A method for model predictive control of temperature and humidity in a ceramic body drying process

Abstract

The application discloses a kind of ceramic body drying process temperature and humidity model predictive control methods, including real-time acquisition drying environment temperature and humidity, wind speed and body surface layer temperature, internal temperature, real-time moisture content, combine body batch characteristics and loading parameters, through moisture-temperature coupling prediction model rolling prediction body moisture content and internal and external temperature difference change trajectory;Automatic identification drying stage and load corresponding safety constraint, temperature and humidity decoupling control is carried out in the constraint range, while according to the prediction deviation trend to carry out feedforward correction on the control amount of actuator;Through the measured and predicted deviation on-line correction model parameter, form prediction type closed loop control, after standard, complete drying by humidification. The present application can avoid the risk of body cracking and bulging in advance, adapt to various types of body and all-season working conditions, with high temperature and humidity control precision, can shorten the drying period, reduce energy consumption, improve the drying uniformity and yield of body, the system can adapt to working condition drift, without manual intervention.

Description

Technical Field

[0001] This invention relates to the field of intelligent drying control technology for ceramic green bodies, and in particular to a method for predictive control of temperature and humidity during the drying process of ceramic green bodies using a model. Background Technology

[0002] Ceramic green bodies are typical porous, water-absorbing media. After forming, the wet green body is filled with pore water. Drying is essentially a thermo-mass coupling process where pore water migrates from the interior of the green body to the surface under the drive of a concentration gradient, and then evaporates from the surface into the environment. This process presents three inherent control challenges. First, water migration is slow and strongly constrained by the thickness of the green body and the porosity of the clay. If the rate of surface water loss and internal water replenishment becomes mismatched, surface shrinkage forms a dry shell, or a stress gradient is generated along the thickness of the green body, manifesting as cracking, bulging, deformation, or even pulverization and scrapping. Second, there is a strong coupling between the temperature and humidity of the drying environment: heating inevitably lowers the relative humidity while increasing the temperature, and humidification lowers the temperature due to heat absorption during vaporization, causing interference between the two control loops. Third, the process has a large lag; the response of the internal moisture state of the green body to changes in environmental parameters takes tens of minutes or even hours, and the effects of adjustments take a long time to manifest in the state of the green body.

[0003] To address the aforementioned challenges, existing technologies have evolved from coarse to fine. Early drying equipment generally employed pre-set drying curves with fixed time, temperature, and humidity. The entire batch of billets was dried according to the same empirical curve, essentially an open-loop operation without measuring the billet's condition. To improve drying quality, a common existing solution divides the drying process into several stages, such as pre-drying, slow drying, and fast drying, with different temperature, humidity, and medium parameters set for each stage, forming a segmented drying curve. This segmented approach is an improvement over a fixed curve, but the stage boundaries are still pre-defined by the process engineer's experience and are not related to the actual dehydration process of the billet. Further improvements have begun to introduce feedback on the billet's condition: humidified hot air is introduced into a closed cavity, and the surface temperature of the billet and the dry-bulb temperature and relative humidity of the drying medium are monitored in real time. The difference between the surface temperature and the medium temperature is used as a controlled variable for feedback adjustment, aiming to match the rate of moisture diffusion from the inside of the billet to the surface evaporation rate, thereby reducing cracking and deformation.

[0004] However, even the better existing solutions, represented by the surface temperature difference feedback, still fail to fundamentally solve the problem of controlling the drying of ceramic blanks. First, such solutions can only sense environmental parameters and a small temperature on the surface of the blank. The actual moisture content and internal temperature of the blank are unknown. "Matching the internal and external water loss rates" can only be indirectly characterized by the surface temperature difference. It is impossible to quantify the degree of matching, let alone predict the risk of cracking in advance. Control is always reactive and indirect.

[0005] Secondly, whether using a fixed curve or a segmented curve, the stage boundaries and parameter values ​​are pre-fixed. However, different batches of clay formulations, green body thickness, loading density, and seasonal environmental changes can cause the critical moisture content and safety boundaries to drift. Fixed curves have poor versatility and are prone to defects such as thick green bodies not drying properly, thin green bodies cracking, and water accumulation inside the dry surface shell. Furthermore, the temperature and humidity loops are usually designed as independent loops. The mutual interference between heating and dehumidifying, and humidifying and cooling, makes conventional control prone to fluctuations and oscillations, making it difficult to guarantee the accuracy of temperature and humidity control. Existing solutions cannot distinguish online between different dehydration stages such as constant-rate drying and falling-rate drying—the moisture migration control links in the constant-rate stage, the first falling-rate stage, and the second falling-rate stage are different, and the corresponding safety boundaries are also different. Using uniform or preset parameters for each stage does not conform to the objective mechanism of moisture migration in the green body.

[0006] Finally, although several mechanism models describing the drying of ceramic blanks have been established in academic research and can be used for offline simulation analysis of the drying process, the parameters of these models are obtained under calibrated working conditions. Once the working conditions change, the predictions become inaccurate and cannot be updated online adaptively. Therefore, they have always remained at the level of process research and have failed to play a role in the real-time control loop. Summary of the Invention

[0007] This invention overcomes the shortcomings of the prior art and provides a method for predicting and controlling the temperature and humidity during the drying process of ceramic blanks using a model.

[0008] To achieve the above objectives, the technical solution adopted by this invention is: a method for predicting and controlling the temperature and humidity during the drying process of ceramic green bodies, comprising the following steps:

[0009] S1. During the drying process, the ambient temperature, ambient humidity and wind speed of the drying environment are collected in real time according to a fixed control cycle, as well as the surface temperature of the ceramic body, the internal temperature of the body and the real-time moisture content of the body.

[0010] S2. Obtain the batch characteristic parameters and loading parameters of the current drying batch of green bodies. The batch characteristic parameters of the green bodies include at least the green body thickness and the porosity of the clay. Input the collected ambient temperature, ambient humidity, wind speed, and batch characteristic parameters and loading parameters of the green bodies into a pre-established moisture-temperature coupling prediction model. The moisture-temperature coupling prediction model characterizes the coupling relationship between the internal moisture migration process and the heat transfer process of the ceramic green body. Based on the moisture-temperature coupling prediction model, using the current collected state as the initial condition, predict the trajectory of the change in the moisture content of the green body and the trajectory of the change in the internal and external temperature difference between the internal temperature and the surface temperature of the green body in future time periods.

[0011] S3. Determine the moisture evaporation rate of the billet based on the real-time moisture content of the billet, automatically identify the current drying stage of the billet based on the change in the moisture evaporation rate, and apply temperature safety constraints, humidity safety constraints, and wind speed safety constraints corresponding to the identified drying stage; the safety constraints corresponding to different drying stages are different.

[0012] S4. Within the boundaries defined by the applied safety constraints, perform temperature and humidity decoupling control based on the deviation of ambient temperature and humidity from the target temperature and humidity of the current drying stage, eliminate the mutual interference between temperature regulation and humidity regulation, and generate the basic control quantities for heating actuator, humidification actuator and fan actuator.

[0013] S5. Based on the deviation trend of the change trajectory of the billet moisture content and the change trajectory of the internal and external temperature difference in the future time period predicted in step S2 relative to the applied safety constraint boundary, the control quantities of the heating actuator, humidification actuator and fan actuator are fed forward to adjust the heating, humidification and wind speed in advance so that the predicted state of the billet does not exceed the safety constraint boundary.

[0014] S6. The continuously collected real-time moisture content, surface temperature, and internal temperature of the billet are compared with the corresponding predicted values ​​of the moisture-temperature coupled prediction model. The model parameters of the moisture-temperature coupled prediction model are corrected online based on the deviation between the measured values ​​and the predicted values. The corrected model is then returned to step S2 for repeated execution to form a predictive closed-loop control until the moisture content of the billet drops to the target moisture content, at which point the drying process ends.

[0015] In a preferred embodiment of the present invention, in step S2, the moisture-temperature coupled prediction model is a lumped parameter coupled model, including a moisture migration sub-model and a heat transfer sub-model.

[0016] The moisture migration sub-model uses the average moisture content of the billet as a state variable to describe the process by which moisture inside the billet migrates to the surface under the drive of the moisture concentration gradient and evaporates into the drying environment at the surface. The rate of moisture migration inside the billet is characterized by the effective moisture diffusion coefficient.

[0017] The heat transfer sub-model uses the internal temperature and surface temperature of the billet as state variables to describe the convective heat transfer, moisture evaporation heat absorption, and internal heat conduction processes between the billet and the drying environment.

[0018] The effective moisture diffusion coefficient is calibrated based on the porosity of the clay: the value of the effective moisture diffusion coefficient is determined by the porosity of the current batch of clay based on a pre-established porosity-effective diffusion coefficient calibration relationship. The porosity-effective diffusion coefficient calibration relationship is pre-calibrated by conducting drying tests on clay samples with different porosities.

[0019] In a preferred embodiment of the present invention, step S2, the rolling prediction includes:

[0020] The moisture migration time constant is determined based on the thickness of the green body and the porosity of the clay material. The moisture migration time constant increases with the increase of the thickness of the green body and decreases with the increase of the effective moisture diffusion coefficient.

[0021] The prediction time domain length is determined by the moisture migration time constant, so that the prediction time domain length covers the complete response process of the green body moisture content.

[0022] Within the prediction time domain, the current real-time moisture content, internal temperature, and surface temperature of the billet are taken as the initial state, and the current ambient temperature, ambient humidity, wind speed, and changes in the control parameters of the actuator are taken as inputs. The predicted values ​​of the billet moisture content and the internal and external temperature differences at each moment within the prediction time domain are obtained through step-by-step recursive calculation using the lumped parameter coupling model.

[0023] After each control cycle, the initial state is updated with the latest collected state and the recursive calculation is performed again to achieve rolling prediction.

[0024] In a preferred embodiment of the present invention, step S3, the automatic identification of the current drying stage of the blank, includes:

[0025] A sliding window filter is applied to the time series of real-time moisture content of the billet, and the time derivative of the filtered moisture content time series is obtained to obtain the moisture evaporation rate series.

[0026] After the drying process enters the constant-rate drying stage, the stable value of the moisture evaporation rate sequence is recorded as the reference rate for the constant-rate segment.

[0027] When the moisture evaporation rate is continuously lower than the first proportion of the constant speed section reference rate and the time derivative of the moisture evaporation rate is continuously negative, it is determined that the green body has entered the first deceleration drying stage from the constant speed drying stage.

[0028] When the moisture evaporation rate is continuously lower than the second proportion of the constant rate section reference rate, and the moisture migration inside the billet changes from surface evaporation control to internal diffusion control, it is determined that the billet has entered the second rate-decreasing drying stage from the first rate-decreasing drying stage; the second proportion is less than the first proportion.

[0029] In a preferred embodiment of the present invention, in step S3, the safety constraints corresponding to each drying stage include:

[0030] Constant-rate drying stage: Set a lower limit constraint on ambient humidity and an upper limit constraint on ambient temperature. The lower limit constraint on ambient humidity is used to prevent the surface of the green body from losing water too quickly and forming a dry shell. The upper limit constraint on ambient temperature is used to limit the intensity of surface moisture evaporation.

[0031] First slow-down drying stage: Set upper limit constraints on the internal and external temperature difference between the internal temperature and the surface temperature of the billet and upper limit constraints on the ambient temperature rise rate. The upper limit constraint on the internal and external temperature difference is used to limit the internal thermal stress of the billet, and the upper limit constraint on the rise rate is used to prevent the internal moisture vaporization pressure of the billet from rising sharply.

[0032] The second stage of slow-down drying involves setting an upper limit constraint on ambient temperature and wind speed adjustment rules. The wind speed adjustment rules include gradually increasing the wind speed as the real-time moisture content of the billet decreases, in order to enhance the removal of surface-migrated moisture.

[0033] In a preferred embodiment of the present invention, in step S4, the temperature and humidity decoupling control is fuzzy decoupling control, including:

[0034] Using temperature deviation and its rate of change, and humidity deviation and its rate of change as inputs, heating control quantities and humidification control quantities are generated respectively through fuzzy reasoning.

[0035] Based on the coupling coefficient characterizing the mutual interference strength between temperature regulation and humidity regulation, the heating control quantity and the humidification control quantity are cross-compensated: the influence of humidity regulation on temperature is converted according to the coupling coefficient and then superimposed on the heating control quantity, and the influence of temperature regulation on humidity is converted according to the coupling coefficient and then superimposed on the humidification control quantity to generate the decoupled basic control quantity;

[0036] The quantization factor, scaling factor, and coupling coefficient of the fuzzy inference constitute a decoupling parameter set. Multiple decoupling parameter sets are pre-stored, each corresponding to a drying stage. When the drying stage switching is identified in step S3, the system synchronously switches to the decoupling parameter set corresponding to the current drying stage.

[0037] In a preferred embodiment of the present invention, step S5, the feedforward correction includes:

[0038] Within the future time period (i.e., the prediction time domain) predicted in step S2, the deviation trend of the predicted value of the billet moisture content and the predicted value of the internal and external temperature difference from the applied safety constraint boundary is examined moment by moment to determine the moment when the predicted state approaches or crosses the safety constraint boundary and the direction of deviation.

[0039] When it is predicted that the rate of decrease in the moisture content of the billet will exceed the allowable rate of water loss in the current drying stage, the heating power should be reduced and / or the humidification rate increased in advance.

[0040] When it is predicted that the internal and external temperature difference will approach the temperature difference boundary in the safety constraint, the increase in heating power is reduced in advance and / or the fan frequency is adjusted to change the heat transfer intensity of the surface layer of the billet.

[0041] When the predicted rate of decrease in the moisture content of the billet is lower than the allowable rate of water loss in the current drying stage and none of the predicted states are close to the safety constraint boundary, the heating power is increased and / or the humidification is reduced in advance to accelerate the drying process while ensuring safety.

[0042] After each control cycle, the feedforward correction is recalculated based on the rolling updated predicted trajectory and executed on a rolling basis.

[0043] In a preferred embodiment of the present invention, step S6, the online correction includes:

[0044] The deviation sequence between the measured values ​​and the corresponding predicted values ​​of the real-time moisture content and internal temperature of the billet is continuously calculated.

[0045] When the magnitude of the deviation sequence continuously exceeds a preset deviation threshold, model parameter correction is triggered;

[0046] The model parameter correction adopts a recursive least squares algorithm with a forgetting factor or an equivalent recursive parameter estimation algorithm: the collected environmental parameters, actuator control quantities and billet state quantities are used as input and output data to recursively estimate and update the model parameters of the moisture-temperature coupled prediction model. The forgetting factor is less than 1, which is used to reduce the weight of historical data and enhance the model parameters' ability to track batch changes, loading changes and seasonal environmental changes.

[0047] The corrected model parameters are used for rolling prediction in the next control cycle.

[0048] In a preferred embodiment of the present invention, in step S1:

[0049] The real-time moisture content of the billet is obtained through one or a combination of the following methods: online weighing of the billet during drying or samples from the same batch being dried, and the real-time moisture content is calculated from the difference between the initial wet weight and the current weight of the billet; resistance method by measuring the resistance value between the surface and interior of the billet and calculating the real-time moisture content according to a pre-calibrated resistance-moisture content relationship; and microwave measurement method by using a microwave moisture sensor to non-contactly measure the moisture content of the billet.

[0050] The internal temperature of the billet is obtained by measuring a temperature measuring element pre-embedded inside the billet during billet forming, and the surface temperature of the billet is obtained by measuring a non-contact temperature measuring element or a temperature measuring element attached to the surface of the billet.

[0051] In a preferred embodiment of the present invention, step S6, the termination of drying includes:

[0052] When both the real-time moisture content and the predicted moisture content of the billet drop to the target moisture content, and the deviation between the two is within the allowable range, the moisture content of the billet is deemed to meet the standard.

[0053] Once the standard is met, the process enters the moisture equalization and final stage: heating is stopped, the fan is kept running at the set speed for the set time, and the ambient humidity is kept at a minimum set final humidity level. This allows the residual moisture inside the billet to be redistributed evenly in the thickness direction of the billet. After the billet temperature drops to near room temperature, the drying process ends.

[0054] This invention addresses the shortcomings of the prior art and has the following beneficial effects:

[0055] (1) Because the present invention introduces direct sensing of the surface temperature, internal temperature and real-time moisture content of the billet into the control loop, and uses the moisture-temperature coupling prediction model driven by the batch characteristic parameters and loading parameters of the billet to predict the development trend of moisture content and internal and external temperature difference, the drying control is upgraded from indirect feedback that "can only measure the environment and cannot see the billet" to predictive control that is oriented towards the real state of the billet. The risk of cracking and bulging is revealed by the predicted trajectory before it occurs. With the help of feedforward correction to adjust heating, humidification and wind speed in advance, the state of the billet never exceeds the safety boundary, thus avoiding surface dry shell, internal water accumulation and thermal stress cracking from the source, and significantly improving the yield of dried billet.

[0056] (2) Since the drying stage is no longer divided according to the preset time, but is automatically identified by the measured moisture evaporation rate and the corresponding temperature, humidity and wind speed safety constraints of each stage are switched accordingly, the stage boundary always follows the actual dehydration process of the batch of green bodies. The drift of critical moisture content caused by changes in clay formula, green body thickness, loading density and seasonal environment is automatically absorbed. The same method can adapt to thick, thin, irregular, multi-variety green bodies and all-season working conditions without manual modification of the drying program, completely getting rid of the dependence on manual experience curves. At the same time, the constraints of each stage are strictly corresponding to the moisture migration mechanism of that stage - the constant speed stage prevents dry shell, the first speed reduction stage prevents thermal stress and vaporization bulging, and the second speed reduction stage speeds up drainage under the premise of ensuring safety. The parameters of each stage are no longer cut off, the drying uniformity is greatly improved, and the consistency of multi-layer and batch loading in the kiln is improved accordingly.

[0057] (3) Since the temperature and humidity decoupling control cross-compensates the heating and humidification outputs according to the coupling coefficient in the control quantity synthesis stage, the mutual interference of heating and humidification and humidification and cooling is eliminated. Moreover, the decoupling parameter group is switched according to the drying stage, and the fluctuation range of temperature and humidity in the drying chamber is greatly reduced. The control accuracy and stability are significantly better than the independent loop scheme. The predictive feedforward channel advances the adjustment action before the risk arrives, overcomes the inherent defect of the feedback adjustment of the large lag process, the kiln working conditions are stable, the ineffective overshoot and callback are reduced, the drying cycle is shortened, the ineffective heating is reduced, and the overall energy consumption is significantly reduced.

[0058] (4) Since the model parameters are driven by the deviation between the measured and the predicted, and the recursive estimation algorithm with forgetting factor is used for online correction, the model can continuously track the drift of mud batch, loading amount and seasonal environment and maintain prediction accuracy for a long time. The prediction basis on which the stage identification and feedforward adjustment depend is always reliable. The whole system realizes fully automatic, no human intervention, and full-condition adaptive precision drying. In addition, the drying ends with the dual standard judgment of measured and predicted and the uniform moisture finishing treatment. The moisture content of the kiln blank is uniform and there is no residual internal stress. The hidden danger of cracking after the kiln is also eliminated. Attached Figure Description

[0059] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0060] Figure 1 This is a flowchart of a preferred embodiment of the present invention; Detailed Implementation

[0061] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0062] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein. Therefore, the scope of protection of the invention is not limited to the specific embodiments disclosed below.

[0063] It should be noted that the main technical problem to be solved by this invention is: how to predict the development trend of the moisture content and internal and external temperature difference of the ceramic body under the conditions of large lag in the drying of the ceramic body, strong coupling of temperature and humidity, frequent changes in batches and loading of clay, and different safety boundaries at each stage of drying, and to proactively and safely adjust the drying environment in stages, so as to shorten the drying cycle, reduce energy consumption, and get rid of dependence on manual experience curves while ensuring that cracking and bulging are not prevented.

[0064] This major technical problem can be further broken down into the following interrelated and detailed technical problems:

[0065] How can the prediction model reflect the coupling mechanism of moisture migration and heat transfer in porous ceramic blanks, and be quickly adapted to batch differences such as clay porosity?

[0066] How to adaptively determine a reasonable prediction time domain for billets of different thicknesses and diffusion capabilities, and complete the rolling prediction calculation of moisture content and internal and external temperature difference;

[0067] How to accurately capture the transition points from the constant speed section to the first deceleration section and from the first deceleration section to the second deceleration section online under noisy conditions in moisture content measurement;

[0068] How to set corresponding safety constraints for different defect mechanisms at each drying stage;

[0069] How to eliminate the mutual interference between temperature regulation and humidity regulation, and make the decoupling parameters adapt to the changes in coupling strength at each stage;

[0070] How to generate feedforward adjustment values ​​for heating, humidification, and fans in advance based on the deviation trend between the predicted trajectory and the safety boundary, so as to ensure both safety and drying efficiency;

[0071] How to ensure that model parameters remain accurate over a long period of time under the drift of mud batches, loading amounts and seasonal environmental changes;

[0072] How to obtain the real-time moisture content and internal temperature of the billet in an engineered manner;

[0073] And how to reliably determine the drying endpoint and avoid cracking due to uneven residual moisture content after kiln exit.

[0074] To solve the above technical problems, such as Figure 1As shown, this invention provides a temperature and humidity model prediction and control method for the drying process of ceramic blanks. Its main process is as follows: During the drying process, the ambient temperature, humidity, and wind speed of the drying environment, as well as the surface temperature, internal temperature, and real-time moisture content of the ceramic blank, are collected in real time. Batch characteristic parameters and loading parameters of the current drying batch are obtained. The collected environmental parameters, along with the batch characteristic parameters and loading parameters, are input into a pre-established moisture-temperature coupled prediction model that characterizes the coupling relationship between the internal moisture migration process and the heat transfer process of the blank. Using the current collected state as the initial condition, the model continuously predicts the trajectory of changes in the moisture content and the internal and external temperature difference of the blank over future periods. The moisture evaporation rate is determined based on the real-time moisture content of the blank, and the current drying stage of the blank is automatically identified according to its changes. A corresponding temperature and humidity model is then applied to the blank. The temperature, humidity, and wind speed safety constraints are corresponding to different drying stages. Temperature and humidity decoupling control is performed within the applied safety constraint boundaries to eliminate mutual interference between temperature and humidity regulation, generating basic control quantities for heating, humidification, and fan actuators. Then, based on the deviation trend of the predicted moisture content change trajectory and the internal and external temperature difference change trajectory relative to the safety constraint boundaries, the actuator control quantities are feedforward corrected to adjust heating, humidification, and wind speed in advance, ensuring that the predicted state of the billet does not exceed the safety constraint boundaries. At the same time, the continuously collected measured values ​​of the billet state are compared with the model prediction values, and the model parameters are corrected online according to the deviation between the measured and predicted values. The corrected model is then run cyclically to form a predictive closed-loop control until the moisture content of the billet drops to the target moisture content, ending the drying process.

[0075] Based on the aforementioned main scheme, this invention further defines each step through subordinate technical solutions. Regarding the green body state prediction step, the moisture-temperature coupled prediction model adopts a lumped parameter coupled model, including a moisture migration sub-model with average moisture content as the state quantity and a heat transfer sub-model with internal temperature and surface temperature as state quantities. The internal moisture migration rate of the green body is characterized by the effective moisture diffusion coefficient, which is determined by the porosity of the current batch of clay based on a pre-established porosity-effective diffusion coefficient calibration relationship. During rolling prediction, the moisture migration time constant is determined according to the green body thickness and clay porosity, and this time constant is used to determine the prediction time domain length. Within the prediction time domain, the predicted moisture content and internal / external temperature difference are calculated iteratively.

[0076] Regarding the identification of drying stages and safety constraints, during drying stage identification, a sliding window filter is applied to the moisture content time series, and the time derivative is calculated to obtain the moisture evaporation rate series. The stable value of the measured rate during the constant-rate drying phase is used as the benchmark rate. When the moisture evaporation rate is continuously lower than a first proportion of the benchmark rate and its time derivative is continuously negative, it is determined that the billet has entered the first decreasing-rate drying stage from the constant-rate drying stage. When the moisture evaporation rate is continuously lower than a second proportion of the benchmark rate and the internal moisture migration of the billet changes from surface evaporation control to internal diffusion control, it is determined that the billet has entered the second decreasing-rate drying stage from the first decreasing-rate drying stage, where the second proportion is less than the first proportion. The specific safety constraints corresponding to the identified stages are: the lower limit of environmental humidity and the upper limit of environmental temperature during the constant-rate drying stage; the upper limit of the internal and external temperature difference and the upper limit of the environmental temperature rise rate during the first decreasing-rate drying stage; and the upper limit of environmental temperature and the adjustment rule of gradually increasing wind speed as the moisture content decreases during the second decreasing-rate drying stage.

[0077] Regarding the control execution within the constraint boundary, the temperature and humidity decoupling control adopts fuzzy decoupling control. The temperature deviation, humidity deviation and their rate of change are taken as inputs and generated by fuzzy inference to produce heating and humidification control quantities. Then, the two control quantities are cross-compensated according to the coupling coefficient that characterizes the mutual interference intensity of temperature and humidity. The quantization factor, proportional factor and coupling coefficient of fuzzy inference constitute the decoupling parameter group, which is synchronously scheduled with the switching of drying stage. During feedforward adjustment, the deviation trend between the predicted trajectory and the safety constraint boundary is checked every moment in the prediction time domain. If the water loss is predicted to be too fast, the heating power is reduced or the humidification is increased in advance. If the internal and external temperature difference is predicted to be close to the temperature difference boundary, the heating power is suppressed or the fan frequency is adjusted in advance. If the water loss is predicted to be too slow and the state margin is sufficient, the speed is increased in advance, and the control is executed in a rolling manner in each control cycle.

[0078] Regarding model maintenance, state acquisition, and drying endpoint, during online model correction, the deviation sequence between measured and predicted values ​​is continuously calculated. When the deviation amplitude continuously exceeds a preset threshold, correction is triggered, and the model parameters are updated using a recursive least squares algorithm with a forgetting factor or its equivalent recursive parameter estimation algorithm. The real-time moisture content of the billet is obtained using one or a combination of weighing, resistance, or microwave measurement methods. The internal temperature of the billet is measured by a temperature measuring element embedded in the billet during forming. At the end of drying, the condition is that both the measured and predicted moisture contents meet the standard and the deviation between them is within the allowable range. After the standard is met, heating is stopped, ventilation is maintained, and the ambient humidity is maintained at no less than the lower limit of the final humidity, so that the residual moisture of the billet is homogenized and the temperature drops to near room temperature before it is removed from the kiln.

[0079] Specifically, the temperature and humidity model prediction and control method for the drying process of ceramic green bodies is implemented in an intelligent drying system for ceramic green bodies, which includes a drying chamber, a multi-source sensing module, a control module, and an actuator.

[0080] The multi-source sensing module includes a temperature and humidity transmitter and an anemometer installed in the drying chamber, a non-contact infrared thermometer for measuring the surface temperature of the billet, a temperature measuring element embedded in the billet, and a weighing device or microwave moisture sensor for obtaining the real-time moisture content of the billet.

[0081] The control module adopts a combination architecture of programmable logic controller and industrial computer, or it can adopt an integrated controller with edge computing capabilities. Model prediction, stage identification and feedforward calculation are completed in industrial computer or edge controller, and the output of basic control quantities and the driving of actuators are completed by programmable logic controller.

[0082] The actuators include heating actuators (electric heating tube assemblies or hot air furnaces), humidification actuators (electrode humidifiers or high-pressure atomizing humidification devices), and fan actuators (variable frequency circulating fans and dehumidification fans).

[0083] The multi-source status acquisition in step S1 is performed synchronously within each control cycle. Specifically, the control module uses... For periodicity (in this embodiment) Take 30 seconds (this can be extended to 1-2 minutes for thick-walled billets), and simultaneously read in the ambient temperature. Ambient relative humidity Wind speed Surface temperature of billet internal temperature of the billet and the real-time moisture content of the billet The validity of each channel signal is checked and outliers are removed. Among these, the surface temperature of the billet... With internal temperature The difference constitutes the internal and external temperature difference that is of concern in subsequent prediction and control:

[0084] ;

[0085] In the formula, The internal and external temperature difference during the k-th control cycle. The internal temperature of the billet is collected during the k-th control cycle. This refers to the surface temperature of the billet collected during the kth control cycle. In the initial drying stage, when the internal moisture of the billet is sufficient and evaporation absorbs heat, the surface temperature is relatively low. Typically positive, its amplitude directly reflects the level of internal thermal stress in the billet. Simultaneously, the control module determines the moisture evaporation rate based on the real-time change in moisture content over time. That is, the negative derivative of moisture content with respect to time:

[0086] ;

[0087] In the formula, This represents the rate of water evaporation. t represents the moisture content of the billet, d represents the time, and d represents the thickness of the billet.

[0088] When each batch of green bodies enters the kiln, the operator enters the batch characteristic parameters and loading parameters of the current drying batch through the human-machine interface, including the green body thickness d and the clay porosity. (Provided by the mud supplier or pre-determined by mercury intrusion porosimetry or boiling water absorption method), number of loaded pieces and initial moisture content .

[0089] In step S2, the control module uses the batch characteristic parameters, loading parameters, and environmental parameters collected in step S1 as inputs to call a pre-established moisture-temperature coupled prediction model. This model characterizes the coupling relationship between the moisture migration process and the heat transfer process inside the ceramic body.

[0090] On the one hand, the heat absorbed by moisture evaporation determines the evolution of the temperature field of the billet; on the other hand, the temperature field affects the speed of moisture migration through the diffusion coefficient.

[0091] In each control cycle, the model uses the currently acquired billet moisture content. Internal temperature Surface temperature Using the current ambient temperature, humidity, wind speed, and changes in the actuator control settings as initial conditions, the trajectory of the change in the moisture content of the billet within future time periods (i.e., the prediction time domain) is predicted in a rolling manner. The trajectory of changes in internal and external temperature differences .

[0092] Step S3 follows from the data collected in step S1. The control module determines the moisture evaporation rate based on the real-time moisture content of the billet. and according to The system automatically identifies the current drying stage of the green body. The drying process of porous ceramic green bodies sequentially goes through a constant-rate drying stage, a first-rate-decreasing drying stage, and a second-rate-decreasing drying stage. The water loss control links and safety boundaries are different in each stage. Therefore, the control module pre-stores temperature safety constraints, humidity safety constraints, and wind speed safety constraints corresponding to each drying stage.

[0093] Whenever a phase transition is identified, the control module immediately unloads the constraint set of the original phase and loads the constraint set of the new phase. The generation of subsequent control variables is restricted to the newly loaded safety constraint boundary.

[0094] Step S4 executes temperature and humidity decoupling control within the applied safety constraints. Since heating inevitably lowers relative humidity while increasing temperature, and humidification inevitably lowers temperature while increasing humidity, there is strong cross-coupling between the temperature and humidity loops. Using two independent conventional control loops would easily cause temperature and humidity oscillations. The control module uses the target temperature for the current drying stage. Target humidity Compared with actual measurement , The deviation is taken as input, and the basic control quantities of heating actuator, humidification actuator and fan actuator are generated through decoupling control, so that the mutual interference between temperature regulation and humidity regulation is canceled in the control quantity synthesis stage.

[0095] Step S5 adds a feedforward correction to the basic control quantity generated in step S4. The large hysteresis characteristic of the drying system determines that adjustment relying solely on deviation feedback is inevitably a "post-correction": by the time the measured state reaches the constraint boundary, the thermal and mass processes inside the billet have already exceeded the safe zone. Therefore, the control module checks the predicted trajectory output in step S2 within each control cycle. , Compared to the deviation trend of the safety constraint boundary loaded in step S3, when the predicted state has not actually crossed the boundary but shows a trend of approaching or crossing the boundary, the control quantities of heating, humidification and wind speed are fed forward in advance to ensure that the predicted state of the billet never crosses the safety constraint boundary.

[0096] Step S6 completes the final two stages of the closed loop. First, the control module continuously collects the real-time moisture content of the billet. Surface temperature and internal temperature The model is compared with the corresponding predicted values ​​of the moisture-temperature coupled prediction model. The model parameters are corrected online based on the deviation between the measured values ​​and the predicted values, so that the model can maintain its prediction accuracy in the long term when the mud batch is changed, the loading amount changes, and the seasonal environment drifts.

[0097] Secondly, the corrected model returns to step S2 to start the next control cycle, thus forming a predictive closed-loop control of "acquisition-prediction-identification-constraint-decoupling-feedforward-correction" until the moisture content of the billet drops to the target moisture content and the drying ends.

[0098] Furthermore, considering that the moisture and temperature distribution of a single ceramic blank in the drying kiln can be approximated by a small number of lumped state quantities under the engineering control precision requirements, this embodiment divides the blank into two lumped units along the thickness direction: an internal core node and a surface node. The model consists of a moisture migration sub-model and a heat transfer sub-model.

[0099] The moisture migration sub-model uses the average moisture content of the green body. Let be a state variable, describing the entire process of moisture migrating from the interior of the billet to the surface driven by a moisture concentration gradient, and then evaporating from the surface into the drying environment. In a lumped sense, this process is equivalent to a first-order inertial element:

[0100] ;

[0101] In the formula, The average moisture content of the green body (dry basis, %). The equilibrium moisture content (%) of the billet under the current environmental conditions. Let be the time constant for moisture migration (s), and t be time. The physical meaning of this formula is: the moisture content of the green body varies with the time constant. The determined rate decays towards the equilibrium value corresponding to its current environment. This comprehensively characterizes the overall rate of moisture diffusion within the green body and surface evaporation. Equilibrium moisture content Determined by environmental conditions, the empirical relationship calibrated through drying equilibrium tests is adopted:

[0102] ;

[0103] In the formula, The saturated equilibrium moisture content coefficient (%). Humidity index The temperature decay coefficient is 1 / ℃. The relative humidity (%RH) is the ambient humidity. The ambient temperature is (°C). , , The results were obtained by fitting the static equilibrium test of the clay sample under different temperature and humidity conditions; for typical daily-use ceramic clay, Approximately 48%, Approximately 0.5-0.8, It is approximately 0.01~0.03 / ℃.

[0104] Moisture migration time constant Based on the structural parameters of the green body and the effective moisture diffusion coefficient The decision is made. For a flat blank of thickness d (an irregularly shaped blank is equivalent to a characteristic thickness), the lumped approximation of Fick's diffusion law yields the following:

[0105] ;

[0106] In the formula, d is the thickness of the billet (m). The effective moisture diffusion coefficient (m² / s) indicates that the thickness of the green body has a quadratic effect on the drying rate, which is the fundamental reason why thick-walled green bodies are difficult to dry. Effective moisture diffusion coefficient Calibration based on clay porosity: Based on a pre-established porosity-effective diffusion coefficient calibration relationship, the porosity of the current batch of clay is used as the basis for calibration. Sure The value of is determined using a calibration formula that considers both the effects of pore structure and temperature.

[0107] ;

[0108] In the formula, Reference porosity Reference temperature The reference effective diffusion coefficient (m² / s) is given below. The porosity of the current batch of clay material. For reference porosity, The porosity calibration index, R is the activation energy for water diffusion (J / mol), and R is the molar gas constant (taken as 8.314 J / (mol·K)). The internal temperature of the billet (°C) is the temperature inside the billet. Reference temperature (°C). , , , Multiple isothermal drying tests were conducted on clay samples with different porosities, and the results were pre-calibrated by inversion fitting of measured drying curves; for daily-use ceramic clay with a porosity of 30%~45%, The typical order of magnitude is ~ m² / s, Approximately 1.5~2.5, It is approximately 20–35 kJ / mol. The exponential term reflects the activating and promoting effect of temperature on water diffusion, and represents the first coupling between the water migration sub-model and the heat transfer sub-model.

[0109] The heat transfer sub-model uses the internal temperature of the billet. and surface temperature Let be a state variable, describing the convective heat transfer between the billet and the drying environment, the heat absorption due to moisture evaporation, and the heat conduction process within the billet. The energy balance is written separately for the internal core nodes and surface nodes, yielding:

[0110] ;

[0111] ;

[0112] In the formula, The apparent density of the wet preform is (kg / m³). Specific heat capacity of wet green body (J / (kg·℃)) , These are the equivalent volumes (m³) of the internal core nodes and the surface nodes, respectively. The effective heat exchange surface area (m²) of a single blank. The thermal conductivity of the wet blank is (W / (m·℃)). The convective heat transfer coefficient between the surface of the billet and the drying environment (W / (m²·℃)). The latent heat of vaporization of water (J / kg). The dry basis weight (kg) of a single blank. To represent the rate of water evaporation, the percentage moisture content is converted to a dry basis mass fraction and expressed in units of 1 / s when used in this calculation. That is, the rate of moisture evaporation of a single blank, kg / s, where d is the thickness of the blank (m). Let represent the ambient temperature (°C). One formula indicates that the temperature change of the internal core is entirely driven by heat conduction between the surface and the core; the other formula's three terms on the right-hand side represent, in order, the convective heating of the surface by the environment, the heat absorption by surface moisture evaporation (i.e., the reverse coupling of moisture migration to the temperature field, which is the second coupling between the two sub-models), and the heat conduction from the surface to the core. Convective heat transfer coefficient. With wind speed The change is expressed using a power function relation:

[0113] ;

[0114] In the formula, The reference heat transfer coefficient for natural convection is (W / (m²·℃)). This is the wind speed coefficient. Wind speed index The ambient wind speed (m / s) during drying is given. These three parameters were determined by the kiln heat exchange calibration test. Typical values ​​range from 0.5 to 0.8.

[0115] The effect of loading parameters on the model is reflected in the environment channel: number of loaded items. The total evaporation rate of the green body inside the kiln determines the total amount of water evaporated, which in turn determines the response characteristics to ambient humidity. The total evaporation rate of the green body inside the kiln is:

[0116] ;

[0117] In the formula, The total moisture evaporation rate of all the billets in the kiln (kg / s). For the number of loaded items, The dry basis weight of a single blank is (kg). The evaporation rate of water (expressed as dry weight fraction, 1 / s) is given. The moisture evaporation rate of a single blank (kg / s). As an input to the environmental humidity dynamic equation, it, together with the environmental humidity measurement channel, forms a forward perturbation model of the humidity loop, explicitly accounting for the influence of load changes on the humidity response. This constitutes a complete lumped parameter coupling model: using the batch characteristic parameters of the billet ( ) and loading parameters ( ) as structural parameters, and environmental parameters ( ) as boundary input, with As a state variable, it fully characterizes the bidirectional coupling of moisture migration and heat transfer during the drying process of porous ceramic blanks.

[0118] Furthermore, a primary issue in prediction is determining the appropriate prediction time domain length: if the time domain is too short, it cannot cover the large hysteresis response process of the ceramic green body's moisture content, rendering feedforward adjustment without a look-ahead basis; if the time domain is too long, the model error accumulates and amplifies during the recursion, and the online computational load increases unnecessarily. This embodiment utilizes the green body's own moisture migration time constant. To determine the prediction time domain length, the prediction span is adaptively matched with the inherent drying rate of the current batch of billets.

[0119] Specifically, based on the thickness of the billet Porosity of mud Determine the water migration time constant , The effective moisture diffusion coefficient increases quadratically with the increase of the billet thickness d. It decreases as the value increases. The prediction time domain length is taken as:

[0120] ;

[0121] In the formula, To predict the time domain length (s). The time-domain coverage coefficient is dimensionless and ranges from 3 to 5, which makes the predicted time-domain coverage of the first-order response of the green body moisture content about 95% to 99% of the complete process. To predict the number of steps, To predict the step size (s), in this embodiment... With control cycle Take the same value. The corresponding integer is between 30 and 300 (for thick billets, take the larger value; this is only a typical reference).

[0122] In each control cycle k, the control module uses the currently acquired real-time moisture content of the billet. internal temperature of the billet and the surface temperature of the billet Let it be the initial state, that is, let And based on the current ambient temperature. Ambient humidity Wind speed The calculation is performed stepwise recursively, taking the changes in the actuator control settings (maintaining the current value when no new control is applied, and following the corrected setting trajectory after feedforward correction) as input. The moisture content channel uses explicit Euler discretization.

[0123] ;

[0124] In the formula, The predicted moisture content (%) of the billet at step k+j is given in the k-th control cycle. Prediction step number , These are the predicted ambient humidity and ambient temperature trajectories (%RH, ℃), respectively. , Calculate using the formula above.

[0125] The temperature channels are discretized using the same Euler method, by , To deduce the next temperature, the evaporation endothermic term... The moisture content is obtained from the difference between two adjacent predicted values; subsequently, the predicted trajectory of the internal and external temperature difference is obtained:

[0126] ;

[0127] In the formula, This represents the predicted internal and external temperature difference (°C) at step k+j. After each control cycle, the control module updates the initial state with the latest acquired data and re-executes the above recursive calculation to achieve rolling prediction. Simultaneously, if the model parameters are corrected online in step S6, the corrected parameters will participate in the recursion from the next control cycle. This results in... and The two predicted trajectories are the inputs for the feedforward adjustment in step S5.

[0128] The input for stage identification is the time series of the real-time moisture content of the billet. The difficulty lies in the fact that moisture content measurement inevitably contains noise (the weighing channel is affected by airflow disturbances in the kiln, and the resistance and microwave channels are affected by temperature drift). Direct differentiation to calculate the rate would amplify the noise to an unusable level. Therefore, this embodiment first applies a sliding window filter to the time series of the real-time moisture content of the billet:

[0129] ;

[0130] In the formula, The moisture content (%) is the filtered value after the kth control cycle. The length of the sliding window (take 5~20, corresponding to a time span of 2.5~10 min, take the larger value for thick billets). For the first The measured moisture content (%) for each control cycle was then calculated. Subsequently, the time derivative of the filtered moisture content time series was obtained to yield the moisture evaporation rate series:

[0131] ;

[0132] In the formula, Let be the water evaporation rate (% / min) in the kth control cycle. To control the cycle length (converted to min for calculation).

[0133] After drying begins, the surface of the green body has sufficient moisture, and the evaporation rate remains essentially constant due to environmental conditions, thus entering the constant-rate drying stage. During this stage, the control module continuously checks the stability of the rate: when the relative fluctuation of the rate between adjacent cycles meets the following... And continue During each control cycle (of which) The initial baseline rate value, which was previously confirmed, is replaced by the current window average value during the initial confirmation. The stationarity criterion coefficient is set between 0.05 and 0.10. (Take a value between 10 and 30), determine that the rate has entered the stable phase, and record the stable value of the water evaporation rate sequence at this time as the reference rate for the constant rate phase:

[0134] ;

[0135] In the formula, The reference speed for the constant speed range is (% / min). Instead of being manually preset, the values ​​are automatically extracted from the measured data of the current batch of blanks. Therefore, they naturally carry information about the current clay formula, thickness, loading density and environmental conditions. Subsequent judgments can be made based on this and automatically adapt to batch differences. This is the key difference between this method and the preset drying curve.

[0136] As the surface evaporable moisture is depleted, the evaporation surface begins to shrink, and the evaporation rate drops. When the moisture evaporation rate remains below a first proportion of the reference rate in the constant-rate drying stage, and the time derivative of the rate remains negative, the green body is determined to have entered the first decreasing-rate drying stage from the constant-rate drying stage. The criterion is:

[0137] ,and Both conditions continue simultaneously One control cycle.

[0138] In the formula, The first proportionality coefficient (taken as 0.80~0.90). To confirm the number of cycles (5~15), the duration is set to suppress instantaneous crossover misjudgment caused by measurement noise.

[0139] After entering the first decreasing rate stage, the surface of the green body is partially dry, and the evaporation surface gradually migrates inward, with internal moisture diffusion gradually becoming the controlling factor for the entire process. When the moisture evaporation rate continues to fall below the second proportion of the reference rate in the constant rate stage, and the internal moisture migration of the green body changes from being controlled by surface evaporation to being controlled by internal diffusion, it is determined that the second decreasing rate drying stage has begun.

[0140] In engineering, the measurable sign of "transitioning to internal diffusion control" is that surface moisture evaporation is no longer sufficient to absorb heat from the environment, and the surface temperature deviates from the wet-bulb temperature and begins to rise continuously. Therefore, the criterion is:

[0141] ,and Both conditions continue simultaneously One control cycle.

[0142] In the formula, The second proportionality coefficient (taken as 0.40~0.60) ), To confirm the number of cycles (take 5~15). Let be the surface temperature (°C) of the billet in the kth control cycle. Whenever the above two conditions are met, the control module confirms a stage transition, then unloads the safety constraint set of the original stage and loads the temperature, humidity, and wind speed safety constraint set corresponding to the new stage; the surface temperature rising trend term also reuses the temperature measurement channel already in step S1, without the need to add any additional sensing means.

[0143] The safety constraints for the constant-rate drying stage are the lower limit of ambient humidity and the upper limit of ambient temperature. During this stage, the evaporation rate of surface moisture in the green body is entirely determined by the heat supplied by the environment and the mass transfer driving force. This is the stage with the fastest water loss in the entire drying process, and it is also the stage most prone to shrinkage and the formation of a dry shell due to excessive surface water loss. Once the dry shell forms, the channels for internal moisture to migrate outward are blocked, inevitably leading to subsequent cracking and bulging. Therefore, a lower limit for relative humidity is set for this stage. (Typically 60%~75% RH) to suppress surface evaporation intensity, while setting an upper limit for ambient temperature. (Typically 45~55℃) to limit the heat intensity supplied from the environment to the blank; both together ensure that the surface evaporation rate does not exceed the migration rate of internal moisture to replenish the surface.

[0144] The safety constraints for the first stage of slow-drying are the upper limit of the internal and external temperature difference and the upper limit of the ambient temperature rise rate. During this stage, the evaporation surface shifts inward, internal diffusion resistance becomes the dominant factor, and the main causes of cracking in the billet become thermal stress and internal vapor pressure. If the internal and external temperature difference... If the difference is too large, the combined effect of thermal expansion and contraction differences along the thickness direction will directly crack the billet; if the ambient temperature rises too quickly, the internal temperature of the billet will rise sharply, causing a sudden increase in the vaporization pressure of residual moisture inside, which can easily lead to bulging. Therefore, an upper limit for the internal and external temperature difference is set at this stage. (Typical values ​​are 8~15℃, with smaller values ​​for thick-walled billets) and the upper limit of the ambient temperature heating rate. (Typically, the speed is 0.5~2℃ / min).

[0145] The safety constraints for the second stage of slow-down drying are the upper limit of ambient temperature and wind speed regulation rules. At this stage, the moisture content of the billet is already low, shrinkage is basically complete, and the risk of cracking is significantly reduced. The process objective shifts to quickly removing residual moisture that has migrated to the surface to shorten the drying cycle. Only the upper limit of ambient temperature is retained as a temperature constraint. (Typically 70~90℃) to prevent the billet from overheating and pulverizing; the air velocity is no longer kept constant, but gradually increases as the real-time moisture content of the billet decreases, in order to enhance the removal of surface moisture:

[0146] ;

[0147] In the formula, The wind speed setpoint (m / s) for the kth control cycle. The starting wind speed for the second deceleration phase (m / s, typically 0.5~1.0). The wind speed increment coefficient (m / s per % moisture content, typically taken as 0.05~0.15). The moisture content (%) of the billet recorded when entering the second deceleration stage. The current real-time moisture content (%) of the billet. The upper limit of wind speed is set at 2-3 m / s (typically 2-3, balancing fan energy consumption and kiln temperature uniformity). Therefore, the constraint sets for the three stages are distinct and automatically switch according to the stage identification results. The control objective of each stage strictly corresponds to the moisture migration mechanism of that stage.

[0148] The coupling between temperature and humidity inside the drying kiln is bidirectional: increasing the heating power raises the temperature while simultaneously increasing the air's water-holding capacity and decreasing the relative humidity; conversely, increasing the humidification rate raises the humidity while simultaneously causing the temperature to drop due to the heat absorbed by moisture vaporization. The decoupling approach in this embodiment is to generate nominal control quantities without considering coupling using fuzzy inference, and then cross-compensate the interaction between the two loops according to the coupling coefficient to the other channel.

[0149] Specifically, the control module calculates the temperature deviation and humidity deviation in each control cycle:

[0150] ;

[0151] In the formula, Temperature deviation (°C) This refers to the humidity deviation (%RH). These represent the target temperature and target humidity (°C, %RH) for the current drying stage, respectively. Their values ​​are selected within safety constraints as the stage changes. Temperature deviation is used as the metric. and its rate of change , humidity deviation and its rate of change The input quantities are converted into fuzzy quantities in the universe of discourse by quantization factors, and then inferred by consulting the fuzzy rule table (the rules are set according to conventional control experience, for example, when the temperature deviation is large and the rate of change of the deviation is positive, the heating output is set to a large negative value to suppress overshoot; when the humidity deviation is large and the change is slow, the humidification output is set to a large positive value to accelerate tracking). After defuzzification and proportional factor conversion, the nominal heating control quantity is obtained. and nominal humidification control amount (All are dimensionless outputs ranging from 0% to 100%).

[0152] Cross-compensation is then performed. The cooling effect of humidification on temperature is calculated using a coupling coefficient and then added to the heating control variable; the dehumidification effect of heating on humidity is calculated using a coupling coefficient and then added to the humidification control variable.

[0153] ;

[0154] ;

[0155] In the formula, , These are the decoupled heating control quantity and humidification control quantity (0~100%), respectively. The coupling coefficient for the effect of humidity regulation on the temperature channel. The coupling coefficient between temperature regulation and the humidity channel was determined beforehand through a step identification test: a unit step was applied to the humidification channel under steady-state conditions, and the steady-state response amplitude and adjustment time of the temperature channel were recorded. The ratio of their gains was... ; The measurement method is symmetrical.

[0156] After compensation, the temperature drop caused by humidification is simultaneously compensated by the heating channel, and the humidity drop caused by heating is simultaneously compensated by the humidification channel. The mutual interference between temperature and humidity regulation is eliminated during the control quantity synthesis stage. Decoupling , These are respectively mapped to the basic power level of the heating actuator and the basic opening degree of the humidifying actuator, and then used as the basic control quantity after being limited by the safety constraints applied in step S3.

[0157] The basic control quantity of the fan actuator is generated by the wind speed setpoint of the current drying stage through an independent wind speed feedback loop: the wind speed setpoint and the wind speed measured in step S1 are used to determine the control quantity. The deviation is adjusted by proportional-integral control to output the basic value of the fan frequency, and then combined with the fan feedforward correction in step S5 to drive the variable frequency fan.

[0158] It should be noted that the coupling strength between temperature and humidity is not constant: evaporation is strong and the humidification and cooling effects are significant during the constant-rate stage, while evaporation is weak and the coupling weakens significantly during the second-rate-decreasing stage. Therefore, this embodiment uses the quantization factor, scaling factor, and coupling coefficient of fuzzy inference. , The whole system consists of a set of decoupling parameters. The controller has three sets of decoupling parameters that correspond one-to-one with the three drying stages of constant speed, first speed reduction and second speed reduction (each set of parameters is tuned under the typical working conditions of the corresponding stage). In step S3, the same control cycle of the drying stage switching is identified, and the controller is synchronously scheduled to switch to the decoupling parameter set corresponding to the current drying stage, so that the decoupling effect remains consistent throughout the entire drying process.

[0159] In this embodiment, the "future time period" mentioned in step S2 is uniformly referred to as the prediction time domain, and its length is [length missing]. The predicted trajectory is and Feedforward correction is carried out in three steps: trend verification, correction amount calculation, and rolling execution.

[0160] Trend testing is performed time-by-time within the prediction time domain. For each prediction step j, the control module checks the predicted evaporation rate corresponding to the predicted moisture content. Will this exceed the allowable rate of water loss during the current drying stage? (This range is part of the stage safety constraints; the values ​​differ between the constant speed and deceleration sections), and the predicted value of the internal and external temperature difference. Will it approach the temperature difference boundary in the safety constraints? This allows us to determine when the predicted state approaches or crosses the safety constraint boundary and the direction of deviation.

[0161] The correction calculation is divided into two categories based on the direction of deviation: suppressive feedforward and accelerating feedforward. Both use time weighting with the principle of "the closer, the greater the weight," and the weights are set as follows:

[0162] ;

[0163] In the formula, Let be the weight for the j-th prediction step, where j is the prediction step number. The weighting is used to predict the number of steps. This weighting gives greater moderating power to near-term out-of-bounds risks, while providing only mild preparation for long-term risks, avoiding overreaction to long-term uncertainties.

[0164] When it is predicted that the rate of decrease in the moisture content of the billet will exceed the upper limit of the allowable water loss rate in the current drying stage, or that the internal and external temperature difference will approach the temperature difference boundary, suppressive feedforward is applied to reduce the heating power in advance. If necessary, the humidification rate is increased and the fan frequency is adjusted to change the heat transfer intensity of the billet surface.

[0165] ;

[0166] In the formula, This is the feedforward correction amount for the suppressive heating power (kW, always non-positive). The correction gain (kW per % / min) for exceeding the water loss rate limit. The correction gain (kW / ℃) is applied when the temperature difference approaches the boundary. This represents the upper limit of the permissible water loss rate (% / min) for the current stage. This represents the upper limit of the internal and external temperature difference (°C). To provide a safety margin for temperature difference (°C, taken as 1~3°C), the feedforward is positioned at a distance from the boundary. (The action begins immediately). This indicates that the value within the parentheses is the larger of zero. The correction values ​​for the humidification channel and the fan channel are generated along the same path: when water loss is predicted to be too rapid, the humidification feedforward is positive (early humidification); when water loss is predicted to be too rapid or the temperature difference is approaching, the fan frequency is corrected in the direction of reducing the surface heat transfer intensity. In this embodiment, the fan correction gain is taken as 0.1 to 0.3 times the heating correction gain and is limited.

[0167] When the predicted rate of decrease in the moisture content of the billet is lower than the lower limit of the allowable rate of water loss at the current stage, and none of the states in the prediction time domain approach any safety constraint boundary, it indicates that the current operating condition is conservative and there is room for acceleration. Acceleration feedforward should be applied to increase the heating power in advance and reduce the humidification rate if necessary.

[0168] ;

[0169] In the formula, The feedforward correction amount for the acceleration heating power (kW, always non-negative). To adjust the gain (kW per % / min, the value is less than) To ensure the safety priority principle of "better to be slow than to break" This represents the lower limit of the permissible water loss rate (% / min) for the current stage. This is the predicted evaporation rate (% / min) for the next control cycle. It is enabled on the premise that the margins of all predicted states from each constraint boundary are greater than the set safety margins; if any margin is insufficient, the acceleration channel is frozen.

[0170] Finally, the feedforward correction and the basic control output from step S4 are combined, limited, and then the output is executed.

[0171] ;

[0172] In the formula, This represents the actual output heating power (kW) during the k-th control cycle. The control quantity is decoupled by step S4. The base heating power (kW) obtained by mapping. The maximum power (kW) of the heating actuator. This is a saturation limiting function, meaning that x is restricted to a lower limit. and upper limit between.

[0173] The synthesis method of humidification amount and fan frequency is symmetrical. After each control cycle, the control module recalculates the above feedforward correction amount based on the rolling updated prediction trajectory and executes it in a rolling manner, so that the adjustment action always arrives before the risk. The feedback loop is responsible for eliminating residual deviations, and the feedforward channel is responsible for avoiding future overshoots. The two converge at the synthesis point.

[0174] The model parameters mentioned above: The equivalent coefficients, heat transfer correlation coefficients, and equilibrium moisture content coefficients are obtained under calibrated operating conditions. However, in actual production, the batch size of the clay, the loading volume, and the seasonal environment continuously drift, inevitably causing the fixed-parameter model to gradually become inaccurate. Furthermore, both stage identification and feedforward adjustment rely on model predictions, and model inaccuracies will amplify along the closed-loop chain. Therefore, the control module continuously calculates the deviation between the measured values ​​and corresponding predicted values ​​of the real-time moisture content and internal temperature of the billet. Taking the moisture content channel as an example:

[0175] ;

[0176] In the formula, Let M(k) be the one-step prediction deviation of the moisture content in the k-th control cycle (%), and M(k) be the measured moisture content (%). Predicted moisture content (%) for this cycle based on the previous control cycle. Deviation of internal temperature channels. Calculate using the same method.

[0177] To avoid erroneous corrections caused by transient measurement disturbances, a continuous over-limit triggering mechanism is adopted: when the moving root mean square of the deviation sequence satisfies And continue During one control cycle ( The deviation thresholds are typically set at 0.3% to 0.8% for the moisture content channel and 1 to 2℃ for the temperature channel. (Take a value of 10~30) to trigger a model parameter correction.

[0178] The corrected algorithm employs a recursive least squares algorithm with a forgetting factor. After discretization, it is rearranged into a linear regression form:

[0179] ;

[0180] In the formula, The estimated output is the measured value of moisture content or internal temperature in the kth control cycle. The regression vector is composed of the collected environmental parameters ( ), and the amount of control by the executing agency ( The historical values ​​of humidification capacity, fan frequency, and billet state parameters are constructed using a discrete model structure. The model parameter vector to be corrected (its elements are...) reciprocal, Correlation coefficient The equivalent coefficients are to be estimated), n(k) is the zero-mean observation noise, and the superscript T indicates transpose. The recursive estimation is performed according to the following three equations:

[0181] ;

[0182] ;

[0183] ;

[0184] In the formula, Let be the gain vector at step k. Let be the covariance matrix at step k (initial value is a large positive definite matrix, such as 10³ times the identity matrix). This is the model parameter estimation vector for the k-th step. This is the forgetting factor, with a value less than 1 (typically 0.95~0.995).

[0185] The forgetting factor adjusts the weights of historical data in the objective function according to... The power-law decay reduces the weight of historical data and enhances the ability of parameter estimation to track changes in mud batches, loading volume, and seasonal environmental changes; during the stable working period, λ can be taken as a large value close to 1 to suppress estimation fluctuations, while during the transition period after batch change, a smaller value can be taken to accelerate convergence.

[0186] After each trigger, the algorithm uses the current... The model is run continuously for a correction window (e.g., 30-100 control cycles) with initial values ​​until the deviation falls back below the threshold, after which the parameters are frozen. The corrected model parameters are used for rolling prediction starting from the next control cycle. As an equivalent alternative, those skilled in the art can also use recursive augmented least squares, Kalman parameter estimation with a forgetting factor, or fading memory recursive algorithm to achieve the same deviation-driven parameter update, all of which are equivalent recursive parameter estimation algorithms.

[0187] The real-time moisture content of the billet can be obtained by weighing: a suspended weighing sensor is installed inside the kiln to weigh representative billets or samples from the same batch during drying. The samples are dried with the same clay, thickness, and kiln position as the final billets, avoiding damage caused by placing sensors on the final product. Let the initial wet weight of the weighed sample be... If the current weight is G(k), then the real-time moisture content on a dry basis is:

[0188] ;

[0189] In the formula, The current weight (kg) of the sample obtained during the k-th control cycle. The sample's oven-dry mass (kg) is given. This represents the real-time moisture content (%) on a dry basis. Oven-dried weight does not require destructive drying testing; it can be determined by the initial wet weight and the initial moisture content measured before kiln loading. (Determination by drying method) Estimation:

[0190] ;

[0191] In the formula, The initial wet weight of the sample is (kg). The initial moisture content is (dry basis, %). The airflow disturbance noise in the weighing channel has been filtered by a sliding window. If necessary, a short-time quasi-static weighing process of "stop airflow - stabilize - weigh - recover" can be used to further improve accuracy.

[0192] The real-time moisture content of the billet can also be obtained by the resistance method: a pair of electrodes are embedded on the surface and inside of the billet, and the resistance value of the billet between the two electrodes is measured. The real-time moisture content is calculated based on the pre-calibrated resistance-moisture content relationship. The conductivity of the ceramic wet body is mainly contributed by ion conduction in the pore liquid phase; the moisture content and resistance are approximately linear in a logarithmic coordinate system.

[0193] ;

[0194] In the formula, The blank resistance value measured in the kth control cycle ( ), The calibration coefficient is obtained by fitting the multi-point resistance measurement of the mud sample under a known moisture content gradient, and lg represents the logarithm to the base 10.

[0195] The resistance method offers fast response and low cost, but its sensitivity decreases at low moisture content, making it suitable for high and medium moisture content ranges. Alternatively, a microwave moisture sensor can be used for non-contact measurement of billet moisture content: utilizing the strong dielectric response of water molecules to microwaves, the moisture content is calculated from the attenuation and phase shift of the microwaves after passing through or being reflected by the billet. This method is unaffected by the surface condition of the billet and is suitable for irregularly shaped billets and kiln cars where wiring is inconvenient. These three methods can be used individually or in combination. For example, the weighing method provides a full-range reference, the resistance method provides a high response rate, and the microwave method provides spatial representativeness. The control module weights and fuses the results of the three methods according to preset confidence levels. If any channel fails, it automatically degrades to the remaining channels, ensuring uninterrupted acquisition in step S1.

[0196] The internal temperature of the billet is measured by a temperature sensing element pre-embedded in the geometric center (or midpoint of the characteristic thickness) of the billet during forming. The preferred temperature sensing element is an armored K-type thermocouple or an NTC thermistor, with its leads extending along the non-surface surface of the billet. In cases where leading leads is inconvenient, the temperature sensing element can be pre-embedded in a batch of accompanying samples. The surface temperature of the billet is measured using a non-contact infrared thermometer, or alternatively, a thin-film thermocouple attached to the surface of the billet. The infrared method, which does not contact the billet and does not interfere with surface moisture evaporation, is the preferred option.

[0197] The determination of whether the moisture content meets the standard is based on dual confirmation of actual measurement and prediction to eliminate misjudgment caused by fluctuations in a single measurement channel: when both the real-time moisture content and the predicted moisture content of the billet drop to the target moisture content and the deviation between the two is within the allowable range, the moisture content of the billet is determined to meet the standard, that is, simultaneously satisfying the following conditions:

[0198] ,and ;

[0199] In the formula, The target moisture content (%) is set according to the requirements of subsequent firing processes; typically, it is 0.5% to 2% for daily-use ceramic blanks. This is the predicted moisture content (%) for the next step. The one-step prediction deviation for moisture content (%). The allowable deviation (%, typically 0.2%~0.5%) for compliance determination. One study confirmed that the model is currently reliable, thus eliminating two potential problems: "actual measurement meets the standard but the model is distorted" or "sensor drift falsely meets the standard".

[0200] After the standard is met, the billet is not immediately removed from the kiln, but instead enters the moisture equalization and finishing stage: heating is stopped, and the fan is maintained at a set speed (typically 30%~50% of the rated speed) for a set time (typically 30~120 min, with the larger value for thick-walled billets), while maintaining the ambient humidity at no lower than the set finishing humidity limit (typically 40%~55%RH), so that the residual moisture inside the billet is redistributed evenly along the thickness of the billet. If the billet is immediately cooled and removed from the kiln at this time, the residual moisture content gradient between the surface and core of the billet will continue to induce internal stress during storage, which is the main cause of "cracking after removal from the kiln". Wait until the surface temperature of the billet drops to near room temperature ( After drying is completed (with the temperature difference from room temperature not exceeding 5°C), the control system automatically records the entire process data and final model parameters of this batch, which serve as the source of initial values ​​for the next batch of the same clay body model, thus realizing the inheritance of experience between batches.

[0201] Based on the preferred embodiments of the present invention described above, those skilled in the art can make various changes and modifications without departing from the inventive concept. The technical scope of this invention is not limited to the contents of the specification, but must be determined according to the scope of the claims.

Claims

1. A method for predicting and controlling temperature and humidity during the drying process of ceramic green bodies, characterized in that, Includes the following steps: S1. During the drying process, the ambient temperature, ambient humidity and wind speed of the drying environment are collected in real time according to a fixed control cycle, as well as the surface temperature of the ceramic body, the internal temperature of the body and the real-time moisture content of the body. S2. Obtain the batch characteristic parameters and loading parameters of the current drying batch of green body. The batch characteristic parameters of the green body include at least the green body thickness and the clay porosity. Input the collected ambient temperature, ambient humidity, wind speed, and batch characteristic parameters and loading parameters of the green body into the pre-established moisture-temperature coupling prediction model. The moisture-temperature coupling prediction model characterizes the coupling relationship between the moisture migration process and the heat transfer process inside the ceramic green body. Based on the moisture-temperature coupled prediction model, the current collected state is used as the initial condition to predict the trajectory of the change in the moisture content of the billet and the trajectory of the change in the internal temperature and the internal and external temperature difference between the billet and the surface temperature of the billet in future time periods. S3. Determine the moisture evaporation rate of the billet based on the real-time moisture content of the billet, automatically identify the current drying stage of the billet based on the change in the moisture evaporation rate, and apply temperature safety constraints, humidity safety constraints, and wind speed safety constraints corresponding to the identified drying stage; the safety constraints corresponding to different drying stages are different. S4. Within the boundaries defined by the applied safety constraints, perform temperature and humidity decoupling control based on the deviation of ambient temperature and humidity from the target temperature and humidity of the current drying stage, eliminate the mutual interference between temperature regulation and humidity regulation, and generate the basic control quantities for heating actuator, humidification actuator and fan actuator. S5. Based on the deviation trend of the change trajectory of the billet moisture content and the change trajectory of the internal and external temperature difference in the future time period predicted in step S2 relative to the applied safety constraint boundary, the control quantities of the heating actuator, humidification actuator and fan actuator are fed forward to adjust the heating, humidification and wind speed in advance so that the predicted state of the billet does not exceed the safety constraint boundary. S6. The continuously collected real-time moisture content, surface temperature, and internal temperature of the billet are compared with the corresponding predicted values ​​of the moisture-temperature coupled prediction model. The model parameters of the moisture-temperature coupled prediction model are corrected online based on the deviation between the measured values ​​and the predicted values. The corrected model is then returned to step S2 for repeated execution to form a predictive closed-loop control until the moisture content of the billet drops to the target moisture content, at which point the drying process ends.

2. The method for predicting and controlling temperature and humidity during the drying process of ceramic green bodies according to claim 1, characterized in that: In step S2, the moisture-temperature coupled prediction model is a lumped parameter coupled model, which includes a moisture migration sub-model and a heat transfer sub-model. The moisture migration sub-model uses the average moisture content of the billet as a state variable to describe the process by which moisture inside the billet migrates to the surface under the drive of the moisture concentration gradient and evaporates into the drying environment at the surface. The rate of moisture migration inside the billet is characterized by the effective moisture diffusion coefficient. The heat transfer sub-model uses the internal temperature and surface temperature of the billet as state variables to describe the convective heat transfer, moisture evaporation heat absorption, and internal heat conduction processes between the billet and the drying environment. The effective moisture diffusion coefficient is calibrated based on the porosity of the clay: the value of the effective moisture diffusion coefficient is determined by the porosity of the current batch of clay based on a pre-established porosity-effective diffusion coefficient calibration relationship. The porosity-effective diffusion coefficient calibration relationship is pre-calibrated by conducting drying tests on clay samples with different porosities.

3. The method for predicting and controlling the temperature and humidity during the drying process of a ceramic body according to claim 2, characterized in that: In step S2, the rolling prediction includes: The moisture migration time constant is determined based on the green body thickness and the clay porosity. The moisture migration time constant increases with the increase of the green body thickness and decreases with the increase of the effective moisture diffusion coefficient. The prediction time domain length is determined by the moisture migration time constant, so that the prediction time domain length covers the complete response process of the billet moisture content. Within the prediction time domain, the current real-time moisture content, internal temperature, and surface temperature of the billet are taken as the initial state, and the current ambient temperature, ambient humidity, wind speed, and changes in the control parameters of the actuator are taken as inputs. The predicted values ​​of the billet moisture content and the internal and external temperature differences at each moment within the prediction time domain are obtained through step-by-step recursive calculation using the lumped parameter coupling model. After each control cycle, the initial state is updated with the latest collected state and the recursive calculation is performed again to achieve rolling prediction.

4. The method for predicting and controlling temperature and humidity during the drying process of ceramic green bodies according to claim 1, characterized in that: In step S3, the automatic identification of the current drying stage of the green body includes: A sliding window filter is applied to the time series of real-time moisture content of the billet, and the time derivative of the filtered moisture content time series is obtained to obtain the moisture evaporation rate series. After the drying process enters the constant-rate drying stage, the stable value of the moisture evaporation rate sequence is recorded as the reference rate for the constant-rate segment. When the moisture evaporation rate is continuously lower than the first proportion of the constant speed section reference rate and the time derivative of the moisture evaporation rate is continuously negative, it is determined that the green body has entered the first deceleration drying stage from the constant speed drying stage. When the moisture evaporation rate is continuously lower than the second proportion of the constant rate section reference rate, and the moisture migration inside the billet changes from surface evaporation control to internal diffusion control, it is determined that the billet has entered the second rate-decreasing drying stage from the first rate-decreasing drying stage; the second proportion is less than the first proportion.

5. The method for predicting and controlling the temperature and humidity during the drying process of a ceramic body according to claim 4, characterized in that: In step S3, the safety constraints corresponding to each drying stage include: Constant-rate drying stage: Set a lower limit constraint on ambient humidity and an upper limit constraint on ambient temperature. The lower limit constraint on ambient humidity is used to prevent the surface of the green body from losing water too quickly and forming a dry shell. The upper limit constraint on ambient temperature is used to limit the intensity of surface moisture evaporation. First slow-down drying stage: Set upper limit constraints on the internal and external temperature difference between the internal temperature and the surface temperature of the billet and upper limit constraints on the ambient temperature rise rate. The upper limit constraint on the internal and external temperature difference is used to limit the internal thermal stress of the billet, and the upper limit constraint on the rise rate is used to prevent the internal moisture vaporization pressure of the billet from rising sharply. The second stage of slow-down drying involves setting an upper limit constraint on ambient temperature and wind speed adjustment rules. The wind speed adjustment rules include gradually increasing the wind speed as the real-time moisture content of the billet decreases, in order to enhance the removal of surface-migrated moisture.

6. The method for predicting and controlling temperature and humidity during the drying process of ceramic green bodies according to claim 4, characterized in that: In step S4, the temperature and humidity decoupling control is a fuzzy decoupling control, including: Using temperature deviation and its rate of change, and humidity deviation and its rate of change as inputs, heating control quantities and humidification control quantities are generated respectively through fuzzy reasoning. Based on the coupling coefficient characterizing the mutual interference strength between temperature regulation and humidity regulation, the heating control quantity and the humidification control quantity are cross-compensated: the influence of humidity regulation on temperature is converted according to the coupling coefficient and then superimposed on the heating control quantity, and the influence of temperature regulation on humidity is converted according to the coupling coefficient and then superimposed on the humidification control quantity to generate the decoupled basic control quantity; The quantization factor, scaling factor, and coupling coefficient of the fuzzy inference constitute a decoupling parameter set. Multiple decoupling parameter sets are pre-stored, each corresponding to a drying stage. When the drying stage switching is identified in step S3, the system synchronously switches to the decoupling parameter set corresponding to the current drying stage.

7. The method for predicting and controlling temperature and humidity during the drying process of ceramic green bodies according to claim 1, characterized in that: In step S5, the feedforward correction includes: Within the future time period (i.e., the prediction time domain) predicted in step S2, the deviation trend of the predicted value of the billet moisture content and the predicted value of the internal and external temperature difference from the applied safety constraint boundary is examined moment by moment to determine the moment when the predicted state approaches or crosses the safety constraint boundary and the direction of deviation. When it is predicted that the rate of decrease in the moisture content of the billet will exceed the allowable rate of water loss in the current drying stage, the heating power should be reduced and / or the humidification rate increased in advance. When it is predicted that the internal and external temperature difference will approach the temperature difference boundary in the safety constraint, the increase in heating power is reduced in advance and / or the fan frequency is adjusted to change the heat transfer intensity of the surface layer of the billet. When the predicted rate of decrease in the moisture content of the billet is lower than the allowable rate of water loss in the current drying stage and none of the predicted states are close to the safety constraint boundary, the heating power is increased and / or the humidification is reduced in advance to accelerate the drying process while ensuring safety. After each control cycle, the feedforward correction is recalculated based on the rolling updated predicted trajectory and executed on a rolling basis.

8. The method for predicting and controlling temperature and humidity during the drying process of ceramic green bodies according to claim 1, characterized in that: In step S6, the online correction includes: The deviation sequence between the measured values ​​and the corresponding predicted values ​​of the real-time moisture content and internal temperature of the billet is continuously calculated. When the magnitude of the deviation sequence continuously exceeds a preset deviation threshold, model parameter correction is triggered; The model parameter correction adopts a recursive least squares algorithm with a forgetting factor or an equivalent recursive parameter estimation algorithm: the collected environmental parameters, actuator control quantities and billet state quantities are used as input and output data to recursively estimate and update the model parameters of the moisture-temperature coupled prediction model. The forgetting factor is less than 1, which is used to reduce the weight of historical data and enhance the model parameters' ability to track batch changes, loading changes and seasonal environmental changes. The corrected model parameters are used for rolling prediction in the next control cycle.

9. The method for predicting and controlling temperature and humidity during the drying process of ceramic green bodies according to claim 1, characterized in that: In step S1: The real-time moisture content of the billet is obtained through one or a combination of the following methods: online weighing of the billet during drying or samples from the same batch being dried, and the real-time moisture content is calculated from the difference between the initial wet weight and the current weight of the billet; resistance method by measuring the resistance value between the surface and interior of the billet and calculating the real-time moisture content according to a pre-calibrated resistance-moisture content relationship; and microwave measurement method by using a microwave moisture sensor to non-contactly measure the moisture content of the billet. The internal temperature of the billet is obtained by measuring a temperature measuring element pre-embedded inside the billet during billet forming, and the surface temperature of the billet is obtained by measuring a non-contact temperature measuring element or a temperature measuring element attached to the surface of the billet.

10. The method for predicting and controlling the temperature and humidity during the drying process of a ceramic body according to claim 1, characterized in that: In step S6, the termination of drying includes: When both the real-time moisture content and the predicted moisture content of the billet drop to the target moisture content, and the deviation between the two is within the allowable range, the moisture content of the billet is deemed to meet the standard. Once the standard is met, the process enters the moisture equalization and final stage: heating is stopped, the fan is kept running at the set speed for the set time, and the ambient humidity is kept at a minimum set final humidity level. This allows the residual moisture inside the billet to be redistributed evenly in the thickness direction of the billet. After the billet temperature drops to near room temperature, the drying process ends.

Images