Spraying equipment automatic control system based on industrial internet

Abstract

The invention discloses a spraying equipment automatic control system based on the industrial internet, and belongs to the technical field of industrial automatic control. The system comprises a state sensing unit, a feature extraction unit, a risk prediction unit and a closed loop correction unit. The state sensing unit obtains a real-time working condition data set; the feature extraction unit generates a core process feature set; the risk prediction unit generates a quality risk prediction set; the closed-loop correction unit calculates a target process set point. The system senses multi-source data such as environment temperature and humidity and coating flow in real time, quantifies the influence of environment and process fluctuation on spraying quality, predicts indexes such as coating film thickness and defect probability in a prospective mode, and adjusts process parameters in a self-adaptive mode on the basis of working condition risk levels. Compared with the prior art, complex working condition insight, quality risk look-ahead prediction, dynamic intelligent optimization and extreme working condition robust control are achieved, the contradiction between high quality and high efficiency in automatic spraying is relieved, and the spraying quality stability and the production efficiency are improved.

Description

Technical Field

[0001] This invention relates to the field of industrial automation control technology, and in particular to an automated control system for spraying equipment based on the Industrial Internet. Background Art

[0002] Automated spraying is a key step in modern discrete manufacturing to improve product appearance quality and production efficiency. Currently, industrial robots combined with automated spray guns are widely used in various spraying production lines. In a typical automated spraying system, a fixed set of process parameters is usually pre-set, such as paint flow rate, atomizing air pressure, and the robot's movement speed and trajectory, in order to achieve consistent coating quality under stable production conditions.

[0003] However, in actual industrial production environments, production conditions are difficult to maintain absolute stability. Especially in large-scale, continuous production operations, the following two core issues are particularly prominent, severely restricting the stability of coating quality and further improvements in production efficiency: First, the dynamic fluctuations of environmental factors have a difficult-to-control impact on the physical properties of coatings. The ambient temperature and humidity in the spraying workshop fluctuate continuously due to seasonal changes, diurnal temperature variations, and the start and stop of ventilation systems. These fluctuations directly affect the physical properties of coatings, especially their core indicator—coating viscosity. For example, increased temperature leads to a significant decrease in coating viscosity and increased fluidity; changes in humidity affect the evaporation rate of solvents in the coating.

[0004] Secondly, fixed process parameters cannot adapt to real-time changes in coating properties and atomization state. When coating viscosity decreases due to increased temperature, continuing to use the existing high coating flow rate to pursue production efficiency can easily lead to excessively thick coatings and sagging defects. Conversely, if the flow rate is drastically reduced to avoid sagging, production efficiency will be affected. Simultaneously, even slight fluctuations in atomizing air pressure can couple with changing coating viscosity, worsening the atomization state and causing defects such as orange peel and pinholes on the workpiece surface. Existing control systems lack an effective mechanism to perceive these combined effects of environmental and process fluctuations in real time and proactively and collaboratively adjust key process parameters such as coating flow rate and atomization pressure to continuously seek the optimal balance between quality and efficiency in a dynamically changing environment.

[0005] Therefore, how to develop a system capable of sensing dynamic changes in the environment and process in real time, accurately predicting their comprehensive impact on coating quality, and dynamically optimizing spraying process parameters based on this to resolve the inherent contradiction between high quality and high efficiency in automated spraying is a technical problem urgently needing to be solved in this field. (Summary of the Invention) The purpose of this invention is to provide an automated control system for spraying equipment based on the Industrial Internet, thereby solving the aforementioned technical problems.

[0006] The purpose of the present invention can be achieved through the following technical solutions: An automated control system for spraying equipment based on the Industrial Internet, characterized in that it includes: A state perception unit is used to acquire a real-time operating condition dataset, which includes: environmental state data characterizing the current spraying environment state, and current process measurement values ​​characterizing the current operating state of the spraying equipment; wherein, the current process measurement values ​​are the values ​​of process parameters measured in real time by sensors within the current control cycle. The feature extraction unit generates a core process feature set based on the real-time working condition dataset and a preset physical model. The core process feature set is used to quantify the comprehensive impact of the real-time working condition dataset on the core state of the spraying process. The risk prediction unit generates a quality risk prediction set based on the core process feature set and the current process measurement value. The quality risk prediction set includes: a predicted quality index for the final coating quality, a predicted defect probability for the occurrence of spraying defects, and a working condition risk level used to characterize whether the real-time working condition dataset is within a preset safety range. The closed-loop correction unit, based on the quality risk prediction set and using the current process measurement value as the correction starting point, calculates the target process setpoint for the next control cycle. The target process setpoint aims to bring the predicted quality index closer to the preset target quality index, while controlling the predicted defect probability within a preset risk threshold. The closed-loop correction unit is further used to adaptively adjust the calculation strategy of the target process setpoint according to the operating condition risk level.

[0007] As a further aspect of the present invention, the environmental status data includes ambient temperature and ambient humidity; the current process measurement values ​​include current atomization pressure and current coating flow rate.

[0008] As a further aspect of the present invention, the core process feature set includes: The coating viscosity deviation index is used to quantify the degree to which the current coating viscosity deviates from the standard state due to the influence of the ambient temperature. The comprehensive leveling index is used to comprehensively evaluate the impact of the current atomization pressure and the ambient humidity on the leveling properties of the coating.

[0009] As a further aspect of the present invention, the predicted quality index in the quality risk prediction set is specifically limited to the predicted coating thickness, and the initial value of the predicted defect probability is specifically limited to the basic defect probability based on historical data. The predicted coating thickness is calculated based on the comprehensive leveling index and the current coating flow rate; the basic defect probability is calculated based on the coating viscosity shift index and the comprehensive leveling index.

[0010] As a further aspect of the present invention, the working condition risk levels in the quality risk prediction set include conventional risk levels and conservative risk levels. Compare the ambient temperature and the current atomization pressure with preset temperature safety ranges and pressure safety ranges; If all conditions are within the safe range, then the risk level of the operating condition is set to the normal risk level. If any parameter exceeds the safety range, the operating condition risk level will be set to a conservative risk level.

[0011] As a further aspect of the present invention, when the operating condition risk level is a conservative risk level, an operating condition deviation degree is calculated to quantify the degree to which the current operating condition deviates from the safe range; based on the basic defect probability and a preset conservative coefficient, a weighted comprehensive defect probability is generated, and the weighted comprehensive defect probability is used as the updated predicted defect probability. The quality risk prediction set further includes the operating condition deviation and the weighted comprehensive defect probability.

[0012] As a further aspect of the present invention, the closed-loop correction unit calculates the target process setpoint by constructing an adaptive objective function; When the working condition risk level is a normal risk level, the adaptive objective function aims to minimize the deviation between the predicted coating thickness and the target coating thickness, as well as the deviation between the basic defect probability and the risk threshold. When the operating condition risk level is a conservative risk level, the adaptive objective function aims to minimize the deviation between the predicted coating thickness and the target coating thickness, as well as the deviation between the weighted comprehensive defect probability and the risk threshold, and adds an additional dynamic penalty term that is positively correlated with the operating condition deviation and the coating flow rate to be optimized.

[0013] As a further aspect of the present invention, the step of the closed-loop correction unit calculating the target process setpoint includes: Based on the aforementioned adaptive objective function, an optimal process correction amount is calculated using the gradient descent method. The optimal process correction amount is summed with the current process measurement value to obtain the target process setpoint, thus forming a closed-loop correction from the current measurement value to the target setpoint of the next cycle.

[0014] The beneficial effects of this invention compared to the prior art are as follows: 1. This invention achieves in-depth insight and quantitative assessment of complex operating conditions, effectively alleviating the problem of insufficient perception of environmental and process fluctuations in existing technologies. The invention acquires multi-source real-time operating condition datasets, including ambient temperature and humidity, coating viscosity, and process parameters, through a state perception unit. Furthermore, through a feature extraction unit, it innovatively refines and couples these multi-dimensional physical quantities into a core set of process features directly related to coating quality. This allows the system to no longer view individual parameters in isolation, providing a scientific and quantitative basis for subsequent accurate prediction and control.

[0015] 2. This invention enables proactive prediction of coating quality and defect risks. Through a risk prediction unit, based on the aforementioned core process feature set, the invention can accurately predict the final quality indicators and defect probabilities before the coating process is executed. This "predictive" capability allows the system to identify potential quality risks such as sagging and orange peel caused by changes in operating conditions in advance. This shifts the control mechanism from reactive response after a problem occurs to proactive prevention before it arises, significantly improving the first-pass yield of products.

[0016] 3. Intelligent optimization and dynamic balancing are achieved under dynamic changes. This invention constructs an adaptive objective function with film thickness accuracy and defect probability as its core through a closed-loop correction unit, and performs closed-loop correction starting from the current process measurement value. This mechanism can calculate the optimal process setpoint that meets high-quality requirements and maximizes production efficiency in real time and dynamically in each control cycle.

[0017] 4. Improved system robustness and production safety under extreme operating conditions. This invention creatively introduces the concepts of operating condition risk level and operating condition deviation. When the system senses that the operating condition exceeds the normal safety range, the closed-loop correction unit automatically switches to a conservative strategy. By adding a dynamic penalty term to the objective function, the system is forced to adopt more cautious process parameters, ensuring that the system will not make incorrect decisions when facing sudden and unforeseen extreme operating conditions, thereby avoiding batch scrap accidents caused by control system failure. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of a module of an automated control system for spraying equipment based on the Industrial Internet of Things according to the present invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0020] This invention proposes an automated control system and method for spraying equipment based on the Industrial Internet, aiming to resolve the contradiction between spraying quality and production efficiency caused by fluctuations in environmental factors and fixed process parameters, as mentioned in the background art. This invention achieves dynamic adaptive optimization of the spraying process by real-time sensing of operating conditions, accurate risk prediction, and closed-loop correction of process parameters.

[0021] The core of the automated control system for spraying equipment of this invention lies in an intelligent closed-loop control loop composed of a state sensing unit, a feature extraction unit, a risk prediction unit, and a closed-loop correction unit. This system relies on an industrial internet platform to achieve real-time acquisition of sensor data within the spraying workshop, intelligent computing at the cloud or edge, and the issuance of real-time control commands to the spraying equipment. Specifically: A state perception unit is used to acquire a real-time operating condition dataset, which includes: environmental state data characterizing the current spraying environment state, and current process measurement values ​​characterizing the current operating state of the spraying equipment; wherein, the current process measurement values ​​are process parameter values ​​measured in real time by sensors within the current control cycle. The environmental condition data specifically includes ambient temperature and humidity measured by temperature and humidity sensors. These data directly reflect the key external factors affecting the physical properties of the coating.

[0022] The current process measurements specifically include the current coating flow rate and current atomizing air pressure, measured by flow meters and pressure sensors. These data are core process parameters that the control system can directly adjust and that directly affect the final coating quality.

[0023] The state-aware unit provides the system with the most basic and immediate information needed for decision-making. Compared to traditional systems that ignore environmental changes or rely on expensive constant temperature and humidity systems, the state-aware unit of this invention can accurately and in real-time capture the dynamic fluctuations of the production environment and the actual operating status of equipment in a low-cost manner. This provides an accurate data foundation for subsequent analysis, prediction, and control.

[0024] The feature extraction unit generates a core process feature set based on the real-time working condition dataset and a preset physical model. The core process feature set is used to quantify the comprehensive impact of the real-time working condition dataset on the core state of the spraying process. The core process feature set specifically includes: Coating viscosity offset index This index is used to quantify the current ambient temperature. The effect on coating viscosity, i.e., the current viscosity compared to a standard or ideal state (e.g., at a reference temperature). The degree of deviation (of viscosity below). Based on the simplified form of the Arrhenius equation: ; in: This refers to the viscosity offset index of the coating. The current ambient temperature is obtained by the state sensing unit; This is the preset reference temperature, i.e., the ideal or standard spraying process temperature; This is the temperature sensitivity coefficient of the coating; this coefficient is determined by the chemical composition of the coating itself and can be obtained through experimental calibration.

[0025] Comprehensive leveling index Used for comprehensive evaluation of current atomization pressure and ambient humidity The overall impact on paint leveling properties is that higher atomization pressure helps paint spread, but excessive humidity slows down solvent evaporation and prolongs leveling time. Both factors jointly determine the final leveling effect. The comprehensive leveling index is calculated as follows: ; in: The overall leveling index; The current atomization pressure obtained by the state sensing unit; The current relative humidity of the environment is obtained by the state sensing unit; These are the weighting coefficients for atomization pressure and ambient humidity, respectively, obtained by fitting experimental data based on specific coating characteristics and process requirements.

[0026] It should be noted that this calculation method is an effective engineering approximation within a specific process window. It simplifies complex fluid dynamics and mass transfer processes into a linear empirical model, and its effectiveness is determined by weighting coefficients. and The experimental data fitting is used to ensure that, while maintaining control accuracy, this simplification greatly reduces the computational complexity of the model, making it very suitable for real-time control systems.

[0027] The feature extraction unit refines multidimensional, raw sensor data into a few key feature indicators with clear physical meaning. This greatly simplifies the complexity of subsequent models and enhances their interpretability. This quantitative assessment enables the system to more accurately determine the true impact of operating condition changes on the core state of the process, providing high-quality input for subsequent risk prediction. In this embodiment, the pre-defined physical model is specifically embodied in the simplified Arrhenius equation for calculating the coating viscosity shift index and the linear empirical formula for calculating the comprehensive leveling index.

[0028] The risk prediction unit generates a quality risk prediction set based on the core process feature set and the current process measurement value. The quality risk prediction set includes: a predicted quality index for the final coating quality, a predicted defect probability for the occurrence of spraying defects, and a working condition risk level used to characterize whether the real-time working condition dataset is within a preset safety range. The quality risk prediction set specifically includes: Predicting coating thickness Based on the current paint flow rate and comprehensive leveling index Calculate the expected final coating thickness. A higher flow rate results in a thicker film; better leveling (more even coating spread) results in a thinner film at the same flow rate. The calculation formula is as follows: ; in: To predict coating thickness; This represents the current paint flow rate; The overall leveling index; This is the flow rate-film thickness conversion coefficient, which is related to relatively fixed parameters such as spray gun model, spraying distance, and robot trajectory speed, and can be calibrated experimentally.

[0029] Predicting defect probability Basic defect probability This is based on historical production data statistics. The system is tailored to the current core technological characteristics. and Correct it. For example, when the viscosity offset index... When the viscosity is too high (too low), the risk of sagging increases; when the overall leveling index is too high... When the temperature is too low, the risk of orange peel and pinholes increases. The calculation can be expressed as: ; function It is obtained through training a machine learning model. Specifically, the function... A feedforward neural network model can be used for construction. This is achieved by collecting a large amount of historical production data (including...). , , As input features, and the corresponding actual defect rate (as labels), the neural network model is trained using supervised learning. The network weights are continuously adjusted through backpropagation and gradient descent optimizer until the model predicts... Minimize the error between the actual value and the actual value.

[0030] Operating condition risk level This is a macro-level assessment of the overall safety of the current operating conditions. The system will consider the current ambient temperature. and atomization pressure With the preset safety range and Comparison: like and ,but =Standard risk level.

[0031] like or ,but =Conservative risk level.

[0032] When the risk level is adjusted to conservative, the system will perform further calculations: Operating condition deviation Used to quantify the severity of the current operating condition exceeding the safe range.

[0033] ; in These are the weighting coefficients. It is a dimensionless value.

[0034] Weighted composite defect probability A conservative penalty is applied to the basic defect probability based on the degree of deviation.

[0035] ; in This is a preset conservative coefficient. Overall defect probability. This will replace the original basic defect probability Used for subsequent calculations.

[0036] The risk prediction unit realizes the transformation from "post-event inspection" to "pre-event early warning." It can not only predict specific quality indicators, but also quantify the probability of defects. More importantly, by introducing operating condition risk levels and operating condition deviations, the system has a graded response capability, which greatly enhances the system's robustness and ability to handle abnormal operating conditions.

[0037] The closed-loop correction unit, based on the quality risk prediction set and using the current process measurement value as the correction starting point, calculates the target process setpoint for the next control cycle. The target process setpoint aims to bring the predicted quality index closer to the preset target quality index, while controlling the predicted defect probability within a preset risk threshold. The closed-loop correction unit is further used to adaptively adjust the calculation strategy of the target process setpoint according to the operating condition risk level.

[0038] Its core is to construct an adaptive objective function. This is achieved through optimization and solution; the objective function is adjusted based on the risk level of the working condition. Its structure is dynamically adjusted.

[0039] When the working condition risk level For a standard risk level: Objective function We are committed to balancing film thickness and defect probability while meeting quality requirements.

[0040] ; in: The target film thickness value; The maximum acceptable defect probability threshold; For film thickness deviation; This is the weighting coefficient for the defect probability deviation; The paint flow rate to be optimized; This is a correction amount for the atomization pressure.

[0041] when When the risk level is conservative: Objective function For the objective function Based on the original system, an additional dynamic penalty is added to suppress behaviors that significantly increase traffic in pursuit of efficiency under high-risk conditions.

[0042] ; in: λ is a dynamic penalty coefficient used to adjust the strength of the penalty term. It is automatically optimized through a cloud platform based on historical defect data. For example, this optimization process can be implemented based on the Q-learning idea in reinforcement learning or a simple adaptive feedback control logic: establish a rule base, and if the defect rate still exceeds the standard in the next cycle after applying the penalty, then appropriately increase λ; if the system efficiency drops significantly due to overly conservative penalties and the defect rate is far below the threshold, then appropriately decrease λ, thereby achieving dynamic optimization of λ. Deviation from operating conditions; This indicates that it only applies to increasing traffic ( Punish the behavior of ( ).

[0043] The steps for calculating the target process setpoint are as follows: Construct an adaptive objective function: Select based on the current risk level of the operating condition. or As an optimization target.

[0044] Optimization Solution: Numerical optimization algorithms such as gradient descent are used to calculate the objective function. Minimum optimal process correction amount The iterative formula for gradient descent is: ; in: The learning rate η is a key hyperparameter of the optimization algorithm. Its initial value and adjustment strategy can be determined using techniques known in the art, such as finding the optimal value within a preset range through grid search or random search, or directly using an optimization algorithm with an adaptive learning rate, where the algorithm automatically adjusts the learning rate. For the objective function pair The gradient; This represents the current iteration number of the gradient descent algorithm; For the The amount of process correction during the next iteration.

[0045] Calculate the target setpoint: Sum the obtained optimal correction value with the current measurement value to obtain the target setpoint for the next cycle. ; ; in: and These are the current process measurement values; and The final calculated target process setpoint forms a closed-loop correction from the current measurement value to the target setpoint for the next cycle.

[0046] The closed-loop correction unit optimizes the objective function, finding a dynamic optimal balance among multiple mutually constraining objectives. In particular, the adaptive objective function design ensures that within the safe zone, the system prioritizes a balance between efficiency and conventional quality indicators; once it enters the risk zone, the system's behavior immediately switches, prioritizing risk control through dynamic penalty terms. This ensures that the system consistently makes the most rational control decisions in complex and ever-changing production environments.

[0047] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the present invention should still fall within the scope of the present invention.

Claims

1. An automated control system for spraying equipment based on the Industrial Internet, characterized in that, include: A state perception unit is used to acquire a real-time operating condition dataset, which includes: environmental state data characterizing the current spraying environment state, and current process measurement values ​​characterizing the current operating state of the spraying equipment; wherein, the current process measurement values ​​are the values ​​of process parameters measured in real time by sensors within the current control cycle. The feature extraction unit generates a core process feature set based on the real-time working condition dataset and a preset physical model. The core process feature set is used to quantify the comprehensive impact of the real-time working condition dataset on the core state of the spraying process. The risk prediction unit generates a quality risk prediction set based on the core process feature set and the current process measurement value. The quality risk prediction set includes: a predicted quality index for the final coating quality, a predicted defect probability for the occurrence of spraying defects, and a working condition risk level used to characterize whether the real-time working condition dataset is within a preset safety range. The closed-loop correction unit, based on the quality risk prediction set and using the current process measurement value as the correction starting point, calculates the target process setpoint for the next control cycle. The target process setpoint aims to bring the predicted quality index closer to the preset target quality index, while controlling the predicted defect probability within a preset risk threshold. The closed-loop correction unit is further used to adaptively adjust the calculation strategy of the target process setpoint according to the operating condition risk level.

2. The automated control system for spraying equipment based on the Industrial Internet as described in claim 1, characterized in that, The environmental status data includes ambient temperature and ambient humidity; the current process measurements include current atomization pressure and current coating flow rate.

3. The system according to claim 2, characterized in that, The core process feature set includes: The coating viscosity deviation index is used to quantify the degree to which the current coating viscosity deviates from the standard state due to the influence of the ambient temperature. The comprehensive leveling index is used to comprehensively evaluate the impact of the current atomization pressure and the ambient humidity on the leveling properties of the coating.

4. The automated control system for spraying equipment based on the Industrial Internet as described in claim 3, characterized in that, The predicted quality index in the quality risk prediction set is specifically limited to the predicted coating thickness, and the initial value of the predicted defect probability is specifically limited to the basic defect probability based on historical data. The predicted coating thickness is calculated based on the comprehensive leveling index and the current coating flow rate. The probability of the basic defect is calculated based on the coating viscosity offset index and the comprehensive leveling index.

5. The automated control system for spraying equipment based on the Industrial Internet as described in claim 4, characterized in that, The operating condition risk levels in the quality risk prediction set include conventional risk levels and conservative risk levels; Compare the ambient temperature and the current atomization pressure with preset temperature safety ranges and pressure safety ranges; If all conditions are within the safe range, then the risk level of the operating condition is set to the normal risk level. If any parameter exceeds the safety range, the operating condition risk level will be set to a conservative risk level.

6. The automated control system for spraying equipment based on the Industrial Internet as described in claim 5, characterized in that, When the operating condition risk level is a conservative risk level, the operating condition deviation degree is calculated to quantify the degree to which the current operating condition deviates from the safe range; based on the basic defect probability and a preset conservative coefficient, a weighted comprehensive defect probability is generated, and the weighted comprehensive defect probability is used as the updated predicted defect probability. The quality risk prediction set further includes the operating condition deviation and the weighted comprehensive defect probability.

7. The automated control system for spraying equipment based on the Industrial Internet as described in claim 6, characterized in that, The closed-loop correction unit calculates the target process setpoint by constructing an adaptive objective function; When the working condition risk level is a normal risk level, the adaptive objective function aims to minimize the deviation between the predicted coating thickness and the target coating thickness, as well as the deviation between the basic defect probability and the risk threshold. When the operating condition risk level is a conservative risk level, the adaptive objective function aims to minimize the deviation between the predicted coating thickness and the target coating thickness, as well as the deviation between the weighted comprehensive defect probability and the risk threshold, and adds an additional dynamic penalty term that is positively correlated with the operating condition deviation and the coating flow rate to be optimized.

8. The automated control system for spraying equipment based on the Industrial Internet as described in claim 7, characterized in that, The steps of the closed-loop correction unit in calculating the target process setpoint include: Based on the aforementioned adaptive objective function, an optimal process correction amount is calculated using the gradient descent method. The optimal process correction amount is summed with the current process measurement value to obtain the target process setpoint, thus forming a closed-loop correction from the current measurement value to the target setpoint of the next cycle.

Images