Multi-station collaborative operation ink-jet printing intelligent production line and control method

Abstract

The invention relates to the technical field of ink-jet printing, in particular to a multi-station collaborative operation ink-jet printing intelligent production line and a control method. According to the technical scheme, the system comprises a multi-station module, a cooperative transmission system, a sensing monitoring system and a central control unit, the multi-station module sequentially comprises a feeding station, a preprocessing station, a printing station, a curing station, a quality detection station and a discharging station, and the cooperative transmission system is composed of a two-way circulating conveying belt and a multi-degree-of-freedom mechanical arm; the sensing monitoring system is used for collecting workpiece, environment and injection state data and transmitting the data to the central control unit. The control method comprises the steps of data collection and environment monitoring, production task modeling, path optimization and task allocation, parallel scheduling and conflict avoidance, defect detection and local compensation, and dynamic feedback and parameter adaptive optimization. Through multi-station parallel operation, bidirectional flexible transmission, real-time defect detection and adaptive scheduling optimization, the efficiency, scheduling and finished product quality of the ink-jet printing production line are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of inkjet printing technology, and in particular to a multi-station collaborative work inkjet printing intelligent production line and control method. BACKGROUND

[0002] The inkjet printing technology is widely used in electronic manufacturing, textile printing, packaging printing and other industries due to its non-contact processing, high resolution and flexibility. However, in the traditional inkjet printing production line, there are several outstanding problems:

[0003] The production cycle is limited by single station. Most inkjet printing production lines adopt serial station arrangement, that is, each station processes the workpiece in a fixed order, and the transmission between stations is completed by a single conveyor. Once a station fails or the processing speed decreases, the entire production cycle will be affected, and even the whole line will be shut down;

[0004] The scheduling and transmission lack flexibility. The scheduling control of existing inkjet printing production lines mostly relies on preset processes, lacks dynamic adjustment ability to real-time state, and is difficult to realize flexible scheduling such as cross-station repair and single insertion operation. At the same time, the one-way transmission structure causes the material to flow in a fixed direction, which is not conducive to quick detour processing in case of production abnormality;

[0005] Defect detection and compensation lag. Traditional quality detection is mostly carried out at the end of production, and after detecting defects, the workpiece often needs to be returned to the previous process, causing production delay and increased energy consumption. And the detection technology has limited recognition ability for complex textures and small defects, and the defect compensation strategy lacks automation and precision;

[0006] Environmental factors are not fully considered in the control model. The inkjet printing process is sensitive to temperature, humidity, air flow rate and other environmental conditions, which will affect the diffusion and solidification quality of ink droplets. The existing system rarely considers environmental parameters in scheduling and path planning, which may cause unstable jetting quality;

[0007] The scheduling optimization lacks adaptive ability. Production parameters and task distribution will change with order type and material characteristics. The existing scheduling system often relies on manual experience adjustment, lacks adaptive optimization mechanism based on historical data, and causes large fluctuations in efficiency in long-period production.

[0008] Therefore, we propose a multi-station collaborative work inkjet printing intelligent production line and control method to solve the existing problems. SUMMARY

[0009] The purpose of the present application is to propose a multi-station collaborative work inkjet printing intelligent production line and control method to solve the problems in the background art.

[0010] To achieve the above object, the present application provides the following technical scheme: a multi-station collaborative work intelligent production line of inkjet printing, comprising a multi-station module, a collaborative transmission system, a sensing and monitoring system and a central control unit, the multi-station module comprises sequentially arranged feeding station, pretreatment station, printing station, curing station, quality detection station and unloading station, each station is connected through a transmission unit which can be independently started and stopped, the transmission unit can realize bidirectional transportation, and each adjacent two stations have an independent driving mechanism to reduce the influence of single point failure on the overall beat;

[0011] The collaborative transmission system comprises a combined transmission structure of bidirectional circulating conveyor belt and multi-degree-of-freedom mechanical arm, wherein the conveyor belt is used for executing main line batch transmission, and the mechanical arm is used for executing cross-station transfer, defect workpiece repair and single insertion operation between different stations;

[0012] The sensing and monitoring system comprises at least one set of high-resolution industrial camera, laser displacement sensor, temperature and humidity sensor and nozzle state sensor, which is used for respectively collecting workpiece surface imaging data, workpiece space position data, environmental condition data and nozzle spraying state data, and transmitting the above data to the central control unit in real time;

[0013] The central control unit comprises a data acquisition module, a scheduling optimization module and an execution control module, the scheduling optimization module calculates the optimal beat of the current batch based on a scheduling optimization formula after receiving the sensing data, and generates an execution sequence and sends it to each station and the transmission system, so as to realize multi-station parallel operation and conflict avoidance.

[0014] Preferably, the scheduling strategy of the central control unit is based on a station load balancing optimization formula:

[0015]

[0016] Wherein, T cycle is the optimal beat time of single batch production, is a set of feasible scheduling schemes, W is a set of stations, t i is the current work remaining time of station i, d i,j is the transmission distance from station i to station j, v t is the average moving speed of the transmission unit, δ i is the station switching time.

[0017] Preferably, the printing station is provided with a nozzle array, and the nozzle arrangement satisfies a minimum spraying interference distance constraint formula:

[0018]

[0019] Wherein, D minis the minimum center distance between the nozzles, k is an empirical coefficient, and σ x ,σ y are the standard deviations of the distribution of the spray points in the X and Y directions, respectively.

[0020] Preferably, the end effector of the mechanical arm is provided with a dual-mode clamping function, including a negative pressure suction module and a mechanical gripper module, the negative pressure suction module is used for flat workpiece handling, and the mechanical gripper module is used for irregular-shaped workpiece handling, and the sensing and monitoring system further includes a nozzle flow detection unit, which acquires the nozzle spray flow by combining micro-weighing and optical drop speed measurement, and is used for predictive maintenance before spray abnormalities occur.

[0021] An intelligent control method for multi-station collaborative work of inkjet printing, comprising the following steps:

[0022] Step 1: Data acquisition and environment monitoring: Through the sensing and monitoring system arranged on each station and the transmission path, real-time production state information and environmental parameters are collected, the information includes real-time three-dimensional position coordinates of the workpiece in the production line; the working state identification of each station, i.e. four states of idle, running, waiting and failure; printing progress percentage and nozzle spraying state parameters, i.e. spraying frequency and ink drop volume; environmental condition parameters, i.e. temperature, humidity and air flow rate, for compensating the inkjet diffusion model;

[0023] Step 2: Production task modeling: after receiving the information transmitted by the data acquisition module, the central control unit converts the current batch production task into a mathematical optimization model, and the optimal cycle time T cycle is minimized as the objective function, and constraint conditions are applied, including station processing capacity constraint, path conflict constraint and environmental stability constraint;

[0024] Step 3: Path optimization and task allocation: based on the station position matrix P and the distance matrix D, a hybrid algorithm of improved dynamic programming combined with heuristic search is used to solve the nozzle movement path π * , considering the spraying interference cost C i,j in path solving, so as to obtain a movement trajectory that is both shortest and low in interference;

[0025] Step 4: Parallel scheduling and conflict avoidance: based on the optimization result, the central control unit generates an execution sequence table for each station and allocates priority to each task, ensuring that the transmission between any two stations will not cause collision and waiting deadlock during scheduling, and conflict avoidance is realized through time slice allocation and safety interval control;

[0026] Step 5 defect detection and local compensation: the quality detection station acquires the printed surface image using a multi-spectral industrial camera, the central control unit calls the defect detection model to calculate the quality score Q, if the quality score is less than the quality threshold, the workpiece is re-assigned to the idle printing station to perform local compensation spraying task, and re-enters the detection process;

[0027] Step 6 dynamic feedback and parameter adaptive optimization: the control system stores historical task data in the database, periodically calls the parameter optimization module, and uses gradient descent or genetic algorithm method to update the scheduling parameter set Θ to continuously improve the production cycle stability and defect repair success rate.

[0028] Preferably, the quality score Q is calculated by the following weighted comprehensive formula:

[0029]

[0030] Wherein, represents the color error rate, C meas is the measured color value, C ref is the reference color value, represents the position deviation rate, ΔL is the deviation length, L ref is the reference length, represents the defect area ratio, α, β, γ are the adjustable weight coefficients of the three indicators, which are dynamically set by the central control unit according to the task type.

[0031] Preferably, the objective function of the path optimization is:

[0032]

[0033] Wherein, π is the nozzle movement path, Π is the feasible path set, d i,j is the path distance of the nozzle moving from position i to j, v p is the average moving speed of the nozzle, C i,j is the spraying interference cost coefficient, which is calculated according to the spraying density and droplet diffusion radius, λ is the interference cost weight factor, which is dynamically adjusted according to different materials and printing accuracy requirements.

[0034] Preferably, the defect detection step includes using a multi-scale convolutional neural network (MSCNN) to extract the texture and edge features of the printed surface defects, inputting the extracted features into a support vector machine (SVM) classifier, and outputting the defect category and severity level according to the defect feature vector and the preset classification boundary. For defects with a severity level higher than the preset threshold, automatically generate a local compensation spraying task instruction, and insert the task into the high-priority queue of the current scheduling sequence.

[0035] Preferably, in the dynamic feedback optimization step, the central control unit dynamically adjusts the scheduling parameter set Θ according to the cumulative production task data set The scheduling parameter set Θ is updated by using the gradient descent method, and the update rule is:

[0036]

[0037] Wherein η is the learning rate, which is automatically adjusted by the system according to the production stability, The beat loss function is defined as the weighted sum of the square difference between the actual beat and the target beat:

[0038]

[0039] Wherein ω1, ω2 are the weights of the two parts of the loss, and DefectRate is the defect rate.

[0040] Compared with the prior art, the beneficial effects of the present application are as follows:

[0041] Multi-station parallel and bidirectional transmission structure: adopt multi-station modules such as feeding, pretreatment, printing, curing, quality detection, and unloading, and connect through bidirectional transmission units that can be independently started and stopped, any adjacent station has an independent driving mechanism, reducing the impact of single point failure on the whole line, realizing multi-station parallel operation, and the collaborative transmission system combines bidirectional circulating conveyor belt with multi-degree-of-freedom mechanical arm, which can not only complete batch linear transmission, but also realize cross-station transfer, defect repair and single insertion processing, greatly improving the flexible production capacity;

[0042] Multi-source sensing monitoring and data-driven scheduling: integrate industrial cameras, laser displacement sensors, temperature and humidity sensors, nozzle state sensors and nozzle flow detection units to comprehensively collect workpiece state, spatial position, environmental parameters and spraying performance, realize real-time monitoring of the whole production process, and the central control unit calculates the optimal beat of the batch based on real-time data and optimization formula, generates a dynamic execution sequence, and ensures load balancing and conflict avoidance of the station;

[0043] Spraying path optimization and interference control: an algorithm combining improved dynamic programming and heuristic search is adopted, which considers the spraying interference cost and path length to generate a short and low interference nozzle motion trajectory, effectively reducing the droplet interference and imaging defect rate;

[0044] Defect detection and local compensation automation: the quality detection station is based on multispectral imaging and multi-scale convolutional neural network (MSCNN) combined with support vector machine (SVM) classification to realize accurate identification and classification of texture and edge defects, the system can automatically generate local compensation spraying tasks and insert high priority execution in the scheduling sequence, significantly improving the repair efficiency;

[0045] Environmental Adaptability and Long-Term Optimization Capability: Environmental compensation factors such as temperature, humidity, and airflow velocity are introduced into the jet diffusion model and scheduling strategy to improve printing stability under different environments. The central control unit uses gradient descent or genetic algorithm to continuously optimize the scheduling parameters, thereby achieving long-term improvement in production cycle stability and defect repair success rate.

[0046] Compared with traditional serial inkjet printing production lines, this invention has significant improvements in production cycle time, flexible scheduling, defect compensation efficiency, and finished product qualification rate, and is especially suitable for inkjet printing tasks with multiple varieties, small batches, and high precision. Attached Figure Description

[0047] Figure 1 This is a schematic diagram of the steps and methods of the present invention. Detailed Implementation

[0048] 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.

[0049] Example 1

[0050] like Figure 1 As shown, the present invention proposes an intelligent inkjet printing production line with multi-station collaborative operation, which includes a multi-station module, a collaborative transmission system, a sensing and monitoring system, and a central control unit.

[0051] The multi-station module includes a loading station, a pre-processing station, a printing station, a curing station, a quality inspection station, and an unloading station arranged in sequence. Each station is connected by a transmission unit that can be started and stopped independently. The transmission unit can realize bidirectional transportation, and each pair of adjacent stations has an independent drive mechanism, thereby reducing the impact of single-point failure on the overall cycle time.

[0052] In this embodiment, the loading station is used to automatically grab, position and stack raw materials or semi-finished products, and can realize workpiece identification through barcode scanning or RFID reading.

[0053] The pretreatment station can perform processes such as cleaning, surface roughening, and primer spraying to improve the adhesion and uniformity of subsequent inkjet printing.

[0054] The printing station is equipped with a multi-row printhead array and an ink circulation and temperature control system to ensure printout stability;

[0055] The curing station can select UV curing, hot air curing, or a combination curing method depending on the type of ink used;

[0056] The quality detection station can perform size measurement, surface defect detection, color difference analysis, and other quality evaluations on the workpiece;

[0057] The unloading station is used to automatically collect qualified products or classify unqualified products into a repair channel.

[0058] The cooperative transmission system is composed of a bidirectional circulating conveyor belt and a multi-degree-of-freedom mechanical arm, which form a combined transmission structure. The conveyor belt is responsible for main line batch transmission; the mechanical arm can perform cross-station transfer, defect workpiece repair, and single insertion operations between different stations to enhance system flexibility.

[0059] The conveyor belt adopts a modular link structure, which facilitates quick replacement or partial maintenance. The mechanical arm is a six-degree-of-freedom structure with end force control function, which can automatically adjust the clamping force during handling to avoid damage to the workpiece. The motion trajectory of the mechanical arm can be adjusted in real time by the central control unit, thereby realizing flexible scheduling across batches and stations.

[0060] The sensing and monitoring system includes at least one set of high-resolution industrial cameras, laser displacement sensors, temperature and humidity sensors, and nozzle state sensors, which are used to collect workpiece surface imaging data, workpiece spatial position data, environmental condition data, and nozzle spraying state data, respectively. These data are transmitted in real time to the central control unit to provide data support for subsequent scheduling optimization and quality detection.

[0061] The central control unit includes a data acquisition module, a scheduling optimization module, and an execution control module. When the scheduling optimization module receives the sensing data, it calculates the optimal cycle time for the current batch based on the scheduling optimization formula and generates an execution sequence to send to each station and the transmission system, realizing multi-station parallel operation and conflict avoidance.

[0062] Its scheduling strategy is based on the work station load balancing optimization formula:

[0063]

[0064] Where,

[0065] T cycle is the optimal cycle time for single batch production,

[0066] is the set of feasible scheduling schemes,

[0067] W is the set of stations,

[0068] t i is the current work remaining time of station i,

[0069] d i,j is the transmission distance from station i to station j,

[0070] v t is the average moving speed of the transport unit,

[0071] δ i is the station switching time.

[0072] This formula realizes the dynamic adjustment of the production rhythm by iterative calculation of real-time station state data. When it detects that a certain station has excessively high load or abnormal downtime, the system will automatically recalculate the scheduling scheme and reduce the fluctuation of the overall production rhythm through strategies such as inserting orders and cross-station handling.

[0073] In the printing station, there is an array of nozzles, and the nozzle arrangement satisfies the minimum ejection interference distance constraint formula:

[0074]

[0075] wherein,

[0076] D min is the minimum center distance between nozzles,

[0077] k is an empirical coefficient,

[0078] σ x ,σ y are the standard deviations of the distribution of the ejection points in the X and Y directions, respectively.

[0079] The layout of the nozzle array not only considers the interference distance but also optimizes it in combination with the ink droplet diffusion model to ensure high precision of the ink droplet landing position even at high-speed ejection.

[0080] The nozzle array can use a partitioned control method, which activates only specific nozzles when local compensation printing is performed, thereby reducing energy consumption and prolonging the service life of the nozzles.

[0081] In addition, the end effector of the mechanical arm has a dual-mode clamping function, including a negative pressure suction module (for flat workpiece handling) and a mechanical gripper module (for irregular workpiece handling).

[0082] The negative pressure suction module has a built-in vacuum generator that can adapt to workpieces of different materials and thicknesses by adjusting the suction force; the mechanical gripper module uses flexible cladding materials to avoid scratching or deforming the workpiece surface during clamping.

[0083] The sensing and monitoring system also includes a nozzle flow detection unit that uses a combination of micro-weighing and optical drop speed measurement to obtain the nozzle ejection flow, so as to perform predictive maintenance before ejection abnormalities occur.

[0084] The trace and speed of the ink drop are captured by a high-speed camera to determine the stability of the ejection, and when a deviation in the ejection beyond a set threshold is detected, the system automatically triggers a cleaning, pressurizing or nozzle replacement operation.

[0085] Embodiment Two

[0086] As Figure 1 As shown in the figure, the present application proposes a multi-station collaborative operation intelligent control method for inkjet printing. The control method proposed in this embodiment includes the following steps:

[0087] Step 1: Data acquisition and environment monitoring

[0088] Real-time production state information and environmental parameters are collected by the sensing monitoring system on each station and the transmission path, including:

[0089] Real-time three-dimensional position coordinates of the workpiece in the production line;

[0090] Station operation status (idle, running, waiting, fault);

[0091] Printing progress percentage and printhead ejection state (ejection frequency, ink drop volume);

[0092] Environmental conditions (temperature, humidity, air flow rate) for compensating the inkjet diffusion model.

[0093] The raw data collected by the sensors are preprocessed by the edge computing nodes, including noise filtering, feature extraction and outlier rejection, to reduce the data load of the central control unit and improve the response speed.

[0094] Step 2: Production task modeling

[0095] After receiving the above data, the central control unit converts the current batch production task into a mathematical optimization model, with the minimum optimal cycle time of single batch production as the objective function, and with the constraints of station processing capacity, path conflict and environmental stability.

[0096] Step 3: Path optimization and task assignment

[0097] Based on the station position matrix and distance matrix, a hybrid algorithm of improved dynamic programming combined with heuristic search is used to solve the printhead motion path. The objective function is:

[0098]

[0099] Where,

[0100] π is the printhead motion path,

[0101] Π is the set of feasible paths,

[0102] d i,j is the path distance of the nozzle moving from position i to j,

[0103] v p is the average moving speed of the nozzle,

[0104] C i,j is the jet interference cost coefficient, calculated according to the jet density and the ink drop diffusion radius,

[0105] λ is the interference cost weight factor, dynamically adjusted according to different materials and printing accuracy requirements.

[0106] Step 4: Parallel scheduling and conflict avoidance

[0107] Generate execution sequence table for each station, assign priority to tasks, ensure that transmission between any two stations does not produce collision or deadlock, and avoid conflict through time slice and safety interval control.

[0108] Step 5: Defect detection and local compensation

[0109] The quality detection station uses a multi-spectral industrial camera to obtain images, calls a multi-scale convolutional neural network (MSCNN) to extract defect features, and classifies them through a support vector machine (SVM), outputting the defect category and severity. If the severity exceeds the threshold, generate a local compensation jet task and insert it into the high-priority queue.

[0110] The quality score is calculated as follows:

[0111]

[0112] where,

[0113] represents the color error rate,

[0114] C meas is the measured color value,

[0115] C ref is the reference color value,

[0116] represents the position deviation rate,

[0117] ΔL is the deviation length,

[0118] L ref is the reference length,

[0119] represents the defect area proportion,

[0120] α, β, γ are adjustable weight coefficients of the three indicators, dynamically set by the central control unit according to the task type.

[0121] Step 6: Dynamic feedback and parameter adaptive optimization

[0122] The system updates the scheduling parameter set according to the cumulative task data set using the gradient descent method, and the update rule is:

[0123]

[0124] Wherein,

[0125] η is the learning rate, which is automatically adjusted by the system according to the production stability,

[0126] is the beat loss function, defined as the weighted sum of the squared difference between the actual beat and the target beat:

[0127]

[0128] Wherein,

[0129] ω1, ω2 are the weights of the two parts of the loss,

[0130] DefectRate is the defect rate.

[0131] Through the above closed-loop optimization mechanism, the embodiment realizes the stable control of the production beat and the improvement of the defect repair success rate.

[0132] The above specific embodiments are only several preferred embodiments of the present application, based on the technical solutions of the present application and the related inspirations of the above embodiments, those skilled in the art can make various alternative improvements and combinations to the above specific embodiments.

[0133] It is obvious for those skilled in the art that the present application is not limited to the details of the above exemplary embodiments, and the present application can be realized in other specific forms without departing from the spirit or essential characteristics of the present application. Therefore, from any point of view, the embodiments should be regarded as exemplary and non-limiting, the scope of the present application is defined by the appended claims rather than the above description, therefore all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be included in the present application.

Claims

1. A multi-station collaborative work intelligent production line of inkjet printing, comprising a multi-station module, a collaborative transmission system, a sensing and monitoring system and a central control unit, characterized in that: The multi-station module comprises, in sequence, a feeding station, a pretreatment station, a printing station, a curing station, a quality detection station and a discharging station, each station being connected through a transmission unit that can be independently started and stopped, the transmission unit being capable of bidirectional transportation, and each of any two adjacent stations having an independent driving mechanism to reduce the influence of single-point failure on the overall beat; The cooperative transmission system comprises a combined transmission structure of a bidirectional circulating conveyor belt and a multi-degree-of-freedom mechanical arm, wherein the conveyor belt is used for performing main line batch transportation, and the mechanical arm is used for performing cross-station transfer, defect workpiece repair and single insertion operation between different stations; The sensing and monitoring system comprises at least one set of high-resolution industrial cameras, laser displacement sensors, temperature and humidity sensors and nozzle state sensors, which are used to respectively collect workpiece surface imaging data, workpiece spatial position data, environmental condition data and nozzle spraying state data, and transmit the above data to a central control unit in real time; The central control unit comprises a data acquisition module, a scheduling optimization module and an execution control module, the scheduling optimization module, after receiving the sensing data, calculates the optimal beat of the current batch based on a scheduling optimization formula, and generates an execution sequence to be sent to each station and the transmission system, so as to realize multi-station parallel operation and conflict avoidance. 2.The multi-station collaborative operation inkjet printing intelligent production line according to claim 1, characterized in that: The scheduling strategy of the central control unit is based on a station load balancing optimization formula: where T cycle is the optimal tact time for single batch production, is the set of feasible scheduling schemes, W is the set of workstations, t i is the remaining time for the current job at workstation i, d i,j is the distance from workstation i to workstation j, v t is the average moving speed of the transport unit, δ i is the workstation switching time. 3.The multi-station collaborative operation inkjet printing intelligent production line according to claim 1, characterized in that: The printing station is provided with a nozzle array, and the nozzle arrangement satisfies a minimum spraying interference distance constraint formula: where D min is the minimum center distance between the nozzles, k is an empirical coefficient, σ x ,σ y are the standard deviations of the distribution of the spray points in the X and Y directions, respectively. 4.The multi-station collaborative operation inkjet printing intelligent production line according to claim 1, characterized in that: The end effector of the mechanical arm has a dual-mode clamping function, including a negative pressure adsorption module and a mechanical gripper module, the negative pressure adsorption module being used for planar workpiece handling, and the mechanical gripper module being used for handling irregularly shaped workpieces, and the sensing and monitoring system further comprises a nozzle flow detection unit, which acquires the nozzle spraying flow by combining micro-weighing and optical drop speed measurement, and is used for predictive maintenance before spraying anomalies occur.

5. A multi-station collaborative work intelligent control method for inkjet printing, characterized in that, The method comprises the following steps: Step 1: Data acquisition and environmental monitoring: real-time acquisition of production state information and environmental parameters through the sensing and monitoring system arranged at each station and the transmission path, the information including real-time three-dimensional position coordinates of the workpiece in the production line, work station state identifiers, i.e. four states of idle, running, waiting and failure, printing progress percentage, nozzle spraying state parameters, i.e. spraying frequency and ink drop volume, and environmental condition parameters, i.e. temperature, humidity and air flow rate, for compensating the inkjet diffusion model; Step 4: Parallel scheduling and conflict avoidance: the central control unit generates an execution sequence table for each station based on the optimization result, and assigns a priority to each task, and ensures that the transmission between any two stations will not cause collision and waiting deadlock in the scheduling process, and conflict avoidance is realized through time slice allocation and safety interval control; Step 2 production task modeling: after receiving the information transmitted by the data acquisition module, the central control unit converts the current batch production task into a mathematical optimization model, so as to obtain the optimal cycle time T of single batch production cycle The minimum is the objective function, and the constraint conditions include the work station processing capacity constraint, the path conflict constraint and the environmental stability constraint; Step 3 Path optimization and task assignment: Based on the position matrix P and the distance matrix D, a hybrid algorithm of improved dynamic programming combined with heuristic search is used to solve the nozzle movement path π * In the path solving, the spray interference cost C is considered i,j So as to obtain the shortest and low interference movement trajectory; Step 5: Defect detection and local compensation: the quality detection station acquires a printing surface image by using a multi-spectral industrial camera, and the central control unit calls a defect detection model to calculate a quality score Q, and if the quality score is less than a quality threshold, the workpiece is re-assigned to an idle printing station to perform a local compensation spraying task, and re-enters the detection process; ​ Step 6 Dynamic feedback and parameter adaptive optimization: the control system stores the historical task data in the database, periodically calls the parameter optimization module, and uses the gradient descent or genetic algorithm method to update the scheduling parameter set Θ to continuously improve the production cycle stability and defect repair success rate.

6. The multi-station collaborative operation intelligent control method for inkjet printing according to claim 5, characterized in that: The quality score Q is calculated by the following weighted comprehensive formula: wherein, represents the color error rate, C meas is the measured color value, C ref is the reference color value, represents the position deviation rate, ΔL is the deviation length, L ref is the reference length, represents the defect area ratio, α, β, and γ are adjustable weight coefficients of the three indicators, which are dynamically set by the central control unit according to the task type.

7. The multi-station collaborative operation intelligent control method for inkjet printing according to claim 5, characterized in that: The objective function of the path optimization is: where v is the motion path of the nozzle, Π is the set of feasible paths, d i,j is the path distance of the nozzle from position i to j, v p is the average moving speed of the nozzle, C i,j is the cost coefficient of the jet interference, which is calculated according to the jet density and the droplet diffusion radius, and λ is the interference cost weight factor, which is dynamically adjusted according to different materials and printing accuracy requirements. 8.The multi-station collaborative operation intelligent control method of inkjet printing according to claim 5, characterized in that: The defect detection step includes using a multi-scale convolutional neural network (MSCNN) to extract texture and edge features of the printing surface defects, inputting the extracted features into a support vector machine (SVM) classifier, outputting the defect category and severity level according to the defect feature vector and the pre-set classification boundary, automatically generating a local compensation jet task instruction for defects with a severity level higher than a pre-set threshold, and inserting the task into a high-priority queue of the current scheduling sequence. 9.The multi-station collaborative operation intelligent control method of inkjet printing according to claim 5, characterized in that: In the dynamic feedback optimization step, the central control unit updates the scheduling parameter set Θ according to the cumulative production task data set The scheduling parameter set Θ is updated by using the gradient descent method, and the update rule is: where η is the learning rate, which is automatically adjusted by the system according to the production stability, is the beat loss function, defined as the weighted sum of squared differences between the actual beats and the target beats: Where ω1, ω2 are the weights of the two parts of the loss, and DefectRate is the defect rate.

Images