Melting furnace control system for metal magnesium ingot refining production and melting furnace

Abstract

The invention belongs to the technical field of metal smelting, and particularly relates to a melting furnace control system for metal magnesium ingot refining production and a melting furnace. The system comprises a high-temperature imaging temperature measurement probe, a laser-induced breakdown spectroscopy sensor, an array thermocouple, a real-time acquisition module, an analysis and prediction module, a decision analysis module, an instruction generation module and an execution driving module, and collects concentration signals and three-dimensional temperature field data of key impurity elements in magnesium melt in real time through cooperation of the high-temperature imaging temperature measurement probe, the laser-induced breakdown spectroscopy sensor and the array thermocouple; constructing a fusion data set with consistent time and space; based on a mass transfer kinetic equation, calculating a diffusion flux vector field and a segregation coefficient space distribution field of impurity elements in real time, constructing an impurity potential enrichment region map, identifying a high-risk region and judging a dominant mechanism of impurity enrichment; through a cooperative instruction of electromagnetic stirrer partition excitation parameters and water jacket cooling water amount and in combination with a space matching and disturbance compensation mechanism, efficient coupling regulation and control of a stirring and cooling system in a target area are ensured.

Description

Technical Field

[0001] This invention belongs to the field of metal smelting technology, specifically a melting furnace control system and melting furnace for refining magnesium ingots. Background Technology

[0002] In the refining process of magnesium ingots, the melting furnace is one of the key pieces of equipment. Its control precision directly affects the removal effect of impurity elements in the magnesium melt and the purity of the final product. In the existing technology, there are some improvement schemes for the control system and method of the melting furnace. For example, the uniformity of the melt can be improved by adjusting the heating power or stirring intensity, or the temperature distribution can be controlled by the cooling system to suppress impurity segregation.

[0003] However, existing technologies still have the following limitations: 1. Existing systems mostly rely on offline sampling or fixed-point temperature measurement, which cannot obtain the three-dimensional distribution of impurity concentration in the melt and its dynamic evolution trend in real time, making it difficult to detect potential impurity enrichment areas in a timely manner.

[0004] 2. Traditional control methods mostly involve independently adjusting stirring or temperature, which fails to fully consider the diffusion, convection, and segregation coupling effects of impurities in the melt, resulting in limited control effects, especially in areas with high impurity concentrations where control blind spots are likely to occur.

[0005] 3. Existing technologies make it difficult to spatially match impurity enrichment areas with specific control methods, thus failing to achieve precise zoned control, resulting in energy waste and low control efficiency. Summary of the Invention

[0006] To overcome the shortcomings of the prior art, the present invention provides a melting furnace control system and melting furnace for refining magnesium ingots, which can effectively solve the problems involved in the prior art.

[0007] The objective of this invention can be achieved through the following technical solutions: In a first aspect, this invention provides a melting furnace control system for the refining of magnesium ingots, comprising: a real-time acquisition module, an analysis and prediction module, a decision analysis module, an instruction generation module, and an execution drive module.

[0008] The real-time acquisition module is connected to the analysis and prediction module, the analysis and prediction module is connected to the decision analysis module, the decision analysis module is connected to the instruction generation module, and the instruction generation module is connected to the execution drive module.

[0009] The real-time acquisition module collects concentration signals of key impurity elements in the magnesium melt and three-dimensional temperature field data of the melt.

[0010] The analysis and prediction module, based on the concentration signal and temperature data, calculates the diffusion flux vector field and segregation coefficient spatial distribution field of key impurity elements in real time through the mass transfer kinetic equation, and generates a map of potential impurity enrichment regions.

[0011] The decision analysis module identifies high-risk areas where the concentration of key impurity elements exceeds a preset permissible threshold based on the impurity potential enrichment region map, and determines the target stirring area and the direction of target temperature gradient adjustment.

[0012] The instruction generation module generates coordinated control instructions, including adjusting the excitation parameters of the electromagnetic stirrer zone and regulating the cooling water volume of the melt water jacket, based on the target stirring area and the target gradient adjustment direction.

[0013] The execution drive module executes the coordinated control command to drive the electromagnetic stirrer and the cooling water flow regulating valve to achieve coupled control of the melt flow field and temperature field.

[0014] Secondly, the present invention provides a melting furnace for refining magnesium ingots, comprising:

[0015] A furnace body for containing molten magnesium, and a heating device coupled to the furnace body for heating and maintaining the temperature of the materials inside the furnace.

[0016] Multiple independently controllable electromagnetic stirrers are disposed at the bottom and / or side of the furnace body for driving the flow of melt when energized.

[0017] The furnace body is embedded between the side wall and the furnace lining. It has cooling water channels inside and is used to cool the melt in sections by adjusting the cooling water flow rate to control the temperature gradient of the melt. The multiple water jacket cooling units are arranged in sections along the circumference of the molten pool. Each water jacket cooling unit has a cooling water channel inside and is connected to an independent flow regulating valve.

[0018] Compared with the prior art, the embodiments of the present invention have at least the following advantages or beneficial effects: (1) The present invention, through the collaborative work of laser-induced breakdown spectroscopy sensor and array thermocouple, combined with high-temperature imaging temperature measurement probe, can collect the concentration signal of key impurity elements and three-dimensional temperature field data in magnesium melt in real time, construct a spatiotemporally consistent fusion dataset, and provide an input basis for subsequent mass transfer analysis and control decision.

[0019] (2) Based on the mass transfer kinetic equation, the system can calculate the diffusion flux vector field and the spatial distribution field of the segregation coefficient of impurity elements in real time, construct the potential enrichment region map of impurities, identify high-risk regions, and determine the dominant mechanism of impurity enrichment by analyzing the proportion of convection and segregation contribution in high-risk regions. Based on this, the target stirring region and the direction of temperature gradient adjustment are determined, thereby realizing the location and targeted intervention of the root cause of the problem.

[0020] (3) The present invention ensures efficient coupling and control of the stirring and cooling system in the target area by coordinating the excitation parameters of the electromagnetic stirrer and the cooling water volume of the water jacket, combined with spatial matching and disturbance compensation mechanisms, thereby avoiding control conflicts and energy waste. Attached Figure Description

[0021] The present invention will be further described with reference to the accompanying drawings, but the embodiments in the drawings do not constitute any limitation on the present invention. For those skilled in the art, other drawings can be obtained based on the following drawings without creative effort.

[0022] Figure 1 This is a schematic diagram of the module connection of the present invention.

[0023] Figure 2 This is a flowchart illustrating the acquisition of key impurity element concentration signals and melt three-dimensional temperature field data in this invention.

[0024] Figure 3 This is a flowchart illustrating the logical judgment process to determine whether the present invention falls within a high-risk area. Detailed Implementation

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

[0026] Reference Figure 1 As shown, in a first aspect, the present invention provides a melting furnace control system for refining magnesium ingots, comprising: a real-time acquisition module, an analysis and prediction module, a decision analysis module, an instruction generation module, and an execution drive module.

[0027] The real-time acquisition module is connected to the analysis and prediction module, the analysis and prediction module is connected to the decision analysis module, the decision analysis module is connected to the instruction generation module, and the instruction generation module is connected to the execution drive module.

[0028] Reference Figure 2 As shown, the real-time acquisition module collects the concentration signals of key impurity elements in the magnesium melt and the three-dimensional temperature field data of the melt.

[0029] Considering that temperature changes during impurity diffusion, segregation, and transport are key factors affecting the accurate spatial location of potential impurity enrichment areas.

[0030] Therefore, by using a laser-induced breakdown spectral sensor installed on the side wall of the melting furnace, the laser beam is sequentially focused onto the surface of magnesium melt at different spatial points in the molten pool according to a preset sampling frequency and a predetermined spatial coordinate sequence. The sensor's built-in spectral acquisition system captures the characteristic spectrum and converts it into an electrical signal containing wavelength and intensity information.

[0031] The preset sampling frequency setting satisfies the Nyquist sampling theorem, that is, it is at least twice the highest frequency component of the melt state parameter change.

[0032] The determination of the predetermined spatial coordinate sequence includes: based on the three-dimensional model of the molten pool, generating a regular grid lattice using a spatial grid division method, refining the grid in key areas (such as solidification interfaces and areas where impurities are easily enriched), and sparsening the grid in the middle of the melt where the flow is sufficient and the temperature is uniform, thus forming a non-uniform spatial coordinate sequence.

[0033] After denoising and background subtraction of the characteristic spectra, the characteristic spectral lines of key impurity elements are identified and converted into concentration values ​​through intensity-concentration calibration curves.

[0034] The key impurity elements include, but are not limited to: iron, nickel, copper, and silicon.

[0035] Based on an array of thermocouples arranged at preset coordinate points in the three-dimensional space of the molten pool, temperature data from multiple discrete points are acquired, along with temperature probes at the observation window on the furnace top, to simultaneously collect temperature data from different depths of the melt.

[0036] The temperature probe installed in the observation window on the top of the furnace uses a CCD high-temperature imaging thermometer to perform a two-dimensional scan of the melt surface according to a preset scanning trajectory, and obtain the apparent temperature data of each point on the melt surface.

[0037] The fixed-point temperature data of the array thermocouple and the scanning temperature data of the temperature probe are subjected to spatiotemporal unification processing. For the fixed-point temperature data, a discrete-point temperature dataset is established based on its preset coordinates. For the scanning temperature data, each temperature measurement point on the scanning trajectory is assigned three-dimensional spatial coordinates based on the probe's spatial position and orientation recorded at the scanning time, forming a high-density point cloud temperature dataset. Through a time synchronization mechanism, the two types of datasets at the same sampling time are merged to ensure that the temperature data participating in the fusion reflects the thermal state of the melt at the same instant.

[0038] Based on the actual geometric dimensions of the molten pool, a structured interpolation calculation grid is established in a three-dimensional spatial coordinate system. The grid density is adaptively set according to the severity of the temperature field gradient change.

[0039] The three-dimensional temperature field data of the melt is constructed using the Kriging interpolation method. The specific steps are as follows:

[0040] Two types of datasets were used as input samples, and a Gaussian model was used to fit the theoretical variogram function to obtain the range, sill value and nugget constant describing the spatial correlation of the temperature field.

[0041] For each grid node to be interpolated, the Kriging equations are solved based on the variogram and the spatial distribution of known sample points to determine the optimal weighting coefficients for each known temperature point participating in the interpolation of that node.

[0042] The temperature value of each known temperature point is summed with its corresponding weighting coefficient to obtain the temperature estimate of that grid node.

[0043] Interpolation calculations are performed sequentially on all nodes in the three-dimensional mesh to obtain continuously distributed temperature values ​​throughout the entire molten pool space, thus constructing three-dimensional spatial temperature field data of the melt.

[0044] By unifying spatial coordinates and synchronizing timestamps, concentration data and temperature data are registered to form a spatiotemporally consistent fused dataset.

[0045] It should be noted that the preset coordinate points refer to the embedded position coordinates of each temperature measuring end of the array thermocouple in the three-dimensional space of the molten pool. They are generated by dividing the basic grid and densifying it in key areas (such as the solidification interface and near the cooling water jacket), while eliminating points that interfere with the furnace structure.

[0046] The preset scanning trajectory is specifically as follows: a spiral path along the geometric contour of the molten pool surface is used as the basic scanning trajectory. In areas with large temperature gradients, thermocouple fixed point projection areas, and key subsurface detection areas, the spacing between adjacent scanning lines is reduced, and the number of reciprocating scans is increased in key areas. The planned scanning path is discretized into a series of continuous spatial coordinate points to form the scanning trajectory.

[0047] The analysis and prediction module, based on concentration signals and temperature data, calculates the diffusion flux vector field and segregation coefficient spatial distribution field of key impurity elements in real time through mass transfer kinetic equations, generating a map of potential impurity enrichment regions.

[0048] The calculation process of the diffusion flux vector field is as follows: based on the three-dimensional temperature field data of the melt, the diffusion coefficient of the key impurity element at each spatial point of the current temperature field is calculated according to the Arrhenius equation, forming a spatial distribution field of diffusion coefficient.

[0049] The specific expression for the Arrhenius equation is as follows:

[0050] ;

[0051] in, The diffusion coefficient represents the rate of diffusion or reaction of a substance under specific conditions. The pre-exponential factor represents the limiting diffusion coefficient or reaction rate at infinitely high temperatures. Activation energy is the energy required to overcome for a substance to diffuse or react. Represents the molar gas constant. Representing absolute temperature, it is a key thermodynamic parameter affecting diffusion or reaction rates. Represented by natural constant An exponential function with base 0.

[0052] In this formula, The exponential part, which constitutes the exponential term, reflects temperature. Significant effect on diffusion or reaction rates: The higher the temperature, the smaller the absolute value of this exponential term. The closer the value is to 1, the higher the diffusion coefficient becomes. Increase; By combining the temperature-independent intrinsic properties with the temperature-dependent exponential factor as a preconditioning factor, the law of diffusion or reaction rate change with temperature is fully described.

[0053] The overall formula reveals that the diffusion or reaction rate increases exponentially with increasing temperature.

[0054] It should be noted that at different temperatures To conduct the diffusion experiment, a series of temperature points were set as absolute temperatures before the experiment. A constant temperature bath is used to maintain a stable temperature, and thermocouples or platinum resistance thermometers are used to monitor and record the temperature in real time. Use accepted values ​​directly. and You can find this information in materials science or chemical kinetics handbooks.

[0055] Based on the concentration signal after spatiotemporal registration, a regular interpolation grid matching the temperature field grid is established in the three-dimensional space of the molten pool using the Kriging interpolation method. The concentration values ​​of each node of the grid are estimated to obtain a three-dimensional concentration distribution field that covers the entire molten pool and is continuously distributed.

[0056] Based on the constructed three-dimensional concentration distribution field, spatial differentiation is performed on the same spatial grid. For internal grid nodes, the partial derivatives of the concentration along the three-dimensional coordinate axes are calculated using the central difference scheme. For grid nodes at the melt pool boundary, the forward or backward difference scheme is used for calculation. The partial derivatives in each direction are combined to form the concentration gradient vector at that node. The above differentiation operation is performed on all grid nodes in sequence to obtain the concentration gradient vector field.

[0057] Using Fick's first law, based on the fact that the diffusion flux vector is proportional to the concentration gradient vector but in opposite directions, the diffusion flux vector of each node is calculated to obtain the diffusion flux vector field covering the entire three-dimensional space of the molten pool.

[0058] The calculation process of the spatial distribution field of the segregation coefficient is as follows: Based on the constructed three-dimensional spatial temperature field data of the melt, the spatial position of the solidification interface at the current moment is determined. The solidification interface is characterized by the liquidus temperature isotherm, that is, the set of spatial points in the melt whose temperature is equal to the liquidus temperature corresponding to the magnesium alloy grade; through the isosurface extraction algorithm, the liquidus temperature isotherm is extracted from the three-dimensional temperature field data and identified as the isotherm of the solid-liquid interface.

[0059] Centered on the identified solid-liquid interface isotherm, extend a certain distance (e.g., 5-20 mm) along the interface normal to the liquid phase side. Define this area as the solid-liquid interface proximity region. Sample the temperature distribution in the proximity region to obtain the temperature gradient information and local temperature value on the liquid phase side near the interface.

[0060] The system retrieves pre-stored multi-element phase diagram data for magnesium alloys from a database containing equilibrium phase diagram thermodynamic data for different magnesium alloy systems. Specifically, the database stores the values ​​of solid-phase equilibrium concentration and liquid-phase equilibrium concentration of each key impurity element under different alloy compositions and temperature conditions.

[0061] For each spatial location in the vicinity of the solid-liquid interface, based on the local temperature value at the spatial location and the matrix composition of the magnesium melt (determined by the initial feed composition and changes in the refining process), a query is performed in the database to obtain the equilibrium concentrations of key impurity elements in the solid phase and the liquid phase under the given temperature and composition conditions; the equilibrium segregation coefficient is defined as the ratio of the equilibrium concentration in the solid phase to the equilibrium concentration in the liquid phase.

[0062] Based on the solid-liquid interface at the current moment, calculate the unit normal vector at each location of the interface. This vector points to one side of the liquid phase and is defined as the normal direction of the solidification interface.

[0063] Extract the normal component of the diffusion flux vector field at the solidification interface. For each spatial location on the interface, read the diffusion flux vector value at that location from the constructed diffusion flux vector field, and perform a dot product operation with the normal vector to obtain the normal component of the diffusion flux.

[0064] By using an array of thermocouples, the temperature change curves at various points over time are monitored to identify the temperature inflection point caused by the release of latent heat of solidification. Combined with the spatial coordinates of the thermocouples, the interface propulsion rate is obtained by Kriging interpolation.

[0065] Based on the mass conservation relationship of the solute at the solidification interface, and combining the normal component of the diffusion flux and the interface propagation rate, the characteristic thickness of the solute diffusion boundary layer near the solidification interface is calculated. The specific calculation process is as follows:

[0066] The mass of solute discharged to the liquid phase side per unit time due to interface advancement should be equal to the mass of solute leaving the boundary layer and entering the distant melt through diffusion per unit time. The normal component of the diffusion flux is the mass of solute leaving the boundary layer per unit area per unit time. The interface advancement rate determines the mass of solute discharged to the interface per unit area per unit time due to solidification. The two have a direct proportional relationship in terms of numerical values.

[0067] The normal component of the diffusion flux is considered as the driving force for solute diffusion, and the interfacial propulsion rate is considered as the source term for solute emission. The ratio of the two can be characterized as the characteristic diffusion scale required to maintain mass conservation, which is the characteristic thickness of the solute diffusion boundary layer.

[0068] For each location on the solidification interface, the normal component of the diffusion flux and the interface propagation rate at that location are extracted. Based on the above physical relationship, the characteristic thickness that matches the conservation conditions is obtained.

[0069] Using the equilibrium segregation coefficient, diffusion boundary layer characteristic thickness, and interface propagation rate obtained from the aforementioned calculations as input parameters, the solute conservation relationship at the solidification interface is described as follows: the amount of solute discharged to the liquid phase side per unit time due to interface propagation is equal to the amount of solute leaving the boundary layer and entering the distant melt per unit time through diffusion. Based on the solute conservation relationship, a solute mass balance equation is established at the interface.

[0070] Using the liquid phase equilibrium concentration obtained from the phase diagram database at the current interface position as the assumed value, the actual solid phase concentration corresponding to the assumed value is calculated based on the solute mass balance relationship.

[0071] The calculated ratio of the actual concentration on the solid side to the actual concentration on the liquid side is compared with the preset residual tolerance value. If the ratio is less than or equal to the preset residual tolerance value, the ratio of the actual concentration on the solid side to the actual concentration on the liquid side corresponding to the current assumed value is taken as the segregation coefficient at that interface location.

[0072] Otherwise, adjust the assumed value of the actual concentration on the liquid side and repeat the above iterative calculation until the requirement is met.

[0073] The preset residual tolerance value is set according to the process control accuracy requirements, for example, it can be set as follows: The implementer can adaptively adjust the numerical value according to specific requirements.

[0074] By solving the solute mass balance equation, the ratio of the actual concentration on the solid phase side to the actual concentration on the liquid phase side at the solidification interface is obtained. This actual concentration ratio is defined as the segregation coefficient at the current interface position. The above solution process is repeated for all discrete positions on the solidification interface to obtain the spatial distribution of the segregation coefficient along the entire solid-liquid interface.

[0075] Based on the constructed three-dimensional spatial temperature field data of the melt, the liquidus temperature isotherm at the current moment is extracted and determined as the spatial morphology of the solid-liquid interface.

[0076] For each discrete location on the solid-liquid interface, these segregation coefficients are used as known data points and associated with their three-dimensional coordinates on the interface to form an interface segregation coefficient dataset.

[0077] Based on the solid-liquid interface isotherm, the spatial extension of the segregation coefficient is carried out along the interface normal direction to both the liquid and solid sides. On the liquid side, the segregation coefficient gradually approaches 1 along the normal direction with increasing distance. On the solid side, the segregation coefficient maintains the same value as at the interface along the normal direction.

[0078] For each grid node in the three-dimensional space of the molten pool, a corresponding segregation coefficient value is assigned according to its relative position to the solid-liquid interface using a distance-weighted method: nodes located on the solid-liquid interface directly use the segregation coefficient value at the interface; nodes located in the solid phase region use the segregation coefficient value at their nearest projection point on the interface; nodes located in the liquid phase region have their segregation coefficient value calculated using a preset exponential decay function based on their distance to the interface and the segregation coefficient value in the normal direction at the interface.

[0079] After assigning the above values ​​to all three-dimensional mesh nodes, a spatial distribution field of segregation coefficients that covers the entire molten pool space and is continuously distributed is obtained.

[0080] It should be noted that the above calculation process is based on the assumption that the concentration distribution within the boundary layer is linear. This assumption has good approximate accuracy when the solidification rate is low and the boundary layer is thin. For high solidification rate cases, a more accurate exponential distribution can be used for correction.

[0081] It should also be noted that each node value in the spatial distribution field of the segregation coefficient characterizes the impurity distribution characteristics corresponding to that spatial location in the current solidification process: for the solid phase region, it reflects the degree of impurity segregation in the solidified part; for the liquid phase region, it reflects the impurity distribution tendency that will occur if solidification occurs at that location; and for the interface region, it describes the actual impurity distribution behavior at the interface where solidification is taking place.

[0082] By constructing a diffusion flux vector field and a segregation coefficient spatial distribution field, we can fully reveal the dynamic behavior mechanism of impurity elements in the three-dimensional space of the melt, providing a basis for identifying potential impurity enrichment regions.

[0083] The decision analysis module identifies high-risk areas where the concentration of key impurity elements exceeds a preset permissible threshold based on the impurity potential enrichment region map, and determines the target stirring area and the direction of target temperature gradient adjustment.

[0084] Considering that the mass transfer and segregation coupling of impurity elements in the melt have a distribution trend, determining the potential enrichment region of impurities provides a clear target location for subsequent regulation; the determination of the target stirring region and the direction of adjustment of the target temperature gradient is to further trace the dominant physical mechanism leading to impurity enrichment.

[0085] Reference Figure 3 As shown, the process of generating a map of potential enrichment regions of impurities and identifying high-risk regions where the concentration of key impurity elements exceeds a preset permissible threshold is as follows: Based on the constructed diffusion flux vector field, at each node of the spatial grid, the inflow and outflow difference of the diffusion flux vector in its neighborhood is calculated. This difference value is the contribution component of convective transport to the net accumulation rate at that node.

[0086] Based on the spatial distribution field of the segregation coefficient and combined with the solidification interface propagation rate, at the grid node near the solid-liquid interface, according to the relative positional relationship between the node and the interface, the segregation effect at the interface is mapped to the node through spatial interpolation, thereby obtaining the contribution component of the impurity concentration change rate caused by segregation at the node.

[0087] For each grid node in the three-dimensional space of the molten pool, its corresponding convective transport contribution component and interfacial segregation contribution component are superimposed to obtain the net accumulation rate value at that node. The superposition follows these principles: for internal melt nodes far from the interface, the net accumulation rate is directly equal to the convective transport contribution component; for nodes in the region adjacent to the solidification interface, the net accumulation rate is the algebraic sum of the convective transport contribution component and the interfacial segregation contribution component; for nodes on the solidified side, the net accumulation rate is set to zero, indicating that the concentration in the solid phase no longer changes.

[0088] After performing the above calculations sequentially on all grid nodes, the net accumulation rate distribution covering the entire molten pool space is obtained.

[0089] Using the current three-dimensional concentration distribution field of key impurity elements as the initial state and the net accumulation rate distribution field as the concentration evolution rate, the forward Eulerian method is used to extrapolate forward at a preset time step, updating the predicted impurity concentration values ​​of each spatial node every moment until the preset total integration time is reached.

[0090] The predicted impurity concentrations are visualized and mapped according to their concentration levels to generate a map of potential impurity enrichment regions.

[0091] The spatial grid nodes covered by the potential enrichment region map of impurities are traversed one by one, the predicted value of impurity concentration at each node is extracted, and the predicted value is compared with the system's preset allowable threshold.

[0092] For all spatial nodes whose predicted impurity concentration exceeds the preset permissible threshold, their three-dimensional spatial coordinates are recorded to form an initial set of exceedance points.

[0093] Centered on each out-of-target point, search for other out-of-target points within its three-dimensional neighborhood. If the Euclidean distance between two points is less than a preset connectivity distance threshold, then the two points are determined to belong to the same connected region. Through recursive search, all mutually connected or indirectly connected out-of-target points are merged into the same cluster, and the same cluster is defined as a high-risk region.

[0094] The process of determining the target stirring area and the target temperature gradient adjustment direction is as follows: perform area integration on the diffusion flux vector along the boundary of the high-risk area to obtain the net dissolved mass flowing into or out of the area per unit time, divide the net dissolved mass by the volume of the area to obtain the average concentration change in the area per unit time caused by convective transport, and define it as the convective concentration increment.

[0095] For the solid-liquid interface section contained in the high-risk area, the mass of solute repelled to the liquid phase side per unit time due to the forward movement of the solidification interface is calculated based on the segregation coefficient value and the interface propagation rate at that section. This mass of solute is divided by the characteristic volume of the liquid phase region adjacent to the interface to obtain the average concentration change in the region per unit time due to the interface segregation effect, which is defined as the accumulation concentration increment.

[0096] The characteristic volume of the interface adjacent to the liquid phase region is defined as the area of ​​the solid-liquid interface segment within the high-risk region multiplied by the characteristic thickness of the diffusion boundary layer.

[0097] If the increase in convection concentration is greater than the increase in accumulation concentration, it indicates that the current melt flow intensity is insufficient to transport impurities released from interface decomposition or locally accumulated to the depths of the melt far from the high-risk area in a timely manner. The main cause of impurity enrichment is insufficient solute convection transport. At this time, based on the dominant direction of impurity migration in the diffusion flux vector field and combined with the three-dimensional spatial coordinates of the high-risk area in the molten pool, the target area that needs to be subjected to or enhanced electromagnetic stirring is determined, so that the forced convection direction generated by stirring matches the impurity transport demand direction, and a first stirring strategy with the core of strengthening convection in this area is generated.

[0098] If the increase in convective concentration is less than the increase in accumulated concentration, it indicates that although there is a certain convective transport effect, the segregation effect at the solidification interface is too strong, causing the mass of solute repelled to the liquid phase side per unit time to exceed the dilution capacity of convection. The main cause of impurity enrichment is the segregation effect at the solidification interface. At this time, based on the gradient change trend of the segregation coefficient in the spatial distribution field of the segregation coefficient, combined with the three-dimensional spatial temperature field data of the melt, the direction of temperature gradient adjustment that can reduce the segregation coefficient at the interface or weaken the segregation intensity is calculated, and the first temperature control strategy with the core of adjusting the thermal state near the solidification interface is generated.

[0099] For most impurity elements, the segregation coefficient approaches 1 as the temperature increases. Therefore, the direction of adjusting the temperature gradient to reduce the segregation intensity is usually to increase the temperature on the liquid side near the interface, so that the actual segregation coefficient approaches 1. The specific adjustment amount can be calculated by combining the temperature sensitivity coefficient of the segregation coefficient with the target reduction of the segregation coefficient.

[0100] If the absolute value of the difference between the convection concentration increment and the accumulation concentration increment is less than or equal to the preset tolerance range, it is determined to be a compound cause, and at the same time, a coordinated instruction for the first stirring strategy and the first temperature control strategy is generated.

[0101] It should be noted that the preset time step is directly taken as an integer fraction of the excitation cycle of the electromagnetic stirrer. For example, when the excitation frequency of the electromagnetic stirrer is 2Hz (cycle 0.5 seconds), the preset time step can be taken as 0.1 seconds, that is, 5 integration calculations are performed in each flow cycle, which ensures the time resolution and avoids excessive calculation due to too small a step.

[0102] The preset tolerance range is based on the quality control requirements for impurity content in magnesium ingot products. The difference between the upper limit of the product impurity limit and the target control value is proportionally converted to the contribution of convection and segregation, and then compared. For example, if the upper limit of the product impurity limit is 100 ppm and the target control value is 80 ppm, then a safety margin of 20 ppm can be allocated to the tolerance range of the control decision. The safety margin is divided by the volume and unit time of the area with the largest proportion of high-risk areas to obtain the corresponding preset tolerance range.

[0103] The preset connectivity distance threshold is set to 2 to 3 times the spacing between adjacent grid nodes. This method ensures that adjacent or second-to-adjacent out-of-range points at the grid scale can be merged into the same region, avoiding fragmentation caused by grid discretization.

[0104] It should also be noted that by adopting a progressive decision-making process of identifying high-risk areas, tracing the main causes, and determining the direction of regulation, the problem of impurity enrichment is accurately diagnosed and targeted. This ensures that the subsequent first stirring strategy, first temperature control strategy, or collaborative command can directly address the root cause of the problem, thereby achieving the best impurity suppression effect with minimal regulation cost and significantly improving the product quality stability of the magnesium ingot refining process.

[0105] The instruction generation module generates coordinated control instructions, including adjusting the excitation parameters of the electromagnetic stirrer zone and regulating the cooling water volume of the melt water jacket, based on the target stirring area and the target gradient adjustment direction.

[0106] The above provides the target stirring area and the target gradient adjustment direction. The above information is converted into specific excitation parameters for each zone of the electromagnetic stirrer and cooling water flow rate settings for each valve of the water jacket cooling system through the instruction generation module.

[0107] The process of adjusting the excitation parameters of the electromagnetic stirrer in each zone is as follows: based on the physical installation position of each excitation coil of the electromagnetic stirrer and the spatial distribution characteristics of the electromagnetic force generated in the molten pool, the range of action of the molten space corresponding to each zone is determined.

[0108] When the first stirring strategy is received, the three-dimensional spatial coordinate range of the target stirring area is extracted. The coordinate range is then spatially superimposed with the spatial action range of each electromagnetic stirrer partition to determine the electromagnetic stirrer partition whose action range covers the target stirring area or partially overlaps with the target stirring area. This partition is then identified as the target excitation partition.

[0109] The diffusion flux vectors at each spatial location of the target stirring area are extracted from the constructed diffusion flux vector field. By vector synthesis of all diffusion flux vectors in the area, the dominant direction of impurity migration in the area is determined. Based on the physical characteristics of the change of electromagnetic force direction with excitation current phase in each zone of the electromagnetic stirrer, the phase value that minimizes the angle between the electromagnetic force direction generated in the zone and the dominant direction is determined.

[0110] Extract the magnitude of the diffusion flux vector within the region and calculate the statistical average as the magnitude of the diffusion flux value.

[0111] Based on the correlation between the desired convection intensity and the diffusion flux characteristic value, the target convection intensity to be achieved along the dominant direction within the target stirring area is determined. This correlation is obtained through experimental calibration, specifically: several typical operating conditions covering the actual production range are selected, including different melt temperatures, initial impurity concentrations, and stirring intensities. Under normal operating conditions, the concentration distribution of key impurity elements in the melt is measured, the diffusion flux characteristic value is calculated, and the convection velocity of the melt under these conditions is measured simultaneously.

[0112] The rate of decrease or the final uniformity of impurity concentration in the target area is used as an evaluation index for convective transport effect to determine the minimum convective intensity required to effectively suppress impurity enrichment.

[0113] The diffusion flux characteristic value is fitted with the corresponding minimum effective convection intensity to establish a fitting curve between the two.

[0114] The target convection intensity is obtained by substituting the real-time calculated diffusion flux characteristic value into the curve.

[0115] The correspondence between the electromagnetic force amplitude generated by each target excitation zone and the melt flow velocity, as well as the velocity superposition characteristics when multiple zones act together, are obtained. The correspondence is obtained in advance through fluid dynamics simulation.

[0116] The goal is to achieve the target convection intensity by reaching the synthetic flow velocity along the dominant direction within the target mixing region. Considering the spatial influence weight of each zone within the target mixing region and the direction of the flow velocity it generates, a set of equations is established regarding the excitation current amplitude of each zone. Solving the set of equations yields the target current amplitude that each zone should provide.

[0117] The difference between the target current amplitude and the current current amplitude of each zone is calculated, and the difference is used as the amplitude adjustment amount for each zone.

[0118] By using pressure sensors installed in the molten pool, instantaneous pressure data of the melt at a set position is continuously collected, forming a time-series signal that reflects the flow fluctuation characteristics of the melt.

[0119] The acquired time series signals are preprocessed to remove abnormal data points caused by sensor noise or external interference, and the signals are then subjected to mean removal processing.

[0120] The signal is converted from the time domain to the frequency domain by using the Fast Fourier Transform to obtain the amplitude spectrum of the signal.

[0121] Peak identification is performed on the amplitude spectrum, and the peak with the largest amplitude is searched. The frequency corresponding to the peak is determined as the dominant fluctuation frequency of the current melt flow.

[0122] With the goal of matching the forced convection generated by electromagnetic stirring with the flow characteristics of the melt itself, the target frequency of the excitation current is set to be equal to the dominant fluctuation frequency.

[0123] The difference between the current excitation current frequency and the dominant fluctuation frequency is calculated, and the difference is used as the frequency adjustment amount.

[0124] When multiple zones act together on the target mixing region, the excitation current amplitude and phase of each zone are used as decision variables, and the projection of the synthetic electromagnetic force in the target mixing region onto the dominant direction is maximized as the main objective. For each set of candidate amplitudes and phases, the synthetic electromagnetic force field generated in the target mixing region is calculated. Through iterative comparison, the amplitude and phase combination that is closest to the main objective is finally obtained.

[0125] It should be noted that the effective range is mainly determined by the decisive condition of the spatial attenuation characteristics of electromagnetic force in the melt. Specifically, based on the physical installation position and excitation parameters of each excitation coil of the electromagnetic stirrer, the distribution of the electromagnetic force generated in the melt pool space under rated excitation conditions of that zone is calculated.

[0126] Numerical calculations based on the Biot-Savart law revealed the law that the electromagnetic force intensity at the geometric center of the partition decreases with increasing spatial distance.

[0127] The critical condition for defining the range of action is that the electromagnetic force intensity decays to 5% of its maximum value. The spatial distance corresponding to this critical condition is defined as the characteristic radius of action of the zone. The position where the electromagnetic force intensity decays to the critical value is calculated along the three-dimensional coordinate direction of the molten pool, forming the spatial boundary of action of the zone in the molten pool.

[0128] Based on the radius of action of this feature and the installation center coordinates of each zone, the effective range of this zone is defined in the three-dimensional spatial coordinate system of the molten pool.

[0129] The process of adjusting the cooling water volume of the melt water jacket is as follows: The melt water jacket consists of multiple sets of independent and controllable cooling water units, arranged in sections along the height and circumference of the molten pool. Each set of water jackets corresponds to a specific spatial region within the molten pool. By adjusting the cooling water volume flowing through the water jacket, the local cooling intensity of the melt in that region can be independently controlled, thereby affecting the temperature gradient of that region. The control system pre-stores data on the spatial range of each water jacket unit and its influence on the temperature gradient.

[0130] Upon receiving the first temperature control strategy, the target temperature gradient adjustment direction determined by the strategy is extracted. For cases requiring an increased longitudinal temperature gradient, water jacket units located at the bottom of the molten pool or the lower part of the sidewall are matched to enhance the cooling intensity from bottom to top. For cases requiring a decreased lateral temperature gradient, water jacket units located at corresponding positions around the molten pool are matched to balance the horizontal cooling distribution. For cases requiring adjustment of the normal temperature gradient at the solidification interface, water jacket units in the adjacent solid-liquid interface projection region are matched. Through spatial superposition judgment, water jacket units that have a dominant influence on the target temperature gradient adjustment direction are selected and identified as the target water jackets.

[0131] According to heat transfer theory, there is a definite correspondence between the change in local temperature gradient and the cooling water volume of the water jacket. The form of the corresponding function is obtained through heat transfer simulation calculation. The gradient change is substituted into the corresponding function to calculate the amount of cooling water that needs to be increased or decreased for each target water jacket.

[0132] It should be noted that when multiple target water jackets jointly influence the target temperature gradient adjustment direction, the goal is to achieve the desired value in the composite gradient change along the target temperature gradient adjustment direction. A priority-based approach is used to allocate the cooling water flow adjustment amount to each target water jacket. The specific process is as follows: the degree of alignment between the spatial region corresponding to the water jacket in the molten pool and the region where the target temperature gradient adjustment direction is located. The higher the degree of alignment, the more direct the impact of the water jacket adjustment on the target gradient, and the higher its priority. For example, for a target requiring an increase in the longitudinal temperature gradient, water jackets located at the bottom of the molten pool or on the lower sidewall have a higher priority than upper water jackets.

[0133] The specific collaborative control command is as follows: when an excitation parameter adjustment command or a cooling water volume adjustment command is generated, the expected disturbance 1 caused by the excitation parameter adjustment command to the three-dimensional temperature field of the melt is analyzed. The expected disturbance 1 includes the local temperature rise and fall caused by the change in convective heat transfer due to the change in stirring intensity, and the temperature field redistribution caused by the change in flow mode.

[0134] The expected disturbance 2 caused by the cooling water flow rate adjustment command to the diffusion flux vector field is analyzed. The expected disturbance 2 includes the change in diffusion coefficient caused by the change in local temperature gradient, and the effect of the change in melt viscosity caused by temperature change on the flow field and impurity transport characteristics.

[0135] By obtaining the correspondence data between excitation parameters and temperature field changes, as well as the correspondence data between cooling water volume and diffusion flux vector field changes through on-site calibration, the proposed adjustments to excitation parameters and cooling water volume are input into the corresponding data to obtain the expected temperature field changes and diffusion flux vector field changes in the target mixing area and related areas. Based on the expected temperature field changes, the compensating water jacket unit that needs adjustment and its required water volume adjustment are determined. Based on the expected diffusion flux vector field changes, the compensating electromagnetic stirrer zone that needs adjustment and its required excitation parameter adjustments are determined, thereby counteracting expected disturbance one and expected disturbance two and maintaining process stability.

[0136] For the same electromagnetic stirrer zone, if both the original excitation adjustment command and the compensating excitation adjustment command are received simultaneously, the adjustment amounts of the two commands are superimposed to obtain the final excitation parameter adjustment value for that zone.

[0137] For the same water jacket unit, if both the original water volume adjustment command and the compensating water volume adjustment command are received simultaneously, the adjustment amounts of the two commands are added together to obtain the final water volume adjustment value for the water jacket unit. For different actuators, sub-commands are generated according to their respective final adjustment values, and the timing of each sub-command is coordinated.

[0138] Secondly, the present invention provides a melting furnace for refining magnesium ingots, comprising:

[0139] A furnace body for containing molten magnesium, and a heating device coupled to the furnace body for heating and maintaining the temperature of the materials inside the furnace.

[0140] Multiple independently controllable electromagnetic stirrers are disposed at the bottom and / or side of the furnace body for driving the flow of melt when energized.

[0141] The furnace body is embedded between the side wall and the furnace lining. It has cooling water channels inside and is used to cool the melt in sections by adjusting the cooling water flow rate to control the temperature gradient of the melt. The multiple water jacket cooling units are arranged in sections along the circumference of the molten pool. Each water jacket cooling unit has a cooling water channel inside and is connected to an independent flow regulating valve.

[0142] The above description is merely an example and illustration of the structure of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the structure of the invention or exceed the scope defined by the present invention, they should all fall within the protection scope of the present invention.

Claims

1. A control system for a melting furnace used in the refining of magnesium ingots, characterized in that, include: The real-time acquisition module collects concentration signals of key impurity elements in the magnesium melt and three-dimensional temperature field data of the melt. The analysis and prediction module, based on the concentration signal and temperature data, calculates the diffusion flux vector field and segregation coefficient spatial distribution field of key impurity elements in real time through the mass transfer kinetic equation, and generates a map of potential impurity enrichment regions. The decision analysis module identifies high-risk areas where the concentration of key impurity elements exceeds a preset permissible threshold based on the impurity potential enrichment region map, and determines the target stirring area and the direction of target temperature gradient adjustment. The instruction generation module generates coordinated control instructions, including adjusting the zone excitation parameters of the electromagnetic stirrer and regulating the cooling water volume of the melt water jacket, based on the target stirring area and the target gradient adjustment direction. The execution drive module executes the coordinated control command to drive the electromagnetic stirrer and the cooling water flow regulating valve to achieve coupled control of the melt flow field and temperature field.

2. The control system for a melting furnace used in the refining of magnesium ingots according to claim 1, characterized in that: The process for obtaining the concentration signals of the key impurity elements and the three-dimensional temperature field data of the melt is as follows: By using sensors installed on the side wall of the melting furnace, the magnesium melt is excited at multiple points and spectral signals are collected at a preset sampling frequency. The concentration signals of key impurity elements are obtained in real time through spectral analysis. Based on an array of thermocouples arranged at preset coordinate points in the three-dimensional space of the molten pool, and a temperature probe set in the observation window on the top of the furnace, temperature data at different depths of the melt are collected simultaneously. The temperature probe acquires temperature data of the melt surface and a certain depth range by scanning. By integrating the fixed-point temperature data of the array thermocouple and the scanning temperature data of the temperature probe, and processing them through a spatial interpolation algorithm, a three-dimensional spatial temperature field data of the melt is constructed. Spatial mapping and registration are performed between the fixed-point sampling coordinates of the sensor and the path coordinates of the temperature scanning probe to align the concentration signal and temperature data in time and space.

3. The control system for a melting furnace used in the refining of magnesium ingots according to claim 2, characterized in that: The calculation process of the diffusion flux vector field is as follows: Based on the three-dimensional spatial temperature field data of the melt, the diffusion coefficient of key impurity elements at each spatial point of the current temperature field is calculated according to the preset formula, forming a spatial distribution field of diffusion coefficient. Spatial three-dimensional interpolation processing is performed on the concentration signals of key impurity elements to construct a three-dimensional concentration distribution field of key impurity elements in the melt; Based on the three-dimensional concentration distribution field, the concentration gradient vector field of key impurity elements in the three-dimensional space of the melt is calculated through spatial differentiation. The diffusion flux vector field of the key impurity element in the three-dimensional space of the melt is calculated by multiplying the spatial distribution field of the diffusion coefficient and the concentration gradient vector field on the same spatial coordinate grid and using Fick's first law.

4. The control system for a melting furnace used in the refining of magnesium ingots according to claim 3, characterized in that: The calculation process for the spatial distribution field of the segregation coefficient is as follows: Based on the three-dimensional spatial temperature field data of the melt, the temperature distribution in the region adjacent to the solid-liquid interface isotherm at the current moment in the melt is extracted. The system calls up the magnesium alloy multi-phase diagram data stored in the database, queries the solid-liquid equilibrium concentration of key impurity elements under the corresponding magnesium melt composition based on the temperature distribution data of the isotherm adjacent region, and calculates the equilibrium segregation coefficient of the key impurity elements at the current interface temperature. Based on the component of the diffusion flux vector field in the normal direction of the solidification interface, combined with the interface propagation rate distribution obtained by the array of thermocouples set in the molten pool, the characteristic thickness of the solute diffusion boundary layer near the solidification interface is calculated. Based on the characteristic thickness of the equilibrium segregation coefficient and the interface propagation rate, the actual solute concentration ratio between the solid phase side and the liquid phase side at the solidification interface is calculated by the solute conservation equation and defined as the segregation coefficient. Based on the spatial morphology of the solid-liquid interface isotherms, the segregation coefficient is mapped to three-dimensional space to form a spatial distribution field of the segregation coefficient.

5. The control system for a melting furnace used in the refining of magnesium ingots according to claim 4, characterized in that: The process of generating the impurity potential enrichment region map and identifying high-risk regions where the concentration of key impurity elements exceeds a preset permissible threshold is as follows: By integrating the diffusion flux vector field and the spatial distribution field of the segregation coefficient, and based on the mass transfer and segregation coupling mechanism of impurity elements in the melt, the net accumulation rate distribution of impurity elements in three-dimensional space is calculated. The spatial distribution evolution of impurity concentration within a preset time period is obtained by integrating the net accumulation rate distribution and the three-dimensional concentration distribution field of key impurity elements over time, thereby generating a map of potential impurity enrichment regions. The impurity concentration values ​​at each spatial coordinate point in the potential impurity enrichment region map are compared with a preset permissible threshold. All spatial points exceeding the preset permissible threshold are identified. Three-dimensional spatial clustering analysis is performed on the identified spatial points, and regions that are spatially connected and have impurity concentration values ​​exceeding the preset permissible threshold are defined as high-risk regions.

6. The control system for a melting furnace used in the refining of magnesium ingots according to claim 4, characterized in that: The process of determining the target stirring region and the target temperature gradient adjustment direction is as follows: Based on the diffusion flux vector field, the spatial distribution field of the segregation coefficient, and the solidification interface propagation rate, the convection concentration increment and the accumulation concentration increment of the segregation effect at the solidification interface per unit time are calculated using flux integral and solidification front solute redistribution theory, respectively. If the increase in convective concentration is greater than the increase in accumulated concentration, the main cause is determined to be insufficient solute convective transport. Based on the vector direction of the diffusion flux vector field and the spatial location of the high-risk area, the target stirring area is determined and the first stirring strategy is generated. If the increase in convection concentration is less than the increase in accumulation concentration, the main cause is determined to be the segregation effect at the solidification interface. Based on the spatial gradient of the spatial distribution field of the segregation coefficient and the three-dimensional spatial temperature field data of the melt, the direction of adjustment of the target temperature gradient is calculated, and the first temperature control strategy is generated. If the absolute value of the difference between the convection concentration increment and the accumulation concentration increment is less than or equal to the preset tolerance range, it is determined to be a compound cause, and at the same time, a coordinated instruction for the first stirring strategy and the first temperature control strategy is generated.

7. A melting furnace control system for refining magnesium ingots according to claim 6, characterized in that: The process of adjusting the zone excitation parameters of the electromagnetic stirrer is as follows: If the first stirring strategy is received, then according to the spatial coordinates of the target stirring area, the electromagnetic stirrer partitions covering the area are matched, and according to the direction and magnitude of the diffusion flux vector field, excitation parameter adjustment instructions for each partition to enhance the convective mass transfer to the high-risk area are calculated and generated.

8. The control system for a melting furnace used in the refining of magnesium ingots according to claim 7, characterized in that: The process of adjusting the cooling water volume of the melt water jacket is as follows: If the received temperature control strategy is the first one, then the direction of adjustment is adjusted according to the target temperature gradient, the specific melt water jacket that affects the heat transfer in the direction is matched, and the cooling water volume adjustment command for each target water jacket is calculated and generated according to the required gradient change.

9. A melting furnace control system for refining magnesium ingots according to claim 8, characterized in that: The specific collaborative control command is as follows: When an excitation parameter adjustment command or a cooling water volume adjustment command is generated, the expected disturbance caused by the excitation parameter adjustment command to the three-dimensional temperature field of the melt is analyzed, and a compensating water volume adjustment command is generated to counteract the thermal disturbance. The expected disturbance to the diffusion flux vector field caused by the cooling water flow regulation command is analyzed, and a compensatory excitation parameter adjustment command is generated to maintain the flow intensity in the target area. The adjustment instructions and compensatory instructions for the same control objective are integrated and output as a coordinated control instruction.

10. A melting furnace for refining magnesium ingots, characterized in that: include: A furnace body for containing molten magnesium, and a heating device coupled to the furnace body for heating and heat preservation of the materials inside the furnace; Multiple independently controllable electromagnetic stirrers are disposed at the bottom and / or side of the furnace body for driving the flow of melt when energized. The furnace body is embedded between the side wall and the furnace lining. It has cooling water channels inside and is used to cool the melt in sections by adjusting the cooling water flow rate to control the temperature gradient of the melt. The multiple water jacket cooling units are arranged in sections along the circumference of the molten pool. Each water jacket cooling unit has a cooling water channel inside and is connected to an independent flow regulating valve.

Images