A kind of ceramic tile drying kiln temperature and humidity coordination precision control system and method

Abstract

The application discloses a kind of ceramic tile drying kiln temperature and humidity coordination accurate regulation and control system and method, it is related to adaptive control technical field, including: obtaining the real-time temperature sequence, real-time humidity sequence of the multiple drying sections of ceramic tile drying kiln along the direction of conveyance division and the moisture content time series data of current batch ceramic tile body;Calculate the target moisture evaporation rate of each drying section;Construct temperature and humidity coordination control feasible region;Generate inter-zone temperature and humidity transfer delay compensation vector;Solve the rolling optimization problem of multivariable model predictive controller, output the hot air temperature set value and relative humidity set value of each drying section;Relative humidity set value is sent to the steam injection valve opening degree execution unit of corresponding drying section;Collect actual kiln moisture content;Based on the deviation of actual kiln moisture content and target moisture content, the temperature and humidity coordination control feasible region of next batch drying process is adaptively contracted and adjusted;The application improves ceramic tile body drying uniformity.

Description

Technical Field

[0001] This invention relates to the field of adaptive control technology, specifically to a system and method for precise control of temperature and humidity in a ceramic tile drying kiln. Background Technology

[0002] Drying the ceramic tile blanks is a core and critical process in ceramic tile production, directly determining the cracking rate, flatness, and overall quality of the finished tiles. Precise temperature and humidity control during the drying process is crucial for ensuring high yield rates. Currently, most ceramic tile drying kilns in China employ a segmented, fixed-parameter constant temperature and humidity control mode, relying on manual experience or a single temperature closed-loop control method to adjust the kiln conditions. This approach suffers from numerous technical shortcomings, including crude control methods, lack of variable coupling, and poor adaptability to different operating conditions. Existing technologies mostly only independently control the kiln temperature, neglecting the temperature of the drying section. The strong coupling relationship between temperature and humidity, without combining the real-time moisture content changes of the terracotta blanks with the drying kinetics to match the optimal evaporation rate, easily leads to problems such as excessively rapid drying in the early stage causing surface cracking of the blanks and incomplete drying in the later stage causing the finished product to have excessive moisture content. At the same time, the traditional control scheme does not consider the delayed characteristics of temperature and humidity transmission between multiple sections of the kiln. The temperature and humidity adjustment actions of the preceding section will lag and affect the working conditions of the subsequent section, resulting in asynchronous temperature and humidity control in each section and a lag in dynamic response. It cannot adapt to complex actual working conditions such as batch moisture content differences and conveyor belt speed fluctuations.

[0003] Furthermore, existing kiln control systems lack effective constraint optimization mechanisms and adaptive correction capabilities, and can only achieve passive adjustment of fixed parameters. They cannot construct a feasible domain for coordinated temperature and humidity control, and are prone to control conflicts such as over-limit heating power and overloaded steam valve opening. Moreover, there is no batch moisture content feedback closed-loop optimization logic, which cannot iteratively optimize subsequent drying strategies based on the moisture content deviation of the kiln-exit blanks. After long-term operation, batch drying stability is poor and the yield fluctuates greatly. In addition, traditional equipment relies on fixed control logic and has not designed adaptation mechanisms for abnormal scenarios such as kiln opening, equipment aging, and fluctuations in operating conditions. The consumption of computing resources is unreasonable and the equipment operation and maintenance early warning capabilities are lacking, making it difficult to meet the needs of large-scale, high-precision, and high-stability modern ceramic tile drying production. Summary of the Invention

[0004] To address the aforementioned technical problems, a system and method for precise temperature and humidity control in ceramic tile drying kilns are provided. This technical solution resolves the problems described above.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A method for precise and coordinated temperature and humidity control in a ceramic tile drying kiln includes: Acquire real-time temperature and humidity sequences of multiple drying zones along the conveying direction in the ceramic tile drying kiln, as well as time-series data of the moisture content of the current batch of ceramic tile blanks; Based on the moisture content time series data and the preset drying kinetics model, the target moisture evaporation rate of each drying section is calculated. Based on the target moisture evaporation rate, and combined with the upper limit of the hot air circulation fan speed and the upper limit of the steam injection valve opening in each drying section, a feasible domain for coordinated temperature and humidity control is constructed. Extract the historical temperature and humidity response delay features of each drying section and generate a temperature and humidity transfer delay compensation vector between sections. Based on the feasible domain of the temperature and humidity coordinated control and the temperature and humidity transfer delay compensation vector between the sections, the rolling optimization problem of the multivariable model predictive controller is solved, and the hot air temperature setpoint and relative humidity setpoint of each drying section are output. The hot air temperature setpoint is sent to the electric heater power adjustment unit of the corresponding drying section, and the relative humidity setpoint is sent to the steam jet valve opening execution unit of the corresponding drying section. Infrared moisture detectors were deployed at the outlets of each drying section to collect the actual moisture content of the kiln exit. Based on the deviation between the actual moisture content exiting the kiln and the target moisture content, an adaptive shrinkage adjustment is made to the feasible range of temperature and humidity coordinated control for the next batch drying process.

[0006] Furthermore, the calculation of the target moisture evaporation rate for each drying section includes: Obtain the initial moisture content and target final moisture content of the current batch of ceramic tile blanks; Based on the real-time position of the ceramic tile blanks in the kiln and the speed of the conveyor belt, the estimated dwell time of the ceramic tile blanks at the exit of each drying section is calculated. A one-dimensional unsteady diffusion equation is constructed using Fick's second law. The average moisture evaporation flux required for each drying section is then calculated by inversion using the expected residence time as the boundary condition. The average moisture evaporation flux is then converted into the target moisture evaporation rate per unit time and per unit area.

[0007] Furthermore, the feasible domain for constructing coordinated temperature and humidity control includes: A dual-input single-output mapping relationship between hot air temperature and relative humidity and the rate of water evaporation is established, and the mapping relationship is obtained through offline identification. The hot air temperature setpoint and relative humidity setpoint are used as decision variables, the target moisture evaporation rate is used as an equality constraint, and the maximum power of the electric heater and the maximum opening of the steam valve are used as inequality constraints. By introducing slack variables to handle constraint conflicts, a feasible region for coordinated temperature and humidity control with soft constraints is constructed.

[0008] Furthermore, the generation of the inter-segment temperature and humidity transfer delay compensation vector includes: Record the time delay of the humidity sensor response in the next drying section to reach steady state after the hot air temperature of the previous drying section changes abruptly under historical operating conditions. Sliding window statistics were performed on multiple historical delay samples to calculate the mean and standard deviation; When the kiln wind speed deviates from the calibrated wind speed under the current operating conditions, the mean value is dynamically corrected based on the inverse relationship between wind speed and delay. The corrected mean delay is used as the transmission delay compensation value for the corresponding segment pair, and combined to form the inter-segment temperature and humidity transmission delay compensation vector.

[0009] Furthermore, solving the rolling optimization problem of the multivariable model predictive controller includes: A coupled state-space equation containing a temperature sub-model and a humidity sub-model is constructed. The state variables of the state-space equation are the temperature and humidity of each section, and the control inputs are the heating power and the steam flow rate. In the prediction time domain, the inter-segment temperature and humidity transfer delay compensation vector is embedded into the state transition matrix to achieve cross-segment dynamic coupling modeling. Define a performance index function, which includes a moisture evaporation tracking error term, a control energy consumption term, and a temperature and humidity overshoot penalty term. The optimal control sequence that satisfies the feasible region constraint of the temperature and humidity coordinated control is solved online using a sequential quadratic programming algorithm.

[0010] Furthermore, before acquiring the real-time temperature and humidity sequences of multiple drying sections divided along the conveying direction in the ceramic tile drying kiln, the following steps are also included: Detect the closing status signal of the kiln's sealing door; When the closed state signal indicates that the sealing door is open, the transmission of new hot air temperature setpoints and relative humidity setpoints is paused, and the emergency balance mode of the kiln environment is activated. The emergency balance mode of the kiln environment includes setting the exhaust fan control signal to the maximum opening command and closing all steam injection valves.

[0011] Furthermore, after adaptively shrinking and adjusting the feasible domain of temperature and humidity coordinated control for the next batch drying process based on the deviation between the actual moisture content exiting the kiln and the target moisture content, the process also includes: Calculate the standard deviation of the actual moisture content exiting the kiln for three consecutive batches; When the standard deviation is less than the preset stability threshold, the kiln is determined to have entered the steady-state operation stage. During the steady-state operation phase, the prediction time domain length of the multivariable model predictive controller is reduced to decrease the computational load; When the standard deviation is greater than or equal to the stability threshold, the prediction time domain length is extended and the control frequency is increased.

[0012] Furthermore, after sending the hot air temperature setpoint to the power adjustment unit of the electric heater in the corresponding drying section, the method further includes: Monitor the actual power consumption of the electric heater; When the actual power consumption exceeds the theoretical power corresponding to the set value by more than 15% within five consecutive control cycles, an aging warning signal for the heating element is triggered and sent to the operation and maintenance management terminal.

[0013] Furthermore, before employing the sequential quadratic programming algorithm to solve online for the optimal control sequence that satisfies the feasible region constraint of the temperature and humidity coordinated control, the method further includes: Using a digital twin simulation platform, based on the current kiln structural parameters and material characteristics, a virtual working condition dataset covering different combinations of moisture content gradients and conveying speeds is generated. A lightweight proxy model is pre-trained on the virtual working condition dataset to initialize the initial solution of the sequential quadratic programming algorithm; The initial solution is input into the sequential quadratic programming algorithm to accelerate the convergence process.

[0014] A precise temperature and humidity control system for a ceramic tile drying kiln includes: The multi-source sensor data acquisition module acquires the real-time temperature sequence, real-time humidity sequence, and moisture content time series data of the current batch of ceramic tile blanks from multiple drying sections divided along the conveying direction in the ceramic tile drying kiln. The target evaporation rate calculation module calculates the target moisture evaporation rate for each drying section based on the moisture content time series data and the preset drying kinetic model. The temperature and humidity coordinated feasible domain construction module constructs a temperature and humidity coordinated control feasible domain based on the target moisture evaporation rate and combined with the upper limit of the hot air circulation fan speed and the upper limit of the steam injection valve opening in each drying section. The segment delay compensation vector generation module extracts the historical temperature and humidity response delay features of each drying segment and generates an inter-segment temperature and humidity transfer delay compensation vector. The multivariable rolling optimization solution module solves the rolling optimization problem of the multivariable model predictive controller based on the feasible domain of the temperature and humidity coordinated control and the temperature and humidity transfer delay compensation vector between the sections, and outputs the hot air temperature setpoint and relative humidity setpoint for each drying section. The actuator command sending module sends the hot air temperature setpoint to the electric heater power adjustment unit of the corresponding drying section, and sends the relative humidity setpoint to the steam jet valve opening execution unit of the corresponding drying section. The kiln exit moisture feedback detection module deploys infrared moisture detectors at the outlet of each drying section to collect the actual kiln exit moisture content; The feasible domain adaptive adjustment module adaptively shrinks and adjusts the feasible domain of temperature and humidity coordinated control for the next batch drying process based on the deviation between the actual moisture content at the kiln outlet and the target moisture content.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. This invention constructs a dynamic model of moisture diffusion in the green body based on Fick's second law, and accurately solves the target moisture evaporation rate of each drying section by combining the time series data of the green body moisture content and the residence time in the kiln, breaking the limitations of traditional fixed parameter control; by constructing a feasible domain for temperature and humidity coordinated control, and using the target evaporation rate as a constraint to achieve coordinated optimization matching of temperature and humidity dual variables, the optimal drying conditions can be dynamically adapted according to the moisture content and conveying speed of different batches of green bodies, effectively avoiding defects such as rapid drying and cracking of the green body surface and incomplete internal drying, and significantly improving the drying uniformity of ceramic tile green bodies and the yield of finished products; 2. This invention extracts the temperature and humidity transmission delay features of multiple zones and constructs a compensation vector, embedding the delay compensation mechanism into the state space model to accurately characterize the dynamic coupling and delayed transmission characteristics of temperature and humidity in multiple zones, solving the problem of operational imbalance caused by traditional independent temperature and humidity control without delay compensation; it solves the rolling optimization problem with soft constraints by using a sequential quadratic programming algorithm to solve the problem, taking into account drying accuracy, energy consumption control and overshoot suppression, realizing the forward-looking and refined dynamic control of temperature and humidity in multiple zones of the kiln, adapting to various dynamic operating condition fluctuations in the production process; 3. This invention achieves batch adaptive shrinkage adjustment within the controllable feasible domain based on the kiln exit moisture content deviation, forming a control system of acquisition, optimization, control, feedback, and iteration to continuously optimize the drying control strategy and effectively suppress batch production deviations. Simultaneously, it combines dynamic adjustment of the controller's prediction time domain with kiln operating stability to balance control accuracy and equipment computational load. It also includes functions such as emergency door opening balancing, heating element aging early warning, and digital twin pre-accelerated solution, solving the problems of poor fault tolerance and low optimization iteration efficiency in traditional systems, and improving the stability of kiln drying production. Attached Figure Description

[0016] Figure 1 This is a flowchart outlining the steps of the present invention. Figure 2 This is a system architecture diagram of the present invention. Detailed Implementation

[0017] The following description is intended to disclose the invention and enable those skilled in the art to implement it. The preferred embodiments described below are merely examples, and other obvious variations will occur to those skilled in the art.

[0018] Example 1: Refer to Figure 1 As shown, a method for precise control of temperature and humidity in a ceramic tile drying kiln includes: Step S100: Obtain the real-time temperature sequence, real-time humidity sequence, and moisture content time series data of the current batch of ceramic tile blanks from multiple drying sections divided along the conveying direction of the ceramic tile drying kiln. Step S200: Based on the time series data of moisture content and the preset drying kinetic model, calculate the target moisture evaporation rate of each drying section; Step S300: Based on the target moisture evaporation rate, and combined with the upper limit of the hot air circulation fan speed and the upper limit of the steam injection valve opening in each drying section, construct a feasible domain for coordinated temperature and humidity control. Step S400: Extract the historical temperature and humidity response delay features of each drying section and generate a temperature and humidity transfer delay compensation vector between sections. Step S500: Based on the feasible region of temperature and humidity coordinated control and the inter-segment temperature and humidity transfer delay compensation vector, solve the rolling optimization problem of the multivariable model predictive controller, and output the hot air temperature setpoint and relative humidity setpoint of each drying section. In step S600, the hot air temperature setting value is sent to the electric heater power adjustment unit of the corresponding drying section, and the relative humidity setting value is sent to the steam jet valve opening execution unit of the corresponding drying section. Step S700: Deploy infrared moisture detectors at the outlet of each drying section to collect the actual moisture content of the kiln outlet. Step S800: Based on the deviation between the actual moisture content exiting the kiln and the target moisture content, adaptive shrinkage adjustment is performed on the feasible domain of temperature and humidity coordinated control for the next batch drying process.

[0019] Through the above scheme, this application accurately matches the moisture evaporation demand of each drying section based on the drying dynamics of ceramic tile blanks, breaking the extensive mode of traditional fixed parameter control; by constructing a feasible domain for multi-variable temperature and humidity collaborative control, it achieves multi-constraint collaborative optimization of temperature, humidity, and equipment execution upper limit; combined with a multi-section temperature and humidity transmission delay compensation mechanism, it solves the problems of cross-section working condition lag and asynchronous control in tunnel kilns; relying on model prediction rolling optimization to achieve forward-looking dynamic parameter adjustment, and completing adaptive optimization of control strategy through batch moisture content feedback iteration, it effectively solves the problems of blank cracking, uneven drying, poor batch stability, and high energy consumption in traditional ceramic tile drying processes, and significantly improves the drying quality of ceramic tile blanks and the level of intelligent production.

[0020] In some embodiments, based on time-series moisture content data and a preset drying kinetics model, the target moisture evaporation rate for each drying zone is calculated, including: Obtain the initial moisture content and target final moisture content of the current batch of ceramic tile blanks; calculate the expected residence time of the ceramic tile blanks at the outlet of each drying section based on the real-time position of the ceramic tile blanks in the kiln and the speed of the conveyor belt; construct a one-dimensional unsteady diffusion equation using Fick's second law, and solve for the average moisture evaporation flux required for each drying section using the expected residence time as the boundary condition; convert the average moisture evaporation flux into the target moisture evaporation rate per unit time per unit area. Specifically, during the continuous drying process of ceramic tiles, the initial moisture content of different batches of green tiles fluctuates slightly. Under normal production conditions, the initial moisture content of the green tile green tiles is generally between 12% and 18%. The process requires that the target final moisture content after leaving the kiln be stably controlled within the range of 0.5% to 1.5%. Before production, the initial moisture content of the current batch of green tile green tiles is collected piece by piece by an online green tile moisture content detection device. The average initial moisture content of the batch is then used as the modeling benchmark, and the target final moisture content preset by the process system is retrieved as the convergence standard. The tunnel-type drying kiln is evenly divided into three independent temperature and humidity control zones along the material conveying direction: a preheating zone, a constant-speed drying zone, and a deceleration drying zone. The normal operating speed range of the production line conveyor belt is 0.2m / min to 0.6m / min. The system locates the coordinate position of the ceramic tile blanks in the kiln in real time. Combined with the real-time conveyor belt speed, it calculates the time nodes when a single blank enters and leaves each zone, accurately calculating the effective residence time of the blank in each drying zone. The normal residence time in a single zone is controlled within the range of 80s to 200s, providing accurate time boundary parameters for drying kinetic modeling. Moisture migration within terracotta tile blanks is a typical non-steady-state molecular diffusion process. Moisture dynamically changes with drying time and the thickness of the blank, perfectly conforming to Fick's second law of diffusion. This application constructs a one-dimensional non-steady-state moisture diffusion solution model based on this law, using the thickness of the blank as the diffusion dimension, the actual residence time of the blank in each drying section as the time boundary constraint, and the initial moisture content and target final moisture content of the batch as the upper and lower concentration boundary constraints. A finite-difference iterative calculation method is used to discretize the continuous moisture diffusion process into multiple short-time iterative processes, successively solving the moisture concentration distribution within the blank at each time step. Finally, the iterative convergence yields the average moisture evaporation flux required for the entire blank within a single drying cycle, which is the total dehydration per unit blank area within a single section. Based on this, the total evaporation flux is evenly distributed across the residence time of the entire section, converting it into the moisture evaporation per unit time and per unit blank surface area, which is the target moisture evaporation rate for each drying section. This provides a quantitative and implementable process constraint standard for subsequent temperature and humidity coordinated control.

[0021] In some embodiments, based on the target moisture evaporation rate, and in conjunction with the upper limit of the hot air circulation fan speed and the upper limit of the steam injection valve opening in each drying section, a feasible domain for coordinated temperature and humidity control is constructed, including: A dual-input single-output mapping relationship between hot air temperature and relative humidity and moisture evaporation rate is established, which is obtained through offline identification. The hot air temperature setpoint and relative humidity setpoint are used as decision variables, the target moisture evaporation rate is used as equality constraint, and the maximum power of the electric heater and the maximum opening of the steam valve are used as inequality constraints. Relaxation variables are introduced to handle constraint conflicts and a feasible region for temperature and humidity coordinated control with soft constraints is constructed. Specifically, the evaporation rate of moisture in the ceramic tile blank is determined by the coupled hot air temperature inside the kiln and the relative humidity of the environment. These two variables are interrelated and jointly affect the drying efficiency. This application uses an offline operating condition identification method to complete model training and parameter solidification. The specific training process is as follows: a stable operating condition range of normal kiln production is selected, with the temperature sampling range set to 40℃~85℃ and the relative humidity sampling range set to 20%~70%. No less than 200 sets of temperature and humidity combination samples are evenly selected, and the actual moisture evaporation rate of the blank corresponding to each set of operating conditions is collected simultaneously. Temperature and humidity are used as input features, and evaporation rate is used as output label. The least squares method is used for fitting combined with shallow neural network iterative training to continuously correct the model weight parameters. After training, the model parameters are solidified to obtain an accurate mapping relationship between temperature and humidity dual inputs and moisture evaporation rate single output, ensuring that the prediction error of the model in the normal operating condition range does not exceed ±3%. During the online construction of the feasible domain, the real-time setpoints for the hot air temperature and relative humidity of each drying section are defined as optimization decision variables. The target moisture evaporation rate obtained from the dynamic model is used as an equality constraint to ensure that the actual evaporation effect after regulation strictly matches the process requirements. At the same time, the equipment hardware limit inequality constraints are superimposed. The exemplary constraint thresholds are as follows: the maximum power constraint of a single section of electric heater is 0kW~30kW, the steam injection valve opening constraint is 0%~100%, and the hot air circulation fan speed constraint is 0r / min~1450r / min. All regulation variables are limited within the executable range of the equipment to avoid overload or regulation failure caused by exceeding the limits of the command. In actual production, there are disturbances such as fluctuations in the moisture content of the billet and fluctuations in wind pressure, which can easily lead to problems where multiple constraints conflict with each other and optimization solutions are infeasible. This application introduces a controllable relaxation variable to make the rigid constraints more flexible. For example, the allowable relaxation deviation range for the evaporation rate is set to ±2%, and the allowable relaxation deviation range for the equipment opening is set to ±3%. Without affecting the overall drying quality, the constraint boundaries are slightly relaxed, effectively avoiding unsolvable problems. Finally, a feasible domain for temperature and humidity collaborative control with flexible constraints is formed that is adapted to real-time operating conditions, providing a legal, solvable, and implementable solution space for subsequent rolling optimization.

[0022] In some embodiments, historical temperature and humidity response delay features of each drying section are extracted to generate an inter-section temperature and humidity transfer delay compensation vector, including: Record the time delay of the humidity sensor response in the next drying section to reach steady state after the hot air temperature of the previous drying section changes abruptly under historical operating conditions; perform sliding window statistics on multiple historical delay samples to calculate the mean and standard deviation; when the kiln wind speed deviates from the calibrated wind speed under the current operating conditions, dynamically correct the mean based on the inverse relationship between wind speed and delay; use the corrected mean delay as the transmission delay compensation value for the corresponding section pair, and combine them into an inter-section temperature and humidity transmission delay compensation vector; Specifically, the drying sections of the tunnel kiln are interconnected. The heating and humidification actions of the preceding section are transmitted to the subsequent section through hot air convection, resulting in a significant delay in the transmission of operating conditions. The delay time is usually in the range of 2s to 15s, which is the core factor causing asynchronous control and overshoot of the operating conditions in multiple sections. This application completes the collection of delayed samples through historical step response testing. The specific process is as follows: a fixed temperature step increment is applied to the preceding drying section. For example, the step increment is set to 5℃ to 10℃. The humidity sensor data of the subsequent section is collected at high frequency throughout the process. The timing starts from the moment the step is applied and continues until the fluctuation of the subsequent humidity data converges and the value change is less than ±0.5%RH for 10 consecutive sampling cycles. The operating condition is then determined to have reached a steady state, and the duration of this segment is counted as a single set of delayed samples. The system uses a fixed-length sliding window to complete sample statistics. For example, the sliding window sample capacity is set to 30 groups. It collects continuous historical delay samples in real time, removes outlier samples that are more than 3 times the standard deviation of the mean, and calculates the mean and standard deviation of the delay of the effective samples in the window. The mean is used as the basic delay parameter between sections. The hot air velocity in the kiln is strictly inversely proportional to the transmission delay. The higher the wind speed, the faster the hot air convection and the shorter the delay time. The standard wind speed calibrated by the equipment is 8 m / s. The system collects the actual wind speed in the kiln in real time. When the actual wind speed deviates from the standard wind speed by more than ±1 m / s, it automatically and dynamically corrects the basic delay mean according to the inverse mapping ratio to adapt to the real-time wind field conditions. The system iterates through all adjacent combinations of preheating-constant speed and constant speed-deceleration sections, and solves the dynamic correction delay value for each pair of sections. All delay compensation values ​​are arranged and combined in the order of kiln section arrangement to generate a multi-dimensional temperature and humidity transmission delay compensation vector containing coupling lag information of the entire section. This vector fully characterizes the lag characteristics of the entire kiln operation and is used for subsequent dynamic coupling modeling and accurate compensation of the controller.

[0023] In some embodiments, solving the rolling optimization problem of a multivariate model predictive controller includes: A coupled state-space equation consisting of a temperature sub-model and a humidity sub-model is constructed. The state variables of the state-space equation are the temperature and humidity of each segment, and the control inputs are heating power and steam flow rate. In the prediction time domain, the temperature and humidity transfer delay compensation vector between segments is embedded into the state transition matrix to achieve dynamic coupling modeling across segments. A performance index function is defined, which includes a moisture evaporation tracking error term, a control energy consumption term, and a temperature and humidity overshoot penalty term. The optimal control sequence that satisfies the feasible region constraint of temperature and humidity coordinated control is solved online using a sequential quadratic programming algorithm. Specifically, considering the strong coupling, large lag, and multivariate correlation of temperature and humidity in kilns, this application constructs a multi-input multi-output coupled state-space model. The real-time temperature and real-time relative humidity of each drying section are defined as system state variables, and the heating power of the electric heaters in each section and the steam supply flow rate corresponding to the steam injection valves are defined as system control input variables. Temperature dynamic response sub-model and humidity dynamic response sub-model are constructed respectively. The influence of temperature changes on humidity and the coupling interference of humidity changes on temperature are characterized by the coupling correlation matrix, realizing bivariate dynamic correlation modeling and accurately restoring the actual operating condition change law of the kiln. The controller is preset to a basic prediction time domain interval of 10 to 30 steps and a control cycle of 1 second. During the prediction time domain iterative calculation, the pre-generated inter-segment delay compensation vector is embedded into the system state transition matrix to correct the traditional ideal state iteration process without delay. The segment transmission lag effect is superimposed in each state iteration step to accurately restore the lag characteristics of cross-segment operating condition transmission and eliminate the control deviation and operating condition oscillation problems caused by lag. This application constructs a multi-objective comprehensive performance index function, which includes three core weight terms. The weight ratio can be adaptively adjusted according to the operating conditions. The conventional benchmark weight configuration is 0.6 for the evaporation tracking error term, 0.2 for the energy consumption term, and 0.2 for the overshoot penalty term. Among them, the moisture evaporation tracking error term is used to constrain the actual evaporation rate to fit the target process value, with an allowable error range of ±2%. The energy consumption control term is used to suppress the ineffective output of heating power and steam flow, so as to achieve energy saving and consumption reduction. The temperature and humidity overshoot penalty term is used to suppress drastic fluctuations in operating conditions. The allowable threshold for temperature overshoot is ≤±2℃, and the allowable threshold for humidity overshoot is ≤±3%RH, which effectively avoids the problems of billet cracking and uneven drying caused by overshoot. Using the feasible region of temperature and humidity coordinated control as the boundary constraint for all optimization variables, a sequential quadratic programming algorithm is used to complete the online iterative solution. During the algorithm iteration, the control variables are updated one by one to reduce the error of the objective function. When the change of the objective function is less than the preset convergence threshold of 0.001 for 5 consecutive iterations, the iteration is considered to have converged, and the optimal heating power and steam flow control sequence in the prediction time domain is output. Only the optimal control quantity at the current moment is extracted for execution, and the remaining sequences are used as the initial values ​​for subsequent rolling iterations, so as to achieve second-by-second rolling optimization and precise control.

[0024] In some embodiments, before acquiring the real-time temperature sequence and real-time humidity sequence of multiple drying sections divided along the conveying direction in the ceramic tile drying kiln, the method further includes: The system detects the closing status signal of the kiln sealing door; when the closing status signal indicates that the sealing door is open, it pauses the transmission of new hot air temperature setpoints and relative humidity setpoints, and activates the emergency balance mode of the kiln environment. The emergency balance mode of the kiln environment includes setting the exhaust fan control signal to the maximum opening command and closing all steam injection valves. Specifically, when the kiln's sealed door is opened, the high-temperature and high-humidity environment inside the kiln rapidly convections with the outside cold air, causing significant instantaneous fluctuations in temperature and humidity inside the kiln. Temperature fluctuations can reach ±8℃ or more, and humidity fluctuations can reach ±15%RH or more. If the controller continuously updates its control commands, it will cause the actuator to operate frequently and the operating conditions to become disordered, seriously affecting the equipment's lifespan and drying quality. The system collects the switch level signal of the sealed door in real time, determining the open state with a high level and the closed state with a low level, achieving millisecond-level state discrimination. When the sealed door is detected to be open, all temperature and humidity settings currently output by the controller are immediately locked, and all new command updates are suspended to avoid ineffective adjustments. At the same time, the emergency balancing mode is activated, and all exhaust fans are set to run at 100% maximum opening to quickly extract the turbulent hot and humid air in the kiln. Simultaneously, all section steam injection valves are closed to completely stop humidity supply and prevent moisture condensation from entering the kiln and causing the billet to become damp. The status of the sealed door is continuously monitored. Once the sealed door is fully closed and the temperature and humidity fluctuations in the kiln have fallen back to the stable range of ±1℃ and ±2%RH, the emergency balancing mode is automatically exited, and the normal collaborative control logic is restored to ensure the safe operation and stability of the equipment.

[0025] In some embodiments, after adaptively shrinking the feasible domain of temperature and humidity coordinated control for the next batch drying process based on the deviation between the actual moisture content exiting the kiln and the target moisture content, the method further includes: Calculate the standard deviation of the actual moisture content of three consecutive batches exiting the kiln; when the standard deviation is less than the preset stability threshold, the kiln is determined to have entered the steady-state operation stage; during the steady-state operation stage, reduce the prediction time domain length of the multivariate model predictive controller to reduce the computational load; when the standard deviation is greater than or equal to the stability threshold, extend the prediction time domain length and increase the control frequency. Specifically, the standard deviation of batch moisture content can accurately characterize the overall stability of the drying conditions. The preset stability threshold of this application is 0.15%, that is, when the standard deviation of moisture content of three consecutive batches is less than 0.15%, the operating conditions are judged to be stable; when it is greater than or equal to 0.15%, the operating conditions are judged to be significantly disturbed. The system stores the kiln moisture content detection data batch by batch. After completing three consecutive production batches, it automatically calculates the standard deviation of the three sets of data to complete the determination of the stability of the operating conditions. When the standard deviation is less than the stability threshold of 0.15%, it indicates that the batch drying consistency is good and there is no obvious disturbance, and the kiln enters the steady-state operation stage. At this time, the controller prediction time domain is reduced from the conventional 20 steps to 10 steps, simplifying the iterative calculation dimension. Without affecting the drying accuracy, the computing power consumption and operating load of the edge controller are greatly reduced, and the equipment operating efficiency is improved. When the standard deviation is greater than or equal to the stability threshold of 0.15%, it is determined that the moisture content of the billet fluctuates greatly and there are dynamic disturbances in the kiln air pressure, temperature and humidity. At this time, the prediction time domain is extended from 20 steps to 30 steps, and the control update frequency is increased from 1 second / time to 0.5 seconds / time to enhance the controller's advanced prediction capability and dynamic anti-interference capability, quickly suppress the fluctuation of operating conditions, and ensure the stability of batch drying quality.

[0026] In some embodiments, after sending the hot air temperature setpoint to the electric heater power regulation unit of the corresponding drying section, the method further includes: Monitor the actual power consumption of the electric heater; when the actual power consumption exceeds the theoretical power corresponding to the set value by more than 15% within five consecutive control cycles, trigger the heating element aging warning signal and send it to the operation and maintenance management terminal. Specifically, the system can accurately calculate the theoretical working power of the electric heater under the current operating conditions based on the current temperature setpoint, the ambient temperature inside the kiln, and real-time heat loss parameters; the system collects the actual active power of the electric heater in real time and compares the deviation ratio between the actual power and the theoretical power in each control cycle. This application sets a preset warning deviation threshold of 15%. When the actual operating power is continuously 15% or more higher than the theoretical power for 5 consecutive control cycles (1 second per cycle, lasting 5 seconds), the heating element is judged to have aging fault characteristics such as increased internal resistance and decreased heating efficiency. The system immediately generates a heating element aging warning signal, including the corresponding section number, power deviation data, and fault time, and uploads it to the on-site operation and maintenance management terminal via industrial Ethernet to remind staff to repair and replace it in time, so as to avoid heating failure and excessive temperature control deviation causing batch blank drying problems, and realize intelligent prediction and operation and maintenance of equipment status.

[0027] In some embodiments, before using a sequential quadratic programming algorithm to solve online for the optimal control sequence that satisfies the feasible region constraint of temperature and humidity coordinated control, the method further includes: Using a digital twin simulation platform, a virtual working condition dataset covering different combinations of moisture content gradients and conveying speeds is generated based on the current kiln structural parameters and material characteristics. A lightweight surrogate model is pre-trained on the virtual working condition dataset to initialize the initial solution of the sequential quadratic programming algorithm. The initial solution is then input into the sequential quadratic programming algorithm to accelerate the convergence process. Specifically, traditional sequential quadratic programming algorithms use random initial values ​​for iteration, which suffers from slow convergence, susceptibility to local optima, and lag in dynamic operating condition response. This application completes the preliminary modeling and dataset construction through a digital twin simulation platform. The specific process is as follows: the hardware structural parameters of the actual drying kiln, such as cavity dimensions, air duct structure, heating layout, sensor positions, and fan parameters, are replicated at a 1:1 scale. At the same time, material characteristic parameters such as the thickness of the terracotta body, porosity, and initial moisture content range are entered to build a virtual simulation kiln that is completely consistent with the actual site. Multiple gradient operating conditions are generated within the simulation platform, with moisture content gradients covering 12% to 18% and conveyor belt speed gradients covering 0.2 m / min to 0.6 m / min. No less than 500 sets of virtual operating condition data are generated by combining them in pairs to construct a fully covered virtual operating condition dataset. Offline training of the lightweight proxy model was completed based on a virtual working condition dataset. The model inputs were the moisture content of the billet, the conveyor belt speed, and the current temperature and humidity inside the kiln. The outputs were the heating power and steam valve opening corresponding to the optimal initial control solution. The training process adopted an iterative error backpropagation mechanism to continuously update the model weights and bias parameters until the error between the model's predicted output and the simulation optimal solution was less than 2%, at which point the model training was completed and the parameters were fixed. During online optimization, the lightweight proxy model was directly called to quickly output the optimal iterative initial solution based on real-time working conditions, replacing the traditional random initial values. This significantly improved the iterative convergence speed of the sequential quadratic programming algorithm, effectively avoided local optima, and met the high real-time and high-precision control requirements of industrial sites.

[0028] The second embodiment of this application provides a system for precise temperature and humidity control in a ceramic tile drying kiln. Example 2: Refer to Figure 2 As shown, a precise temperature and humidity control system for a ceramic tile drying kiln includes: The multi-source sensor data acquisition module acquires the real-time temperature sequence, real-time humidity sequence, and moisture content time series data of the current batch of ceramic tile blanks from multiple drying sections divided along the conveying direction in the ceramic tile drying kiln. The target evaporation rate calculation module calculates the target moisture evaporation rate for each drying section based on time-series moisture content data and a preset drying kinetics model. The temperature and humidity coordinated feasible domain construction module constructs a feasible domain for temperature and humidity coordinated control based on the target moisture evaporation rate and the upper limit of the hot air circulation fan speed and the upper limit of the steam injection valve opening in each drying section. The segment delay compensation vector generation module extracts the historical temperature and humidity response delay features of each drying segment and generates an inter-segment temperature and humidity transfer delay compensation vector. The multivariable rolling optimization solution module solves the rolling optimization problem of the multivariable model predictive controller based on the feasible region of temperature and humidity coordinated control and the temperature and humidity transfer delay compensation vector between sections, and outputs the hot air temperature setpoint and relative humidity setpoint for each drying section. The actuator command sending module sends the hot air temperature setpoint to the electric heater power adjustment unit of the corresponding drying section, and sends the relative humidity setpoint to the steam jet valve opening execution unit of the corresponding drying section. The kiln exit moisture feedback detection module deploys infrared moisture detectors at the outlet of each drying section to collect the actual kiln exit moisture content; The feasible domain adaptive adjustment module adaptively shrinks and adjusts the feasible domain of temperature and humidity coordinated control for the next batch drying process based on the deviation between the actual moisture content at the kiln outlet and the target moisture content. Specifically, the multi-source sensor data acquisition module consists of distributed high-precision temperature sensors, capacitive humidity sensors, and online billet infrared moisture content detectors. These are distributed at the air inlet, middle air outlet, and air outlet of each drying section of the kiln, respectively, to collect real-time operating data. The data is then denoised using a moving average filtering algorithm with a filter window length of 10 groups to eliminate abrupt interference data and ensure that the accuracy error of the collected data is less than ±1%. The target evaporation rate calculation module incorporates Fick's second law iterative solution logic and an adaptive boundary condition matching algorithm. It can automatically iteratively solve for the accurate target moisture evaporation rate based on the real-time billet moisture content and the residence time in each section, adapting to different batch production conditions. The temperature and humidity collaborative feasible domain construction module integrates an offline identification and training temperature and humidity-evaporation rate mapping model with flexible relaxation constraint logic, which can dynamically match real-time operating conditions to generate a legal and controllable solution space; the segment delay compensation vector generation module has the functions of historical sample statistics, outlier removal, dynamic wind speed correction, and multi-dimensional vector reconstruction, which accurately adapts to the large lag and cross-segment coupling characteristics of the kiln; the multivariable rolling optimization solution module integrates a coupled state space model, a multi-objective weight adjustable performance function, and an accelerated convergence optimization algorithm to achieve synchronous and precise optimization and control of multiple variables. The actuator command sending module supports continuous power adjustment from 0% to 100% and closed-loop control of valve opening, with command output accuracy up to 0.1 level, ensuring precise execution of control commands. The kiln exit moisture feedback detection module adopts a non-contact infrared detection principle, which has high detection accuracy and no damage to the green body, and can collect finished product moisture content data in real time. The feasible domain adaptive adjustment module realizes batch closed-loop iterative optimization, dynamically shrinks and optimizes the feasible domain according to batch deviation, continuously improves the system control accuracy and production stability, and is suitable for the large-scale, continuous, and intelligent drying production of ceramic tiles.

[0029] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention. The scope of protection claimed by the appended claims and their equivalents is defined.

Claims

1. A method for precise and coordinated temperature and humidity control in a ceramic tile drying kiln, characterized in that, include: Acquire real-time temperature and humidity sequences of multiple drying zones along the conveying direction in the ceramic tile drying kiln, as well as time-series data of the moisture content of the current batch of ceramic tile blanks; Based on the moisture content time series data and the preset drying kinetics model, the target moisture evaporation rate of each drying section is calculated. Based on the target moisture evaporation rate, and combined with the upper limit of the hot air circulation fan speed and the upper limit of the steam injection valve opening in each drying section, a feasible domain for coordinated temperature and humidity control is constructed. Extract the historical temperature and humidity response delay features of each drying section and generate a temperature and humidity transfer delay compensation vector between sections. Based on the feasible domain of the temperature and humidity coordinated control and the temperature and humidity transfer delay compensation vector between the sections, the rolling optimization problem of the multivariable model predictive controller is solved, and the hot air temperature setpoint and relative humidity setpoint of each drying section are output. The hot air temperature setpoint is sent to the electric heater power adjustment unit of the corresponding drying section, and the relative humidity setpoint is sent to the steam jet valve opening execution unit of the corresponding drying section. Infrared moisture detectors were deployed at the outlets of each drying section to collect the actual moisture content of the kiln exit. Based on the deviation between the actual moisture content exiting the kiln and the target moisture content, an adaptive shrinkage adjustment is made to the feasible range of temperature and humidity coordinated control for the next batch drying process.

2. The method for precise temperature and humidity control in a ceramic tile drying kiln according to claim 1, characterized in that, The calculation of the target moisture evaporation rate for each drying section includes: Obtain the initial moisture content and target final moisture content of the current batch of ceramic tile blanks; Based on the real-time position of the ceramic tile blanks in the kiln and the speed of the conveyor belt, the estimated dwell time of the ceramic tile blanks at the exit of each drying section is calculated. A one-dimensional unsteady diffusion equation is constructed using Fick's second law. The average moisture evaporation flux required for each drying section is then calculated by inversion using the expected residence time as the boundary condition. The average moisture evaporation flux is then converted into the target moisture evaporation rate per unit time and per unit area.

3. The method for precise temperature and humidity control in a ceramic tile drying kiln according to claim 1, characterized in that, The feasible domain for constructing coordinated temperature and humidity control includes: A dual-input single-output mapping relationship between hot air temperature and relative humidity and the rate of water evaporation is established, and the mapping relationship is obtained through offline identification. The hot air temperature setpoint and relative humidity setpoint are used as decision variables, the target moisture evaporation rate is used as an equality constraint, and the maximum power of the electric heater and the maximum opening of the steam valve are used as inequality constraints. By introducing slack variables to handle constraint conflicts, a feasible region for coordinated temperature and humidity control with soft constraints is constructed.

4. The method for precise temperature and humidity control in a ceramic tile drying kiln according to claim 1, characterized in that, The inter-segment temperature and humidity transfer delay compensation vector includes: Record the time delay of the humidity sensor response in the next drying section to reach steady state after the hot air temperature of the previous drying section changes abruptly under historical operating conditions. Sliding window statistics were performed on multiple historical delay samples to calculate the mean and standard deviation; When the kiln wind speed deviates from the calibrated wind speed under the current operating conditions, the mean value is dynamically corrected based on the inverse relationship between wind speed and delay. The corrected mean delay is used as the transmission delay compensation value for the corresponding segment pair, and combined to form the inter-segment temperature and humidity transmission delay compensation vector.

5. The method for precise temperature and humidity control in a ceramic tile drying kiln according to claim 1, characterized in that, The problem of solving the rolling optimization of the multivariate model predictive controller includes: A coupled state-space equation containing a temperature sub-model and a humidity sub-model is constructed. The state variables of the state-space equation are the temperature and humidity of each section, and the control inputs are the heating power and the steam flow rate. In the prediction time domain, the inter-segment temperature and humidity transfer delay compensation vector is embedded into the state transition matrix to achieve cross-segment dynamic coupling modeling. Define a performance index function, which includes a moisture evaporation tracking error term, a control energy consumption term, and a temperature and humidity overshoot penalty term. The optimal control sequence that satisfies the feasible region constraint of the temperature and humidity coordinated control is solved online using a sequential quadratic programming algorithm.

6. The method for precise temperature and humidity control in a ceramic tile drying kiln according to claim 1, characterized in that, Before acquiring the real-time temperature and humidity sequences of multiple drying sections along the conveying direction of the ceramic tile drying kiln, the following steps are also included: Detect the closing status signal of the kiln's sealing door; When the closed state signal indicates that the sealing door is open, the transmission of new hot air temperature setpoints and relative humidity setpoints is paused, and the emergency balance mode of the kiln environment is activated. The emergency balance mode of the kiln environment includes setting the exhaust fan control signal to the maximum opening command and closing all steam injection valves.

7. The method for precise temperature and humidity control in a ceramic tile drying kiln according to claim 1, characterized in that, After adaptively shrinking and adjusting the feasible domain of temperature and humidity coordinated control for the next batch drying process based on the deviation between the actual moisture content exiting the kiln and the target moisture content, the process also includes: Calculate the standard deviation of the actual moisture content exiting the kiln for three consecutive batches; When the standard deviation is less than the preset stability threshold, the kiln is determined to have entered the steady-state operation stage. During the steady-state operation phase, the prediction time domain length of the multivariable model predictive controller is reduced to decrease the computational load; When the standard deviation is greater than or equal to the stability threshold, the prediction time domain length is extended and the control frequency is increased.

8. The method for precise temperature and humidity control in a ceramic tile drying kiln according to claim 1, characterized in that, After sending the hot air temperature setpoint to the electric heater power adjustment unit of the corresponding drying section, the method further includes: Monitor the actual power consumption of the electric heater; When the actual power consumption exceeds the theoretical power corresponding to the set value by more than 15% within five consecutive control cycles, an aging warning signal for the heating element is triggered and sent to the operation and maintenance management terminal.

9. The method for precise temperature and humidity control in a ceramic tile drying kiln according to claim 5, characterized in that, Before employing a sequential quadratic programming algorithm to solve online for the optimal control sequence that satisfies the feasible region constraint of the temperature and humidity coordinated control, the method further includes: Using a digital twin simulation platform, based on the current kiln structural parameters and material characteristics, a virtual working condition dataset covering different combinations of moisture content gradients and conveying speeds is generated. A lightweight proxy model is pre-trained on the virtual working condition dataset to initialize the initial solution of the sequential quadratic programming algorithm; The initial solution is input into the sequential quadratic programming algorithm to accelerate the convergence process.

10. A temperature and humidity coordinated precise control system for a ceramic tile drying kiln, used to implement the temperature and humidity coordinated precise control method for a ceramic tile drying kiln as described in any one of claims 1 to 9, characterized in that, include: The multi-source sensor data acquisition module acquires the real-time temperature sequence, real-time humidity sequence, and moisture content time series data of the current batch of ceramic tile blanks from multiple drying sections divided along the conveying direction in the ceramic tile drying kiln. The target evaporation rate calculation module calculates the target moisture evaporation rate for each drying section based on the moisture content time series data and the preset drying kinetic model. The temperature and humidity coordinated feasible domain construction module constructs a feasible domain for temperature and humidity coordinated control based on the target moisture evaporation rate and the upper limit of the hot air circulation fan speed and the upper limit of the steam injection valve opening in each drying section. The segment delay compensation vector generation module extracts the historical temperature and humidity response delay features of each drying segment and generates an inter-segment temperature and humidity transfer delay compensation vector. The multivariable rolling optimization solution module solves the rolling optimization problem of the multivariable model predictive controller based on the feasible domain of the temperature and humidity coordinated control and the temperature and humidity transfer delay compensation vector between the sections, and outputs the hot air temperature setpoint and relative humidity setpoint for each drying section. The actuator command sending module sends the hot air temperature setpoint to the electric heater power adjustment unit of the corresponding drying section, and sends the relative humidity setpoint to the steam jet valve opening execution unit of the corresponding drying section. The kiln exit moisture feedback detection module deploys infrared moisture detectors at the outlet of each drying section to collect the actual kiln exit moisture content; The feasible domain adaptive adjustment module adaptively shrinks and adjusts the feasible domain of temperature and humidity coordinated control for the next batch drying process based on the deviation between the actual moisture content at the kiln outlet and the target moisture content.

Images