A casting process for low-oxygen copper

Abstract

This invention relates to the technical field of low-oxygen copper production, and in particular to a casting process for low-oxygen copper, comprising: S1 setting important properties of the copper billet, analyzing the influence of adjustable parameters, and setting an influence threshold; S2 inputting real-time fixed parameters and required performance values ​​into a prediction model to obtain predicted values ​​of adjustable parameters; S3 producing the copper billet; S4 acquiring abnormally important properties; S5 merging the optional sets corresponding to abnormally important properties to form a union; correcting the order of adjustable parameter calls to obtain a priority call set; S6 serializing the priority call set to obtain a priority adjustment sequence; S7 calling at least one adjustable parameter in the priority adjustment sequence for adjustment; by constructing a prediction model, the parameter values ​​that meet the specific performance requirements of low-oxygen copper can be predicted in advance, significantly improving the certainty and efficiency of the production process, and at the same time, enabling rapid location and resolution of performance anomalies.

Description

Technical Field

[0001] This invention relates to the technical field of low-oxygen copper production, and in particular to a casting process for low-oxygen copper. Background Technology

[0002] Low-oxygen copper is a copper material with extremely low oxygen content, typically not exceeding 0.003%. It possesses excellent electrical conductivity, good mechanical properties, and high thermal stability. These characteristics make low-oxygen copper widely used in fields such as power transmission, electronic component manufacturing, and communication equipment.

[0003] SCR (Sequencing Continuous Casting) is an advanced metal forming technology that can continuously cast molten copper into copper billets, offering good continuity and high production efficiency.

[0004] In the long-term practice of using the SCR method for continuous casting to prepare low-oxygen copper, traditional processes have revealed a series of problems that seriously restrict the development of the industry. Regarding the handling of adjustable parameters, traditional methods have been deeply mired in difficulties, relying solely on extremely limited past experience to make trial adjustments repeatedly in actual production. This approach not only fails to predict the parameter values ​​that meet the specific performance requirements of low-oxygen copper in advance, but also makes the entire production process full of uncertainty. When significant performance abnormalities occur in the produced low-oxygen copper billets, the shortcomings of traditional processes are even more exposed. On the one hand, it is difficult to quickly and accurately identify the core factors causing the performance abnormalities. The production process involves numerous links and factors, and traditional processes lack an effective troubleshooting mechanism, leaving technicians helpless and searching for solutions. On the other hand, even if some potential problem points are identified, there is no scientific and reasonable method to adjust the relevant parameters specifically. Usually, only simple and conventional measures can be taken, which often fail to address the essence of the problem and cannot fundamentally solve the performance abnormality issue.

[0005] The information disclosed in this background section is intended only to enhance the understanding of the overall background of the present invention and should not be construed as an admission or in any way implying that the information constitutes prior art known to those skilled in the art. Summary of the Invention

[0006] To address the aforementioned technical problems, this invention provides a casting process for low-oxygen copper. By constructing a predictive model, the parameter values ​​required to meet the specific performance requirements of low-oxygen copper can be predicted in advance, significantly improving the certainty and efficiency of the production process. At the same time, by prioritizing the use of sets and serialized adjustments, performance anomalies can be quickly located and resolved, fundamentally optimizing the low-oxygen copper casting process and improving product quality and production stability.

[0007] The casting process for low-oxygen copper of the present invention includes:

[0008] S1 sets the important properties of the copper billet, analyzes the influence of adjustable parameters on the important properties, sets the influence threshold, and obtains multiple selectable sets;

[0009] S2 inputs real-time fixed parameters and desired performance values ​​into a pre-built prediction model to obtain adjustable parameter prediction values;

[0010] S3 produces copper billets based on real-time fixed parameters and adjustable parameter predictions.

[0011] The S4 test detects important properties of copper billets, obtains performance values, and compares them with required performance values ​​to identify exceptionally important properties.

[0012] S5 merges the optional sets corresponding to exceptionally important performance into a union; based on the frequency, impact and deviation of the adjustable parameters in the union, the order of adjustable parameter calls is corrected to obtain the priority call set;

[0013] S6 uses preset external conditions to serialize and adjust the priority call set to obtain the priority adjustment sequence;

[0014] S7 calls the priority adjustment sequence to adjust at least one adjustable parameter.

[0015] The aforementioned solution, by constructing a predictive model and scientifically analyzing the impact of adjustable parameters on key performance characteristics, can predict in advance the values ​​of adjustable parameters that meet the specific performance requirements of low-oxygen copper. This significantly improves the certainty and efficiency of the production process, avoiding the inefficiency of relying on experience and repeated trial and error in traditional processes. By obtaining the optional set corresponding to abnormal key performance characteristics and correcting them by combining the frequency, impact, and deviation of adjustable parameters, the solution can quickly locate the core factors causing performance abnormalities and adjust relevant parameters accordingly, fundamentally solving the problem. Furthermore, by sequentially adjusting the priority call set through preset external conditions, the order and strategy of adjustable parameter adjustments are further optimized, ensuring the scientific nature and effectiveness of the adjustments. This solution not only improves the quality and performance stability of low-oxygen copper billets but also significantly increases production efficiency and reduces production costs, providing reliable technical support for the large-scale, high-quality production of low-oxygen copper.

[0016] As a preferred embodiment of the present invention, the important properties of the copper billet in S1 include: electrical conductivity, oxygen content, tensile strength, elongation, hardness and microstructure.

[0017] Conductivity is a key indicator for measuring the ability of copper materials to conduct current. For fields such as power transmission and electronic component manufacturing, high conductivity means lower resistance loss and higher efficiency. Therefore, it is crucial to ensure that copper billets have excellent conductivity.

[0018] Controlling oxygen content helps reduce oxide inclusions, improve the purity and conductivity of materials, and avoid the decline in mechanical properties caused by oxidation.

[0019] Tensile strength, elongation, and hardness reflect the response characteristics of a material under different stress conditions; tensile strength determines the maximum tensile force that a material can withstand; elongation indicates the material's deformation capacity before fracture; hardness affects the material's wear resistance and processing performance; good mechanical properties ensure the reliability and durability of copper castings in various application scenarios.

[0020] Microstructure directly affects the internal structure of materials, such as grain size, distribution, and the presence of defects; it not only affects the mechanical properties and electrical conductivity of materials, but also relates to the uniformity and consistency of the cast billet, thus affecting the quality of the final product.

[0021] Adjustable parameters include: cooling water flow rate and temperature on the crystallizing wheel side, cooling water flow rate and temperature on the steel strip side, casting temperature, billet pulling speed, and casting angle;

[0022] The flow rate and temperature of the cooling water on both the crystallizing wheel side and the steel strip side can affect the aforementioned important properties: an appropriate cooling rate can promote the formation of fine and uniform grains, which is beneficial to improving conductivity; too fast or too slow a cooling rate may affect the removal of oxygen, indirectly affecting the oxygen content; rapid cooling can inhibit the formation of coarse grains, which helps to improve mechanical properties such as tensile strength, elongation, and hardness; cooling conditions directly affect the crystal growth mode and size, thereby changing the microstructure.

[0023] Casting temperature can also affect the above-mentioned important properties: casting temperature directly affects the solidification process of molten copper, thereby affecting the microstructure and conductivity of the final product; higher casting temperature may lead to more oxygen dissolving in molten copper, increasing the oxygen content in the finished product; casting temperature that is too high or too low will affect the phase transformation path during solidification, thus affecting mechanical properties.

[0024] The casting speed can also affect the above-mentioned important properties: a faster casting speed may lead to uneven internal structure of the billet, affecting electrical conductivity; the casting speed affects the cooling rate, thus indirectly affecting the oxygen content; the casting speed also affects the uniformity and density of the internal structure of the billet, thereby affecting mechanical properties and possible internal defects.

[0025] The casting angle also affects the aforementioned important properties: the casting angle affects how the molten metal enters the five-wheel casting machine; an improper angle may lead to poor filling or the formation of air bubbles, indirectly affecting conductivity; the casting angle affects the flow of molten metal and the discharge of gas, thus affecting oxygen content; the casting angle affects the forming quality of the billet and the distribution of thermal stress; an incorrect angle may cause stress concentration points or cracks inside the billet, weakening mechanical properties; a reasonable casting angle can ensure that the molten metal flows smoothly into the mold, forming a dense and uniform microstructure, reducing internal defects, and improving overall performance.

[0026] As a preferred embodiment of the present invention, the method for obtaining the influence degree in S1 includes the following steps:

[0027] S11 obtains the setting values ​​of each adjustable parameter of copper billet from different production batches and the corresponding important performance results of the copper billet, and obtains historical production data.

[0028] Specifically, historical production data is obtained from platforms such as the Enterprise Production Management System (MES) and the Distributed Control System (DCS). Important performance test results for each batch of copper billets are extracted from the quality inspection system, and the timestamps of all data are ensured to be consistent. To ensure the quality of the collected data, data cleaning can be performed, missing values ​​can be filled by means of imputation and interpolation, outliers can be identified and corrected or removed by box plots and Z-scores, and if the dimensions of the parameters differ greatly, the data can be standardized.

[0029] Based on historical production data, S12 sets up a quantitative method to evaluate the independent influence of each adjustable parameter on the important properties of the copper billet, and obtains the parameter independent influence.

[0030] Based on the principles of metallurgy and the theory of materials science, S13 predicts the influence of adjustable parameters on the important properties of copper billets and obtains the prediction results of parameter influence.

[0031] Specifically, we consulted the latest metallurgical and materials science literature to understand the influence mechanism of each parameter on the performance of copper billets, and invited experts in the field to consult to ensure the scientific validity and rationality of the predictions.

[0032] Based on historical production data and parameter impact prediction results, S14 selects a quantitative method to analyze the joint influence of each adjustable parameter on the important properties of the copper billet, and obtains the joint influence of the parameters.

[0033] S15 assigns weights to the independent and joint influence of parameters and calculates the influence of each adjustable parameter on the important properties of the copper billet.

[0034] As a preferred embodiment, by combining the independent influence of a single parameter and the combined influence of multiple parameters, the impact of each parameter on the important properties of the copper billet can be comprehensively and scientifically quantified. Single-parameter analysis can assess the independent contribution of each adjustable parameter to the important properties, preliminarily screen out key parameters, and exclude irrelevant parameters with little impact on performance. Multi-parameter analysis can study the combined influence of multiple parameters on performance, especially the interaction between parameters. Before multi-parameter joint analysis, the influence of adjustable parameters on the important properties of the copper billet can be predicted. Its main purpose is to preliminarily determine which parameters may have interactive relationships from a theoretical and data perspective, narrow the scope of analysis, optimize experimental design and data analysis process. Through prediction, the interference of irrelevant parameters can be reduced, and targeted joint analysis can be carried out.

[0035] The quantitative methods include: statistical analysis, effect size measurement, regression analysis, machine learning, multi-criteria decision analysis, and response surface methodology.

[0036] Statistical analysis is a data-based approach used to describe and interpret patterns and trends in data. By calculating statistics such as mean, standard deviation, and correlation coefficient, the relationships between different parameters and their impact on external conditions can be assessed. Statistical analysis also supports hypothesis testing, helping to determine whether observed effects are statistically significant, thus providing reliable data support for decision-making.

[0037] Effect size measurement is used to quantify the specific impact of a factor on the outcome. This method not only focuses on statistical significance but also emphasizes the importance of actual effects. By measuring effect size, the actual impact of parameter adjustments can be assessed more accurately, ensuring that optimization strategies are not only statistically effective but also significant in practical applications.

[0038] Regression analysis is a statistical method used to model and analyze the relationships between variables. By establishing linear or nonlinear regression models, it can predict the impact of one or more independent variables (such as adjustable parameters) on a dependent variable (such as performance indicators). Regression analysis helps identify key influencing factors and provides quantitative predictive power, making the optimization process more scientific and reasonable.

[0039] Machine learning uses algorithms and statistical models to automatically learn from large amounts of data and make predictions or decisions. By training models, machine learning can identify complex patterns, discover hidden relationships in data, and make high-precision predictions. Common machine learning methods include decision trees, random forests, and support vector machines. These methods can help optimize parameter selection and improve production efficiency and product quality.

[0040] Multi-criteria decision analysis (MCDA) is a systematic approach to address decision problems involving multiple conflicting objectives. By defining different evaluation criteria (such as cost and response time) and using weighted summation or other comprehensive scoring methods, MCDA can assess the overall merits of different options. This approach ensures that various factors are comprehensively considered during the decision-making process and provides a scientific and reasonable basis for decision-making.

[0041] Response surface methodology (RSM) is a statistical and mathematical tool used for process design and improvement. By constructing designs of experiments (DOEs), RSM can systematically explore the parameter space and find optimal operating conditions. RSM can not only determine the best combination of parameters but also reveal the interactions between parameters, helping to understand complex processes and achieve efficient optimization.

[0042] Through the above steps, the present invention can systematically acquire and quantify the impact of each adjustable parameter on the important properties of low-oxygen copper billets, providing a scientific basis for subsequent adjustments and optimizations. This method not only improves the accuracy of the evaluation, but also provides solid data support for process improvement.

[0043] As a preferred embodiment of the present invention, the method for constructing the prediction model in S2 includes:

[0044] S21 collects historical production data, including fixed parameters, adjustable parameters, and corresponding key performance data of copper billets; the collected data is cleaned to remove outliers and noise, and then standardized.

[0045] S22 randomly samples the processed historical production data to generate several subset datasets;

[0046] Specifically, in order to enhance the robustness of the model and avoid excessive reliance on a single data distribution, the processed historical production data can be randomly sampled to generate several subsets. The sampling method can be sampling with replacement or sampling without replacement. The size of the subsets should be set in combination with the actual amount of data and the complexity of the model, usually 60%-80% of the original dataset.

[0047] S23 selects several differentiated base learners, trains the base learners on each subset of the dataset, and combines the prediction results of several base learners to obtain a prediction model;

[0048] Specifically, this step can select differentiated base learners, including decision trees, logistic regression, neural networks, etc. The goal is to make each model complementary in error distribution. Differentiated base learners can capture different patterns in the data, which helps to improve the generalization ability of the model. By combining the prediction results of multiple base learners, such as averaging and voting, the bias of a single model can be reduced and the overall performance can be improved.

[0049] S24 integrates the constructed prediction model into the production environment to predict adjustable parameters.

[0050] The predictive model construction method of this invention significantly improves the intelligence level and prediction accuracy of low-oxygen copper casting forming process, ensures the stability and reliability of the model, greatly improves production efficiency and product quality, reduces trial and error costs, promotes process optimization and technological progress, realizes data-driven accurate decision-making, adapts to the needs of different scenarios, and thus provides enterprises with more competitive and efficient solutions.

[0051] As a preferred embodiment of the present invention, the correction method in S5 includes:

[0052] S51 counts the number of times each adjustable parameter appears in the set;

[0053] Specifically, all optional sets corresponding to exceptionally important performance are merged into a union, and frequency statistics are performed on each adjustable parameter in the union to record the number of times each adjustable parameter appears.

[0054] By counting the occurrences, we can intuitively see which adjustable parameters repeatedly appear under multiple abnormal conditions, indicating that these adjustable parameters may be the core factors causing the problem. This helps technicians quickly identify key parameters that may affect multiple performance indicators and ensure that adjusting these parameters can improve performance in multiple aspects at the same time.

[0055] S52 obtains the impact of each adjustable parameter in the set on exceptionally important performance;

[0056] S53 calculates the deviation of abnormally important performance;

[0057] S54 assigns weights to the frequency of occurrence, influence, and deviation, respectively.

[0058] When assigning weights to frequency of occurrence, impact, and deviation, the following factors should be considered:

[0059] 1) Assess the impact of different deviations on product quality and production stability, determine which performance anomalies are most urgent, and appropriately increase the weight of deviation for performance problems with large deviations;

[0060] 2) By combining data from multiple performance anomalies, identify the key parameters that frequently appear in multiple anomalies. For parameters with high prevalence, appropriately increase the weight of their frequency of occurrence.

[0061] 3) Adjust the weights of frequency of occurrence, impact, and deviation based on the importance that enterprises attach to different performance indicators;

[0062] 4) If some performance values ​​are better than the optimal value of the required performance value, and the set of alternative performance values ​​and the union of the set of alternative performance values ​​appear in a union, then there may be some room for the corresponding performance values ​​to decrease. In this case, the influence of the adjustable parameters in the union can be appropriately reduced.

[0063] S55 calculates the overall priority of each adjustable parameter to obtain the priority call set:

[0064] More specifically, the aggregation priority is calculated using the following formula:

[0065] S = C × W1 + D × W2 + E × W3;

[0066] Where S is the overall priority of each adjustable parameter, C is the number of times the adjustable parameter appears in the set, W1 is the weight of the number of occurrences, D is the influence, W2 is the weight of the influence, E is the deviation, W3 is the weight of the deviation, and W1+W2+W3=1.

[0067] By comprehensively considering the frequency of occurrence, impact, and deviation, the influence of each adjustable parameter on the important properties of low-oxygen copper billets can be fully evaluated. A single indicator cannot fully reflect the importance of parameters in complex processes, while multi-dimensional evaluation can provide a more accurate judgment. The calculation of comprehensive priority ensures that high-priority adjustable parameters are given priority, thereby improving the overall process optimization effect, which is crucial for improving product quality, reducing scrap rate, and increasing production efficiency.

[0068] As a preferred embodiment of the present invention, the serialization adjustment method in S6 includes:

[0069] S61 evaluates the effect of each adjustable parameter in the priority call set on the applied condition;

[0070] When evaluating the effect value, the following factors should be considered: Adjusting certain adjustable parameters may bring unexpected risks or side effects. For example, while increasing the cooling rate can improve conductivity, it may lead to equipment overload or increased energy consumption, so its effect value should be appropriately reduced. By evaluating the effect value, these potential problems can be identified before adjustment, and corresponding preventive measures can be taken. At the same time, the adjustment of different adjustable parameters may involve different resource inputs (such as time, manpower, and funds). By evaluating the effect value, the rational allocation of resources can be ensured, and unnecessary waste and over-investment can be avoided. When certain adjustable parameters bring more resource input, their effect value should be appropriately reduced.

[0071] S62 calculates the call priority for each adjustable parameter in the priority call set;

[0072] More specifically, the calling priority of each adjustable parameter is calculated using the following formula:

[0073] S1 = S × W4;

[0074] Where S1 is the call priority of each adjustable parameter, S is the overall priority, and W4 is the effect value of the external condition.

[0075] The overall priority reflects the potential contribution of the adjustable parameter to important performance, while the additional conditional value reflects the risks and additional costs that may be brought about by adjusting the parameter. Multiplying the two can achieve a balance between risk and benefit, ensuring that parameters that can significantly improve performance without bringing excessive risks or costs are selected for adjustment.

[0076] S63 sorts the call priority of each adjustable parameter in descending order to obtain the priority adjustment sequence;

[0077] The call priority of all adjustable parameters is compared numerically and arranged in descending order. The higher the value of the adjustable parameter, the greater the product of its overall priority and the effect of the applied conditions, indicating that the adjustable parameter is more important to the overall optimization and should be called first.

[0078] As a preferred embodiment of the present invention, the external conditions in S6 include cost and response time;

[0079] Costs include energy consumption, material usage, and equipment wear and tear; response time is the time required to reach a stable state after adjusting adjustable parameters.

[0080] If adjusting a certain adjustable parameter would lead to a significant increase in energy consumption (such as increasing the cooling rate), its external condition effect value should be reduced accordingly. This helps to avoid additional costs caused by high energy consumption and ensures that the optimization process is economically feasible.

[0081] Adjusting certain adjustable parameters may lead to material waste or additional raw material requirements (such as increasing the proportion of additives), which will increase production costs. Therefore, their effective values ​​should be appropriately reduced.

[0082] If adjusting an adjustable parameter may cause equipment overload or accelerate equipment wear (such as by increasing pressure or temperature), its value should be reduced accordingly. This helps to extend the equipment's lifespan and reduce maintenance costs.

[0083] If adjusting a certain parameter requires a long response time to reach a stable state, this will affect production efficiency, increase production cycle and cost. Therefore, its value should be appropriately reduced.

[0084] When evaluating the effect of the above-mentioned external conditions, a weighted summation method is used for quantification. Specifically, for each adjustable parameter, its specific impact on cost and response time is defined, and then a reasonable weight is assigned to each specific impact value. Finally, a weighted calculation is performed to obtain the effect value.

[0085] As a preferred embodiment of the present invention, at least one adjustable parameter includes the adjustable parameter that is first in the priority adjustment sequence;

[0086] The most important adjustable parameter is the adjustment point that is most critical to solving the current performance anomaly. By adjusting the most important adjustable parameter first, the scope of the problem can be quickly narrowed down, the most effective solution can be found, and time and resources can be avoided by wasting time and resources on unimportant parameters. At the same time, the most important parameter can also take into account production costs and response time, which helps to achieve performance improvement with minimal cost input and improve economic efficiency.

[0087] As a preferred embodiment of the present invention, the fixed parameters in S2 include: crystallizing wheel diameter, crystallizing wheel groove thickness, crystallizing wheel groove carbon layer thickness, and steel strip carbon layer thickness;

[0088] Since different plants may have different fixed parameters, it is crucial to use these parameters as input variables for the predictive model.

[0089] The diameter of the crystallizing wheel directly affects the forming process of the billet and the dimensional accuracy of the final product. A larger crystallizing wheel diameter can provide a longer cooling path, which helps to improve the cooling uniformity and structural stability of the billet, thereby improving electrical conductivity and mechanical strength.

[0090] The thickness of the crystallizer groove determines the thickness and forming quality of the billet. An appropriate groove thickness can ensure that the billet has good surface quality and dimensional accuracy, while reducing the occurrence of defects (such as cracks and porosity). If the groove thickness is too large, it may slow down the cooling rate and affect the microstructure of the billet. If the groove thickness is too small, it may cause the cooling rate to be too fast, resulting in stress concentration and deformation.

[0091] The carbon layer thickness in the crystallizing wheel groove serves as insulation and lubrication, helping to reduce friction between the billet and the crystallizing wheel and prevent sticking. An appropriate carbon layer thickness can improve the surface finish of the billet and the demolding effect, while reducing equipment wear. If the carbon layer is too thick, it may affect the cooling effect of the billet and lead to a decrease in performance; while if the carbon layer is too thin, it may cause the billet to stick together and increase the difficulty of production.

[0092] The thickness of the carbon layer on the steel strip also serves as insulation and lubrication, protecting the steel strip from high-temperature damage and ensuring that the billet can be smoothly separated from the steel strip. An appropriate carbon layer thickness can improve the surface quality of the billet and production efficiency, and reduce the wear and maintenance needs of the steel strip. If the carbon layer is too thick, it may lead to uneven heat transfer, affecting the cooling effect and performance of the billet; while if the carbon layer is too thin, it may cause adhesion problems and reduce production stability.

[0093] As a preferred embodiment of the present invention, the method for calculating the deviation in S53 includes:

[0094] S531 obtains the required performance values ​​and the performance values ​​of the copper billet;

[0095] S532 uses the following formula to calculate the deviation:

[0096] R = |Q1-Q2| / Q2;

[0097] Where R is the deviation, Q1 is the performance value of the copper billet, and Q2 is the required performance value.

[0098] Deviation can accurately quantify the risk level of each performance indicator. A higher deviation means a larger deviation, which may bring higher risks or potential problems. By monitoring changes in deviation, potential risks can be identified and dealt with in advance, ensuring the stability and reliability of the production process.

[0099] Compared with existing technologies, the beneficial effects of this invention are as follows: By introducing advanced predictive models and scientific analysis methods, the intelligent level of the low-oxygen copper casting process is greatly improved. First, the solution can predict the parameter values ​​that meet specific performance requirements in advance, thereby significantly improving the certainty and efficiency of the production process. In traditional processes, due to the lack of effective predictive tools, the selection of adjustable parameters often relies on limited experience and repeated trial and error, which not only wastes a lot of time and resources, but also makes the production process full of uncertainty. The new solution uses predictive models built with advanced technologies such as machine learning, which can accurately predict the expected performance results under different parameter settings based on historical data, enabling production enterprises to have a clear understanding of the final product quality before actual casting, thereby optimizing parameter selection and reducing unnecessary adjustments and tests.

[0100] Secondly, when faced with significant performance anomalies in the produced low-oxygen copper billets, this invention provides a systematic method to quickly locate and resolve the core issues. Unlike traditional processes, the new solution effectively identifies key factors by merging and analyzing the available set of options corresponding to the abnormal performance, and combining factors such as the frequency, impact, and deviation of adjustable parameters. This method not only accelerates troubleshooting but also ensures the pertinence and effectiveness of adjustment measures, fundamentally solving the performance anomaly problem. Furthermore, by evaluating the value of external conditions, it further ensures that the adjustment measures taken can significantly improve performance without incurring excessive risks or costs, achieving a balance between risk and benefit. Attached Figure Description

[0101] Figure 1 This is a flowchart illustrating the present invention;

[0102] Figure 2This is a flowchart illustrating the method in step S5 for correcting the order of adjustable parameter calls based on the frequency, impact, and deviation of abnormally important performance of the adjustable parameters. Detailed Implementation

[0103] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0104] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0105] Secondly, the term "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places throughout this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments. Example

[0106] Reference Figure 1 This embodiment provides a casting process for low-oxygen copper, including:

[0107] S1 sets the important properties of the copper billet, analyzes the influence of adjustable parameters on the important properties, sets the influence threshold, and obtains multiple selectable sets;

[0108] Analyzing the impact of adjustable parameters on key performance is to gain a deeper understanding of how and to what extent various operating conditions affect product quality. This analysis helps establish a causal relationship between parameter adjustments and performance changes, providing a theoretical basis for developing scientific and reasonable adjustment strategies.

[0109] The important properties of copper billets include: electrical conductivity, oxygen content, tensile strength, elongation, hardness, and microstructure.

[0110] Conductivity is a key indicator for measuring the ability of copper materials to conduct current. For fields such as power transmission and electronic component manufacturing, high conductivity means lower resistance loss and higher efficiency. Therefore, it is crucial to ensure that copper billets have excellent conductivity.

[0111] Controlling oxygen content helps reduce oxide inclusions, improve the purity and conductivity of materials, and avoid the decline in mechanical properties caused by oxidation.

[0112] Tensile strength, elongation, and hardness reflect the response characteristics of a material under different stress conditions; tensile strength determines the maximum tensile force that a material can withstand; elongation indicates the material's deformation capacity before fracture; hardness affects the material's wear resistance and processing performance; good mechanical properties ensure the reliability and durability of copper castings in various application scenarios.

[0113] Microstructure directly affects the internal structure of materials, such as grain size, distribution, and the presence of defects; it not only affects the mechanical properties and electrical conductivity of materials, but also relates to the uniformity and consistency of the cast billet, thus affecting the quality of the final product.

[0114] Adjustable parameters include: cooling water flow rate and temperature on the crystallizing wheel side, cooling water flow rate and temperature on the steel strip side, casting temperature, billet pulling speed, and casting angle;

[0115] The flow rate and temperature of the cooling water on both the crystallizing wheel side and the steel strip side can affect the aforementioned important properties: an appropriate cooling rate can promote the formation of fine and uniform grains, which is beneficial to improving conductivity; too fast or too slow a cooling rate may affect the removal of oxygen, indirectly affecting the oxygen content; rapid cooling can inhibit the formation of coarse grains, which helps to improve mechanical properties such as tensile strength, elongation, and hardness; cooling conditions directly affect the crystal growth mode and size, thereby changing the microstructure.

[0116] Casting temperature can also affect the above-mentioned important properties: casting temperature directly affects the solidification process of molten copper, thereby affecting the microstructure and conductivity of the final product; higher casting temperature may lead to more oxygen dissolving in molten copper, increasing the oxygen content in the finished product; casting temperature that is too high or too low will affect the phase transformation path during solidification, thus affecting mechanical properties.

[0117] The casting speed can also affect the above-mentioned important properties: a faster casting speed may lead to uneven internal structure of the billet, affecting electrical conductivity; the casting speed affects the cooling rate, thus indirectly affecting the oxygen content; the casting speed also affects the uniformity and density of the internal structure of the billet, thereby affecting mechanical properties and possible internal defects.

[0118] The casting angle also affects the aforementioned important properties: the casting angle affects how the molten metal enters the five-wheel casting machine; an improper angle may lead to poor filling or the formation of air bubbles, indirectly affecting conductivity; the casting angle affects the flow of molten metal and the discharge of gas, thus affecting oxygen content; the casting angle affects the forming quality of the billet and the distribution of thermal stress; an incorrect angle may cause stress concentration points or cracks inside the billet, weakening mechanical properties; a reasonable casting angle can ensure that the molten metal flows smoothly into the mold, forming a dense and uniform microstructure, reducing internal defects, and improving overall performance;

[0119] Preferably, the method for obtaining the degree of influence includes the following steps:

[0120] S11 is based on metallurgical principles and materials science theory to predict the potential impact of each adjustable parameter on important performance.

[0121] Specifically, we consulted the latest metallurgical and materials science literature to understand the influence mechanism of each parameter on the performance of copper billets, and invited experts in the field to consult to ensure the scientific validity and rationality of the predictions.

[0122] S12 obtains the setting values ​​of each adjustable parameter of copper billet from different production batches and the corresponding important performance results of the copper billet.

[0123] Specifically, historical production data is obtained from platforms such as the Enterprise Production Management System (MES) and the Automated Control System (DCS), and the important performance test results of each batch of copper billets are extracted from the quality inspection system to ensure that the timestamps of all data are consistent, which facilitates subsequent data cleaning and analysis.

[0124] S13 cleans and preprocesses the acquired historical production data; sets up quantitative methods to evaluate the impact of each adjustable parameter on the important properties of the copper billet; the quantitative methods include statistical analysis, effect size measurement, regression analysis, machine learning, multi-criteria decision analysis, and response surface methodology;

[0125] Specifically, by setting an influence threshold, multiple selectable sets are obtained. The main purpose of this step is to use scientific methods to screen out the adjustable parameters that have a significant impact on the important properties of copper billets, thereby removing those adjustable parameters with low influence and little contribution to the quality of the final product.

[0126] Preferably, when setting the impact threshold, experts in the fields of metallurgy and materials science are invited to participate in the discussion. Based on their experience and expertise, a minimum impact threshold is set for each key performance indicator. These thresholds will serve as screening criteria to distinguish which adjustable parameters have a significant impact on performance and which have a negligible impact. For the remaining high-impact adjustable parameters, multiple different configuration combinations are generated to form multiple optional sets.

[0127] S2 inputs real-time fixed parameters and the required performance values ​​into a pre-built prediction model to obtain the predicted values ​​of adjustable parameters;

[0128] Methods for constructing predictive models include:

[0129] S21 collects historical production data, including fixed parameters, adjustable parameters, and corresponding key performance data of copper billets; the collected data is cleaned to remove outliers and noise, and then standardized.

[0130] S22 selects machine learning models as the basic architecture of its predictive models. These machine learning models include support vector machines, gradient boosting trees, long short-term memory networks, random forests, and neural networks.

[0131] S23 divides the processed dataset into training, validation, and test sets; uses the training set to train the prediction model and continuously adjusts the model parameters; uses the validation set to evaluate the model's stability and prediction performance; uses the test set to evaluate the model's accuracy and reliability; and deploys the trained model into production.

[0132] S3 produces copper billets based on real-time fixed parameters and adjustable parameter predictions.

[0133] In this embodiment, the method for producing copper billets is the SCR method. The copper billets are continuously produced by SCR according to the predicted values ​​of real-time fixed parameters and adjustable parameters. This not only ensures the high efficiency and stability of the production process, but also provides a key opportunity to verify the output results of the prediction model. By comparing the performance of the copper billets produced in actual production with the expected results of the prediction model, the accuracy and reliability of the model can be evaluated, thereby optimizing the model and production process, and providing data support for continuous technological improvement.

[0134] The S4 test detects important properties of copper billets, obtains performance values, and compares them with required performance values ​​to identify exceptionally important properties.

[0135] Specifically, the performance values ​​refer to the actual test results of key performance indicators (such as conductivity, oxygen content, tensile strength, elongation, hardness, and microstructure) of copper billets continuously cast using the SCR method; the required performance values ​​for each important performance indicator are clearly defined, and these required performance values ​​are usually set based on industry standards, customer requirements, or the company's internal quality control specifications.

[0136] The actual performance values ​​are compared one by one with the required performance values ​​to determine whether each performance is within the acceptable range. If the actual value of a certain performance indicator exceeds the preset tolerance range, it is marked as an abnormally important performance. For example, if the conductivity is lower than expected or the oxygen content is higher than the maximum allowable value, these are abnormal situations.

[0137] S5 merges the optional sets corresponding to exceptionally important performance into a union; based on the frequency, impact and deviation of the adjustable parameters in the union, the order of adjustable parameter calls is corrected to obtain the priority call set;

[0138] Specifically, the methods of correction, such as Figure 2 As shown, it includes:

[0139] S51 counts the number of times each adjustable parameter appears in the set;

[0140] Specifically, the optional sets corresponding to all the abnormally important performances are merged into a union, and the frequency of each adjustable parameter in the union is counted to record the number of times each parameter appears.

[0141] By counting the occurrences, we can intuitively see which adjustable parameters repeatedly appear under multiple abnormal conditions, indicating that these adjustable parameters may be the core factors causing the problem. This helps technicians quickly identify key parameters that may affect multiple performance indicators and ensure that adjusting these parameters can improve performance in multiple aspects at the same time.

[0142] S52 obtains the influence degree of each adjustable parameter in the set on the abnormally important performance. The influence degree mentioned here is the influence degree obtained in step S1.

[0143] S53 calculates the deviation of abnormally important performance;

[0144] Methods for calculating deviation include:

[0145] S531 obtains the required performance values ​​and the performance values ​​of the copper billet;

[0146] S532 uses the following formula to calculate the deviation:

[0147] R = |Q1-Q2| / Q2;

[0148] Where R is the deviation, Q1 is the performance value of the copper billet, and Q2 is the required performance value;

[0149] S54 assigns weights to the frequency of occurrence, influence, and deviation, respectively.

[0150] When assigning weights to frequency of occurrence, impact, and deviation, the following factors should be considered:

[0151] 1) Assess the impact of different deviations on product quality and production stability, determine which performance anomalies are most urgent, and appropriately increase the weight of deviation for performance problems with large deviations;

[0152] 2) By combining data from multiple performance anomalies, identify the key parameters that frequently appear in multiple anomalies. For parameters with high prevalence, appropriately increase the weight of their frequency of occurrence.

[0153] 3) Adjust the weights of frequency of occurrence, impact, and deviation based on the importance that enterprises attach to different performance indicators;

[0154] 4) If some performance values ​​are better than the optimal value of the required performance value, and if the union of the alternative set and the union of the corresponding performance values ​​occurs, the corresponding performance values ​​may have some room for decrease. In this case, the influence of the adjustable parameters in the union can be appropriately reduced.

[0155] S55 uses the following formula to calculate the overall priority of each adjustable parameter:

[0156] S = C × W1 + D × W2 + E × W3;

[0157] Where S is the overall priority of each adjustable parameter, C is the number of times the adjustable parameter appears in the set, W1 is the weight of the number of occurrences, D is the influence, W2 is the weight of the influence, E is the deviation, W3 is the weight of the deviation, and W1+W2+W3=1.

[0158] S6 uses preset external conditions to serialize and adjust the priority call set to obtain the priority adjustment sequence;

[0159] Specifically, the external conditions include cost and response time; where cost includes energy consumption, material usage, and equipment wear and tear; and response time is the time required to reach a steady state after adjusting the adjustable parameters.

[0160] Methods for serialization adjustment include:

[0161] S61 evaluates the effect of each adjustable parameter in the priority call set on the applied condition;

[0162] When evaluating the effect value, the following factors should be considered: Adjusting certain adjustable parameters may bring unexpected risks or side effects. For example, while increasing the cooling rate can improve conductivity, it may lead to equipment overload or increased energy consumption, so its effect value should be appropriately reduced. By evaluating the effect value, these potential problems can be identified before adjustment, and corresponding preventive measures can be taken. At the same time, the adjustment of different adjustable parameters may involve different resource inputs (such as time, manpower, and funds). By evaluating the effect value, the rational allocation of resources can be ensured, and unnecessary waste and over-investment can be avoided. When certain adjustable parameters bring more resource input, their effect value should be appropriately reduced.

[0163] S62 uses the following formula to calculate the calling priority of each adjustable parameter in the priority calling set:

[0164] S1 = S × W4;

[0165] Where S1 is the call priority of each adjustable parameter, S is the overall priority, and W4 is the effect value of the external condition.

[0166] The overall priority reflects the potential contribution of the adjustable parameter to important performance, while the additional conditional value reflects the risks and additional costs that may be brought about by adjusting the parameter. Multiplying the two can achieve a balance between risk and benefit, ensuring that parameters that can significantly improve performance without bringing excessive risks or costs are selected for adjustment.

[0167] S63 sorts the call priority of each adjustable parameter in descending order to obtain the priority adjustment sequence;

[0168] The call priority of all adjustable parameters is compared numerically and arranged in descending order. The higher the value of the adjustable parameter, the greater the product of its overall priority and the effect of the external conditions, indicating that the adjustable parameter is more important to the overall optimization and should be called first.

[0169] S7 calls at least one adjustable parameter in the priority adjustment sequence for adjustment; wherein, at least one adjustable parameter includes the first adjustable parameter in the priority adjustment sequence;

[0170] Specifically, it is assumed that the oxygen content of the copper billet is higher than the required performance value;

[0171] Based on the foregoing analysis, the priority adjustment sequence is: casting temperature, cooling water flow rate on the crystallizing wheel side, and billet pulling speed;

[0172] First, select the first parameter, "casting temperature," and adjust it accordingly.

[0173] Based on historical data and expert advice, it has been decided to reduce the casting temperature from the current 1140℃ to 1138℃. This adjustment aims to reduce the oxygen solubility in the molten copper, thereby reducing the oxygen content in the finished product.

[0174] The heating power of the smelting furnace was adjusted according to the new settings to ensure that the casting temperature accurately reached the predetermined new value. During the adjustment process, the changes in the casting temperature were closely monitored, and the specific time and magnitude of the adjustment were recorded.

[0175] After the initial adjustments are completed, a new batch of copper billets is immediately produced and subjected to rapid testing, with a focus on changes in oxygen content. If the oxygen content is found to have decreased but still not fully met the expected standard, further fine-tuning of the casting temperature or adjustment of the cooling water flow rate on the crystallizing wheel side can be considered.

[0176] Collect more batches of data to verify whether the adjustment effect is consistent and provide a reliable reference for future decisions. Feed the results of this adjustment back to the predictive model to help it learn the new parameter combination and the corresponding product performance change pattern.

[0177] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A casting forming process for low oxygen copper, characterized in that, The method comprises the following steps: S1: setting important properties of copper cast billets, analyzing the influence degree of adjustable parameters on important properties, setting an influence degree threshold, and obtaining a plurality of optional sets; S2: inputting real-time fixed parameters and required property values into a pre-constructed prediction model to obtain adjustable parameter prediction values; S3: preparing copper cast billets according to real-time fixed parameters and adjustable parameter prediction values; S4: detecting important properties of copper cast billets to obtain property values, comparing the property values with required property values, and obtaining abnormal important properties; S5: merging optional sets corresponding to abnormal important properties to form a union set, correcting the calling order of adjustable parameters according to the number of occurrences of adjustable parameters in the union set, the influence degree, and the deviation degree of abnormal important properties, and obtaining a priority calling set; S6: using a preset external condition to sequence the priority calling set to obtain a priority adjustment sequence; S7: calling at least one adjustable parameter in the priority adjustment sequence for adjustment.

2. The cast forming process for low-oxygen copper of claim 1, wherein, The important properties of the copper cast billets in S1 include electrical conductivity, oxygen content, tensile strength, elongation, hardness, and microstructure. The adjustable parameters include crystallization wheel side cooling water flow and temperature, steel strip side cooling water flow and temperature, casting temperature, casting speed, and casting angle.

3. The cast forming process for low-oxygen copper of claim 1, wherein, The method for obtaining the influence degree in S1 comprises the following steps: S11: obtaining setting values of each adjustable parameter of different production batches of copper cast billets and corresponding important property results of the copper cast billets to obtain historical production data; S12: setting a quantitative method to evaluate the independent influence degree of each adjustable parameter on the important properties of copper cast billets according to the historical production data to obtain parameter independent influence degrees; S13: predicting the influence of the adjustable parameters on the important properties of the copper cast billets based on metallurgical principles and material science theories to obtain parameter influence prediction results; S14: selecting a quantitative method to analyze the joint influence degree of each adjustable parameter on the important properties of the copper cast billets according to the historical production data and the parameter influence prediction results to obtain parameter joint influence degrees; S15: performing weight distribution on the parameter independent influence degrees and the parameter joint influence degrees to calculate the influence degree of each adjustable parameter on the important properties of the copper cast billets.

4. The cast forming process for low-oxygen copper of claim 1, wherein, The method for constructing the prediction model in S2 comprises the following steps: S21: collecting historical production data, including fixed parameters, adjustable parameters, and corresponding copper cast billet important property data; and performing cleaning and standardization processing on the collected data; S22: randomly sampling the processed historical production data to generate a plurality of sub-data sets; S23: selecting a plurality of differentiated base learners, training the base learners on each sub-data set, combining the prediction results of the plurality of base learners, and obtaining a prediction model; S24: integrating the constructed prediction model into a production environment to predict the adjustable parameters.

5. The cast forming process for low-oxygen copper of claim 1, wherein, The method for correction in S5 comprises the following steps: S51: counting the number of occurrences of each adjustable parameter in the set; S52: obtaining the influence degree of each adjustable parameter in the set on abnormal important properties; S53: calculating the deviation degree of abnormal important properties; S54: respectively assigning weights to the number of occurrences, the influence degree, and the deviation degree. S55 calculates the comprehensive priority of each adjustable parameter to obtain a priority calling set.

6. The cast forming process for low-oxygen copper of claim 5, wherein, The method of S6 includes: S61 evaluates the effect value of each adjustable parameter in the priority calling set on the external condition; S62 calculates the calling priority of each adjustable parameter in the priority calling set; S63 sorts the calling priority of each adjustable parameter in descending order to obtain a priority adjustment sequence.

7. The cast forming process for low-oxygen copper of claim 1, wherein, The external condition in S6 includes cost and response time. The cost includes energy consumption, material usage, and equipment wear and tear; the response time is the time required to reach a stable state after adjusting the adjustable parameter.

8. The cast forming process for low-oxygen copper of claim 3, wherein, The at least one adjustable parameter includes the most advanced adjustable parameter in the priority adjustment sequence.

9. The cast forming process for low-oxygen copper of claim 1, wherein, The fixed parameters in S2 include the diameter of the crystallization wheel, the thickness of the crystallization wheel groove, the thickness of the carbon layer of the crystallization wheel groove, and the thickness of the carbon layer of the steel belt.

10. The cast forming process for low-oxygen copper of claim 5, wherein, The calculation method of the deviation in S53 includes: R=∣Q1-Q2∣ / Q2; Wherein, R is the deviation, Q1 is the performance value of the copper casting blank, and Q2 is the required performance value.

Images