Silicon-zirconium smelting temperature self-adaptive control method based on temperature field dynamic prediction

Abstract

The present application relates to the field of silicon-zirconium smelting temperature adaptive control, and particularly relates to a silicon-zirconium smelting temperature adaptive control method based on temperature field dynamic prediction. A continuous temperature field is constructed based on temperature data and equipment running state data; change data of historical regional heat input power and change response data of corresponding spatial position temperature values are obtained, a non-migration temperature rise rate is constructed, and a temperature structure apparent migration speed vector is calculated, and then in combination with the temperature field of the current control period, a predicted temperature field and a predicted effectiveness check are obtained; a regional heat input redistribution control strategy is constructed, and a candidate regional power adjustment scheme is generated, the candidate regional power adjustment scheme is screened and evaluated, and a prediction feedback closed-loop optimization control of the temperature field of the direct current electric mineral heat furnace is realized. The problems of incomplete temperature field perception, external power disturbance affecting temperature judgment and inability to realize spatial heat optimization configuration under total input power constraint are solved.

Description

Technical Field

[0001] This invention relates to the field of adaptive control of silicon-zirconium melting temperature, and more particularly to an adaptive control method for silicon-zirconium melting temperature based on dynamic prediction of the temperature field. Background Technology

[0002] DC electric submerged arc furnaces, such as those used for silicon-zirconium smelting, are typical high-temperature electrothermal smelting equipment. Their smelting quality, energy consumption, and operational stability are all closely related to the spatial distribution of the temperature field within the furnace. In actual production, DC electric submerged arc furnaces typically employ multiple zone heat input regulating units configured in different areas to heat the materials inside the furnace. Due to the complex influence of various factors such as heating location, furnace structure, heat conduction path, and the flow state of the melt in the molten pool, uneven spatial temperature distribution is highly likely to occur within the DC electric submerged arc furnace.

[0003] Traditional DC electric arc furnace temperature control methods typically rely on limited temperature sensing devices to collect localized temperature information, adjusting the overall heating power or localized heating status accordingly. However, the temperature field within a DC electric arc furnace exhibits significant nonlinearity and spatial distribution characteristics. A limited number of temperature sensing points cannot fully reflect the true thermal state of the entire molten pool region, easily leading to the failure to detect localized temperature anomalies in a timely manner. Furthermore, adjustments to the heat input power in different areas of the DC electric arc furnace directly cause temperature changes at corresponding spatial locations. These changes are intertwined with and superimposed on factors such as heat propagation and temperature structure migration within the melt, making it difficult to accurately distinguish the source of detected temperature changes. Specifically, it is impossible to differentiate between non-migratory temperature rises caused by external power regulation and those caused by internal thermal structure migration, thus severely impacting the accuracy of judging the trend of temperature changes within the furnace.

[0004] Furthermore, existing technologies mostly rely on simple feedback adjustments based on current temperature deviations, failing to predict future trends in furnace temperature distribution after power adjustments. When insufficient melting temperature occurs in localized areas of a DC electric submerged arc furnace, simply increasing the overall input power, while improving the low-temperature region, can easily lead to excessively high temperatures in other normal areas, causing localized overheating risks, accelerated furnace lining erosion, and energy waste. Therefore, existing DC electric submerged arc furnace control methods struggle to achieve adaptive and optimized heat distribution across different areas within the furnace when total power supply is limited. Summary of the Invention

[0005] This invention provides an adaptive control method for silicon-zirconium melting temperature based on dynamic temperature field prediction, in order to solve the problems of incomplete temperature field perception, external power disturbance affecting temperature judgment, and inability to achieve optimal spatial heat allocation under total input power constraints in the prior art.

[0006] The adaptive control method for silicon-zirconium smelting temperature based on dynamic temperature field prediction of the present invention includes the following steps:

[0007] S1. Acquire temperature data and equipment operating status data of the DC electric submerged arc furnace equipped with regional heat input regulation units, and divide the smelting area inside the furnace into two or more spatial calculation units. Construct a temperature field for each spatial region of the DC electric submerged arc furnace, consisting of the temperature values ​​at the spatial locations of each spatial calculation unit in each control cycle. Based on the acquired changes in regional heat input power of each regional heat input regulation unit in different control cycles, and combined with the temperature field in the corresponding control cycle, calculate the non-migrating temperature rise rate. Based on the temperature field and non-migrating temperature rise rate in the current control cycle, calculate the apparent migration velocity vector of the temperature structure in the current control cycle. Based on the temperature field and apparent migration velocity vector of the temperature structure in the current control cycle, predict the trend of temperature distribution change in the furnace in the next control cycle, and obtain the predicted temperature field for the next control cycle.

[0008] S2. Based on the predicted temperature field for the next control cycle, risk assessment is performed for different spatial regions; a regional heat input redistribution control strategy is constructed, and candidate regional power adjustment schemes are generated; the candidate regional power adjustment schemes are screened, and for the retained candidate regional power adjustment schemes, the predicted temperature field for the next control cycle after the implementation of different candidate regional power adjustment schemes is calculated; based on the predicted temperature field for the next control cycle after the implementation of different candidate regional power adjustment schemes, a future abnormal temperature evaluation index is constructed; based on the abnormal temperature evaluation index, the retained candidate regional power adjustment schemes are evaluated, and the actual execution scheme for the next control cycle is obtained.

[0009] Preferably, S1 specifically includes:

[0010] Based on the temperature field, the difference between the temperature value of the next control cycle and the temperature value of the previous control cycle after the change in regional heat input power is calculated, and the ratio of the difference to the control cycle difference is calculated to obtain the actual temperature change rate of each spatial location after the change in regional heat input power; and the change in regional heat input power of the same regional heat input regulation unit and the corresponding actual temperature change rate of the spatial location are recorded to form a historical correspondence between the change in regional heat input power and the actual temperature change rate of the spatial location.

[0011] Preferably, S1 specifically includes:

[0012] In the current control cycle, calculate the change in regional heat input power and its corresponding spatial location. In the historical correspondence, if the change in regional heat input power is the same as the change in regional heat input power in a certain historical record, directly call the actual temperature change rate of the spatial location corresponding to the change in regional heat input power as the non-migrating temperature rise rate. When there is no regional heat input power change that is exactly the same as the one in the historical record, use linear interpolation to calculate the temperature change rate of the corresponding spatial location as the non-migrating temperature rise rate.

[0013] Preferably, S1 specifically includes:

[0014] A non-migration temperature rise separation mechanism is introduced. Based on the non-migration temperature rise rate, combined with the time change of the temperature field and the spatial gradient of the temperature field, the apparent migration velocity vector of the temperature structure is calculated.

[0015] Preferably, S1 specifically includes:

[0016] Based on the temperature field and apparent migration velocity vector of the temperature structure in the current control cycle, a spatial displacement vector is generated, and the spatial position is updated along the temperature field gradient direction to obtain the predicted position, thereby obtaining the predicted temperature field for the next control cycle.

[0017] Preferably, S1 specifically includes:

[0018] Based on the actual temperature field of the next control cycle and the predicted temperature field generated in the previous cycle, the temperature deviation of each spatial computing unit is calculated; the average absolute error of the temperature deviation of each spatial computing unit is calculated based on the temperature deviation of each spatial computing unit; the validity of the prediction result is determined based on the average absolute error; if the prediction result is valid, regional power adjustment based on the predicted temperature field is allowed; otherwise, regional power migration control based on the predicted temperature field is stopped.

[0019] Preferably, S2 specifically includes:

[0020] By introducing a target temperature range, the temperature field is predicted in the next control cycle based on the power adjustment schemes of different candidate regions, and high-temperature risk terms and low-temperature risk terms are constructed. Spatial evaluation coefficients are introduced, and the high-temperature risk terms and low-temperature risk terms are weighted and summed. The results are then normalized by a normalization term composed of spatial evaluation coefficients to obtain an abnormal temperature evaluation index.

[0021] Preferably, S2 specifically includes:

[0022] When there are abnormal regions in the predicted temperature field corresponding to the candidate region power adjustment scheme that exceed the target temperature range, the candidate region power adjustment scheme with the smallest abnormal temperature evaluation index is selected.

[0023] Preferably, S2 specifically includes:

[0024] When the abnormal temperature evaluation index of the candidate region power adjustment scheme is zero, any candidate region power adjustment scheme that meets the target temperature range is selected as the actual execution scheme for the next control cycle.

[0025] The beneficial effects of the technical solution of the present invention are:

[0026] 1. By constructing a continuous temperature field inside the furnace and separating the non-migrating temperature rise caused by changes in external heat input based on the historical regional heat input power change data and the corresponding spatial location temperature value change response data, the influence of power disturbance on temperature change judgment is reduced, and the accuracy of temperature structure migration trend prediction is improved.

[0027] 2. By predicting the future temperature distribution, potential high-temperature risk areas and insufficient melting areas are identified. While keeping the total input electrical power constant, the thermal input power of different areas is spatially reconfigured to optimize the heat source distribution.

[0028] 3. By predicting the temperature field, evaluate the future temperature stability of different candidate region power adjustment schemes, select a control scheme that can reduce the risk of abnormal temperature, thereby improving the temperature field uniformity and operational stability of the DC electric arc furnace.

[0029] 4. By verifying the predicted temperature field against the actual temperature state, unreasonable adjustments caused by prediction deviations are avoided, improving the reliability of system control and realizing the transformation of the silicon-zirconium melting process from traditional temperature feedback control to predictive spatial thermal field optimization control. Attached Figure Description

[0030] Figure 1 This is a flowchart of the adaptive control method for silicon-zirconium melting temperature based on dynamic temperature field prediction as described in this invention. Detailed Implementation

[0031] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments. 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.

[0032] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0033] The following description, in conjunction with the accompanying drawings, details the specific scheme of the adaptive control method for silicon-zirconium melting temperature based on dynamic temperature field prediction provided by this invention.

[0034] See attached document Figure 1 The diagram illustrates a flowchart of an adaptive control method for silicon-zirconium melting temperature based on dynamic temperature field prediction, provided by an embodiment of the present invention. The method includes the following steps:

[0035] S1. Acquire temperature data and equipment operating status data of the DC electric submerged arc furnace equipped with regional heat input regulation units, and divide the smelting area inside the furnace into two or more spatial calculation units. Construct a temperature field for each spatial region of the DC electric submerged arc furnace, consisting of the temperature values ​​at the spatial locations of each spatial calculation unit in each control cycle. Based on the acquired changes in regional heat input power of each regional heat input regulation unit in different control cycles, and combined with the temperature field in the corresponding control cycle, calculate the non-migrating temperature rise rate. Based on the temperature field and non-migrating temperature rise rate of the current control cycle, calculate the apparent migration velocity vector of the temperature structure in the current control cycle. Based on the temperature field and apparent migration velocity vector of the temperature structure in the current control cycle, predict the trend of temperature distribution change in the furnace in the next control cycle, and obtain the predicted temperature field for the next control cycle.

[0036] First, during the operation of a DC electric submerged arc furnace equipped with multiple zone heat input regulation units, furnace temperature data is acquired through multiple temperature detection units located at different spatial positions within the furnace body. There are at least two temperature detection units, each with fixed spatial coordinates, used to collect temperature change information for its corresponding spatial area. The temperature data output by each detection unit includes the coordinates of the measurement location, the temperature acquisition time, and the corresponding temperature value. The coordinates of the measurement location are determined by the installation position of the temperature detection unit, the temperature value is directly acquired by on-site temperature detection equipment such as high-temperature thermocouples and infrared temperature measuring devices installed at different spatial positions within the DC electric submerged arc furnace, and the acquisition time is synchronously recorded by the DC electric submerged arc furnace's electrothermal control system.

[0037] Simultaneously, the operating status data of the heat input regulating units in each zone is acquired through the DC electric submerged arc furnace electrothermal control system. This operating status data includes the actual regional heat input power output by each zone's heat input regulating unit in the current control cycle and the historical heat input power of the corresponding zone in the previous control cycle. The difference between the regional heat input power in the current control cycle and the previous control cycle is used to determine the change in regional heat input power between adjacent control cycles. The regional heat input power is the heat output from the regional heat input regulating unit into the furnace. The historical records are automatically collected and stored by the DC electric submerged arc furnace electrothermal control system.

[0038] After acquiring the above data, the temperature values ​​at different spatial locations are synchronized according to the temperature acquisition time, ensuring that the data from each temperature measurement point within the same control cycle correspond to the same time state. Furthermore, a correspondence between temperature values ​​and spatial calculation units is established based on the spatial location of the temperature measurement points. Specifically, a three-dimensional Cartesian coordinate system is established based on the actual smelting space within the DC electric submerged arc furnace. Using the furnace height, radial direction, and circumferential direction as spatial division directions, the smelting area within the furnace is divided into multiple regular hexahedral spatial calculation units. The size of each spatial calculation unit is determined based on the maximum spatial distance between temperature detection units and the heating influence range corresponding to the regional heat input adjustment unit, ensuring that the distance between adjacent spatial calculation units meets the temperature gradient calculation requirements. Fixed spatial intervals are set along the three coordinate directions for each spatial calculation unit, and the center position of each spatial calculation unit corresponds to a spatial position index. This is used to represent the temperature state of the spatial region. For incomplete areas such as near the furnace wall, furnace bottom, and furnace top, the regular spatial calculation units are truncated according to the actual boundary shape of the DC electric submerged arc furnace, retaining only the spatial calculation units located within the smelting area. For boundary spatial calculation units adjacent to the furnace wall, the temperature state of the adjacent internal spatial calculation units is used as the boundary temperature calculation reference to avoid the loss of the temperature field calculation area due to the furnace structure boundary. In this embodiment, the furnace temperature calculation area is divided into 27 spatial calculation units of 3×3×3. The heating influence range corresponding to the regional heat input adjustment unit is determined according to the rated output power of the regional heat input adjustment unit, the effective radius of the electrode, and the thermal conductivity of the melt in the furnace. Furthermore, the temperature measurement position coordinates of the temperature detection units installed in the furnace smelting area are taken as known temperature nodes, and the center positions of all spatial calculation units in the furnace smelting area without temperature detection units are taken as nodes to be calculated. For each node to be calculated, the multiple known temperature nodes with the closest spatial distance to the corresponding node to be calculated are first determined, and the temperature value and spatial position coordinate information of the known temperature nodes are obtained. As a specific embodiment, the distance is selected. The nearest known temperature nodes, not exceeding the total number of known temperature nodes, are used in the calculation. Then, a temperature influence relationship is established based on the spatial distance between the node to be calculated and each known temperature node. The closer a known temperature node is to the node to be calculated, the greater its influence on the current node. As the distance between the node to be calculated and the known temperature nodes increases, the influence of the temperature state on the current node gradually decreases. Finally, based on the spatial positional relationship of multiple known temperature nodes and their corresponding temperature values, spatial interpolation is performed on the temperature state of the node to be calculated. Specifically, an inverse distance-weighted spatial interpolation method based on distance weights is used. First, the three-dimensional Euclidean distance between the node to be calculated and each known temperature node is calculated as the spatial distance. Spatial influence weights are established based on these spatial distances, with higher weights for known temperature nodes closer to the node to be calculated and lower weights for known temperature nodes farther away. Subsequently, a weighted calculation is performed based on the temperature values ​​of each known temperature node and their corresponding spatial influence weights to obtain the temperature value at the spatial location of the node to be calculated. This ensures that each spatial calculation unit's spatial location index has a corresponding temperature state, forming a temperature field covering all spatial regions of the DC electric arc furnace. ,in, Indicates the spatial location index of the temperature field. Indicates the control cycle index. Indicates the first Spatial position of each control cycle The temperature value at that location.

[0039] Furthermore, based on the historical data collected during the operation of the DC electric submerged arc furnace regarding changes in regional heat input power and corresponding spatial location temperature response data, the impact of equipment operating status data changes on spatial temperature changes is determined. Anomaly detection is performed on the historically collected regional heat input power and temperature response data, removing historical data that exceeds the specified operating range of the DC electric submerged arc furnace, exceeds the allowable temperature range of the smelting process, or is missing, to ensure the validity of the historical response relationship. Specifically, during the historical operation of the DC electric submerged arc furnace, the changes in regional heat input power of each regional heat input adjustment unit within different control cycles are acquired. Simultaneously, based on the reconstructed temperature field within the corresponding control cycle, the difference between the temperature value of the subsequent control cycle after the change in regional heat input power and the temperature value of the previous control cycle is calculated. The ratio of this difference to the control cycle difference is then calculated to obtain the actual temperature change rate of each spatial location after the change in regional heat input power. The changes in regional heat input power of the same regional heat input adjustment unit within different control cycles and the corresponding actual temperature change rate of the spatial location are recorded accordingly, forming a historical correspondence between the changes in regional heat input power and the actual temperature change rate of the spatial location. The establishment of the above historical correspondence requires that, within each sampling control cycle, only a single area heat input adjustment unit is allowed to change its regional heat input power, while the heat input adjustment units of other areas maintain constant power. This ensures that the actual temperature change rate of the recorded spatial location is caused only by the power change of that single area heat input adjustment unit, thus eliminating the coupling interference of power changes of multiple units.

[0040] In the current control cycle, the change in regional heat input power and its corresponding spatial location are calculated based on the detection data. In the aforementioned historical correspondence, the two historical regional heat input power changes closest to the change in regional heat input power in the current control cycle are selected as interpolation data points. If the current regional heat input power change is the same as a historical change in regional heat input power, the actual temperature change rate at the corresponding spatial location is directly used. If no identical historical regional heat input power change exists, the temperature change rate at the corresponding spatial location is calculated using linear interpolation based on the distance between the current regional heat input power change and the two adjacent historical regional heat input power changes, and this temperature change rate is used as the spatial location. Non-migration temperature rise rate caused by changes in regional heat input power .in, Indicates the spatial location caused by changes in the current device operating status data. Regarding the temperature change portion; the historical correspondence is established under the constraint that only a single region's heat input regulation unit is allowed to change power, while the heat input regulation units in other regions maintain constant power. This ensures that the measured temperature change rate at the corresponding spatial location is mainly caused by the direct change in the external power input of that single unit, and is therefore regarded as the non-migrating temperature rise rate. .

[0041] Obtaining the non-migration temperature rise rate Subsequently, based on the theory of temperature field spatial migration, the migration trend of the temperature structure, i.e., the apparent migration velocity vector of the temperature structure, is determined by the relationship between the temporal change of the temperature field and the spatial gradient. Addressing the issue of non-migration temperature rise caused by changes in regional heat input power during silicon-zirconium melting, a non-migration temperature rise separation mechanism is introduced to remove temperature changes caused by equipment status changes from the actual temperature changes. Combined with constraints on the spatial gradient direction and stabilization processing based on temperature resolution, the apparent migration velocity vector of the temperature structure after removing the influence of external heat input is obtained. The formula for calculating the apparent migration velocity vector of temperature structure is as follows: , in, Indicates the first Within each control cycle, spatial location The apparent migration velocity vector of the temperature structure along the temperature gradient direction is used to describe the tendency of the temperature structure to move spatially along the temperature gradient direction. Indicates the first Spatial position of each control cycle Temperature value at; Indicates the first Spatial position of each control cycle Temperature value at; when hour, , The initial temperature field is obtained by collecting temperature data from each temperature detection unit at the initial moment before control begins, and reconstructing the field based on the temperature data and corresponding spatial locations. To control the periodic time interval; Indicates the first Within each control cycle, at the current spatial location The spatial gradient of the temperature field at that location, based on the temperature field of the current control cycle. It is determined that the ratio of the temperature difference between adjacent spatial computing units to the corresponding spatial distance is used to characterize the degree and direction of temperature change along the spatial direction at the current location. This indicates the minimum temperature change that the temperature detection unit can distinguish, and is determined by the factory-rated parameters of the temperature detection equipment. The characteristic spatial scale representing the spatial division of the temperature field is taken as the minimum value among the fixed spatial intervals in the three coordinate directions; in calculating the spatial gradient of the temperature field... When calculating the directional components, a fixed spatial interval corresponding to the coordinate direction is used as the position difference step size for that direction. This represents the change in temperature over time. Indicates the current spatial location The combined rate of temperature change detected between adjacent control cycles; This represents the effective temperature structure change rate after removing the influence of regional thermal input power. It indicates the migration trend of effective temperature change along the spatial gradient direction of the temperature field; Indicates the current spatial location The intensity of the spatial gradient change in the temperature field at that location; It represents the smallest identifiable temperature gradient formed by the combined temperature measurement accuracy and spatial computation scale; This represents the smallest identifiable squared temperature gradient term, which serves as a stabilization constraint for the squared temperature gradient term. Its squared form can be used in conjunction with... Maintaining the same mathematical form and dimensions, it directly participates in denominator normalization calculations; This represents the normalized constraint term formed by the intensity of the spatial gradient change in the temperature field and the minimum identifiable temperature gradient.

[0042] Furthermore, based on the temperature field of the current control cycle and the apparent migration velocity vector of the temperature structure at the current location. This involves predicting the trend of furnace temperature distribution changes during the next control cycle. Specifically, it involves predicting the temperature field of the current control cycle. As the initial temperature state, according to The migration trend of the temperature structure at the current spatial location along the temperature field gradient direction is characterized, and the spatial displacement vector of the corresponding temperature structure at the next control cycle time interval is determined. Furthermore, based on the spatial displacement vector corresponding to each spatial location, the spatial position of the temperature state in the current temperature field is updated. That is, the predicted position after migration along the temperature field gradient direction is determined as the center position of the spatial computing unit plus the spatial displacement vector. Further, the predicted position after migration of each spatial computing unit and its corresponding current temperature state are taken as known temperature nodes, and the center positions of each spatial computing unit in the next control cycle are taken as nodes to be calculated. For each node to be calculated, multiple known temperature nodes with the closest spatial distance are determined, and the known temperature nodes with the closest distance and not exceeding the total number of known temperature nodes are selected to participate in the calculation. Spatial influence weights are established based on the spatial distance between the node to be calculated and each known temperature node. The above-mentioned inverse distance weighted spatial interpolation method is used to obtain the temperature value of the spatial position corresponding to the node to be calculated, so that the center position of each spatial computing unit has a corresponding predicted temperature state, forming the predicted temperature field for the next control cycle. Specifically, when a single known temperature node coincides with the node to be calculated, the temperature state corresponding to the known temperature node is directly assigned to the corresponding spatial computing unit; when the predicted positions of two or more known temperature nodes coincide with the center position of the same spatial computing unit, the arithmetic mean of the temperature states corresponding to the two or more known temperature nodes is taken as the predicted temperature value of the center position of the spatial computing unit.

[0043] After the next control cycle begins, temperature data at different spatial locations inside the furnace are re-acquired through the temperature detection unit, and the actual temperature field is obtained according to the aforementioned temperature field construction method. The actual temperature field is compared with the predicted temperature field generated in the previous period, and the temperature deviation of each spatial calculation unit is obtained. Based on the temperature deviation of each spatial calculation unit, the average absolute error of the temperature deviation of the spatial calculation unit is calculated. The average absolute error between the actual temperature field and the predicted temperature field does not exceed the minimum temperature change that the temperature detection unit can distinguish. The current predicted temperature field is considered If the predicted temperature field can reflect the actual temperature change trend, then the prediction result is valid and allows for subsequent regional power adjustments based on the predicted temperature field; otherwise, the predicted temperature field is considered invalid. It cannot accurately describe the actual temperature structure changes and stops performing regional power migration control based on the predicted temperature field.

[0044] S2. Based on the predicted temperature field for the next control cycle, risk assessment is performed for different spatial regions; a regional heat input redistribution control strategy is constructed, and candidate regional power adjustment schemes are generated; the candidate regional power adjustment schemes are screened, and for the retained candidate regional power adjustment schemes, the predicted temperature field for the next control cycle after the implementation of different candidate regional power adjustment schemes is calculated; based on the predicted temperature field for the next control cycle after the implementation of different candidate regional power adjustment schemes, a future abnormal temperature evaluation index is constructed; based on the abnormal temperature evaluation index, the retained candidate regional power adjustment schemes are evaluated, and the actual execution scheme for the next control cycle is obtained.

[0045] After completing the prediction of the temperature distribution trend for the next control cycle and confirming the validity of the prediction results, based on the predicted temperature field... The future temperature distribution trend is analyzed to determine whether there is a risk of overheating or insufficient melting in different spatial regions. Specifically, the temperature range corresponding to the molten state is determined according to the requirements of the silicon-zirconium smelting process. The liquidus temperature of silicon-zirconium smelting material is determined by differential thermal analysis based on the actual raw materials used in production. The upper limit of the safe operating temperature for a DC electric submerged arc furnace is determined by the furnace body's temperature resistance rating and the allowable operating temperature of the heating elements, both specified in the furnace design parameters. The smaller of these two values ​​is selected as the upper limit of the safe operating temperature. This is based on the spatial location within the predicted temperature field. The predicted temperature at the location meets the requirements. When the future temperature state of the space region is considered to meet the smelting requirements; when the spatial position in the predicted temperature field... Predicted temperature at the location Higher than At that time, it was believed that there was a risk of excessively high temperatures in the space region in the future; when the spatial location in the predicted temperature field... Predicted temperature at the location Below At that time, it was believed that there was a risk of insufficient temperature and insufficient melting in the space region in the future.

[0046] Furthermore, a regional heat input redistribution control strategy is constructed. Based on the spatial distribution of abnormal temperatures in the predicted temperature field and the spatial relationship between the corresponding heat input adjustment units in each region, the spatial regions that need to reduce heat input and those that need to increase heat input are determined. Specifically, the regional heat input adjustment units corresponding to the spatial regions predicted to have excessively high temperatures are designated as heat input adjustment output regions, and the regional heat input adjustment units corresponding to the spatial regions predicted to have insufficient temperatures are designated as heat input compensation regions. Temperature field adjustment is achieved through the redistribution of heat input between regions.

[0047] The spatial position relationship of each area heat input regulation unit is determined as follows: the rated heating center position of each area heat input regulation unit in the furnace is taken as the spatial reference point. When the three-dimensional Euclidean distance between the center position of the spatial calculation unit and the spatial reference point is less than or equal to the heating influence range corresponding to the area heat input regulation unit, it is determined that the spatial calculation unit is located within the heating influence range of the area heat input regulation unit, and the attribution correspondence between the spatial calculation unit and the corresponding area heat input regulation unit is established.

[0048] After determining the heat input adjustment output region and the heat input compensation region, instead of improving the low-temperature region by increasing the overall heat input power of the DC electric arc furnace, the heat input power of different regions is redistributed among the heat input adjustment units while keeping the total input power of the DC electric arc furnace constant. Specifically, based on the actual regional heat input power of each region's heat input adjustment unit in the current control cycle, the regional heat input power of all regions' heat input adjustment units predicted to have an excessively high temperature risk is reduced according to the temperature deviation ratio, and the adjusted candidate region heat input power is... The heat input adjustment units in the predicted temperature-insufficient areas are increased according to the temperature deviation ratio to form new regional heat input power schemes, i.e., candidate regional power adjustment schemes. The maximum allowable power adjustment range of the furnace regional heat input adjustment units is used as the upper limit of the candidate regional heat input power. Multiple candidate regional heat input power values ​​are gradually generated from zero to the upper limit of the constrained value within the range that meets the equipment operating constraints, with a preset power adjustment step size as the interval, until the upper limit of the constrained value is reached and generation stops. The preset power adjustment step size is determined based on the power response resolution of the regional heat input adjustment units and the temperature stability requirements within the control cycle, and is a conventional parameter that can be reasonably set by those skilled in the art based on the equipment characteristics. The temperature deviation ratio refers to the calculation of the power adjustment weight of the corresponding region based on the degree of deviation of the predicted temperature of each abnormal space region from the target temperature range when there are multiple over-temperature regions and multiple under-temperature regions. The regional heat input power adjustment amount is then allocated according to the power adjustment weight of each region, so that each regional heat input adjustment unit receives a power adjustment amount that matches the degree of its predicted temperature anomaly. Specifically, over-temperature regions are defined as those where the predicted temperature exceeds the upper limit of the safe operating temperature. The deviation determines the power output adjustment weight, and the temperature deficiency area is determined based on the predicted temperature being below the liquidus temperature. The deviation determines the power compensation weight. During the adjustment process, the total output power of the heat input regulation units in all regions remains consistent before and after the adjustment, only changing the distribution of heat input within the spatial region of the DC electric arc furnace. Let the first... The power output adjustment weight of the heat input regulation unit corresponding to each overheated area is: , No. The power compensation weight of the heat input regulation unit corresponding to each temperature-deficient region is: ,in , These are the indices of the heat input adjustment units corresponding to the overheated and underheated areas, respectively. , These are the sets of abnormal spatial locations associated with the corresponding region's heat input regulation unit; the heat input power of the candidate region... The power reduction for each region is allocated to the heat input regulation unit according to the proportion of each region's weight to the total weight of the corresponding category. , The total number of heat input regulation units corresponding to the overheated area; the power increase for the underheated area is... , This represents the total number of regional heat input regulation units corresponding to areas with insufficient temperature.

[0049] Then, based on the maximum and minimum output power limits of the regional heat input regulating units, the maximum power change limit per unit control cycle, and the constraint of maintaining constant total input power, the generated candidate regional power adjustment schemes are screened. Schemes that cannot meet the equipment operation requirements are eliminated, while those that do meet the equipment operation conditions are retained. If none of the candidate regional power adjustment schemes meet the equipment operation requirements, the current output power of each regional heat input regulating unit remains unchanged, and a power adjustment constraint conflict warning is output. In the next control cycle, temperature data is re-acquired, and temperature field reconstruction, temperature structure migration analysis, and regional power optimization processes are executed. The maximum and minimum output power limits are determined by the rated output capacity and stable operation requirements of the regional heat input regulating units, respectively; the maximum power change limit per unit control cycle is determined by the response capability of the regional heat input regulating units and the temperature change stability requirements of the DC electric arc furnace; the constraint of maintaining constant total input power is determined by the regional heat input redistribution control strategy to ensure that the overall heat input level of the DC electric arc furnace does not change during the adjustment process. For each retained candidate region power adjustment scheme, based on the historical relationship between regional heat input power changes and spatial temperature response, the non-migratory temperature rise rate caused by regional heat input power changes at the corresponding spatial location of the candidate region power adjustment scheme is obtained. Wherein, when the candidate region power adjustment scheme... When multiple regional heat input regulation units are adjusted simultaneously, the non-migrating temperature rise rate corresponding to each unit is synthesized into a spatial location according to the following weighted linear superposition rule. The unique non-migrating temperature rise rate; further, the apparent migration velocity vector of the temperature structure in S1. Using the current temperature structure migration trend as the basis for prediction, and based on the temperature field state of the current control cycle, the temperature structure migration trend, and the non-migration temperature rise rate corresponding to the power adjustment schemes in the candidate regions, the predicted temperature field for the next control cycle after the implementation of different candidate region power adjustment schemes is obtained. ,in, This represents the candidate region power adjustment scheme. (Candidate region power adjustment scheme) Thermal input power of candidate region The heat input adjustment units for each region are allocated according to the aforementioned temperature deviation ratio, whereby... One of several discrete values ​​generated according to a preset power adjustment step size; candidate region power adjustment scheme The corresponding power change is Predicting the temperature field for the next control cycle. The mathematical expression is as follows: , in, Power adjustment scheme for candidate regions (The corresponding power change is) Under the influence of ) spatial position The non-migratory temperature rise rate caused by changes in regional heat input power is based on historical correlations and determined by candidate regional power adjustment schemes. The power change between the actual regional heat input power and the current control cycle is obtained by linear interpolation.

[0050] When candidate region power adjustment scheme When the regional heat input power of multiple regional heat input regulation units is changed simultaneously, the spatial position The non-migrating temperature rise rate caused by the power adjustment scheme in this candidate region is as follows: , in, Based on the aforementioned historical correspondence, by the first Changes in regional heat input power of each regional heat input regulation unit The non-migratory temperature rise rate obtained by linear interpolation, which is mainly caused by the change in external power input of the single unit; This represents the total number of regional heat input regulation units; For the first Each regional heat input adjustment unit controls spatial location. The non-migrating temperature rise influence weight is used to combine the individual response results of multiple regional heat input regulation units without changing the spatial location response already obtained in the historical correspondence of a single regional heat input regulation unit. Influence Weight Determined according to the following rules:

[0051] First, with the first The rated heating center position of each zone heat input regulating unit within the furnace is used as a spatial reference point. The relationship between this reference point and the spatial position is calculated. The three-dimensional Euclidean distance between them The rated heating center location is determined based on the spatial installation coordinates of the heat input regulating units in each zone of the DC electric submerged arc furnace design and installation data. An inverse distance weighting method is used. ; This is the index for the regional heat input adjustment unit. When At that time, take And the weights of the remaining units are reset to zero. When the spatial position Beyond the first When determining the heating influence range of a regional heat input regulation unit, the non-migratory temperature rise rate of the regional heat input regulation unit for its spatial location is set to zero, and it is not included in the weight normalization calculation.

[0052] The range of values ​​is determined by the spatial position of the current control cycle. Temperature value at and temperature structure apparent migration velocity vector Common constraints, when When it approaches 0, take ,at this time . The value range is constrained by the actual temperature change rate corresponding to the maximum and minimum regional heat input power changes recorded in the historical correspondence. When the power change of any regional heat input adjustment unit in the candidate regional power adjustment scheme exceeds the range of regional heat input power changes covered by the historical correspondence, the regional heat input power change is truncated to the maximum or minimum regional heat input power change in the historical correspondence. Based on the truncated regional heat input power change, the corresponding non-migrating temperature rise rate is determined according to the historical correspondence. When the regional heat input power change is within the coverage of the historical correspondence, the corresponding non-migrating temperature rise rate is determined by linear interpolation between two adjacent historical regional heat input power changes. If the predicted position exceeds the boundary of the furnace melting area after the spatial position update, the temperature state of the calculation unit is assigned according to the boundary space closest to it.

[0053] The relationship between the predicted temperature field obtained in the next control cycle and the predicted temperature field obtained in S1 during the calculation process of the predicted temperature field: The predicted temperature field obtained from step S1 As a fundamental term for temperature structure migration, candidate region power adjustment scheme The corresponding predicted temperature field is obtained by superimposing the non-migrating temperature rise caused by the candidate region power adjustment scheme onto this basic term; that is... It only carries the spatial migration information of the temperature structure after removing external power disturbances, and does not carry the non-migration temperature rise information caused by power changes in the candidate region power adjustment scheme.

[0054] The action mode and duration of the non-migration temperature rise rate corresponding to the candidate scheme: It acts on the corresponding spatial location in the form of a temperature rise rate, and integrates at a constant rate over the complete control cycle; its action is in the form of an algebraic superposition term. The duration of action is equal to the control cycle time interval. ;

[0055] The synthesis order and method of the temperature structure migration and non-migration temperature rise components: First, synthesize the temperature field of the current control cycle. Perform temperature structure space migration to obtain Then, the non-migrating temperature rise term is superimposed at each spatial location. and Perform linear algebraic addition.

[0056] After obtaining the predicted temperature field for the next control cycle after the implementation of different candidate region power adjustment schemes, an evaluation index for future abnormal temperatures is constructed based on predictive control principles to evaluate the temperature state after the implementation of each candidate region power adjustment scheme. Based on the predicted temperature field corresponding to the candidate region power adjustment scheme, and considering the material melting requirements and equipment safety operation limitations during silicon-zirconium smelting, the liquidus temperature of the smelting material is determined. Upper limit of safe operating temperature As the target temperature range, the deviation of the predicted temperature at each spatial location from the target temperature range is normalized. Simultaneously, a spatial evaluation coefficient is introduced based on temperature changes at different spatial locations to increase the impact of areas with significant temperature changes on the overall evaluation results and reduce the interference from areas with slow temperature changes. Finally, by summarizing the degree of abnormal temperature at each spatial location, an abnormal temperature evaluation index is obtained corresponding to the power adjustment schemes for candidate regions. This index evaluates the impact of different regional power adjustment schemes on the stability of the future temperature field and selects the regional power adjustment scheme that can reduce the risk of future abnormal temperatures. The calculation method for the abnormal temperature evaluation index is as follows: , in, Indicates the power adjustment scheme for the candidate region An abnormal temperature evaluation index for the degree of future temperature anomalies is used to compare the merits of different candidate region power adjustment schemes. The smaller the value, the closer the future temperature distribution is to the target temperature range. The spatial evaluation coefficient represents the degree of influence of temperature state changes at different spatial locations on the overall evaluation result. It is determined based on the temperature field gradient theory and is obtained by analyzing the current temperature field. Temperature gradient amplitude at various spatial locations The maximum value normalization process is performed to give higher evaluation weights to areas with significant temperature changes and lower evaluation weights to areas with gradual temperature changes. It is a very small constant, and can be taken as This is used to prevent division by zero errors; This represents the positive value cutoff function, used to extract abnormal deviations caused by the predicted temperature exceeding the upper limit of the safe operating temperature or falling below the liquidus temperature. When the input value of the positive value cutoff function is greater than zero, the original value is retained; when the input value of the positive value cutoff function is less than or equal to zero, the positive value cutoff function is set to zero, that is, when the temperature of the corresponding spatial location is within the target temperature range, the value is 0. For high temperature risk items, it means that if the predicted temperature of a spatial location exceeds the upper limit of the safe operating temperature after the power adjustment scheme for the candidate region is implemented, the degree to which the corresponding spatial location exceeds the upper limit of the safe operating temperature is calculated. The low-temperature risk item indicates that if the predicted temperature at a spatial location is lower than the melting requirement after the power adjustment scheme for the candidate region is implemented, the degree of low-temperature deficiency at the corresponding spatial location will be calculated. This is a normalization term used to normalize the spatial evaluation results, making the evaluation results of candidate region power adjustment schemes obtained under different temperature field division conditions and different spatial evaluation coefficients comparable.

[0057] Finally, based on the abnormal temperature evaluation index All candidate regional power adjustment schemes that meet the constraints of maximum output power limit, minimum output power limit, maximum power change limit per unit control cycle, and constant total input power of the regional heat input regulation unit are evaluated. When abnormal regions exceeding the safe operating temperature limit or below the liquidus temperature still exist in the predicted temperature field corresponding to the candidate regional power adjustment scheme, an abnormal temperature evaluation index is selected. The minimum candidate region power adjustment scheme is selected to reduce the risk of future temperature anomalies; when multiple candidate region power adjustment schemes are compared, the abnormal temperature evaluation index is considered. When all values ​​are zero, it indicates that the predicted temperature field after the implementation of each candidate region power adjustment scheme meets the target temperature range requirement. Then, any candidate region power adjustment scheme that meets the condition is selected from the candidate region power adjustment schemes where all the above abnormal temperature evaluation indicators are zero as the actual execution scheme for the next control cycle.

[0058] In summary, an adaptive control method for silicon-zirconium smelting temperature based on dynamic temperature field prediction has been developed.

[0059] The order of the embodiments is for illustrative purposes only and does not represent the superiority or inferiority of the embodiments. The processes depicted in the drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0060] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0061] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. An adaptive control method for silicon-zirconium smelting temperature based on dynamic temperature field prediction, characterized in that, Includes the following steps: S1. Acquire temperature data and equipment operating status data of the DC electric submerged arc furnace equipped with regional heat input regulation units, and divide the smelting area inside the furnace into two or more spatial calculation units. Construct a temperature field for each spatial region of the DC electric submerged arc furnace, consisting of the temperature values ​​at the spatial locations of each spatial calculation unit in each control cycle. Based on the acquired changes in regional heat input power of each regional heat input regulation unit in different control cycles, and combined with the temperature field in the corresponding control cycle, calculate the non-migrating temperature rise rate. Based on the temperature field and non-migrating temperature rise rate in the current control cycle, calculate the apparent migration velocity vector of the temperature structure in the current control cycle. Based on the temperature field and apparent migration velocity vector of the temperature structure in the current control cycle, predict the trend of temperature distribution change in the furnace in the next control cycle, and obtain the predicted temperature field for the next control cycle. S2. Based on the predicted temperature field for the next control cycle, risk assessment is performed for different spatial regions; a regional heat input redistribution control strategy is constructed, and candidate regional power adjustment schemes are generated; the candidate regional power adjustment schemes are screened, and for the retained candidate regional power adjustment schemes, the predicted temperature field for the next control cycle after the implementation of different candidate regional power adjustment schemes is calculated; based on the predicted temperature field for the next control cycle after the implementation of different candidate regional power adjustment schemes, a future abnormal temperature evaluation index is constructed; based on the abnormal temperature evaluation index, the retained candidate regional power adjustment schemes are evaluated, and the actual execution scheme for the next control cycle is obtained.

2. The adaptive control method for silicon-zirconium smelting temperature based on dynamic temperature field prediction according to claim 1, characterized in that, S1 specifically includes: Based on the temperature field, the difference between the temperature value of the next control cycle and the temperature value of the previous control cycle after the change in regional heat input power is calculated, and the ratio of the difference to the control cycle difference is calculated to obtain the actual temperature change rate of each spatial location after the change in regional heat input power; and the change in regional heat input power of the same regional heat input regulation unit and the corresponding actual temperature change rate of the spatial location are recorded to form a historical correspondence between the change in regional heat input power and the actual temperature change rate of the spatial location.

3. The adaptive control method for silicon-zirconium smelting temperature based on dynamic temperature field prediction according to claim 2, characterized in that, S1 specifically includes: In the current control cycle, calculate the change in regional heat input power and its corresponding spatial location. In the historical correspondence, if the change in regional heat input power is the same as the change in regional heat input power in a certain historical record, directly call the actual temperature change rate of the spatial location corresponding to the change in regional heat input power as the non-migrating temperature rise rate. When there is no regional heat input power change that is exactly the same as the one in the historical record, use linear interpolation to calculate the temperature change rate of the corresponding spatial location as the non-migrating temperature rise rate.

4. The adaptive control method for silicon-zirconium smelting temperature based on dynamic temperature field prediction according to claim 3, characterized in that, S1 specifically includes: A non-migration temperature rise separation mechanism is introduced. Based on the non-migration temperature rise rate, combined with the time change of the temperature field and the spatial gradient of the temperature field, the apparent migration velocity vector of the temperature structure is calculated.

5. The adaptive control method for silicon-zirconium smelting temperature based on dynamic temperature field prediction according to claim 4, characterized in that, S1 specifically includes: Based on the temperature field and apparent migration velocity vector of the temperature structure in the current control cycle, a spatial displacement vector is generated, and the spatial position is updated along the temperature field gradient direction to obtain the predicted position, thereby obtaining the predicted temperature field for the next control cycle.

6. The adaptive control method for silicon-zirconium smelting temperature based on dynamic temperature field prediction according to claim 5, characterized in that, S1 specifically includes: Based on the actual temperature field of the next control cycle and the predicted temperature field generated in the previous cycle, the temperature deviation of each spatial computing unit is calculated; the average absolute error of the temperature deviation of each spatial computing unit is calculated based on the temperature deviation of each spatial computing unit; the validity of the prediction result is determined based on the average absolute error; if the prediction result is valid, regional power adjustment based on the predicted temperature field is allowed; otherwise, regional power migration control based on the predicted temperature field is stopped.

7. The adaptive control method for silicon-zirconium smelting temperature based on dynamic temperature field prediction according to claim 1, characterized in that, S2 specifically includes: By introducing a target temperature range, the temperature field is predicted in the next control cycle based on the power adjustment schemes of different candidate regions, and high-temperature risk terms and low-temperature risk terms are constructed. Spatial evaluation coefficients are introduced, and the high-temperature risk terms and low-temperature risk terms are weighted and summed. The results are then normalized by a normalization term composed of spatial evaluation coefficients to obtain an abnormal temperature evaluation index.

8. The adaptive control method for silicon-zirconium smelting temperature based on dynamic temperature field prediction according to claim 7, characterized in that, S2 specifically includes: When there are abnormal regions in the predicted temperature field corresponding to the candidate region power adjustment scheme that exceed the target temperature range, the candidate region power adjustment scheme with the smallest abnormal temperature evaluation index is selected.

9. The adaptive control method for silicon-zirconium smelting temperature based on dynamic temperature field prediction according to claim 7, characterized in that, S2 specifically includes: When the abnormal temperature evaluation index of the candidate region power adjustment scheme is zero, any candidate region power adjustment scheme that meets the target temperature range is selected as the actual execution scheme for the next control cycle.

Images