Cooperative actuator linkage control method and system based on spatial humidity distribution mapping

Abstract

The application discloses a cooperative actuator linkage control method and system based on spatial humidity distribution mapping, relates to the technical field of actuator linkage control, and comprises the following steps: initializing and establishing a cup body fixed coordinate system and a nozzle array coordinate system, and calibrating a conversion matrix of the two coordinate systems; the rotating bearing table and the nozzle run according to set parameters during drying, the humidity of each region of the cup body is collected in real time and is mapped to the corresponding coordinate system; when the humidity of a certain region exceeds the standard, the controller solves and realizes accurate blowing; meanwhile, a digital twin model is trained based on historical data, the humidity change is predicted in real time, and abnormal regions are fed forward and compensated. The system monitors the execution state of the nozzle in real time, starts fault tolerance reconstruction when a fault occurs, adjusts the parameters of adjacent nozzles, and alarms. The application realizes dynamic tracking type accurate drying, introduces digital twin feedforward and fault tolerance reconstruction, reduces the required energy consumption and drying time, greatly improves the drying coverage rate and the standard deviation of the residual humidity of the cup wall, and has the abilities of process tracing and fault tolerance.

Description

Technical Field

[0001] This invention relates to the field of actuator linkage control technology, specifically to a collaborative actuator linkage control method and system based on spatial humidity distribution mapping. Background Technology

[0002] During experiments such as water bath digestion and heat incubation, water bath liquid will inevitably remain on the outer wall of the sample cup. If this residual moisture is not removed in time, it will cause a series of technical problems: contaminating the experimental environment, causing cross-contamination, interfering with precise detection results such as spectral analysis, and reducing experimental efficiency.

[0003] In existing technologies, for example, patent CN111944689A discloses a cell water bath oscillation culture device, which uses a single airflow in a fixed direction for purging. This results in obvious purging dead zones, failing to cover key areas such as the bottom curved surface of the cup wall and the inner side of the cup mouth. The measured residual moisture rate is as high as 15%-25%. The airflow parameters are completely fixed and cannot be dynamically adjusted according to the actual degree of drying, lacking real-time monitoring and closed-loop control. Patent CN107490584A discloses a sample cup clamping device for a spectrometer, but it only involves clamping and fixing the sample cup, completely omitting the drying process. This results in extremely low automation and poor adaptability. Conventional manual wiping is inefficient and easily introduces secondary contamination; natural air drying is too time-consuming; hot air drying carries the risk of high-temperature deformation (PP material deforms by more than 5% after 5 minutes at temperatures above 80°C) and produces uneven drying. Patent CN201637624U discloses a dryer for drying sample cups by spraying airflow, including a heating chamber, a control panel, an adjustable air nozzle, and a temperature controller. However, the "adjustable direction" of this patent is essentially manual adjustment; that is, the operator manually adjusts the air nozzle to a fixed direction before the experiment begins, and the device then continuously blows air in that direction. This static preset method cannot be adjusted according to real-time changes in operating conditions during the drying process, making it difficult to solve the problem of uneven drying in certain areas.

[0004] In addition, although automated drying equipment with multiple nozzles and rotating tables has appeared on the market, each actuator (rotating table, nozzle angle, flow valve) mostly adopts an independent control strategy. That is, if the sensor detects that the humidity in a certain area is too high, the corresponding nozzle is controlled to increase the blowing. This modular splicing method is essentially still a rough "point-to-point" mapping, ignoring the dynamic coordinate transformation caused by the rotation of the cup and the superposition effect of the airflow field of multiple nozzles. As a result, there is still room for improvement in drying efficiency and uniformity, and it lacks fault tolerance for actuator failures.

[0005] In summary, achieving seamless drying, adaptive adjustment, process traceability, and compatibility with multiple cup types has become a key technological requirement for overcoming the bottlenecks in laboratory automation. Summary of the Invention

[0006] The purpose of this invention is to overcome the shortcomings of the prior art and provide a collaborative actuator linkage control method and system based on spatial humidity distribution mapping.

[0007] The objective of this invention is achieved through the following technical solution: In a first aspect, this application discloses a collaborative actuator linkage control method based on spatial humidity distribution mapping, comprising the following steps: S1. The system initializes and establishes a virtual coordinate system, which includes a cup-body fixed coordinate system (C-System) and a nozzle array coordinate system (N-System). The cup-body fixed coordinate system (C-System) has its origin at the center of the rotating support platform and rotates with the platform. The nozzle array coordinate system (N-System) has its origin at the multi-angle jet assembly mounting base. The transformation matrix between the cup-body fixed coordinate system (C-System) and the nozzle array coordinate system (N-System) is determined through a calibration program. This transformation matrix is ​​a function of the real-time angle of the rotating support platform. Simultaneously, the height position of the multi-angle jet assembly is adjusted using a height adjustment mechanism. S2. Start the drying operation. The rotating support platform runs at an initial speed N0. The multi-angle jet assembly delivers pulsed air at an initial flow rate Q0. The humidity sensor array collects humidity data of each area of ​​the cup in real time and maps it to the cup's fixed coordinate system C-System. S3. If, in the fixed coordinate system C-System of the cup, the humidity sensor array detects that the humidity in the spatial region P exceeds a preset humidity threshold, then a collaborative calculation is performed in the controller to determine the optimal spraying time. And through inverse kinematics calculation, a coordinated command is generated to determine the combination of nozzles involved in the purging and their deflection angle, so as to accurately purge the spatial region P; S4. Train a digital twin model of the drying process using several sets of historical drying data to predict the rate of change of humidity over time in each region of the cup wall under a combination of coordinated parameters. The combination of coordinated parameters includes the set values ​​of nozzle angle, rotation speed, and flow rate at the current time and in the future time. Solidify the trained model into the central controller and call it in real time during the drying process. If the rate of change of humidity over time in a certain region is lower than the preset threshold, adjust the corresponding nozzle angle or the rotation speed of the rotating support table in advance to perform feedforward compensation. S5. In real time, compare the command angle of each nozzle with the actual feedback angle, air pressure and flow sensor data. If the data is abnormal and causes the nozzle actuator to malfunction, immediately start the fault-tolerant reconstruction mode, including recalculating, generating new collaborative commands, reducing the rotation speed of the rotating platform, and adjusting the angle and flow of a preset number of adjacent nozzles to compensate for the failure area, while issuing an alarm prompt. S6. A segmented adaptive control strategy is adopted until the humidity of all areas does not exceed the preset humidity threshold within a preset time and the humidity standard deviation is less than the preset relative humidity; if a timeout occurs, timeout protection and fault alarm are activated.

[0008] Based on the first aspect, the calibration procedure described in step S1 includes scanning a laser target or using a standard cup for multi-point scanning.

[0009] Based on the first aspect, the collaborative solution performed in the controller in step S3 includes the following sub-steps: S31, Set look-ahead time window Discretize the window into M equally spaced time intervals. , Where t represents time, and the discrete time indices are k=1,...,M. Represents discrete time intervals, based on the current rotational speed. And the expected speed adjustment, predict the angle of the bearing platform. , Where θ(t) represents the angle of the rotating platform at time t, and n represents the rotational speed of the rotating platform; calculate the time intervals. The mapped coordinates P' of the lower spatial region P in the nozzle array coordinate system N-System. ,in The transformation matrix represents the coordinate system C-System of the cup body and the coordinate system N-System of the nozzle array. Then, the distance between the mapped coordinate P' and the center line of each nozzle is calculated. The earliest time when the distance is less than the preset threshold is selected as the optimal spraying time t+Δt. If there is no time that meets the condition, the rotation speed n of the rotating platform is adjusted so that P enters the effective range faster and the prediction is repeated. S32. Based on the mapped coordinates P', determine the nozzle combination involved in the purging and its deflection angle through geometric inverse kinematics calculation; specifically including: To make the nozzle i Pointing to the mapped coordinates P', using the formula Calculate the unit direction vector ,in Indicates nozzle i The fixed installation position in the nozzle array coordinate system N-System, and then through the formula and Reverse solution nozzle i Horizontal swing angle and pitch angle ,in Represents the unit direction vector Coordinates on the y-axis Represents the unit direction vector Coordinates on the x-axis Represents the unit direction vector The coordinates on the z-axis; select the preset number of nozzles closest to the mapped coordinate P', calculate their required horizontal swing angle and pitch angle respectively, and adjust the rotation speed of the rotating platform to... If the velocity of the spatial region P within the prediction window exceeds the preset velocity threshold, then fine-tune. The speed at which the spatial region P sweeps across the nozzle focusing area does not exceed a preset speed threshold, wherein the nozzle focusing area includes the nozzle. i And the selected preset number of nozzles.

[0010] Based on the first aspect, the step S4 of training a digital twin model of the drying process using several sets of historical drying data includes the following sub-steps: S41. Data acquisition: Record the input parameter sequence and output parameters of the complete drying process for different cup shapes and different initial humidity distributions; the input parameter sequence includes the nozzle angle, the rotation speed of the rotating support table, and the airflow rate; the output parameters include the humidity change curve of each sensor area over time. S42. Data preprocessing: normalize, denoise, and align the collected input parameter sequence and output parameters. S43. Feature engineering: Extracting key features from the preprocessed input parameter sequence and output parameters, including the current humidity distribution, humidity change rate and actuator status; S44. Model selection and training: A time-series prediction model is constructed using a Long Short-Term Memory (LSTM) network or a gated recurrent unit (GRU). The Adam optimizer is used to train the model with the goal of minimizing the prediction error. The ratio of the training set to the validation set is 8:2. The current state vector and the future control parameter sequence are used as its inputs, and the predicted humidity values ​​of each sensor area in the future are used as its outputs. The future control parameter sequence includes the nozzle angle, rotation speed, and flow rate at each future time. S45. Model Deployment: The trained model is embedded into the central controller and called in real time during the drying process to predict the rate of change of humidity over time in various regions of the cup wall under the combined parameters.

[0011] Based on the first aspect, the fault in step S5 caused by abnormal data to result in nozzle actuator failure includes nozzle jamming or angle deviation exceeding the preset threshold of nozzle angle.

[0012] Based on the first aspect, the segmented adaptive control strategy described in step S6 specifically includes the following sub-steps: S61. Rapid drying stage: Run with default coordination parameters and calculate the first humidity change rate in sensor region j. , ,in Indicates the sampling period. This represents the humidity value of sensor region j at time t. Indicates that sensor region j is in Humidity value at time, the first humidity change rate The temperature will fluctuate due to instantaneous changes in the sensor; therefore, the first humidity change rate R of the current and the previous L-1 sampling points is taken. j The average value is used as the output to obtain the second humidity change rate. , Where L represents the window length, Indicates the offset index of the sampling point within the window; if the second humidity change rate Three consecutive times the humidity level is below the preset threshold. And the current humidity If this is the case, it indicates the presence of stubborn water stains, and the area enters the localized intensive drying phase; among which... Indicates the hysteresis interval. Indicates the preset humidity threshold; S62, Localized Intensive Drying Stage: Identify the center of the stubborn water stain area. Based on the determination of the nozzle to be deflected and its target angle in step S3, generate coordinated commands, including deflecting the nozzle to the target angle and reducing the rotation speed of the rotating platform. At this time, the angle of the bearing platform The calculation is performed using uniform angular velocity motion, i.e. Increase the airflow of the multi-angle jet assembly to 1.5Q0 and continuously track the humidity of stubborn water stains until it subsides; S63, Slow drying stage, when the average humidity of all sensor areas... At that time, reduce the rotation speed of the rotating platform to... This reduces the airflow rate of the multi-angle jet assembly to The PID controller is activated to fine-tune the flow rate, while the residual humidity uniformity is predicted through a digital twin model of the drying process. When it is predicted that the humidity of a certain area will deviate from that of other areas by more than 0.5% relative humidity, or when the current sensor detects that the humidity of a certain area deviates from the average value by more than 1% relative humidity, the small angle adjustment amount is automatically calculated and the corresponding nozzle is adjusted to provide additional airflow compensation to that area to eliminate the small deviation.

[0013] Based on the first aspect, in step S61, the center position of the stubborn water stain area and the corresponding nozzle angle are recalculated in each control cycle to achieve dynamic focusing. If the center position of the stubborn water stain area moves, the nozzle angle will move accordingly in real time.

[0014] Secondly, this application discloses a coordinated actuator linkage control system based on spatial humidity distribution mapping, which utilizes the aforementioned coordinated actuator linkage control method based on spatial humidity distribution mapping, including: The initialization module is used to complete system initialization and establish the cup body fixed coordinate system C-System and the nozzle array coordinate system N-System; The humidity sensing module drives the humidity sensor array to collect humidity data of the cup in real time; based on the coordinate system transformation matrix, it maps the raw humidity data to the cup's fixed coordinate system C-System; and preprocesses the humidity data. The cooperative motion control module receives coordinate information of areas with excessive humidity, performs cooperative calculations to determine the optimal spraying time, and generates actuator cooperative commands through inverse kinematics calculations. The feedforward compensation module is used to train a digital twin model of the drying process and predict the rate of change of humidity over time in each region of the cup wall under the combined parameters; when the rate of change of humidity in a region is lower than the threshold, feedforward compensation is automatically performed. The monitoring module is used to collect and compare the actual feedback angle, air pressure sensor, and flow sensor data of the nozzle actuator in real time, providing a basis for fault diagnosis. The reconfiguration module is used to automatically identify actuator faults; when a fault occurs, the fault-tolerant reconfiguration mode is immediately activated: the kinematics are recalculated and new cooperative instructions are generated. The segmented adaptive control module is used to execute the segmented adaptive control strategy until the humidity and humidity standard deviation of all areas meet the drying completion conditions. The alarm module is used to receive fault signals and perform timeout protection and fault alarm.

[0015] The beneficial effects of this invention are: 1) This application uses spatial mapping to directly use the "three-dimensional spatial distribution information" provided by the humidity sensor array to drive the deep collaboration of multiple actuators (nozzle angle, rotary table speed, flow rate), thereby realizing dynamic tracking precision drying. By introducing digital twin feedforward and fault-tolerant reconstruction into the drying equipment, this application reduces the required energy consumption and drying time compared with the prior art, significantly improves the drying coverage and the standard deviation of residual humidity on the cup wall, and has the ability to trace the process and tolerate faults. Attached Figure Description

[0016] Figure 1 This is a schematic diagram illustrating the steps of the collaborative actuator linkage control method based on spatial humidity distribution mapping according to an embodiment of the present invention. Detailed Implementation

[0017] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] This application discloses a collaborative actuator linkage control method based on spatial humidity distribution mapping. By establishing a precise mapping between the fixed coordinate system of the cup and the coordinate system of the nozzle array, the airflow can dynamically track water stain areas, achieving true full-space coverage and eliminating drying dead zones. The rotating platform speed, multiple nozzle angles, and airflow rate are used as mutually coupled control variables. A collaborative instruction set is generated based on real-time humidity spatial distribution, achieving a high degree of temporal and spatial synchronization between "perception-decision-execution." A digital twin model is constructed using historical data to predict the drying process and adjust parameters in advance, eliminating control lag and improving drying consistency and energy efficiency. Multi-sensor feedback is used to monitor the actuator status in real time, automatically reconstructing the control strategy in case of failure, and using adjacent nozzles to compensate for the failed area, ensuring equipment robustness. A modular cup adapter component is provided to support interface with a Manufacturing Execution System (MES), seamlessly integrating into laboratory automated production lines. A schematic diagram of the steps of the method is shown below. Figure 1 As shown, the specific steps include: S1. The system initializes and establishes a virtual coordinate system, which includes a cup-body fixed coordinate system (C-System) and a nozzle array coordinate system (N-System) (calibrated automatically when a specific cup type is used for the first time). The cup-body fixed coordinate system (C-System) has its origin at the center of the rotating platform and rotates with the platform. The nozzle array coordinate system (N-System) has its origin at the multi-angle jet assembly mounting base. The transformation matrix between the cup-body fixed coordinate system (C-System) and the nozzle array coordinate system (N-System) is determined through a calibration program. This transformation matrix is ​​a function of the real-time angle of the rotating platform. Simultaneously, the height position of the multi-angle jet assembly is adjusted using a height adjustment mechanism. S2. Start the drying operation. The rotating platform operates at an initial speed N0 (representing the initial speed of the rotating platform in the rapid drying stage, in rpm) (used to drive the sample cup to rotate at a constant speed at the start of drying). The multi-angle jet assembly pulses air at an initial flow rate Q0 (representing the initial airflow flow rate of the jet assembly in the rapid drying stage, in L / min) (used to spray drying gas onto the sample cup wall at the start of drying). The humidity sensor array collects humidity data of each area of ​​the cup in real time and maps it to the cup's fixed coordinate system C-System. S3. If, in the fixed coordinate system C-System of the cup body, the humidity sensor array detects that the humidity in a spatial region P (e.g., a point on the curved surface at the bottom of the cup wall) exceeds a preset humidity threshold, then a collaborative calculation is performed in the controller to determine the optimal spraying time. And through inverse kinematics calculation, a coordinated command is generated to determine the combination of nozzles involved in the purging and their deflection angle, so as to accurately purge the spatial region P; S4. Train a digital twin model of the drying process using several sets (≥1000 sets) of historical drying data to predict the rate of change of humidity over time in each region of the cup wall under a combination of coordinated parameters. The combination of coordinated parameters includes the set values ​​of nozzle angle, rotation speed, and flow rate at the current time and in future time. Solidify the trained model into the central controller and call it in real time during the drying process. If the rate of change of humidity over time in a certain region is lower than the preset threshold, adjust the corresponding nozzle angle or the rotation speed of the rotating support table in advance to perform feedforward compensation. S5. Real-time comparison of the command angle and actual feedback angle of each nozzle, as well as air pressure and flow sensor data. If abnormal data causes the nozzle actuator to malfunction, the fault-tolerant reconfiguration mode is immediately activated, including recalculation, generating new collaborative commands, reducing the rotation speed of the rotating platform, and adjusting the angle and flow of a preset number of adjacent nozzles to compensate for the failure area. At the same time, an alarm is issued. This mechanism greatly improves the robustness of the system. S6. A segmented adaptive control strategy is adopted until the humidity of all areas does not exceed the preset humidity threshold within a preset time (e.g., 2 minutes) and the humidity standard deviation is less than the preset relative humidity (e.g., 2% relative humidity). If a timeout occurs, timeout protection and fault alarm are activated.

[0019] For example, the rotating support stage is used to place and securely fix the sample cups to be dried. It includes a main rectangular platform (diameter 150-300mm), with 4-6 pneumatic / electric adjustable grippers evenly distributed along the edge, and a food-grade silicone anti-slip pad (2mm thick) attached to the inside. It is suitable for sample cups with a diameter of 20-50mm. The clamping force is steplessly adjustable from 0.5-5N (accuracy 0.1N). It is driven by a high-precision servo motor (50W), with stepless speed regulation from 0-400rpm and an accuracy of ±1rpm. It has a built-in absolute encoder for real-time angle feedback.

[0020] For example, the frame is made of high-quality carbon structural steel as the main material, and the surface is treated with electrostatic powder coating; it is equipped with four sets of height-adjustable vibration-damping feet, with an adjustment range of ±50mm; the jet assembly is fixedly installed on one side of the frame, maintaining a working distance of 100-150mm from the rotating support platform; it includes 2-4 coaxially arranged annular jet pipes (adjacent spacing 50mm), with 8-16 micro jet nozzles (304 stainless steel, Ra≤0.8μm) evenly distributed in the inner ring of each ring. Each nozzle achieves automatic angle adjustment through an independent micro stepper motor (step angle 1.8°) and worm gear transmission mechanism, with an adjustment range of 0-60° and a control accuracy of ±0.5°. Each nozzle is equipped with an angle encoder for closed-loop control.

[0021] For example, the height adjustment mechanism is rigidly connected to the annular jet pipe via an aluminum alloy bracket, employing a direct-drive lifting system consisting of a precision ball screw (5mm lead) and a servo motor. The vertical travel is 0-200mm, with a positioning accuracy of ±0.1mm. The humidity sensor array uses 4-8 infrared non-contact humidity sensors, distributed at multiple points along the axial (top, middle, bottom) and radial directions of the sample cup, forming a three-dimensional monitoring network. The resolution is 0.1%RH, the sampling rate is 100Hz, the response time is <50ms, and the accuracy is ±2%RH. Each sensor is calibrated during installation, and its spatial coordinates in the fixed coordinate system of the cup are known.

[0022] For example, the calibration procedure described in step S1 includes laser target or multi-point scanning using a standard cup.

[0023] For example, the collaborative solution performed in the controller in step S3 includes the following sub-steps: S31, Set look-ahead time window (For example, 1 second), discretize the window into M equally spaced time intervals. , Where t represents time, and the discrete time indices are k=1,...,M. Represents the discrete time interval (the time difference between two adjacent discrete moments, in seconds), based on the current rotational speed. And the expected speed adjustment, predict the angle of the bearing platform. , Where θ(t) represents the angle of the rotating platform at time t, and n represents the rotational speed of the rotating platform; calculate the time intervals. The mapped coordinates P' of the lower spatial region P in the nozzle array coordinate system N-System. ,in The transformation matrix represents the coordinate system C-System of the cup body and the coordinate system N-System of the nozzle array; then, the distance between the mapped coordinate P' and the center line pointed to by each nozzle is calculated, and the earliest time when the distance is less than a preset threshold (e.g., 5mm) is selected as the optimal spraying time. If there is no time that meets the conditions, adjust the rotation speed n of the rotating platform to make P enter the effective range faster and re-predict (by changing the rotation speed of the rotating platform, the stubborn water stain area enters the effective purging range of the nozzle faster). S32. Based on the mapped coordinates P', determine the nozzle combination involved in the purging and its deflection angle through geometric inverse kinematics calculation; specifically including: To make the nozzle i Pointing to the mapped coordinates P', using the formula Calculate the unit direction vector ,in Indicates nozzle i The fixed installation position in the nozzle array coordinate system N-System, and then through the formula and Reverse solution nozzle i Horizontal swing angle and pitch angle ,in Represents the unit direction vector Coordinates on the y-axis Represents the unit direction vector Coordinates on the x-axis Represents the unit direction vector The coordinates on the z-axis; select the nozzles closest to the mapped coordinate P' (e.g., 2-4, sorted by Euclidean distance), calculate their required horizontal swing angle and pitch angle respectively, and adjust the rotation speed of the rotating platform to... If the velocity of the spatial region P within the prediction window exceeds the preset velocity threshold, then fine-tune. The speed at which the spatial region P sweeps across the nozzle focusing area does not exceed a preset speed threshold, wherein the nozzle focusing area includes the nozzle. i And the selected preset number of nozzles.

[0024] For example, the step S4 of training a digital twin model of the drying process using several sets of historical drying data includes the following sub-steps: S41. Data acquisition: Record the input parameter sequence and output parameters of the complete drying process for different cup shapes and different initial humidity distributions; the input parameter sequence includes the nozzle angle, the rotation speed of the rotating support table, and the airflow rate; the output parameters include the humidity change curve of each sensor area over time. S42. Data preprocessing: normalize, denoise, and align the collected input parameter sequence and output parameters. S43. Feature engineering: Extracting key features from the preprocessed input parameter sequence and output parameters, including the current humidity distribution, humidity change rate and actuator status; S44. Model Selection and Training: A time-series prediction model is constructed using a Long Short-Term Memory (LSTM) network or a gated recurrent unit (GRU). The Adam optimizer is used to train the model with the objective of minimizing the prediction error (e.g., root mean square error). The ratio of the training set to the validation set is 8:2. The current state vector and the sequence of future control parameters (e.g., nozzle angles, rotation speeds, and flow rates within the next 2 seconds) are used as inputs, and the predicted humidity values ​​for each sensor region in the future are used as outputs. The sequence of future control parameters includes the nozzle angles, rotation speeds, and flow rates at future times. S45. Model Deployment: The trained model is embedded into the central controller and called in real time during the drying process to predict the rate of change of humidity over time in various regions of the cup wall under the combined parameters.

[0025] For example, the "if abnormal data causes the nozzle actuator to malfunction" mentioned in step S5 includes nozzle jamming and angle deviation exceeding the preset threshold of nozzle angle.

[0026] For example, the segmented adaptive control strategy described in step S6 specifically includes the following sub-steps: S61. Rapid drying stage: Run with default coordination parameters and calculate the humidity change rate of sensor area j. , ,in Indicates the sampling period (e.g., 0.1s). This represents the humidity value of sensor region j at time t. Indicates that sensor region j is in Humidity value at time, the first humidity change rate The temperature will fluctuate due to instantaneous changes in the sensor; therefore, the first humidity change rate of the current and the previous L-1 sampling points is taken. The average value is used as the output to obtain the second humidity change rate. , Where L represents the window length, Indicates the offset index of the sampling point within the window (used to accumulate the original rate of change of each sampling point within the window); if the second humidity rate of change... Three consecutive times the humidity level is below the preset threshold. (e.g., 0.5% relative humidity / s) and current humidity If this is the case, it indicates the presence of stubborn water stains, and the area enters the localized intensive drying phase; among which... Indicates the hysteresis range (e.g., 2% relative humidity). Indicates the preset humidity threshold; S62, Localized Intensive Drying Stage: Identify the center of the stubborn water stain area. Based on the determination of the nozzle to be deflected and its target angle in step S3, generate coordinated commands, including deflecting the nozzle to the target angle and reducing the rotation speed of the rotating platform. At this time, the angle of the bearing platform The calculation is performed using uniform angular velocity motion, i.e. Increase the airflow of the multi-angle jet assembly to 1.5Q0 and continuously track the humidity of stubborn water stains until it subsides; S63, Slow drying stage, when the average humidity of all sensor areas... At that time, reduce the rotation speed of the rotating platform to... This reduces the airflow rate of the multi-angle jet assembly to The PID controller is activated to fine-tune the flow rate, while the residual humidity uniformity is predicted through a digital twin model of the drying process. When it is predicted that the humidity of a certain area will deviate from that of other areas by more than 0.5% relative humidity, or when the current sensor detects that the humidity of a certain area deviates from the average value by more than 1% relative humidity, the small angle adjustment amount (within ±2°) is automatically calculated and the corresponding nozzle is adjusted to provide additional airflow compensation to that area to eliminate the small deviation.

[0027] For example, in step S61, the center position of the stubborn water stain area and the corresponding nozzle angle are recalculated in each control cycle (e.g., 0.1s) to achieve dynamic focusing. If the center position of the stubborn water stain area moves, the nozzle angle moves accordingly in real time.

[0028] For example, a standard glass sample cup (30mm in diameter, 100mm in height) with preset parameters: N0=200rpm, Q0=15L / min. =2% relative humidity, maximum allowable drying time T max =30s. The system first calls the calibration parameters (transformation matrix T) for this cup type. At the 5th second, the rate of change of humidity in the bottom area over time is lower than the preset threshold, and the sensor locates the coordinates of this area in the C-System (0, -15, -40) (unit: mm). The controller combines this with the current rotation angle of the support platform θ(t) = 120°, and the look-ahead window... =1s, after discretization, the prediction is in At 5.3 seconds, the area rotates to the optimal windward side, with mapped coordinates P' = (-20, 30, -40). Calculations show that nozzles 3 and 4 are closest to P'. It is determined that nozzle 3 needs to deflect (horizontal swing angle 12°, pitch angle 5°), and nozzle 4 needs to deflect (horizontal swing angle 18°, pitch angle 3°). The rotational speed of the rotating platform is reduced to 100 rpm, and the flow rate of the multi-angle jet assembly is increased to 22 L / min. Focused purging begins at 5.3 seconds and continues until the humidity drops to 2.1% at 8 seconds. At 13 seconds, a slow drying phase begins, with PID control adjusting the flow rate, completing at 18 seconds. The residual humidity is 2.0-2.1% relative humidity, and the energy consumption is 0.40 kWh.

[0029] For example, a polypropylene centrifuge tube (15 mm in diameter, 40 mm in height); preset parameters: N0 = 150 rpm, Q0 = 10 L / min. =3% relative humidity, clamping force 1.5N. The rapid drying phase is stable with no stubborn water stains. However, the digital twin model predicts a slight lag in the top annular area during the drying process (predicting a humidity drop rate of only 0.3% relative humidity / s in the next 2 seconds). The upper nozzles are pre-adjusted upwards by 5°, and the process directly enters slow drying at the 8th second, completing the process at the 16th second. The residual humidity is 2.9-3.1% relative humidity, and the energy consumption is 0.27kWh.

[0030] For example, a long, narrow test tube (8 mm in diameter, 150 mm in height), N0 = 250 rpm, Q0 = 18 L / min, =2% relative humidity. Adjust the height to 5mm from the bottom of the cup. At the 10th second, the humidity in the bottom area is relatively high. After mapping and calculation, the coordinates of the bottom area are (0, -4, -70). The predicted optimal time is t+Δt=10.4s. Drive the corresponding bottom nozzles 1 and 2 to deflect by 25° and 20° respectively. The rotation speed of the rotating support platform is 125rpm. The flow rate of the multi-angle jet assembly is 27L / min. At the same time, the upper nozzles maintain an auxiliary angle to form a coordinated airflow from top to bottom. The process is completed at the 28th second. The residual humidity is 1.9-2.1% relative humidity, and the energy consumption is 0.48kWh.

[0031] For example, in a fault-tolerant demonstration, during the drying process, nozzle #5 became stuck due to a foreign object, and the feedback angle remained at 0°. The controller immediately identified the fault and entered fault-tolerant mode: it recalculated and assigned the task of nozzle #5 to the adjacent nozzles #4 and #6, which deflected to 22° and 19° respectively, and slightly reduced the rotation speed to 90 rpm. The drying process was not significantly affected, and a maintenance alarm was issued, demonstrating the robustness of the system.

[0032] For example, this application discloses a coordinated actuator linkage control system based on spatial humidity distribution mapping, which utilizes the aforementioned coordinated actuator linkage control method based on spatial humidity distribution mapping, including: The initialization module is used to complete system initialization and establish the cup body fixed coordinate system C-System and the nozzle array coordinate system N-System; The humidity sensing module drives the humidity sensor array to collect humidity data of the cup in real time; based on the coordinate system transformation matrix, it maps the raw humidity data to the cup's fixed coordinate system C-System; and preprocesses the humidity data. The cooperative motion control module receives coordinate information of areas with excessive humidity, performs cooperative calculations to determine the optimal spraying time, and generates actuator cooperative commands through inverse kinematics calculations. The feedforward compensation module is used to train a digital twin model of the drying process and predict the rate of change of humidity over time in each region of the cup wall under the combined parameters; when the rate of change of humidity in a region is lower than the threshold, feedforward compensation is automatically performed. The monitoring module is used to collect and compare the actual feedback angle, air pressure sensor, and flow sensor data of the nozzle actuator in real time, providing a basis for fault diagnosis. The reconfiguration module is used to automatically identify actuator faults; when a fault occurs, the fault-tolerant reconfiguration mode is immediately activated: the kinematics are recalculated and new cooperative instructions are generated. The segmented adaptive control module is used to execute the segmented adaptive control strategy until the humidity and humidity standard deviation of all areas meet the drying completion conditions. The alarm module is used to receive fault signals and perform timeout protection and fault alarm.

[0033] The above description is merely a preferred embodiment of the present invention. It should be understood that the present invention is not limited to the forms disclosed herein and should not be construed as excluding other embodiments. It can be used in various other combinations, modifications, and environments, and can be altered within the scope of the concept described herein through the above teachings or related technologies or knowledge. Modifications and variations made by those skilled in the art that do not depart from the spirit and scope of the present invention should be within the protection scope of the appended claims.

Claims

1. A coordinated actuator linkage control method based on spatial humidity distribution mapping, characterized in that, Includes the following steps: S1. The system initializes and establishes a virtual coordinate system, which includes a cup-body fixed coordinate system (C-System) and a nozzle array coordinate system (N-System). The cup-body fixed coordinate system (C-System) has its origin at the center of the rotating support platform and rotates with the platform. The nozzle array coordinate system (N-System) has its origin at the multi-angle jet assembly mounting base. The transformation matrix between the cup-body fixed coordinate system (C-System) and the nozzle array coordinate system (N-System) is determined through a calibration program. This transformation matrix is ​​a function of the real-time angle of the rotating support platform. Simultaneously, the height position of the multi-angle jet assembly is adjusted using a height adjustment mechanism. S2. Start the drying operation. The rotating support platform runs at an initial speed N0. The multi-angle jet assembly delivers pulsed air at an initial flow rate Q0. The humidity sensor array collects humidity data of each area of ​​the cup in real time and maps it to the cup's fixed coordinate system C-System. S3. If, in the fixed coordinate system C-System of the cup, the humidity sensor array detects that the humidity in the spatial region P exceeds a preset humidity threshold, then a collaborative calculation is performed in the controller to determine the optimal spraying time. And through inverse kinematics calculation, a coordinated command is generated to determine the combination of nozzles involved in the purging and their deflection angle, so as to accurately purge the spatial region P; S4. Train a digital twin model of the drying process using several sets of historical drying data to predict the rate of change of humidity over time in each region of the cup wall under a combination of coordinated parameters. The combination of coordinated parameters includes the set values ​​of nozzle angle, rotation speed, and flow rate at the current time and in the future time. Solidify the trained model into the central controller and call it in real time during the drying process. If the rate of change of humidity over time in a certain region is lower than the preset threshold, adjust the corresponding nozzle angle or the rotation speed of the rotating support table in advance to perform feedforward compensation. S5. In real time, compare the command angle of each nozzle with the actual feedback angle, air pressure and flow sensor data. If the data is abnormal and causes the nozzle actuator to malfunction, immediately start the fault-tolerant reconstruction mode, including recalculating, generating new collaborative commands, reducing the rotation speed of the rotating platform, and adjusting the angle and flow of a preset number of adjacent nozzles to compensate for the failure area, while issuing an alarm prompt. S6. A segmented adaptive control strategy is adopted until the humidity of all areas does not exceed the preset humidity threshold within a preset time and the humidity standard deviation is less than the preset relative humidity; if a timeout occurs, timeout protection and fault alarm are activated.

2. The collaborative actuator linkage control method based on spatial humidity distribution mapping according to claim 1, characterized in that: The calibration procedure described in step S1 includes scanning a laser target or using a standard cup for multi-point scanning.

3. The collaborative actuator linkage control method based on spatial humidity distribution mapping according to claim 1, characterized in that, Step S3, which describes performing collaborative solution in the controller, includes the following sub-steps: S31, Set look-ahead time window Discretize the window into M equally spaced time intervals. , Where t represents time, and the discrete time indices are k=1,...,M. Represents discrete time intervals, based on the current rotational speed. And the expected speed adjustment, predict the angle of the bearing platform. , Where θ(t) represents the angle of the rotating platform at time t, and n represents the rotational speed of the rotating platform; calculate the time intervals. The mapped coordinates P' of the lower spatial region P in the nozzle array coordinate system N-System. ,in The transformation matrix represents the coordinate system C-System of the cup body and the coordinate system N-System of the nozzle array; then, the distance between the mapped coordinate P' and the center line pointed to by each nozzle is calculated, and the earliest time when the distance is less than a preset threshold is selected as the optimal spraying time. If there is no moment that meets the conditions, the rotation speed n of the rotating bearing platform is adjusted so that P enters the effective range more quickly and the prediction is repeated. S32. Based on the mapped coordinates P', determine the nozzle combination involved in the purging and its deflection angle through geometric inverse kinematics calculation; specifically including: To make the nozzle i Pointing to the mapped coordinates P', using the formula Calculate the unit direction vector ,in Indicates nozzle i The fixed installation position in the nozzle array coordinate system N-System, and then through the formula and Reverse solution nozzle Horizontal swing angle and pitch angle ,in Represents the unit direction vector Coordinates on the y-axis Represents the unit direction vector The coordinates of v on the x-axis i,z Represents the unit direction vector v i The coordinates on the z-axis; select the preset number of nozzles closest to the mapped coordinate P', calculate their required horizontal swing angle and pitch angle respectively, and adjust the rotation speed of the rotating platform to... ; If the velocity of the spatial region P within the prediction window exceeds the preset velocity threshold, then fine-tuning is performed. The speed at which the spatial region P sweeps across the nozzle focusing area does not exceed a preset speed threshold, wherein the nozzle focusing area includes the nozzle. i And the selected preset number of nozzles.

4. The collaborative actuator linkage control method based on spatial humidity distribution mapping according to claim 1, characterized in that, Step S4, which involves training a digital twin model of the drying process using several sets of historical drying data, includes the following sub-steps: S41. Data acquisition: Record the input parameter sequence and output parameters of the complete drying process for different cup shapes and different initial humidity distributions; the input parameter sequence includes the nozzle angle, the rotation speed of the rotating support table, and the airflow rate; the output parameters include the humidity change curve of each sensor area over time. S42. Data preprocessing: normalize, denoise, and align the collected input parameter sequence and output parameters. S43. Feature engineering: Extracting key features from the preprocessed input parameter sequence and output parameters, including the current humidity distribution, humidity change rate and actuator status; S44. Model selection and training: A time-series prediction model is constructed using a Long Short-Term Memory (LSTM) network or a gated recurrent unit (GRU). The Adam optimizer is used to train the model with the goal of minimizing the prediction error. The ratio of the training set to the validation set is 8:

2. The current state vector and the future control parameter sequence are used as its inputs, and the predicted humidity values ​​of each sensor area in the future are used as its outputs. The future control parameter sequence includes the nozzle angle, rotation speed, and flow rate at each future time. S45. Model Deployment: The trained model is embedded into the central controller and called in real time during the drying process to predict the rate of change of humidity over time in various regions of the cup wall under the combined parameters.

5. The collaborative actuator linkage control method based on spatial humidity distribution mapping according to claim 1, characterized in that: If abnormal data causes the nozzle actuator to malfunction as described in step S5, it includes nozzle jamming or angle deviation exceeding the preset threshold for nozzle angle.

6. The coordinated actuator linkage control method based on spatial humidity distribution mapping according to claim 3, characterized in that, The segmented adaptive control strategy described in step S6 specifically includes the following sub-steps: S61. Rapid drying stage: Run with default coordination parameters and calculate the first humidity change rate in sensor region j. , ,in Indicates the sampling period. This represents the humidity value of sensor region j at time t. Indicates that sensor region j is in Humidity value at time, the first humidity change rate The temperature will fluctuate due to instantaneous changes in the sensor; therefore, the first humidity change rate of the current and the previous L-1 sampling points is taken. The average value is used as the output to obtain the second humidity change rate. , Where L represents the window length, Indicates the offset index of the sampling point within the window; if the second humidity change rate Three consecutive times the humidity level is below the preset threshold. And the current humidity If so, it is determined that there is a stubborn water stain area, and the process enters the localized intensive drying stage; in Indicates the hysteresis interval. Indicates the preset humidity threshold; S62, Localized Intensive Drying Stage: Identify the center of the stubborn water stain area. Based on the determination of the nozzle to be deflected and its target angle in step S3, generate coordinated commands, including deflecting the nozzle to the target angle and reducing the rotation speed of the rotating platform. At this time, the angle of the bearing platform The calculation is performed using uniform angular velocity motion, i.e. Increase the airflow of the multi-angle jet assembly to 1.5Q0 and continuously track the humidity of stubborn water stains until it subsides; S63, Slow drying stage, when the average humidity of all sensor areas... At that time, reduce the rotation speed of the rotating platform to... This reduces the airflow rate of the multi-angle jet assembly to The PID controller is activated to fine-tune the flow rate, while the residual humidity uniformity is predicted through a digital twin model of the drying process. When it is predicted that the humidity of a certain area will deviate from that of other areas by more than 0.5% relative humidity, or when the current sensor detects that the humidity of a certain area deviates from the average value by more than 1% relative humidity, the small angle adjustment amount is automatically calculated and the corresponding nozzle is adjusted to provide additional airflow compensation to that area to eliminate the small deviation.

7. The collaborative actuator linkage control method based on spatial humidity distribution mapping according to claim 6, characterized in that: In step S61, the center position of the stubborn water stain area and the corresponding nozzle angle are recalculated in each control cycle to achieve dynamic focusing. If the center position of the stubborn water stain area moves, the nozzle angle will move accordingly in real time.

8. A coordinated actuator linkage control system based on spatial humidity distribution mapping, employing the coordinated actuator linkage control method based on spatial humidity distribution mapping as described in any one of claims 1-7, characterized in that, include: The initialization module is used to complete system initialization and establish the cup body fixed coordinate system C-System and the nozzle array coordinate system N-System; The humidity sensing module drives the humidity sensor array to collect humidity data of the cup in real time; based on the coordinate system transformation matrix, it maps the raw humidity data to the cup's fixed coordinate system C-System; and preprocesses the humidity data. The cooperative motion control module receives coordinate information of areas with excessive humidity, performs cooperative calculations to determine the optimal spraying time, and generates actuator cooperative commands through inverse kinematics calculations. The feedforward compensation module is used to train a digital twin model of the drying process and predict the rate of change of humidity over time in each region of the cup wall under the combined parameters; when the rate of change of humidity in a region is lower than the threshold, feedforward compensation is automatically performed. The monitoring module is used to collect and compare the actual feedback angle, air pressure sensor, and flow sensor data of the nozzle actuator in real time. Provide a basis for fault diagnosis; The reconfiguration module is used to automatically identify actuator faults; when a fault occurs, the fault-tolerant reconfiguration mode is immediately activated: the kinematics are recalculated and new cooperative instructions are generated. The segmented adaptive control module is used to execute the segmented adaptive control strategy until the humidity and humidity standard deviation of all areas meet the drying completion conditions. The alarm module is used to receive fault signals and perform timeout protection and fault alarm.

Images