Infrared hot air cooperative drying control method based on uniformity driving cooperative control architecture

Abstract

The invention provides an infrared hot air cooperative drying control method based on a uniformity driving cooperative control framework. The method comprises the following steps: acquiring temperature and water content states on a unified cavity coordinate grid, generating multiple groups of infrared power and hot air temperature / air volume alternative schemes according to a partition adjustable range, and mapping the alternative schemes into an equipment action boundary diagram; inputting the current state and the boundary diagram into a prediction network containing space coordinate coding, infrared small neighborhood convolution, hot air fusion building function, large neighborhood convolution and field propagation to obtain future temperature and moisture content distribution; under the constraints of the maximum allowable temperature, the target water content and the adjustable range, sorting is carried out according to uniformity evaluation values, optimization is selected according to the combined energy consumption related quantity during paralleling, a partition instruction and an infrared hot air proportion are generated and issued, and a closed loop is formed by periodic acquisition and updating; according to the method, infrared and hot air collaboration, space alignment prediction and reproducible mapping are achieved, the drying uniformity is improved, and the control consistency under threshold constraint is guaranteed.

Description

Technical Field

[0001] This invention relates to the field of drying process control technology, and in particular to an infrared hot air collaborative drying control method based on a uniformity-driven collaborative control architecture. Background Technology

[0002] Infrared and hot air combined drying is widely used in belt and box-type equipment for dehydrating and heating continuous or batch materials. The equipment is typically divided into multiple zones along the conveying direction, with infrared power, hot air temperature, and airflow adjusted separately to adapt to the initial temperature and humidity state of the materials and the conveying speed. While ensuring the maximum allowable temperature and target moisture content, it is necessary to pay attention to the spatial uniformity of temperature and moisture content distribution within the chamber, and establish a spatial representation consistent with the area array observation data and a traceable control link. In actual processes, factors such as equipment layout, zone coverage length, and arrival / dwelling sequence can easily lead to localized overheating or moisture retention.

[0003] Existing technologies often employ zone-level control combined with empirical settings: PID or feedforward control of power, air temperature, and air volume is configured within the infrared and hot air zones, with speed switching and sequential control referenced to inlet / outlet measurements and conveyor belt speed; some schemes introduce simplified heat and mass transfer models or digital twins to maintain temperature and moisture content at the zone scale; others generate control schemes using predefined infrared / hot air ratios and zone setting tables, applying effects to the cavity according to the zone coverage area, and evaluating and switching based on indicators such as average temperature and average moisture content. Furthermore, common schemes use rule bases to set infrared array switches or duty cycles, fan and valve openings, and heating unit settings, combining arrival / dwell time calculations to achieve timing coordination, maintaining output within the control cycle and collecting observation data at the end of the cycle for the next adjustment.

[0004] The aforementioned solutions generally lack the representation of equipment action boundaries aligned with the material state space. Inconsistent conflict handling occurs when partitions overlap, making it difficult to evaluate effects on a unified grid. Partitioned or overall models struggle to provide grid-level future temperature and humidity field evolution and lack an optimization mechanism centered on spatial uniformity under constraints of maximum temperature and target moisture content. Furthermore, the lack of clear constraints on the proportion of infrared and hot air synergy during implementation leads to insufficient consistency between decision-making and implementation.

[0005] Therefore, an infrared hot air synergistic drying control method that can overcome the shortcomings of the existing technology is a problem that needs to be solved by those skilled in the art. Summary of the Invention

[0006] One objective of this invention is to propose an infrared hot air collaborative drying control method based on a uniformity-driven collaborative control architecture. The core technical problem to be solved by this application is: in the infrared hot air collaborative drying scenario, how to accurately map the adjustable parameters and spatial layout of the equipment partitions to the action boundary aligned with the material state grid, and based on this, to carry out spatiotemporal evolution prediction and uniformity-driven constraint optimization, select a reproducible and consistent control scheme, and iterate in a closed loop until the temperature and moisture content thresholds are met. This is the core technical problem that this application intends to solve.

[0007] The infrared hot air collaborative drying control method based on a uniformity-driven collaborative control architecture according to embodiments of the present invention includes:

[0008] S1. Obtain the temperature distribution and moisture content distribution of the material in the drying chamber to form the current material state diagram. Obtain the spatial layout of the infrared zone and the hot air zone and their respective adjustable ranges to form the adjustable parameters of the equipment.

[0009] S2. Generate multiple sets of alternative control schemes for infrared zone power and hot air zone temperature and volume based on the adjustable parameters of the equipment, and map each alternative control scheme to the coordinates of the drying chamber according to the spatial layout to form the corresponding equipment action boundary diagram.

[0010] S3. Input the current material state diagram and the equipment action boundary diagram corresponding to the alternative control scheme into the spatiotemporal neural field drying evolution prediction network to obtain the corresponding future temperature distribution and future moisture content distribution. The spatiotemporal neural field drying evolution prediction network includes a spatial coordinate input unit, an infrared action input unit, a hot air action input unit, and a field evolution output unit. The spatial coordinate input unit generates position features, the infrared action input unit and the hot air action input unit generate boundary features respectively, and the field evolution output unit merges the position features and boundary features to output the future temperature distribution and future moisture content distribution.

[0011] S4. Calculate the uniformity evaluation value based on the future temperature distribution and future moisture content distribution corresponding to the alternative control schemes. Under the constraints of the maximum allowable temperature threshold and the target moisture content threshold, screen the alternative control schemes in conjunction with the adjustable range defined by the equipment action boundary diagram, and determine the control commands to be executed. The control commands to be executed include infrared zone power, hot air zone air temperature and air volume, and infrared hot air ratio.

[0012] S5. Send the execution control command to the infrared zone heater and the hot air zone fan and execute it. At the end of the control cycle, obtain the new temperature distribution and the new moisture content distribution to form an updated material state diagram.

[0013] S6. Calculate the measured uniformity evaluation value based on the updated material state diagram. When the new moisture content distribution meets the target moisture content threshold and the measured uniformity evaluation value is less than the uniformity threshold, a drying completion judgment is formed. When the condition is not met, a new alternative control scheme is generated based on the updated material state diagram and the control instruction to be executed is determined.

[0014] Optionally, S1 is as follows:

[0015] The partition temperature, partition power, and emission array status of each partition of the infrared drying equipment are obtained, and the partition temperature, partition power, and emission array status are respectively composed into partition temperature vector, partition power vector, and emission array status vector according to the number of partitions.

[0016] The system acquires the conveyor belt speed and air volume, as well as the material inlet temperature, material inlet moisture content, material outlet temperature, and material outlet moisture content. The system then uses zone temperature vector, zone power vector, transmitter array state vector, air volume, conveyor belt speed, material inlet temperature, material inlet moisture content, material outlet temperature, and material outlet moisture content to form the current state of the equipment process.

[0017] Based on the adjacent order and coverage length of the infrared drying equipment zones, and combined with the conveyor belt speed, calculate the material zone arrival order table, and align the material zone arrival order table with the current state of the equipment process.

[0018] The actual control quantities are formed by the partition power vector, conveyor belt speed and air volume, and the current state of the equipment process and the actual control quantities are used as inputs to the dynamic evolution digital twin model.

[0019] Optionally, S2 is as follows:

[0020] Based on the spatial layout of the infrared zone and the hot air zone, determine the zone coverage range under the coordinates of the drying chamber, and convert the zone coverage range into a zone coverage grid identifier that uses the same number of grid rows and grid columns as the current material state diagram.

[0021] Based on the adjustable range of infrared zone power, hot air zone temperature, and hot air zone air volume in the adjustable parameters of the equipment, multiple candidate setting values ​​are generated for each infrared zone and each hot air zone, and the candidate setting values ​​of each zone are combined in the order of adjacent zones to form multiple sets of alternative control schemes.

[0022] For each alternative control scheme, the proportion of infrared hot air is determined. The proportion of infrared hot air is calculated by the combination relationship between the infrared zone power and the hot air zone temperature and air volume within the alternative control scheme, and the proportion of infrared hot air is bound to the alternative control scheme.

[0023] The individual alternative control schemes are mapped to the drying chamber coordinates according to the partition coverage grid identifier. The partition settings that fall into the same grid position are written according to the priority of the partition coverage range, so as to obtain the distribution of infrared partition power on the drying chamber coordinates, the distribution of hot air partition temperature on the drying chamber coordinates, and the distribution of hot air partition air volume on the drying chamber coordinates.

[0024] The distribution of infrared zone power in the drying chamber coordinates, the distribution of hot air zone temperature in the drying chamber coordinates, and the distribution of hot air zone air volume in the drying chamber coordinates are superimposed by channels to form a device action boundary map, which is then used as the spatial alignment input of the infrared action input unit and the hot air action input unit.

[0025] Repeatedly perform partition coverage grid identification mapping and device action boundary map generation for all candidate control schemes to obtain a set of device action boundary maps corresponding to the candidate control schemes and output them to the spatiotemporal neural field drying evolution prediction network.

[0026] Optionally, S3 specifically refers to:

[0027] The coordinates of the drying chamber are used to generate a spatial coordinate map based on the number of grid rows and columns. Each grid position in the spatial coordinate map contains horizontal and vertical coordinates and corresponds to the grid position in the current material state map.

[0028] The spatial coordinate map is input into the spatial coordinate input unit, and the spatial coordinate map is encoded to obtain the position feature map, so that the position feature map is aligned with the position of each grid in the drying cavity;

[0029] The distribution of infrared partition power in the drying cavity coordinates in the equipment action boundary map is input into the infrared action input unit. Small neighborhood convolutional layers are stacked to generate an infrared boundary feature map, so that the infrared boundary feature map can characterize the spatial action of infrared radiation within the local coverage area.

[0030] The distribution of hot air zone temperature in the drying chamber coordinates and the distribution of hot air zone air volume in the drying chamber coordinates in the equipment action boundary map are input into the hot air action input unit. A hot air boundary feature map is generated by stacking large neighborhood convolutional layers, and the channel dimension of the hot air boundary feature map is aligned with the channel dimension of the infrared boundary feature map.

[0031] The current material state map is input into the initial state encoding layer, and the current material state map is convolutionally encoded to generate an initial state feature map, so that the initial state feature map represents the spatial initial state of the current temperature distribution and the current moisture content distribution.

[0032] The location feature map, infrared boundary feature map, hot air boundary feature map and initial state feature map are input into the field evolution output unit and spliced ​​through channels to form a fused feature map. The fused feature map is then input into the field propagation layer for spatial coupling propagation and outputs the propagated feature map.

[0033] The propagated feature map is input into the output layer, and pointwise convolution mapping is used to obtain the predicted output map. The predicted output map contains two channels: future temperature distribution and future moisture content distribution. The aforementioned input and inference are performed on the equipment action boundary map corresponding to each alternative control scheme to obtain the future temperature distribution and future moisture content distribution corresponding to that alternative control scheme.

[0034] Optionally, when inputting the distribution of hot air zone temperature and hot air zone air volume in the drying chamber coordinates from the equipment action boundary diagram into the hot air action input unit, a construction function is used to process the hot air zone temperature and hot air zone air volume into the input of the hot air action input unit. Specifically, the construction function is:

[0035] ;

[0036] in, The hot air fusion input diagram is shown in the first... Line number The numerical value of the column grid position. Represents the grid row index and its value range is to , Indicates the grid column index and its value range is to , Represents the boundary diagram of equipment operation The distribution of hot air zone temperature on the drying chamber coordinates at the grid position The wind temperature value, Represents the boundary diagram of equipment operation The distribution of hot air zoning air volume in the drying chamber coordinates at the grid position The air volume value, This represents the global minimum air temperature within the adjustable range of the hot air zone defined by the adjustable parameters of the device. This represents the global maximum air temperature within the adjustable range of the hot air zone defined by the adjustable parameters of the device. This represents the global minimum airflow rate within the adjustable range of the hot air zone as defined by the device's adjustable parameters. This represents the global maximum airflow within the adjustable range of the hot air zone defined by the device's adjustable parameters. This represents the preset hot air volume spatial variation weighting coefficient, used to adjust the contribution of air volume variations at adjacent grid positions to the hot air fusion input map. , , , When the location is outside the grid boundary, the corresponding outside grid boundary location will be... Replace with .

[0037] Optionally, S4 specifically refers to:

[0038] Using the alternative control scheme as an index, read the future temperature distribution and future moisture content distribution corresponding to the alternative control scheme respectively, and read the equipment action boundary diagram corresponding to the alternative control scheme as the adjustable range constraint input;

[0039] The maximum allowable temperature threshold constraint is determined for the future temperature distribution. The predicted maximum temperature value is extracted by traversing the grid positions of the future temperature distribution point by point. The candidate control scheme with the predicted maximum temperature value greater than the maximum allowable temperature threshold is eliminated.

[0040] The target moisture content threshold constraint is used to determine the future moisture content distribution. The predicted maximum moisture content value is extracted by traversing the grid positions of the future moisture content distribution point by point. The candidate control schemes with the predicted maximum moisture content value greater than the target moisture content threshold are eliminated.

[0041] For the alternative control schemes that pass the threshold constraint, the adjustable range consistency judgment is performed. The distribution of infrared zone power in the drying chamber coordinates, the distribution of hot air zone temperature in the drying chamber coordinates, and the distribution of hot air zone air volume in the drying chamber coordinates in the equipment action boundary diagram are compared point by point with the upper and lower limits of the adjustable parameters of the equipment. Alternative control schemes with grid positions that exceed the boundaries are eliminated.

[0042] For the alternative control schemes that pass the consistency judgment, calculate the uniformity evaluation value, use the temperature spatial dispersion of the future temperature distribution to obtain the temperature uniformity sub-index, use the moisture content spatial dispersion of the future moisture content distribution to obtain the moisture content uniformity sub-index, and combine the temperature uniformity sub-index and the moisture content uniformity sub-index according to the preset weight to form the uniformity evaluation value.

[0043] The uniformity evaluation value is used as the sorting key to sort the candidate control schemes that have passed the screening. The candidate control scheme with the smallest uniformity evaluation value is selected as the target candidate control scheme. When the uniformity evaluation values ​​are the same, the target candidate control scheme with the smaller correlation between the combined energy consumption of infrared zone power and hot air zone temperature and air volume is selected.

[0044] The target alternative control scheme is converted into an execution control command. The execution control command includes the infrared zone power, hot air zone temperature and air volume that are consistent with the target alternative control scheme. After determining the proportion of infrared hot air based on the combination relationship between infrared zone power and hot air zone temperature and air volume in the target alternative control scheme, the execution control command is output.

[0045] Optionally, when calculating the uniformity evaluation value for alternative control schemes that pass the threshold constraint, an optimization function is used to synthesize the temperature uniformity sub-index and the moisture content uniformity sub-index according to a preset weight to obtain the uniformity evaluation value, wherein the optimization function is specifically:

[0046] ;

[0047] in, Indicates alternative control schemes The uniformity evaluation value, This represents the temperature uniformity sub-index and corresponds to the value in the above formula. Weighted bracketed items, This represents the sub-index of moisture content uniformity and corresponds to the expression in the above formula. Weighted bracketed items, This represents the preset weight of the temperature uniformity sub-index. This indicates the preset weight of the moisture content uniformity sub-index. Indicates the number of grid rows. Indicates the number of grid columns. Indicates the grid row index. Indicates the grid column index. and This represents the traversal index used to calculate the grid average. This indicates the future temperature distribution at the grid location. Temperature value, This indicates the future moisture content distribution at the grid locations. The moisture content value, Indicates the maximum permissible temperature threshold. Indicates the target moisture content threshold. This represents the preset weighting coefficients for adjacent temperature difference terms. This represents the preset weighting coefficient for adjacent difference terms in moisture content, when the index... or When the location is outside the grid boundary, the corresponding outside grid boundary location will be... or Replace with or .

[0048] Optional, S5 specifically includes:

[0049] The execution control command is parsed into infrared zone power, hot air zone temperature and air volume, and infrared hot air ratio. The infrared zone power is sent to the corresponding infrared zone heater, and the hot air zone temperature and air volume are sent to the corresponding hot air zone fan.

[0050] Within the control cycle, the infrared zone heater and hot air zone fan are driven to output according to the executed control command, so that the infrared effect and hot air effect in the drying chamber are consistent with the executed control command.

[0051] At the end of the control cycle, acquire the new temperature distribution and the new moisture content distribution, and map the new temperature distribution and the new moisture content distribution to the current material state according to the drying chamber coordinates. Figure 1 The number of grid rows and columns;

[0052] The mapped new temperature distribution and the new moisture content distribution are superimposed on each channel to form an updated material state diagram. This updated material state diagram is then used as input for the subsequent regeneration of alternative control schemes and the prediction of future temperature and moisture content distributions.

[0053] Optionally, step S6 specifically includes:

[0054] The updated material state diagram is split into channels to obtain new temperature distribution and new moisture content distribution. The highest measured moisture content value of the new moisture content distribution is extracted by point-by-point traversal. The measured uniformity evaluation value is calculated based on the spatial dispersion of the new temperature distribution and the new moisture content distribution at the grid position.

[0055] The measured maximum moisture content value is constrained by the target moisture content threshold, and the measured uniformity evaluation value is constrained by the uniformity threshold. When both the target moisture content threshold and the uniformity threshold are satisfied, the drying completion determination is formed.

[0056] Before a drying completion determination is formed, the updated material state diagram is used as the current material state diagram to generate alternative control schemes and generate equipment action boundary diagrams corresponding to the alternative control schemes. The current material state diagram and equipment action boundary diagram are input into the spatiotemporal neural field drying evolution prediction network to obtain the future temperature distribution and future moisture content distribution. Under the constraints of the maximum allowable temperature threshold and the target moisture content threshold, the control command to be executed is selected based on the uniformity evaluation value.

[0057] The beneficial effects of this invention are:

[0058] (1) This proposal puts forward an improved spatiotemporal neural field drying evolution prediction method and network structure. It adopts multi-scale feature coupling of spatial coordinate point-by-point encoding, infrared small neighborhood convolution and hot air large neighborhood convolution, and introduces a hot air fusion construction function that normalizes the range of wind temperature and air volume and weights the air volume changes of adjacent grids. The field propagation layer realizes the coupled evolution of position features, boundary features and initial state features. After the improvement, it can more accurately predict the future temperature and moisture content distribution on a unified grid, improve the consistency of spatial alignment, identify local overheating and moisture content retention in advance, and provide a reliable prediction basis for subsequent threshold constraints and uniformity optimization. Compared with traditional algorithms that infer from partition average or overall model, this network uses the device action boundary map as spatial alignment input, reducing the estimation bias caused by partition coverage overlap and write conflict.

[0059] (2) This proposal puts forward a novel uniformity-driven infrared-hot air coordinated control method and technology. It uses the equipment action boundary map as an adjustable range constraint, combines point-by-point judgment of the maximum allowable temperature and target moisture content threshold, and employs a uniformity evaluation function including grid average deviation and adjacent difference terms for ranking. In the case of parallel operation, it introduces the energy consumption related to the infrared and hot air combination as the selection criterion, and outputs the infrared-hot air ratio to constrain the coordination ratio during the execution period. This method, while ensuring the consistency of safety thresholds and zoning upper and lower limits, selects a control scheme with better spatial uniformity and more reasonable energy consumption, reducing the risk of local overheating and uneven drying, improving the reproducibility and consistency of decision-making, and demonstrating a difference from traditional experience-based or single-objective optimization methods.

[0060] (3) This proposal puts forward an overall closed-loop control method and technology based on a unified drying chamber coordinate grid, equipment action boundary map mapping and dynamic evolution digital twin. It realizes an iterative closed loop of process acquisition, alternative scheme generation, spatiotemporal prediction, constraint screening, partition command issuance and status update within the S1 to S6 cycle. The writing conflict is resolved by partition coverage priority rules, and observation mapping and missing sampling filling are realized by grid aggregation or weighted averaging. The overall method maintains spatial alignment and temporal traceability under different equipment layouts and partition scales, reduces trial and error and parameter dependence, and enables infrared and hot air to stably coordinate according to proportion constraints within the control cycle. When the measured uniformity and moisture content reach the threshold, the process is terminated in time to improve the drying quality stability and control link consistency. Attached Figure Description

[0061] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0062] Figure 1This is a flowchart of an infrared hot air collaborative drying control method based on a uniformity-driven collaborative control architecture proposed in this invention.

[0063] Figure 2 This is a flowchart showing the alignment of the current state of the equipment process with the material arrival order table for an infrared hot air collaborative drying control method based on a uniformity-driven collaborative control architecture proposed in this invention.

[0064] Figure 3 This is a flowchart of the alternative control scheme and equipment action boundary diagram generation process for an infrared hot air collaborative drying control method based on a uniformity-driven collaborative control architecture proposed in this invention.

[0065] Figure 4 Drawings for a preheated fluidized bed. Detailed Implementation

[0066] In Example 1, reference Figures 1 to 3 An infrared hot air collaborative drying control method based on a uniformity-driven collaborative control architecture, wherein the equipment uses a preheated fluidized bed, such as... Figure 4 As shown, it includes:

[0067] S1. Obtain the temperature distribution and moisture content distribution of the material in the drying chamber to form the current material state diagram. Obtain the spatial layout of the infrared zone and the hot air zone and their respective adjustable ranges to form the adjustable parameters of the equipment.

[0068] S2. Generate multiple sets of alternative control schemes for infrared zone power and hot air zone temperature and volume based on the adjustable parameters of the equipment, and map each alternative control scheme to the coordinates of the drying chamber according to the spatial layout to form the corresponding equipment action boundary diagram.

[0069] S3. Input the current material state diagram and the equipment action boundary diagram corresponding to the alternative control scheme into the spatiotemporal neural field drying evolution prediction network to obtain the corresponding future temperature distribution and future moisture content distribution. The spatiotemporal neural field drying evolution prediction network includes a spatial coordinate input unit, an infrared action input unit, a hot air action input unit, and a field evolution output unit. The spatial coordinate input unit generates position features, the infrared action input unit and the hot air action input unit generate boundary features respectively, and the field evolution output unit merges the position features and boundary features to output the future temperature distribution and future moisture content distribution.

[0070] S4. Calculate the uniformity evaluation value based on the future temperature distribution and future moisture content distribution corresponding to the alternative control schemes. Under the constraints of the maximum allowable temperature threshold and the target moisture content threshold, screen the alternative control schemes in conjunction with the adjustable range defined by the equipment action boundary diagram, and determine the control commands to be executed. The control commands to be executed include infrared zone power, hot air zone air temperature and air volume, and infrared hot air ratio.

[0071] S5. Send the execution control command to the infrared zone heater and the hot air zone fan and execute it. At the end of the control cycle, obtain the new temperature distribution and the new moisture content distribution to form an updated material state diagram.

[0072] S6. Calculate the measured uniformity evaluation value based on the updated material state diagram. When the new moisture content distribution meets the target moisture content threshold and the measured uniformity evaluation value is less than the uniformity threshold, a drying completion judgment is formed. When the condition is not met, a new alternative control scheme is generated based on the updated material state diagram and the control instruction to be executed is determined.

[0073] In this embodiment, step S1 specifically includes:

[0074] The infrared drying equipment is divided into zones along the conveyor belt direction, with the following number of zones: The system comprises multiple infrared zones, each equipped with a zone temperature acquisition channel, a zone power acquisition channel, and a transmitter array status acquisition channel. The controller reads the zone temperature value from each infrared zone at a control cycle interval. to Read partition power value to Read the status value of the transmission array to Among them, the transmit array status value The system employs a numerical representation that can be directly used for modeling and control. This numerical representation includes switch state values ​​and duty cycle state values. Switch state values ​​indicate the on / off state of the corresponding transmitter array, while duty cycle state values ​​indicate the conduction ratio of the corresponding transmitter array within a cycle. The controller will... to The temperature vectors of the zones are arranged in order of their zone numbers. ,Will to The partition power vector is formed according to the partition number order. ,Will to The transmit array state vector is composed according to the partition number order. This ensures that the vector subscripts are in the same order as the physical adjacency of the infrared partitions;

[0075] The controller further reads the conveyor belt speed. With air volume And obtain the material inlet temperature at the material inlet location. With the moisture content of the material inlet Obtain the material outlet temperature at the material outlet location. With the moisture content of the material outlet The controller will partition the temperature vector. Partition power vector Transmit array state vector Air volume Conveyor belt speed Material inlet temperature Moisture content of material inlet Material outlet temperature Moisture content of material outlet The current state of the device process is formed by combining fixed fields in a specific order. Among them, the current state of the equipment process The partition field is determined by , , Composed of, single-value fields are , , , , , This configuration enables the dynamic evolution digital twin model to receive input simultaneously at both the partition scale and the whole-machine scale;

[0076] To generate a material arrival sequence table, the controller reads the coverage length of each infrared zone in the conveyor belt direction from the equipment space layout parameters. to It also reads the cumulative distance from the material inlet to the starting boundary of each infrared zone. to The controller is based on the conveyor belt speed The arrival time of each infrared zone is calculated, where the arrival time is the ratio of the cumulative distance to the conveyor belt speed. The controller then determines the arrival time based on the coverage length. The controller calculates the dwell time for each infrared zone, which is the ratio of the coverage length to the conveyor belt speed. The controller further calculates the departure time, which is the sum of the arrival time and the dwell time. The controller then combines the zone number, arrival time, and departure time in adjacent order to form a material zone arrival order table. ;

[0077] The controller uses a material zone arrival sequence table. The index mapping method will determine the arrival order table of material partitions. Current status of the equipment process Alignment, the alignment including with Middle partition number and Establish a correspondence between vector subscripts and use Arrival time and departure time are used as corresponding partitions in the current conveyor belt speed. The time window under the action ensures that the action sequence of the equipment side and the heating sequence of the material side have a traceable consistency on the same time axis, thereby supporting the subsequent mapping of alternative control schemes into the equipment action boundary diagram according to the spatial layout and spatially aligning them with the current material state diagram.

[0078] Current state of the aligned device process Subsequently, the controller uses partitioned power vectors. Conveyor belt speed With air volume Composition of actual control quantity The actual control quantity In accordance with the execution amount of the control commands, the controller will maintain the current state of the equipment process. With actual control quantity As input to the dynamic evolution digital twin model, it enables the model to maintain the implicit state of the material partition temperature vector and the material partition moisture content vector in a rolling update, and provides a process-side input basis for subsequent spatial distribution prediction and threshold constraint screening in combination with the equipment action boundary map.

[0079] In this embodiment, step S2 specifically includes:

[0080] A unified coordinate system for the drying chamber is established, and the coordinates are arranged according to the number of grid rows. The number of grid columns is The two-dimensional grid discretization ensures that the equipment action boundary diagram and the current material state diagram use the same spatial resolution. The number of infrared partitions is denoted as... The number of hot air zones is recorded as The controller reads the coverage area of ​​each infrared zone and each hot air zone in the drying chamber coordinates from the equipment space layout parameters. The coverage area is described by the start and end boundaries in the drying chamber coordinates. The controller converts the coverage area of ​​each zone into a zone coverage grid identifier. The zone coverage grid identifier of the infrared zone is denoted as... The hot air zone's zone coverage grid identifier is denoted as ,in Infrared zone numbering, Numbering of hot air zones and Each represents the set of grid index pairs covered by that partition;

[0081] The controller reads the adjustable range of power for each infrared zone, the adjustable range of air temperature for each hot air zone, and the adjustable range of airflow for each hot air zone based on the adjustable parameters of the equipment. This information is then used to adjust the infrared zone settings. The controller discretizes the adjustable range of the infrared zone power into a sequence of candidate setpoints according to a preset number of levels. This sequence of candidate setpoints is denoted as... ,in Each candidate setting value is denoted as , Assign gear number to each hot air zone The controller discretizes the adjustable range of the hot air zone temperature into a sequence of candidate setpoints. Each candidate setting value is denoted as The controller discretizes the adjustable range of the hot air zone's airflow into a sequence of candidate setpoints. Each candidate setting value is denoted as The candidate setting values ​​are generated using a fixed increment method, so that each candidate setting value is an execution setting value that can be directly issued.

[0082] The controller combines candidate setpoints for each partition in adjacent order to form multiple sets of alternative control schemes. A single set of alternative control schemes is denoted as […]. ,in Alternative control scheme number, alternative control scheme The infrared part consists of the power setting value of each infrared zone, and the hot air part consists of the air temperature setting value and the air volume setting value of each hot air zone. The setting value sequence index is kept consistent with the infrared zone number and the hot air zone number, so as to ensure that spatial writing can be performed according to the zone coverage grid identifier in the future.

[0083] For each alternative control scheme The controller calculates the proportion of infrared hot air, and the proportion of infrared hot air is recorded as follows: The controller The infrared input intensity index is obtained by summing the power settings of all infrared zones. The controller The temperature and air volume setpoints of all hot air zones are merged. The fusion method uses the air volume setpoint as the weight to sum the temperature setpoints, resulting in a hot air input intensity index. Among them, hot air zoning Contribution amount The setting value and The controller multiplies the setpoints and sums them over all hot air zones, then... and Determining the normalized ratio ,Pick for and The ratio, and Alternative control schemes Binding enables the execution control commands obtained through subsequent filtering to directly carry the proportion of infrared hot air;

[0084] The controller will select a single alternative control scheme. Mapping to the drying chamber coordinates generates a distribution map. The controller first creates three two-dimensional distribution maps consistent with the grid, representing the distribution of infrared zone power, hot air zone temperature, and hot air zone airflow in the drying chamber coordinates, respectively. These three distribution maps are then initialized to zero values. Subsequently, the controller traverses the distribution maps in the order of infrared zone numbers. ,Will Write the power setting value of the corresponding infrared zone into The controller traverses the covered grid locations in order of hot air zone number. ,Will The temperature and airflow settings for the corresponding hot air zones are written into the system respectively. The location of the covered grid;

[0085] When multiple partitions conflict in writing to the same grid location, the controller writes according to the partition coverage priority. The partition coverage priority is generated by the geometry given by the device space layout parameters, and the geometry includes the distance of each infrared partition to the material surface. Distance from each hot air zone to the material surface The controller generates priorities using fixed rules, which are then sorted by distance from smallest to largest, by coverage length from largest to smallest, and by partition number from smallest to largest. Only the settings of the highest priority partition are retained at the conflict grid locations, thus ensuring the consistency and reproducibility of the spatial distribution of the device action boundary map in multiple generation processes.

[0086] The controller superimposes the distributions of infrared zone power, hot air zone temperature, and hot air zone airflow on the drying chamber coordinates to form a device action boundary diagram by channel. The action boundary diagram of a single device is denoted as... and make The grid position corresponds to the grid position of the current material state diagram, and the controller supports all alternative control schemes. Repeatedly execute the partition coverage grid identification mapping and equipment action boundary map generation to obtain the equipment action boundary map set corresponding to the alternative control scheme. The equipment action boundary map set is then output to the spatiotemporal neural field drying evolution prediction network. This allows the infrared action input unit to receive the distribution channel of infrared partition power on the drying cavity coordinates, and the hot air action input unit to receive the distribution channel of hot air partition temperature and hot air partition air volume on the drying cavity coordinates. This completes the conversion from alternative control schemes to spatially aligned boundary inputs.

[0087] In this embodiment, step S3 specifically includes:

[0088] The spatiotemporal neural field drying evolution prediction network uses a two-dimensional grid of drying cavity coordinates as a unified spatial reference, assuming the number of grid rows is... The number of grid columns is The controller determines the horizontal and vertical coordinate ranges of the grid in the coordinate system of the drying chamber, and linearly discretizes the coordinate ranges according to the grid rows and columns: generating a horizontal coordinate value for each column and a vertical coordinate value for each row, and combining the horizontal and vertical coordinate values ​​according to the grid positions to obtain a spatial coordinate map. Spatial coordinate diagram Each grid position contains two numerical channels: horizontal coordinate and vertical coordinate, and corresponds to the grid position of the current material state diagram;

[0089] The controller will display the spatial coordinate map. The input spatial coordinate input unit performs point-by-point encoding on the horizontal and vertical coordinates of each grid location. The point-by-point encoding adopts a cascaded structure of point-by-point convolution and nonlinear mapping, so that each grid location outputs a multi-channel position feature vector. The spatial coordinate input unit concatenates the point-by-point encoding results of all grid locations to form a position feature map. ,in Each grid location contains positional features obtained by mapping horizontal and vertical coordinates;

[0090] For any alternative control scheme, the controller reads the equipment action boundary diagram and records it as follows: ,in For alternative control scheme numbers, see equipment action boundary diagram. It has three channels: one for the distribution of infrared zone power in the drying chamber coordinates, one for the distribution of hot air zone temperature in the drying chamber coordinates, and one for the distribution of hot air zone airflow in the drying chamber coordinates. The controller will... The infrared partition power distribution channel is input to the infrared action input unit. The infrared action input unit is encoded using stacked small-neighbor convolutional layers. The neighborhood size of the small-neighbor convolution is denoted as... ,in The edge length of the convolution kernel on the grid is represented and fixed as a preset network parameter, so that the coverage range of the convolutional neighborhood corresponds to the local coverage range of the infrared partition, and the infrared boundary feature map is output. ;

[0091] The controller will The hot air zone temperature distribution channel and the hot air zone air volume distribution channel are input into the hot air action input unit. Before entering the large neighborhood convolutional layer stack, the hot air action input unit first fuses the distribution of hot air zone temperature in the drying cavity coordinates and the distribution of hot air zone air volume in the drying cavity coordinates into a single-channel hot air fusion input map. and will As input to the stack of large neighborhood convolutional layers, the fusion is constructed using a function, and the specific calculation method is as follows:

[0092] ;

[0093] in, The hot air fusion input diagram is shown in the first... Line number The numerical value of the column grid position. Represents the grid row index and its value range is to , Indicates the grid column index and its value range is to , Represents the boundary diagram of equipment operation The distribution of hot air zone temperature on the drying chamber coordinates at the grid position The wind temperature value, Represents the boundary diagram of equipment operation The distribution of hot air zoning air volume in the drying chamber coordinates at the grid position The air volume value, This represents the global minimum air temperature within the adjustable range of the hot air zone defined by the adjustable parameters of the device. This represents the global maximum air temperature within the adjustable range of the hot air zone defined by the adjustable parameters of the device. This represents the global minimum airflow rate within the adjustable range of the hot air zone as defined by the device's adjustable parameters. This represents the global maximum airflow within the adjustable range of the hot air zone defined by the device's adjustable parameters. This represents the preset hot air volume spatial variation weighting coefficient, used to adjust the contribution of air volume variations at adjacent grid positions to the hot air fusion input map. , , , When the location is outside the grid boundary, the corresponding outside grid boundary location will be... Replace with ;

[0094] Hot air input unit for hot air fusion input diagram Encoding is performed using stacked large-neighbor convolutional layers, where the neighborhood size of the large-neighbor convolution is denoted as . ,in The convolution kernel's edge length on the grid is represented and fixed as a preset network parameter, so that the convolutional neighborhood coverage corresponds to the spatial diffusion range of the thermal air mixing and transport effects, and the thermal air boundary feature map is output. The hot air input unit introduces a channel alignment layer at the output end. The channel alignment layer uses point-by-point convolution to... The number of channels is mapped to the number of channels. Same number of channels;

[0095] The controller inputs the current material state map into the initial state encoding layer. The current material state map contains two channels: a temperature distribution channel and a moisture content distribution channel. The initial state encoding layer performs convolutional encoding on the two channels and outputs the initial state feature map. ,in Each grid location contains spatial initial state characteristics determined by the current temperature distribution and the current moisture content distribution;

[0096] The controller will use the location feature map Infrared boundary feature map Hot air boundary feature map With initial state feature map The input field evolution output unit first performs channel concatenation on the four types of feature maps in the channel dimension to form a fused feature map. The feature maps will then be fused. The input field propagation layer performs spatial coupling propagation and outputs the propagated feature map. ;

[0097] The controller will propagate the feature map Input and output layers, the output layer uses pointwise convolution pairs Perform channel mapping to obtain the predicted output image. Predicted output image It has two channels: a future temperature distribution channel and a future moisture content distribution channel, and maintains the same number of grid rows as the current material state diagram. With the number of grid columns The controller's boundary diagram for each alternative control scheme. Repeatedly execute the spatial coordinate map input, boundary feature generation, initial state feature generation, fusion propagation and output mapping to obtain the future temperature distribution and future moisture content distribution corresponding to the alternative control scheme.

[0098] In this embodiment, step S4 specifically includes:

[0099] The controller organizes the inference results of the spatiotemporal neural field drying evolution prediction network for each alternative control scheme into filterable data pairs, and numbers the alternative control schemes as follows: In this situation, the controller uses As an index, read the future temperature distribution corresponding to the alternative control scheme. With future moisture content distribution And read the equipment action boundary diagram corresponding to the alternative control scheme. As an adjustable range constraint input, the future temperature distribution With future moisture content distribution All are based on the number of grid rows. The number of grid columns is The grid representation of the device's operational boundary diagram The three channels represent the distribution of infrared zone power in the drying cavity coordinates, the distribution of hot air zone temperature in the drying cavity coordinates, and the distribution of hot air zone air volume in the drying cavity coordinates.

[0100] Controller for future temperature distribution To determine the maximum permissible temperature threshold constraint, the controller scans the grid position point by point. All grid locations Update the predicted highest temperature value during the traversal process. ,in The controller will take the maximum value among the temperature values ​​found in the traversed grid. With the maximum allowable temperature threshold When comparing, Greater than Alternative control schemes will be selected at that time. Mark it as a candidate to be eliminated and terminate the subsequent decision-making process for this alternative control scheme;

[0101] The controller further performs target moisture content threshold constraint determination on the alternative control schemes constrained by temperature. The controller scans the future moisture content distribution by traversing the grid positions point by point. All grid locations Update the predicted maximum moisture content value during the traversal process. ,in The controller will take the maximum value of the moisture content values ​​found in the traversed grid. With the target moisture content threshold When comparing, Greater than Alternative control schemes will be selected at that time. Mark as a culled object;

[0102] For alternative control schemes that pass the threshold constraint, the controller performs an adjustable range consistency determination to ensure the equipment's action boundary diagram. The controller reads the parameters consistent with the adjustable parameters of the equipment. Infrared partition power distribution channel values Hot air zone temperature distribution channel value Hot air zone air volume distribution channel value The controller determines the location of each grid based on the partition coverage grid identifier and the partition coverage priority. Determine the source partition number to write to; the determination process includes... Perform a partition coverage set retrieval and select a unique source partition number based on partition coverage priority, ensuring that each grid location... , , Each corresponds to a unique infrared zone number or hot air zone number. After obtaining the source zone number, the controller reads the upper and lower limits of the adjustable range of the infrared zone power, the upper and lower limits of the adjustable range of the hot air zone temperature, and the upper and lower limits of the adjustable range of the hot air zone airflow from the adjustable parameters of the device, and then... , , Point-by-point comparison; when any grid position exceeds the limit, alternative control schemes are selected. Mark as a culled object;

[0103] For alternative control schemes that pass the consistency criterion, the controller calculates the uniformity evaluation value. And obtain the temperature uniformity sub-index within the same computational link. Sub-index of moisture content uniformity Uniformity evaluation value Construct a single value that can be directly used for sorting based on the optimization function:

[0104] ;

[0105] in, Indicates alternative control schemes The uniformity evaluation value, This represents the temperature uniformity sub-index and corresponds to the value in the above formula. Weighted bracketed items, This represents the sub-index of moisture content uniformity and corresponds to the expression in the above formula. Weighted bracketed items, This represents the preset weight of the temperature uniformity sub-index. This indicates the preset weight of the moisture content uniformity sub-index. Indicates the number of grid rows. Indicates the number of grid columns. Indicates the grid row index. Indicates the grid column index. and This represents the traversal index used to calculate the grid average. This indicates the future temperature distribution at the grid location. Temperature value, This indicates the future moisture content distribution at the grid locations. The moisture content value, Indicates the maximum permissible temperature threshold. Indicates the target moisture content threshold. This represents the preset weighting coefficients for adjacent temperature difference terms. This represents the preset weighting coefficient for adjacent difference terms in moisture content, when the index... or When the location is outside the grid boundary, the corresponding outside grid boundary location will be... or Replace with or ;

[0106] The controller uses uniformity evaluation values The remaining alternative control schemes are sorted using the sorting key, and the alternative control scheme with the smallest uniformity evaluation value is selected as the target alternative control scheme. To handle cases where the uniformity evaluation values ​​are the same, the controller calculates the combined energy consumption related quantity for the alternative control schemes participating in the ranking. The controller sums the infrared power setpoints of the infrared zones within the candidate control schemes according to their infrared zone numbers to obtain a total infrared power value. It then normalizes this total infrared power value using the sum of the maximum power of each infrared zone to obtain the infrared energy consumption component. The controller also normalizes the hot air zone temperature and airflow setpoints for each hot air zone according to their hot air zone numbers. The normalized temperature is the ratio of the setpoint to the upper and lower limits of the adjustable temperature range for that zone, and the normalized airflow is the ratio of the setpoint to the upper and lower limits of the adjustable airflow range for that zone. The controller multiplies the normalized temperature and normalized airflow to obtain the hot air energy consumption component for that zone and sums these values ​​across all hot air zones to obtain the total hot air energy consumption component. Finally, the controller combines the infrared energy consumption component and the hot air energy consumption component according to a preset energy consumption weight to obtain the combined energy consumption correlation quantity. , and select The smaller alternative control scheme is selected as the target alternative control scheme;

[0107] The controller converts the target alternative control scheme into execution control commands. It extracts the infrared zone power setpoint sequence from the target alternative control scheme to form the infrared zone power control quantity, and extracts the hot air zone temperature setpoint sequence and hot air zone airflow setpoint sequence from the target alternative control scheme to form the hot air zone temperature and airflow control quantity. It also reads the infrared hot air percentage bound to the target alternative control scheme. As the infrared hot air ratio control quantity, the controller combines the infrared zone power control quantity, hot air zone air temperature and air volume control quantity and infrared hot air ratio control quantity to output the execution control command, so that the execution control command is consistent with the equipment execution quantity field and can be directly sent to the infrared zone heater and hot air zone fan.

[0108] In this embodiment, step S5 specifically includes:

[0109] The controller receives the execution control command output from step four and parses it into directly executable partition quantities. The controller also parses the infrared partition power into an infrared partition power command sequence. to ,in This represents the number of infrared zones. Indicates the first The controller interprets the power setpoints for each infrared zone into a sequence of hot air zone temperature and airflow command values. to Hot air zone air volume command sequence to ,in Number of hot air zones Indicates the first Temperature settings for each hot air zone Indicates the first The controller interprets the airflow setting value for each hot air zone into a percentage command. Percentage instruction It is a numerical parameter and is bound to the target alternative control scheme. It is used to constrain the coordinated output ratio of infrared zone power command and hot air zone temperature and air volume command within the controller.

[0110] The controller establishes corresponding distribution channels according to the partition number, and... The power execution interface of the corresponding infrared zone heater is sent respectively. and The controller will send the control signals to the corresponding hot air zone fan's temperature and air volume execution interfaces, respectively. The power execution interface for the infrared zone heater will be implemented using a duty cycle control or a power setting register. The output is converted to an acceptable execution value for the heater and written to the corresponding partition. The air temperature execution interface of the hot air partition fan is implemented using the heating unit setpoint or the heat exchange valve opening setpoint. The controller will... Converted to the corresponding execution value and written to the corresponding partition, the air volume execution interface uses the variable frequency setting value or the damper opening setting value, and the controller will... The corresponding execution quantity is converted and written to the corresponding partition. After the controller completes the issuance, it records the execution control instruction index within the control cycle, which is used to form a data closed loop of the same control cycle with the control input and the subsequently collected temperature distribution and moisture content distribution.

[0111] The controller has a control cycle duration of Within the control cycle, the controller drives the infrared zone heater and hot air zone fan to output power according to the executed control commands. The controller maintains power control of the infrared zone heater within the execution cycle, ensuring that the actual output power of the infrared zone heater revolves around... Stable, the controller maintains stable temperature and airflow control for the hot air zone fans during their execution cycles, ensuring that the actual temperature and airflow of the hot air zone remain stable. and Stable, the controller uses percentage commands As a collaborative constraint parameter, the execution priority of the infrared zone power command and the hot air zone temperature and air volume command is fixed within the control cycle, so that the infrared effect and hot air effect in the drying chamber remain consistent with the execution control command within the control cycle.

[0112] At the end of the control cycle, the controller acquires the new temperature distribution. With the new moisture content distribution New temperature distribution The data was collected by an array temperature acquisition device positioned at the observation location within the drying chamber. The acquisition results are expressed in coordinates of the drying chamber, showing a new moisture content distribution. The moisture content data was collected by an array of sensors positioned at the observation points within the drying chamber. The results were expressed in coordinates of the drying chamber, and the controller will... and Mapped to the current material state Figure 1 The number of grid rows is With the number of grid columns A two-dimensional grid, in which Indicates the number of grid rows. Indicates the number of grid columns and the grid position index. middle Represents the grid row index and its value range is to , Indicates the grid column index and its value range is to During mapping, the controller maps each grid location. The grid coverage area in the coordinates of the drying chamber is calculated, and the sampling points falling into this grid coverage area are aggregated. The aggregation method is to average the sampling point values ​​or to calculate the weighted average based on the distance from the sampling point to the grid center, so that each grid location obtains a unique temperature and moisture content value. When there are no sampling points in a grid coverage area, the controller uses the mapped values ​​of adjacent grids to fill the gap, so that the mapped grid results completely cover the area. Spatial range;

[0113] The controller will map the new temperature distribution. With the new moisture content distribution An updated material status diagram is generated by overlaying channels. Among them, updating the material status diagram The two channels correspond to the temperature distribution and moisture content distribution, respectively, and the grid positions correspond to the coordinates of the drying chamber. The controller will update the material state diagram. Write the current material status diagram to the buffer to replace the current material status diagram of the previous control cycle, and update the material status diagram. As inputs for the subsequent regeneration of the equipment action boundary map corresponding to the alternative control scheme, and for the input spatiotemporal neural field drying evolution prediction network to predict the future temperature distribution and future moisture content distribution, the execution of the control command and the prediction and screening of the next control cycle are connected in a closed loop.

[0114] In this embodiment, step S6 specifically includes:

[0115] In this embodiment, the controller receives the updated material state diagram after the control cycle ends and records the updated material state diagram as follows: Update the material status diagram Using a grid row number of The number of grid columns is The two-dimensional grid representation, where the grid position index is , The grid row index and the value range is to , The grid column index has a value range of 1. to The controller will update the material status diagram. A new temperature distribution is obtained by splitting the channels. With the new moisture content distribution ,in For grid position The measured temperature value, For grid position The measured moisture content value;

[0116] The controller adapts to the new moisture content distribution. Perform a point-by-point traversal, initialize the highest measured moisture content value to the moisture content value of the first grid cell, and then iterate through each grid cell. The current grid moisture content value is compared with the highest measured moisture content value. If the current grid moisture content value is greater, the highest measured moisture content value is updated. After the traversal is completed, the highest measured moisture content value is obtained. The controller is based on the new temperature distribution. With the new moisture content distribution Calculate the measured uniformity evaluation value based on the spatial dispersion at the grid location. The spatial dispersion is calculated by aggregating the grid average and point-by-point deviation: the controller traverses all grid locations to calculate the grid average temperature of the new temperature distribution. And calculate for each grid location. The average is then calculated to obtain the spatial dispersion of temperature as a sub-index of temperature uniformity. The controller traverses all grid locations to calculate the grid average moisture content of the new moisture content distribution. And calculate for each grid location. The average was then calculated to obtain the spatial dispersion of moisture content as a sub-index of moisture content uniformity. The controller uses preset weights and Temperature uniformity sub-index Sub-index of moisture content uniformity The weighted average is used to obtain the measured uniformity evaluation value. ,in and The controller is given a preset numerical weight that satisfies the condition that the sum of the weights is one.

[0117] The controller will measure the highest moisture content value. With the target moisture content threshold Perform constraint determination, the determination rule is as follows: Not greater than The controller will evaluate the measured uniformity value based on the target moisture content threshold. With uniformity threshold Perform constraint determination, the determination rule is as follows: Not greater than When the uniformity threshold is determined, the controller forms a drying completion determination when both the target moisture content threshold and the uniformity threshold are passed simultaneously, and outputs a drying completion flag to the host computer of the equipment. At the same time, it stops generating subsequent alternative control schemes and issuing control commands.

[0118] The controller will update the material status diagram before a drying completion determination is made. Write to the current material state diagram cache and update the material state diagram. As the current material state diagram participates in subsequent control decisions, the controller generates multiple candidate setpoints for each zone based on the adjustable ranges of infrared zone power, hot air zone temperature, and hot air zone airflow from the adjustable parameters of the equipment. These setpoints are then combined according to the zone number to form multiple sets of alternative control schemes. The controller records the alternative control scheme number as follows: And generate a corresponding equipment action boundary diagram for each alternative control scheme. The controller compares the current material status diagram with the equipment action boundary diagram. By inputting a spatiotemporal neural field drying evolution prediction network, the future temperature distribution corresponding to the alternative control scheme is obtained. With future moisture content distribution ;

[0119] The controller operates at the maximum permissible temperature threshold. With the target moisture content threshold The controller performs filtering under constraints for each future temperature distribution. Extracting the predicted highest temperature value by traversing point by point. ,when Greater than Eliminating alternative control schemes The controller for each future moisture content distribution Extracting the predicted maximum moisture content value by traversing point by point. ,when Greater than Eliminating alternative control schemes For alternative control schemes that pass the threshold constraint, the controller... Calculate the average temperature of the grid And calculate for all grid locations. The average value is used to obtain the temperature uniformity sub-index. The controller Calculate the average moisture content of the grid And calculate for all grid locations. The average value is used to obtain the moisture content uniformity sub-index. The controller uses preset weights and Will and The weighted summation is used to obtain the uniformity evaluation value of the alternative control scheme. The controller The selected alternative control schemes are sorted using the sorting key, and then... The smallest alternative control scheme is taken as the target alternative control scheme, and the target alternative control scheme is converted into an execution control instruction, so that the execution control instruction enters the parsing and distribution process of the next control cycle.

[0120] Example 1:

[0121] The equipment in this embodiment is a preheated fluidized bed. The bed is divided into six hot air chambers along the material flow direction. Each chamber is equipped with a heating unit and a damper / variable frequency fan, allowing for independent adjustment of the zone's air temperature and air volume. Six infrared heating zones (with adjustable power or duty cycle of the emitting array) are correspondingly located above the bed. A surface array temperature measurement device is arranged above the bed to obtain the temperature distribution, and a surface array moisture content measurement device (or multi-point moisture content + interpolation) is arranged at the outlet section to obtain the moisture content distribution. The controller establishes a unified cavity coordinate system within the bed and maps the observed data to a "current material state map" (temperature channel + moisture content channel) with the same grid size.

[0122] Within the control cycle T, the controller reads the current power and array status of each infrared zone, the current air temperature / airflow of each air chamber zone, and their adjustable upper and lower limits to form adjustable parameters for the equipment. The adjustable range of each zone is discretized into multiple setpoints and combined to generate multiple sets of alternative control schemes. For each set of alternative schemes, based on the coverage of the infrared zone and air chamber zone in the cavity coordinates, the zone setpoints are written into the corresponding grid to obtain the infrared power distribution, hot air temperature distribution, and hot air flow distribution, which are then superimposed by channel to form an "equipment action boundary map." When zone coverage overlaps, it is written according to a preset coverage priority rule to ensure consistent and reproducible mapping. Simultaneously, the infrared-hot air ratio is calculated and bound.

[0123] The controller inputs the current material state diagram and the equipment action boundary diagrams of each alternative scheme into the spatiotemporal neural field drying evolution prediction network: spatial coordinates are encoded point by point to generate positional features; infrared boundaries are extracted using small-neighborhood convolution to extract local action features; hot air boundaries are first normalized for air temperature and air volume and fused into hot air action input, and then subjected to large-neighborhood convolution to extract diffusion and transport features; after fusion, the future temperature distribution and future moisture content distribution corresponding to the alternative scheme are output through the field propagation layer. Subsequently, under the constraints of the maximum allowable temperature threshold and the target moisture content threshold, schemes that do not meet the requirements are eliminated, and the uniformity evaluation value (temperature dispersion + moisture content dispersion and adjacent difference terms weighted) is calculated for the remaining schemes. The scheme with the smallest uniformity is selected as the target scheme; if the uniformity is tied, the optimal scheme is selected based on the combined energy consumption.

[0124] Finally, the target scheme is converted into execution control commands, which are sent to the corresponding infrared partition and air chamber partition (air temperature / air volume) for execution and maintained until the end of the cycle. At the end of the cycle, the temperature / moisture content distribution is re-acquired to form an updated material state map, and the measured uniformity and measured maximum moisture content are calculated. When the target moisture content threshold and uniformity threshold are met at the same time, the preheating / drying is determined to be completed; otherwise, the next control cycle closed loop iteration begins.

[0125] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for controlling infrared hot air collaborative drying based on a uniformity-driven collaborative control architecture, characterized in that, include: S1. Obtain the temperature distribution and moisture content distribution of the material in the drying chamber to form the current material state diagram. Obtain the spatial layout of the infrared zone and the hot air zone and their respective adjustable ranges to form the adjustable parameters of the equipment. S2. Generate multiple sets of alternative control schemes for infrared zone power and hot air zone temperature and volume based on the adjustable parameters of the equipment, and map each alternative control scheme to the coordinates of the drying chamber according to the spatial layout to form the corresponding equipment action boundary diagram. S3. Input the current material state diagram and the equipment action boundary diagram corresponding to the alternative control scheme into the spatiotemporal neural field drying evolution prediction network to obtain the corresponding future temperature distribution and future moisture content distribution. The spatiotemporal neural field drying evolution prediction network includes a spatial coordinate input unit, an infrared action input unit, a hot air action input unit, and a field evolution output unit. The spatial coordinate input unit generates position features, the infrared action input unit and the hot air action input unit generate boundary features respectively, and the field evolution output unit merges the position features and boundary features to output the future temperature distribution and future moisture content distribution. S4. Calculate the uniformity evaluation value based on the future temperature distribution and future moisture content distribution corresponding to the alternative control schemes. Under the constraints of the maximum allowable temperature threshold and the target moisture content threshold, screen the alternative control schemes in conjunction with the adjustable range defined by the equipment action boundary diagram, and determine the control commands to be executed. The control commands to be executed include infrared zone power, hot air zone air temperature and air volume, and infrared hot air ratio. S5. Send the execution control command to the infrared zone heater and the hot air zone fan and execute it. At the end of the control cycle, obtain the new temperature distribution and the new moisture content distribution to form an updated material state diagram. S6. Calculate the measured uniformity evaluation value based on the updated material state diagram. When the new moisture content distribution meets the target moisture content threshold and the measured uniformity evaluation value is less than the uniformity threshold, a drying completion judgment is formed. When the condition is not met, a new alternative control scheme is generated based on the updated material state diagram and the control instruction to be executed is determined.

2. The infrared hot air collaborative drying control method based on a uniformity-driven collaborative control architecture according to claim 1, characterized in that, S1 specifically refers to: The partition temperature, partition power, and emission array status of each partition of the infrared drying equipment are obtained, and the partition temperature, partition power, and emission array status are respectively composed into partition temperature vector, partition power vector, and emission array status vector according to the number of partitions. The system acquires the conveyor belt speed and air volume, as well as the material inlet temperature, material inlet moisture content, material outlet temperature, and material outlet moisture content. The system then uses zone temperature vector, zone power vector, transmitter array state vector, air volume, conveyor belt speed, material inlet temperature, material inlet moisture content, material outlet temperature, and material outlet moisture content to form the current state of the equipment process. Based on the adjacent order and coverage length of the infrared drying equipment zones, and combined with the conveyor belt speed, calculate the material zone arrival order table, and align the material zone arrival order table with the current state of the equipment process. The actual control quantities are formed by the partition power vector, conveyor belt speed and air volume, and the current state of the equipment process and the actual control quantities are used as inputs to the dynamic evolution digital twin model.

3. The infrared hot air collaborative drying control method based on a uniformity-driven collaborative control architecture according to claim 1, characterized in that, S2 specifically refers to: Based on the spatial layout of the infrared zone and the hot air zone, determine the zone coverage range under the coordinates of the drying chamber, and convert the zone coverage range into a zone coverage grid identifier that uses the same number of grid rows and grid columns as the current material state diagram. Based on the adjustable range of infrared zone power, hot air zone temperature, and hot air zone air volume in the adjustable parameters of the equipment, multiple candidate setting values ​​are generated for each infrared zone and each hot air zone, and the candidate setting values ​​of each zone are combined in the order of adjacent zones to form multiple sets of alternative control schemes. For each alternative control scheme, the proportion of infrared hot air is determined. The proportion of infrared hot air is calculated by the combination relationship between the infrared zone power and the hot air zone temperature and air volume within the alternative control scheme, and the proportion of infrared hot air is bound to the alternative control scheme. The individual alternative control schemes are mapped to the drying chamber coordinates according to the partition coverage grid identifier. The partition settings that fall into the same grid position are written according to the priority of the partition coverage range, so as to obtain the distribution of infrared partition power on the drying chamber coordinates, the distribution of hot air partition temperature on the drying chamber coordinates, and the distribution of hot air partition air volume on the drying chamber coordinates. The distribution of infrared zone power in the drying chamber coordinates, the distribution of hot air zone temperature in the drying chamber coordinates, and the distribution of hot air zone air volume in the drying chamber coordinates are superimposed by channels to form a device action boundary map, which is then used as the spatial alignment input of the infrared action input unit and the hot air action input unit. Repeatedly perform partition coverage grid identification mapping and device action boundary map generation for all candidate control schemes to obtain a set of device action boundary maps corresponding to the candidate control schemes and output them to the spatiotemporal neural field drying evolution prediction network.

4. The infrared hot air collaborative drying control method based on a uniformity-driven collaborative control architecture according to claim 1, characterized in that, S3 specifically refers to: The coordinates of the drying chamber are used to generate a spatial coordinate map based on the number of grid rows and columns. Each grid position in the spatial coordinate map contains horizontal and vertical coordinates and corresponds to the grid position in the current material state map. The spatial coordinate map is input into the spatial coordinate input unit, and the spatial coordinate map is encoded to obtain the position feature map, so that the position feature map is aligned with the position of each grid in the drying cavity; The distribution of infrared partition power in the drying cavity coordinates in the equipment action boundary map is input into the infrared action input unit. Small neighborhood convolutional layers are stacked to generate an infrared boundary feature map, so that the infrared boundary feature map can characterize the spatial action of infrared radiation within the local coverage area. The distribution of hot air zone temperature in the drying chamber coordinates and the distribution of hot air zone air volume in the drying chamber coordinates in the equipment action boundary map are input into the hot air action input unit. A hot air boundary feature map is generated by stacking large neighborhood convolutional layers, and the channel dimension of the hot air boundary feature map is aligned with the channel dimension of the infrared boundary feature map. The current material state map is input into the initial state encoding layer, and the current material state map is convolutionally encoded to generate an initial state feature map, so that the initial state feature map represents the spatial initial state of the current temperature distribution and the current moisture content distribution. The location feature map, infrared boundary feature map, hot air boundary feature map and initial state feature map are input into the field evolution output unit and spliced ​​through channels to form a fused feature map. The fused feature map is then input into the field propagation layer for spatial coupling propagation and outputs the propagated feature map. The propagated feature map is input into the output layer, and pointwise convolution mapping is used to obtain the predicted output map. The predicted output map contains two channels: future temperature distribution and future moisture content distribution. The aforementioned input and inference are performed on the equipment action boundary map corresponding to each alternative control scheme to obtain the future temperature distribution and future moisture content distribution corresponding to that alternative control scheme.

5. The infrared hot air collaborative drying control method based on a uniformity-driven collaborative control architecture according to claim 4, characterized in that, When inputting the distribution of hot air zone temperature and hot air zone air volume in the drying chamber coordinates from the equipment action boundary diagram into the hot air action input unit, a construction function is used to process the hot air zone temperature and hot air zone air volume into the input of the hot air action input unit. Specifically, the construction function is as follows: ； in, The hot air fusion input diagram is shown in the first... Line 1 The numerical value of the column grid position. Represents the grid row index and its value range is to , Indicates the grid column index and its value range is to , Represents the boundary diagram of equipment operation The distribution of hot air zone temperature on the drying chamber coordinates at the grid position The wind temperature value, Represents the boundary diagram of equipment operation The distribution of hot air zoning air volume in the drying chamber coordinates at the grid position The air volume value, This represents the global minimum air temperature within the adjustable range of the hot air zone defined by the adjustable parameters of the device. This represents the global maximum air temperature within the adjustable range of the hot air zone defined by the adjustable parameters of the device. This represents the global minimum airflow rate within the adjustable range of the hot air zone as defined by the device's adjustable parameters. This represents the global maximum airflow within the adjustable range of the hot air zone defined by the device's adjustable parameters. This represents the preset hot air volume spatial variation weighting coefficient, used to adjust the contribution of air volume variations at adjacent grid positions to the hot air fusion input map. , , , When the location is outside the grid boundary, the corresponding outside grid boundary location will be... Replace with .

6. The infrared hot air collaborative drying control method based on a uniformity-driven collaborative control architecture according to claim 1, characterized in that, S4 specifically refers to: Using the alternative control scheme as an index, read the future temperature distribution and future moisture content distribution corresponding to the alternative control scheme respectively, and read the equipment action boundary diagram corresponding to the alternative control scheme as the adjustable range constraint input; The maximum allowable temperature threshold constraint is determined for the future temperature distribution. The predicted maximum temperature value is extracted by traversing the grid positions of the future temperature distribution point by point. The candidate control scheme with the predicted maximum temperature value greater than the maximum allowable temperature threshold is eliminated. The target moisture content threshold constraint is used to determine the future moisture content distribution. The predicted maximum moisture content value is extracted by traversing the grid positions of the future moisture content distribution point by point. The candidate control schemes with the predicted maximum moisture content value greater than the target moisture content threshold are eliminated. For the alternative control schemes that pass the threshold constraint, the adjustable range consistency judgment is performed. The distribution of infrared zone power in the drying chamber coordinates, the distribution of hot air zone temperature in the drying chamber coordinates, and the distribution of hot air zone air volume in the drying chamber coordinates in the equipment action boundary diagram are compared point by point with the upper and lower limits of the adjustable parameters of the equipment. Alternative control schemes with grid positions that exceed the boundaries are eliminated. For the alternative control schemes that pass the consistency judgment, calculate the uniformity evaluation value, use the temperature spatial dispersion of the future temperature distribution to obtain the temperature uniformity sub-index, use the moisture content spatial dispersion of the future moisture content distribution to obtain the moisture content uniformity sub-index, and combine the temperature uniformity sub-index and the moisture content uniformity sub-index according to the preset weight to form the uniformity evaluation value. The uniformity evaluation value is used as the sorting key to sort the candidate control schemes that have passed the screening. The candidate control scheme with the smallest uniformity evaluation value is selected as the target candidate control scheme. When the uniformity evaluation values ​​are the same, the target candidate control scheme with the smaller correlation between the combined energy consumption of infrared zone power and hot air zone temperature and air volume is selected. The target alternative control scheme is converted into an execution control command. The execution control command includes the infrared zone power, hot air zone temperature and air volume that are consistent with the target alternative control scheme. After determining the proportion of infrared hot air based on the combination relationship between infrared zone power and hot air zone temperature and air volume in the target alternative control scheme, the execution control command is output.

7. The infrared hot air collaborative drying control method based on a uniformity-driven collaborative control architecture according to claim 6, characterized in that, When calculating the uniformity evaluation value for alternative control schemes that pass the threshold constraint, an optimization function is used to synthesize the temperature uniformity sub-index and the moisture content uniformity sub-index according to a preset weight to obtain the uniformity evaluation value. Specifically, the optimization function is as follows: ； in, Indicates alternative control schemes The uniformity evaluation value, This represents the temperature uniformity sub-index and corresponds to the value in the above formula. Weighted bracketed items, This represents the sub-index of moisture content uniformity and corresponds to the expression in the above formula. Weighted bracketed items, This indicates the preset weight of the temperature uniformity sub-index. This indicates the preset weight of the moisture content uniformity sub-index. Indicates the number of grid rows. Indicates the number of grid columns. Indicates the grid row index. Indicates the grid column index. and This represents the traversal index used to calculate the grid average. This indicates the future temperature distribution at the grid location. Temperature value, This indicates the future moisture content distribution at the grid locations. The moisture content value, Indicates the maximum permissible temperature threshold. Indicates the target moisture content threshold. This represents the preset weighting coefficients for adjacent temperature difference terms. This represents the preset weighting coefficient for adjacent difference terms in moisture content, when the index... or When the location is outside the grid boundary, the corresponding outside grid boundary location will be... or Replace with or .

8. The infrared hot air collaborative drying control method based on a uniformity-driven collaborative control architecture according to claim 1, characterized in that, S5 specifically refers to: The execution control command is parsed into infrared zone power, hot air zone temperature and air volume, and infrared hot air ratio. The infrared zone power is sent to the corresponding infrared zone heater, and the hot air zone temperature and air volume are sent to the corresponding hot air zone fan. Within the control cycle, the infrared zone heater and hot air zone fan are driven to output according to the executed control command, so that the infrared effect and hot air effect in the drying chamber are consistent with the executed control command. At the end of the control cycle, acquire the new temperature distribution and the new moisture content distribution, and map the new temperature distribution and the new moisture content distribution to the grid row number and grid column number consistent with the current material state diagram according to the drying chamber coordinates; The mapped new temperature distribution and the new moisture content distribution are superimposed on each channel to form an updated material state diagram. This updated material state diagram is then used as input for the subsequent regeneration of alternative control schemes and the prediction of future temperature and moisture content distributions.

9. The infrared hot air collaborative drying control method based on a uniformity-driven collaborative control architecture according to claim 1, characterized in that, Step S6 is as follows: The updated material state diagram is split into channels to obtain new temperature distribution and new moisture content distribution. The highest measured moisture content value of the new moisture content distribution is extracted by point-by-point traversal. The measured uniformity evaluation value is calculated based on the spatial dispersion of the new temperature distribution and the new moisture content distribution at the grid position. The measured maximum moisture content value is constrained by the target moisture content threshold, and the measured uniformity evaluation value is constrained by the uniformity threshold. When both the target moisture content threshold and the uniformity threshold are satisfied, the drying completion determination is formed. Before a drying completion determination is formed, the updated material state diagram is used as the current material state diagram to generate alternative control schemes and generate equipment action boundary diagrams corresponding to the alternative control schemes. The current material state diagram and equipment action boundary diagram are input into the spatiotemporal neural field drying evolution prediction network to obtain the future temperature distribution and future moisture content distribution. Under the constraints of the maximum allowable temperature threshold and the target moisture content threshold, the control command to be executed is selected based on the uniformity evaluation value.

Images