control strategy. If the adaptive adjustment parameters match the information of the current melting stage, the final power factor value for the current melting stage is determined by integrating the real-time monitored resistance characteristic change trend.

[0020] The sequence generation unit is used to evaluate the equipment loss level based on the equipment loss monitoring data obtained during operation of the final power factor value using the support vector machine algorithm. If the loss level is lower than the preset loss threshold, an optimized power input sequence is generated.

[0021] The fluctuation calibration unit is used to calibrate the resistance characteristic fluctuation through the optimized power input sequence to obtain a stable melting efficiency index.

[0022] The model update unit is used to extract feedback data on the frequency and magnitude of power mutations from the smelting efficiency index. If the feedback data indicates a reduction in power mutations, the initial power factor adjustment model is updated.

[0023] Thirdly, this application provides a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the above-described method for processing operating data in electric arc furnace steelmaking.

[0024] Compared with the prior art, the beneficial effects of the technical solution of this application are at least as follows:

[0025] 1. By precisely adjusting the power input and real-time resistance fluctuations, the sudden power fluctuations of the electric furnace are reduced, thereby effectively reducing energy waste. The optimized power input sequence can smoothly transition and avoid ineffective energy consumption caused by power fluctuations.

[0026] 2. Through adaptive adjustment and dynamic optimization control strategies, the electric furnace can maintain a stable power input at different melting stages, thereby improving the efficiency of the melting process, effectively suppressing resistance fluctuations, and significantly improving the stability and predictability of the melting process.

[0027] 3. Optimizing the power input sequence and reducing power fluctuations not only improves smelting efficiency but also reduces equipment losses caused by overload and frequent fluctuations. Long-term application of this method can effectively extend the service life of the equipment and reduce maintenance and repair costs.

[0028] 4. By updating the model with real-time feedback data, the system can automatically adapt to the characteristics of different scrap steel batches, ensuring that the power adjustment strategy has high flexibility and long-term stability. This adaptive adjustment function enables the system to handle complex and ever-changing smelting environments, ensuring that optimal performance is achieved in every smelting operation.

[0029] 5. Through real-time equipment loss monitoring and precise power factor adjustment, equipment energy efficiency can be optimized, power consumption can be reduced, and the working efficiency of the equipment can be maximized, thereby improving the energy efficiency ratio of the entire smelting process.

[0030] 6. This invention constructs a complete closed-loop control system of "perception-decision-execution-feedback". By using the smelting efficiency index as feedback to drive the online update of the initial model, the control strategy can adapt to scrap steel raw materials with different compositions and changing operating conditions, realizing intelligent and continuous optimization of the smelting process, and ensuring long-term operational stability and product quality consistency. Attached Figure Description

[0031] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0032] Figure 1 This is a flowchart of the operation data processing method for refining electric arc furnace steelmaking according to this application;

[0033] Figure 2 This is a schematic diagram of the operating data processing system for refining electric arc furnace steelmaking according to this application. Detailed Implementation

[0034] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a particular order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms “comprising” or “having,” and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0035] For ease of understanding, the specific process of the embodiments of this application is described below. Figure 1The diagram shows a flowchart of the operation data processing method for refining electric arc furnace steelmaking provided by the present invention. The flowchart specifically includes the following steps:

[0036] Step S1: Obtain resistance characteristic data and smelting stage information during the scrap steel smelting process, use support vector machine algorithm to analyze the impact of composition diversity on resistance characteristics, and establish an initial power factor adjustment model.

[0037] In one specific embodiment, the process of performing step S1 may specifically include the following steps:

[0038] The resistance characteristics data and smelting stage information of scrap steel are collected in real time by sensors. The resistance characteristics data include voltage and current values, and the smelting stage information includes preheating stage, melting stage or refining stage.

[0039] The resistance characteristic data and smelting stage information are input into a pre-trained support vector machine model. The support vector machine model classifies the nonlinear effects of component diversity on resistance characteristics and outputs classification results to characterize the degree of resistance fluctuation.

[0040] Based on the classification results, an initial power factor adjustment model is generated using a mapping function or lookup table. The initial power factor adjustment model characterizes the mapping relationship between the diversity of scrap steel composition and the ideal power factor.

[0041] Specifically, sensors installed on the smelting furnace are used to monitor voltage and current values ​​in real time. These values ​​reflect the resistance changes of scrap steel at different temperatures. The current smelting stage is also recorded, such as the low-temperature heating in the preheating stage or the high-temperature melting in the melting stage. The collected voltage, current, and smelting stage information are used as input data and transmitted to the computing unit for preprocessing, including normalization, to ensure data consistency. For example, in the preheating stage of scrap steel smelting, the sensor collects data with a voltage of 220V and a current of 100A and marks it as the preheating stage. This data is used for subsequent analysis of resistance fluctuations caused by compositional diversity.

[0042] Support Vector Machine (SVM) is a supervised learning algorithm that constructs a classification hyperplane by maximizing the margin between support vectors. Here, historical scrap steel smelting datasets are used for pre-training, including resistance data for various composition combinations. The training process involves selecting a radial basis function as the kernel function to handle the nonlinear effects of compositional diversity. During training, sample data containing different scrap steel compositions, such as iron, carbon, and impurity ratios, are collected and labeled with resistance characteristic changes, for example, the category where high impurity content leads to increased resistance. Cross-validation is used to optimize model parameters, such as a regularization parameter C of 1.0 and a kernel parameter gamma of 0.1, ensuring the model achieves an accuracy of over 95% in classifying the impact of compositional diversity on resistance. Real-time collected voltage, current, and smelting stage information are input into this trained SVM model, which automatically extracts features, such as the voltage-to-current ratio, as resistance indicators. For example, during the smelting stage, if the input voltage is 380V and the current is 150A, the model will determine whether it belongs to the high compositional diversity category based on the trained hyperplane, thus identifying abnormal resistance. It should be noted that this input process ensures real-time performance. Through the kernel function mapping of the model, the nonlinear resistance changes caused by the diversity of components, such as the different metal proportions in scrap steel, are effectively captured.

[0043] When classifying, the Support Vector Machine (SVM) model first calculates the similarity between the feature vectors and support vectors of the input data. It then uses a kernel function to map the data to a high-dimensional space and outputs classification labels based on a decision function, such as low diversity, medium diversity, or high diversity categories, corresponding to the degree of resistance variation. For the low diversity category, the model identifies that uniform scrap steel composition leads to stable resistance, classifying it as category 1. For high diversity, the mixed composition causes large resistance fluctuations, classifying it as category 3. The classification result also includes a confidence score; for example, a confidence score of 0.8 indicates high reliability, used to filter noisy data. Classification is applied at different smelting stages. For example, in the refining stage, the model focuses on analyzing the impact of impurity composition on resistance, outputting classification results specific to that stage. For instance, in a batch of scrap steel containing 60% iron, 20% carbon, and 20% impurities, the model classifies it as high diversity, showing a 15% increase in resistance.

[0044] It should be noted that this classification mechanism reduces overfitting by maximizing the interval, effectively distinguishing the resistance patterns caused by composition in the scrap steel smelting scenario. The beneficial effect is to improve the response speed to sudden resistance changes and reduce energy waste.

[0045] After obtaining the classification results, an initial power factor adjustment model is generated using a mapping function or lookup table. For example, when the classification result is "high diversity," impurities in the components may cause a significant increase in resistance. In this case, the model maps to a lower initial power factor setting (e.g., 0.85) to cope with the expected resistance instability; conversely, the "low diversity" category corresponds to a higher power factor setting (e.g., 0.9). This model stores the mapping relationship between component diversity and the ideal power factor in a structured manner, providing a benchmark for dynamic power adjustment in subsequent stages.

[0046] The technical solution of this invention can accurately analyze and respond to resistance fluctuations caused by scrap steel composition, achieving refined power factor adjustment. This not only reduces energy waste but also significantly improves the efficiency of the smelting process and the service life of the equipment. The application of this technology provides a highly efficient and intelligent power regulation solution for refining electric furnaces, solving technical bottlenecks such as inaccurate power factor adjustment and severe energy waste.

[0047] Step S2: Extract key feature vectors representing the dynamic changes in resistance during stage switching from the initial power factor adjustment model. If the key feature vectors exceed the preset vector threshold, predict the risk of power mutation through a neural network algorithm and determine the potential energy waste points in the smelting process.

[0048] In one specific embodiment, the process of performing step S2 may specifically include the following steps:

[0049] From the real-time data of the initial power factor adjustment model, the resistance characteristic data sequence within a set time window before and after the stage switching moment is extracted. The resistance change rate is calculated based on the resistance characteristic data sequence, and the corresponding stage switching time point is recorded. A key feature vector is constructed based on the resistance change rate and the stage switching time point.

[0050] Calculate the magnitude of the key feature vector and determine whether the magnitude exceeds the preset vector threshold. If so, input the key feature vector into the pre-trained neural network model.

[0051] The neural network model outputs a risk probability value that represents the possibility of a power mutation based on the nonlinear relationship in the key feature vectors;

[0052] Determine whether the risk probability value exceeds the preset risk threshold. If so, determine that there is a potential energy waste point near the current stage switching time, and record the location information of the potential energy waste point. The location information includes at least the timestamp of its occurrence and the smelting stage to which it belongs.

[0053] Specifically, resistance characteristic data is extracted from the initial power factor adjustment model. This data is collected within a set time window before and after the stage switching moments in the smelting process. The resistance change rate is calculated based on this data, and the stage switching time points are accurately recorded. For example, in the smelting process from solid to liquid, the resistance change typically exhibits significant fluctuations. By analyzing the resistance data sequence, the resistance change rate can be calculated, representing the magnitude of resistance change per unit time. Key feature vectors are constructed by combining the stage switching time points, forming important parameters describing the dynamic changes in the smelting process.

[0054] The magnitude of the eigenvector is calculated by taking the square root of the sum of squares of its components. This value comprehensively reflects the severity and timing of resistance changes. The calculated magnitude is compared to a preset vector threshold. If the magnitude exceeds the threshold, it indicates abnormal resistance fluctuations during the current switching process, requiring the initiation of a power surge risk prediction procedure. For example, if the magnitude is greater than a preset threshold (e.g., 0.5), it indicates significant resistance fluctuations that could lead to a power surge. This process identifies potential power surge risks by calculating the magnitude of resistance changes, ensuring timely response to potential risks during the smelting process.

[0055] The neural network model processes key feature vectors from the input and outputs a risk probability value based on the nonlinear relationship between the rate of change of resistance and the stage switching time. This risk probability value reflects the likelihood of a power surge near the current stage switching moment. For example, in scrap steel smelting, if the scrap steel has a high compositional diversity, a large rate of change of resistance, and a modulus exceeding a threshold, the neural network model will output a high risk probability value, such as 0.8, indicating an 80% probability of a power surge. This risk prediction helps identify potential energy waste points and optimize power regulation strategies. Conversely, for scrap steel with a more homogeneous composition, the resistance change is smaller, the risk probability value is lower, and therefore, immediate adjustment of the power input is not necessary.

[0056] The neural network employs a multilayer perceptron structure, and its training data comes from a dataset of resistance characteristics marked with power surge events during historical smelting processes. The neural network model maps and learns the nonlinear relationships in the input feature vector through hidden layers, learning the impact of different smelting stages and scrap steel composition on resistance fluctuations. The output layer generates a risk probability value between 0 and 1, representing the likelihood of a power surge occurring under the current feature conditions. During the training phase, input data such as the rate of resistance change and time points are mapped to a high-dimensional space. The model constructs decision boundaries by calculating the similarity of support vectors, thereby achieving accurate prediction of power surge risk.

[0057] The risk probability value is compared with a preset risk threshold. If the probability value exceeds the risk threshold, a potential energy waste point is identified near the current stage switching time, and the location information of the potential energy waste point, including the timestamp and the smelting stage to which it belongs, is automatically recorded. For example, during the switching process from the refining stage to the tapping stage, when the resistance change rate is detected to be 0.8 ohms per minute and the stage switching time is at the 45th minute, the calculated key feature vector magnitude is 0.9, exceeding the preset threshold of 0.5. After analysis by the neural network model, the output risk probability value is 0.85, exceeding the risk threshold of 0.7, thus marking the area near this time point as a potential energy waste area.

[0058] This technology enables real-time monitoring of resistance fluctuations and prediction of potential power surges, thereby optimizing power regulation. Particularly in scrap steel smelting, the diversity of components can cause resistance variations. Through neural network-based risk prediction, proactive measures can be taken during actual smelting to avoid energy waste caused by power surges, thus improving overall smelting efficiency and equipment stability. This solves the problems of traditional power regulation methods failing to accurately predict power surges and responding slowly to resistance fluctuations, ensuring intelligent, precise, and energy-efficient optimization of the smelting process.

[0059] Step S3: Construct a dynamic adjustment sequence based on potential energy waste points, and use the gradient descent algorithm to optimize the parameters of the dynamic adjustment sequence to generate a smooth transition control strategy.

[0060] In one specific embodiment, the process of performing step S3 may specifically include the following steps:

[0061] The location information of potential energy waste points is obtained, and a dynamic adjustment sequence is generated accordingly. Each waste point in the dynamic adjustment sequence corresponds to an adjustment instruction, which includes the power adjustment magnitude and the adjustment time interval.

[0062] With the optimization objective of minimizing the power spikes and energy consumption deviations caused by the dynamically adjusted sequence during execution, a corresponding loss function is defined.

[0063] The gradient descent algorithm is used to iteratively update the power adjustment magnitude and adjustment time interval of each adjustment command in the sequence, starting with the dynamically adjusted sequence, until the loss function converges to a predetermined range, and the optimized sequence parameters are obtained.

[0064] Based on the optimized sequence parameters, a smooth transition control strategy is generated.

[0065] Specifically, the location information of each potential energy waste point includes the timestamp of that point and its corresponding smelting stage (such as preheating, melting, refining, etc.). After acquiring the waste point location data containing the timestamp and the corresponding smelting stage, a corresponding adjustment instruction is generated for each marked waste point. These instructions together constitute the initial dynamic adjustment sequence. Each adjustment instruction includes a power adjustment range and an adjustment time interval. The power adjustment range indicates the amount of adjustment required to the current power input, and the adjustment time interval specifies how long the adjustment should take to complete. For example, the power adjustment range is set to an initial value, such as a gradual range of 5% to 10%, and the adjustment time interval is set to an interval sequence of 1 to 5 seconds to ensure that the sequence covers the locations of all waste points, flexibly responding to fluctuations in resistance characteristics, thereby generating the initial dynamic adjustment sequence.

[0066] To improve the smoothness of power regulation, a corresponding loss function is established with the optimization objective of minimizing power fluctuations and energy consumption deviations caused during the execution of the dynamic adjustment sequence. This loss function quantifies the degree of fluctuation in the power output sequence and its deviation from the theoretical energy consumption, forming a numerical representation of the adjustment strategy's effectiveness. Based on this, a gradient descent algorithm is used to optimize the dynamic adjustment sequence. Gradient descent is an iterative optimization method that aims to optimize the parameters of power adjustment amplitude and time interval by minimizing the loss function. The optimization process starts with the initial dynamic adjustment sequence and gradually updates the parameter values ​​of each adjustment command along the negative gradient direction by calculating the gradient direction of the loss function with respect to the power adjustment amplitude and time interval parameters. During the parameter update process, the loss function value is continuously monitored. When the function value decreases to a predetermined range and the change tends to stabilize, the optimization process is considered to have converged, at which point the optimized sequence parameters are obtained. For example, when an energy waste point with a timestamp of 25 minutes is detected during the refining stage, the initial settings may be a power adjustment range of 10% and an adjustment time interval of 2 seconds. After multiple iterations of the gradient descent algorithm, the parameters may be optimized to a power adjustment range of 6% and an adjustment time interval of 3 seconds. This combination of parameters significantly reduces the variability of power output.

[0067] Based on the optimized sequence parameters, a smooth transition control strategy is finally generated. This strategy establishes a continuous mapping relationship between power values ​​and time variables, ensuring that the power output exhibits smooth changes during the smelting process, i.e., a smooth transition from one value to another, avoiding equipment shocks and energy losses caused by abrupt parameter changes. This strategy avoids power abrupt changes and instabilities during the transition process by describing the continuous change law of power adjustment. Specifically, the generated strategy is like a smooth interpolation curve, ensuring stable power input changes during the smelting process and reducing equipment pressure or energy waste caused by abrupt changes.

[0068] For example, in the initial stage of scrap steel smelting, if gradient descent optimization reduces the adjustment amplitude from 10% to 6% and extends the adjustment time interval from 2 seconds to 4 seconds, the optimized sequence will effectively mitigate the impact of resistance characteristic fluctuations, stabilize smelting efficiency, and reduce energy consumption. Similarly, in the refining stage, if the waste point occurs at the 20-minute mark, after five rounds of optimization, the adjustment amplitude may be reduced to 4%, and the time interval extended to 1.5 seconds. The resulting control strategy ensures continuous power changes, avoids abrupt changes, and effectively extends the equipment's lifespan.

[0069] In this technical solution, the gradient descent algorithm and the smelting process parameters form a functionally mutually supportive relationship. The gradient descent algorithm can optimize power regulation strategies based on real-time feedback and resistance changes in a dynamic environment, smoothly transitioning and reducing energy loss. This not only improves equipment operational stability and smelting efficiency but also significantly reduces energy waste.

[0070] Step S4: Extract adaptive adjustment parameters from the smooth transition control strategy. If the adaptive adjustment parameters match the information of the current smelting stage, then integrate the real-time monitored resistance characteristic change trend to determine the final power factor value for the current smelting stage.

[0071] In one specific embodiment, the process of performing step S4 may specifically include the following steps:

[0072] Adaptive adjustment parameters are extracted from the smooth transition control strategy. These adaptive adjustment parameters include the power adjustment coefficient and the stage matching factor.

[0073] The stage matching factor is compared with the current smelting stage information. If the matching degree is higher than the preset matching threshold, it is determined that the smooth transition control strategy is applicable to the current smelting stage.

[0074] Real-time acquisition of resistance characteristic data, quantitative analysis of its changing trend through the slope of resistance change, and calculation of power factor adjustment value based on the slope of resistance change using a preset adjustment function;

[0075] The power adjustment coefficient, power factor adjustment value, and reference power factor are combined to generate the final power factor value.

[0076] Specifically, the adaptive adjustment parameters extracted from the smooth transition control strategy include a power adjustment coefficient and a stage matching factor. The power adjustment coefficient is used to scale the value for resistance fluctuations, dynamically adjusting the power factor to adapt to energy consumption changes in different stages of the smelting process. Its value is typically set between 0.8 and 1.2, reflecting the degree of adaptation to the current resistance fluctuation level. The stage matching factor reflects the similarity between the current smelting stage and the stage in the smooth transition control strategy, i.e., the adaptability. Its value is obtained by comparing the consistency between the strategy's preset stage and the actual operating stage.

[0077] After extracting the parameters, the stage matching factor is compared with a preset matching threshold. When the matching degree is higher than the threshold, it is determined that the current control strategy is applicable to the ongoing smelting stage, and the dynamic power factor adjustment process is initiated. This determination mechanism ensures the stage-specificity of the control strategy and avoids adjustment failures caused by stage mismatch. After confirming stage matching, the system collects resistance characteristic data sequences in real time and obtains the resistance change slope by calculating the rate of change of resistance value within a continuous time window. This slope quantifies the dynamic trend of resistance characteristic changes. If the matching degree is lower than the threshold, it indicates that the control strategy is not applicable to this smelting stage, and the strategy needs to be reselected or adjusted.

[0078] The slope of resistance change is derived by calculating the rate of change of resistance over time. This slope effectively reflects the dynamic characteristics of resistance during the smelting process, thus revealing the trend of energy consumption changes. In scenarios such as scrap steel smelting, resistance changes are usually closely related to factors such as furnace temperature and composition fluctuations. Therefore, monitoring the trend of resistance changes can help accurately identify energy waste points in the smelting process and make optimization adjustments. Noise filtering techniques, such as the moving average method, also need to be applied when calculating the slope to ensure the stability and accuracy of the data, thereby avoiding erroneous decisions caused by instantaneous fluctuations.

[0079] Based on the slope of the resistance change, the system calculates the power factor adjustment value using a preset adjustment function. This function maps the resistance change slope to a corresponding power factor correction, and the mapping relationship is obtained through training with historical smelting data. For example, when a positive resistance change slope is detected during the melting stage, the adjustment function outputs a corresponding power factor increment; conversely, when a negative slope is encountered during the refining stage, it outputs a decrement. This mapping relationship reflects the process correlation between resistance characteristics and power parameters. This process allows the power factor to be optimized and adjusted according to the actual needs of the smelting stage, ensuring maximum energy efficiency during the smelting process.

[0080] Finally, the power adjustment coefficient, the power factor adjustment value, and the baseline power factor are fused to generate the final power factor value. The baseline power factor originates from the output of the initial power factor adjustment model, the power adjustment coefficient provides a macroscopic adjustment scale, and the power factor adjustment value reflects real-time fluctuation requirements. The weighted fusion of these three values ​​generates the final power factor value, which maintains the baseline characteristics required by the process while incorporating a dynamic response to real-time operating conditions. This fusion process enables the final power factor value to simultaneously respond to macroscopic power demands caused by compositional diversity (reflected by the power adjustment coefficient) and real-time resistance fluctuations (reflected by the power factor adjustment value), thereby achieving adaptive power regulation for the current smelting stage.

[0081] By employing the aforementioned technical methods, dynamic adjustments can be made in real time during scrap steel smelting based on the changing trends of resistance characteristics, optimizing the power factor, reducing energy waste, and improving smelting efficiency. Especially in the initial and refining stages of the smelting process, the different resistance trends allow the power factor adjustment value to accurately reflect the actual needs of the current stage, effectively improving smelting efficiency and reducing equipment load. This addresses the adverse effects of high component diversity on smelting efficiency and significantly improves the accuracy and stability of energy efficiency control. By integrating adaptive adjustment parameters with resistance characteristic changing trends, a highly efficient and stable energy consumption optimization method is provided. This not only meets the precise control requirements of high-energy-consuming production processes such as scrap steel smelting but also effectively reduces energy waste, extends equipment life, and improves production efficiency.

[0082] Step S5: Based on the equipment loss monitoring data obtained during operation of the final power factor value, the support vector machine algorithm is used to evaluate the equipment loss level. If the loss level is lower than the preset loss threshold, an optimized power input sequence is generated.

[0083] In one specific embodiment, the process of performing step S5 may specifically include the following steps:

[0084] During the operation of the smelting equipment based on the final power factor value, equipment loss monitoring data is acquired, which includes equipment temperature and current loss values.

[0085] Input the equipment loss monitoring data into a pre-trained support vector machine model to obtain the loss level assessment results;

[0086] Determine whether the loss level assessment result is lower than the preset loss threshold. If so, determine that the current operating strategy based on the final power factor value is effective.

[0087] The smooth transition control strategy and the final power factor value are converted and serialized to generate an optimized power input sequence.

[0088] Specifically, during the operation of the equipment based on the final power factor value, real-time sensors collect data on equipment temperature and current loss. This data is closely related to the actual conditions of the smelting process; equipment temperature reflects the furnace's thermal load status, and current loss characterizes the energy conversion efficiency. For example, in scrap steel smelting, when the power factor is at a certain value, the equipment temperature may reach 1500 degrees Celsius, while the current loss is 5 amperes. By collecting this data, the equipment's operating status can be monitored in a timely manner, reflecting the impact of the power factor on the equipment's thermal stress and energy loss.

[0089] Support Vector Machine (SVM) is a supervised learning method that classifies data by finding the maximum margin hyperplane. The SVM model has been trained on historical smelting data, including samples of normal and abnormal losses. A kernel function maps the input features to a high-dimensional feature space, establishing the classification boundary between normal and abnormal losses. The input equipment temperature and current loss values, after being classified by the SVM model, generate a loss level assessment result. This assessment result determines the equipment's loss level by calculating the distance between the feature vector and the support vectors, thus reflecting the equipment's energy efficiency. For example, when the equipment temperature is 1500 degrees Celsius and the current loss is 5 amps, the model might classify this data as "high loss" or "low loss" and generate a corresponding loss level assessment result, characterizing the current energy efficiency status of the equipment.

[0090] The preset loss threshold can be set according to actual production needs; for example, the preset loss threshold can be 0.5. When the loss level assessment result output by the support vector machine model is lower than 0.5, it indicates that the equipment has high energy efficiency, effectively reducing the risk of power fluctuations and further optimizing the generation of the power input sequence. If the assessment result is lower than the preset loss threshold, it indicates that the equipment is in a high-efficiency, low-consumption operating range at the current power factor, and the operating strategy based on the final power factor value is deemed effective. This judgment mechanism establishes the correlation between equipment status and power strategy, providing a decision-making basis for power sequence optimization. For example, when the equipment temperature is 1420 degrees Celsius and the current loss is 4.3 amperes, the loss level assessment result output by the support vector machine model is 0.42, which is lower than the preset threshold of 0.5, thus confirming that the current power adjustment strategy has achieved the expected effect.

[0091] After confirming the effectiveness of the operating strategy, the power sequence generation process is initiated. The smooth transition control strategy is integrated and transformed with the final power factor value. Through time-dimensional serialization, an optimized power input sequence that is continuous in time and matches the smelting stage is generated. This sequence retains the continuity of the smooth transition control strategy while incorporating a validated power factor setpoint, ensuring a stable transition of power output over time. For example, combining a power factor of 0.88 with a smooth adjustment parameter generates a power sequence that smoothly increases from 800 kW to 950 kW within 180 seconds, ensuring a smooth transition in equipment power demand and avoiding large fluctuations. This optimized sequence allows for precise control of equipment power input during production processes such as scrap steel smelting, mitigating the impact of resistance fluctuations on the smelting process, further stabilizing smelting efficiency indicators, and enhancing the cyclic optimization effect of the adaptive adjustment mechanism.

[0092] This invention significantly improves equipment energy efficiency management, especially in the complex scrap steel smelting process. By acquiring real-time equipment loss monitoring data and combining it with support vector machine algorithms for evaluation, it can accurately determine the equipment's energy efficiency and promptly identify and resolve potential energy waste issues. Compared to traditional smelting control methods, this invention improves equipment operational stability, reduces energy waste, and enhances overall production efficiency through intelligent loss assessment and optimization adjustments. In particular, it effectively controls energy fluctuations caused by the diversity of components during the smelting process.

[0093] Step S6: Calibrate the resistance characteristic fluctuations using the optimized power input sequence to obtain a stable smelting efficiency index.

[0094] In one specific embodiment, the process of performing step S6 may specifically include the following steps:

[0095] The optimized power input sequence is applied to the melting equipment to perform power control, while the resistance characteristic fluctuation data caused by the diversity of components is monitored in real time.

[0096] The resistance characteristic fluctuation data is compared with the expected effect of the optimized power input sequence. By dynamically adjusting the subtle parameters of the power output, the resistance fluctuation is suppressed and compensated, and calibrated resistance characteristic data is generated.

[0097] Based on the calibrated resistance characteristic data, the smelting efficiency index, which characterizes the overall performance of the smelting process, is calculated. The smelting efficiency index includes energy consumption efficiency and smelting stability.

[0098] Determine whether the smelting efficiency index simultaneously meets the preset energy consumption efficiency threshold and smelting stability threshold. If so, determine that the current smelting process has reached a stable state and output the smelting efficiency index.

[0099] Specifically, an input sequence containing continuous power values ​​is applied to the actuator of the smelting equipment, and the power regulator enables precise control of the electric furnace power supply parameters. Simultaneously with power control, a resistance monitoring device continuously collects resistance characteristic fluctuation data caused by the diversity of scrap steel composition, forming a resistance time-series record synchronized with the power sequence.

[0100] During the smelting process, resistance fluctuations are typically influenced by raw material composition, temperature variations, and current load fluctuations. For example, resistance fluctuations in scrap steel smelting may be caused by metals with different compositions, which can affect equipment energy efficiency and smelting stability. By monitoring these fluctuations in real time, power input can be adjusted promptly to compensate for resistance fluctuations. The collected resistance characteristic fluctuation data is compared in real time with the expected effect of the power input sequence. This comparison process identifies the fluctuation range requiring compensation by calculating the deviation between the actual and expected resistance values. Based on the deviation analysis results, subtle parameters of the power output are dynamically adjusted, employing a proportional compensation mechanism to suppress resistance fluctuations. This dynamic adjustment is achieved by modifying local parameters of the power sequence, maintaining the overall power change trend while smoothing instantaneous fluctuations, ultimately generating calibrated resistance characteristic data with significantly reduced fluctuation amplitude. Each adjustment calculates a calibration coefficient based on the specific circumstances of the resistance characteristic fluctuations and applies it to the power input control of the smelting equipment, thereby smoothing the resistance fluctuation curve and reducing energy waste caused by fluctuations.

[0101] After the resistance characteristic fluctuations are effectively calibrated, the calibrated resistance characteristic data is used to calculate the smelting efficiency index, which characterizes the overall performance of the smelting process. This efficiency index comprehensively considers energy consumption efficiency and smelting stability, enabling a comprehensive evaluation of the overall performance of the smelting process. Energy consumption efficiency is calculated by the ratio of input power to the heat generated during the smelting process, reflecting the smelting output per unit of power. Smelting stability is measured by evaluating the stability of resistance fluctuations; smaller fluctuations in resistance characteristics indicate a more stable smelting process, thereby improving production efficiency and reducing energy waste. For example, the variance of the calibrated resistance characteristic data is calculated, and the smelting stability is calculated based on the reciprocal of the variance; a higher value indicates better stability.

[0102] Once the smelting efficiency indicators are calculated, it will be determined whether these indicators meet preset standards. For example, the set energy efficiency threshold is 80%, and the smelting stability threshold is 90%. If the calculated energy efficiency and smelting stability both meet these thresholds, it indicates that the current smelting process has reached a stable state, and a stable smelting efficiency indicator containing timestamps and indicator values ​​can be output.

[0103] The technical solution of this invention can effectively address the resistance fluctuation problem during the smelting process. By adjusting the power input and calibrating the resistance characteristics in real time, it optimizes the equipment's operating status and achieves stable smelting efficiency. This process not only improves smelting efficiency and stability but also effectively reduces energy consumption, providing a new solution for the intelligent control of smelting equipment.

[0104] Step S7: Extract feedback data on the frequency and magnitude of power mutations from the smelting efficiency index. If the feedback data indicates a reduction in power mutations, update the initial power factor adjustment model.

[0105] In one specific embodiment, the process of performing step S7 may specifically include the following steps:

[0106] Feedback data for quantifying power stability is extracted from stable smelting efficiency indicators. The feedback data includes the frequency and magnitude of power fluctuations.

[0107] The frequency and magnitude of power mutations are compared with their respective preset mutation thresholds. If both are lower than their thresholds, the power mutations are determined to have decreased, triggering a model update.

[0108] Based on the feedback data, an iterative optimization algorithm is used to adjust the internal parameters of the initial power factor adjustment model, obtain the updated initial power factor adjustment model, and use it for power regulation in the next round of scrap steel smelting process.

[0109] Specifically, feedback data on power fluctuations is obtained by stabilizing smelting efficiency indicators. This data includes the frequency and amplitude of power fluctuations. The frequency of power fluctuations is obtained by counting the number of abnormal fluctuations in power value exceeding a set threshold per unit time, while the amplitude of power fluctuations is determined by calculating the peak-to-valley difference between these abnormal fluctuation points. This feedback data directly reflects the changes in power stability during the smelting process and can effectively reflect the stability and energy efficiency of the electric furnace operation.

[0110] After the feedback data is extracted, the frequency and amplitude of power fluctuations are compared with preset thresholds. If both are lower than their preset thresholds, it is considered that the power fluctuation phenomenon in the current smelting process has been effectively suppressed, and the power stability has reached the expected level. In this case, the model update process is triggered, that is, the parameters of the initial power factor adjustment model are adjusted according to the feedback data.

[0111] After confirming the need to update the model, an iterative optimization algorithm was used to correct the internal parameters of the initial power factor adjustment model. The optimization process used the frequency and magnitude of power fluctuations as optimization objectives. By calculating the deviation between the model output and the desired stability index, the weights and bias parameters in the model were gradually adjusted along the gradient direction. For example, when the frequency of power fluctuations decreased from 1.5 times per minute to 0.8 times per minute, the model parameters were adjusted accordingly to enhance adaptability to stable operating conditions; when the magnitude of power fluctuations decreased from 40 kW to 25 kW, the parameter adjustments focused on improving the smoothness of power output. After completing the parameter optimization, an updated initial power factor adjustment model was generated. This model retained the original structural framework, but the internal parameters were specifically adjusted based on the actual operating results. The updated model was stored in the system database and called upon when the next round of scrap steel smelting process was started, serving as the initial benchmark for the new round of power regulation.

[0112] For example, during scrap steel smelting, if feedback data shows that the power fluctuation frequency is once per minute and the amplitude is 30 kW, both lower than preset thresholds (e.g., a frequency threshold of twice per minute and an amplitude threshold of 50 kW), then it is determined that the power fluctuations have decreased, triggering an update to the initial power factor adjustment model. Based on the feedback data, the model parameters are optimized; for example, by increasing the weights to enhance the model's adaptability to low-fluctuation scenarios, and by reducing the bias according to the amplitude, to smooth the transition of power input. This adjustment can effectively reduce power fluctuations, improve the stability of the smelting process, thereby reducing energy waste and equipment wear.

[0113] Through this adaptive adjustment mechanism, the initial power factor adjustment model can be continuously optimized based on feedback data from different smelting stages, ensuring more refined and intelligent power control during the smelting process. For example, in scrap steel with a high iron content, feedback data shows a frequency of 0.8 times / minute and an amplitude of 20 kW. The system will optimize the model parameters based on this data, improving the power factor from 0.85 to 0.9, thereby improving smelting efficiency and reducing energy loss. In another scrap steel scenario, feedback data shows a power fluctuation frequency of 1.2 times / minute and an amplitude of 15 kW. The system will adjust only the model's bias parameters based on the frequency change, without changing the weights, thus maintaining the model's stability under high-frequency fluctuations and avoiding unnecessary power fluctuations.

[0114] This invention's technical solution, through continuous monitoring and optimization of the power factor adjustment model, not only improves the stability of the smelting process but also reduces energy waste caused by power fluctuations, thereby significantly improving the energy efficiency and production efficiency of the scrap steel smelting process. Through an intelligent feedback mechanism, it can adaptively adjust under different smelting conditions, ensuring that the power input during the smelting process is always in an optimal state, further enhancing the automation and accuracy of the smelting process and effectively solving the problems of frequent power fluctuations and low efficiency in traditional smelting processes.

[0115] The above describes the operation data processing method for refining electric arc furnace steelmaking in the embodiments of this application. The following describes the operation data processing system for refining electric arc furnace steelmaking in the embodiments of this application. Please refer to [link to relevant documentation]. Figure 2 The present application provides a schematic diagram of the structure of an operating data processing system for refining electric arc furnace steelmaking. The system includes:

[0116] Model generation unit 10 is used to acquire resistance characteristic data and smelting stage information during the scrap steel smelting process, and uses the support vector machine algorithm to analyze the influence of composition diversity on resistance characteristics and establish an initial power factor adjustment model.

[0117] The risk prediction unit 20 is used to extract key feature vectors representing the dynamic changes in resistance during stage switching from the initial power factor adjustment model. If the key feature vectors exceed the preset vector threshold, the risk of power mutation is predicted by a neural network algorithm, and potential energy waste points in the smelting process are identified.

[0118] The strategy generation unit 30 is used to construct a dynamic adjustment sequence based on potential energy waste points, and to optimize the parameters of the dynamic adjustment sequence using a gradient descent algorithm to generate a smooth transition control strategy.

[0119] The power determination unit 40 is used to extract adaptive adjustment parameters from the smooth transition control strategy. If the adaptive adjustment parameters match the information of the current melting stage, the final power factor value for the current melting stage is determined by integrating the real-time monitored trend of resistance characteristic changes.

[0120] The sequence generation unit 50 is used to evaluate the equipment loss level using a support vector machine algorithm based on the equipment loss monitoring data obtained during operation of the final power factor value. If the loss level is lower than the preset loss threshold, an optimized power input sequence is generated.

[0121] The fluctuation calibration unit 60 is used to calibrate the resistance characteristic fluctuation through the optimized power input sequence to obtain a stable melting efficiency index.

[0122] The model update unit 70 is used to extract feedback data on the frequency and magnitude of power mutations from the smelting efficiency index. If the feedback data indicates a reduction in power mutations, the initial power factor adjustment model is updated.

[0123] This application also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium, wherein the computer-readable storage medium stores instructions that, when executed on a computer, cause the computer to perform the steps of the operating data processing method for refining electric arc furnace steelmaking.

[0124] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for processing operating data in electric arc furnace steelmaking, characterized in that, The method includes: Step S1: Obtain resistance characteristic data and smelting stage information during the scrap steel smelting process, use support vector machine algorithm to analyze the impact of composition diversity on resistance characteristics, and establish an initial power factor adjustment model. Step S2: Extract key feature vectors representing the dynamic changes in resistance during stage switching from the initial power factor adjustment model. If the key feature vectors exceed the preset vector threshold, predict the risk of power mutation through a neural network algorithm and determine the potential energy waste points in the smelting process. Step S3: Construct a dynamic adjustment sequence based on potential energy waste points, and use the gradient descent algorithm to optimize the parameters of the dynamic adjustment sequence to generate a smooth transition control strategy; Step S4: Extract adaptive adjustment parameters from the smooth transition control strategy. If the adaptive adjustment parameters match the information of the current melting stage, then integrate the real-time monitored resistance characteristic change trend to determine the final power factor value for the current melting stage. Step S5: Based on the equipment loss monitoring data obtained during operation of the final power factor value, the support vector machine algorithm is used to evaluate the equipment loss level. If the loss level is lower than the preset loss threshold, an optimized power input sequence is generated. Step S6: Calibrate the resistance characteristic fluctuations using the optimized power input sequence to obtain a stable smelting efficiency index; Step S7: Extract feedback data on the frequency and magnitude of power mutations from the smelting efficiency index. If the feedback data indicates a reduction in power mutations, update the initial power factor adjustment model.

2. The method for processing operating data in electric arc furnace steelmaking according to claim 1, characterized in that, Step S1 includes: The resistance characteristics data and smelting stage information of scrap steel are collected in real time by sensors. The resistance characteristics data include voltage and current values, and the smelting stage information includes preheating stage, melting stage or refining stage. The resistance characteristic data and smelting stage information are input into a pre-trained support vector machine model. The support vector machine model classifies the nonlinear effects of component diversity on resistance characteristics and outputs classification results to characterize the degree of resistance fluctuation. Based on the classification results, an initial power factor adjustment model is generated using a mapping function or lookup table. The initial power factor adjustment model characterizes the mapping relationship between the diversity of scrap steel composition and the ideal power factor.

3. The method for processing operating data in electric arc furnace steelmaking according to claim 1, characterized in that, Step S2 includes: From the real-time data of the initial power factor adjustment model, the resistance characteristic data sequence within a set time window before and after the stage switching moment is extracted. The resistance change rate is calculated based on the resistance characteristic data sequence, and the corresponding stage switching time point is recorded. A key feature vector is constructed based on the resistance change rate and the stage switching time point. Calculate the magnitude of the key feature vector and determine whether the magnitude exceeds the preset vector threshold. If so, input the key feature vector into the pre-trained neural network model. The neural network model outputs a risk probability value that represents the possibility of a power mutation based on the nonlinear relationship in the key feature vectors; Determine whether the risk probability value exceeds the preset risk threshold. If so, determine that there is a potential energy waste point near the current stage switching time, and record the location information of the potential energy waste point. The location information includes at least the timestamp of its occurrence and the smelting stage to which it belongs.

4. The method for processing operating data in electric arc furnace steelmaking according to claim 1, characterized in that, Step S3 includes: The location information of potential energy waste points is obtained, and a dynamic adjustment sequence is generated accordingly. Each waste point in the dynamic adjustment sequence corresponds to an adjustment instruction, which includes the power adjustment magnitude and the adjustment time interval. With the optimization objective of minimizing the power spikes and energy consumption deviations caused by the dynamically adjusted sequence during execution, a corresponding loss function is defined. The gradient descent algorithm is used to iteratively update the power adjustment magnitude and adjustment time interval of each adjustment command in the sequence, starting with the dynamically adjusted sequence, until the loss function converges to a predetermined range, and the optimized sequence parameters are obtained. Based on the optimized sequence parameters, a smooth transition control strategy is generated.

5. The method for processing operating data in electric arc furnace steelmaking according to claim 1, characterized in that, Step S4 includes: Adaptive adjustment parameters are extracted from the smooth transition control strategy. These adaptive adjustment parameters include the power adjustment coefficient and the stage matching factor. The stage matching factor is compared with the current smelting stage information. If the matching degree is higher than the preset matching threshold, it is determined that the smooth transition control strategy is applicable to the current smelting stage. Real-time acquisition of resistance characteristic data, quantitative analysis of its changing trend through the slope of resistance change, and calculation of power factor adjustment value based on the slope of resistance change using a preset adjustment function; The power adjustment coefficient, power factor adjustment value, and reference power factor are combined to generate the final power factor value.

6. The method for processing operating data in electric arc furnace steelmaking according to claim 1, characterized in that, Step S5 includes: During the operation of the smelting equipment based on the final power factor value, equipment loss monitoring data is acquired, which includes equipment temperature and current loss values. Input the equipment loss monitoring data into a pre-trained support vector machine model to obtain the loss level assessment results; Determine whether the loss level assessment result is lower than the preset loss threshold. If so, determine that the current operating strategy based on the final power factor value is effective. The smooth transition control strategy and the final power factor value are converted and serialized to generate an optimized power input sequence.

7. The method for processing operating data in electric arc furnace steelmaking according to claim 1, characterized in that, Step S6 includes: The optimized power input sequence is applied to the melting equipment to perform power control, while the resistance characteristic fluctuation data caused by the diversity of components is monitored in real time. The resistance characteristic fluctuation data is compared with the expected effect of the optimized power input sequence. By dynamically adjusting the subtle parameters of the power output, the resistance fluctuation is suppressed and compensated, and calibrated resistance characteristic data is generated. Based on the calibrated resistance characteristic data, the smelting efficiency index, which characterizes the overall performance of the smelting process, is calculated. The smelting efficiency index includes energy consumption efficiency and smelting stability. Determine whether the smelting efficiency index simultaneously meets the preset energy consumption efficiency threshold and smelting stability threshold. If so, determine that the current smelting process has reached a stable state and output the smelting efficiency index.

8. The method for processing operating data in electric arc furnace steelmaking according to claim 1, characterized in that, Step S7 includes: Feedback data for quantifying power stability is extracted from stable smelting efficiency indicators. The feedback data includes the frequency and magnitude of power fluctuations. The frequency and magnitude of power mutations are compared with their respective preset mutation thresholds. If both are lower than their thresholds, the power mutations are determined to have decreased, triggering a model update. Based on the feedback data, an iterative optimization algorithm is used to adjust the internal parameters of the initial power factor adjustment model, obtain the updated initial power factor adjustment model, and use it for power regulation in the next round of scrap steel smelting process.

9. A data processing system for refining electric arc furnace steelmaking, used to implement the method described in any one of claims 1 to 8, characterized in that, The system includes: The model generation unit is used to acquire resistivity data and smelting stage information during the scrap steel smelting process, and uses the support vector machine algorithm to analyze the impact of compositional diversity on resistivity characteristics and establish an initial power factor adjustment model. The risk prediction unit is used to extract key feature vectors representing the dynamic changes in resistance during stage switching from the initial power factor adjustment model. If the key feature vectors exceed the preset vector threshold, the power change risk is predicted by a neural network algorithm, and potential energy waste points in the smelting process are identified. The strategy generation unit is used to construct a dynamic adjustment sequence based on potential energy waste points, and to optimize the parameters of the dynamic adjustment sequence using a gradient descent algorithm to generate a smooth transition control strategy. The power determination unit is used to extract adaptive adjustment parameters from the smooth transition control strategy. If the adaptive adjustment parameters match the information of the current melting stage, the final power factor value for the current melting stage is determined by integrating the real-time monitored resistance characteristic change trend. The sequence generation unit is used to evaluate the equipment loss level based on the equipment loss monitoring data obtained during operation of the final power factor value using the support vector machine algorithm. If the loss level is lower than the preset loss threshold, an optimized power input sequence is generated. The fluctuation calibration unit is used to calibrate the resistance characteristic fluctuation through the optimized power input sequence to obtain a stable melting efficiency index. The model update unit is used to extract feedback data on the frequency and magnitude of power mutations from the smelting efficiency index. If the feedback data indicates a reduction in power mutations, the initial power factor adjustment model is updated.

10. A computer-readable storage medium storing instructions thereon, characterized in that, When the instructions are executed by the processor, they implement the operating data processing method for refining electric arc furnace steelmaking as described in any one of claims 1 to 8.

Images