Aluminum smelting process precision temperature control method and system

Abstract

This invention relates to the field of metallurgical furnace process control, and particularly to a precise temperature control method and system for aluminum smelting processes. Step S1: Collect and time-align nine types of data; Step S2: Input the data into a model to obtain composite tank temperature and wall heat dissipation indices; Step S3: Compare the composite tank temperature with the upper and lower limits of the target temperature zone to determine deviation, and decompose it into convective heat dissipation components, material parameters, and electrical parameter components; Step S4: If deviation occurs, adjust the exhaust air volume and waste heat bypass ratio according to the first control sequence to bring the composite tank temperature back to the target temperature zone; Step S5: If the temperature has not yet entered the target zone after reaching the main control evaluation time threshold, adjust the alumina feed rate, electrolytic cell operating current, and anode-cathode spacing in a coordinated manner until the temperature enters the target zone. This invention, through a model-filtered fusion of "multi-source observation → composite tank temperature" and coordinated main control of "exhaust air volume + waste heat bypass ratio" (with limited linkage when necessary), maintains the composite tank temperature stably within the target temperature zone while meeting environmental protection and anti-condensation constraints.

Description

Technical Field

[0001] This invention relates to the field of metallurgical furnace process control, and in particular to a method and system for precise temperature control in aluminum smelting processes. Background Technology

[0002] The stable operation of aluminum electrolytic cells depends on maintaining the thermal balance of the electrolyte (cryolite-alumina system) within a narrow temperature range. In industrial settings, the thermal state of the cell is typically characterized and controlled indirectly through two methods: first, calculating the "cell temperature" based on a small number of fixed temperature measuring points or periodic sampling; second, adjusting process parameters such as alumina feed rate, electrolytic cell operating current, and anode-cathode spacing. However, the convective heat dissipation of the gas phase above the electrolytic cell is strongly influenced by the exhaust air volume and the opening of the exhaust hood's partition baffles, and changes rapidly with the induced draft fan speed and pipeline resistance. Simultaneously, an increasing number of production lines are equipped with indirect closed-loop waste heat exchange-preheating loops. By changing the waste heat bypass ratio through the opening of the waste heat bypass valve, the sensible heat of the material entering the cell is affected, as well as the effective temperature field and convective heat transfer conditions on the gas phase side. These two sets of devices (exhaust and waste heat) often operate within independent control loops, primarily focusing on waste gas capture efficiency and equipment protection, without being incorporated into the unified quantitative and coordinated control of the overall thermal state of the electrolytic cell.

[0003] The existing practices have three common problems in engineering:

[0004] The observations are discrete and asynchronous. Data sources for various parameters, such as temperature inside the exhaust hood, exhaust gas flow rate in each zone branch, heat flux on the tank sidewall, exhaust gas dew point temperature, preheater outlet temperature, inlet temperature and flow rate of the purification unit, zone baffle opening, induced draft fan speed, and waste heat bypass valve opening, are diverse, have different sampling rates, and inconsistent timestamps, making it difficult to make a reliable judgment on the heat balance at the same moment.

[0005] There is a lack of a unified temperature state quantity that can be used as a control criterion. The single-point temperature or empirical "tank temperature" index commonly used in industry is affected by spatial non-uniformity, as well as by the instantaneous operating conditions of extraction and discharge and the arrangement of measuring points, making it difficult to serve as a stable criterion within a narrow target temperature range. At the same time, the heat dissipation of the sidewalls varies significantly with time and space, and there is a lack of quantitative indicators of wall heat dissipation that are closely related to the overall thermal state of the tank.

[0006] Constraints and temperature control targets are mutually restrictive. To avoid acidic condensation and equipment corrosion, the inlet temperature of the purification device must not be lower than the sum of the exhaust gas dew point temperature and the dew point safety margin, and the exhaust gas collection efficiency must meet environmental protection boundaries. However, without control logic coupled with thermal conditions, simply increasing the exhaust air volume or arbitrarily changing the waste heat bypass ratio can easily cause passive fluctuations in the complex thermal balance, resulting in conflicts between "temperature control, collection, and anti-condensation." Furthermore, coupled with rapid events such as anode replacement, shell breaking and feeding, and current disturbances, traditional single-variable or decentralized control strategies often exhibit limit exceedances and oscillations.

[0007] Against this backdrop, industrial sites face a core challenge: on the one hand, it is necessary to construct a unified temperature criterion that can represent the overall thermal state of the electrolytic cell under conditions of multi-source measurement noise and spatial non-uniformity; on the other hand, it is necessary to incorporate the exhaust system and the waste heat exchange-preheating loop into the same control framework, while strictly meeting constraints such as the waste gas capture efficiency and the inlet temperature of the purification device not being lower than the waste gas dew point temperature plus the dew point safety margin, to quickly and verifiably adjust the thermal balance of the electrolytic cell.

[0008] In summary, the technical problem actually solved by this invention is how to integrate multi-source observations into a "composite tank temperature" and coordinately adjust the exhaust air volume and waste heat bypass ratio to keep the composite tank temperature stably maintained in the target temperature range, under the premise of meeting the requirements of exhaust gas capture efficiency and the inlet temperature of the purification device not being lower than the exhaust gas dew point temperature + dew point safety margin threshold. Summary of the Invention

[0009] To overcome the above-mentioned technical defects, the purpose of this invention is to provide a precise temperature control method and system for aluminum smelting process. Through the model of "multi-source observation → composite tank temperature" - filtering fusion and the coordinated main control of "exhaust air volume + waste heat bypass ratio" (with limited linkage when necessary), the composite tank temperature is stably maintained within the target temperature range while meeting environmental protection and anti-condensation constraints.

[0010] This invention discloses a method for precise temperature control in the aluminum smelting process, comprising:

[0011] Step S1: Data Acquisition and Time Alignment: Collect and time-align the following data items in the exhaust hood above the electrolyte in the aluminum electrolysis cell, on the outer side wall of the cell, and in the area adjacent to the cathode:

[0012] (1) Temperature data at multiple points inside the enclosure; (2) Waste gas flow data of each zone branch; (3) Heat flux data of the tank sidewall; (4) Waste gas dew point temperature data; (5) Preheater outlet temperature data; (6) Purification device inlet temperature data and inlet flow data; (7) Zone baffle opening data; (8) Exhaust fan speed data; (9) Waste heat bypass valve opening data.

[0013] Step S2 Composite tank temperature estimation: Input the data from (1) to (9) of step S1 into the reduced-order model of thermal-electric coupling of aluminum electrolysis tank, and correct the model state through statistical filtering to obtain the composite tank temperature and wall heat dissipation index.

[0014] Step S3 Target Judgment and Component Identification: When the composite tank temperature is less than or equal to the preset target temperature zone lower limit threshold or greater than or equal to the preset target temperature zone upper limit threshold, it is determined to be a deviation from the target temperature zone, and the temperature difference is decomposed into convection heat dissipation component, material parameter component and electrical parameter component.

[0015] Step S4 Main control quantity adjustment: When a deviation is determined, the controller executes according to the first control sequence: (a) Adjust the exhaust air volume according to the partition baffle opening data and the induced draft fan speed data, with the adjustment range being greater than or equal to the preset exhaust increase threshold; (b) Adjust the waste heat bypass ratio and change the preheater outlet temperature according to the waste heat bypass valve opening data, with the adjustment range being greater than or equal to the preset bypass ratio reduction threshold, so as to change the convective heat dissipation and the sensible heat of the material entering the tank, so that the composite tank temperature returns to the target temperature zone;

[0016] Step S5 Backup linkage and time criterion: If the composite cell temperature does not enter the target temperature zone after the first control sequence has been completed for a period of time greater than or equal to the preset main control evaluation time threshold, then execute the second control sequence: linkage adjustment of alumina feed amount, electrolytic cell working current and anode-cathode distance until the composite cell temperature enters the target temperature zone.

[0017] Preferably, the exhaust hood is divided into multiple independently controlled partitions along the length of the trough, and the number of partitions is greater than or equal to a preset threshold number of partitions. The controller determines the target partition based on the spatial distribution of the composite trough temperature and wall heat dissipation index of each partition, and executes an instruction to adjust the exhaust air volume of the target partition.

[0018] Preferably, the waste heat exchange adopts an indirect closed-loop method, in which the clean circulating working fluid is heated and then used for preheating alumina particles; the controller maintains the temperature at the low-temperature end of the heat exchanger greater than or equal to the sum of the waste gas dew point temperature and the preset dew point safety margin threshold based on the waste gas dew point temperature data, and inputs the preheater outlet temperature into the control algorithm.

[0019] Preferably, the control algorithm satisfies the following constraints throughout the process: the exhaust gas capture efficiency is greater than or equal to the preset lower limit threshold of capture efficiency, and the inlet temperature of the purification device is greater than or equal to the sum of the exhaust gas dew point temperature and the preset dew point safety margin threshold; the controller executes the constraints at the computational level through constrained predictive control or safety filtering with barrier functions.

[0020] Preferably, the calculation of the composite tank temperature adopts a sensitivity-driven adaptive weighting of the multi-point temperature data inside the hood, the exhaust gas flow data of each zone branch, and the heat flux data of the tank sidewall, and uses information gain as the weight update criterion: the contribution of each sensing quantity to the composite tank temperature and the wall heat dissipation index is calculated in real time and the weight is dynamically updated.

[0021] Preferably, for anode replacement, power modulation, or shell breaking and feeding events, the event handling process includes: (a) when the time advance before the event starts is greater than or equal to a preset event advance time threshold, the baseline of the exhaust air volume of the adjacent zone is increased according to the zone baffle opening data and the induced draft fan speed data, and the increase is greater than or equal to a preset exhaust increase threshold; at the same time, the waste heat bypass ratio is reduced according to the waste heat bypass valve opening data, and the reduction is greater than or equal to a preset bypass ratio reduction threshold; (b) when the event ends and the composite tank temperature no longer meets the deviation judgment condition of step S3, and the exhaust gas collection efficiency is greater than or equal to a preset collection efficiency lower limit threshold, and the purification device inlet temperature is greater than or equal to the sum of the exhaust gas dew point temperature and the preset dew point safety margin threshold, the event is restored to the pre-event set value by a linear ramp within a duration greater than or equal to a preset event recovery time threshold.

[0022] Preferably, residual-based sequence verification is used to identify anomalies in the temperature sensor, heat flow sensor, and air volume meter; when an anomaly is detected, a degradation strategy is executed: the weight of the corresponding data item in the weight update is frozen; the rate of change of the exhaust air volume is limited to be less than or equal to a preset rate of change limit threshold per execution cycle; and the absolute value of the offset of the waste heat bypass ratio is less than or equal to a preset bypass ratio offset threshold.

[0023] Preferably, the composite tank temperature is predicted in the short term within a preset prediction time threshold based on the physical-data fusion model; when the prediction shows that it will exceed the target temperature zone after a preset advance triggering time threshold, the controller pre-executes the adjustment amount of the first control sequence based on the partition baffle opening data, induced draft fan speed data and waste heat bypass valve opening data, the pre-execution amplitude is greater than or equal to the preset pre-execution amplitude threshold and the duration is greater than or equal to the preset pre-execution hold time threshold.

[0024] Preferably, the execution cycle of the exhaust air volume and the partition baffle opening is greater than or equal to a preset first execution cycle threshold, and the execution cycle of the waste heat bypass ratio is greater than or equal to a preset second execution cycle threshold; both are respectively set with preset dead zone thresholds and preset reverse hysteresis band thresholds.

[0025] A second objective of this invention is to provide a system for implementing the above-described method, characterized in that it comprises:

[0026] The system includes: exhaust hoods and their electric baffles divided along the length of the tank; variable frequency induced draft fans and main exhaust pipes connected to the exhaust hoods; an indirect closed-loop waste heat exchange-preheating circuit connected in parallel with the main exhaust pipe, including a heat exchanger, a bypass valve, and a material preheater; a data acquisition device for acquiring multiple temperatures inside the hood, exhaust gas flow rates in each zone branch, heat flux through the tank sidewall, exhaust gas dew point temperature, preheater outlet temperature, and purification device inlet temperature and inlet flow rate; a status acquisition device for acquiring the opening degree of the zone baffles, the speed of the induced draft fan, and the opening degree of the waste heat bypass valve; and a data processing and control unit, which executes steps S1 to S5 of the above method and establishes a restricted linkage interface with the feeding actuator, the current regulating device, and the anode-cathode spacing regulating device; wherein, the system uses the exhaust air volume and the waste heat bypass ratio as the main control variables, and controls the composite tank temperature within the target temperature range determined by the preset lower limit threshold and the preset upper limit threshold of the target temperature range, provided that various threshold conditions are met.

[0027] Compared with existing technologies, the above technical solution has the following advantages:

[0028] 1. In existing technologies, the temperature of individual thermocouples, artificial infrared thermometry, or indirect voltage-based temperature estimation at aluminum electrolysis cells is often used as a reference for "cell temperature". These quantities are extremely sensitive to the arrangement of measuring points, instantaneous operating conditions of exhaust air volume, and local disturbances; readings at different locations at the same time vary greatly, and sampling from different channels is not synchronized, making it difficult to form a unified temperature quantity that can be directly used for closed-loop threshold discrimination. In this invention, step S1 synchronously collects and aligns the nine types of process quantities over time; step S2 inputs these quantities into a reduced-order model of the thermo-electric coupling of the aluminum electrolysis cell, corrects the state using statistical filtering, and outputs the composite cell temperature (unified temperature criterion) and wall heat dissipation index online. Sensitivity-driven adaptive weighting is applied to the temperature data at multiple points inside the enclosure, the exhaust gas flow data of each zone branch, and the heat flux data of the tank sidewall. The weights are updated in real time with information gain, thereby integrating discrete and multi-source observations into a temperature state quantity with the same physical meaning. This weakens the influence of single-point bias and instantaneous fluctuations in the pumping and exhausting processes, and provides a closed-loop criterion that can be directly compared with the lower and upper thresholds of the target temperature zone.

[0029] 2. In existing technologies, multiple points of temperature within the enclosure, zoned flow rate, sidewall heat flux, induced draft fan speed, and waste heat bypass valve opening are all assigned to different subsystems (process instrumentation, dust removal, frequency conversion, preheating circuit), and their timestamps, sampling periods, and units are inconsistent. In engineering practice, this often relies on manual comparison or post-event statistics, leading to "out-of-time data" entering decision-making. In this invention, by using the data processing and control unit as a unified time source, setting unified rules for common sampling periods and units, and adding quality markers (valid, disconnected, out of range, interpolated) to each data point, all inputs enter the model and control algorithm with a unified time base, avoiding the superposition of out-of-time errors. The quality markers also provide a basis for subsequent statistical filtering and anomaly degradation.

[0030] 3. In existing technologies, exhaust systems primarily focus on waste gas capture efficiency, while waste heat circuits mainly prioritize equipment protection (anti-condensation). Temperature control is independent of both. When conflicts arise, compromises are often made based on experience, easily leading to mutual constraints between temperature control, waste gas capture, and anti-condensation. In this invention, a hard constraint is set at the computational level, combining the lower limit threshold of waste gas capture efficiency with the sum of the purification device inlet temperature ≥ waste gas dew point temperature + dew point safety margin threshold. Constrained predictive control or safety filtering with a barrier function ensures a solution within the feasible region during command generation, achieving temperature control while satisfying environmental protection and anti-condensation boundaries, eliminating the need for manual compromises. The constraint-integrated algorithm reduces the risk of exceeding limits.

[0031] 4. In existing technologies, corrections are often made using slow-response quantities such as alumina feed rate, electrolytic cell operating current, and anode-cathode spacing. However, there is a lack of direct and rapid closed-loop control over convective heat dissipation and the sensible heat of the material entering the tank, resulting in slow response and a tendency to overshoot. In this invention, the exhaust air volume and waste heat bypass ratio are established as the main control quantities. When the composite tank temperature deviates from the target temperature range, the first control priority is given: adjusting the exhaust air volume based on the partition baffle opening data and induced draft fan speed data (amplitude ≥ preset exhaust increase threshold); adjusting the waste heat bypass ratio based on the waste heat bypass valve opening data (amplitude ≥ preset bypass ratio decrease threshold when decreasing), and linking the preheater outlet temperature data to change the sensible heat of the material entering the tank and the effective temperature field of the gas phase; and using execution cycle threshold, dead zone threshold, and reverse hysteresis band threshold to suppress jitter, directly affecting the two main factors (convective heat dissipation and sensible heat of the material entering the tank), resulting in fast response, strong controllability, and reduced reliance on slow channels.

[0032] 5. In the prior art, the exhaust air volume is often increased as a whole to deal with local overheating, which is energy-intensive and can easily "absorb cold" into other zones. In this invention, the target zone is identified by using the spatial distribution of wall heat dissipation index and temperature data at multiple points inside the enclosure. The change in synthetic air volume is weighted and allocated to the target zone for priority execution, so that the execution is directed to the source of the problem, reducing the increase in the system's synthetic air volume and reducing secondary disturbances and energy consumption.

[0033] 6. In existing technologies, indirect closed-loop waste heat exchange-preheating loops are mostly independent of temperature control, the preheater outlet temperature is not included in the tank temperature criterion, and bypass regulation only considers heat exchanger protection. In this invention, by incorporating the preheater outlet temperature data into the algorithm, the waste heat bypass ratio is taken as one of the main control variables. The low-temperature end of the heat exchanger is constrained by the sum of the exhaust gas dew point temperature and the dew point safety margin threshold, so as to achieve stable recovery of waste heat within the safety boundary and controllably increase the sensible heat of the material entering the tank, reduce thermal shock, and improve the adjustability of thermal balance.

[0034] 7. In the prior art, fixed or empirical weights are not sensitive to changes in operating conditions, sensor noise and missing data, and the estimation results are prone to drift. In this invention, sensitivity-driven adaptive weighting is implemented for key observations, and the weights are updated with information gain. Combined with quality marking, abnormal channels are automatically downweighted or eliminated, so as to maintain the stability of composite tank temperature estimation when there is measurement noise or single-point anomaly, and avoid the chain amplification of "measurement anomaly → control anomaly".

[0035] 8. In the prior art, events such as anode replacement, power modulation, and shell breaking and feeding often cause disturbances before they are rectified, resulting in amplified fluctuations and slow recovery. In this invention, by setting event advance time thresholds and event recovery time thresholds, the baseline of the target zone exhaust air volume is increased and the waste heat bypass ratio is reduced before the event. After the event, the recovery is carried out according to a linear ramp, thereby suppressing the disturbances in advance, smoothing the recovery, reducing secondary oscillations, and shortening the recovery time.

[0036] 9. In existing technologies, sensor drift or malfunction often only triggers an alarm, while the control system continues to operate based on erroneous data, causing actuator jitter or out-of-bounds operation. In this invention, anomalies (such as three times the standard deviation) are identified by using a residual-based sequence check. Upon triggering, the following steps are taken: the weight of the channel in the weight update is frozen; a preset rate-of-change threshold is set for the rate of change of the exhaust air volume; and a preset bypass ratio offset threshold is set for the waste heat bypass ratio offset. This prevents anomalies from overflowing into large-scale erroneous execution, maintains closed-loop stability and equipment safety, and buys time for on-site maintenance.

[0037] 10. In existing technologies, the main approach is still to correct deviations after exceeding the limit, often involving "exceeding the limit first and then catching up." In this invention, short-term prediction is made within a preset prediction time domain threshold. When it is expected that the limit will be exceeded after a preset advance triggering time threshold, the first control sequence action is executed in advance according to a preset pre-execution amplitude threshold and a preset pre-execution hold time threshold. This reduces the probability of exceeding the limit, narrows the extent of the limit exceedance, and improves dynamic quality. Often, it can return to the limit without triggering backup linkage.

[0038] 11. In the prior art, high-frequency fine-tuning leads to actuator wear, energy consumption and system oscillation; in this invention, by setting a first execution cycle threshold for the exhaust air volume and the opening of the partition baffle, setting a second execution cycle threshold for the waste heat bypass ratio, and configuring a dead zone threshold and a reverse hysteresis band threshold, meaningless small movements are filtered out, the actuator life is extended, energy consumption is reduced and smoothness is improved.

[0039] 12. In the prior art, when the fast master control is insufficient to return to the zone, manual adjustment of the feed amount, current or anode-cathode spacing is often required, resulting in poor repeatability. In this invention, if the system still fails to return to the zone after reaching the preset master control evaluation time threshold, the second control sequence is triggered. The alumina feed amount, electrolytic cell operating current and anode-cathode spacing are optimized under constraints or solved by constrained least squares, and the system converges in a rolling manner. This provides a deterministic correction channel within the constrained feasible region, ensuring that the composite cell temperature automatically exits the standby linkage after entering the target temperature zone.

[0040] 13. In the existing technology, there is a lack of indicators directly related to the heat loss of the sidewall, making it difficult to locate "thermal deviation" and the adjustment is mostly based on experience. In this invention, by outputting the heat dissipation index of the wall surface and decomposing the temperature deviation of the composite tank into convective heat dissipation components, material parameter components and electrical parameter components, the control decision has physical interpretability, which makes it easier to accurately identify the target zone and optimize the weight parameters, thereby improving the adjustment efficiency.

[0041] 14. In the prior art, some solutions rely on highly customized equipment or closed-source systems, which are difficult to expand and have high maintenance costs. In this invention, the data processing and control unit adopts an industrial control computer or a programmable logic controller, and the sensing and execution use general industrial components. The threshold and parameters can be configured according to different slot types and scales, so as to achieve small modification, scalability, easy maintenance, and easy promotion in existing production lines. Attached Figure Description

[0042] Figure 1 This is a schematic diagram of the steps of a precise temperature control method and system for aluminum smelting process according to the present invention. Detailed Implementation

[0043] The advantages of the present invention will be further illustrated below with reference to the accompanying drawings and specific embodiments.

[0044] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numerals in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure as detailed in the appended claims.

[0045] The terminology used in this disclosure is for the purpose of describing particular embodiments only and is not intended to be limiting of the disclosure. The singular forms “a,” “the,” and “the” as used in this disclosure and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any and all possible combinations of one or more of the associated listed items.

[0046] It should be understood that although the terms first, second, third, etc., may be used in this disclosure to describe various information, such information should not be limited to these terms. These terms are used only to distinguish information of the same type from one another. For example, without departing from the scope of this disclosure, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to determination."

[0047] In the description of this invention, it should be understood that the terms "longitudinal", "lateral", "up", "down", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0048] In the description of this invention, unless otherwise specified and limited, it should be noted that the terms "installation", "connection" and "linking" should be interpreted broadly. For example, they can refer to mechanical or electrical connections, or internal connections between two components. They can be direct connections or indirect connections through an intermediate medium. Those skilled in the art can understand the specific meaning of the above terms according to the specific circumstances.

[0049] In the following description, suffixes such as "module," "part," or "unit" used to denote elements are used only for the convenience of the description of the invention and have no specific meaning in themselves. Therefore, "module" and "part" can be used interchangeably.

[0050] See Figure 1As shown, this embodiment will describe in detail a precise temperature control method for aluminum smelting process, including: Step S1 Data acquisition and time alignment: In the exhaust hood above the electrolyte of the aluminum electrolysis cell, outside the side wall of the cell and near the cathode, the following data items are collected and time aligned respectively: (1) temperature data at multiple points inside the hood, (2) exhaust gas flow data of each zone branch, (3) heat flux data of the side wall of the cell, (4) exhaust gas dew point temperature data, (5) preheater outlet temperature data, (6) inlet temperature data and inlet flow data of the purification device, (7) zone baffle opening data, (8) induced draft fan speed data, (9) waste heat bypass valve opening data; Step S2 Composite cell temperature estimation: The data from (1) to (9) of step S1 are input into the reduced-order model of the thermo-electric coupling of the aluminum electrolysis cell, and the model state is corrected by statistical filtering to obtain the composite cell temperature and wall heat dissipation index; Step S3 Target determination and component identification: When the composite cell temperature is less than or equal to the preset target temperature zone When the threshold is greater than or equal to the preset upper limit of the target temperature zone, it is determined to be a deviation from the target temperature zone, and the temperature difference is decomposed into convective heat dissipation component, material parameter component and electrical parameter component; Step S4 Main control quantity adjustment: When a deviation is determined, the controller executes according to the first control sequence: (a) Adjust the exhaust air volume according to the partition baffle opening data and the induced draft fan speed data, and the adjustment range is greater than or equal to the preset exhaust increase threshold; (b) Adjust the waste heat bypass ratio and change the preheater outlet temperature according to the waste heat bypass valve opening data, and the adjustment range is greater than or equal to the preset bypass ratio reduction threshold, so as to change the convective heat dissipation and the sensible heat of the material entering the tank, so that the composite tank temperature returns to the target temperature zone; Step S5 Backup linkage and time criterion: If the composite tank temperature still does not enter the target temperature zone after the completion of the first control sequence and the time is greater than or equal to the preset main control evaluation time threshold, the second control sequence is executed: Linkage adjustment of alumina feed amount, electrolytic cell working current and anode-cathode distance until the composite tank temperature enters the target temperature zone.

[0051] This embodiment will describe step S1 in detail. Step S1, data acquisition and time alignment, is used to provide spatiotemporally consistent, dimensionally unified, and quality-traceable input data for the subsequent composite bath temperature estimation and control algorithm. A data processing and control unit is configured at the system level. This unit can be an industrial control computer or a programmable logic controller, equipped with analog quantity acquisition modules, differential pressure acquisition modules, and digital communication modules, and serves as the clock source for the entire system. The raw data output by each sensor or transmitter is timestamped uniformly on the data processing and control unit side; time synchronization can use a network time protocol or a precision time protocol to ensure that the time synchronization deviation of different acquisition channels is within the engineering allowable range. To facilitate subsequent calculations, a common equally spaced time axis is set as a common sampling period (e.g., one second); channels with a frequency higher than the common sampling frequency are first resampled after anti-aliasing low-pass filtering, while channels with a frequency lower than the common sampling frequency are mapped to the common time axis using hold or linear interpolation, and an "interpolation mark" is retained in the data record. Each data entry carries a quality mark field, which includes at least an over-range flag, a disconnection or communication interruption flag, sensor self-test status, and the most recent calibration time. All temperature quantities are expressed in Celsius, flow rates are expressed in volumetric flow rates under uniform metering conditions or standard volumetric flow rates, heat flux is expressed in watts per square meter, opening degree is expressed as a percentage, and rotation speed is expressed in revolutions per minute; units are standardized before entering the time alignment process.

[0052] First, multiple temperature data points are collected inside the hood. Multiple temperature measuring points are arranged along the length of the aluminum electrolysis cell inside the exhaust hood, with each measuring point at a consistent height from the free surface of the electrolyte to reflect the spatial distribution of the gas phase temperature above the electrolyte. K-type thermocouple temperature sensors can be used, with an external corundum protective tube to adapt to the high-temperature flue gas environment containing fluorine and dust; alternatively, platinum resistance temperature sensors with a corrosion-resistant protective sleeve can be used. The temperature sensors are connected to the isolated thermocouple acquisition module via a high-temperature shielded cable, with the shielding layer grounded at one end. Two- or three-point temperature calibration is performed in the constant temperature bath before installation, and on-site comparison is conducted according to the maintenance plan after commissioning. This data reflects the intensity of convective heat transfer and the local extraction status, and is one of the important inputs for estimating the composite cell temperature.

[0053] Secondly, waste gas flow data for each zone branch is collected. A flow measurement device is installed on the straight pipe section of each zone branch, with the straight pipe section length meeting the instrument's straight pipe section requirements. A Venturi flow meter or a Pitot-static differential pressure flow meter (equipped with a high-temperature and corrosion-resistant pressure tap) can be used, or a high-temperature thermal mass flow meter can be used depending on the conditions. The differential pressure or flow signal is input to the data processing and control unit via an isolated analog quantity acquisition module, with temperature and pressure compensation measuring points added if necessary. Calibration is performed according to metrological regulations, and verification can be performed through multi-point velocity measurement across the entire cross-section. This data directly represents the actual exhaust air volume of each zone branch, used for exhaust capacity calculation and subsequent constraint determination.

[0054] Next, collect heat flux data from the sidewall of the aluminum electrolysis cell. Several measuring points are arranged along the height and length directions on the outer surface of the sidewall. Thin-film heat flux sensors (such as Garden type heat flux sensors) can be used to directly measure the heat flux; alternatively, two or three embedded thermocouple temperature sensors can be buried along the thickness direction. The heat flux is calculated using Fourier's law by combining the measured temperature gradient with the material's thermal conductivity and geometric parameters. Good thermal contact must be ensured during sensor attachment or burial, and mechanical fixation and heat-resistant sealing must be performed. The signal enters the isolated acquisition module. Thin-film heat flux sensors are calibrated according to the manufacturer's sensitivity coefficient. For the temperature gradient method, the material's thermal conductivity, thickness, and burial location must be recorded. This data reflects the heat loss per unit area of ​​the sidewall, providing a direct or indirect source for the wall's heat dissipation indicators.

[0055] Subsequently, exhaust gas dew point temperature data are collected. A sampling probe and a heated sampling pipeline are installed between the main exhaust duct and the inlet of the purification unit to avoid condensation and compositional changes during sampling. The instrument can be a cold mirror dew point meter or a laser absorption spectroscopy dew point measurement device, equipped with corrosion-resistant sampling components and a dust filtration unit. The sampling probe is fixed to the pipe wall, and the sampling pipeline is heated throughout. The signal is connected to the data processing and control unit via digital communication or analog signal transmission. The dew point meter is calibrated at multiple points according to regulations and periodically verified. This data provides the dew point temperature of the exhaust gas system, which is the direct basis for determining the temperature limits at the low-temperature end of the heat exchanger and the inlet temperature limits of the purification unit.

[0056] Synchronously collect preheater outlet temperature data. A temperature measuring point is set at the preheater outlet; if monitoring the gas side, a sheathed temperature probe is placed in the central area of ​​the pipeline; if monitoring the material side, a wear-resistant sheathed temperature probe is installed in the material outlet channel. K-type thermocouple temperature sensors or platinum resistance temperature sensors can be used, with the range and accuracy selected according to the temperature zone. The signal is connected to the data processing and control unit via an isolated acquisition module; if the measurement is on the material side, heat transfer hysteresis compensation parameters can be provided in the data record. This data characterizes the current thermal state of the waste heat exchange-preheating loop and can be used as a feedforward input or limiting variable in subsequent operations.

[0057] For the inlet temperature and flow rate data of the purification unit, temperature measuring points and flow measurement devices are installed in the inlet pipe section of the purification unit. Temperature measurement can use a type K thermocouple temperature sensor or a platinum resistance temperature sensor; flow measurement can use a Venturi flow meter, a Pitot-static differential pressure flow meter, or a high-temperature thermal mass flow meter. During installation, ensure that the straight pipe section, pressure tap location, and flow rectification components meet the instrument requirements, and that the signal and exhaust gas dew point temperature data are collected and recorded synchronously. These two sets of data, combined with the exhaust gas dew point temperature data, are used to determine the safety margin of the purification unit's inlet temperature and its compatibility with the capture conditions.

[0058] To collect data on the opening degree of the zone baffles, a position feedback device, such as a potentiometer-type position sensor, a magnetic encoder-type position sensor, or an opening feedback device with a valve positioner, needs to be installed on the actuator of each zone baffle. The device is coaxially mounted with the actuator, and mechanical clearance is calibrated. The output is sent to the data processing and control unit in percentage form. Zero-position and full-stroke calibrations are performed, and the full-stroke characteristics are recorded. This data reflects the geometric channel status of each zone branch and is a direct state quantity for subsequent estimation and adjustment of exhaust air volume.

[0059] The induced draft fan speed data can be obtained from a magnetoelectric speed sensor at the fan shaft end, or it can be directly acquired from the speed feedback signal provided by the frequency converter. The speed signal is converted by the frequency converter control cabinet and then input to the data processing and control unit. Before commissioning, a handheld tachometer is used for comparison and verification to form a comparison coefficient. This data, together with the zone baffle opening data, reflects the exhaust capacity setting and dynamic response of the exhaust system.

[0060] Finally, the opening degree data of the waste heat bypass valve is collected. A position feedback device, such as a valve positioner opening feedback or an encoder-type position sensor, is installed on the actuator of the waste heat bypass valve to ensure that the hysteresis of the mechanical structure connected to the valve stem is within the specified range. The opening degree is entered into the data processing and control unit in the form of a percentage, and three-point calibrations are completed at zero point, half stroke, and full stroke. This data represents the distribution state of the flue gas entering the heat exchanger and bypassing it, and is a key state quantity for determining the operating point of the waste heat exchange-preheating loop.

[0061] To achieve "time alignment," the data processing and control unit maps the nine types of raw data to a common sampling period using a unified timestamp. Asynchronous samples are processed using forward hold or linear interpolation, with interpolation markers retained in the records. Samples exhibiting disconnection, over-range errors, or self-test failures are marked as invalid at the specified timestamp, while their previous values ​​are retained and an event record is generated; no extrapolation is performed. After time alignment, unit and range checks are performed; samples failing the check are also marked as invalid. Raw data, aligned data, and quality markers are archived hierarchically for use in subsequent steps.

[0062] This embodiment will explain step S2 in detail. In this embodiment of the invention, the composite tank temperature estimation in step S2 is performed by the data processing and control unit. The purpose of this step is to perform state estimation and statistical filtering correction using a reduced-order model of the thermal-electric coupling of the aluminum electrolysis tank, based on the nine types of time-series data obtained from data acquisition and time alignment in step S1, thereby obtaining two types of process quantities in real time: the composite tank temperature and the corresponding wall heat dissipation index. The data processing and control unit can be an industrial control computer or a programmable logic controller, and its built-in computing tasks include model solving, statistical filtering, sensor quality label parsing, and data input / output stack management.

[0063] The so-called reduced-order model of thermal-electric coupling in aluminum electrolysis cells refers to a control-oriented model with a small number of state variables and parameters suitable for online calculation, which lumps together and equivalently represents the components such as the electrolyte body, anode assembly, cathode region, and upper gas phase heat exchange region at the energy and electro-thermal interaction level. The energy balance term of the model includes at least the internal energy change term of the electrolyte body, the convective heat dissipation term caused by the exhaust air volume, the conductive heat dissipation term caused by the heat flux of the tank sidewall, and the radiative heat dissipation term caused by radiative heat transfer. The "electro-thermal coupling" term of the model adopts lumped equivalence, and the Joule heating caused by the current and the electrode reaction heat are uniformly regarded as the disturbance power term to be estimated in this step, which is estimated and corrected online by the statistical filtering process, thereby avoiding the introduction of new measurement quantities other than those in step S1. The boundary conditions and driving quantities of the model are all derived from the nine types of data in step S1: multi-point temperature data inside the hood are used to define the boundary temperature and convective heat transfer intensity of the gas phase heat exchange zone; exhaust gas flow data of each zone branch, zone baffle opening data, and induced draft fan speed data are used together to characterize the instantaneous extraction capacity of the extraction system and drive the convective heat dissipation term; heat flux data of the tank sidewall is directly entered into the calculation of the wall heat dissipation index and participates in the state correction; exhaust gas dew point temperature data is used to determine the flue gas properties and low-temperature end limitations, and correct the gas phase side specific heat and moisture content related parameters; preheater outlet temperature data and waste heat bypass valve opening data jointly characterize the distribution state and heat transfer strength of the waste heat transfer-preheating loop, and correct the effective gas phase temperature and convective heat transfer coefficient in the model; purification device inlet temperature data and inlet flow data are used to perform mass conservation and energy conservation closure checks at the system level, and are used as auxiliary observations in statistical filtering.

[0064] To ensure online solvability and real-time performance, the reduced-order model employs lumped equivalence for spatial distribution: the electrolyte bulk is represented by single- or double-capacity heat capacity equivalence, the adjacent regions of the anode and cathode are represented by additional heat capacity and thermal resistance equivalence, and the upper gas phase region is described by equivalent convective heat transfer coefficients and equivalent radiative heat transfer coefficients. Equivalent parameters are determined through trial operation calibration before the device is put into operation and are slowly adaptively corrected during operation using statistical filtering. The model uses the unified time axis after time alignment in step S1 as the integration step size; each calculation cycle includes one model time advance and one observation correction. During the time advance phase, the model uses the state of the previous cycle and the driving quantities of the current cycle (including exhaust gas flow data of each zone branch, multi-point temperature data inside the hood, waste heat bypass valve opening data, induced draft fan speed data, etc.) to calculate the uncorrected predicted state and the corresponding predicted composite tank temperature and predicted wall heat dissipation index. During the observation and calibration phase, the model reads observational data such as heat flux data of the tank sidewall, temperature data of multiple points inside the hood, and inlet temperature data of the purification device. Combined with sensor quality marks, the model performs statistical filtering to update the measurements and obtains the calibrated composite tank temperature and wall heat dissipation index.

[0065] The statistical filtering in this step is a probabilistic statistical method based on state space, which can be implemented using extended Kalman filtering, unscented Kalman filtering, or an equivalent Bayesian filtering method. The process noise covariance and observation noise covariance of the statistical filtering are initialized according to the sensor's technical specifications and calibration results, and adaptively scaled during operation based on quality markers and residual statistics. When an observation channel experiences out-of-range, disconnection, or anomalies accompanied by large residuals, the data processing and control unit reduces the weight of that channel in the measurement update or temporarily removes the observation based on the quality marker, ensuring that the state correction remains solvable. Statistical testing of the residuals can use the three-standard-deviation criterion or an equivalent robustness criterion to identify instantaneous outliers. Through the closed-loop iteration of the above time advancement and measurement update, the estimated disturbance power term in the reduced-order model converges online, thereby ensuring the temporal and physical consistency of the composite tank temperature and wall heat dissipation indicators.

[0066] In terms of output definition, the composite cell temperature is a single scalar state quantity derived from the data processing and control unit, representing the equivalent temperature of the overall thermal state of the aluminum electrolysis cell. Its calculation is not a simple geometric or arithmetic average, but rather the result of statistical filtering of the main temperature state estimated by the model based on energy conservation and heat transfer mechanisms, using observational corrections based on multi-point temperature data within the enclosure and heat flux data from the cell sidewalls. This quantity maintains comparability under different operating conditions and sensor availability, and is used for direct comparison with the preset lower and upper threshold values ​​of the target temperature zone. The wall heat dissipation index is a process quantity of sidewall heat loss, defined as the area-weighted average heat flux of the region where the sidewall measuring points are located, or the equivalent total heat dissipation power density per unit length. The specific form used is uniformly determined during equipment commissioning and maintained consistent within the data processing and control unit. Cell sidewall heat flux data is directly incorporated into the calculation of this index; when some measuring points are unavailable, statistical filtering performs bounded interpolation on the index based on spatial correlation and historical data, retaining validity markers in the results.

[0067] To ensure feasibility, the data processing and control unit must complete parameter initialization and consistency checks before executing step S2. Parameter initialization includes setting the initial covariance of equivalent heat capacity, equivalent thermal resistance, equivalent convective heat transfer coefficient, equivalent radiative heat transfer coefficient, observation noise and process noise, and the initial residual threshold. Consistency checks include unit checks, sensor channel mapping checks, and timestamp integrity checks. During operation, the data processing and control unit will archive uncorrected predicted values, statistically filtered corrected estimated values, and quality markers hierarchically for direct retrieval in step S3 for target determination and component identification.

[0068] In this embodiment, step S3 will be described in detail. In the implementation of the present invention, the target determination and component identification in step S3 is performed by the data processing and control unit in each common sampling period. This step takes the composite tank temperature and wall heat dissipation index output in step S2 as the core input, and combines the exhaust gas flow data of each zone branch, multi-point temperature data inside the hood, heat flux data of the tank sidewall, exhaust gas dew point temperature data, preheater outlet temperature data, purification device inlet temperature and inlet flow data, zone baffle opening data, induced draft fan speed data, and waste heat bypass valve opening data obtained by data acquisition and time alignment in step S1 to complete the objective determination of whether the current thermal state deviates from the target temperature zone, and to calculate the decomposition of the temperature difference into convective heat dissipation component, material parameter component, and electrical parameter component.

[0069] First, the target temperature zone is determined. The data processing and control unit reads the preset lower and upper threshold values ​​of the target temperature zone from the parameter library; these two values ​​together define the allowable range of the target temperature zone. The current composite bath temperature is compared with these two threshold values: if the composite bath temperature is less than or equal to the preset lower threshold value, or greater than or equal to the preset upper threshold value, it is determined to deviate from the target temperature zone. To avoid instability in the determination caused by instantaneous disturbances at a single sampling point, the data processing and control unit can set a joint criterion of determination window period and outlier labeling during implementation: within a finite time window containing several common sampling periods, the determination result is confirmed according to the majority principle; at the same time, data samples with invalid quality labels are not included in the determination.

[0070] After confirming the deviation, the composite bath temperature deviation is defined. When the composite bath temperature is less than or equal to the preset lower threshold of the target temperature zone, the composite bath temperature deviation is the difference between the composite bath temperature and the preset lower threshold of the target temperature zone; when the composite bath temperature is greater than or equal to the preset upper threshold of the target temperature zone, the composite bath temperature deviation is the difference between the composite bath temperature and the preset upper threshold of the target temperature zone. The sign and magnitude of the composite bath temperature deviation are used for subsequent component calculations and direction determination.

[0071] Subsequently, component identification is performed, decomposing the composite tank temperature deviation into three types of process components with clear physical meanings: convective heat dissipation components, material parameter components, and electrical parameter components. The decomposition process is based on the local linearization information of the reduced-order model of the aluminum electrolysis cell thermo-electric coupling in step S2 at the current operating point. The data processing and control unit fixes the model state and equivalent parameters at the current moment and calculates the sensitivity coefficients of the composite tank temperature to the following set of driving variables: convective heat transfer related exhaust airflow, equivalent convective heat transfer coefficient, and effective gas phase temperature proxy (determined jointly by exhaust gas flow data from each zone branch, zone baffle opening data, induced draft fan speed data, and multi-point temperature data within the hood); material and inlet physical quantities related to alumina feed rate and preheater outlet temperature (determined by preheater outlet temperature data and material feed record signals); and electrical parameters related to the electrolysis cell operating current and anode-cathode spacing (determined by the status acquisition signals of the current regulating device and the anode-cathode spacing regulating device). Under this local linearization approximation, a small change in the composite tank temperature can be expressed as a linear combination of the changes in the above driving variables and their corresponding sensitivity coefficients.

[0072] To obtain the three components, the data processing and control unit divides the aforementioned driving variables into three groups based on their physical meaning: a convection heat dissipation group, a material parameter group, and an electrical parameter group. Within each group, the sensitivity coefficient of the driving variable within that group is weighted and summed with its change relative to the approximate steady state at the current moment to obtain the equivalent contribution of that group to the composite tank temperature. The weight values ​​are derived from two types of information: one is the observation noise and state uncertainty output from the statistical filtering in step S2, used to reflect the reliability of different channels; the other is the quality label from step S1, used to reduce or remove the participation of the corresponding channel in the presence of disconnection, over-range, or communication interruption. Through this weighted synthesis, the values ​​and signs of the convection heat dissipation component, material parameter component, and electrical parameter component are obtained respectively. If it is necessary to improve the numerical stability, the data processing and control unit can use a least squares solution method with non-negative constraints or regularization terms, so that the algebraic sum of the three components statistically approximates the composite tank temperature deviation, while avoiding excessive amplification of the uncertainty of a single channel.

[0073] Regarding the correlation with the wall heat dissipation index, the data processing and control unit uses the heat flux data of the tank sidewall and the wall heat dissipation index to verify the consistency of the direction and magnitude of the convective heat dissipation component. Specifically, when the wall heat dissipation index shows an increasing trend relative to its short-term smooth reference, and the exhaust gas flow data of each zone branch and the zone baffle opening data show enhanced extraction, the sign of the convective heat dissipation component is consistent with this trend; when the wall heat dissipation index is stable but the composite tank temperature deviation is significant, component compensation is preferentially completed through the sensitivity contribution of the material parameter group and the electrical parameter group. This consistency check is based on the statistical filtering residual of step S2, and triggers data quality backtracking and channel weight reduction processing when contradictions occur.

[0074] After component identification is completed, the data processing and control unit outputs four distinct calculation results for step S4: first, the deviation flag and deviation direction; second, the absolute value of the composite tank temperature deviation; third, the values ​​of the convective heat dissipation component, material parameter component, and electrical parameter component; and fourth, the relative proportions of the three components and their corresponding confidence levels. All outputs are timestamped and quality-marked, and are archived in the data record along with the composite tank temperature and wall heat dissipation indicators at the same time. It should be noted that step S3 does not introduce new measurement channels; all inputs come from steps S1 and S2. The sensitivity coefficient required for component identification is calculated by a reduced-order model of the aluminum electrolysis tank's thermal-electric coupling under the current operating conditions or by updating parameters through statistical filtering, ensuring consistency and traceability.

[0075] Through the above continuous process, while objectively completing the target temperature zone deviation judgment, step S3 decomposes the composite tank temperature deviation into convection heat dissipation component, material parameter component and electrical parameter component on the same time scale based on model sensitivity and statistical weight, providing a direct and calculable basis for the main control quantity adjustment in step S4.

[0076] In this embodiment, step S4 will be described in detail. In this embodiment of the invention, the adjustment of the main control quantity in step S4 is executed by the data processing and control unit in each common sampling period, provided that step S3 has determined that the temperature has "deviated from the target temperature zone". The data processing and control unit reads the deviation flag, composite tank temperature deviation, convection heat dissipation component, material parameter component, and electrical parameter component output by step S3, and retrieves the preset extraction increase threshold and the preset bypass ratio decrease threshold from the parameter library. According to the "first control sequence", it calculates and issues two types of adjustment commands for the main control quantity: one is the adjustment command for the extraction air volume (achieved by the opening of the partition baffle and the speed of the induced draft fan), and the other is the adjustment command for the waste heat bypass ratio (achieved by the opening of the waste heat bypass valve). This step does not introduce new measurement channels; all input quantities come from the outputs of steps S1, S2, and S3.

[0077] First, the data processing and control unit establishes a mapping relationship between the deviation and the control quantity. To ensure consistency with the component identification in step S3, the data processing and control unit combines the absolute value of the composite tank temperature deviation with the proportion of the convective heat dissipation component to obtain the "calculated value" of the target change in the exhaust air volume. Similarly, it combines the absolute value of the composite tank temperature deviation with the proxy quantity related to gas phase heat transfer (the effective gas phase temperature proxy corresponding to the multi-point temperature data inside the shroud and the waste heat bypass valve opening data) to obtain the "calculated value" of the target change in the waste heat bypass ratio. When the deviation direction requires increasing or decreasing a certain control quantity, the data processing and control unit provides a directional target change quantity accordingly. If the magnitude of the calculated target change quantity is less than the preset exhaust increase threshold or the preset bypass ratio decrease threshold, it is increased to the corresponding threshold magnitude to satisfy "the adjustment magnitude is greater than or equal to the preset threshold". The above increase only changes the execution magnitude of the current cycle and does not change the direction determination of the control quantity.

[0078] For the adjustment command of the exhaust air volume, the data processing and control unit first performs spatial allocation, and then decomposes the actuator. The spatial allocation stage distributes the target change in exhaust air volume to each zone branch based on the spatial distribution of heat flux data and heat dissipation indicators of the tank sidewall. The allocation weight is related to the heat flux level of each zone, the temperature data at multiple points inside the hood, and the exhaust gas flow data of each zone branch, with the weights summed to one. The actuator decomposition stage maps the target change in exhaust air volume for each zone to two types of commands: one is the target change in the zone damper opening, and the other is the target change in the induced draft fan speed. The target change in the zone damper opening is calculated based on the calibrated air volume-opening curve of the zone branch, and the upper and lower limits of the opening and the minimum resolvable step size are checked within the data processing and control unit; the target change in the induced draft fan speed is calculated based on the system total resistance characteristics and the fan characteristic curve, and is issued through the frequency converter control interface. To avoid mutual cancellation, the data processing and control unit adopts a "district baffle first, then induced draft fan" sequence: first, the target change in the baffle opening of each target zone is executed, and then the target change in the induced draft fan speed is executed based on the synthesized remaining airflow deviation. The entire process is repeated within each common sampling period, but the command issuance follows a fixed execution cycle and change rate constraint; when a zone baffle has reached its mechanical limit or the minimum resolvable step size is insufficient to achieve the target change, the unfinished portion is automatically transferred to the induced draft fan speed target change for processing, and the saturation state of that zone is marked in the data record.

[0079] For the waste heat bypass ratio adjustment command, the data processing and control unit converts the target change in waste heat bypass ratio into the target change in waste heat bypass valve opening based on the allocation relationship between the bypass loop and the heat exchanger loop. The static mapping used for the conversion is given by the "opening-allocation ratio" calibration curve identified during the unit's trial operation. When there is significant hysteresis or nonlinear segment, the data processing and control unit introduces piecewise mapping and hysteresis compensation coefficients in the calculation of the target change in opening. The target change in opening is sent to the valve positioner after upper and lower limit checks and minimum resolvable step size checks. When the calculation result of this cycle is a change in the direction of reducing the waste heat bypass ratio, the data processing and control unit ensures that its amplitude is not less than the preset bypass ratio reduction threshold. When the calculation result is not in the direction of reduction, but the algorithm still needs fine-tuning, it is executed according to the minimum resolvable step size allowed by the equipment, and the fact that the preset bypass ratio reduction threshold is not applied is recorded in the log.

[0080] To ensure numerical stability and traceability, the data processing and control unit executes a unified inspection process during the command generation of two types of main control quantities. The first is a quality flag check: when a critical channel involved in mapping (e.g., partition baffle opening data, induced draft fan speed data, waste heat bypass valve opening data) shows a disconnection, over-range, or communication failure flag, the weight of that channel is reduced or the most recently valid data is used; if the execution conditions cannot be met, the current cycle is skipped. The second is a unit and range check: all intermediate and target quantities are checked for unit consistency before issuance, and the target quantity is pruned to its upper and lower limits; the pruned difference is re-evaluated in the next common sampling cycle. The third is execution status readback: the partition baffle opening, induced draft fan speed, and waste heat bypass valve opening are read back in real time, the deviation between the actual executed quantity and the target quantity is calculated, and this deviation is fed back to the command calculation for the next cycle. The above inspection process retains timestamps, target quantities, actual quantities, and quality flags in the data records for subsequent steps and post-event traceability.

[0081] Regarding the connection with step S3, at the end of this step, the data processing and control unit archives the target changes in exhaust air volume, the target changes in the opening degree of each zone damper, the target changes in the induced draft fan speed, the target changes in the waste heat bypass ratio, and the target changes in the opening degree of the waste heat bypass valve for this cycle, and outputs "the total magnitude of the main control quantity changes executed in this cycle". These quantities are used in subsequent steps for time criteria and triggering of backup linkage, without changing the execution criteria of "exhaust air volume first, then waste heat bypass ratio" for the two types of main control quantities and "adjustment magnitude greater than or equal to the preset threshold" in this step.

[0082] In this embodiment, step S5 will be described in detail. In the embodiment of the present invention, step S5, the backup linkage and time criterion, is executed by the data processing and control unit within each common sampling cycle. The execution premise is that step S3 determines "deviation from the target temperature zone", and step S4 has been continuously executed according to the first control sequence, with the cumulative duration reaching or exceeding the preset master control evaluation time threshold, but the composite tank temperature has not yet entered the target temperature zone defined by the preset lower limit threshold and the preset upper limit threshold of the target temperature zone. Therefore, the data processing and control unit initiates the calculation and issuance of the "second control sequence", which is to jointly adjust the alumina feed rate, the electrolytic cell operating current, and the anode-cathode spacing.

[0083] The data processing and control unit first establishes a quantitative description of the heat gap. Using the composite bath temperature obtained in step S2 as the controlled variable, the difference between the current composite bath temperature and the target temperature zone boundary is read to form the composite bath temperature deviation. Combining the equivalent heat capacity and heat loss terms of the reduced-order model under the current operating conditions, the required equivalent heat power correction within a finite correction time domain is calculated. This equivalent heat power correction does not depend on the addition of new measurement channels; its parameters are derived from the model parameters and statistical filter output of step S2, and are updated according to the latest state in each common sampling period.

[0084] After obtaining the equivalent thermal power correction, the data processing and control unit maps it into the linkage increment of three types of execution quantities. The influence of alumina feed rate consists of two parts: first, the enthalpy effect related to the dissolution and reaction of alumina particles after entering the electrolyte; second, the sensible heat of the feed material determined by the preheater outlet temperature data; the two are combined into the equivalent heat exchange per unit feed rate. The influence of the electrolytic cell operating current is entered into the model through the electro-thermal coupling relationship, including the increase or decrease of the power term related to resistive Joule heating and reaction potential. The anode-cathode spacing changes the equivalent resistance and voltage of the electrolytic cell by affecting the electrolytic gap and polarization state, thereby changing the electrical input power. The data processing and control unit establishes a sensitivity matrix for the above three types of execution quantities at the current operating point. The sensitivity coefficients are obtained through local linearization of the reduced-order model and historical identification results, and maintain the same time base as the calculation of the convective heat dissipation component, material parameter component, and electrical parameter component in step S3.

[0085] The joint solution employs constrained least squares or an equivalent convex optimization method. The optimization variables are the increment of alumina feed rate, the increment of the electrolytic cell operating current, and the increment of the anode-cathode distance. The objective is to make the equivalent thermal power change, projected by the sensitivity matrix, approximate the required equivalent thermal power correction in the current common sampling period. Constraints fall into three categories: physical upper and lower limits and mechanical limits for each actuator; maximum rate of change and minimum resolvable step size for each actuator; and process safety-related operating boundaries (e.g., the electrolytic cell operating current does not exceed the rated current and allowable rate of change, the anode-cathode distance does not exceed the allowable process limit, and the alumina feed rate does not exceed the allowable upper or lower limit and remains consistent with the feeding cycle). When the optimization solution encounters any constraint boundary, the data processing and control unit automatically re-solves within the feasible region, and any unachieved portion is compensated for in the next common sampling period. To ensure numerical robustness and traceability, timestamps of the objective, constraints, solution vectors, and convergence flags are retained during the solution process.

[0086] The target increment obtained from the joint solution needs to be converted into executable device commands. The increment of alumina feed rate is converted into a feed command or feed rate adjustment command for the current cycle through the cycle time and single feed quantity parameters of the feed actuator; the increment of the electrolytic cell operating current is converted into a change in the active power setting of the converter system through the current regulating device and forms the electrolytic cell operating current setting value; the increment of the anode-cathode distance is converted into a change in the anode lifting position setting value through the anode-cathode distance regulating device. Before issuing the commands, the data processing and control unit performs unit consistency checks, upper and lower limit clipping, and change rate checks on the three types of commands respectively; for quantities that fail to meet acceptance standards due to mechanical hysteresis or transmission clearance, error backfilling compensation is performed in the next common sampling cycle in combination with position and current readback. After the commands are issued, the data processing and control unit synchronously collects alumina feed rate feedback, electrolytic cell operating current feedback, and anode-cathode distance position feedback, and forms a closed-loop execution error with the target value as the known quantity for the optimization problem in the next cycle.

[0087] In terms of operational logic, the execution of the second control sequence is a continuous rolling process. Each common sampling period is based on the latest composite tank temperature, wall heat dissipation indicators, and execution feedback, recalculating the equivalent heat power correction and the increments of the three types of execution quantities until the condition of "composite tank temperature entering the target temperature zone" is met. To avoid misjudgment caused by a single sample point, the data processing and control unit can use persistent confirmation for the "entering the target temperature zone" criterion, requiring the composite tank temperature to fall between the preset lower threshold and the preset upper threshold of the target temperature zone within several consecutive common sampling periods. Once confirmed to have entered the target temperature zone, the data processing and control unit freezes the incremental calculation of the second control sequence, writes the new baseline values ​​of the three types of execution quantities into the current operating condition, and automatically reverts to executing only step S4 of the first control sequence; if the composite tank temperature subsequently deviates again and meets the time criterion, this step is re-entered.

[0088] To ensure consistency with upstream steps, the input for this step remains solely from the data acquisition and time alignment in step S1, the composite tank temperature and wall heat dissipation parameters in step S2, the component identification results in step S3, and the execution record in step S4. No additional measurement channels are introduced. Quality markers remain in effect throughout this step: when any critical channel experiences a disconnection, over-range, or communication failure, the data processing and control unit downweights or temporarily removes the corresponding sensitivity or feedback; if removal renders the optimization problem unsolvable in the current common sampling period, the system retains the feasible solution from the previous period and records the event. All inputs, objectives, constraints, solutions, and execution feedback related to step S5 are archived with timestamps for post-event traceability and parameter tuning.

[0089] This embodiment also provides a system for implementing the method provided in the above embodiments, including: exhaust hoods and their electric baffles divided along the length of the tank; a variable frequency induced draft fan and a main exhaust pipe connected to the exhaust hoods; an indirect closed-loop waste heat exchange-preheating loop connected in parallel with the main exhaust pipe, including a heat exchanger, a bypass valve, and a material preheater; a data acquisition device for acquiring multiple point temperatures inside the hood, exhaust gas flow rates of each zone branch, heat flux of the tank sidewall, exhaust gas dew point temperature, preheater outlet temperature, purification device inlet temperature, and inlet flow rate; and a data acquisition device for acquiring zone baffles. The system includes a status acquisition device for plate opening degree, induced draft fan speed and waste heat bypass valve opening degree; and a data processing and control unit. The data processing and control unit is used to execute steps S1 to S5 in the method provided in the above embodiments and establish a restricted linkage interface with the feeding actuator, current adjustment device and anode-cathode spacing adjustment device. The system uses the exhaust air volume and waste heat bypass ratio as the main control variables. Under the condition of meeting various threshold conditions, the composite tank temperature is controlled within the target temperature range determined by the preset target temperature range lower limit threshold and the preset target temperature range upper limit threshold.

[0090] This embodiment will describe the system in detail. The system consists of a gas extraction and zone control subsystem, an indirect closed-loop waste heat exchange-preheating subsystem, a sensing and data acquisition subsystem, and a data processing and control unit and actuators. Zoned exhaust hoods are set along the length of the aluminum electrolysis cell. Each zone is equipped with an electric baffle, and the exhaust gas flows into the main exhaust duct through corresponding zone branches. The main exhaust duct is connected to a variable frequency induced draft fan, which provides extraction capacity and sends the flue gas to subsequent devices. The zoning allows for independent spatial adjustment of the extraction air volume by zone. The electric baffles provide continuously adjustable geometric channel states, and the induced draft fan speed provides system-level extraction capacity adjustment. Both correspond to the main control quantity of the extraction air volume in the method.

[0091] An indirect closed-loop waste heat exchange-preheating loop is connected in parallel to the main exhaust duct. This loop includes a heat exchanger, a bypass valve, and a material preheater. The working medium does not directly contact the dust- and fluorine-containing flue gas; recoverable heat is transferred to the material preheater through indirect heat exchange. The bypass valve is used to allocate the ratio of flue gas entering the heat exchanger to that passing through the bypass branch, thus forming the waste heat bypass ratio. The data processing and control unit outputs the target opening degree of the waste heat bypass valve, thereby controlling the waste heat bypass ratio, which is consistent with the main control variable of the waste heat bypass ratio in the method. A temperature measuring point is set at the outlet of the material preheater to provide a source of preheater outlet temperature data.

[0092] The sensing and data acquisition subsystem covers the following steps in step S19: Multiple temperature measurement points are set up inside the exhaust hood to collect temperature data; flow measurement devices are set up on the straight pipe sections of each zone branch to collect exhaust gas flow data; heat flux or temperature gradient measurement points are set up on the outer surface of the aluminum electrolysis cell sidewall to collect heat flux data; sampling points and dew point meters are set up between the main exhaust pipe and the purification device inlet to obtain exhaust gas dew point temperature data; temperature measurement points are set up at the material preheater outlet to obtain preheater outlet temperature data; temperature and flow measurement points are set up at the purification device inlet to obtain purification device inlet temperature and inlet flow data; zone baffle opening data and waste heat bypass valve opening data are read at each zone baffle actuator and valve positioner; and induced draft fan speed data is obtained at the frequency converter control cabinet or induced draft fan shaft end. All the above sensors and transmitters are uniformly connected to an analog quantity acquisition module, differential pressure acquisition module, or digital communication module. The data is sent to the data processing and control unit with a unified timestamp to complete the data acquisition and time alignment in step S1.

[0093] The data processing and control unit can be an industrial control computer or a programmable logic controller, internally configured to implement the software functions of steps S1 to S5: completing time alignment, unit unification, and quality mark management; running a reduced-order model and statistical filtering of the aluminum electrolysis cell's thermal-electric coupling to output composite cell temperature and wall heat dissipation indicators; performing target determination and component identification based on the preset lower and upper threshold values ​​of the target temperature zone; calculating and issuing the zone baffle opening target, induced draft fan speed target, and waste heat bypass valve opening target according to the first control sequence; when the preset main control evaluation time threshold is reached but the composite cell temperature has not entered the target temperature zone, calculating the alumina feed increment, electrolysis cell operating current increment, and anode-cathode distance increment jointly according to the second control sequence and constrained optimization strategy, and issuing commands to the corresponding actuators.

[0094] The actuators include a feeding actuator, a current regulating device, and an anode-cathode spacing regulating device. The feeding actuator adjusts the alumina feeding amount and feeding cycle according to the feeding command from the data processing and control unit; the current regulating device adjusts the active power output of the converter system according to the electrolytic cell operating current setpoint; and the anode-cathode spacing regulating device drives the anode to rise and fall according to the position setpoint, resulting in controlled changes in the anode-cathode spacing. These actuators, along with the partition electric damper, waste heat bypass valve, and variable frequency induced draft fan mechanism, form a closed-loop link, feeding back the partition damper opening, waste heat bypass valve opening, induced draft fan speed, feeding amount, electrolytic cell operating current, and anode-cathode spacing position to the data processing and control unit for command verification and rolling calculation in the next sampling cycle.

[0095] The system's piping connections, power supply, and control communication are implemented according to the process layout and electrical specifications: each zone branch and main exhaust duct uses corrosion-resistant and temperature-resistant materials, and necessary straight pipe sections and rectifier components are installed; shielded cables and control cables are laid separately between the variable frequency induced draft fan and the variable frequency control cabinet; the sensor cable shielding layer is grounded at one end; valve positioners and electric damper actuators use local control boxes and communicate with the data processing and control unit via fieldbus. Before commissioning, the opening-airflow curve, fan characteristic curve, and opening-distribution ratio curve are calibrated and imported on the system side, enabling the system to execute steps S1 to S5 according to the method logic provided in the above embodiment.

[0096] For example, in one specific embodiment, the object of this embodiment is a prebaked anode aluminum electrolytic cell, with four partitioned enclosures A / B / C / D set along the length of the cell; the preset threshold for the number of partitions in the parameter library is 3. An indirect closed-loop waste heat exchange-preheating loop (the clean circulating working fluid does not directly contact the dust- and fluorine-containing flue gas) is connected in parallel on the main exhaust duct, and the preheater preheats the alumina before it enters the cell. The common sampling period is set to 1 second, and the timing deviation is ≤100ms. This embodiment uses the following global threshold and execution cycle:

[0097] Lower threshold of target temperature zone: 950℃; Upper threshold of target temperature zone: 960℃.

[0098] Sampling increase threshold: Single target partition ≥ 1500 Its system's synthetic air volume is ≥3000 per cycle. ( (Based on 0℃, 101.325kPa).

[0099] Bypass ratio reduction threshold: absolute value ≥ 5%.

[0100] The master control evaluation time threshold is 10 minutes.

[0101] Dew point safety margin threshold: 12℃; during operation, the inlet temperature of the forced purification device is ≥ the dew point temperature of the exhaust gas + 12℃, and the exhaust gas collection efficiency is ≥ 95%.

[0102] First execution cycle (exhaust / exhaust air volume and zone baffle opening): ≥2s; Second execution cycle threshold (waste heat bypass ratio): ≥5s; Dead zone and hysteresis:

[0103] Dead zone threshold for exhaust air volume 400 Reverse hysteresis threshold 300 ;

[0104] The absolute value of the dead zone threshold for the waste heat bypass ratio is 0.01, and the absolute value of the reverse hysteresis band threshold is 0.008.

[0105] Abnormal degradation threshold: Exhaust air volume change rate limit ≤ 800 • Cycle. The absolute value of the waste heat bypass ratio offset is ≤3%.

[0106] Predictive control parameters: Prediction time threshold 300s; Advance trigger time threshold 60s; Pre-execution amplitude threshold (synthetic air volume) ≥2000 The pre-execution hold time threshold is 120 seconds.

[0107] Step S1: Data Acquisition and Time Alignment. Data after alignment at a representative moment: Multi-point temperature data inside the enclosure: Center points of sections A / B / C are 345℃ / 337℃ / 343℃; Exhaust gas flow data for each zone branch: A 6200, B 7500, C 6800, D 5900. Heat flux data for the tank sidewalls (area-weighted): A5.7, B6.4, C6.1, D5.5 Exhaust gas dew point temperature: 68℃; Preheater outlet temperature: 320℃; Purification unit inlet temperature and flow rate: 85℃ and 180,000. The data for the partition baffle openings are: A 55%, B 62%, C 50%, D 56%; the data for the induced draft fan speed is 72% of the rated speed; and the data for the waste heat bypass valve opening is 40%. The data processing and control unit completes resampling and interpolation marking with a common time axis of 1 second, synchronizes the unit, saves the quality mark, and prepares to enter the estimation stage.

[0108] Step S2, composite cell temperature estimation, employs a combined operation of a reduced-order model of the aluminum electrolysis cell's thermal-electric coupling and statistical filtering (unscented Kalman filter). Current cycle output:

[0109] Composite bath temperature: 962℃; Wall heat dissipation index: 5.9 (Weighted representation of sidewall area).

[0110] To improve robustness and interpretability, sensitivity-driven adaptive weighting is estimated for three key observation types, and the weights are updated using information gain. The information gain ratios within the current rolling window are: 0.48 for multi-point temperature data within the enclosure; 0.32 for exhaust gas flow data in each zone branch; and 0.20 for heat flux data on the tank sidewall. The normalized weights are updated accordingly to 0.48 / 0.32 / 0.20. Abnormal channels will be automatically downweighted and recorded.

[0111] Step S3, target determination and component identification, compares 962℃ with the target temperature threshold: 962℃ ≥ 960℃, indicating a deviation from the target temperature range (too high). Local linearization is applied to the reduced-order model at the current operating point to obtain the sensitivity of the three sets of driving variables: convection heat transfer, material parameters, and electrical parameters. Combined with the S2 weights, the composite tank temperature deviation of 2.0℃ is decomposed into:

[0112] Convection heat dissipation component: -0.8℃;

[0113] Material parameter quantity: +1.2℃;

[0114] Electrical parameter component: +1.6℃.

[0115] Based on the spatial distribution of wall heat dissipation indicators and heat flux data of the tank sidewalls, the heat loss in zones B / C is high (6.4 / 6.1). ).

[0116] Simultaneously, the system reads the exhaust gas dew point temperature as 68℃ and the dew point safety margin threshold as 12℃, locking the low-temperature end of the heat exchanger at a constraint of ≥80℃; combined with the purification device inlet temperature of 85℃ and inlet flow rate of 180,000... The verification meets the constraints of capture efficiency ≥ 95% and inlet temperature ≥ (dew point + 12℃).

[0117] Step S4: Main control quantity adjustment:

[0118] Execute the first control sequence:

[0119] (1) Adjustment of exhaust air volume: Calculate the change in the target of the system's synthetic air volume + 4000 (Meets the requirement that "the synthetic air volume is ≥3000 per batch") () Assigned to the target partition according to spatial weight: B+2500 C+1500 The conversion is divided into target changes in the baffle opening: B+5%, C+4%, which are issued after upper and lower limit checks; to compensate for the remaining deviation, the target change in the induced draft fan speed is issued simultaneously at +3%. The command issuance cycle meets the first execution cycle threshold ≥2s, and 400 is applied. Dead Zone and 300 Reverse hysteresis band to avoid high-frequency jitter.

[0120] (2) Waste heat bypass ratio adjustment: The direction is to decrease; directly reduce by 5% according to the threshold (from 40%→35%), satisfying "the absolute value of the bypass ratio reduction threshold is ≥5"; after the issuance, monitor the preheater outlet temperature and it changes from 320℃→328℃ within a few minutes; the cycle time meets the second execution cycle threshold ≥5s and applies a dead zone and reverse hysteresis of 0.01 / 0.008. The entire process is ensured by constrained predictive control + safety filtering with barrier function to guarantee that the inlet temperature of the purification device is ≥68℃+12℃=80℃ and the collection efficiency is ≥95%.

[0121] After 10 minutes of execution, the composite tank temperature reached 960.2℃ (still hitting the upper limit). Since the main control evaluation time threshold of 10 minutes had been reached and the tank had not yet entered the zone, step S5 was triggered.

[0122] Step S5: Backup Linkage and Time Criteria: The data processing and control unit calculates the equivalent thermal power correction of -2.5kW required in the next 5 minutes based on the reduced-order model and current heat loss. Under constraints, the linkage increment of the second control sequence is obtained by solving for:

[0123] Alumina feed rate set at +8%;

[0124] The electrolytic cell operating current is set to -0.6%;

[0125] Anode-cathode spacing set to -1mm.

[0126] After the command was issued and verified through feedback closed loop, the temperature of the composite bath dropped from 960.2℃ to 957℃ within 5 minutes. The system uses a 30-second continuous confirmation window to confirm that the temperature of the composite bath is stable within the target temperature range of 950–960℃. Then, the second control sequence is frozen, and only the first control sequence is kept running.

[0127] Additionally, in some scenarios:

[0128] Scenario 1: Anode replacement event handling. The anode is scheduled to be replaced at 14:30. Parameter settings: event advance time threshold 120s, pumping increase threshold ≥1500. (Per target partition), bypass ratio reduction threshold of 5%, event recovery time threshold of 240s.

[0129] Starting at 14:28:30, the baseline for exhaust air volume in adjacent B / C zones will be increased to B+2000. C+1500 Meanwhile, the waste heat recovery rate was further reduced by 5% year-on-year (e.g., from 38% to 33%). The preheater outlet temperature decreased from 326℃ to 334℃ within 2 minutes.

[0130] When the event ends and the conditions of the composite tank temperature being within the zone, the collection efficiency being ≥95%, and the inlet temperature being ≥(dew point + 12℃) are met, the system will be restored to the baseline setting before the event using a 240s linear ramp.

[0131] Scenario 2: Sensor anomaly and degradation, 15:05, residual error of heat flux sensor on the side wall of a tank in section C is continuously >3. An alarm has been triggered. Immediately implement a downgrade:

[0132] Freeze the weight of this channel in adaptive weighting (press it down to the lower limit of 0.05);

[0133] The rate of change of exhaust air volume is limited to ≤800. ·cycle;

[0134] The waste heat bypass ratio offset is limited to an absolute value of ≤3%;

[0135] Continue operation until on-site calibration is successful, then remove the downgrade.

[0136] Scenario 3: Threshold and Early Execution. The prediction model's time-domain threshold is 300 seconds. At 15:25, the prediction indicates that the threshold will be exceeded after 60 seconds. The system performs early pre-execution:

[0137] Synthetic air volume +2000 (≥Pre-execution amplitude threshold);

[0138] The waste heat bypass ratio will be further reduced by 5%;

[0139] Hold for 120 seconds.

[0140] Result: The predicted over-limit range changed from ≈2.0℃ to ≈0.6℃. Subsequently, conventional control can keep the composite tank temperature in the target temperature range without triggering the backup linkage.

[0141] It should be noted that the embodiments of the present invention have better implementability and are not intended to limit the present invention in any way. Any person skilled in the art may use the above-disclosed technical content to change or modify it into equivalent effective embodiments. However, any modifications or equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of the technical solution of the present invention.

Claims

1. A method for precise temperature control in an aluminum smelting process, characterized in that, include: Step S1: Data Acquisition and Time Alignment: Collect and time-align the following data items in the exhaust hood above the electrolyte in the aluminum electrolysis cell, on the outer side wall of the cell, and in the area adjacent to the cathode: (1) Temperature data at multiple points inside the enclosure; (2) Waste gas flow data of each zone branch; (3) Heat flux data of the tank sidewall; (4) Waste gas dew point temperature data; (5) Preheater outlet temperature data; (6) Purification device inlet temperature data and inlet flow data; (7) Zone baffle opening data; (8) Exhaust fan speed data; (9) Waste heat bypass valve opening data. Step S2 Composite tank temperature estimation: Input the data from (1) to (9) of step S1 into the reduced-order model of thermal-electric coupling of aluminum electrolysis tank, and correct the model state through statistical filtering to obtain the composite tank temperature and wall heat dissipation index. Step S3 Target Determination and Component Identification: When the temperature of the composite tank is less than or equal to the preset lower threshold of the target temperature zone or greater than or equal to the preset upper threshold of the target temperature zone, it is determined to be a deviation from the target temperature zone, and the temperature difference is decomposed into convection heat dissipation component, material parameter component and electrical parameter component. Step S4 Main control quantity adjustment: When a deviation is determined, the controller executes according to the first control sequence: (a) Adjust the exhaust air volume according to the partition baffle opening data and the induced draft fan speed data, with the adjustment range being greater than or equal to the preset exhaust increase threshold; (b) Adjust the waste heat bypass ratio and change the preheater outlet temperature according to the waste heat bypass valve opening data, with the adjustment range being greater than or equal to the preset bypass ratio reduction threshold, so as to change the convective heat dissipation and the sensible heat of the material entering the tank, so that the composite tank temperature returns to the target temperature zone; Step S5 Backup linkage and time criterion: If the composite cell temperature does not enter the target temperature zone after the first control sequence is completed and the time is greater than or equal to the preset main control evaluation time threshold, then execute the second control sequence: linkage adjustment of alumina feed amount, electrolytic cell working current and anode-cathode distance until the composite cell temperature enters the target temperature zone.

2. The precise temperature control method for aluminum smelting process according to claim 1, characterized in that, The exhaust hood is divided into multiple independently controlled partitions along the length of the trough, with the number of partitions being greater than or equal to a preset threshold. The controller determines the target partition based on the spatial distribution of the composite trough temperature and the wall heat dissipation index of each partition, and executes an instruction to adjust the exhaust air volume of the target partition.

3. The precise temperature control method for aluminum smelting process according to claim 1, characterized in that, Waste heat exchange adopts an indirect closed-loop method, where the clean circulating working fluid is heated and then used to preheat alumina particles; the controller maintains the temperature at the low-temperature end of the heat exchanger greater than or equal to the sum of the waste gas dew point temperature and the preset dew point safety margin threshold based on the waste gas dew point temperature data, and inputs the preheater outlet temperature into the control algorithm.

4. The precise temperature control method for aluminum smelting process according to claim 3, characterized in that, The control algorithm satisfies the following constraints throughout the process: the exhaust gas capture efficiency is greater than or equal to the preset lower limit threshold of capture efficiency, and the inlet temperature of the purification device is greater than or equal to the sum of the exhaust gas dew point temperature and the preset dew point safety margin threshold; the controller executes the constraints at the computational level through constrained predictive control or safety filtering with barrier functions.

5. The precise temperature control method for aluminum smelting process according to claim 4, characterized in that, The calculation of the composite tank temperature uses a sensitivity-driven adaptive weighting of the multi-point temperature data inside the hood, the exhaust gas flow data of each zone branch, and the heat flux data of the tank sidewall. Information gain is used as the weight update criterion to calculate the contribution of each sensor quantity to the composite tank temperature and the heat dissipation index of the wall in real time and dynamically update the weights.

6. The method for precise temperature control in the aluminum smelting process according to claim 5, characterized in that, For incidents involving anode replacement, power modulation, or shell breakage during feeding, the incident handling process includes: (a) When the advance time before the start of the event is greater than or equal to the preset event advance time threshold, the baseline of the exhaust air volume of the adjacent zone is raised according to the partition baffle opening data and the exhaust fan speed data, and the increase is greater than or equal to the preset exhaust increase threshold; at the same time, the waste heat bypass ratio is reduced according to the waste heat bypass valve opening data, and the reduction is greater than or equal to the preset bypass ratio reduction threshold. (b) When the event ends and the composite tank temperature no longer meets the deviation judgment condition of step S3, and the exhaust gas collection efficiency is greater than or equal to the preset lower limit threshold of collection efficiency, and the inlet temperature of the purification device is greater than or equal to the sum of the exhaust gas dew point temperature and the preset dew point safety margin threshold, the temperature is restored to the pre-event set value by a linear ramp within a duration greater than or equal to the preset event recovery time threshold.

7. The method for precise temperature control in the aluminum smelting process according to claim 1, characterized in that, Anomalies in temperature sensors, heat flow sensors, and air volume meters are identified using a residual-based sequence check. When anomalies are detected, a degradation strategy is implemented: the weight of the corresponding data item in the weight update is frozen; and the rate of change of the exhaust air volume is limited to be less than or equal to a preset rate of change limit threshold per execution cycle. The absolute value of the offset of the waste heat bypass ratio is less than or equal to a preset bypass ratio offset threshold.

8. The method for precise temperature control in the aluminum smelting process according to claim 1, characterized in that, Based on the physical-data fusion model, the temperature of the composite tank is predicted in the short term within a preset prediction time domain threshold. When the prediction shows that the temperature will exceed the target temperature zone after a preset advance trigger time threshold, the controller pre-executes the adjustment amount of the first control sequence according to the partition baffle opening data, the induced draft fan speed data and the waste heat bypass valve opening data. The pre-execution amplitude is greater than or equal to the preset pre-execution amplitude threshold and the duration is greater than or equal to the preset pre-execution hold time threshold.

9. The method for precise temperature control in the aluminum smelting process according to claim 1, characterized in that, The execution cycle of the exhaust air volume and the partition baffle opening is greater than or equal to a preset first execution cycle threshold, and the execution cycle of the waste heat bypass ratio is greater than or equal to a preset second execution cycle threshold; both are respectively set with preset dead zone thresholds and preset reverse hysteresis band thresholds.

10. A system for implementing the method of claim 1, characterized in that, include: The system comprises: exhaust hoods and their electric baffles partitioned along the length of the tank; a variable frequency induced draft fan and a main exhaust duct connected to the exhaust hoods; an indirect closed-loop waste heat exchange-preheating circuit connected in parallel with the main exhaust duct, including a heat exchanger, a bypass valve, and a material preheater; a data acquisition device for acquiring multiple temperatures within the hoods, exhaust gas flow rates in each partition branch, heat flux through the tank sidewalls, exhaust gas dew point temperature, preheater outlet temperature, and purification device inlet temperature and inlet flow rate; a status acquisition device for acquiring partition baffle openings, induced draft fan speeds, and waste heat bypass valve openings; and a data processing and control unit, which executes steps S1 to S5 of claim 1 and establishes a restricted linkage interface with the feeding actuator, current regulating device, and anode-cathode spacing regulating device; wherein, the system uses the exhaust air volume and the waste heat bypass ratio as the main control variables, and, under the condition of satisfying various threshold conditions, controls the composite tank temperature within the target temperature range determined by the preset target temperature range lower threshold and the preset target temperature range upper threshold.

Images