Intelligent Control Method and System for Multi-Source Data Environment of Spray Painting Booth

Abstract

This invention provides a multi-source data environment intelligent control method and system for spray booths, relating to the field of industrial automation control technology. By acquiring the process action sequence of the spray production line control system, it calculates the single-action disturbance curve of the process action on environmental parameters before the process action is executed. Combined with the disturbance mode and the instantaneous evaporation rate of the coating, the disturbance amplitude is dynamically corrected to generate a composite disturbance prediction result. Before the disturbance arrives, the feedforward execution timestamp is calculated and the guide vane angle command is issued. Simultaneously, a direct control of the airflow path is used in conjunction with the guide vanes to adjust the airflow, greatly reducing overall energy consumption, improving control precision, achieving refined proactive operation and energy saving. This ensures that the actuator has completed pre-adjustment actions when the disturbance peak arrives, significantly reducing the environmental parameter response delay to the millisecond level, essentially achieving zero hysteresis. This reduces the probability of VOCs concentration exceeding limits and the rate of spray quality defects, improving the safety and process stability of the spray booth.

Description

Technical Field

[0001] This invention relates to the field of industrial automation control technology, specifically to a method and system for intelligent control of multi-source data environment in spray booths. Background Technology

[0002] Spray painting booths are critical workstations in industrial coating production lines. Their internal environmental parameters, such as temperature, humidity, VOC concentration, airflow organization, and pressure distribution, directly affect coating quality, operational safety, and energy consumption. Existing spray painting booth environmental control systems primarily employ feedback control strategies: sensors deployed at exhaust vents or operating areas monitor VOC concentration or pressure differential in real time. When the detected values ​​exceed preset thresholds, a PID controller adjusts the exhaust fan speed or damper opening to maintain environmental parameters within permissible ranges. For example, CN119395981A discloses a spray painting system capable of intelligently adjusting the overall line's airflow balance. This system, based on an airflow balance control algorithm, collects airflow, pressure differential, and VOC concentration data using multi-dimensional sensors and utilizes a PID control strategy to adjust the fan and proportional valves to achieve overall line airflow balance. Additionally, other technical solutions involve deploying multiple monitoring points within the spray painting booth to analyze the air pollution coefficients of each area and identify hazardous monitoring points, then selectively activating exhaust and intake air treatments. Existing technologies have the following inherent drawbacks:

[0003] First, existing spray booth environmental control systems employ a feedback control architecture, which inherently introduces a delay of several to tens of seconds from VOCs release, diffusion, sensor response to actuator action. When processes such as spray gun startup, color change cleaning, or workpiece door opening and closing occur, VOCs concentration rises rapidly, but the control system only begins to adjust after the concentration exceeds the standard, leading to increased safety risks. Furthermore, it cannot provide feedforward suppression of environmental disturbance sources, and the feedback control for spraying exhibits lag and low predictability, failing to adjust before environmental disturbances occur, resulting in poor spraying quality.

[0004] Secondly, existing technologies alter airflow by adjusting exhaust fan speed or damper opening, indirectly affecting airflow distribution and pressure balance within the spray booth. Because the fans operate in inefficient zones, energy consumption remains high, and the impact of airflow changes on the overall airflow field is non-linear, making precise control of local airflow paths difficult. When the streamline direction of a certain area needs to be changed, increasing the airflow often causes airflow turbulence in other areas, resulting in slow response, high energy consumption, and an inability to achieve precise spatial control of airflow paths with low energy consumption. Summary of the Invention

[0005] To achieve the above objectives, the present invention provides the following technical solution: a multi-source data environment intelligent control method for spray booths, the method comprising: Acquire the process sequence of the spraying production line, the dynamic data of paint volatilization at the spray gun outlet, the airflow field data and environmental feedback data in the spray booth; Based on the process action sequence and a pre-built action disturbance mapping library, the single action disturbance curves of each process action on the environmental parameters inside the spray booth are predicted. The amplitude of the single action disturbance curves is corrected based on the coating volatilization dynamic data and then superimposed on the time axis to generate the composite disturbance prediction result. The current airflow path is determined based on the airflow field state data, the ideal airflow path is determined based on the composite disturbance prediction results, and the path deviation field between the current airflow path and the ideal airflow path is calculated. The feedforward execution time is calculated based on the composite disturbance prediction results. The target angle and execution timestamp of each guide vane are calculated based on the path deviation field, and the feedforward execution command is generated. The feedforward execution instructions are issued according to the execution timestamp, and the execution status feedback and post-execution environmental feedback data are obtained. The environmental feedback data is compared with the composite disturbance prediction results to calculate the deviation signal. Based on the deviation signal, a feedback correction command is generated and issued for execution. Based on the deviation signal and execution status feedback, update the motion disturbance mapping library and the calculation parameters of the feedforward execution time.

[0006] Furthermore, the present invention provides a multi-source data environment intelligent control system for spray booths, used to implement the above-mentioned multi-source data environment intelligent control method for spray booths, the system comprising: The data acquisition module is used to acquire the process sequence of the spraying production line, the dynamic data of paint volatilization at the spray gun outlet, the airflow field status data and environmental feedback data in the spray booth; The disturbance prediction module is used to predict the single-action disturbance curve of each process action based on the process action sequence and a pre-built action disturbance mapping library, and to generate composite disturbance prediction results by combining the coating volatilization dynamic data. The path planning module is used to determine the current airflow path based on airflow field state data, determine the ideal airflow path based on composite disturbance prediction results, and calculate the path deviation field between the current airflow path and the ideal airflow path. The control quantity calculation module is used to calculate the feedforward execution time based on the composite disturbance prediction results and generate feedforward execution instructions based on the path deviation field. The execution control module is used to issue feedforward execution instructions according to the execution timestamp, and to obtain execution status feedback and post-execution environmental feedback data; The feedback correction module is used to compare environmental feedback data with composite disturbance prediction results, calculate deviation signals, generate feedback correction instructions based on deviation signals, and execute them. The parameter update module is used to update the calculation parameters of the motion disturbance mapping library and feedforward execution time based on the deviation signal and execution status feedback.

[0007] This invention provides a method and system for intelligent control of multi-source data environment in spray booths. It has the following beneficial effects: 1. This invention acquires the process action sequence of the spraying production line control system and predicts the impact curves of disturbance sources such as spray gun start-up, color change cleaning, and workpiece door opening and closing on VOCs concentration, pressure, temperature, and humidity before these actions occur. By reading the disturbance mode of each action from the action disturbance mapping library and combining it with the instantaneous evaporation rate of the paint monitored in real time at the spray gun outlet, the disturbance amplitude is dynamically corrected to generate a composite disturbance prediction result. Based on the composite disturbance prediction result, the feedforward execution timestamp is calculated and the guide vane angle command is issued before the disturbance arrives, so that the actuator has completed the pre-adjustment action when the disturbance peak arrives. This greatly reduces the response delay of environmental parameters to the millisecond level, basically achieving zero lag, while reducing the probability of VOCs concentration exceeding the limit and the spraying quality defect rate. Compared with the lag adjustment that relies on sensor feedback in the prior art, this invention achieves a fundamental shift from post-compensation to pre-adjustment, improving the safety and process stability of the spraying booth.

[0008] 2. This invention abandons the indirect control mode of existing technologies that rely solely on fan airflow adjustment. Instead, it uses adjustable guide vanes as the main actuator to directly manipulate the spatial shape of the airflow path. By reconstructing the airflow velocity vector field in the spray booth in real time, it generates the current airflow path and compares it with the ideal airflow path dynamically determined based on the operating conditions. The path deviation field calculated by the comparison is converted into the target angle of each guide vane. The airflow path is corrected primarily through low-power blade angle adjustment. The total airflow is only adjusted as an auxiliary measure when the blade adjustment range is insufficient. This significantly increases the time percentage of the exhaust fan's high-efficiency zone and greatly reduces overall energy consumption. Furthermore, because it directly controls the airflow direction rather than indirectly adjusting the airflow, it improves the suppression accuracy of problems such as local VOCs accumulation and airflow short-circuiting, achieving the dual goals of refined active control of airflow organization and energy saving. Attached Figure Description

[0009] Figure 1 This is a flowchart illustrating the steps of the intelligent control method for multi-source data environment of a spray booth according to the present invention. Figure 2 This is a data flow diagram of the intelligent control method for multi-source data environment of the spray booth according to the present invention; Figure 3 This is an architecture diagram of the intelligent control system for multi-source data environment of the spray booth according to the present invention. Detailed Implementation

[0010] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0011] like Figures 1 to 2 As shown, the intelligent control method for multi-source data environment of the spray booth includes: Step S100 involves acquiring the process sequence of the spraying production line, dynamic data on paint evaporation at the spray gun outlet, airflow field data within the spray booth, and environmental feedback data. The process sequence reflects the deterministic operational instructions of the spraying production line over a future period, representing the root cause of environmental disturbances within the spray booth and providing a basis for subsequent feedforward predictions in terms of timing and action types. The dynamic data on paint evaporation, through real-time monitoring of the instantaneous evaporation rate of paint mist at the spray gun outlet, reflects the actual evaporation characteristics of the currently used paint at the moment of spraying. The airflow field data reflects the true distribution of airflow organization at the current moment, providing a benchmark for judging the rationality of the airflow path. The environmental feedback data reflects the actual effect of environmental control.

[0012] Step S200: Based on the process action sequence and a pre-built action disturbance mapping library, predict the single-action disturbance curves of each process action on the environmental parameters inside the spray booth. After amplitude correction based on paint volatilization dynamic data, the single-action disturbance curves are superimposed along the time axis to generate a composite disturbance prediction result. The action disturbance mapping library records the specific disturbance patterns generated by each process action on the environmental parameters at each monitoring point inside the spray booth, including the area, magnitude, and duration of the disturbance impact. This reflects the inherent correlation between various process actions and the environmental response of the spray booth. The composite disturbance prediction result reflects the overall trend and peak occurrence time of the environmental parameters inside the spray booth over a future period under the combined effect of multiple disturbance sources.

[0013] Step S300: Determine the current airflow path based on airflow field state data, determine the ideal airflow path based on composite disturbance prediction results, and calculate the path deviation field between the current airflow path and the ideal airflow path. The path deviation field is used to quantify the spatial difference between the current airflow path and the ideal airflow path, reflecting in which areas the current airflow path deviates from the ideal route, whether it deviates from the exhaust outlet direction, or whether vortices have formed in the operating area. This provides a precise and quantitative spatial basis for adjusting the angle of each guide vane to correct the airflow direction, enabling airflow control to transform from blind airflow adjustment to spatial and precise control of airflow direction.

[0014] In step S400, the feedforward execution time is calculated based on the composite disturbance prediction results. The target angle and execution timestamp of each guide vane are calculated based on the path deviation field, and a feedforward execution command is generated. The feedforward execution command is used to adjust the actuator to a predetermined state in advance before the actual arrival of the environmental disturbance, reflecting the entire control system's ability to predict future disturbances and its intention to actively intervene.

[0015] Step S500: Issue feedforward execution instructions according to the execution timestamp, and acquire execution status feedback and post-execution environmental feedback data. Execution status feedback monitors the physical execution of the feedforward execution instructions, reflecting whether the actuator accurately completed the system's instructions. Post-execution environmental feedback data reflects the changes and dynamic responses of the actual environmental state inside the spray booth after the feedforward action is executed.

[0016] Step S600: The environmental feedback data is compared with the composite disturbance prediction results to calculate the deviation signal. Based on the deviation signal, a feedback correction command is generated and issued for execution. The feedback correction command performs real-time fine-tuning and compensation for the residual error of the feedforward control, reflecting the robustness and adaptability of the system in pursuing high-precision control. It ensures that even when the prediction is not completely accurate, the actual environmental parameters can be effectively controlled within the target range, realizing an organic combination of feedforward active prediction and feedback passive correction.

[0017] Step S700: Based on the deviation signal and execution status feedback, update the motion disturbance mapping library and the calculation parameters of the feedforward execution time. By updating the motion disturbance mapping library, the deviation between the actual environmental feedback data and the predicted data generated in the current control cycle is used to dynamically correct and optimize the correspondence between process actions and disturbance modes, making the motion disturbance mapping library closer to the actual response characteristics of the current production line, thereby improving the accuracy of the next prediction. By updating the calculation parameters of the feedforward execution time, the physical response deviation of the actuator itself is eliminated, making the calculation of the feedforward execution timestamp more consistent with the true characteristics of the physical actuator, thus improving accuracy.

[0018] In this embodiment, the implementation steps for acquiring the process sequence of the spraying production line, the dynamic data of paint evaporation at the spray gun outlet, the airflow field state data and environmental feedback data in the spray booth include: Step S101: Connect to the PLC controller of the spraying production line via an industrial Ethernet interface. Read the process action instructions and their corresponding timestamps and status values ​​from the PLC controller of the spraying production line at a set sampling period. The sampling period is generally set to be no less than 100ms. Sort the read process action instructions by time to generate a process action sequence within a future time window. The process action sequence includes action type, trigger time, and action parameters. The action types of the process action instructions include spray gun start / stop instructions, robot color change and cleaning instructions, workpiece door opening / closing instructions, and workpiece conveying instructions. The spray gun start / stop instructions include the spray gun number, position coordinates, and spraying duration, used to accurately locate the position, intensity, and duration of the disturbance source (VOCs generation point). Robot color-changing cleaning instructions should include at least the cleaning duration and solvent type to predict the intensity and duration of secondary disturbances caused by solvent evaporation during the cleaning process. Solvent types typically include water-based cleaning solvents, organic solvents (such as acetone, toluene, etc.), and strong cleaning agents. Solvent type distinguishes the evaporation rate and toxicity of different solvents, as different types of solvents exhibit significant differences in VOC concentration and composition during cleaning, directly affecting the amplitude correction of disturbance predictions. Workpiece door opening / closing instructions should include at least the opening duration and opening angle to predict the disturbance impact of external air influx on airflow and pressure within the booth. Workpiece conveying instructions should include at least the workpiece type and entry / exit direction to determine the area where disturbances occur and the workpiece's obstruction effect on the flow field. Workpiece type, such as car bumpers, dashboards, or body side panels, is used to determine the workpiece's dimensions and aerodynamic characteristics. Different workpieces have different obstruction and disturbance effects on airflow when moving within the spray booth, affecting the planning of local airflow paths.

[0019] Step S102: At the spray gun outlet, the instantaneous evaporation rate of the solvent in the paint mist is collected at a set sampling frequency. After the instantaneous evaporation rate is processed by sliding window filtering, it is matched with a preset paint feature library to obtain dynamic data on paint evaporation. The instantaneous evaporation rate refers to the change in mass or volume of solvent evaporating from the paint mist into the air per unit time at the spray gun outlet. It dynamically reflects the real-time evaporation intensity of the sprayed paint after atomization, providing an accurate basis for correcting the VOCs disturbance amplitude caused by the spray gun operation. The instantaneous evaporation rate is obtained in real time by continuously collecting fluorescence signals of specific wavelengths in the paint mist at a frequency of not less than 50Hz using an optical detection probe installed at the outlet of each spray gun, employing laser-induced fluorescence technology. The signal intensity is calculated in real time based on the calibrated relationship between the signal intensity and the solvent concentration. The pre-built coating feature library is a pre-established database that stores the baseline volatility characteristic curves, characteristic parameters, and corresponding coating type labels of different coatings (such as water-based paints, solvent-based paints, and high-solids paints) under standard conditions. This is used to match the real-time measured evaporation rate curves with the standard curves in the library, thereby quickly identifying the type of coating currently in use and providing a benchmark for subsequent amplitude correction. The baseline evaporation characteristic curve contains the baseline evaporation rate of that coating type under standard test conditions. The pre-built coating feature library is established based on a large amount of offline experimental data, which includes evaporation tests conducted on coatings of different brands and formulations under standard temperature and wind speed conditions. Common characteristics were extracted and categorized accordingly.

[0020] When generating dynamic data on coating evaporation, the original instantaneous evaporation rate data is first filtered by a sliding window (window length 0.2 seconds) to eliminate impulse noise. A continuous curve is plotted with time as the x-axis and instantaneous evaporation rate as the y-axis, which is the evaporation rate curve, reflecting the change of solvent evaporation rate with time during a single spraying process. Next, the filtered evaporation rate curve is matched with various benchmark evaporation characteristic curves in the coating feature library to find the most similar benchmark evaporation characteristic curve. The corresponding function model and characteristic parameters (such as initial evaporation rate, decay time constant, etc.) are extracted from this benchmark evaporation characteristic curve. Then, the amplitude of the function model is scaled according to the actual measured peak evaporation rate to generate a evaporation characteristic curve function that precisely matches the spraying process and includes the coating type label. The generated evaporation characteristic curve function is a mathematical model (such as an exponential decay function, polynomial function, etc.) that describes the law of evaporation rate change with time in analytical form, digitizing and parameterizing the discrete evaporation rate curve, facilitating rapid numerical calculation and comparison by the control system. The coating type label includes at least water-based paint, solvent-based paint, and high-solids paint. Finally, the instantaneous evaporation rate, coating type label, and evaporation characteristic curve function are integrated to form dynamic data of coating evaporation.

[0021] Step S103: Collect wind speed vectors and static pressure values ​​at N measuring points within the spray booth, where N is an integer. Calculate the airflow velocity vector field and pressure field of the continuous space based on the wind speed vectors and static pressure values ​​as airflow field state data. First, array-type micro-pressure sensors (including 64 measuring points) and hot-wire anemometers (including 16 measuring points) are arranged at key sections within the spray booth, including above the painting area, in front of the exhaust vent, and the operating area. Wind speed vectors and static pressure values ​​at each measuring point are collected at a frequency of at least 20Hz, and the wind speed vector includes components in the X, Y, and Z directions. Then, the collected discrete data with spatial coordinates is reconstructed using spatial interpolation methods (such as Kriging interpolation, radial basis function interpolation, etc.) to calculate the wind speed vector and pressure value at any point within the continuous space, thereby constructing a complete airflow velocity vector field and pressure field. The final airflow velocity vector field is a spatialized dataset describing the flow velocity and direction (including components in the X, Y, and Z directions) of the airflow at any point (X, Y, Z coordinates) within the spray booth. The pressure field describes the pressure value at any point inside the booth. The airflow velocity vector field and the pressure field together constitute a complete digital description of the airflow organization morphology inside the spray booth, forming a set of airflow field state data. This data forms the basis for streamline tracing and path deviation analysis. The airflow field state data is a snapshot of the overall state of the airflow organization inside the spray booth at a certain moment, including the airflow velocity vector field and pressure field in continuous space. It serves as a benchmark for judging whether the current airflow path is reasonable and provides the current state input for generating feedforward execution commands.

[0022] Spatial interpolation is a method that uses known points to estimate the values ​​of unknown points, based on the first law of geography, which states that points that are spatially closer are more similar in their attribute values. In step S103, the following steps are taken: First, the coordinates and measured values ​​(e.g., wind speed in the X direction) of all discrete measuring points within the spray booth are used as input. Then, for the unknown spatial point to be estimated, several known measuring points within a certain distance range are found. Based on the spatial distance between these known measuring points and the unknown point, the weight of each known point is calculated; the closer the distance, the greater the weight. Finally, the measured values ​​of the known points and their corresponding weights are weighted and summed to obtain the estimated value of the unknown point. This process is repeated for all points in space that need to be estimated, ultimately forming a continuous field distribution.

[0023] Step S104: Environmental parameters are collected from M measuring points in the operating area, exhaust vents, and near the workpiece surface within the spray booth as environmental feedback data. This environmental feedback data includes temperature, humidity, and VOC concentration. Specifically, three measuring points are set up in the operating area, two at the exhaust vents, and four near the workpiece surface. Temperature and humidity sensors and VOC concentration sensors are installed at these points, and environmental parameters are collected at a frequency of at least 10Hz. Outlier removal and median filtering are performed on the collected environmental parameters to obtain the final environmental feedback data. Temperature data reflects the heat distribution within the spray booth and is used in the correction calculation of VOC concentration (because the evaporation rate is affected by temperature) and in assessing worker comfort. Humidity data reflects the water vapor content in the air, affecting the drying speed of water-based paints and the evaporation characteristics of solvent-based paints. VOC concentration data reflects the concentration of combustible gases and safety risks within the spray booth; it is the core target parameter for the entire environmental control process, providing the most direct comparison basis for feedforward prediction and feedback correction. Environmental feedback data is a collection of real-time monitoring values ​​of temperature, humidity, and VOCs concentration at key locations such as the operating area, exhaust vents, and workpiece surface within the spray booth. As the actual output of closed-loop control, it is compared with feedforward prediction results to calculate deviation signals, thereby driving the generation of feedback correction commands. It also serves as the basis for updating the action disturbance mapping library, reflecting the actual environmental effects after the implementation of control measures.

[0024] The outlier removal calculation logic identifies and removes obviously unreasonable measurement values ​​caused by occasional sensor malfunctions, signal interference, etc., based on the Laida criterion in statistics or judgments based on physical thresholds. For example, if the VOCs concentration at a certain measuring point suddenly jumps to a value far exceeding the normal range (such as a negative value or exceeding the sensor's range), it is judged as an outlier and directly removed from the data sequence. In the calculation of median filtering, for a continuous measurement data sequence, a fixed-length sliding window is set. Each time, the median of all data within the window (i.e., the middle number after sorting by size) is taken as the output value of the current point. This is mainly based on the fact that the median can effectively resist the influence of impulse noise (i.e., individual abnormally large or small spikes). Compared with mean filtering, it can better preserve the edge and detail information of the signal while smoothing the data, avoiding data distortion caused by individual outliers.

[0025] In this embodiment, based on the process action sequence and a pre-built action disturbance mapping library, the implementation steps for predicting the single-action disturbance curves of each process action on the environmental parameters inside the spray booth, and then superimposing the single-action disturbance curves on the time axis after amplitude correction based on the paint volatilization dynamic data to generate the composite disturbance prediction result include:

[0026] Step S201: Construction and updating of the action disturbance mapping library. First, based on historically output process action sequences and historical environmental feedback data, each process action in the historical data is temporally aligned and spatially correlated with the changes in environmental feedback data over a period of time after its occurrence, extracting the disturbance pattern corresponding to each action type. The extraction process involves first retrieving a process action sequence and synchronously recorded environmental feedback data from the historical database for a certain period. For each process action (e.g., "spray gun start"), the environmental feedback data change sequence within a fixed time window is extracted, starting from its trigger time. Then, by analyzing the peak value of environmental parameters at each monitoring point within this time window, the start time of the change, the duration of the change, and the distribution of affected monitoring points, the influence amplitude matrix, influence time history vector, and influence spatial vector of the process action are calculated. The peak value change of each monitoring point is filled into the corresponding position to obtain the influence amplitude matrix. The influence time history matrix is ​​obtained by recording the time difference from the start of the action to the first response at each monitoring point and the total duration of the disturbance from its start to its decay to stability. The influence spatial vector is obtained by marking the spatial area range of monitoring points where the response peak exceeds a preset threshold. Finally, the calculated influence amplitude matrix, influence time history vector, and influence space vector for each group are extracted and bound to the type of the process action, and stored in the action disturbance mapping library. This ultimately forms the action disturbance mapping library representing the relationship between process actions and disturbance modes. The library is updated at the end of each control cycle.

[0027] The disturbance mode is a digital description of the spatial and temporal response of environmental parameters within the spray booth after a specific process action occurs. The disturbance mode includes an influence amplitude matrix, an influence time history vector, and an influence spatial vector. The influence amplitude matrix records the peak values ​​of the influence of the process action on VOCs concentration, pressure, temperature, and humidity at each monitoring point. The influence time history vector records the delay time from the occurrence of the process action to the start of response at each monitoring point, and the duration of the disturbance. The influence spatial vector records the main affected area of ​​the disturbance, represented by spatial coordinates.

[0028] Step S202: Traverse each process action within the future time window and read the perturbation mode corresponding to the process action from the action perturbation mapping library.

[0029] If the process action is the "spray gun start" command in the spray gun start / stop instructions, the volatilization correction coefficient is calculated based on the paint volatilization dynamic data, and the influence amplitude matrix is ​​corrected based on the volatilization correction coefficient. When the process action is spray gun start, the instantaneous volatilization rate at the current spray gun start time is directly read from the paint volatilization dynamic data obtained in step S102. The reference volatilization rate of the paint under standard test conditions is obtained from the volatilization characteristic curve function extracted from the same set of data. The instantaneous volatilization rate is divided by the reference volatilization rate, and the resulting ratio is the volatilization correction coefficient of the paint. The influence amplitude matrix is ​​multiplied by the volatilization correction coefficient to obtain the corrected influence amplitude matrix. By correcting the influence amplitude matrix, the influence of differences in the volatilization characteristics of different paints on the prediction accuracy is eliminated. Since the perturbation patterns stored in the action perturbation mapping library are based on historical data (possibly based on a certain standard paint), if no correction is made, when using paints with significantly different volatilization rates, the predicted VOCs concentration peak will deviate significantly from the actual value, causing feedforward control to fail.

[0030] Based on the trigger time and impact time history vector of each process action, a single-action disturbance curve is generated. The single-action disturbance curve is a predicted curve describing the change of specific environmental parameters within the spray booth over time after a single process action occurs. It serves as the basic unit for predicting composite disturbances, reflecting the impact process of a single disturbance source on the environment. The single-action disturbance curve includes VOCs concentration disturbance curves, pressure disturbance curves, and temperature and humidity disturbance curves. The VOCs concentration disturbance curve reflects the change of VOCs concentration at each spatial point over time, the pressure disturbance curve reflects the change of pressure over time, and the temperature and humidity disturbance curve reflects the change of temperature and humidity over time. The generation process of the VOCs concentration disturbance curve is as follows: First, the influence amplitude matrix and influence time history vector corresponding to the process action are read from the action disturbance mapping library. Taking the process action as the trigger time point, an initial disturbance curve shape (e.g., rapid rise followed by exponential decay) is generated according to the response delay and disturbance duration recorded in the influence time history vector. Next, if the process action is spray gun start-up, the curve amplitude is scaled by multiplying the VOCs-related amplitude in the influence amplitude matrix by the volatilization correction coefficient. Finally, the complete curve of VOCs concentration change over time at various points in space is obtained. The generation process of the pressure disturbance curve and the temperature and humidity disturbance curve is similar. Since pressure and temperature and humidity are not directly affected by the paint volatilization rate, the amplitude correction does not depend on the volatilization correction coefficient. Instead, the corresponding pressure and temperature and humidity values ​​in the influence amplitude matrix are directly read and combined with the influence time history vector to generate the corresponding temperature and humidity curve and pressure disturbance curve.

[0031] The baseline evaporation rate refers to the standard mass or volume of solvent that evaporates into the air per unit time for a specific coating under preset standard conditions (such as standard temperature, standard airflow speed, and standard atomization pressure). It is used to calculate the evaporation correction factor and is determined by averaging multiple evaporation tests conducted on the coating under strictly controlled environmental conditions in a laboratory setting. Compared to the instantaneous evaporation rate, the baseline evaporation rate is a fixed constant representing the typical characteristics of the coating, while the instantaneous evaporation rate is a value measured in real time at the production site, reflecting the actual change in the evaporation state of the coating at the current moment.

[0032] Step S203: Align the single-action disturbance curves of all process actions along the time axis and superimpose them to generate a composite disturbance prediction curve. Extract the peak time point and peak amplitude of the composite disturbance prediction curve. During extraction, the composite disturbance prediction curve is treated as a set of discrete data points with time as the horizontal axis and VOCs concentration as the vertical axis. Traverse all data points, compare their concentration values, record the maximum concentration value and the corresponding time, use the maximum concentration value as the peak amplitude, and output the corresponding time as the peak time point. The composite disturbance prediction curve is obtained by aligning the single-action disturbance curves of all process actions within a future time window according to their respective occurrence times, and superimposing the predicted values ​​of the same type of parameter (such as VOCs concentration) on the same time axis. This results in a comprehensive prediction curve reflecting the change of environmental parameters over time under the combined effect of multiple disturbances. It provides a complete predictive view of the environmental changes in the spray booth over a future period, providing a global quantitative basis for the timing and intensity of feedforward control, and reflecting the comprehensive impact trend of multiple disturbance sources on the environment. The peak time point refers to the time point on the composite disturbance prediction curve when the VOCs concentration reaches its maximum value. It is used to calculate the feedforward execution timestamp to ensure that the actuator completes its pre-adjustment action precisely at the moment of the most severe disturbance. The peak amplitude is the magnitude of this maximum value, used to determine the control priority under the current operating condition (such as whether the safety threshold is exceeded) and for subsequent selection of the ideal airflow path.

[0033] In this embodiment, the steps of determining the current airflow path based on airflow field state data, determining the ideal airflow path based on composite disturbance prediction results, and calculating the path deviation field between the current airflow path and the ideal airflow path include: Step S301: Obtain the current airflow velocity vector field and pressure field from the airflow field state data. Starting from the center point of the spray painting area, trace the streamline along the velocity vector direction to generate a sequence of spatial coordinate points for the current airflow path. First, select a starting point at the center of the spray painting area. Starting from the starting point, obtain the velocity vector (including components in the X, Y, and Z directions) in the airflow velocity vector field at the starting point to determine the airflow direction. Then, move forward a small step along this direction to reach the next spatial point. At the new spatial point, obtain the velocity vector again to determine the new flow direction. Repeat the process of determining the direction and moving a step to reach a new point until the airflow reaches the exhaust vent area or the boundary of the spray booth. Finally, connect all the spatial points in sequence to form a continuous spatial trajectory line, which is the current airflow path. The current airflow path includes a series of continuous spatial coordinate points traced from the center point of the spray painting area along the velocity vector direction to the exhaust vent, showing the actual spatial trajectory of pollutants from their source to their discharge under the current airflow organization state. Streamline tracing is a method that simulates the trajectory of air particles in a flow field based on the airflow velocity vector field. It transforms the abstract airflow velocity vector field into an intuitive and quantifiable airflow trajectory based on the fundamental fluid dynamics principle that the tangent direction of a streamline at any point is consistent with the velocity vector direction at that point.

[0034] Step S302: Determine the control priority based on the peak amplitude in the composite disturbance prediction results and the current operating conditions. Select an ideal airflow path matching the control priority from a preset ideal path library. The ideal airflow path is a sequence of spatial coordinate points starting from the painting area and ending at the exhaust vent. Control priority refers to the weighted ranking of the three control objectives—safety, quality, and energy consumption—under the current operating conditions. It guides the system in selecting the ideal airflow path that best matches the current production objectives from the ideal path library, making the control strategy more targeted and flexible. Control priority includes three modes: safety priority, quality priority, and energy consumption priority. The control priority weights are determined based on the current operating conditions, which include all data output from step S100. When determining the control priority, firstly, the peak value of VOCs concentration in the composite disturbance prediction results is obtained. If the peak value exceeds the preset safety threshold, the "safety" weight is set to the highest. If it does not exceed the safety threshold, the recent spraying quality inspection data is obtained. If the paint film defect rate (such as orange peel, pinholes, etc.) is increased, the "quality" weight is set to the highest. If the quality indicators are normal, the production plan is obtained. If the current production season is off-season or the system has entered energy-saving mode, the "energy consumption" weight is set to the highest.

[0035] Based on the control priority, the path that best matches the current control priority is selected from the preset ideal path library as the ideal airflow path: If the safety weight is set to the highest, then the safety priority mode is selected, and the streamline shape that transports VOCs to the exhaust vent through the shortest path is selected as the ideal airflow path. If the mass weight is set to the highest, then the mass priority mode is selected, and the streamline shape that makes the airflow velocity uniform and the vortex minimized in the operating area is selected as the ideal airflow path. If the energy consumption weight is set to the highest, then the energy consumption priority mode is selected, and the streamline shape that minimizes the total air volume requirement is selected as the ideal airflow path.

[0036] The final ideal airflow path is stored in the form of a spatial streamline, which contains a sequence of continuous spatial coordinate points starting from the paint spraying area and ending at the exhaust vent. This provides an optimized and quantified target reference for the current airflow path, which is used to calculate the path deviation field and guide the guide vanes to perform precise path correction.

[0037] The preset ideal path library is a pre-established spatial streamline database containing the optimal airflow path from the paint spraying area to the exhaust vent under different control objectives. Based on computational fluid dynamics simulation results or historical best control experience, it generates corresponding streamline shapes for three control objectives: safety, quality, and energy consumption. For example, in the safety priority mode, the path is the shortest and points directly to the exhaust vent, while in the quality priority mode, the path is evenly distributed in the operating area and has the minimum eddy current. This provides the control system with a series of preset and verified optimal objectives, avoiding the generation of complex paths online in real time and reducing the computational burden.

[0038] Step S303: Discretize the current airflow path and the ideal airflow path into the same number of spatial sampling points, so that the points on the two paths correspond one-to-one. For each pair of corresponding points, calculate the spatial Euclidean distance between the corresponding sampling points to obtain the position deviation value; calculate the vector direction from the current point to the ideal point to obtain the deviation direction angle; combine the position deviations and deviation direction angles of all corresponding points to form a complete path deviation field, and statistically analyze the maximum deviation value, average deviation value, and deviation direction angle of the path deviation field; based on the path deviation field, sort the Euclidean distances calculated for all corresponding points, take the maximum value as the maximum deviation value, and calculate the average value of all Euclidean distances as the average deviation value; directly take the deviation direction angle of the pair of corresponding points that caused the maximum deviation value as the main deviation direction angle output.

[0039] If the Euclidean distance between the current airflow path point and the ideal airflow path point in a certain spatial region is greater than a preset tolerance threshold, then the region is considered to require an airflow direction adjustment force. The calculation of the airflow direction adjustment force is not directly expressed as a mechanical quantity, but rather transformed into the target angle of the guide vanes. The calculation process is as follows: First, based on the location of the region, determine the guide vane number responsible for influencing the airflow in that region; then, extract the deviation direction angle for that region; next, through a preset mapping relationship between vane angles and streamline directions, calculate the target angle that needs adjustment by back-looking up or interpolating the deviation direction angle. The magnitude and direction of this target angle represent the required airflow direction adjustment force.

[0040] The path deviation field is a spatialized dataset used to quantify the difference between the current airflow path and the ideal airflow path at various points in space. It includes the Euclidean distance (i.e., positional deviation) and deviation direction angle (i.e., the direction of deviation) between corresponding sampling points. This serves as the direct input for calculating the target angle of the guide vanes, transforming the abstract path correction requirement into a specific spatial adjustment requirement, reflecting the distribution of the gap between the current airflow organization and the ideal state. The maximum deviation value reflects the area and degree of the most severe deviation between the current airflow path and the ideal path, used to identify key locations requiring focused adjustment; the average deviation value reflects the overall severity of the deviation, used to determine the overall control pressure of the current control cycle; and the deviation direction angle reflects the specific direction of the deviation, used to determine which direction the guide vanes should be adjusted.

[0041] In this embodiment, the steps for calculating the feedforward execution time based on the composite disturbance prediction results, calculating the target angle and execution timestamp of each guide vane based on the path deviation field, and generating the feedforward execution command include: Step S401: Based on the peak time point of the current composite disturbance prediction result and the preset guide vane response delay value, calculate the feedforward execution timestamp. The feedforward execution timestamp equals the peak time point minus the guide vane response delay value. The feedforward execution timestamp is used to accurately determine the time when to issue the execution command, ensuring that actuators such as guide vanes can complete the pre-adjustment action just before the peak of the environmental disturbance arrives, thereby achieving zero-hysteresis control. The feedforward execution timestamp contains the specific execution time (absolute time or delay relative to the current time). In order for the blades to be in the correct position when the disturbance peak arrives, the execution command must be issued one guide vane response time before the peak arrives.

[0042] Step S402: Traverse each guide vane, extract the local adjustment requirements of the corresponding region from the path deviation field based on the influence space range of the guide vane, and convert the local adjustment requirements into the target angle of the guide vane through the preset mapping relationship between the blade angle and the streamline direction. By obtaining the deviation direction angle in the local adjustment requirements of the guide vane, and using the current angle of the guide vane as the initial value, query the mapping relationship table between the blade angle and the streamline direction to find a set of blade angles that make the change in airflow direction within the influence space range of the guide vane closest to the required deviation direction angle; if the mapping relationship between the blade angle and the streamline direction is in the form of a function, directly substitute the deviation direction angle to solve for the target angle; finally, output the calculated target angle.

[0043] The system determines whether the target angles of all guide vanes are within the adjustable angle range. If any guide vane's target angle exceeds the adjustable angle range, the excess adjustment requirement is converted into an auxiliary airflow regulation, generating an airflow regulation command. When the blade adjustment reaches its mechanical limit and still cannot completely eliminate the path deviation, it is necessary to further adjust the airflow pattern by changing the overall airflow kinetic energy (i.e., airflow) as a supplementary control method to ensure that the airflow path can be controlled within an acceptable range even under extreme conditions. The conversion process first calculates the excess adjustment requirement, i.e., the remaining deviation value in the local adjustment requirement not covered by the blade adjustment; then, the remaining deviation value is multiplied by a preset deviation airflow conversion coefficient to obtain the initial auxiliary airflow regulation. The deviation airflow conversion coefficient is calibrated through simulation or experiment and represents the auxiliary airflow regulation required per unit deviation; finally, following the order of "adjusting exhaust first, then supply air," the exhaust airflow is increased first. If the exhaust airflow has reached its upper limit, the supply airflow is increased to ensure overall pressure balance. The adjustable angle range is determined by the mechanical structure limitations of the guide vanes and the physical stroke setting of the drive actuator. It is generally set within the physical limit angle range of the drive actuator with a certain safety margin. For example, if the mechanical limit of the drive actuator is 0 degrees to 90 degrees, the adjustable angle range is usually set to 5 degrees to 85 degrees to avoid the vanes hitting the limit or entering the nonlinear region.

[0044] After all guide vanes are adjusted to the target angle, it is determined whether the path deviation field still exceeds the preset path deviation threshold. If so, the required auxiliary airflow adjustment is calculated. After the guide vane angle calculation is completed, a virtual adjustment evaluation is performed. First, the calculated target angles of all guide vanes are substituted into the mapping relationship between vane angle and streamline direction to simulate the adjusted airflow path. Then, the path deviation field is recalculated by comparing the simulated adjusted airflow path with the ideal airflow path. Finally, if the maximum deviation value or average deviation value in the recalculated path deviation field is still greater than the preset path deviation threshold, it is determined that the auxiliary airflow adjustment needs to be calculated; otherwise, the auxiliary airflow adjustment does not need to be calculated. The preset path deviation threshold is set according to the accuracy requirements of the airflow organization in the painting process. For example, in the painting area, the airflow is required to be strictly directed towards the exhaust vent to prevent paint mist rebound, so the path deviation threshold is set relatively small (e.g., 5 cm). In areas far from the painting area, the requirements can be appropriately relaxed, and the path deviation threshold is set relatively large (e.g., 20 cm).

[0045] The process of calculating the auxiliary airflow adjustment amount first involves obtaining the remaining path deviation field after the virtual adjustment, and extracting the average or maximum deviation value of this deviation field as the remaining deviation amount. Next, according to the preset deviation airflow conversion relationship, the remaining deviation amount is mapped to the required total airflow increment. The deviation airflow conversion relationship is usually non-linear and is determined by fluid mechanics principles. Finally, based on the current fan operating status, the specific adjustment amounts of the supply fan and exhaust fan are calculated to ensure that the airflow increment is within the allowable range of the equipment. The calculation is based on the fact that the larger the airflow, the stronger the airflow inertia, and the more obvious the path correction effect.

[0046] The airflow adjustment command includes the target speed (or frequency) of the exhaust fan and the target speed (or frequency) of the supply fan. It is used to assist in path correction by changing the overall airflow kinetic energy when the guide vane adjustment range is insufficient. When the airflow adjustment command is generated, the increase or decrease in total airflow is first determined based on the calculated auxiliary airflow adjustment amount. Then, following the principle of "prioritizing exhaust adjustment," the required increase in airflow for the exhaust fan is calculated first and converted into the exhaust fan's target speed. If the total airflow adjustment demand is still not met, the adjustment amount for the supply fan is calculated and converted into the supply fan's target speed. Finally, the target speeds of the exhaust fan and the supply fan are encapsulated into an airflow adjustment command.

[0047] The calculation of the influence space range of the guide vane involves first obtaining its installation coordinates, blade size, and maximum adjustable angle; then, based on fluid dynamics principles or pre-performed computational fluid dynamics simulations, simulating the influence area of ​​the vane on the downstream airflow direction at different angles; finally, the influence areas at all angles are combined, and the minimum bounding box or spatial boundary of this combination is extracted as the influence space range of the guide vane. The influence space range of the guide vane refers to the airflow area that a single guide vane can effectively control within the spray booth, represented by a spatial coordinate range (such as a cubic region or a curved surface region).

[0048] The process of extracting the local adjustment requirements for a corresponding region involves first traversing every sampling point in the path deviation field; then, determining whether the location of the sampling point falls within the influence space of a certain guide vane; if it does, extracting the deviation direction angle and Euclidean distance value of that point; finally, summarizing the deviation direction angles and deviation values ​​of all sampling points within the influence space of that guide vane (e.g., taking the average or maximum value) to form the local adjustment requirements for that vane. Local adjustment requirements refer to the amount and intensity of airflow direction adjustment required within a specific spatial region to bring the current airflow path closer to the ideal airflow path. This includes the location coordinates of the region, the required adjustment direction (deviation direction angle), and the urgency of the adjustment (e.g., the magnitude of the deviation value). It is based on the deviation data of the corresponding region in the path deviation field, decomposing the global path deviation field into local control tasks that each guide vane needs to independently undertake.

[0049] The mapping relationship between blade angle and streamline direction is a pre-generated lookup table or function that records how a certain combination of angles of a specific guide vane will change the airflow direction at a certain location within its influence space. This includes blade angle input (such as blade angle 1, blade angle 2) and corresponding airflow direction output (such as velocity increment in the X direction, velocity increment in the Y direction). It is established through a large amount of offline computational fluid dynamics simulation or wind tunnel experimental data. As the core converter of "control quantity calculation", it transforms the abstract local adjustment requirement of "needing to adjust the airflow to a certain direction" into a specific, executable execution instruction of "how many degrees the blade should rotate".

[0050] Step S403 encapsulates the target angle of each guide vane, the airflow adjustment command, and the feedforward execution timestamp into a feedforward execution command. The feedforward execution command is a comprehensive control command issued by the control system to the lower-level actuator to proactively respond to future disturbances. It includes the target angle of each guide vane (i.e., the specific angle value to which each vane should rotate), the airflow adjustment command (i.e., the target speed of the exhaust fan and the supply fan), and the feedforward execution timestamp (i.e., the specific time when the command should take effect). The results of the feedforward prediction and control quantity calculation are converted into executable physical commands, and the execution timing is precisely controlled to ensure that the actuator has completed its action in advance when the disturbance peak arrives, thereby achieving proactive pre-adjustment of environmental parameters.

[0051] In this embodiment, the implementation steps of issuing feedforward execution instructions according to the execution timestamp and obtaining execution status feedback and post-execution environmental feedback data include: Step S501: Upon reaching the feedforward execution timestamp, the target angle of each guide vane is sent to the corresponding drive actuator of the guide vane. The target angle of the guide vane is an absolute angle based on the mechanical zero position (i.e., the initial installation position) of the vane. For example, assuming that the mechanical zero position of a guide vane is defined as the position where the vane is parallel to the airflow direction (0 degrees), when the target angle calculated by the system is 30 degrees, it means that the vane needs to rotate 30 degrees from the zero position in a preset direction (such as clockwise) to form a 30-degree angle with the airflow direction, thereby guiding the airflow in the predetermined direction.

[0052] The system receives execution status feedback from each drive actuator, including execution success status and actual angle of arrival. This feedback monitors the physical execution of feedforward commands, including execution success status (e.g., whether the command was received, whether the actuator is fault-free) and actual angle of arrival (i.e., the actual angle at which the blade finally stops). This allows the system to distinguish between "command not executed" and "command executed but with poor results," preventing actuator malfunctions from being misdiagnosed as prediction model bias.

[0053] Step S502: After the feedforward execution command is issued, continuously collect environmental parameters at M measurement points near the operation area, exhaust vent, and workpiece surface, where M is an integer, as environmental feedback data after execution.

[0054] The environmental feedback data after execution, the execution status feedback, and the feedforward execution command for this operation are stored in the execution log. The execution log records complete execution information for each feedforward control action, providing historical data support for subsequent parameter optimization, fault tracing, and prediction model updates. It includes the feedforward execution command (i.e., what command was issued), the execution status feedback (i.e., how the actuator actually responded), and the environmental feedback data after execution (i.e., the actual environmental effect after regulation). By storing these three types of data in a related manner, the system can analyze the causal relationship between commands, responses, and effects afterward, providing analytical samples for parameter optimization and model updates in step S700.

[0055] In this embodiment, the implementation steps of comparing environmental feedback data with composite disturbance prediction results, calculating deviation signals, generating feedback correction instructions based on deviation signals, and issuing them for execution include: Step S601 involves continuously collecting environmental feedback data at set intervals (e.g., 200ms) within a time period following the occurrence of the process action. This continuously collected environmental feedback data provides a continuous time series for point-by-point comparison with the feedforward prediction results, thereby accurately calculating the control deviation at each moment. Compared to the environmental feedback data in step S502, which represents an initial state used to determine whether the environment begins to change in the predicted direction after the feedforward action, the continuously collected environmental feedback data in step S601 records the entire process and is used to evaluate the sustained accuracy and dynamic response of the control throughout the entire disturbance cycle.

[0056] The collected environmental feedback data at each time point is compared point-by-point with the predicted values ​​at the corresponding time points in the composite disturbance prediction results, and the deviation value at each time point is calculated. The predicted value at the corresponding time point in the composite disturbance prediction results refers to the predicted environmental parameter value (such as VOCs concentration) that should be reached at a specific time. It is obtained directly by reading the composite disturbance prediction curve generated in step S203 according to the timestamp. Point-by-point comparison includes VOCs concentration value comparison, pressure value comparison, and temperature and humidity value comparison. The point-by-point comparison process first aligns the time axis to ensure that the time points of the actual collected data correspond one-to-one with the time points of the prediction curve; then, at each time point, the actual VOCs concentration and the predicted VOCs concentration are read respectively, and the difference between the two is calculated to obtain the VOCs concentration deviation value; finally, the above steps are repeated for the pressure value and temperature and humidity value to obtain the pressure deviation value and temperature and humidity deviation value at each time point. The deviation values ​​at each time point are used to quantify the difference between the actual environment and the predicted environment at each moment, providing accurate, time-varying input for the calculation of feedback corrections. It can perform differentiated corrections for errors at different stages. The deviation values ​​at each time point include the VOCs concentration deviation, pressure deviation, and temperature and humidity deviation values ​​obtained after point-by-point comparison.

[0057] The feedback correction is calculated based on the deviation values ​​at each time point. This feedback correction includes fine-tuning of the guide vane angle and airflow. The feedback correction is an incremental command calculated based on the real-time deviation signal, used for fine-tuning the actuator. It compensates in real-time for residual errors that feedforward control could not completely eliminate, improving control accuracy. The feedback correction is based on a preset mapping relationship between the deviation value and the correction amount, typically a proportional control logic, meaning the correction is proportional to the deviation value. The deviation values ​​at each time point are converted into fine-tuning amounts for the vane angle and airflow, respectively.

[0058] In step S602, the feedback correction amount is superimposed with the current angle and current airflow of the guide vanes to generate a real-time correction command, which is immediately sent to the corresponding drive actuator for execution. The real-time correction command is a dynamic command used to adjust the actuator state in real time, and to compensate for the residual error after feedforward control, thereby achieving fine-grained closed-loop regulation. After the real-time correction command is issued and executed, the process from step S502 to step S602 needs to be executed repeatedly. As long as the process action sequence is still in progress, the disturbance is still ongoing, or the environmental parameters have not yet stabilized within the preset allowable range, environmental feedback data needs to be continuously collected and feedback corrections need to be continuously performed to form a dynamic closed-loop control cycle. When the time window covered by the composite disturbance prediction result has passed, and the continuously collected environmental feedback data shows that the VOCs concentration, pressure, temperature, and humidity of all key measuring points have stabilized within the preset process allowable range, and no new process actions are triggered, the control cycle is determined to end, and the process enters the parameter update stage in step S700.

[0059] In this embodiment, the implementation steps for updating the calculation parameters of the motion perturbation mapping library and the feedforward execution time based on the deviation signal and execution status feedback include: Step S701 involves statistically analyzing the deviation signal to determine its type. Deviation types include systematic positive offset, systematic negative offset, and random fluctuations. This statistical analysis helps diagnose the root cause of the deviation, allowing for targeted parameter update strategies and preventing a one-size-fits-all approach that could lead to prediction divergence. Different types of deviations reflect different levels of problems: systematic positive offset indicates that the actual value is consistently higher than the predicted value, suggesting the prediction model underestimates the disturbance intensity; systematic negative offset indicates that the actual value is consistently lower than the predicted value, suggesting the prediction model overestimates the disturbance intensity; random fluctuations indicate that the deviation value alternates between positive and negative values ​​irregularly, suggesting the presence of unmodeled random disturbances or inconsistent actuator responses.

[0060] When the deviation type is a systematic positive or negative offset, it is judged as a prediction model deviation. The update weights of the disturbance mode parameters for the corresponding process action in the action disturbance mapping library are adjusted, increasing the weight of the actual disturbance mode in the weighted average update. When the deviation signal exhibits a systematic offset, it indicates that the disturbance mode in the current action disturbance mapping library is no longer accurate, and the parameters in the library need to be updated more significantly using the actual disturbance data. The adjustment process is as follows: First, based on the deviation signal analysis results, if it is determined to be a systematic positive or negative offset, a higher update weight coefficient is set (e.g., 0.7, indicating that the actual disturbance mode accounts for 70% of the updated parameter weight). Then, based on the actual environmental feedback data collected, the actual disturbance mode (i.e., the amplitude matrix and time history vector actually generated by the action) is extracted. Next, the original disturbance mode parameters in the action disturbance mapping library and the actual disturbance mode are weighted and averaged according to the set weights to calculate the new disturbance mode parameters. Finally, the new parameters replace the old parameters in the action disturbance mapping library. The prediction model refers to the entire computational framework used in step S200 to generate the composite disturbance prediction results. Its core is the parameters of the disturbance patterns stored in the action disturbance mapping library.

[0061] When the deviation type is random fluctuation and the deviation amplitude exceeds the preset amplitude threshold, it is determined that the actuator response is insufficient or there is external interference. Based on the deviation between the actual arrival angle and the target angle in the execution status feedback, the preset guide vane response delay value is adjusted. When the deviation signal exhibits random fluctuation and the amplitude exceeds the preset amplitude threshold, combined with the deviation between the actual arrival angle and the target angle in the execution status feedback, it is determined that the actuator's physical response speed does not match the preset value. The adjustment process first extracts the deviation value between the actual arrival angle and the target angle and the actual time taken to reach the target angle from the execution status feedback; then, the difference between the actual response time and the preset response delay value is calculated; next, a first-order filter or moving average is used to smooth the actual response time and historical response time to obtain the corrected response delay value; finally, the corrected value is updated in the control parameters for subsequent feedforward execution timestamp calculation.

[0062] The adjusted disturbance mode parameters and guide vane response delay values ​​are applied to subsequent control cycles.

[0063] Step S702 involves statistically analyzing the control performance indicators in the historical execution log at a set period. These indicators quantify the system's control effectiveness over long-term operation, guiding self-optimization and parameter adjustments. Control performance indicators include safety indicators, quality indicators, and energy consumption indicators. Safety indicators include the duration of VOC concentration exceeding limits and peak concentration, reflecting the system's ability to control the risk of VOC concentration exceeding limits. The duration of VOC concentration exceeding limits is calculated in seconds, representing the total time (in seconds) that the VOC concentration at all measuring points exceeds the safety threshold within a statistical period. The longer the duration, the higher the system's safety risk, and this is used to assess the system's safety protection capabilities. Peak concentration is obtained by calculating the maximum VOC concentration at all measuring points within the statistical period, reflecting the control effectiveness under extreme operating conditions. Quality indicators include the coating quality pass rate and the process window deviation time, reflecting the degree to which control measures ensure the stability of the coating process. The coating quality pass rate is obtained by dividing the number of qualified coated workpieces by the total number of coated workpieces within the statistical period, and directly reflects the stability of the coating process. The process window deviation time is obtained by the total time during which temperature and humidity exceed the process setting range within the statistical period, and temperature and humidity deviations directly affect the paint film curing quality. Energy consumption indicators include the cumulative energy consumption of the fan and the cumulative number of guide vane actions, reflecting the economic efficiency of system operation. The cumulative energy consumption of the fan is obtained by the cumulative power consumption of the supply fan and exhaust fan within the statistical period (which can be calculated through power integration), and the cumulative energy consumption of the fan directly reflects the operating cost. The cumulative number of guide vane actions is obtained by the total number of times the execution angle of all guide vanes is adjusted within the statistical period, and the number of actions reflects the wear and energy consumption of the actuator.

[0064] The control performance indicators are compared with preset target thresholds. Parameter optimization is initiated when any indicator falls below the target threshold. A weighted summation method is used to calculate a comprehensive performance score based on a set periodicity. The weights are dynamically adjusted according to the current production mode. If the comprehensive performance score falls below a preset score threshold, parameter optimization is triggered. The triggering conditions for parameter optimization employ a dual mechanism: periodic triggering and event triggering. Periodic triggering calculates a comprehensive performance score by weighting and summing the control performance indicators at a set periodicity (e.g., daily, weekly). Parameter optimization is triggered when this score falls below a preset score threshold. Event triggering involves continuous monitoring of individual indicators within a period. Parameter optimization is immediately triggered when the VOCs concentration in the safety indicator exceeds the limit for an extended period or the peak concentration exceeds the preset safety target threshold, or when the coating quality pass rate in the quality indicator falls below the preset quality target threshold, without waiting for periodic evaluation. This dual mechanism ensures both systematic optimization and timely response to urgent safety or quality issues.

[0065] During parameter optimization, each ideal airflow path in the pre-set ideal path library is traversed, and the overall performance score of each ideal airflow path under the current operating conditions is evaluated. The ideal airflow path with the highest overall performance is selected as the ideal airflow path for subsequent control cycles. The mapping relationship between blade angle and streamline direction is traversed to evaluate whether there are better mapping parameters. The optimized ideal airflow path and mapping parameters are gradually applied to the ideal airflow path determination step of subsequent control cycles using a soft update method.

[0066] The process of evaluating the comprehensive performance score of an ideal airflow path under current operating conditions includes: First, acquiring the characteristics of the current operating conditions, including the predicted peak VOC concentration, the current workpiece type, and the production mode (safety / quality / energy consumption priority); then, for each candidate path in the ideal path library, predicting the safety, quality, and energy consumption indicators that the path can achieve under the current operating conditions using a pre-established path performance mapping model. The path performance mapping model is established based on historical operating data or simulation data. Next, according to the control priority weights determined under the current operating conditions, the predicted values ​​of the three indicators are weighted and summed. Finally, the comprehensive performance score of the path under the current operating conditions is obtained.

[0067] The assessment examines whether better mapping parameters exist. Based on the current mapping parameters (i.e., the mapping relationship between blade angle and streamline direction), the target angle calculated after actual execution exhibits a systematic deviation from the expected path correction effect. The assessment process involves the following steps: First, P sets of data pairs (P being an integer) showing the changes in blade target angle and actual airflow direction are selected from the execution log. Then, the actual airflow direction change is compared with the direction change predicted by the current mapping parameters, and the prediction error is calculated. Next, if the prediction error shows a systematic trend (e.g., consistently large or small), it indicates a calibration bias in the mapping parameters. Finally, using least-squares fitting or online learning algorithms, the mapping parameters are recalibrated using historical data to generate a new set of mapping parameters, and the effectiveness of the new parameters in reducing the prediction error is evaluated. If the new parameters significantly reduce the prediction error, a better mapping parameter is determined to exist.

[0068] In this embodiment, during actual use, the motion disturbance mapping library in step S200, the preset coating feature library in step S102, and the preset ideal path library in step S302 are all static or semi-static preset knowledge bases, which can be stored and managed using relational databases (such as SQLite, MySQL) or real-time historical databases (such as InfluxDB, PISystem). Three data tables are established in the database. In the motion disturbance mapping library table, "motion type" is used as the primary key, storing the influence amplitude matrix as a JSON array, the influence time history vector as a time field (delay time, duration), and the influence spatial vector as a spatial coordinate range string. In the coating feature library table, "coating type label" is used as the primary key, storing the baseline evaporation rate value of the coating, and parameterizing the evaporation characteristic curve function, such as storing the initial value and decay coefficient of the exponential decay function. In the ideal path library table, "control mode" is used as the primary key, storing the spatial coordinate point sequence of the ideal airflow path as a binary flow or JSON coordinate array. When the system is running, the control software reads this data through standard database query interfaces (such as ODBC and JDBC) and loads it into memory for quick access. The mapping relationship between blade angle and streamline direction in step S303, the preset deviation air volume conversion coefficient in step S402, the preset deviation air volume conversion relationship, and the preset deviation value and correction amount mapping relationship in step S601 are all conversion rules required for online rapid calculation. They can be implemented using function block libraries in industrial automation configuration software or mathematical function modules written in high-level languages ​​(C++, Python).

[0069] For the mapping relationship between blade angle and streamline direction in step S303, a multi-dimensional look-up table (LUT) can be constructed. In the software, this table is stored as a two-dimensional array. The row index represents the combination of blade angles (or the angle range of multiple blades), the column index represents different position points within the affected spatial range, and the array elements are the changes in airflow direction (X, Y, Z direction components) at that position point. When it is necessary to convert the local adjustment requirement into a target angle, the software module executes the following logic: First, using the current blade angle as input, the closest row is located in the look-up table; then, the matching degree between the deviation direction angle in the local adjustment requirement and the changes in airflow direction in each column of that row is calculated; finally, the blade angle corresponding to the row with the highest matching degree is selected as the target angle output; if the mapping relationship needs to be continuous and differentiable, a linear interpolation function or radial basis function fitting is used to fit the discrete look-up table into a continuous mathematical function, and the target angle is directly output from the input deviation direction angle.

[0070] To implement the preset mapping relationship between the deviation value and the correction amount in step S601, proportional control logic is used. In the control software, a proportional controller module is written, whose input is the real-time deviation value and output is the correction amount. By setting a proportional coefficient, which can be set empirically or self-tuned online, the output correction amount = proportional coefficient × input deviation value. To avoid high-frequency oscillation, the software module also needs to add dead zone setting (when the absolute value of the deviation is less than the dead zone threshold, the correction amount output is zero) and output limiting (the correction amount does not exceed the preset maximum fine-tuning step size). For the preset deviation airflow conversion coefficient and preset deviation airflow conversion relationship in step S402, a piecewise linear function or rule-based reasoning module is used. In the software, this relationship is encapsulated as an independent function block, with the input being the remaining path deviation value (e.g., average deviation distance) and the output being the auxiliary airflow adjustment amount. By preseting a set of deviation threshold intervals, each interval corresponding to an airflow adjustment coefficient, the function block first determines which interval the input deviation value falls into, and then calculates the auxiliary airflow adjustment amount by multiplying the coefficient of that interval by the deviation value (or by directly looking up a table). If a continuous function form is used, an exponential or logarithmic function can be used to fit the physical relationship between deviation and airflow; the function block substitutes the deviation value into the function for calculation.

[0071] The path performance mapping model in step S702 is used to evaluate the comprehensive performance of different ideal airflow paths under the current operating conditions. It is a predictive analysis model and can be trained offline using machine learning frameworks (such as Scikit-learn, TensorFlowLite) or statistical regression analysis software. The trained model can then be exported to a format that can run on the control platform (such as ONNX, PMML). In practical use, firstly, a large amount of historical operating data is collected on an offline simulation platform or historical data analysis platform. This includes all historical operating conditions constituted by the data collected in step S100, the types of ideal airflow paths used, and the actual control performance indicators generated. Then, this data is used to train a multi-output regression model, such as random forest regression or neural network regression. The input is the operating condition features and candidate path labels, and the output is the predicted values ​​of the three performance indicators. After the model is trained, it is deployed to the parameter update module of the real-time control software. When it is necessary to evaluate the comprehensive performance score of an ideal airflow path under the current operating conditions, the system's parameter update module inputs the current operating condition characteristics and the path label into the model. The model outputs the predicted safety index, quality index and energy consumption index values, and then performs a weighted sum according to the current control priority weight to obtain the comprehensive performance score. The model can be updated by replacing the model file after offline retraining.

[0072] like Figure 3 As shown, the spray booth multi-source data environment intelligent control system includes: The data acquisition module is used to acquire the process sequence of the spraying production line, the dynamic data of paint volatilization at the spray gun outlet, the airflow field status data and environmental feedback data in the spray booth; The disturbance prediction module is used to predict the single-action disturbance curves of each process action on the environmental parameters inside the spray booth based on the process action sequence and a pre-built action disturbance mapping library. After the amplitude of the single-action disturbance curves is corrected based on the coating volatilization dynamic data, they are superimposed on the time axis to generate composite disturbance prediction results. The path planning module is used to determine the current airflow path based on airflow field state data, determine the ideal airflow path based on composite disturbance prediction results, and calculate the path deviation field between the current airflow path and the ideal airflow path. The control quantity calculation module is used to calculate the feedforward execution time based on the composite disturbance prediction results, calculate the target angle and execution timestamp of each guide vane based on the path deviation field, and generate the feedforward execution command. The execution control module is used to issue feedforward execution instructions according to the execution timestamp, and to obtain execution status feedback and post-execution environmental feedback data; The feedback correction module is used to compare environmental feedback data with composite disturbance prediction results, calculate deviation signals, generate feedback correction instructions based on deviation signals, and issue them for execution. The parameter update module is used to update the calculation parameters of the motion disturbance mapping library and feedforward execution time based on the deviation signal and execution status feedback.

[0073] In this embodiment, by acquiring the process action sequence of the spraying production line control system, the impact curves on VOCs concentration, pressure, temperature, and humidity are predicted before disturbance source actions such as spray gun start-up, color change cleaning, and workpiece door opening and closing occur. By reading the disturbance mode of each action from the action disturbance mapping library and combining it with the instantaneous evaporation rate of the paint monitored in real time at the spray gun outlet, the disturbance amplitude is dynamically corrected to generate a composite disturbance prediction result. Based on the composite disturbance prediction result, the feedforward execution timestamp is calculated and the guide vane angle command is issued before the disturbance arrives, so that the actuator has completed the pre-adjustment action when the disturbance peak arrives. This greatly reduces the response delay of environmental parameters to the millisecond level, basically achieving zero lag, while reducing the probability of VOCs concentration exceeding the limit and the spraying quality defect rate. Compared with the lag adjustment that relies on sensor feedback in the prior art, this achieves a fundamental shift from post-compensation to pre-adjustment, improving the safety and process stability of the spraying booth. This invention abandons the indirect control mode of existing technologies that rely solely on fan airflow adjustment. Instead, it uses adjustable guide vanes as the main actuator to directly manipulate the spatial shape of the airflow path. By reconstructing the airflow velocity vector field in the spray booth in real time, it generates the current airflow path and compares it with the ideal airflow path dynamically determined based on the operating conditions. The path deviation field calculated by the comparison is converted into the target angle of each guide vane. The airflow path is corrected primarily through low-power blade angle adjustment, and the total airflow is only adjusted as an auxiliary measure when the blade adjustment range is insufficient. This significantly increases the time proportion of the exhaust fan's high-efficiency zone while greatly reducing overall energy consumption. Furthermore, because it directly controls the airflow direction rather than indirectly adjusting the airflow, it improves the suppression accuracy of problems such as local VOCs accumulation and airflow short-circuiting, achieving the dual goals of refined active control of airflow organization and energy saving.

[0074] This application also provides an electronic device. The electronic device may include one or more processors and one or more memories. The memories store computer-readable code, which, when executed by the one or more processors, can perform the intelligent control method and system for multi-source data environment of the spray booth as described above.

[0075] The methods and systems according to the embodiments of this application can also be implemented using the architecture of the electronic device shown in this application. The electronic device may include a bus, one or more CPUs, ROM, RAM, a communication port connected to a network, input / output, a hard disk, etc. The storage device in the electronic device, such as a ROM or hard disk, may store the intelligent control method and system for multi-source data environment of the spray booth provided in this application. Furthermore, the electronic device may also include a user interface. Of course, the architecture shown in this application is merely exemplary; when implementing different devices, one or more components of the electronic device shown in this application may be omitted according to actual needs.

[0076] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising a reference structure" does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes the element.

[0077] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for intelligent control of multi-source data environment in a spray booth, characterized in that, The method includes: Acquire the process sequence of the spraying production line, the dynamic data of paint volatilization at the spray gun outlet, the airflow field data and environmental feedback data in the spray booth; Based on the process sequence and the pre-built motion disturbance mapping library, the single motion disturbance curves of each process motion on the environmental parameters inside the spray booth are predicted. The amplitude of the single motion disturbance curves is corrected based on the dynamic data of paint volatilization and then superimposed on the time axis to generate the composite disturbance prediction result. The current airflow path is determined based on the airflow field state data, the ideal airflow path is determined based on the composite disturbance prediction results, and the path deviation field between the current airflow path and the ideal airflow path is calculated. The feedforward execution time is calculated based on the composite disturbance prediction results. The target angle and execution timestamp of each guide vane are calculated based on the path deviation field, and the feedforward execution command is generated. The feedforward execution instructions are issued according to the execution timestamp, and the execution status feedback and post-execution environmental feedback data are obtained. The environmental feedback data is compared with the composite disturbance prediction results to calculate the deviation signal. Based on the deviation signal, a feedback correction command is generated and issued for execution. Based on the deviation signal and execution status feedback, update the motion disturbance mapping library and the calculation parameters of the feedforward execution time.

2. The intelligent control method for multi-source data environment of a spray booth according to claim 1, characterized in that, The acquisition of the process sequence of the spraying production line, dynamic data of paint evaporation at the spray gun outlet, airflow field data and environmental feedback data in the spray booth includes: The process action instructions in the spraying production line controller are read at a set sampling period. The read process action instructions are sorted by time to generate a process action sequence within a future time window. The process action sequence includes action type, trigger time and action parameters. The instantaneous evaporation rate of solvent in paint mist is collected at the spray gun outlet at a set sampling frequency. The instantaneous evaporation rate is then processed by sliding window filtering and matched with a preset paint feature library to obtain dynamic data of paint evaporation. Collect wind speed vectors and static pressure values ​​at N measuring points inside the spray booth, and calculate the airflow velocity vector field and pressure field of the continuous space based on the wind speed vectors and static pressure values ​​as airflow field state data; Environmental parameters were collected from M measuring points in the spray booth's operating area, exhaust vents, and near the workpiece surface as environmental feedback data. The environmental feedback data included temperature data, humidity data, and VOCs concentration.

3. The intelligent control method for multi-source data environment of a spray booth according to claim 2, characterized in that, Based on the process sequence and a pre-built motion disturbance mapping library, the system predicts the single-motion disturbance curves of each process motion on the environmental parameters inside the spray booth. After amplitude correction of the single-motion disturbance curves based on paint volatilization dynamic data, the curves are superimposed along the time axis to generate a composite disturbance prediction result, including: Traverse each process action within the future time window and read the perturbation mode corresponding to the process action from the action perturbation mapping library. The perturbation mode includes an influence amplitude matrix, an influence time history vector, and an influence space vector. The volatilization correction coefficient is calculated based on the dynamic data of coating volatilization, and the influence amplitude matrix is ​​corrected based on the volatilization correction coefficient; Based on the trigger time and influence time history vector of each process action, a single-action disturbance curve is generated; The single-action disturbance curves of all process actions are aligned and superimposed along the time axis to generate a composite disturbance prediction curve, and the peak time point and peak amplitude of the composite disturbance prediction curve are extracted.

4. The intelligent control method for multi-source data environment of a spray booth according to claim 3, characterized in that, The process of determining the current airflow path based on the airflow field state data, determining the ideal airflow path based on the composite disturbance prediction results, and calculating the path deviation field between the current airflow path and the ideal airflow path includes: The airflow velocity vector field and pressure field at the current moment are obtained from the airflow field state data. Streamline tracing is performed along the velocity vector direction starting from the center point of the spray painting area to generate a sequence of spatial coordinate points of the current airflow path. The control priority is determined based on the peak amplitude in the composite disturbance prediction result and the current operating condition. An ideal airflow path matching the control priority is selected from the preset ideal path library. The ideal airflow path is a sequence of spatial coordinate points that starts from the paint spraying area and ends at the exhaust vent. The current airflow path and the ideal airflow path are discretized into the same number of spatial sampling points. The Euclidean distance between the corresponding sampling points is calculated to generate a path deviation field. The maximum deviation value, average deviation value and deviation direction angle of the path deviation field are then statistically analyzed.

5. The intelligent control method for multi-source data environment of a spray booth according to claim 1, characterized in that, The step of calculating the feedforward execution time based on the composite disturbance prediction result, calculating the target angle and execution timestamp of each guide vane based on the path deviation field, and generating feedforward execution instructions includes: Based on the peak time point of the current composite disturbance prediction result and the preset guide vane response delay value, calculate the feedforward execution timestamp; Each guide vane is traversed, and the local adjustment requirements of the corresponding region are extracted from the path deviation field according to the influence space range of the guide vane. The local adjustment requirements are converted into the target angle of the guide vane through the preset mapping relationship between the blade angle and the streamline direction. Determine whether the target angle of all guide vanes is within the adjustable angle range. If the target angle of any guide vane exceeds the adjustable angle range, convert the adjustment requirement of the excess part into an auxiliary air volume adjustment amount and generate an air volume adjustment command. The target angle and feedforward execution timestamp of each guide vane are encapsulated into feedforward execution instructions.

6. The intelligent control method for multi-source data environment of a spray booth according to claim 5, characterized in that, The step of issuing feedforward execution instructions according to the execution timestamp and obtaining execution status feedback and post-execution environmental feedback data includes: When the feedforward execution timestamp is reached, the target angle of each guide vane is sent to the corresponding guide vane drive actuator; Receive execution status feedback returned by each drive actuator, the execution status feedback including execution success status and actual angle reached; After the feedforward execution command is issued, environmental parameters of M measuring points near the operation area, exhaust vent and workpiece surface are continuously collected as environmental feedback data after execution. The environmental feedback data, execution status feedback, and feedforward execution instructions after execution are stored in the execution log.

7. The intelligent control method for multi-source data environment of a spray booth according to claim 6, characterized in that, The step of comparing environmental feedback data with composite disturbance prediction results, calculating deviation signals, generating feedback correction instructions based on deviation signals, and issuing them for execution includes: During the time period following the occurrence of the process action, environmental feedback data is continuously collected at a set cycle. The collected environmental feedback data at each time point is compared with the predicted values ​​at the corresponding time points in the composite disturbance prediction results, and the deviation value at each time point is calculated. The feedback correction amount is calculated based on the deviation value at each time point, and the feedback correction amount includes the guide vane angle fine adjustment amount and the air volume fine adjustment amount; The feedback correction amount is superimposed with the current angle and current airflow of the current guide vane to generate a real-time correction command, which is immediately sent to the corresponding drive actuator for execution.

8. The intelligent control method for multi-source data environment of a spray booth according to claim 7, characterized in that, The step of updating the motion disturbance mapping library and the calculation parameters of the feedforward execution time based on the deviation signal and execution status feedback includes: Statistical analysis is performed on the deviation signal to determine the type of deviation, which includes systematic positive offset, systematic negative offset, and random fluctuation. When the deviation type is a systematic positive offset or a systematic negative offset, adjust the update weight of the disturbance mode parameter of the corresponding process action in the action disturbance mapping library. When the deviation type is random fluctuation and the deviation amplitude exceeds the preset amplitude threshold, the preset guide vane response delay value is adjusted according to the deviation between the actual arrival angle and the target angle in the execution status feedback.

9. The intelligent control method for multi-source data environment of a spray booth according to claim 8, characterized in that, The method further includes: The control performance indicators in the historical execution log are statistically analyzed at a set period, and the control performance indicators include safety indicators, quality indicators and energy consumption indicators. The control performance index is compared with a preset target threshold, and parameter optimization is performed when any index is lower than the target threshold.

10. A multi-source data environment intelligent control system for spray booths, characterized in that: The system includes: The data acquisition module is used to acquire the process sequence of the spraying production line, the dynamic data of paint volatilization at the spray gun outlet, the airflow field status data and environmental feedback data in the spray booth; The disturbance prediction module is used to predict the single-action disturbance curve of each process action based on the process action sequence and a pre-built action disturbance mapping library, and to generate composite disturbance prediction results by combining the coating volatilization dynamic data. The path planning module is used to determine the current airflow path based on airflow field state data, determine the ideal airflow path based on composite disturbance prediction results, and calculate the path deviation field between the current airflow path and the ideal airflow path. The control quantity calculation module is used to calculate the feedforward execution time based on the composite disturbance prediction results and generate feedforward execution instructions based on the path deviation field. The execution control module is used to issue feedforward execution instructions according to the execution timestamp, and to obtain execution status feedback and post-execution environmental feedback data; The feedback correction module is used to compare environmental feedback data with composite disturbance prediction results, calculate deviation signals, generate feedback correction instructions based on deviation signals, and execute them. The parameter update module is used to update the calculation parameters of the motion disturbance mapping library and feedforward execution time based on the deviation signal and execution status feedback.

Images