Intelligent regulation and control method for combustion system of aluminum smelting furnace based on deep learning

Abstract

The application discloses an intelligent regulation and control method for an aluminum smelting furnace combustion system based on deep learning, and relates to the technical field of intelligent control of aluminum metallurgy industry. The method comprises the following steps: acquiring an oxidation loss rate, an in-furnace atmosphere component and a temperature distribution parameter of each smelting stage as an aluminum smelting parameter set, and classifying and labeling the aluminum smelting parameter set based on a smelting stage to obtain a plurality of smelting stage sets; using the plurality of smelting stage sets to respectively perform combustion parameter optimization to obtain optimized combustion parameters of a plurality of stages; using the optimized combustion parameters to perform intelligent regulation and control of the aluminum smelting furnace combustion system, and correcting the optimized combustion parameters of the next stage based on the smelting stage set of the last smelting stage to obtain corrected combustion parameters and obtain interval combustion parameters between stages, and continuously performing intelligent regulation and control of the aluminum smelting furnace combustion system. The application effectively reduces the oxidation loss rate in the aluminum smelting process.

Description

Technical Field

[0001] This invention relates to the field of intelligent control technology in the aluminum metallurgical industry, specifically to an intelligent control method for the combustion system of an aluminum smelting furnace based on deep learning. Background Technology

[0002] As the aluminum casting industry develops towards intelligence, low carbon emissions, and high efficiency, aluminum smelting, as a pre-process in aluminum casting, faces the challenge of metal loss and energy waste due to aluminum molten metal oxidation. This has become a key area for industry optimization. The control effect of the smelting furnace combustion system directly affects key factors affecting molten aluminum temperature, oxygen levels within the furnace, and the contact state between the molten aluminum and air. Existing technologies typically rely on manual experience or simple control strategies based on fixed models to manage the combustion process in the smelting furnace.

[0003] However, existing technologies lack the ability to deeply perceive and dynamically compensate for aluminum melt burn-off during the smelting process, resulting in persistently high metal oxidation losses in actual production. Specifically, aluminum melt oxidizes upon contact with air at high temperatures, and the degree of burn-off is directly affected by the uniformity of temperature distribution within the furnace, the oxygen content of the atmosphere, and the exposure time of the aluminum melt. Existing extensive combustion control methods are difficult to coordinate precisely, failing to provide optimal combustion parameters tailored to the characteristics of each smelting stage, and lacking adaptive and smooth transitions between stages. This exacerbates unnecessary oxidation burn-off and affects the stability and energy efficiency of the entire smelting process. Summary of the Invention

[0004] This invention provides an intelligent control method for the combustion system of an aluminum smelting furnace based on deep learning, aiming to solve the technical problem of high oxidation loss rate in aluminum smelting in the prior art.

[0005] In view of the above problems, the present invention provides an intelligent control method for the combustion system of an aluminum smelting furnace based on deep learning, comprising: The oxidation loss rate, furnace atmosphere composition and temperature distribution parameters of each smelting stage are obtained as aluminum smelting parameter sets, and the aluminum smelting parameter sets are classified and labeled based on the smelting stage to obtain multiple smelting stage sets. Combustion parameters are optimized using multiple sets of smelting stages to obtain optimized combustion parameters for multiple stages; The optimized combustion parameters are used to intelligently control the combustion system of the aluminum smelting furnace. Based on the smelting stage set of the previous smelting stage, the optimized combustion parameters for the next stage are corrected, the corrected combustion parameters are obtained, and the stage interval combustion parameters are obtained, so as to continue the intelligent control of the combustion system of the aluminum smelting furnace.

[0006] One or more technical solutions provided in this invention have at least the following technical effects or advantages: This invention provides an intelligent control method for the combustion system of an aluminum smelting furnace based on deep learning. By collecting parameters such as oxidation loss rate, furnace atmosphere composition, and temperature distribution, and classifying and labeling them according to smelting stages to form a smelting stage set, it provides staged and precise parameter support for combustion parameter optimization, avoiding optimization deviations caused by disordered parameters. Combustion parameter optimization is carried out separately for each smelting stage set, achieving targeted adaptation of combustion parameters for each smelting stage, ensuring a high degree of matching between combustion parameters and the smelting process and loss control requirements of different stages. By optimizing combustion parameters and implementing control, and using the smelting data of the previous stage to correct the parameters of the next stage and designing stage interval combustion parameters, the entire process of dynamic intelligent control of the aluminum smelting furnace combustion system is realized. This ensures the continuity of control at each smelting stage and stage interval, effectively avoids sudden changes in the furnace environment, and ultimately achieves precise control of the oxidation loss rate. Simultaneously, it improves the stability and adaptability of the aluminum smelting furnace combustion system control, and optimizes smelting conditions and energy utilization efficiency. Attached Figure Description

[0007] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0008] Figure 1 A flowchart illustrating the intelligent control method for the combustion system of an aluminum smelting furnace based on deep learning, provided in an embodiment of the present invention. Figure 2 This is a flowchart illustrating the process of obtaining optimized combustion parameters at multiple stages in the intelligent control method for the combustion system of an aluminum smelting furnace based on deep learning, as provided in an embodiment of the present invention. Detailed Implementation

[0009] This invention provides a deep learning-based intelligent control method for the combustion system of an aluminum smelting furnace, which is used to address the technical problem of high oxidation loss rate in aluminum smelting in the prior art.

[0010] Examples, such as Figure 1 As shown, this invention provides an intelligent control method for the combustion system of an aluminum smelting furnace based on deep learning, the method comprising: S100: Obtain the oxidation loss rate, furnace atmosphere composition and temperature distribution parameters of each smelting stage as an aluminum smelting parameter set, and classify and label the aluminum smelting parameter set based on the smelting stage to obtain multiple smelting stage sets.

[0011] In this embodiment of the invention, the oxidation loss rate, furnace atmosphere composition, and temperature distribution parameters of each smelting stage are acquired as an aluminum smelting parameter set. Based on the smelting stage, the aluminum smelting parameter set is classified and labeled to obtain multiple smelting stage sets. Traditional aluminum smelting process control relies on preset fixed parameters, which cannot sense and respond to real-time changes in operating conditions, and is one of the fundamental reasons for high oxidation loss rates and low energy efficiency. To achieve intelligent control, it is first necessary to construct a data foundation that can comprehensively and dynamically depict the smelting state. Therefore, this step aims to construct a parameter set reflecting key process indicators through multi-source information fusion, and to use data-driven methods to accurately and dynamically identify and label the actual smelting stages, providing precise input for subsequent stage-based optimization.

[0012] Step S100 in the method provided in this embodiment of the invention includes: Specifically, the oxidation loss rate, furnace atmosphere composition, and temperature distribution parameters for each smelting stage are obtained as a set of aluminum smelting parameters, including: Obtain the total mass of raw materials fed and the mass of molten aluminum produced during the smelting process, calculate the ratio of the deviation between the total mass of raw materials fed and the mass of molten aluminum produced to the total mass of raw materials, and obtain the oxidation loss rate; The composition of the furnace atmosphere is obtained, wherein the composition of the furnace atmosphere includes at least the oxygen concentration and the carbon monoxide concentration. Temperature data from multiple points inside the furnace are collected by multiple temperature measurement points to form temperature distribution parameters; The oxidation loss rate, furnace atmosphere composition, and temperature distribution parameters of each smelting stage are integrated into a set of aluminum smelting parameters.

[0013] First, obtain the total mass of raw materials fed and the mass of molten aluminum produced during the smelting process. Calculate the ratio of the deviation between the total mass of raw materials fed and the mass of molten aluminum produced to the total mass of raw materials to obtain the oxidation loss rate. The oxidation loss rate refers to the percentage of metal loss in the molten aluminum during aluminum smelting due to high-temperature oxidation and contact with air, relative to the total mass of raw materials fed. It is a quantitative indicator of the degree of metal loss in the molten aluminum. The total mass of raw materials fed refers to the total weight of the aluminum raw materials input, excluding impurities. The mass of molten aluminum produced refers to the weight of qualified molten aluminum obtained after this stage. Obtain the total mass of raw materials fed and the mass of molten aluminum produced for a certain smelting stage; the mass deviation of aluminum oxidation loss in this stage = total mass of raw materials fed - mass of molten aluminum produced; the oxidation loss rate in this stage = (mass deviation / total mass of raw materials) × 100%.

[0014] For example, taking the melting stage of an aluminum smelting furnace as an example: the total mass of aluminum ingots fed in is 5000 kg, that is, the total mass of raw materials fed in is 5000 kg. After this stage of smelting, 4850 kg of qualified molten aluminum is produced, that is, the mass of molten aluminum produced is 4850 kg; the mass deviation value = 5000 - 4850 = 150 kg; the oxidation loss rate = (150 / 5000) × 100% = 3%, that is, the oxidation loss rate of this melting stage is 3%.

[0015] Secondly, the furnace atmosphere composition is obtained, which includes at least oxygen concentration and carbon monoxide concentration. The furnace atmosphere composition refers to the gaseous components and corresponding concentrations that participate in the combustion reaction and affect the oxidation of molten aluminum in the aluminum smelting furnace, including oxygen concentration and carbon monoxide concentration. Oxygen concentration characterizes the degree of oxidation in the furnace, while carbon monoxide concentration reflects the degree of combustion completeness; these are key indicators for determining the furnace oxidation environment and combustion conditions. At least two atmosphere detection points are pre-set in the aluminum smelting furnace chamber, located in the middle of the furnace chamber and above the molten aluminum surface, respectively, to avoid local data deviations. Oxygen and carbon monoxide concentration data for a specific smelting stage are collected in real time using an atmosphere detection device such as an infrared gas analyzer. The collected gas concentration values ​​are recorded to form the furnace atmosphere composition data for that stage.

[0016] For example, following the above example of the melting stage: one infrared gas analyzer is deployed in the middle of the furnace chamber and above the surface of the molten aluminum to collect gas data during the melting stage. The results show that the oxygen concentration is 3.2% and the carbon monoxide concentration is 1.5%, which means that the atmosphere composition in the furnace during this stage is O2=3.2% and CO=1.5%.

[0017] Furthermore, temperature data from multiple points within the furnace are collected through various temperature measurement points to form temperature distribution parameters. These parameters, representing a set of temperature data collected from multiple points within the furnace, reflect the temperature differences at different locations and the overall temperature field distribution, avoiding misjudgments of operating conditions caused by single-point temperature measurements. Temperature measurement points are set up in the aluminum smelting furnace according to a layered and uniform principle, typically with two points each in the upper, middle, and lower layers of the furnace, for a total of six points covering the entire furnace area. Real-time temperature data for a specific smelting stage is simultaneously collected using temperature-sensing elements such as thermocouples at each point. The temperature data from all measurement points are then compiled and summarized to form the temperature distribution parameters for that stage.

[0018] For example, two thermocouple temperature measurement points are set in each of the upper, middle and lower layers of the furnace chamber, for a total of six points. The temperature data collected at each point during the melting stage are as follows: upper layer 780℃, 790℃; middle layer 850℃, 860℃; lower layer 920℃, 930℃. By integrating these six sets of temperature data, the temperature distribution parameters for this stage are obtained: 780℃, 790℃, 850℃, 860℃, 920℃, 930℃.

[0019] Finally, the oxidation loss rate, furnace atmosphere composition, and temperature distribution parameters of each smelting stage are integrated to form an aluminum smelting parameter set. The aluminum smelting parameter set refers to a set of parameters formed by integrating the oxidation loss rate, furnace atmosphere composition, and temperature distribution parameters of a certain smelting stage according to a standardized format. It serves as the basic data carrier for subsequent stage determination and combustion parameter optimization. The three types of parameter data are verified to ensure data completeness and absence of anomalies; the three types of parameters for the same smelting stage are integrated in the order of oxidation loss rate, furnace atmosphere composition, and temperature distribution parameters; the integrated parameter set is named according to the smelting stage, clearly identifying the corresponding smelting stage.

[0020] For example, first verify the three types of parameter data: oxidation loss rate, furnace atmosphere composition, and temperature distribution parameters are all normal. Then, integrate the three types of parameters according to a standardized format and name it the "Melting Stage - Aluminum Smelting Parameter Set," specifically: oxidation loss rate 3%, furnace atmosphere composition O2=3.2%, CO=1.5%, temperature distribution 780℃, 790℃, 850℃, 860℃, 920℃, 930℃. Similarly, follow the above steps to obtain the aluminum smelting parameter sets for the preheating, refining, and holding stages, respectively.

[0021] Specifically, the aluminum smelting parameter set is classified and labeled based on the smelting stage to obtain multiple smelting stage sets, including: Based on the aluminum smelting scheme, obtain the preset smelting stages; Based on the furnace atmosphere composition and the temperature distribution parameters, the smelting stage is obtained; Based on the preset smelting stage and the timing deviation of the smelting stage, obtain the stage deviation parameter; The aluminum smelting parameter set is classified and labeled based on the smelting stage and the preset smelting stage to obtain multiple smelting stage sets, wherein the smelting stage includes the preheating stage, the melting stage, the refining stage and the heat holding stage.

[0022] First, based on the aluminum smelting scheme, the pre-set smelting stages are obtained. The pre-set smelting stages refer to the smelting process and stage division standards pre-defined according to the aluminum smelting process scheme, including four stages: preheating, melting, refining, and holding. The basic parameters such as the time interval and temperature range of each stage are clearly defined. The process requirements for this aluminum smelting are then reviewed; based on the process requirements, the division standards for the time interval and temperature range of each smelting stage are set; and a clear list of pre-set smelting stages is formed.

[0023] For example, the process requirements for this aluminum smelting are: the aluminum ingot material is 6061 aluminum alloy, and the smelting capacity is 5 tons / furnace. Based on the 6061 aluminum alloy smelting process requirements, the following preset smelting stages are obtained: Preheating stage: time 0-1 hour, furnace temperature 400-800℃, key characteristic is aluminum ingot heating up, no obvious melting; Melting stage: time 1-3 hours, furnace temperature 800-950℃, key characteristic is aluminum ingot melting, significant oxidation and burn-off; Refining stage: time 3-3.5 hours, furnace temperature 750-800℃, key characteristic is removal of impurities from molten aluminum, low burn-off; Holding stage: time 3.5-5 hours, furnace temperature 700-750℃, key characteristic is maintaining stable aluminum liquid temperature, awaiting casting.

[0024] Secondly, based on the furnace atmosphere composition and temperature distribution parameters, the smelting stage is obtained. The actual smelting stage refers to the current smelting stage dynamically evaluated based on real-time furnace operating parameters and preset smelting stage characteristics. It is an accurate reflection of the actual smelting progress, distinct from preset stages determined solely by time. The core characteristics of the furnace atmosphere composition and temperature distribution parameters at the current stage are extracted, such as temperature range and oxygen concentration range. The extracted real-time characteristics are compared and matched with the core characteristics of the preset smelting stage. Based on the matching results, the current actual smelting stage is determined.

[0025] For example, the core features of the current melting stage are extracted as follows: temperature distribution parameters 780-930℃ and furnace atmosphere O2=3.2%; this feature is compared with the core features of the preset melting stage, where the temperature distribution parameters are in the range of 800-950℃ of the preset melting stage, the oxidation reaction is relatively active, and the features of the melting stage are matched, so the current actual melting stage is determined to be the melting stage.

[0026] Furthermore, based on the preset smelting stage and the timing deviation of the smelting stage, a stage deviation parameter is obtained. The stage deviation parameter refers to the deviation between the judgment time of the actual smelting stage and the corresponding time interval of the preset smelting stage. It is used to measure the matching degree between the actual smelting progress and the preset process, providing a progress reference for subsequent parameter optimization. The judgment time of the current actual smelting stage is recorded, i.e., the total smelting time when the stage judgment is completed; the preset time interval corresponding to the actual stage is found by referring to the preset smelting stage list, including the start and end times; the stage deviation parameter = actual judgment time - start time of the preset stage; if the actual time is within the preset interval, the deviation is positive; otherwise, it is negative.

[0027] For example, the actual time determined to be in the melting stage is the total melting time of 1.2 hours. Compared to the preset melting stage, the preset start time for the melting stage is 1 hour and the preset end time is 3 hours. The stage deviation parameter = 1.2h - 1h = 0.2h, meaning the current melting stage progress is 0.2 hours behind the preset progress, and the deviation is positive, within the preset range. If the actual determination time is 0.8 hours, it should be in the preheating stage, but is determined to be in the melting stage, then the deviation parameter = 0.8h - 1h = -0.2h, indicating that the melting stage has started prematurely.

[0028] Finally, the aluminum smelting parameter set is classified and labeled based on the smelting stage and the preset smelting stage to obtain multiple smelting stage sets. The smelting stages include the preheating stage, melting stage, refining stage, and holding stage. A smelting stage set refers to a unique set of parameters formed by classifying and labeling the aluminum smelting parameter sets of the same actual smelting stage, combined with the preset smelting stage and stage deviation parameters. Each stage corresponds to one smelting stage set. All aluminum smelting parameter sets are initially collected according to the actual smelting stage; for each stage's collected aluminum smelting parameter set, the corresponding preset smelting stage name and stage deviation parameter are labeled; the labeled parameter sets of the same stage are then integrated to form a dedicated smelting stage set for each stage.

[0029] For example, the preheating stage smelting stage set integrates the preheating stage aluminum smelting parameter set: oxidation loss rate 0.5%, oxygen concentration 5.0%, carbon monoxide concentration 0.8%, temperature distribution 500℃, 520℃, 580℃, 600℃, 630℃, 650℃, marked with a preset preheating stage: 0-1h, stage deviation parameter 0.5h, actual judgment time 0.5h, forming a dedicated smelting stage set for the preheating stage. The melting stage smelting stage set integrates the melting stage aluminum smelting parameter set: oxidation loss rate 3%, oxygen concentration 3.2%, carbon monoxide concentration 1.5%, temperature distribution 780℃, 790℃, 850℃, 860℃, 920℃, 930℃, marked with a preset melting stage 1-3h, stage deviation parameter 0.2h, actual judgment time 1.2h, forming a dedicated smelting stage set for the melting stage. The refining stage and the holding stage smelting stages are similar.

[0030] In this embodiment of the invention, a standardized process was used to collect and integrate parameters such as oxidation loss rate, furnace atmosphere composition, and temperature distribution at each stage of aluminum smelting. This constructed a set of aluminum smelting parameters that comprehensively reflects the smelting conditions. Simultaneously, by combining preset smelting stages with actual furnace operating parameters, the actual smelting stage was accurately determined. After calculating the stage deviation parameters, the aluminum smelting parameter set was classified and labeled to form a set of smelting stages that precisely matches each smelting stage. This achieved a dynamic correlation between key smelting parameters and actual smelting stages, providing accurate data support that fits the actual operating conditions for subsequent staged combustion parameter optimization and dynamic intelligent control of the aluminum smelting furnace combustion system. It also laid a data foundation for effectively controlling aluminum molten oxidation loss and improving smelting energy efficiency.

[0031] S200: Combustion parameters are optimized using multiple sets of the smelting stages to obtain optimized combustion parameters for multiple stages.

[0032] In this embodiment of the invention, combustion parameters are optimized using multiple smelting stage sets to obtain optimized combustion parameters for each stage. The optimization of combustion parameters in an aluminum smelting furnace must closely align with the operating characteristics of each smelting stage. Existing technologies often involve uniform, indiscriminate adjustments to combustion parameters, failing to incorporate oxidation loss rate and stage deviation parameters for targeted optimization. This results in poor combustion performance and an inability to effectively balance aluminum melt burn-off control with smelting energy efficiency improvement. Therefore, based on the smelting stage set obtained in S100, combustion parameters such as air-fuel ratio, combustion air volume, and burner power distribution ratio need to be obtained. Combustion performance parameters are constructed by correlating oxidation loss rate with stage deviation parameters. A particle swarm optimization method is then used to achieve precise optimization of combustion parameters for each stage, thereby improving the adaptability and effectiveness of the combustion system control.

[0033] like Figure 2 As shown, step S200 in the method provided in this embodiment of the invention includes: Based on multiple smelting stage sets, multiple stage combustion parameters are obtained, wherein the stage combustion parameters include air-fuel ratio, combustion air volume and burner power distribution ratio; Based on the oxidation burn-off rate and stage deviation parameters, the combustion effect parameters are obtained; With the goal of maximizing the combustion effect parameters, the combustion parameters for each stage are iteratively optimized to obtain optimized combustion parameters for multiple stages.

[0034] First, based on multiple smelting stage sets, multiple stage combustion parameters are obtained, including the air-fuel ratio, combustion air volume, and burner power distribution ratio. Stage combustion parameters are parameters that directly affect the combustion conditions of the aluminum smelting furnace, including the air-fuel ratio, combustion air volume, and burner power distribution ratio, and are fundamental parameters for achieving precise control of the combustion system. Specifically, the air-fuel ratio refers to the ratio of air to fuel, the combustion air volume refers to the volumetric flow rate of combustion air supplied to the furnace, and the burner power distribution ratio refers to the percentage of power of each burner relative to the total power. A smelting stage set refers to the set of core parameters obtained in step S100, categorized and labeled according to preset smelting stages and actual smelting stages, including information such as oxidation loss rate, furnace atmosphere composition, and temperature distribution parameters for each stage. From the multiple smelting stage sets obtained in step S100, parameters such as the air-fuel ratio, combustion air volume, and burner power distribution ratio corresponding to each stage are extracted. The extracted parameters are then categorized and organized according to the chronological order of the smelting stages, forming a set of stage combustion parameters corresponding one-to-one with each smelting stage.

[0035] For example, from the four melting stages of S100—preheating, melting, refining, and heat preservation—the corresponding stage combustion parameters are extracted for each stage. In the preheating stage, the air-fuel ratio is 1:1 and the combustion air volume is 300 m³ / s. 3 / h, burner power distribution ratio is 0.5; air-fuel ratio during melting stage is 1.2:1, combustion air volume is 500m³ / h. 3 / h, burner power distribution ratio is 0.7; air-fuel ratio in the refining stage is 0.8:1, combustion air volume is 400m³ / h. 3 / h, burner power distribution ratio is 0.4; air-fuel ratio during the heat preservation stage is 1:1, combustion air volume is 350m³ / h. 3 / h, the burner power distribution ratio is 0.3, and the stage combustion parameters are arranged in order to form a set of four stages.

[0036] Secondly, combustion effect parameters are obtained based on the oxidation burn-off rate and stage deviation parameters.

[0037] Among them, combustion effect parameters are obtained based on the oxidation burn-off rate and stage deviation parameters, including: Calculate the sum of the oxidation burn-off rate and the stage deviation parameter to obtain the burn-off deviation parameter; Calculate 2 and subtract the burn-off deviation parameter to obtain the combustion effect parameter.

[0038] First, calculate the sum of the oxidation burn-off rate and the stage deviation parameter to obtain the burn-off deviation parameter. The burn-off deviation parameter is the sum of the oxidation burn-off rate and the stage deviation parameter, with a value ranging from 0 to 1. The larger the value, the higher the degree of aluminum melt burn-off and the greater the deviation between the actual melting progress and the preset process. Add the values ​​of the oxidation burn-off rate and the stage deviation parameter to obtain the burn-off deviation parameter. For example, if the oxidation burn-off rate of a certain melting stage is 0.3 and the stage deviation parameter is 0.2, the burn-off deviation parameter = 0.3 + 0.2 = 0.5.

[0039] Next, subtract the burn-off deviation parameter from 2 to obtain the combustion effect parameter. The combustion effect parameter is an indicator of the quality of combustion conditions, ranging from 0 to 1. A higher value indicates a better combustion effect, which is more conducive to reducing aluminum melt burn-off and improving smelting efficiency. Subtracting the burn-off deviation parameter from the value 2 yields the combustion effect parameter, which also ranges from 0 to 1; a higher value indicates a better combustion effect. For example, the combustion effect parameter = 2 - 0.5 = 1.5, meaning the combustion effect parameter for this melting stage is 1.5.

[0040] Furthermore, with the goal of maximizing the combustion effect parameters, the combustion parameters for each stage are iteratively optimized to obtain optimized combustion parameters for multiple stages.

[0041] Specifically, with the optimization objective of maximizing the combustion effect parameter, iterative optimization of the combustion parameters at each stage is performed to obtain optimized combustion parameters for multiple stages, including: Select one combustion stage according to the process sequence as the first combustion stage; Using the combustion parameters of the first combustion stage as initial particles, and generating multiple candidate combustion parameters based on iterative compensation, an initial particle swarm is formed. The initial particle swarm is iteratively optimized with the goal of maximizing the combustion effect parameters. Based on the historical best position of each particle, the global historical best position of the initial particle swarm, and the first optimization step size, iterative optimization is performed until the number of iterations is reached or there is a particle whose combustion effect parameter is greater than the combustion effect threshold. The particle with the largest combustion effect parameter is obtained as the optimized combustion parameter for the first combustion stage, and iterative optimization is continued for other combustion stages to obtain the optimized combustion parameters for multiple stages.

[0042] The iteration optimization rounds are obtained based on the stage deviation parameter.

[0043] First, select a combustion stage according to the process sequence as the first combustion stage. The first combustion stage refers to the first combustion stage to be optimized, selected according to the process sequence of preheating, melting, refining, and holding. It is the starting stage for iterative optimization of combustion parameters. The combustion process sequence of the aluminum smelting furnace is clearly defined as preheating, melting, refining, and holding; the first combustion stage in this sequence is selected as the first combustion stage. For example, following the process sequence of preheating, melting, refining, and holding, the preheating stage is selected as the first combustion stage.

[0044] Secondly, using the stage combustion parameters of the first combustion stage as initial particles, and generating multiple candidate stage combustion parameters based on iterative compensation, an initial particle swarm is formed. The initial particles refer to the stage combustion parameters of the first combustion stage, serving as the initial data carrier for particle swarm optimization. Candidate stage combustion parameters refer to combustion parameters generated based on the initial particles through iterative compensation, exhibiting slight differences from the initial particles. The initial particle swarm is a set of parameters composed of the initial particles and multiple candidate stage combustion parameters, forming the basis for iterative particle swarm optimization. The process involves using the stage combustion parameters of the first combustion stage as initial particles; generating multiple candidate stage combustion parameters with slight differences from the initial particles based on iterative compensation rules; and integrating the initial particles with all candidate stage combustion parameters to form the initial particle swarm.

[0045] For example, the combustion parameters for the first combustion stage are: air-fuel ratio 1:1, combustion air volume 300 m³ / h, and burner power distribution ratio 0.5, which are used as initial particles; based on the iterative compensation rule, three candidate stage combustion parameters are generated, namely [1:1.05, 300 m³ / h, 1:1.05 ... 3 / h, 0.52], [0.95:1, 290m 3 / h, 0.48], [1:1, 310m 3 / h、0.5], the above 4 parameters constitute the initial particle swarm.

[0046] Next, with the goal of maximizing the combustion effect parameters, the initial particle swarm is iteratively optimized. The number of iterations is determined based on the stage deviation parameters. Iterative optimization refers to the process of adjusting and filtering the parameters in the initial particle swarm round by round, with the goal of maximizing the combustion effect parameters, until a preset condition is met. The number of iterations refers to the total number of iterations, and its value is determined by the stage deviation parameters of the first combustion stage. The stage deviation parameters reflect the degree of deviation between the actual process and the preset path. A larger stage deviation parameter indicates a more complex current operating condition or a greater deviation from the ideal state; therefore, more iterations are needed to search for a better combustion parameter solution to ensure optimization quality. The optimization goal is clearly defined as maximizing the combustion effect parameters, and iterative optimization is carried out on the initial particle swarm based on this goal. The number of iterations is obtained based on the stage deviation parameters of the first combustion stage. After determining the specific number of iterations according to preset corresponding rules, the iteration is initiated.

[0047] For example, the optimization goal is set to maximize the combustion effect parameter, and iterative optimization is carried out on the initial particle swarm in the preheating stage; the stage deviation parameter of this stage is 0.2. According to the preset rule, a deviation of 0.2 corresponds to 4 iterations. The iteration optimization round is determined to be 4 rounds, and the optimization round by round begins.

[0048] The acquisition of the first optimization step size includes: Obtain the first oxidation burn-off rate of the first combustion stage, and obtain the average oxidation burn-off rate of the same type of aluminum smelting furnace; Calculate the ratio of the first oxidation burn-off rate to the average oxidation burn-off rate, multiply it by the basic optimization step size, and obtain the first optimization step size.

[0049] First, obtain the first oxidation loss rate of the first combustion stage and the average oxidation loss rate of the same type of aluminum smelting furnace. The first oxidation loss rate refers to the oxidation loss rate of the molten aluminum corresponding to the first combustion stage, with a value ranging from 0 to 1. The larger the value, the higher the degree of oxidation loss of the molten aluminum in that stage. The average oxidation loss rate of the same type of aluminum smelting furnace refers to the average oxidation loss rate of the same type of aluminum smelting furnace under the same specifications and process conditions in the same smelting stage. It is a reference indicator for measuring the smelting stability of this type of furnace. Extract the first oxidation loss rate of the first combustion stage: Extract the oxidation loss rate value corresponding to the first combustion stage from the smelting stage set obtained from S100; Retrieve the average oxidation loss rate of the same type of aluminum smelting furnace: Retrieve the average oxidation loss rate value of the corresponding smelting stage from the historical operating data of the same type of aluminum smelting furnace.

[0050] For example, the first combustion stage is the preheating stage, and the first oxidation loss rate extracted from the smelting stage in this stage is 0.005; the average oxidation loss rate of the corresponding preheating stage retrieved from the historical data of the same type of aluminum smelting furnace is 0.01.

[0051] Next, the ratio of the first oxidation burn-off rate to the average oxidation burn-off rate is calculated, and multiplied by the basic optimization step size to obtain the first optimization step size. The basic optimization step size refers to the preset benchmark value for the adjustment range of the iterative optimization parameters, which is a fixed coefficient controlling the adjustment range of parameters in each iteration. The first optimization step size refers to the adjustment range of the iterative optimization parameters set for the first combustion stage, and is the core basis for parameter adjustment during particle swarm iteration optimization. The first oxidation burn-off rate is divided by the average oxidation burn-off rate of the same type of aluminum smelting furnace to obtain the ratio between the two; the calculated ratio is multiplied by the preset basic optimization step size to obtain the first optimization step size for the first combustion stage. For example, the first oxidation burn-off rate / average oxidation burn-off rate = 0.005 / 0.01 = 0.5; if the preset basic optimization step size is 0.1, then the first optimization step size = 0.5 × 0.1 = 0.05.

[0052] Further, based on the historical optimal position of each particle, the global historical optimal position of the initial particle swarm, and the first optimization step size, iterative optimization is performed until the number of iterations is reached or a particle with a combustion effect parameter greater than the combustion effect threshold exists. The particle with the largest combustion effect parameter is then obtained as the optimized combustion parameter for the first combustion stage, and iterative optimization continues for other combustion stages to obtain the optimized combustion parameters for multiple stages. The historical optimal position of a particle refers to the parameter combination corresponding to the maximum combustion effect parameter for a single particle in each iteration. The global historical optimal position refers to the parameter combination corresponding to the maximum combustion effect parameter for all particles in the entire initial particle swarm in each iteration. The combustion effect threshold is a preset critical value for achieving the combustion effect target, used to determine whether the iteration can be terminated early.

[0053] Specifically, based on the historical best position of each particle, the global historical best position of the initial particle swarm, and the first optimization step size, the particle swarm parameters are adjusted round by round; this is continued until the preset number of iterations is reached, or the combustion effect parameter of a certain particle is greater than the combustion effect threshold; from the iterated particle swarm, the particle with the largest combustion effect parameter is selected as the optimized combustion parameter for the first combustion stage; according to the process sequence, the above process is repeated for the remaining combustion stages, iterative optimization is carried out one by one, and finally the optimized combustion parameters for multiple stages are obtained.

[0054] For example, following the preheating stage (first combustion stage): The combustion effect threshold is set to 1.9. Based on the individual particle's historical best position, the global historical best position, and a first optimization step size of 0.05, particle parameters are adjusted round by round. In the third iteration, a particle's combustion effect parameter reaches 1.93, which is greater than the combustion effect threshold of 1.9, thus meeting the termination condition and stopping the iteration. This particle is selected as the optimized combustion parameter for the preheating stage: air-fuel ratio 1.03:1, combustion air volume 305 m³ / h. 3 / h, power distribution ratio 0.51; according to the process sequence, the above process is repeated for the melting stage, refining stage, and heat preservation stage in sequence, and finally the optimized combustion parameters of the four stages are obtained: melting stage [1.18:1, 485m 3 / h, 0.69], refining stage [0.82:1, 395m 3 / h, 0.41], heat preservation stage [1.01:1, 345m 3 / h, 0.32].

[0055] In this embodiment of the invention, combustion parameters for each stage are obtained based on the smelting stage set, and combustion effect parameters are constructed by combining oxidation loss rate and stage deviation parameters. The particle swarm optimization method is used to achieve precise optimization of combustion parameters for each stage, thereby achieving precise adaptation of combustion parameters to the operating conditions of each smelting stage. This effectively improves the control efficiency of the combustion system, reduces the oxidation loss rate of molten aluminum, and provides precise parameter support for the intelligent control of the combustion system of the aluminum smelting furnace in the future.

[0056] S300: Using the optimized combustion parameters, intelligent control of the aluminum smelting furnace combustion system is performed, and the optimized combustion parameters for the next stage are corrected based on the smelting stage set of the previous smelting stage. The corrected combustion parameters are obtained, and the stage interval combustion parameters are obtained, and intelligent control of the aluminum smelting furnace combustion system continues.

[0057] In this embodiment of the invention, the optimized combustion parameters are used for intelligent control of the aluminum smelting furnace combustion system. Based on the smelting stage set of the previous smelting stage, the optimized combustion parameters for the next stage are corrected, and the corrected combustion parameters and stage interval combustion parameters are obtained, continuing the intelligent control of the aluminum smelting furnace combustion system. The different stages of aluminum smelting do not operate independently; the combustion effect of the previous stage directly affects the smelting conditions of the next stage. If the combustion effect of the previous stage is poor, it can easily lead to insufficient adaptability of the optimized combustion parameters for the current stage, further aggravating aluminum molten metal oxidation and burn-off, and reducing smelting energy efficiency. Existing technologies do not consider the inter-stage correlation and lack combustion parameters for smooth transitions between stages, easily causing sudden changes in furnace conditions. Therefore, it is necessary to implement basic control by optimizing combustion parameters, correcting the current stage parameters based on the combustion effect of the previous stage, and designing stage interval combustion parameters to achieve continuous and precise intelligent control throughout the entire process, ensuring stable operating conditions at each stage and stage interval, and continuously optimizing the combustion effect.

[0058] Step S300 in the method provided in this embodiment of the invention includes: Specifically, the optimized combustion parameters are used for intelligent control of the aluminum smelting furnace combustion system, and the optimized combustion parameters for the next stage are corrected based on the smelting stage set of the previous smelting stage to obtain the corrected combustion parameters, including: The optimized combustion parameters are used to intelligently control the combustion system of the aluminum smelting furnace and obtain the smelting stage set of the current smelting stage. Based on the set of smelting stages, the combustion effect parameters of the previous stage are obtained, and the stage influence parameters are obtained based on the combustion effect parameters of the previous stage. When there is no previous stage in the current stage, the stage influence parameters are obtained based on the set of smelting stages of the current smelting stage. The combustion effect parameters corresponding to the optimized combustion parameters of the current stage are corrected using the stage influence parameters to obtain the corrected combustion effect parameters; Based on the corrected combustion effect parameters, the optimized combustion parameters are iteratively corrected and optimized to obtain the corrected combustion parameters.

[0059] First, the optimized combustion parameters are used to intelligently control the combustion system of the aluminum smelting furnace, and the smelting stage set for the current smelting stage is obtained. Optimized combustion parameters refer to the specific combustion parameters for each smelting stage obtained through iterative optimization in S200, including air-fuel ratio, combustion air volume, and burner power distribution ratio. The current smelting stage refers to the smelting stage for which combustion control is being implemented. The smelting stage set for the current smelting stage refers to the set of parameters collected, classified, and labeled according to the S100 method during the current stage control process, containing information such as oxidation loss rate, furnace atmosphere composition, and temperature distribution parameters.

[0060] Specifically, the optimized combustion parameters of the current stage obtained by S200 are imported into the combustion control system of the aluminum smelting furnace. Based on these optimized combustion parameters, the control system automatically adjusts the air-fuel ratio, combustion air volume and burner power distribution to implement intelligent control of the combustion system. During the control process, the smelting stage set of the current smelting stage is obtained according to the acquisition and classification labeling process of S100.

[0061] For example, the optimized combustion parameters for the melting stage obtained by S200 are: air-fuel ratio 1.18:1, combustion air volume 485m³ / h. 3 / h, power distribution ratio 0.69; import this parameter into the combustion control system to automatically adjust the combustion conditions in the furnace; during the control process, collect the oxidation loss rate of 0.03, oxygen concentration 3.2%, carbon monoxide concentration 1.5%, and temperature distribution of 780℃, 790℃, 850℃, 860℃, 920℃, and 930℃ for this stage, and label them according to the S100 method to form a melting stage set. If the current stage is the first stage, i.e., the preheating stage, then import the optimized combustion parameters for the preheating stage [1.03:1, 305m 3 / h、0.51] Implement regulation and control to obtain the preheating stage and smelting stage set.

[0062] Secondly, based on the set of smelting stages, the combustion effect parameters of the previous stage are obtained, and the stage influence parameters are obtained based on the combustion effect parameters of the previous stage. When there is no previous stage in the current stage, the stage influence parameters are obtained based on the set of smelting stages for the current smelting stage. The combustion effect parameters of the previous stage refer to the combustion effect parameters of the previous smelting stage calculated by S200; the smaller the value, the worse the combustion effect of the previous stage. The stage influence parameter is a coefficient characterizing the degree of influence of the combustion effect of the previous stage on the current stage, with a value range of 0-1. The worse the combustion effect parameter of the previous stage, the smaller the value, and the larger the stage influence parameter.

[0063] If the current stage is the preheating stage, i.e., the first stage, then there is no previous stage; if it is the melting, refining, or heat preservation stage, then there is a previous stage, corresponding to the preheating, melting, and refining stages respectively; if there is a previous stage, the corresponding value is extracted from the combustion effect parameters of the previous stage calculated by S200; the calculation is based on the rule that the worse the combustion effect parameters of the previous stage, the greater the stage influence parameter, with the preset calculation logic: stage influence parameter = 1 - previous stage combustion effect parameter / 2, ensuring that the smaller the combustion effect parameter, the greater the influence parameter; if there is no previous stage, the stage influence parameter can be obtained based on the melting stage set of the current melting stage. For example, if the current melting stage set is similar to the ideal melting stage set of the current stage, indicating that the current melting situation is good, it can be taken as 0.1 to appropriately consider the influence of operating condition fluctuations.

[0064] For example, if the current stage is the preheating stage and there is no preceding stage, the stage influence parameter is set to 0.1. If the current stage is the melting stage and the preceding stage is the preheating stage, S200 calculates the combustion effect parameter for the preheating stage to be 1.8, and the stage influence parameter = 1 - 1.8 / 2 = 0.1; if the combustion effect in the preheating stage is poor, the combustion effect parameter is 1.2, and the stage influence parameter = 1 - 1.2 / 2 = 0.4. The worse the effect of the preceding stage, the larger the influence parameter.

[0065] Next, the combustion effect parameters corresponding to the optimized combustion parameters of the current stage are corrected using the stage influence parameters to obtain the corrected combustion effect parameters. The combustion effect parameters corresponding to the optimized combustion parameters of the current stage refer to the combustion effect parameters calculated in S200. The corrected combustion effect parameters refer to the values ​​after adjusting the original combustion effect parameters of the current stage downwards using the stage influence parameters. The worse the combustion effect of the previous stage, the larger the stage influence parameter, and the smaller the corrected value. It is assumed that the combustion effect of the current stage is worse, providing a basis for further optimization. From the calculation results of S200, the combustion effect parameters corresponding to the optimized combustion parameters of the current stage are extracted; the corrected combustion effect parameter = original combustion effect parameter × (1 - stage influence parameter), ensuring that the larger the stage influence parameter, the smaller the corrected value, and thus adjusting downwards; thus, the corrected combustion effect parameters of the current stage are obtained.

[0066] For example, the current stage is the melting stage: if the combustion effect in the previous stage is good, the original combustion effect parameter = 1.9, the stage influence parameter = 0.1, and the corrected combustion effect parameter = 1.9 × (1 - 0.1) = 1.71; if the combustion effect in the previous stage is poor, the original combustion effect parameter = 1.9, the stage influence parameter = 0.4, and the corrected combustion effect parameter = 1.9 × (1 - 0.4) = 1.14.

[0067] Furthermore, based on the corrected combustion effect parameters, the optimized combustion parameters are iteratively corrected and optimized to obtain corrected combustion parameters. Iterative correction and optimization refers to the process of readjusting and optimizing the optimized combustion parameters of the current stage based on the corrected combustion effect parameters. For the worse-case scenario of the corrected combustion effect parameters being too small due to poor combustion performance in the previous stage, the parameters are further optimized to improve combustion performance. Corrected combustion parameters refer to the combustion parameters obtained after iterative correction and optimization that are suitable for the current operating conditions and are more closely aligned with actual operating requirements than the original optimized combustion parameters. The optimization goal is to improve the corrected combustion effect parameters to a reasonable range. Since the corrected values ​​are too small, the goal is to adjust the parameters to make the actual combustion effect parameters approach or exceed the original values. Based on the corrected combustion effect parameters, referring to the iterative optimization logic of S200, the air-fuel ratio, combustion air volume, and power distribution ratio of the optimized combustion parameters of the current stage are slightly adjusted and iterated. The iteration continues until the corrected combustion effect parameters reach a preset reasonable range, and the combustion parameters at this point are obtained as the corrected combustion parameters.

[0068] For example, if the previous stage's effect is poor, the combustion effect parameter is adjusted to 1.14: The optimization goal is to increase the adjusted combustion effect parameter from 1.14 to above 1.8; referring to the S200 logic, the original optimized combustion parameters [1.18:1, 485m] are used. 3 Based on [ / h, 0.69], adjustments are made round by round, such as fine-tuning the air-fuel ratio to 1.15:1 and the combustion air volume to 470m³ / h. 3 / h, power distribution ratio to 0.67; after 3 iterations, the combustion effect parameters were improved to 1.82, at which point the corresponding parameters are [air-fuel ratio 1.15:1, combustion air volume 470m³ / ... 3 / h, power distribution ratio 0.67] are the corrected combustion parameters for the melting stage.

[0069] This includes acquiring stage interval combustion parameters and continuing intelligent control of the aluminum smelting furnace combustion system, including: Based on the optimization and correction of combustion parameters in the previous stage, the difference in combustion parameters is obtained; Based on the combustion parameter difference and the time parameter of the preset smelting stage, the interval time length is obtained, and combined with the combustion parameter difference, the stage interval combustion parameter is obtained, wherein the stage interval combustion parameter includes multiple transition combustion parameters; The combustion system of the aluminum smelting furnace is intelligently controlled using the aforementioned stage interval combustion parameters.

[0070] First, based on the optimized and corrected combustion parameters from the previous stage, the combustion parameter difference is obtained. The optimized combustion parameters from the previous stage refer to the optimized combustion parameters obtained through S200 in the previous smelting stage. The combustion parameter difference refers to the numerical difference between the optimized combustion parameters from the previous stage and the corrected combustion parameters from the current stage, corresponding to the differences in the air-fuel ratio, combustion air volume, and power distribution ratio, respectively. The optimized combustion parameters from the previous stage and the corrected combustion parameters from the current stage are extracted separately. For the air-fuel ratio, combustion air volume, and power distribution ratio in both parameter sets, the numerical difference is calculated one by one: Difference = Optimized combustion parameter from the previous stage - Corrected combustion parameter from the current stage. The differences of the three parameters are then integrated to form a combustion parameter difference set.

[0071] For example, the current stage is the melting stage, and the corrected combustion parameters are [1.15:1, 470m]. 3 / h, 0.67]; The optimized combustion parameters in the previous stage were [1.03:1, 305m 3 / h, 0.51]; Calculate the difference in combustion parameters: air-fuel ratio difference = 1.03 - 1.15 = -0.12, combustion air volume difference = 305 - 470 = -165m 3 / h, power distribution ratio difference = 0.51 - 0.67 = -0.16; integrating these values ​​yields a set of combustion parameter differences [-0.12, -165m]. 3 / h, -0.16].

[0072] Secondly, based on the combustion parameter difference and the preset melting stage time parameters, the interval time length is obtained. Combined with the combustion parameter difference, the stage interval combustion parameters are obtained, whereby the stage interval combustion parameters include multiple transitional combustion parameters. The preset melting stage time parameters refer to the preset time intervals for each melting stage in S100, such as preheating 0-1h, melting 1-3h, taking the time interval between the end time of the previous stage and the start time of the current stage as the baseline value. The interval time length refers to the duration of the stage interval, determined jointly by the combustion parameter difference and the time parameters; the larger the difference, the longer the interval. The stage interval combustion parameters refer to the transitional combustion parameters used within the stage interval, including multiple transitional combustion parameters, used to achieve a smooth switch between the combustion parameters of the previous stage and the current stage. The transitional combustion parameters refer to the combustion parameters adjusted sequentially within the stage interval, evenly distributed according to the interval time length, gradually transitioning from parameters close to those of the previous stage to the corrected combustion parameters of the current stage.

[0073] Specifically, from the preset smelting stage list, extract the time interval benchmark value corresponding to the end time of the previous stage and the start time of the current stage; the interval length = time benchmark value × (1 + mean absolute value of combustion parameter difference); divide the combustion parameter difference evenly according to the interval length, generate multiple transition combustion parameters one by one, gradually transitioning from the optimized combustion parameters of the previous stage to the corrected combustion parameters of the current stage; integrate the stage interval combustion parameters: integrate all transition combustion parameters to form the stage interval combustion parameters.

[0074] For example, in the preset melting stage, the preheating stage ends at 1 hour and the melting stage begins at 1 hour, with a time base value of 0.2 hours; the average absolute value of the combustion parameter difference is calculated as (|-0.12|+|-165|+|-0.16|) / 3 ≈ 55.09; the interval length is calculated as 0.2 hours × (1+55.09) ≈ 11.02 hours; the combustion parameter differences of -0.12 and -165 are then considered. 3 / h, -0.16 is evenly divided over 30 minutes to generate 3 transitional combustion parameters: Transition 1 (minutes 0-10): air-fuel ratio 1.07:1, combustion air volume 360m³ / h. 3 / h, power distribution ratio 0.56; Transition 2 (minutes 10-20): air-fuel ratio 1.11:1, combustion air volume 415m³ / h 3 / h, power distribution ratio 0.61; Transition 3 (minutes 20-30): air-fuel ratio 1.15:1, combustion air volume 470m³ / h 3 / h, power distribution ratio 0.67; the three transition combustion parameters are integrated to form the stage interval combustion parameters.

[0075] Finally, the staged interval combustion parameters are used to continue intelligent control of the aluminum smelting furnace combustion system. The staged interval combustion parameters are imported into the combustion control system; the control system switches the transition combustion parameters sequentially according to the time nodes within the stage interval to implement intelligent control; after the stage interval ends, it automatically switches to the current stage correction combustion parameters to continue intelligent control of the combustion system in the current stage.

[0076] For example, the above three transitional combustion parameters are imported into the control system. From 0 to 10 minutes, transitional parameter 1 is used; from 10 to 20 minutes, transitional parameter 2 is used; and from 20 to 30 minutes, transitional parameter 3 is used. After 30 minutes, the system automatically switches to the melting stage to correct the combustion parameters [1.15:1, 470m]. 3 [ / h、0.67], continue to implement intelligent control of the combustion system during the melting stage to ensure a smooth transition of temperature and atmosphere in the furnace and avoid sudden changes in operating conditions.

[0077] In this embodiment of the invention, basic intelligent control is implemented by optimizing combustion parameters. Combined with dynamic calculation of stage-influence parameters based on the combustion effect of the previous stage, targeted corrections are made to the optimized combustion parameters for the current stage. This achieves precise adaptation of the influence of operating conditions between stages, effectively avoiding the negative impact of poor combustion effects in the previous stage on the current stage. By subtracting smaller corrections to combustion effect parameters to preset worse results, further optimization of parameters is promoted. Simultaneously, stage-interval combustion parameters are designed based on parameter differences and time parameters, achieving a smooth transition of combustion parameters between stages and avoiding sudden changes in furnace operating conditions. This forms a closed loop of basic control, parameter correction, interval transition, and continuous control, ensuring the consistency and accuracy of the entire process control of the aluminum smelting furnace combustion system, and providing full-process technical support for reducing aluminum molten oxidation loss and improving smelting energy efficiency.

[0078] Through the specific implementation methods described above, the embodiments of the present invention achieve the following technical effects: This invention provides an intelligent control method for the combustion system of an aluminum smelting furnace based on deep learning. By constructing a standardized set of aluminum smelting parameters covering all smelting stages and a set of smelting stages precisely matching actual operating conditions, reliable data support is provided for subsequent combustion parameter optimization. Iterative optimization of combustion parameters is carried out based on the smelting stage set, accurately matching the operating conditions of each stage and improving the adaptability of combustion parameters to the smelting process. Basic intelligent control is implemented based on optimized combustion parameters. The impact parameters of the current stage are dynamically calculated by combining the combustion effect of the previous stage, and the parameters of the current stage are specifically corrected. Stage interval combustion parameters are designed to achieve a smooth transition, effectively avoiding the negative impact of poor combustion effects in the previous stage on the current stage, while ensuring the stability of operating conditions in each stage and stage interval. Ultimately, a complete technical closed loop of data construction, parameter optimization, and intelligent control is formed. This not only achieves precise control of the entire process of the aluminum smelting furnace combustion system but also effectively reduces the aluminum molten metal oxidation loss rate and improves energy utilization efficiency, providing technical support for intelligent and efficient production in the aluminum smelting industry.

[0079] It should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for intelligent control of combustion systems in aluminum smelting furnaces based on deep learning, characterized in that, include: The oxidation loss rate, furnace atmosphere composition and temperature distribution parameters of each smelting stage are obtained as aluminum smelting parameter sets, and the aluminum smelting parameter sets are classified and labeled based on the smelting stage to obtain multiple smelting stage sets. Combustion parameters are optimized using multiple sets of smelting stages to obtain optimized combustion parameters for multiple stages; The optimized combustion parameters are used to intelligently control the combustion system of the aluminum smelting furnace. Based on the smelting stage set of the previous smelting stage, the optimized combustion parameters for the next stage are corrected, the corrected combustion parameters are obtained, and the stage interval combustion parameters are obtained, so as to continue the intelligent control of the combustion system of the aluminum smelting furnace.

2. The intelligent control method for the combustion system of an aluminum smelting furnace based on deep learning according to claim 1, characterized in that, Obtain the oxidation loss rate, furnace atmosphere composition, and temperature distribution parameters for each smelting stage to form an aluminum smelting parameter set, including: Obtain the total mass of raw materials fed and the mass of molten aluminum produced during the smelting process, calculate the ratio of the deviation between the total mass of raw materials fed and the mass of molten aluminum produced to the total mass of raw materials, and obtain the oxidation loss rate; The composition of the furnace atmosphere is obtained, wherein the composition of the furnace atmosphere includes at least the oxygen concentration and the carbon monoxide concentration. Temperature data from multiple points inside the furnace are collected by multiple temperature measurement points to form temperature distribution parameters; The oxidation loss rate, furnace atmosphere composition, and temperature distribution parameters of each smelting stage are integrated into a set of aluminum smelting parameters.

3. The intelligent control method for the combustion system of an aluminum smelting furnace based on deep learning according to claim 1, characterized in that, The aluminum smelting parameter set is classified and labeled based on the smelting stage to obtain multiple smelting stage sets, including: Based on the aluminum smelting scheme, obtain the preset smelting stages; Based on the furnace atmosphere composition and the temperature distribution parameters, the smelting stage is obtained; Based on the preset smelting stage and the timing deviation of the smelting stage, obtain the stage deviation parameter; The aluminum smelting parameter set is classified and labeled based on the smelting stage and the preset smelting stage to obtain multiple smelting stage sets, wherein the smelting stage includes the preheating stage, the melting stage, the refining stage and the heat holding stage.

4. The intelligent control method for the combustion system of an aluminum smelting furnace based on deep learning according to claim 1, characterized in that, Combustion parameters are optimized using multiple sets of smelting stages to obtain optimized combustion parameters for multiple stages, including: Based on multiple smelting stage sets, multiple stage combustion parameters are obtained, wherein the stage combustion parameters include air-fuel ratio, combustion air volume and burner power distribution ratio; Based on the oxidation burn-off rate and stage deviation parameters, the combustion effect parameters are obtained; With the goal of maximizing the combustion effect parameters, the combustion parameters for each stage are iteratively optimized to obtain optimized combustion parameters for multiple stages.

5. The intelligent control method for the combustion system of an aluminum smelting furnace based on deep learning according to claim 4, characterized in that, Based on the oxidation burn-off rate and stage deviation parameters, combustion effect parameters are obtained, including: Calculate the sum of the oxidation burn-off rate and the stage deviation parameter to obtain the burn-off deviation parameter; Calculate 2 and subtract the burn-off deviation parameter to obtain the combustion effect parameter.

6. The intelligent control method for the combustion system of an aluminum smelting furnace based on deep learning according to claim 4, characterized in that, With the optimization objective of maximizing the combustion effect parameter, iterative optimization of the combustion parameters at each stage is performed to obtain optimized combustion parameters for multiple stages, including: Select one combustion stage according to the process sequence as the first combustion stage; Using the combustion parameters of the first combustion stage as initial particles, and generating multiple candidate combustion parameters based on iterative compensation, an initial particle swarm is formed. The initial particle swarm is iteratively optimized with the goal of maximizing the combustion effect parameters. Based on the historical best position of each particle, the global historical best position of the initial particle swarm, and the first optimization step size, iterative optimization is performed until the number of iterations is reached or there is a particle whose combustion effect parameter is greater than the combustion effect threshold. The particle with the largest combustion effect parameter is obtained as the optimized combustion parameter for the first combustion stage, and iterative optimization is continued for other combustion stages to obtain the optimized combustion parameters for multiple stages.

7. The intelligent control method for the combustion system of an aluminum smelting furnace based on deep learning according to claim 6, characterized in that, Obtaining the first optimization step size includes: Obtain the first oxidation burn-off rate of the first combustion stage, and obtain the average oxidation burn-off rate of the same type of aluminum smelting furnace; Calculate the ratio of the first oxidation burn-off rate to the average oxidation burn-off rate, multiply it by the basic optimization step size, and obtain the first optimization step size.

8. The intelligent control method for the combustion system of an aluminum smelting furnace based on deep learning according to claim 6, characterized in that, The iteration optimization rounds are obtained based on the stage deviation parameter.

9. The intelligent control method for the combustion system of an aluminum smelting furnace based on deep learning according to claim 1, characterized in that, The optimized combustion parameters are used for intelligent control of the aluminum smelting furnace combustion system, and the optimized combustion parameters for the next stage are corrected based on the smelting stage set of the previous smelting stage, obtaining the corrected combustion parameters, including: The optimized combustion parameters are used to intelligently control the combustion system of the aluminum smelting furnace and obtain the smelting stage set of the current smelting stage. Based on the set of smelting stages, the combustion effect parameters of the previous stage are obtained, and the stage influence parameters are obtained based on the combustion effect parameters of the previous stage. When there is no previous stage in the current stage, the stage influence parameters are obtained based on the set of smelting stages of the current smelting stage. The combustion effect parameters corresponding to the optimized combustion parameters of the current stage are corrected using the stage influence parameters to obtain the corrected combustion effect parameters; Based on the corrected combustion effect parameters, the optimized combustion parameters are iteratively corrected and optimized to obtain the corrected combustion parameters.

10. The intelligent control method for the combustion system of an aluminum smelting furnace based on deep learning according to claim 1, characterized in that, Obtain the combustion parameters for each stage interval and continue intelligent control of the aluminum smelting furnace combustion system, including: Based on the optimization and correction of combustion parameters in the previous stage, the difference in combustion parameters is obtained; Based on the combustion parameter difference and the time parameter of the preset smelting stage, the interval time length is obtained, and combined with the combustion parameter difference, the stage interval combustion parameter is obtained, wherein the stage interval combustion parameter includes multiple transition combustion parameters; The combustion system of the aluminum smelting furnace is intelligently controlled using the aforementioned stage interval combustion parameters.

Images