An adaptive compensation intelligent industrial coating closed-loop control method and system

Abstract

The application discloses a kind of self-adapting compensation intelligent industrial coating closed-loop control method, comprising the following steps: workpiece pretreatment and reference calibration, multi-modal perception data acquisition, start multi-modal perception network, the multi-modal heterogeneous data of workpiece is synchronously collected, multi-source data fusion and intelligent analysis, the multi-modal heterogeneous data collected is denoising, alignment, integration, adaptive control parameter generation and execution, coating quality real-time detection and feedback optimization and coating post-processing and data archiving: complete coating after adaptive adjustment drying parameter, this coating whole-process data is archived, forms traceable production record.The application integrates vision, laser and multiple sensors, synchronously collects five kinds of core data of workpiece, paint, environment, equipment and coating, defect recognition accuracy reaches 0.1mm, coating thickness measurement accuracy is ±2 μm, color difference ΔE<0.5, perception dimension is improved by 4 times compared with single scheme, significantly reduces the influence of perception blind area.

Description

Technical Field

[0001] This application relates to the field of industrial coating, and more specifically, to an adaptive compensation intelligent industrial coating closed-loop control method and system. Background Technology

[0002] Industrial coating is a core process for achieving surface protection, decoration, and functionality in products; its quality directly determines the product's lifespan, appearance, and market competitiveness. Currently, the industry faces core technological bottlenecks that have become key obstacles hindering the high-quality development of coating processes. Incomplete perception information and prominent blind spot hazards: Existing coating systems mostly use single vision or sensor perception, which can only obtain local information of the workpiece surface and cannot simultaneously capture multi-dimensional data such as the three-dimensional shape of the workpiece, surface defects, coating status, spraying environment and equipment operating parameters.

[0003] The control mode is rigid and lacks flexibility: Traditional coating process parameters are mostly preset fixed values, which can only be adjusted manually. They cannot be dynamically adjusted according to differences in workpieces, environmental changes, coating conditions, and equipment wear. When faced with workpiece size deviations, fluctuations in ambient temperature and humidity, and changes in coating viscosity, problems such as uneven coating thickness, excessive color difference, and coating waste are likely to occur. When switching between multiple products, the debugging cycle is long and the efficiency is low, making it difficult to adapt to the needs of flexible production.

[0004] Lagging quality control and lack of closed-loop optimization: Current coating quality inspections are mostly post-event sampling inspections, which cannot achieve real-time process detection and feedback. By the time defects are discovered, a large number of non-conforming products have already been generated. Furthermore, the detection data is disconnected from the control parameters, making it impossible to optimize process parameters through defect data, thus hindering continuous quality improvement.

[0005] Breakthroughs in high-end technologies are difficult, and the pressure of green transformation is significant: High-end coating equipment and core algorithms have long relied on imports, with foreign brands holding over 90% market share. Standardized process routes are poorly adapted to my country's reality of "multi-variety, small-batch" production, and equipment maintenance costs are high, while process upgrades are subject to external constraints. Domestic systems lack in sensing accuracy, control response speed, and algorithm optimization capabilities. Furthermore, due to inaccurate parameter control, paint utilization is generally below 40%, VOC emission control is difficult, and green transformation is costly. In the field of electrostatic coating for complex geometric workpieces such as automotive bumpers, the problem of uneven coating thickness caused by the Faraday cage effect is particularly prominent: sensitive areas such as deep grooves, sharp inner corners, and narrow gaps suffer from low paint particle deposition efficiency due to electric field shielding, resulting in coating thicknesses that are generally 30%-40% lower than the baseline area. Meanwhile, adjacent flange areas are prone to over-deposition. Existing solutions can only alleviate this through post-compensation methods such as increasing the number of sprays, failing to address the problem at the physical mechanism level. This affects coating quality consistency and causes paint waste. Therefore, there is an urgent need to develop an intelligent coating solution that can proactively predict and intervene in the distribution of electrostatic fields, so as to fundamentally solve the coating quality bottleneck of complex geometric workpieces. Summary of the Invention

[0006] In view of the above-mentioned defects and shortcomings of the existing technology, the core objective of the present invention is to provide an adaptive compensation intelligent industrial coating closed-loop control method and system, specifically achieving the following objectives.

[0007] An adaptive compensation intelligent industrial coating closed-loop control method includes the following steps: S1. Workpiece pretreatment and reference calibration: The workpiece to be coated is cleaned and plasma treated, the reference image of the workpiece is collected and compared with the preset standard model, the reference calibration of the workpiece position and posture is completed, and the spraying start position and range are determined. S2. Multimodal sensing data acquisition: Start the multimodal sensing network to synchronously acquire multimodal heterogeneous data of the workpiece, with a sampling frequency of 10-20Hz; S3. Multi-source data fusion and intelligent analysis: First, multi-source heterogeneous data are fused by weighted average to eliminate noise. Then, the fused feature vector is input into the trained CNN model for intelligent analysis of the workpiece. S4. Adaptive control parameter generation and execution: Based on intelligent analysis results, the optimal spraying parameters are automatically generated and sent to the execution module. The spraying path, spraying process parameters and auxiliary equipment operating status are dynamically adjusted. The electrostatic voltage / current of the spray gun is dynamically adjusted synchronously for the Faraday cage effect sensitive area to achieve active intervention of the electric field. The response delay of the execution module is <0.1s. S5. Real-time detection and feedback optimization of coating quality: Real-time detection of coating quality, graded feedback according to the degree of defect, real-time adjustment of parameters to repair minor defects, triggering shutdown warning for serious defects, and storing all detection data in the database. S6. Post-coating processing and data archiving: After coating is completed, the drying parameters are adaptively adjusted, and the entire coating process data is archived to form a traceable production record. Furthermore, in step S1, a high-pressure airflow with a pressure of 0.3-0.5MPa is used to remove impurities from the workpiece surface, and a plasma treatment device with a power of 500-800W is used to treat the workpiece surface for 10-20 seconds, with a reference calibration error ≤ ±0.05mm.

[0008] Furthermore, in step S2, the multimodal heterogeneous data of the workpiece includes five categories: three-dimensional morphology and surface defect data, coating status data, spraying environment data, equipment operation status data, and coating real-time status data.

[0009] Furthermore, the three-dimensional morphology and surface defect data are acquired using a line laser scanner and an industrial high-resolution camera. The line laser scanner has a scanning accuracy of 0.01 mm, and the smallest identifiable defect size is 0.1 mm × 0.1 mm. Coating status data includes viscosity, density, flow rate, and temperature, with viscosity measured in the range of 10-100 mPa·s and an accuracy of ±1 mPa·s. Spraying environment data includes workshop temperature and humidity, air velocity, and particulate matter content, with temperature measured in the range of 0-50℃ and humidity in the range of 20%-80%RH. Equipment operating status data includes the joint torque of the spraying robot, vibration frequency, spray gun atomization pressure, and electrostatic voltage / current, with torque measured in the range of 0-100 N·m and an accuracy of ±0.5 N·m, and electrostatic voltage measured in the range of 0-100 kV and an accuracy of ±1 kV. Real-time coating status data is acquired using multispectral imaging technology, a spectrophotometer, and laser-induced breakdown spectroscopy, with coating thickness measurement accuracy of ±2 μm and color difference accuracy ΔE < 0.5. Furthermore, in step S3, the intelligent analysis specifically includes: adaptability analysis, coating status analysis, environmental impact analysis, equipment status analysis and coating quality prediction, output parameter adjustment suggestions, and data processing delay ≤0.2s; for complex geometric workpieces such as automobile bumpers, a three-dimensional reconstruction model of the workpiece is constructed simultaneously based on three-dimensional point cloud data, the sensitive area of ​​Faraday cage effect is identified, the local electric field distribution and coating deposition efficiency are predicted, and an electrostatic parameter adjustment strategy is generated.

[0010] Furthermore, in step S3, the multi-source data fusion adopts a weighted average method, with the weight allocation as follows: workpiece data 0.3, coating data 0.25, environmental data 0.15, equipment data 0.15, and coating data 0.15; the process parameter database covers spraying strategies for at least 2,000 typical workpieces, including a feature library of Faraday cage effect sensitive areas for complex geometric workpieces such as car bumpers.

[0011] The specific logic for identifying sensitive areas and predicting electric fields in the Faraday cage effect is as follows: Based on the three-dimensional point cloud data collected by the line laser scanner, a three-dimensional reconstruction model of the workpiece is constructed. Through curvature analysis and calculation of local surface normal deviation, deep grooves, sharp inner angles, narrow gaps, and shading areas are identified, and the geometric complexity and electric field shielding coefficient are quantified. A pre-trained real-time electric field prediction model or a pre-calculated lookup table is called, and combined with the current spray gun position, attitude, and initial electrostatic parameters, the local electric field intensity and coating deposition efficiency of the sensitive area are quickly estimated, and the deposition deviation and the risk of over-deposition in adjacent areas are quantified and predicted.

[0012] Further, in step S4, the specific logic of the dynamic adjustment includes: spraying path adjustment: automatically adjusting the motion trajectory of the spraying robot and the angle of the spray gun to ensure that the distance between the spray gun and the workpiece surface is maintained at 100-300mm, and the spray overlap rate in complex curved areas is adjusted to 50%-70%; spraying process parameter adjustment: paint flow rate is adjusted to 0.5-5L / min, atomization pressure is adjusted to 0.2-0.8MPa, spraying speed is adjusted to 0.5-2m / s, spraying times are adjusted to 1-5 times, and the paint flow rate in defective areas is increased by 20%-50%. Increase the number of spray coats by 1-2 times; dynamically adjust electrostatic parameters: when the spray gun enters the Faraday cage sensitive area, temporarily reduce the electrostatic voltage by 10%-30% or the current by 5%-15%, allowing charged paint particles to penetrate the recessed area by inertia; when the spray gun leaves the sensitive area or enters the flange area, restore or increase the electrostatic voltage / current by 5%-10% to maintain paint transfer efficiency; adjust auxiliary equipment: automatically control the workshop temperature and humidity to maintain a temperature of 20-25℃ and a humidity of 40%-60%, and control the paint state to maintain a viscosity of 20-30 mPa·s and a temperature of 20-25℃. Furthermore, in step S5, the specific rules for graded feedback are as follows: if the coating quality meets the standard, a qualified signal is output and the system operates according to the current parameters; if minor defects are detected, including coating thickness deviations exceeding ±5μm in the Faraday cage area, the spraying parameters or electrostatic parameters are adjusted in real time to complete the repair without stopping the machine; if serious defects are detected, including large-area thickness deviations, sagging, and severe color differences, a stop signal is immediately triggered to prompt the operator to troubleshoot the fault.

[0013] Furthermore, in step S6, the drying temperature is adjusted within the range of 60-120℃, the drying time is adjusted within the range of 10-30min, and the coating adhesion level is ≥1; all production data storage period is ≥3 years, and the entire process information can be traced through the unique identifier of the workpiece.

[0014] Furthermore, an intelligent industrial coating system with multimodal sensing and adaptive control includes: The conveying and pretreatment module is used to complete workpiece conveying, surface cleaning, plasma treatment and reference calibration; The multimodal sensing module integrates the workpiece sensing unit, coating sensing unit, environment sensing unit, equipment sensing unit, and coating sensing unit to simultaneously collect five types of core sensing data. The data processing unit is used to complete multi-source data fusion and intelligent analysis, output parameter adjustment suggestions, and has a data processing delay of ≤0.2s. It has a built-in predictive electrostatic field adjustment submodule, which is used to identify Faraday cage effect sensitive areas, predict electric field distribution and generate electrostatic parameter adjustment strategies. An adaptive control unit is used to generate optimal control parameters and send them to the execution module, with a response time of ≤0.1s; The execution module, including the spraying robot, paint supply system, electrostatic generator, environmental control equipment, and drying equipment, is used to execute control commands to complete spraying and post-processing. The detection module is used to detect coating quality in real time and output feedback signals according to the defect level; the database and traceability module is used to store production data throughout the entire process and support quality traceability. The human-machine interaction module is used for parameter setting, status monitoring and fault handling; all modules are connected via industrial Ethernet with a communication rate of ≥100Mbps.

[0015] Furthermore, in the multimodal sensing module: the workpiece sensing unit includes a line laser scanner with a scanning accuracy of 0.01 mm and an industrial high-resolution camera with a resolution of 1920×1080; the paint sensing unit includes a viscosity sensor, a density meter, a flow sensor, and a temperature sensor installed on the paint supply pipe and the paint storage tank; the environmental sensing unit includes a temperature and humidity sensor, an air velocity meter, and a particulate matter detector evenly distributed around the painting station; the equipment sensing unit includes a torque sensor, an encoder, and a vibration sensor installed on the joints of the spraying robot, as well as a pressure sensor and an electrostatic voltage sensor installed on the spray gun and the paint supply pipe. / Current sensor; The coating sensing unit includes a multispectral imager, a spectrophotometer, and a laser-induced breakdown spectrometer, with a coating thickness measurement accuracy of ±2μm. The predictive electrostatic field adjustment submodule includes: a geometric feature analysis unit, used to construct a 3D reconstruction model of the workpiece based on 3D point cloud data, identify and classify Faraday cage effect sensitive areas, and quantify the electric field shielding coefficient; an electric field prediction unit, with a built-in pre-trained electric field prediction model or pre-calculated lookup table, used to quickly estimate the local electric field strength and coating deposition efficiency of sensitive areas; and a parameter generation unit, used to generate electrostatic parameter adjustment instructions for the corresponding area based on the electric field prediction results.

[0016] Furthermore, a computer-readable storage medium having a computer program stored thereon that, when executed by a processor, implements the steps of the method described.

[0017] Beneficial effects:

[0018] 1. This invention integrates vision, laser and multiple types of sensors to simultaneously collect five core data: workpiece, coating, environment, equipment and coating. The defect identification accuracy reaches 0.1mm, the coating thickness measurement accuracy is ±2μm, the color difference ΔE<0.5, and the perception dimension is increased by 4 times compared with a single solution, significantly reducing the impact of perception blind spots.

[0019] 2. This invention is based on deep learning and iterative updates of models based on historical data to achieve real-time dynamic adjustment of process parameters, automatically adapt to workpiece deviations, environmental fluctuations, coating status and equipment wear, shorten the commissioning cycle of multi-product line changeover by more than 30%, and achieves automated configuration of process parameters without manual intervention, thereby reducing manual commissioning costs.

[0020] 3. This invention constructs a complete closed loop of "perception-analysis-decision-execution-detection-feedback-optimization", which can detect and repair minor defects in real time, greatly improving the first-pass yield and greatly reducing the defect rate.

[0021] 4. This invention transforms the experience of senior technicians into digital models, constructs a process parameter database, realizes the digital inheritance and reuse of experience, improves the replicability and stability of production lines, and facilitates large-scale promotion.

[0022] 5. This invention can be widely applied to coating of various types of workpieces such as automotive parts, construction machinery, and 3C electronics. It supports flexible adjustment of parameters and sensing accuracy, adapts to different industry coating standards, and its application scope far exceeds that of a single-scenario system. Attached Figure Description

[0023] The invention will now be further described with reference to the accompanying drawings; Figure 1 This is a flowchart of the method of the present invention; Figure 2 This is a framework diagram of the system in this invention. Detailed Implementation

[0024] 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.

[0025] The technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. The components of this application described and shown in the accompanying drawings can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0026] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0027] Example 1

[0028] Reference Figure 1 This invention discloses an adaptive compensation intelligent industrial coating closed-loop control method, comprising: S1. Workpiece Pre-treatment and Reference Calibration: The car bumper to be painted is transported to the painting station via a belt conveyor. The conveyor speed is set to 0.2 m / s, the load capacity is 0-500 kg, and the positioning accuracy is ±0.05 mm. A high-pressure airflow cleaning device is activated, using 0.4 MPa high-pressure airflow from six 8 mm diameter nozzles on both sides of the conveyor to remove dust and impurities from the workpiece surface. Subsequently, a plasma treatment device is activated, with a power of 650 W, a treatment width of 300 mm, and a treatment time of 15 seconds. The device is installed above the conveyor, 80 mm from the workpiece surface, and uses plasma to bombard the workpiece surface, improving coating adhesion. Simultaneously, a 3D vision sensor acquires a reference image of the workpiece and compares it with a standard bumper model stored in the database to complete the reference calibration of the workpiece's position and orientation. The calibration error is controlled within ±0.05 mm, determining the starting position and range of the spraying to accommodate workpiece dimensional deviations (≤±0.5 mm).

[0029] S2. Multimodal Sensing Data Acquisition: Activate the multimodal sensing network and simultaneously acquire five core data types at a sampling frequency of 15Hz: Workpiece 3D morphology and surface defect data: A line laser scanner with a scanning accuracy of 0.01mm and a scanning speed of 500mm / s is used to collect 3D point cloud data of the workpiece, construct a 3D digital reconstruction model of the workpiece, and identify substrate defects such as depressions and protrusions of 0.1mm×0.1mm on the workpiece surface; At the same time, an industrial high-resolution camera with a resolution of 1920×1080 and a frame rate of 30fps is used with special optical filters to eliminate the interference of spray fog, capture the details of the workpiece surface, assist in defect identification, minimize the perception blind spot, and effectively solve the problem of coating coverage on the back of complex workpieces.

[0030] Paint condition data: Viscosity, density, flow rate and temperature of the paint are monitored in real time by viscosity sensors, densitometers and flow rate sensors installed on the paint supply pipeline and temperature sensors installed on the paint storage tank. This ensures that the paint is in the best working condition and avoids problems such as coating sagging and orange peel caused by abnormal paint condition.

[0031] Spraying environment data: Temperature and humidity sensors, air velocity meters, and particulate matter detectors are evenly distributed around the painting station to collect workshop temperature and humidity, air velocity, and particulate matter content, avoiding painting defects caused by environmental factors and providing a basis for adaptive parameter adjustment.

[0032] Equipment operating status data: Through torque sensors, encoders, and vibration sensors installed on the joints of the painting robot, data such as the robot's running trajectory, joint torque, and vibration frequency are collected in real time to identify problems such as equipment wear and trajectory deviation. At the same time, pressure sensors and electrostatic voltage / current sensors installed on the spray gun and paint supply pipeline are used to detect whether the spray gun atomization pressure, paint supply pressure, and electrostatic voltage / current are stable, allowing for early maintenance intervention and avoiding quality fluctuations caused by equipment problems.

[0033] Real-time coating status data: Multispectral imaging technology is used to measure coating thickness non-contactly, a spectrophotometer is used to detect coating color parameters, and laser-induced breakdown spectroscopy (LIBS) is used to measure the thickness of a single layer in multi-layer coatings. Real-time feedback on coating quality provides data support for coating quality prediction, ensuring that the coating thickness deviation is ≤ ±2μm and the color difference ΔE < 0.5.

[0034] S3. Multi-source data fusion and intelligent analysis: The multimodal heterogeneous data collected in step S2 is transmitted to the data processing unit. This unit uses an industrial controller with a CPU frequency of ≥2.5GHz, memory of ≥8GB, and storage capacity of ≥128GB. The specific processing procedure is as follows: Multi-source data fusion: A weighted average fusion algorithm is used to reduce noise, align and integrate workpiece data (weight 0.3), coating data (weight 0.25), environmental data (weight 0.15), equipment data (weight 0.15), and coating data (weight 0.15), eliminating the limitations of single sensing methods and outputting comprehensive and accurate integrated data.

[0035] Intelligent Analysis: Based on CNN convolutional neural networks and iterative model updates based on historical data, combined with a pre-set process parameter database (covering spraying strategies for 2000+ typical workpieces, including a feature library of Faraday cage effect sensitive areas for complex geometric workpieces such as car bumpers), the following analysis is performed with a data processing latency of ≤0.2s: Workpiece compatibility analysis: By comparing the workpiece 3D reconstruction model with the standard model, the workpiece size deviation, surface defect location and degree are determined, and it is determined whether the spraying path and parameters need to be adjusted.

[0036] Coating condition analysis: Determine whether the coating viscosity, density, and temperature deviate from the standard range, analyze the reasons for the deviation, and provide parameter adjustment suggestions.

[0037] Environmental impact analysis: Based on data such as workshop temperature and humidity, and air velocity, analyze the impact of the environment on the leveling properties and drying speed of the coating, and determine whether it is necessary to adjust the spraying parameters or environmental control equipment.

[0038] Equipment status analysis: By analyzing equipment operating data, we can predict the wear and tear of mechanical parts, identify trajectory deviations, and take early maintenance interventions to avoid quality fluctuations caused by equipment problems; at the same time, we can determine whether the spray gun atomization pressure, paint supply pressure, and electrostatic voltage / current are stable.

[0039] Coating quality prediction: Based on real-time coating thickness and color data, predict whether the coating has defects such as sagging, orange peel, color difference, and pinholes, analyze the causes of defects, and output parameter adjustment instructions in reverse.

[0040] Faraday cage effect prediction and analysis: For complex geometric workpieces such as automotive bumpers, curvature analysis and local surface normal deviation calculation are performed based on 3D point cloud data acquired by a line laser scanner. Sensitive areas of the Faraday cage effect, such as deep grooves, sharp inner corners, narrow gaps, and shading areas, are identified, and the electric field shielding coefficient of each area is quantified. A pre-trained real-time electric field prediction model is called, and combined with the current spray gun position, attitude, and initial electrostatic parameters, the local electric field intensity and coating deposition efficiency of sensitive areas are quickly estimated within 0.05s. The reduction ratio of deposition in the recessed area and the over-deposition risk in the adjacent flange area are quantified and predicted, and an electrostatic parameter adjustment strategy is generated to provide a basis for subsequent active intervention of the electric field.

[0041] S4. Adaptive control parameter generation and execution: Based on the analysis results of step S3, the adaptive control unit automatically generates the optimal spraying parameters and sends them to the execution module to achieve dynamic control without manual intervention. The specific control logic is as follows: Adaptive adjustment of spraying path: Based on the three-dimensional shape and dimensional deviation of the workpiece, the movement trajectory of the spraying robot and the angle of the spray gun are automatically adjusted to ensure that the distance between the spray gun and the workpiece surface is always maintained within the optimal range of 200mm; for the complex curved surface area of ​​the bumper, the spray overlap rate is automatically optimized to 60% to ensure the uniformity of the coating.

[0042] Adaptive adjustment of spraying process parameters: Paint flow rate: The paint flow rate is automatically adjusted to 2L / min according to the degree of surface defects and coating thickness requirements of the workpiece; for defective areas, the flow rate is increased by 30% to 2.6L / min, while the standard flow rate is maintained in non-defective areas to avoid paint waste and increase the paint utilization rate from less than 40% to more than 80%.

[0043] Atomization pressure: The atomization pressure is automatically adjusted to 0.5MPa based on the viscosity of the coating and the ambient temperature and humidity. If the viscosity is too high, the pressure is increased by 15% to 0.575MPa to ensure uniform atomization of the coating. If the humidity is too high, the pressure is increased by 8% to 0.54MPa to avoid coating defects caused by environmental factors.

[0044] Spraying speed: The spraying speed is automatically adjusted to 1.2m / s according to the workpiece size and coating thickness requirements; if a thicker coating is required, the speed is reduced by 25% to 0.9m / s to improve coating uniformity.

[0045] Number of sprays: The number of sprays is automatically adjusted to 2 based on the degree of surface defects and coating thickness requirements of the workpiece; for defective areas, 1 more spray is added to bring the total to 3 to ensure defect repair while avoiding over-spraying.

[0046] Adaptive adjustment of electrostatic parameters: Based on the prediction results of the Faraday cage effect, the electrostatic parameters of the spray gun are dynamically adjusted for different regions: 0.1s before the spray gun enters the sensitive area, the electrostatic voltage is temporarily reduced by 20% to 56kV to weaken the electric field's binding force on charged paint particles, allowing the particles to penetrate into the recessed area by inertia and improve deposition efficiency; when the spray gun leaves the sensitive area or enters the flange area, the electrostatic voltage is immediately restored to 70kV to maintain high paint transfer efficiency, avoid excessive thickness in the flange area, solve the problem of uneven coating thickness caused by the Faraday cage effect, and control the coating thickness deviation between the sensitive area and the reference area within ±5μm.

[0047] Auxiliary equipment adaptive adjustment: Based on environmental data, the system automatically controls the operation of dehumidifiers and temperature control equipment to maintain the workshop temperature and humidity within the optimal range of 22℃ and 50% humidity. Based on coating condition data, the system automatically controls the operation of coating heating / cooling equipment and thinner addition equipment to ensure that the coating condition remains stable at a viscosity of 25 mPa·s and a temperature of 23℃. The response latency of the execution module is less than 0.1 seconds, ensuring that parameter adjustments can keep up with changes in operating conditions in real time. The changeover and debugging cycle for multiple workpieces is shortened by more than 30%, adapting to flexible production needs.

[0048] S5. Real-time detection and feedback optimization of coating quality: During the coating process, the detection module simultaneously performs real-time detection of the coating quality, as detailed below: Real-time inspection: Industrial cameras and image recognition algorithms are used to detect defects such as drips, orange peel, bubbles, pinholes, and particles on the coating surface in real time; multispectral imaging technology is used to detect the coating thickness in real time, and a spectrophotometer is used to detect the color difference to ensure that the coating thickness deviation is ≤±2μm and the color difference ΔE<0.5; for Faraday cage sensitive areas, the sampling frequency is increased to 50Hz to focus on detecting thickness deviation.

[0049] Tiered feedback: Qualified feedback: If the coating quality meets the preset standard, a "qualified" signal is fed back, and the system continues to coat according to the current parameters.

[0050] Minor Defect Feedback: If a minor defect is detected, the defect data is immediately transmitted to the data processing unit. The adaptive control unit quickly adjusts the relevant parameters and repairs the defect in real time without stopping the machine, thus avoiding the generation of a large number of defective products. The first-pass yield rate is increased from 92% to 98.5%, and the defect rate is reduced by 65%.

[0051] Critical Defect Feedback: If a critical defect is detected, a stop signal is immediately sent, the system stops painting, and the operator is prompted to investigate the cause. After the problem is resolved, the system optimizes the parameters based on the defect data and restarts painting.

[0052] Continuous optimization: All test data is stored in the database. Combined with iterative updates of the model based on historical data, the process parameter correlation model and process parameter database are continuously optimized. The Faraday cage effect prediction model and electrostatic parameter adjustment rules are updated simultaneously. For every 500 sets of bumper coating data accumulated, the prediction accuracy is improved by 8%-12%, enabling the system to have self-learning and iterative capabilities, improving the self-learning ability of control accuracy, achieving continuous improvement in coating quality and reduction of defect rate. The quality control efficiency is improved by more than 80% compared with post-inspection.

[0053] S6. Post-coating processing and data archiving: After coating, the workpiece is sent to the drying station. Depending on the coating type, the drying temperature is adaptively controlled to 80℃ and the drying time to 20 minutes. A hot air drying oven is used, with a temperature uniformity of ±2℃, to ensure uniform coating drying and satisfactory adhesion. At the same time, all data from this coating process is archived and stored for ≥3 years. The entire coating process data of the workpiece can be queried through its unique identifier, enabling quality traceability, problem investigation, and facilitating subsequent quality analysis and process optimization.

[0054] Example 2

[0055] Reference Figure 2 The adaptive compensation intelligent industrial coating closed-loop control system of the present invention includes a conveying and pre-processing module, a multimodal sensing module, a data processing unit, an adaptive control unit, an execution module, a detection module, a database and traceability module, and a human-machine interaction module. Each module is connected via an industrial Ethernet with a communication rate of ≥100Mbps to ensure stable and real-time data transmission.

[0056] Conveying and Pre-treatment Module: Includes a conveyor line, a high-pressure airflow cleaning device, and a plasma treatment device, with the following specific configuration: Conveyor line: The conveyor line adopts a belt conveyor with an adjustable conveying speed of 0.1-0.5m / s, a load capacity of 0-500kg, and a positioning accuracy of ±0.05mm. It is used to deliver the workpieces to be coated to each station and can realize the continuous conveying of multiple types of workpieces.

[0057] High-pressure airflow cleaning device: Equipped with an air compressor and airflow nozzles, with 4-6 nozzles evenly distributed on both sides of the conveyor line, used to remove dust and impurities from the surface of workpieces.

[0058] Plasma treatment device: power 500-800W, processing width 100-500mm, processing time 10-20s adjustable, installed above the conveyor line, 50-100mm away from the workpiece surface, used to improve the adhesion of the workpiece surface.

[0059] Multimodal sensing module: Integrates 5 types of sensing devices, all of which communicate bidirectionally with the data processing unit. The specific configuration is as follows: Workpiece sensing unit: includes a line laser scanner with a scanning accuracy of 0.01mm, a scanning speed of 500mm / s, and a scanning range of 50-500mm, as well as an industrial high-resolution camera with a resolution of 1920×1080, a frame rate of 30fps, and a lens focal length of 16-25mm. It is installed above and on both sides of the painting station to collect the three-dimensional shape, surface defects, and position information of the workpiece.

[0060] Paint sensing unit: including viscosity sensor, density meter, flow sensor and temperature sensor, installed on paint supply pipeline and paint storage tank, used to detect paint status in real time.

[0061] Environmental sensing unit: including temperature and humidity sensors, air velocity meter, and particulate matter detector, installed in the painting workshop and evenly distributed around the painting station to collect environmental parameters.

[0062] Equipment sensing unit: includes torque sensor, encoder, vibration sensor, pressure sensor, and electrostatic voltage / current sensor, which are installed on the joints, spray gun, and paint supply pipe of the painting robot, respectively, to collect equipment operating status data.

[0063] Coating sensing unit: including multispectral imager, spectrophotometer, and laser-induced breakdown spectrometer (LIBS), installed behind the coating station to detect coating thickness, color and multilayer coating distribution in real time.

[0064] Data Processing Unit: Serving as the system's "data hub," this unit includes an industrial controller, a data fusion module, a deep learning analysis module, and a predictive electrostatic field adjustment submodule. It communicates bidirectionally with the multimodal sensing module, adaptive control unit, and database. Its specific functions are as follows: Data reception and preprocessing: Receive all data transmitted by the multimodal sensing module, use wavelet denoising algorithm to eliminate data interference, and perform alignment and standardization processing on data of different types and dimensions to ensure data consistency.

[0065] Multi-source data fusion: Employing a weighted average fusion algorithm, it deeply integrates five types of data—workpiece, coating, environment, equipment, and coating layer—to output comprehensive and accurate integrated data, eliminating the limitations of single-source perception.

[0066] Intelligent Analysis: Based on CNN convolutional neural network and iterative model updates based on historical data, combined with a pre-set process parameter database, the fused data is analyzed to determine workpiece adaptability, coating condition, environmental impact, equipment condition and coating quality prediction, and output analysis results and parameter adjustment suggestions.

[0067] Predictive electrostatic field adjustment: For complex geometric workpieces, a 3D reconstruction model of the workpiece is constructed based on 3D point cloud data, sensitive areas of Faraday cage effect are identified, the local electric field distribution and coating deposition efficiency are predicted, and corresponding electrostatic parameter adjustment strategies are generated for the corresponding areas.

[0068] Data transmission: The analysis results are transmitted to the adaptive control unit in real time, and all raw data, analysis data and control parameters are transmitted to the database for storage.

[0069] Adaptive Control Unit: As the "brain" of the system, it communicates bidirectionally with the data processing unit and execution module. It adopts a PLC controller and has a built-in adaptive control algorithm. Its specific functions are as follows: Parameter generation: Based on the analysis results of the data processing unit, the optimal spraying path, spraying process parameters, electrostatic parameters, and auxiliary equipment control parameters are automatically generated.

[0070] Dynamic control: The generated control parameters are sent to the execution module in real time to control the operation of each execution device. At the same time, the system receives feedback data from the execution module and adjusts the parameters in real time to ensure control accuracy.

[0071] Closed-loop feedback: Receives quality inspection data from the detection module, adjusts control parameters based on the inspection results, and achieves closed-loop control of "detection-feedback-adjustment".

[0072] Self-learning optimization: By combining historical data stored in the database, the parameter generation model is continuously optimized through iterative updates based on historical data, thereby improving the accuracy and efficiency of adaptive control and adapting to complex working conditions and various types of workpieces.

[0073] Execution Module: As the "execution organ" of the system, it includes the painting robot, paint supply system, electrostatic generator, environmental control equipment, and drying equipment. It receives control commands from the adaptive control unit, and its specific configuration is as follows: Spraying robot: It adopts a 6-axis industrial robot with a repeatability of ±0.02mm and a load of 5-10kg. It is equipped with a high-precision spray gun and installed on one side of the painting station to execute the spraying path and spraying parameters, and can achieve precise spraying of complex curved surfaces.

[0074] Paint supply system: includes paint storage tank, paint pump, thinner adding device, and stirring device. It receives control commands, adjusts paint flow rate and paint viscosity, and ensures a stable supply of paint.

[0075] Electrostatic generator: It communicates bidirectionally with the adaptive control unit, with a response time of ≤0.05s. It can adjust the electrostatic voltage and current output to the spray gun in real time according to the command, so as to realize active intervention of the electric field.

[0076] Environmental control equipment, including dehumidifiers and constant temperature air conditioners, is used to maintain optimal environmental parameters in the painting workshop.

[0077] Drying equipment: A hot air drying oven is used, with an adjustable drying temperature of 60-120℃ and an adjustable drying time of 10-30 minutes. The temperature uniformity is ±2℃. It is used for drying workpieces after coating to ensure that the coating dries to the required standard.

[0078] The detection module includes a real-time detection unit and an offline detection unit, and communicates bidirectionally with the data processing unit and the adaptive control unit. The specific configuration is as follows: Real-time detection unit: including industrial camera, image recognition module, multispectral imager, spectrophotometer, installed behind the coating station, used to detect coating defects, thickness and color difference in real time during the coating process.

[0079] Offline testing unit: including adhesion tester and hardness tester, used to sample and test the adhesion and hardness of the coating after coating is completed, to ensure that the product quality meets the standards.

[0080] Defect alarm unit: When a serious defect is detected, it will automatically issue an audible and visual alarm and send a stop signal to prompt the operator to handle the situation.

[0081] Database and Traceability Module: Includes an industrial database and a traceability management module, with the following specific functions: Data storage: Stores all relevant data such as workpiece information, multimodal sensing data, control parameters, detection results, and production time, with a storage time of ≥3 years, and supports data backup and recovery.

[0082] Traceability Management: By using the unique identifier of the workpiece, the entire coating process data of the workpiece can be queried, enabling quality traceability and problem investigation.

[0083] Data sharing: Supports integration with enterprise MES and ERP systems to achieve production data sharing, facilitating overall production management and optimization for enterprises.

[0084] Human-computer interaction module: Includes touch screen, operation buttons, and emergency stop switch, used by operators for parameter setting, status monitoring, and fault handling. Specific functions are as follows: Parameter settings: You can manually set the workpiece standard model, coating quality standards, process parameter range, etc., and save and recall parameters.

[0085] Status monitoring: Real-time display of the operating status, sensing data, control parameters, and detection results of each module of the system, making it easy for operators to keep track of the production situation in real time.

[0086] Troubleshooting: When a system malfunctions, the display screen shows the fault type, fault location, and troubleshooting suggestions. Operators can troubleshoot using the operation buttons. The emergency stop switch can quickly stop the machine in an emergency to ensure the safety of equipment and personnel.

[0087] Example 3

[0088] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the aforementioned intelligent industrial coating method with multimodal sensing and adaptive control. The steps include workpiece preprocessing and reference calibration, multimodal sensing data acquisition, multi-source data fusion and intelligent analysis, adaptive control parameter generation and execution, real-time detection and feedback optimization of coating quality, and coating post-processing and data archiving. This method can be widely applied to intelligent control scenarios in the field of industrial coating.

[0089] The above are merely embodiments of this application and are not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

[0090] The above formulas are all derived from software simulation using a large amount of data, and are selected to be close to the actual values. The coefficients in the formulas are set by those skilled in the art based on the actual situation. The above are only preferred embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or changes made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. An adaptive compensation intelligent industrial coating closed-loop control method, characterized in that, Includes the following steps: S1. Workpiece pretreatment and reference calibration: The workpiece to be coated is cleaned and plasma treated, the reference image of the workpiece is collected and compared with the preset standard model, the reference calibration of the workpiece position and posture is completed, and the spraying start position and range are determined. S2. Multimodal sensing data acquisition: Start the multimodal sensing network to synchronously acquire multimodal heterogeneous data of the workpiece, with a sampling frequency of 10-20Hz; S3. Multi-source data fusion and intelligent analysis: First, multi-source heterogeneous data are fused by weighted average to eliminate noise. Then, the fused feature vector is input into the trained CNN model for intelligent analysis of the workpiece. S4. Adaptive control parameter generation and execution: Based on intelligent analysis results, the optimal spraying parameters are automatically generated and sent to the execution module to dynamically adjust the spraying path, spraying process parameters and auxiliary equipment operating status. S5. Real-time detection and feedback optimization of coating quality: Real-time detection of coating quality, graded feedback according to the degree of defect, real-time adjustment of parameters to repair minor defects, triggering shutdown warning for serious defects, and storing all detection data in the database. S6. Post-coating processing and data archiving: After coating is completed, the drying parameters are adaptively adjusted, and the data of the entire coating process is archived to form a traceable production record.

2. The adaptive compensation intelligent industrial coating closed-loop control method according to claim 1, characterized in that, In step S1, a high-pressure airflow of 0.3-0.5MPa is used to remove impurities from the workpiece surface, and a plasma treatment device with a power of 500-800W is used to treat the workpiece surface for 10-20 seconds, with a reference calibration error ≤ ±0.05mm.

3. The adaptive compensation intelligent industrial coating closed-loop control method according to claim 2, characterized in that, In step S2, the multimodal heterogeneous data of the workpiece includes five categories: three-dimensional morphology and surface defect data, coating status data, spraying environment data, equipment operation status data, and coating real-time status data.

4. The adaptive compensation intelligent industrial coating closed-loop control method according to claim 3, characterized in that, The three-dimensional morphology and surface defect data were acquired using a line laser scanner and an industrial high-resolution camera. The line laser scanner had a scanning accuracy of 0.01 mm and a minimum identifiable defect size of 0.1 mm × 0.1 mm. The coating status data included coating viscosity, density, flow rate, and temperature, with the viscosity measurement range being 10-100 mPa·s and an accuracy of ±1 mPa·s. The spraying environment data included workshop temperature and humidity, air velocity, and particulate matter content, with the temperature and humidity measurement range being 0-50℃ and 20%-80%RH. The equipment operating status data included spraying robot joint torque, vibration frequency, spray gun atomization pressure, and electrostatic voltage / current, with the torque measurement range being 0-100 N·m and an accuracy of ±0.5 N·m, and the electrostatic voltage measurement range being 0-100 kV and an accuracy of ±1 kV. The real-time coating status data was acquired using multispectral imaging technology, a spectrophotometer, and laser-induced breakdown spectroscopy, with a coating thickness measurement accuracy of ±2 μm and a color difference accuracy of ΔE < 0.

5.

5. The adaptive compensation intelligent industrial coating closed-loop control method according to claim 1, characterized in that, In step S3, the intelligent analysis specifically includes: adaptability analysis, coating status analysis, environmental impact analysis, equipment status analysis and coating quality prediction, output parameter adjustment suggestions, and data processing delay ≤0.2s; for complex geometric workpieces such as automobile bumpers, a three-dimensional reconstruction model of the workpiece is constructed simultaneously based on three-dimensional point cloud data, the sensitive area of ​​Faraday cage effect is identified, the local electric field distribution and coating deposition efficiency are predicted, and an electrostatic parameter adjustment strategy is generated.

6. The adaptive compensation intelligent industrial coating closed-loop control method according to claim 5, characterized in that, In step S3, the multi-source data fusion adopts a weighted average method, with the weights allocated as follows: workpiece data 0.3, coating data 0.25, environmental data 0.15, equipment data 0.15, and coating data 0.

15. The process parameter database covers spraying strategies for at least 2,000 typical workpieces, including a feature library of Faraday cage effect sensitive areas for complex geometric workpieces such as car bumpers.

7. The adaptive compensation intelligent industrial coating closed-loop control method according to claim 5, characterized in that, In step S5, the specific rules for graded feedback are as follows: if the coating quality meets the standard, a qualified signal is output and the system operates according to the current parameters; if minor defects are detected, including coating thickness deviations exceeding ±5μm in the Faraday cage area, the spraying parameters or electrostatic parameters are adjusted in real time to complete the repair without stopping the machine; if serious defects are detected, including large-area thickness deviations, sagging, and severe color differences, a stop signal is immediately triggered to prompt the operator to troubleshoot the fault.

8. An adaptive compensation intelligent industrial coating closed-loop control system, characterized in that, include: The conveying and pretreatment module is used to complete workpiece conveying, surface cleaning, plasma treatment and reference calibration; The multimodal sensing module integrates the workpiece sensing unit, coating sensing unit, environment sensing unit, equipment sensing unit, and coating sensing unit to simultaneously collect five types of core sensing data. The data processing unit is used to complete multi-source data fusion and intelligent analysis, output parameter adjustment suggestions, and has a data processing delay of ≤0.2s. It has a built-in predictive electrostatic field adjustment submodule, which is used to identify Faraday cage effect sensitive areas, predict electric field distribution and generate electrostatic parameter adjustment strategies. An adaptive control unit is used to generate optimal control parameters and send them to the execution module, with a response time of ≤0.1s; The execution module, including the spraying robot, paint supply system, electrostatic generator, environmental control equipment, and drying equipment, is used to execute control commands to complete spraying and post-processing. The detection module is used to detect coating quality in real time and output feedback signals according to the defect level; The database and traceability module is used to store production data throughout the entire process and supports quality traceability. The human-computer interaction module is used for parameter setting, status monitoring, and fault handling. Each module is connected via an industrial Ethernet network with a communication rate of ≥100Mbps.

9. The adaptive compensation intelligent industrial coating closed-loop control system according to claim 8, characterized in that, The multimodal sensing module includes: a workpiece sensing unit comprising a line laser scanner with a scanning accuracy of 0.01 mm and an industrial high-resolution camera with a resolution of 1920×1080; a coating sensing unit comprising a viscosity sensor, a density meter, a flow sensor, and a temperature sensor installed on the paint supply pipe and the paint storage tank; an environmental sensing unit comprising a temperature and humidity sensor, an air velocity meter, and a particulate matter detector uniformly distributed around the coating station; an equipment sensing unit comprising a torque sensor, an encoder, and a vibration sensor installed on the joints of the spraying robot, as well as a pressure sensor and an electrostatic voltage / current sensor installed on the spray gun and the paint supply pipe; and a coating sensing unit comprising a multispectral imager, a spectrophotometer, and a laser-induced breakdown spectrometer, with a coating thickness measurement accuracy of ±2 μm.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps of the method according to any one of claims 1 to 7.

Images