A tunnel kiln roasting control method and system, electronic equipment and storage medium

Abstract

The application provides a tunnel kiln roasting control method and system, electronic equipment and a storage medium, which are suitable for the technical field of tunnel kiln roasting control. The tunnel kiln roasting control method comprises the following steps: based on a production target, a current comprehensive state feature, a partition working condition of all partitions and a historical optimal roasting curve feature, an optimization model is used for optimization training to obtain a current optimal roasting curve feature; target control parameters of each time period are calculated according to the current optimal roasting curve feature; the tunnel kiln is controlled to roast according to time periods: taking the target control parameters of the current time period as a target, based on external disturbance variables and current control states of each partition, rolling optimization adjustment is performed to obtain current optimal cooperative control instructions, and the tunnel kiln is controlled to roast in the current time period; the roasting curve optimization and the multi-parameter control execution depth are coupled and integrated to improve the production efficiency and the overall performance of the tunnel kiln.

Description

Technical Field

[0001] This application relates to the technical field of tunnel kiln roasting control, specifically to a tunnel kiln roasting control method, system, electronic equipment, and storage medium. Background Technology

[0002] Tunnel kiln firing control is a core element in ensuring product quality and reducing energy consumption.

[0003] While existing technologies have achieved automated control in some aspects, they still have the following key shortcomings:

[0004] 1. Disconnect between roasting curve optimization and control execution: Traditional methods often use fixed roasting curves or fine-tune optimization based on human experience, with the curve optimization process independent of real-time control execution. Even if an optimized theoretical roasting curve is obtained, the control execution still relies on the independent operation of each control loop (temperature, pressure, air-fuel ratio), which cannot guarantee that the optimized theoretical roasting curve can be executed accurately and in a coordinated manner, resulting in "optimization not being implemented".

[0005] 2. Lack of a coordinated mechanism for multi-parameter control oriented towards global objectives: Existing control methods often employ independent PID control loops or simple decoupling control to regulate multiple parameters such as temperature, pressure, atmosphere, and flow rate. When changes occur (such as adjustments and optimizations to the calcination curve), there is a lack of a unified coordinated strategy among the control loops to allocate control quantities, which easily leads to control conflicts (for example, increasing the airflow to rapidly raise the temperature may disrupt kiln pressure stability), resulting in system oscillations or the forced sacrifice of some calcination parameters.

[0006] 3. Insufficient adaptability: The existing solution has poor adaptability to changes in operating conditions such as changes in raw material characteristics, equipment status drift, and fluctuations in production load. It cannot maintain optimal performance continuously under changing operating conditions.

[0007] 4. Single optimization dimension: Most optimization methods only target a single indicator (such as temperature tracking accuracy or gas consumption), lacking the ability to comprehensively weigh and optimize multiple objectives such as product quality, energy consumption, efficiency, and stability. Summary of the Invention

[0008] To address one of the aforementioned technical deficiencies, this application provides a tunnel kiln roasting control method, system, electronic equipment, and storage medium.

[0009] According to the first aspect of this application, a method for controlling the roasting of a tunnel kiln is provided, comprising:

[0010] The production target, current comprehensive status characteristics, and historical optimal roasting curve characteristics are obtained; the current comprehensive status characteristics include the primary characteristics of the raw materials and real-time operating conditions.

[0011] The tunnel kiln body is divided into multiple zones along its length, and the operating conditions of each zone are obtained.

[0012] Based on production targets, current overall status characteristics, operating conditions of all zones, and historical best roasting curve characteristics, the current best roasting curve characteristics are obtained through optimization training using an optimization model.

[0013] Calculate the target control parameters for each time period based on the current optimal roasting curve characteristics;

[0014] The firing process in the tunnel kiln is controlled according to time periods, specifically including:

[0015] Real-time acquisition of external disturbance variables and the current control status of each partition ;

[0016] Using the target control parameters for the current time period as the objective, based on external disturbance variables and the current control status of each partition The optimal collaborative control command is obtained by rolling optimization adjustment through prediction model and optimization function, and the calcination control of tunnel kiln for the current time period is carried out according to the optimal collaborative control command.

[0017] According to a second aspect of this application, a tunnel kiln roasting control system is provided, including modules for implementing the tunnel kiln roasting control method as described above.

[0018] According to a third aspect of this application, an electronic device is provided, comprising:

[0019] Memory;

[0020] Processor; and

[0021] Computer programs;

[0022] The computer program is stored in the memory and configured to be executed by the processor to implement the tunnel kiln firing control method as described above.

[0023] According to a fourth aspect of this application, a computer-readable storage medium is provided having a computer program stored thereon; the computer program is executed by a processor to implement the tunnel kiln roasting control method as described above.

[0024] The beneficial effects of this application are as follows:

[0025] This application employs a closed-loop architecture of perception-decision-execution-optimization, providing a method for deeply coupling and integrating roasting curve optimization with multi-parameter control execution into a unified and collaborative control system. This achieves integrated roasting control from production targets to target control parameters, and then to the current optimal collaborative control command. It fundamentally solves the disconnect between optimization and control, ensuring that the current optimal roasting curve characteristics can be executed accurately and collaboratively, improving production efficiency and enhancing the overall performance of the tunnel kiln. It achieves global collaborative regulation of target control parameters, effectively resolving control conflicts and significantly improving overall stability and control quality. It possesses strong adaptive and self-learning capabilities, automatically optimizing the current optimal roasting curve characteristics based on the primary characteristics of the raw materials and real-time and zoned operating conditions; using the target control parameters for the current time period as the objective, and based on the current control state... and external disturbance variables By using predictive models and optimization functions for rolling optimization adjustments, the system obtains the current optimal collaborative control command, continuously learning the closed loop, and exhibits excellent adaptability to complex and ever-changing production environments. It considers multiple dimensions of indicators such as product quality, energy consumption, efficiency, and stability, and achieves comprehensive trade-offs in actual control through an integrated collaborative approach, contributing to maximizing economic benefits. It reduces reliance on senior operators by embedding expert optimization curves and experience in handling control conflicts into the algorithm model, achieving automation and standardization, reducing dependence on human experience, and ensuring production consistency.

[0026] Other features and advantages of this application will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing the application. The objectives and other advantages of this application may be realized and obtained by means of what is pointed out in the written description and the accompanying drawings. Attached Figure Description

[0027] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments of this application and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0028] Figure 1 A flowchart illustrating a method for controlling the firing of a tunnel kiln provided in this application;

[0029] Figure 2 A schematic diagram illustrating the process of obtaining the current optimal calcination curve characteristics provided for this application;

[0030] Figure 3 A schematic diagram illustrating the calculation process of the target control parameters provided in this application;

[0031] Figure 4This is a schematic diagram of the process for controlling the roasting of a tunnel kiln during the current time period, as provided in this application.

[0032] Figure 5 This application provides a functional structure diagram of a tunnel kiln roasting control system.

[0033] Figure 6 for Figure 5 A schematic diagram of the functional structure of the optimized training module;

[0034] Figure 7 for Figure 5 A functional structure diagram of the target control parameter calculation module;

[0035] Figure 8 for Figure 5 Functional structure diagram of the calcination control unit;

[0036] In the picture:

[0037] 10 is the acquisition module, 20 is the partitioning module, 30 is the optimization training module, 40 is the target control parameter calculation module, 50 is the roasting control module, 301 is the construction unit, 302 is the initialization unit, 303 is the execution unit, 401 is the theoretical net heating power calculation unit, 402 is the input heat calculation unit, 403 is the target control parameter calculation unit, 404 is the verification unit, 405 is the first output unit, 501 is the real-time acquisition unit, 502 is the roasting control unit, 3031 is the simulation unit, 3032 is the optimization parameter calculation unit, 3033 is the reward value calculation unit, 3034 is the first judgment unit, 3035 is the training optimization unit, 3036 is the curve feature acquisition unit, 5021 is the prediction unit, 5022 is the control command acquisition unit, 5023 is the current roasting unit, 5024 is the second judgment unit, 5025 is the second output unit, and 5026 is the update unit. Detailed Implementation

[0038] To make the technical solutions and advantages of the embodiments of this application clearer, the exemplary embodiments of this application will be described in further detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not an exhaustive list of all embodiments. It should be noted that, unless otherwise specified, the embodiments and features in the embodiments of this application can be combined with each other.

[0039] like Figure 1 As shown, to address the above problems, this application provides a method for controlling the roasting of a tunnel kiln, including:

[0040] The system acquires production targets, current comprehensive status characteristics, and historical best roasting curve characteristics. The current comprehensive status characteristics include the primary characteristics of the raw materials and real-time operating conditions. The primary characteristics of the raw materials include chemical composition, moisture content, particle size distribution, and thermal analysis data. The real-time operating conditions include the basic frequency of the blowers in the tunnel kiln, the kiln car advance speed (i.e., the car entry speed, which is usually uniform and is the advance speed of the entire kiln car queue), and the calorific value of the fuel gas.

[0041] The tunnel kiln body is divided into multiple zones along its length, and the operating conditions of each zone are obtained; the operating conditions of each zone include temperature, pressure, and atmosphere composition.

[0042] Based on production targets, current overall status characteristics, operating conditions of all zones, and historical best roasting curve characteristics, the current best roasting curve characteristics are obtained through optimization training using an optimization model.

[0043] Calculate the target control parameters for each time period based on the current optimal roasting curve characteristics;

[0044] The firing process in the tunnel kiln is controlled according to time periods, specifically including:

[0045] Real-time acquisition of external disturbance variables and the current control status of each partition External disturbance variables (global variables) include kiln car advance speed, ambient temperature, and fuel gas calorific value; the current control state includes temperature and pressure, and may also include kiln hot body temperature; external disturbance variables are measurable but uncontrollable, and are used as feedforward information input to the prediction model to compensate for changes in fuel energy input in advance; the fuel gas calorific value here is the same data source as the fuel gas calorific value collected in real-time operating conditions, but plays the role of feedforward disturbance variable in the control architecture;

[0046] Using the target control parameters for the current time period as the objective, based on external disturbance variables and the current control status of each partition The current optimal collaborative control command is obtained by rolling optimization adjustment through prediction model and optimization function (the current optimal collaborative control command is the execution command that can be directly manipulated, including the opening command of natural gas proportional regulating valve, the opening command of combustion air regulating valve, and the frequency control command of key fans, including exhaust fan, cooling fan and undercarriage balancing fan). The calcination control of tunnel kiln for the current time period is carried out according to the current optimal collaborative control command.

[0047] In the above scheme, the primary characteristics of the raw materials can be used to determine the heat demand, reaction stage, and suitable temperature range during the roasting process. Real-time operating conditions can be acquired in real time by sensors, providing a foundation for optimizing the current optimal roasting curve characteristics. Based on the above scheme, a closed-loop architecture of perception-decision-execution-optimization is adopted, providing a method for deeply coupling and integrating roasting curve optimization with multi-parameter control execution, achieving integrated roasting control from production targets to target control parameters, and then to the current optimal collaborative control command. This fundamentally solves the disconnect between optimization and control, ensuring that the current optimal roasting curve characteristics can be executed accurately and collaboratively, improving production efficiency and enhancing the overall performance of the tunnel kiln. It achieves global collaborative regulation of target control parameters, effectively resolving control conflicts and significantly improving overall stability and control quality. It possesses strong adaptive and self-learning capabilities, automatically optimizing the current optimal roasting curve characteristics based on the primary characteristics of the raw materials, real-time operating conditions, and zone operating conditions; taking the target control parameters of the current time period as the target, based on the current control state... and external disturbance variables By using predictive models and optimization functions for rolling optimization adjustments, the system obtains the current optimal collaborative control command, continuously learning the closed loop, and exhibits excellent adaptability to complex and ever-changing production environments. It considers multiple dimensions of indicators such as product quality, energy consumption, efficiency, and stability, and achieves comprehensive trade-offs in actual control through an integrated collaborative approach, contributing to maximizing economic benefits. It reduces reliance on senior operators by embedding expert optimization curves and experience in handling control conflicts into the algorithm model, achieving automation and standardization, reducing dependence on human experience, and ensuring production consistency.

[0048] like Figure 2 As shown, further, the optimization training of the optimization model based on production targets, current comprehensive status characteristics, partition conditions of all partitions, and historical optimal roasting curve characteristics to obtain the current optimal roasting curve characteristics specifically includes:

[0049] Construct a comprehensive benefit function and adjust it according to production targets; the resulting comprehensive benefit function is as follows: ;

[0050] In the formula, Indicates quality deviation. Indicates cumulative energy consumption, Indicates production time (one cycle). Indicates a stability penalty. They represent Weighting coefficients;

[0051] Determine heat requirements and reaction stages based on the primary characteristics of the raw materials;

[0052] The real-time operating condition and the operating conditions of all partitions are used as the initial state features. The historical best roasting curve feature is used as the first roasting curve feature. ;

[0053] Starting from t=1, perform the following steps:

[0054] The (t-1)th state feature Input into the simulation module (which can be a high-fidelity simulation environment built from the digital twin module), and based on the characteristics of the t-th calcination curve. Perform a simulation (run a complete calcination cycle) to obtain the t-th state feature. and the t-th virtual parameter; where the t-th virtual parameter includes the t-th product quality. The t-th virtual energy consumption The t-th virtual period and the fluctuation data of the t-th process ;

[0055] The t-th optimized parameter is calculated based on the t-th virtual parameter; where the t-th optimized parameter includes the t-th quality deviation. The t-th cumulative energy consumption The t-th production time and the t-th stability penalty The optimization parameters calculated based on the t-th virtual parameter specifically include:

[0056] Based on the quality of the t-th product The t-th quality deviation was calculated. The calculation formula is:

[0057] ,in, This represents the m-th product quality index (including strength, density, etc.) in the t-th product quality obtained from the simulation. This represents the target value (a preset value) for the m-th product quality indicator. This represents the weight value of the m-th product quality indicator (a preset value).

[0058] The t-th virtual energy consumption The t-th virtual period These are respectively used as the t-th cumulative energy consumption The t-th production time Specifically:

[0059] make ;

[0060] make ;

[0061] Fluctuation data from the t-th process The stability penalty of the t-th term is calculated. The calculation formula is:

[0062] ;

[0063] In the formula, This represents the fluctuation data of the t-th process during the simulation. The sequence of the j-th key controlled parameter changing over time. This represents the variance calculation function. This represents the weight of the j-th key controlled parameter; the key controlled parameters specifically include the operating temperature of each zone, the operating pressure at key locations (such as the exhaust port, the pressure under the vehicle), etc. The key controlled parameters are a direct reflection of the system's energy and mass balance.

[0064] The t-th optimization parameter is input into the comprehensive benefit function to calculate the t-th reward value. , ;

[0065] Determine the t-th reward value Whether the difference between the reward value and the previous n reward values ​​is less than a preset threshold, or determine... Has the preset number of iterations been reached?

[0066] If not, then the t-th state feature The input optimization model (which can be a Bayesian optimization parameter search model or a Deep Deterministic Policy Gradient (DDPG) reinforcement learning model) is used for optimization training to obtain the t+1th roasting curve feature. And repeat the above steps;

[0067] If so, then start from the t-th reward value Select the maximum value from the first n reward values, then take the maximum value as the optimal reward value, and take the roasting curve feature corresponding to the optimal reward value as the current optimal roasting curve feature.

[0068] Specifically, the current optimal roasting curve features include vectors of heating rate, cooling rate, target temperature, and holding time for each stage (preparation, heating, roasting, holding, and cooling). The independent variable (horizontal axis) of the current optimal roasting curve, which is composed of these features, is usually time or position within the kiln (position along the length of the tunnel kiln body), while the dependent variable (vertical axis) is usually temperature, but can also be pressure or atmosphere composition.

[0069] Based on the above scheme, a comprehensive benefit function is constructed by considering quality deviation, cumulative energy consumption, production time, and stability penalty. This ensures that the calculated reward value is a more accurate balance point between quality, energy consumption, efficiency, and fluctuation, thereby guaranteeing that the current optimal roasting curve characteristics can adapt to the current raw material roasting requirements and production targets. This solves the limitations of existing technologies that use fixed roasting curves or rely on manual experience for fine-tuning optimization, which makes it difficult to quantify the balance. Among these, quality deviation reflects the degree of deviation (weighted mean square error) between the physicochemical properties (such as strength, density, water absorption, and color) of the roasted product and the target value. The weights of the target value and physicochemical properties are preset according to the product specifications. Cumulative energy consumption refers to the combined cost of natural gas and electricity consumed to complete one roasting cycle, which is directly derived from energy metering data (which can be directly obtained). Production time refers to the total time required from the kiln car entering the kiln to exiting the kiln to complete a full roasting cycle. The stability penalty is used to suppress the fluctuation amplitude of key parameters such as temperature and pressure (the weighted sum of the variance of fluctuations around their set values). Reduce reliance on expert experience: This involves leveraging senior process engineers' understanding of the correlation between raw materials, processes, and results; their decision-making logic for adjusting curve characteristics under real-time operating conditions; and their experience in balancing quality and energy consumption. Through a simulation module (where state characteristics form the state space, the calcination curve characteristics form the action space, and the negative value of the comprehensive benefit function is the reward function), the system can adjust based on the input (t-1)th state characteristic. and the characteristics of the t-th roasting curve Simulation was performed to obtain the t-th state feature. Given the t-th virtual parameter, the t-th optimized parameter can be calculated based on the t-th virtual parameter, and then input into the comprehensive benefit function to calculate the t-th reward value. If the t-th reward value The difference between the reward value and the previous n reward values ​​is not less than a preset threshold, and If the preset number of iterations is not reached, the optimization model is trained to obtain the (t+1)th roasting curve feature. Then repeat the above operation, based on the t-th state feature. and the characteristics of the (t+1)th roasting curve Then perform the simulation again; until the t-th reward value. The difference between the value and the first n reward values ​​is less than the preset threshold, or Once the preset number of iterations is reached, the reward value starts from the t-th value. Select the maximum value from the first n reward values, then take the maximum value as the optimal reward value, and take the roasting curve feature corresponding to the optimal reward value as the current optimal roasting curve feature.

[0070] The above solution supports data-driven continuous improvement. Based on historically optimal roasting curve characteristics and real-time feedback, it continuously learns, forming a closed loop of "application-evaluation-optimization." This allows the roasting curve characteristics to continuously evolve with the accumulation of production data, achieving continuous optimization of the production process, maximizing reward values, and ultimately obtaining the current optimal roasting curve characteristics that adapt to the primary characteristics of raw materials, real-time operating conditions, and zoned operating conditions, significantly improving production flexibility. It achieves standardization and automation of process optimization decisions, reducing the impact of human differences on product quality and ensuring production consistency. It solves the problem in existing technologies that only address single indicators (such as temperature tracking accuracy or gas consumption), lacking the ability to comprehensively balance and optimize multiple objectives such as product quality, energy consumption, efficiency, and stability. It also addresses the insufficient adaptability of existing solutions: they are poorly adaptable to changes in raw material characteristics, equipment state drift, and production load fluctuations, and cannot continuously maintain optimal performance under changing operating conditions.

[0071] Specifically, adjustments are made based on production targets, including:

[0072] When the production target is high quality, increase the quality deviation. Weighting coefficients Adaptively reduce cumulative energy consumption Weighting coefficients and maintain stability penalty Weighting coefficients At a relatively high level; (for example, to Increase to or 1.5 to 2 times; Reduce by 10% to 20% to allow for a slight increase in energy consumption to meet more precise process control; while maintaining (At a higher level to suppress process fluctuations).

[0073] When the production target is energy conservation, increase cumulative energy consumption. Weighting coefficients Adaptively reduce quality deviation Weighting coefficients and slightly reduce production time. Weighting coefficients ; (for example, to Set as the highest weight; Reduce by 20% to 30% to accept minor fluctuations in product quality indicators within the acceptable range; and slightly reduce... This allows for a gentler heating / cooling rate, thereby reducing peak energy consumption.

[0074] When the production target is rapid delivery, increase production time. Weighting coefficients Adaptively reduce stability penalty Weighting coefficients and maintain quality deviation Weighting coefficients To ensure the basic level of product quality; (e.g., improve) ;Will Reduce by 15% to 25% to allow for more aggressive but slightly more volatile control strategies during process transitions (such as heating and cooling); while maintaining (Based on ensuring product quality at a basic level).

[0075] The production objectives in this application include product type, specifications, quality requirements, delivery time, etc. The weights in the comprehensive benefit function can be adaptively and dynamically adjusted according to the production objectives, so that the reward value calculated by the comprehensive benefit function can adapt to the requirements of the production objectives. This makes the current optimal roasting curve characteristics more accurate, ensures the accuracy of roasting control, and improves the qualification rate of the roasted product.

[0076] Specifically, taking the DDPG reinforcement learning model as an example, the optimization model is trained based on multiple sets of historical data (the relevant data in this application can also be used as multiple sets of historical data storage and participate in the optimization training of the optimization model);

[0077] Historical data is ;

[0078] in, Represents the characteristics of the t-th historical state. express Input the characteristics of the (t+1)th historical roasting curve obtained from the optimization model, Indicates according to The historical reward value obtained from the simulation at the (t+1)th time... Indicates according to The (t+1)th historical state feature obtained from the simulation;

[0079] During the optimization training process using multiple sets of historical data, a critic network (evaluating the value of actions) is trained to more accurately assess the value of state-action pairs, and an actor network (learning the optimal policy) is trained to output actions that yield higher predicted rewards. The (t+1)th historical roasting curve feature is generated by the updated policy network, thereby guiding the exploration towards a better direction.

[0080] like Figure 3 As shown, further, the target control parameters for each time period are calculated based on the current optimal roasting curve characteristics (wherein the target control parameters for each time period include the target gas flow rate for each zone). Target combustion air volume for each zone Target temperature for each zone The target pressure value P at key locations includes:

[0081] Obtain the secondary properties of the raw material (including specific heat capacity) ,density Based on the secondary properties of the raw materials and the kiln car advance speed Based on the characteristics of the current optimal roasting curve Calculate the theoretical net heating power required for each zone in each time period. The calculation formula is:

[0082] (Based on the principle of energy balance)

[0083] In the formula, This indicates the cross-sectional area of ​​the material carried by the kiln car. Indicates the optimal calcination curve at position Time period The rate of heating (or cooling) at that time;

[0084] Based on the theoretical net heating power required by each zone in each time period and the estimated total heat loss of each zone in each time period The required input heat to be added to the kiln body for each zone during each time period was calculated. , ; wherein, the estimated comprehensive heat loss Based on the kiln structure parameters, insulation material properties, flue gas flow rate and flue gas temperature, the calculation is performed using empirical formulas or steady-state heat transfer models (this part belongs to existing technology), mainly including heat dissipation from the kiln wall and heat carried away by the flue gas.

[0085] Input heat in each zone during each time period The target control parameters for each time period were calculated. Specifically, it includes:

[0086] The temperature in the optimal calcination curve is used as the target temperature. ;

[0087] Calculate the target gas flow rate for each zone within each time period. The calculation formula is:

[0088] ;

[0089] In the formula, Indicates burner efficiency (representing the effectiveness of combustion of the burner, which is based on empirical or model parameters of burner model, historical operating data and maintenance status, and can be calibrated periodically). This indicates the lower heating value of the current fuel;

[0090] Based on the target gas flow rate of each zone in each time period Calculate the corresponding target combustion air volume Specifically, it includes:

[0091] Calculate the optimal air-fuel ratio for the current fuel. Specifically, it includes:

[0092] Key process parameters of each zone (including gas flow rate, combustion air volume, kiln temperature and flue gas temperature, etc.) are collected in real time, and online optimization is performed using algorithms such as reinforcement learning or extreme value search. Specifically, within a preset safety range, the combustion air volume is automatically adjusted and explored in a small and continuous manner.

[0093] Through online optimization, the optimal air-fuel ratio that maximizes instantaneous thermal efficiency or minimizes overall energy consumption (gas + electricity) in the current zone is identified and output in real time.

[0094] The optimal air-fuel ratio is calculated and updated in real time by an independent dynamic optimization model for air-fuel ratio. The dynamic optimization model is a closed-loop adaptive optimizer that continuously seeks and locks the most economical operating point in the combustion process while meeting the process requirements.

[0095] Calculate the target combustion air volume The calculation formula is:

[0096] ;

[0097] In the formula, The stoichiometric constant representing the amount of air required for the complete combustion of a theoretical unit of current fuel;

[0098] Calculate the target pressure value P at key locations (such as exhaust vents, undercarriage, etc.) within each time period, specifically including:

[0099] Construct a simplified network model of gas flow inside the kiln;

[0100] Based on the target combustion air volume The amount of flue gas generated is used as the flow boundary condition for the cut-off point in the simplified network model;

[0101] By using a set of nonlinear equations describing the pressure-gas flow relationship in the pipeline network (such as the Hardy-Cross method), the target pressure value that needs to be maintained at key locations (such as exhaust outlets and under the vehicle) to meet the gas flow distribution is calculated.

[0102] Perform safety checks on the target control parameters for each time period to determine whether they are simultaneously satisfied:

[0103] Target gas flow rate Within the safe operating flow range of the corresponding burner and proportional valve;

[0104] Target combustion air volume Within the capacity and adjustment range of the corresponding fan and damper;

[0105] Target pressure value Within the safe pressure limits of the kiln body structure;

[0106] The rate of change of each target control parameter between adjacent time periods did not exceed the equipment's allowable adjustment rate limit;

[0107] If all conditions are met, the security check passes and the target control parameters for each time period are output.

[0108] Otherwise, if the safety verification fails, the roasting control will end, and the user will be notified that the safety verification failed and the roasting control failed.

[0109] Based on the above scheme, the target control parameters for each time period are calculated according to the characteristics of the current optimal roasting curve. As an intelligent translator and collaborative planner between the production target and the lower-level roasting control, its core function is to accurately and consistently convert the upper-level abstraction of the current optimal roasting curve (temperature target) into the target control parameters (a multi-parameter collaborative setpoint sequence) required by the various actuators at the lower level (gas valves, air valves, fans, etc.), including the target temperature for each zone. Target pressure value P at key locations, target gas flow rate for each zone Target combustion air volume for each zone This ensures that macro-level production goals are accurately and collaboratively executed. It guarantees global control consistency by using a simplified network model of gas flow within the kiln (abstracting the kiln body, pipes, and fans as resistance elements and power sources). By solving the nonlinear equations describing the pressure-flow relationship in the pipeline network, it calculates the target pressure values ​​required at key locations within the kiln (such as exhaust vents and the bottom of the kiln) to meet the flow distribution. This ensures that airflow occurs as needed, guaranteeing that all target control parameters serve the same heat target, eliminating the possibility of conflicts arising from inconsistent targets in different control loops. It enables feedforward precision setting, providing the target trajectory of control parameters for a future period, offering clear and advanced tracking targets for underlying actuators, exhibiting foresight, and significantly reducing lag and overshoot. It improves response speed and stability, with smooth changes in target control parameters conforming to physical laws, avoiding sudden command changes that may occur with manual settings or independent PID control, resulting in smoother actuator operation, longer equipment lifespan, and safer processes. It provides an interface for advanced optimization, and the target control parameters provide a clear optimization objective for subsequent prediction models and optimization functions.

[0110] This solution addresses the disconnect between calcination curve optimization and control execution in existing technologies: traditional methods often employ fixed calcination curves or rely on manual experience for fine-tuning, with the optimization process independent of real-time control execution; even when an optimized theoretical calcination curve is obtained, control execution still depends on the independent operation of each control loop (temperature, pressure, air-fuel ratio), failing to guarantee precise and coordinated execution of the optimized theoretical calcination curve, resulting in "optimization not being implemented"; and the lack of a coordinated mechanism for multi-parameter control oriented towards global objectives: existing control methods often use independent PID control loops or simple decoupling control to regulate multiple parameters such as temperature, pressure, gas flow, and combustion air volume; when changes occur (such as calcination curve adjustment and optimization), there is a lack of a unified coordinated strategy among the control loops to allocate control quantities, easily leading to control conflicts (for example, increasing air volume for rapid heating disrupts kiln pressure stability), causing system oscillations or forcing the sacrifice of some calcination indicators.

[0111] like Figure 4 As shown, further, the target control parameter for the current time period is taken as the objective, based on external disturbance variables. and the current control status of each partition The optimal collaborative control command is obtained through rolling optimization adjustments using a predictive model and optimization function. Based on this optimal command, the tunnel kiln is then subjected to roasting control for the current time period. Specifically, this includes:

[0112] The external disturbance variables and the current control state of each partition are input into the predictive model to obtain the predictive control state of each partition. and predictive control parameters for each partition ;

[0113] The prediction model is:

[0114] ;

[0115] ;

[0116] In the formula, This represents the current control state estimate after correction based on measurements over time period k. Represents external disturbance variables. This represents the current optimal cooperative control command to be solved in the current k-time period, where A, B, and D represent the state transition matrix, control input matrix, and disturbance input matrix, respectively. Represents the output matrix; A describes the impact of internal system dynamics on the predictive control state; B describes... Impact on predictive control state; D description Impact on predictive control state; C describes the predictive control state. To predictive control parameters The mapping relationship;

[0117] Predictive control status for each partition and predictive control parameters for each partition Collision detection includes:

[0118] Calculate the tracking error of the predictive control parameters of each partition in time period k compared to the predictive control parameters of each partition in time period k-1;

[0119] If the tracking errors of multiple controlled variables (temperature, pressure, and possibly kiln hot body temperature) in the predictive control parameters are all positive and all exceed the preset threshold; or if the tracking errors of multiple controlled variables include both positive and negative values ​​(for example, the temperature tracking error increases in the positive direction while the pressure tracking error increases in the negative direction), then it is determined to be a control conflict, and the symmetric positive definite weight matrix in the optimization function is dynamically adjusted.

[0120] Otherwise, it is determined that no control conflict has occurred;

[0121] Based on the target control parameters for the current time period and the predictive control parameters for each zone, the minimum cost is calculated using an optimization function, and the current optimal cooperative control command corresponding to the minimum cost is obtained. ;

[0122] The optimized function is: ;

[0123] The constraints of the optimization function are: ; ;

[0124] In the formula, N represents in The total number of time periods for which prediction is performed. Indicates in Time period prediction Forecast control parameters for a time period express Target control parameters for the time period , express Time period prediction The control increment relative to the previous time period is used to smooth control actions. express Time period prediction Time-based coordinated control instructions, express Time period prediction Time-based coordinated control instructions, To preset the threshold range of control commands, To preset the control increment threshold range, Let Q and R represent the squares of the weighted Euclidean norm, respectively; Q and R represent the symmetric positive definite weight matrices; Q and R penalize the tracking error and the change in control increment, respectively.

[0125] Based on the current optimal cooperative control command Control the roasting of the tunnel kiln for the current time period;

[0126] Determine whether the current time period is the last time period for the tunnel kiln firing control;

[0127] If it is the last time period, a prompt message indicating that the roasting control is complete will be output to the user;

[0128] If it is not the last time period, then obtain the actual control parameters of each partition in real time. And calculate its predictive control parameters compared to the previous time period. Current prediction error Based on the current prediction error The prediction model is updated in real time.

[0129] Furthermore, the statement based on the current prediction error The prediction model is updated in real time, specifically including:

[0130] Current prediction error Input state observer, estimate of the current control state. Real-time updates are performed using the following formula:

[0131] ;

[0132] In the formula, Represents the observer gain matrix. This represents a predicted value for the current state based on information from the previous time period.

[0133] Based on the current prediction error Driving parameter estimation algorithms (such as recursive least squares) adjust the parameters in the state transition matrix A, control input matrix B, and disturbance input matrix D of the prediction model (online fine-tuning).

[0134] Based on the above scheme, firstly, the estimated value of the current control state is corrected by a state observer, and the prediction model is updated in real time to ensure that the prediction for the next time period is based on a more accurate actual state. Secondly, for slow time-varying processes, the state transition matrix A, control input matrix B, and disturbance input matrix D in the prediction model can be fine-tuned online using a parameter estimation algorithm; where A, B, and D respectively describe the influence of the system's internal dynamics on the predicted control state, ... Impact on predictive control state The parameterized relationships affecting the predictive control state include, for example, matrix A containing parameters describing temperature inertia and pressure transmission characteristics, matrix B containing parameters describing the influence coefficients of valve opening on temperature and pressure, and matrix D containing parameters describing the influence of disturbances such as kiln car advance speed, ambient temperature, and fuel gas calorific value on the state. By continuously fine-tuning these parameters, the predictive model gradually approximates the dynamic characteristics of the actual object.

[0135] Furthermore, the dimension of the symmetric positive definite weight matrix Q is the same as the dimension of the target control parameters, and its diagonal elements... These correspond to the target gas flow rate in the target control parameters. Target combustion air volume Target temperature The penalty weight for tracking error of the target pressure value P. The symmetric positive definite weight matrix R has the same dimension as the current optimal cooperative control command, and its diagonal elements... These correspond to the opening commands of the natural gas proportional control valve, the opening commands of the combustion air control valve, and the frequency control commands of the inverters for the key fans, which include the exhaust fan, the cooling fan, and the under-vehicle balancing fan.

[0136] Furthermore, the symmetric positive definite weight matrix is ​​dynamically adjusted, specifically as follows:

[0137] Based on the current process stage or control conflicts detected in real time, modify the corresponding diagonal elements in the Q matrix. The value.

[0138] Based on the above scheme, intelligent priority management can be achieved through dynamic arbitration. When different target control parameters conflict irreconcilably due to physical limits or strong disturbances (such as rapid heating versus pressure stabilization), dynamic adjustments can be initiated according to the current process stage (such as heating stage or holding stage) and preset priority rules. This involves adjusting the weights of corresponding diagonal elements in the Q matrix to achieve strategic compromise, ensuring the achievement of core process objectives, and arbitrating and resolving conflicts. For example, in the rapid heating stage, the weights of the diagonal elements corresponding to temperature tracking errors in the Q matrix are increased, while the weights of the diagonal elements corresponding to pressure tracking errors are appropriately decreased; in the holding stage, the balance between the two is restored. This achieves rule-based multi-objective dynamic priority management.

[0139] The above scheme takes the target control parameters for each time period as the objective, and based on the current control state and external disturbance variables, performs rolling optimization adjustments through a predictive model and optimization function to obtain the current optimal cooperative control command. Based on this optimal cooperative control command, the tunnel kiln is controlled for the current time period for roasting. As a bottom-level real-time cooperative execution and conflict resolution method, its core function is to receive the target control parameters. Faced with complex multivariate coupling, disturbances, and physical constraints in actual processes, it calculates and outputs the current optimal cooperative control command in real time through a predictive model, optimization function, and dynamic arbitration, enabling the mutual coordination of various target control parameters for roasting control of the tunnel kiln during the current time period. It actively handles the problem of strong multivariate coupling. The predictive model can explicitly handle the mutual influence between loops such as temperature, pressure, gas flow, and combustion air volume. Through the optimization function, it actively compensates for coupling effects, fundamentally solving the problem of "paying attention to one thing but losing another" in traditional single-loop PID control. It can explicitly handle constraints, ensuring safe operation. , As constraints on the optimization function, it ensures that the calculated optimal coordinated control commands remain within a safe and executable range, avoiding the risk of exceeding limits. It exhibits strong robustness, re-optimizing based on the target control parameters at each time interval. This rolling "plan-execute-correction" mode provides strong suppression and robustness against preset model errors and unforeseen disturbances, maintaining long-term stable operation. The current optimal coordinated control commands are directly operable commands, including the opening commands of the natural gas proportional regulating valve, the opening commands of the combustion air regulating valve, and the frequency control commands of the inverters for key fans, including the exhaust fan, cooling fan, and under-vehicle balancing fan.

[0140] It should be understood that although the steps in the flowchart are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the diagram may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed in the same time period, but can be executed in different time periods. The execution order of these sub-steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the sub-steps or stages of other steps.

[0141] According to a second aspect of this application, a tunnel kiln roasting control system is provided, including modules for implementing the tunnel kiln roasting control method as described above.

[0142] like Figure 5 As shown, specifically, the control system includes:

[0143] The acquisition module 10 is used to acquire production targets, current comprehensive status characteristics, and historical optimal roasting curve characteristics; the current comprehensive status characteristics include the primary characteristics of the raw materials and real-time operating conditions;

[0144] The partitioning module 20 is used to divide the kiln body of the tunnel kiln into multiple partitions along the length direction and obtain the partitioning conditions of each partition.

[0145] The optimization training module 30 is used to optimize and train the model based on production targets, current comprehensive status characteristics, partition conditions of all partitions and historical best roasting curve characteristics to obtain the current best roasting curve characteristics.

[0146] The target control parameter calculation module 40 is used to calculate the target control parameters for each time period based on the characteristics of the current optimal roasting curve.

[0147] The roasting control module 50 is used to control the roasting of the tunnel kiln according to time periods, specifically including:

[0148] Real-time acquisition unit 501 is used to acquire external disturbance variables in real time. and the current control status of each partition ;

[0149] The calcination control unit 502 is used to target the control parameters for the current time period based on external disturbance variables. and the current control status of each partition The optimal coordinated control command is obtained through rolling optimization adjustments using a predictive model and optimization function. Based on this optimal command, the tunnel kiln is then subjected to calcination control for the current time period. The target control parameters for the time period are as follows:

[0150] like Figure 6 As shown, more specifically, the optimized training module 30 includes:

[0151] Construction unit 301 is used to construct the comprehensive benefit function and adjust it according to production targets;

[0152] Initialization unit 302 is used to take the real-time operating condition and the partition operating conditions of all partitions as the initial state features. The historical best roasting curve feature is used as the first roasting curve feature. ;

[0153] Execution unit 303 is used to execute the following steps starting from t=1:

[0154] Simulation unit 3031 is used to process the (t-1)th state feature Input into the simulation module, and based on the characteristics of the t-th calcination curve. Simulation was performed to obtain the t-th state feature. and the t-th virtual parameter;

[0155] The optimization parameter calculation unit 3032 is used to calculate the t-th optimization parameter based on the t-th virtual parameter;

[0156] The reward value calculation unit 3033 is used to input the t-th optimization parameter into the comprehensive benefit function and calculate the t-th reward value. ;

[0157] The first judgment unit 3034 is used to judge the t-th reward value. Whether the difference between the reward value and the previous n reward values ​​is less than a preset threshold, or determine... Has the preset number of iterations been reached?

[0158] Training optimization unit 3035 is used to, if not, then the t-th state feature The input is used for optimization training in the optimization model to obtain the calcination curve features of the (t+1)th time. And repeat the above steps;

[0159] Curve feature acquisition unit 3036, used to determine if the reward value is t. Select the maximum value from the first n reward values, then take the maximum value as the optimal reward value, and take the roasting curve feature corresponding to the optimal reward value as the current optimal roasting curve feature.

[0160] like Figure 7 As shown, specifically, the target control parameter calculation module 40 includes:

[0161] Theoretical net heating power calculation unit 401 is used to obtain the second characteristic of the raw material, including specific heat capacity. and density Based on the secondary properties of the raw materials and the kiln car advance speed Based on the characteristics of the current optimal roasting curve Calculate the theoretical net heating power required for each zone in each time period. ;

[0162] Input heat calculation unit 402 is used to calculate the theoretical net heating power required by each zone in each time period. and the estimated total heat loss of each zone in each time period The required input heat to be added to the kiln body for each zone during each time period was calculated. , ;

[0163] The target control parameter calculation unit 403 is used to calculate the input heat of each zone within each time period. The target control parameters for each time period were calculated. ;

[0164] in, , Indicates gas flow rate, Indicates the burner efficiency. This indicates the current lower heating value of the fuel. , Indicates the combustion air volume. This represents the optimal air-fuel ratio. The stoichiometric constant representing the theoretical amount of air required for the complete combustion of a unit of fuel gas. Indicates the target temperature. This indicates the target pressure value at a critical location;

[0165] Verification unit 404 is used to perform safety verification on the target control parameters for each time period to determine whether the following conditions are met simultaneously:

[0166] Target gas flow rate Within the safe operating flow range of the corresponding burner and proportional valve;

[0167] Target combustion air volume Within the capacity and adjustment range of the corresponding fan and damper;

[0168] Target pressure value Within the safe pressure limits of the kiln body structure;

[0169] The rate of change of each target control parameter between adjacent time periods did not exceed the equipment's allowable adjustment rate limit;

[0170] The first output unit 405 is used to output the target control parameters for each time period if the security check passes when all conditions are met simultaneously. It is also used to terminate the roasting control if the safety verification fails, and output the information that the safety verification failed and the roasting control failed to the user.

[0171] like Figure 8 As shown, specifically, the calcination control unit 502 includes:

[0172] Prediction unit 5021 is used to input external disturbance variables and the current control state of each partition into the prediction model to obtain the predictive control state of each partition. and predictive control parameters for each partition ;

[0173] The control command acquisition unit 5022 is used to calculate the minimum cost value through an optimization function based on the target control parameters of the current time period and the predicted control parameters of each partition, and obtain the current optimal cooperative control command corresponding to the minimum cost value.

[0174] The current roasting unit 5023 is used to control the roasting of the tunnel kiln for the current time period according to the current optimal cooperative control command;

[0175] The second judgment unit 5024 is used to determine whether the current time period is the last time period of the tunnel kiln roasting control;

[0176] The second output unit 5025 is used to output a prompt message to the user indicating that the roasting control is complete if it is the last time period.

[0177] Update unit 5026 is used to obtain the actual control parameters of each partition in real time if it is not the last time period. And calculate its predictive control parameters compared to the previous time period. Current prediction error Based on the current prediction error The prediction model is updated in real time.

[0178] According to a third aspect of this application, an electronic device is provided, comprising:

[0179] Memory;

[0180] Processor; and

[0181] Computer programs;

[0182] The computer program is stored in the memory and configured to be executed by the processor to implement the tunnel kiln firing control method as described above.

[0183] According to a fourth aspect of this application, a computer-readable storage medium is provided having a computer program stored thereon; the computer program is executed by a processor to implement the tunnel kiln roasting control method as described above.

[0184] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. The solutions in the embodiments of this application can be implemented in various computer languages, such as C, VHDL, Verilog, the object-oriented programming language Java, and the interpreted scripting language JavaScript.

[0185] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0186] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0187] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0188] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0189] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A method for controlling the firing process in a tunnel kiln, characterized in that, include: The production target, current comprehensive status characteristics, and historical optimal roasting curve characteristics are obtained; the current comprehensive status characteristics include the primary characteristics of the raw materials and real-time operating conditions. The tunnel kiln body is divided into multiple zones along its length, and the operating conditions of each zone are obtained. Based on production targets, current overall status characteristics, operating conditions of all zones, and historical best roasting curve characteristics, the current best roasting curve characteristics are obtained through optimization training using an optimization model. Calculate the target control parameters for each time period based on the current optimal roasting curve characteristics; The firing process in the tunnel kiln is controlled according to time periods, specifically including: Real-time acquisition of external disturbance variables and the current control status of each partition ; Using the target control parameters for the current time period as the objective, based on external disturbance variables and the current control status of each partition The optimal collaborative control command is obtained by rolling optimization adjustment through prediction model and optimization function, and the calcination control of tunnel kiln for the current time period is carried out according to the optimal collaborative control command.

2. The tunnel kiln roasting control method according to claim 1, characterized in that, The optimal roasting curve features are obtained by optimizing the model based on production targets, current overall status characteristics, zoning conditions of all zones, and historical best roasting curve features. Specifically, these features include: Construct a comprehensive benefit function and adjust it according to production targets; The real-time operating condition and the operating conditions of all partitions are used as the initial state features. The historical best roasting curve feature is used as the first roasting curve feature. ; Starting from t=1, perform the following steps: The (t-1)th state feature Input into the simulation module, and based on the characteristics of the t-th calcination curve. Simulation was performed to obtain the t-th state feature. and the t-th virtual parameter; The t-th optimized parameter is calculated based on the t-th virtual parameter; The t-th optimization parameter is input into the comprehensive benefit function to calculate the t-th reward value. ; Determine the t-th reward value Whether the difference between the reward value and the previous n reward values ​​is less than a preset threshold, or determine... Has the preset number of iterations been reached? If not, then the t-th state feature The input is used for optimization training in the optimization model to obtain the calcination curve features of the (t+1)th time. And repeat the above steps; If so, then start from the t-th reward value Select the maximum value from the first n reward values, then take the maximum value as the optimal reward value, and take the roasting curve feature corresponding to the optimal reward value as the current optimal roasting curve feature.

3. The tunnel kiln roasting control method according to claim 1, characterized in that, The calculation of target control parameters for each time period based on the current optimal roasting curve characteristics specifically includes: Obtain the second property of the raw material, including specific heat capacity. and density Based on the secondary properties of the raw materials and the kiln car advance speed Based on the characteristics of the current optimal roasting curve Calculate the theoretical net heating power required for each zone in each time period. ; Based on the theoretical net heating power required by each zone in each time period and the estimated total heat loss of each zone in each time period The required input heat to be added to the kiln body for each zone during each time period was calculated. , ; Input heat in each zone during each time period The target control parameters for each time period were calculated. ; in, , Indicates gas flow rate, Indicates the burner efficiency. This indicates the current lower heating value of the fuel. , Indicates the combustion air volume. This represents the optimal air-fuel ratio. The stoichiometric constant representing the theoretical amount of air required for the complete combustion of a unit of fuel gas. Indicates the target temperature. This indicates the target pressure value at a critical location; Perform safety checks on the target control parameters for each time period to determine whether they are simultaneously satisfied: Target gas flow rate Within the safe operating flow range of the corresponding burner and proportional valve; Target combustion air volume Within the capacity and adjustment range of the corresponding fan and damper; Target pressure value Within the safe pressure limits of the kiln body structure; The rate of change of each target control parameter between adjacent time periods did not exceed the equipment's allowable adjustment rate limit; If all conditions are met, the security check passes and the target control parameters for each time period are output. Otherwise, if the safety verification fails, the roasting control will end, and the user will be notified that the safety verification failed and the roasting control failed.

4. The tunnel kiln roasting control method according to claim 1, characterized in that, The target control parameter for the current time period is taken as the objective, based on external disturbance variables. and the current control status of each partition The optimal collaborative control command is obtained through rolling optimization adjustments using a predictive model and optimization function. Based on this optimal command, the tunnel kiln is then subjected to calcination control for the current time period. Specifically, this includes: The external disturbance variables and the current control state of each partition are input into the predictive model to obtain the predictive control state of each partition. and predictive control parameters for each partition ; Based on the target control parameters for the current time period and the predicted control parameters for each partition, the minimum cost value is calculated through an optimization function, and the current optimal cooperative control command corresponding to the minimum cost value is obtained. The tunnel kiln is controlled for roasting during the current time period based on the current optimal collaborative control command; Determine whether the current time period is the last time period for the tunnel kiln firing control; If it is the last time period, a prompt message indicating that the roasting control is complete will be output to the user; If it is not the last time period, then obtain the actual control parameters of each partition in real time. And calculate its predictive control parameters compared to the previous time period. Current prediction error Based on the current prediction error The prediction model is updated in real time.

5. The tunnel kiln roasting control method according to claim 4, characterized in that, The prediction model is as follows: ； ； In the formula, This represents the current control state estimate after correction based on measurements over time period k. Represents external disturbance variables. This represents the current optimal cooperative control command to be solved in the current k-time period, where A, B, and D represent the state transition matrix, control input matrix, and disturbance input matrix, respectively. Indicates the output matrix; Based on the current prediction error The prediction model is updated in real time, and the update formula is as follows: ； In the formula, Represents the observer gain matrix. This represents the predicted value of the current state based on information from the previous period; Based on the current prediction error The driving parameter estimation algorithm adjusts the elements in the state transition matrix A, control input matrix B, and disturbance input matrix D in the prediction model.

6. The tunnel kiln roasting control method according to claim 4, characterized in that, The optimization function is: ; The constraints of the optimization function are: ; ; In the formula, N represents in The total number of time periods for which prediction is performed. Indicates in Time period prediction Forecast control parameters for a time period express Target control parameters for the time period , express Time period prediction The control increment of the time period relative to the previous time period express Time period prediction Time-based coordinated control instructions, express Time period prediction Time-based coordinated control instructions, Let Q and R represent the squares of the weighted Euclidean norm, respectively. Let Q and R represent the symmetric positive definite weight matrices, respectively penalizing the tracking error and the change in control increment.

7. The tunnel kiln roasting control method according to claim 2, characterized in that, The t-th virtual parameter includes the t-th product quality. The t-th virtual energy consumption The t-th virtual period and the fluctuation data of the t-th process ; The t-th optimization parameter includes the t-th quality deviation. The t-th cumulative energy consumption The t-th production time and the t-th stability penalty ; The calculation of the t-th optimized parameter based on the t-th virtual parameter specifically includes: Based on the quality of the t-th product The t-th quality deviation was calculated. The calculation formula is: ,in, This represents the quality index of the m-th product in the t-th product quality obtained from the simulation. This represents the target value of the m-th product quality indicator. This represents the weight value of the m-th product quality indicator; The t-th virtual energy consumption The t-th virtual period These are respectively used as the t-th cumulative energy consumption The t-th production time Specifically: make ; make ; Fluctuation data from the t-th process The stability penalty of the t-th term is calculated. The calculation formula is: ； In the formula, This represents the fluctuation data of the t-th process during the simulation. The sequence of the j-th key controlled parameter changing over time. This represents the variance calculation function. This represents the weight of the j-th key controlled parameter.

8. A tunnel kiln firing control system, characterized in that, It includes a module for implementing the tunnel kiln firing control method as described in any one of claims 1 to 7.

9. An electronic device, characterized in that, include: Memory; processor; as well as Computer programs; The computer program is stored in the memory and configured to be executed by the processor to implement the tunnel kiln firing control method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, It stores a computer program; the computer program is executed by a processor to implement the tunnel kiln firing control method as described in any one of claims 1 to 7.

Images