A printed matter quality traceability and defect prediction system based on big data

Abstract

The application discloses a printing quality traceability and defect prediction system based on big data, and relates to the technical fields of printing quality detection and industrial automation control.The system comprises a data acquisition module, a traceability analysis module and an adaptive control module.The data acquisition module is used for extracting a spindle clock and emitting a synchronous trigger signal, and synchronously locking a visual image and a bottom layer mechanical parameter.The traceability analysis module is used for analyzing the spatial coordinates of film printing defects and establishing a reverse mapping matrix, thereby providing a basis for defect root cause tracing.The adaptive control module is used for real-time prediction and dynamic adjustment of mechanical compensation parameters of a corresponding color group according to the chromatic aberration decay gradient of a continuous film substrate.The data acquisition module and the adaptive control module are electrically connected with the traceability analysis module.The application solves the problems of difficult defect root cause tracing and lack of dynamic prediction in high-speed film screen printing.

Description

Technical Field

[0001] This invention relates to the field of printing quality inspection and industrial automation control technology, specifically a big data-based system for tracing the source of printed materials and predicting defects. Background Technology

[0002] In the field of IMD / IML film screen printing and processing for high-end home appliances (such as smart control panels for washing machines), machine vision-based surface defect detection systems serve as crucial quality control infrastructure, undertaking important tasks such as detecting misregistration, ink leakage, and color deviation. However, traditional film printing quality control systems have revealed many serious problems in actual operation, greatly limiting the energy-saving and consumption-reducing capabilities and zero-waste management capabilities of production lines.

[0003] Traditional printing quality inspection systems lack the ability to align underlying data in both time and space, and to trace root causes. In high-speed continuous film printing, the vision inspection system and the programmable logic controller (PLC) at the mechanical execution layer typically operate in two independent time domains. When the vision camera detects an image defect at the feeding end, the system can only output the two-dimensional geometric coordinates of the defect on the film, but it cannot know the actual mechanical parameters of the substrate as it passed through each upstream screen printing ink group, such as the doctor blade pre-press pressure, screen spacing, or transient ink viscosity. This long-term disconnect between visual data and underlying mechanical data means that the system is unable to perform pixel-level physical root cause tracing after a quality problem is detected, often leaving maintenance personnel to rely on blind, experience-based troubleshooting.

[0004] In existing technologies, mechanisms relying on static threshold over-limit alarms suffer from severe passive lag. Quality degradation in film printing production lines is typically a slow, non-linear drift process accompanied by squeegee wear, screen tension fatigue, or ink evaporation. Current vision systems often only trigger alarms when defects exceed the national standard tolerance limit. Due to this delayed, passive detection mechanism, by the time the system triggers an alarm and manual intervention is required to stop the machine, a large amount of continuous waste has already been generated on the production line, resulting in expensive waste of film materials and energy.

[0005] Furthermore, traditional detection systems employ relatively simple control methods and lack intelligent adaptive prediction strategies based on continuous quality degradation gradients. During continuous operation, the system cannot extract the drift rate and degradation acceleration of color difference in real time, making it difficult to calculate the optimal mechanical feedforward compensation before defects actually occur, and it cannot automatically perform closed-loop correction. Therefore, designing a big data-based system for traceability and defect prediction of printed materials that can achieve hard synchronization of heterogeneous data, accurately trace the physical root cause, and adaptively predict and feedforward compensate is of significant practical importance. Summary of the Invention

[0006] The purpose of this invention is to provide a big data-based system for tracing the quality of printed materials and predicting defects, in order to solve the problems mentioned in the background art.

[0007] To address the aforementioned technical problems, this invention provides the following technical solution: a big data-based system for tracing the source of printed materials and predicting defects, comprising a data acquisition module, a source analysis module, and an adaptive control module. The data acquisition module is used to extract the spindle clock and transmit a synchronous trigger signal, synchronously locking the visual image with the underlying mechanical parameters. The source analysis module is used to analyze the spatial coordinates of defects in thin-film printed materials and establish a reverse mapping matrix, providing a basis for tracing the root causes of defects. The adaptive control module is used to predict and dynamically adjust the mechanical compensation parameters of the corresponding color group in real time based on the color difference decay gradient of the continuous thin-film substrate. Both the data acquisition module and the adaptive control module are electrically connected to the source analysis module.

[0008] According to the above technical solution, the data acquisition module includes a clock reference submodule, an image acquisition submodule, and a parameter latching submodule. The clock reference submodule further includes an incremental spindle encoder, a high-speed counter, and a clock frequency divider circuit. The clock reference submodule is used to generate and transmit a global pulse signal characterizing the feed length. The image acquisition submodule further includes a linear CCD camera, a light source strobe controller, and an image acquisition card. The image acquisition submodule is used to receive a trigger signal at the feeding end and capture a two-dimensional surface quality image of the thin film substrate. The parameter latching submodule further includes an industrial Ethernet parser, a direct memory access controller, and a PLC underlying data buffer. The parameter latching submodule is used to synchronously capture the ink viscosity, screen spacing, doctor blade pre-press pressure, and oven temperature data of each printing color group under the same pulse timestamp.

[0009] According to the above technical solution, the source tracing analysis module includes a spacing analysis submodule, a reverse deduction submodule, a data addressing submodule, and a feature comparison submodule. The spacing analysis submodule is used to obtain the absolute physical material path from each printing color group to the visual camera at the unloading end. The reverse deduction submodule is used to calculate the specific timestamp of the current image feature point passing through the historical color group based on the current speed and physical material path. The data addressing submodule is used to use the historical timestamp to address and match the transient mechanical parameters that induce defects in the underlying database. The feature comparison submodule is used to match and analyze the extracted transient mechanical parameters with the cause threshold of historical typical defects.

[0010] According to the above technical solution, the adaptive control module includes a gradient monitoring submodule, a prediction output submodule, and an actuator control submodule. The gradient monitoring submodule is used to monitor the color difference offset and overprinting deviation vector change of the continuous film substrate in real time. The prediction output submodule is used to formulate the evaluation result of the remaining safe print volume and the corresponding feedforward compensation strategy based on the deviation vector change. The actuator control submodule is used to control the operation of the production line by adjusting the tension torque of the servo motor and the execution step parameters of the pre-printing pressure of the bottom blade.

[0011] According to the above technical solution, the operation method of the data acquisition module includes the following steps: Step S1: The incremental spindle encoder in the clock reference submodule acquires the high-frequency rotation pulses of the printing press main drive shaft in real time. After processing by the high-speed counter and clock frequency divider circuit, the pulses are generated to correspond to the physical feed length. A global clock pulse sequence that exhibits a strictly linear mapping relationship; Step S2: The global clock pulse sequence is transmitted in parallel to the image acquisition submodule and the parameter latch submodule. When the pulse counter reaches the preset single-area detection interval, a synchronization interrupt signal is triggered. Step S3: After the linear CCD camera of the image acquisition submodule captures the synchronization interrupt signal, it instantly exposes the image and, in conjunction with the light source strobe controller, completes the line frequency scan of the current image area to extract the absolute length coordinates of the current image area on the global feed axis. ; Step S4: Within the same hardware clock cycle, the parametric latch submodule bypasses the CPU through the direct memory access controller, forcibly reads and latches the ink viscosity of each printing color group temporarily stored in the programmable logic controller. Pre-printing pressure of the scraper Oven temperature and transient vehicle speed and compare it with the absolute length coordinates Bind them with the same hardware timestamp, encapsulate them into heterogeneous synchronous data frames and store them in the underlying cache.

[0012] According to the above technical solution, the operation method of the source tracing analysis module includes the following steps: Step A1: During continuous film printing operations, input the basic topology parameters of the current production line, including parameters from the first... The precise physical feed length between the center of the screen printing line of each printing color group and the vertical projection point of the optical axis of the linear CCD camera lens at the feeding end. ,in This represents the total number of color groups on the current production line. Step A2: The image acquisition submodule performs pattern recognition on the acquired two-dimensional surface quality image. When a defect is detected in a certain area, the local coordinates of the defect image within the image area are extracted. And combined with the absolute length coordinates of the image area. The precise physical mileage calibration point at the location of the defect feature was calculated. ; Step A3: After calibrating the current defect mileage location, obtain the real-time time series of equipment speed recorded by the data acquisition module. Then, using the reverse derivation submodule, through integral equations... Calculate and obtain the calibration point of the defect feature at a historical moment when it crosses the first... Time difference during the printing of each color group's printing line ,in It is a velocity distribution function with respect to the material feed length; Step A4: Based on the current absolute time when the defect was detected. Combined with the calculated time difference Calculate the time the defective area experiences during the [number]th [period]. Exact historical transient timestamps during the printing of color units Repeat steps A3 and A4 until the set of historical timestamps of all associated color groups for the defective area is obtained. ; Step A5: The data addressing submodule uses the historical timestamp set. The pointer address offset is precisely extracted from the heterogeneous synchronized data frames in the underlying cache. The set of transient mechanical parameters at the corresponding moment; Step A6: Identify the extracted set of transient mechanical parameters, analyze the abnormal characteristics of adjacent printing color groups after parameter changes such as doctor blade pre-printing pressure fluctuation, ink viscosity mutation or temperature step change, and match the real physical root cause that induces the image defect in the feature comparison submodule. Step A7: Use the source analysis module to output a complete quality source tracing report that includes the historical status of each color group, root cause mechanical parameters, and the current defect mapping.

[0013] According to the above technical solution, step A6 further includes: Step A61: Establish a database of root cause features of thin film printing defects. Record the abnormal fluctuations of mechanical parameters of the corresponding color groups when various typical surface image defects occur in history into the feature database. The abnormal fluctuations include: the drop amplitude of doctor blade pre-printing pressure, the slope of ink viscosity decay, and the step variance of oven temperature. Step A62: Perform time-domain waveform analysis on the set of transient mechanical parameters extracted in step A5 to extract the actual abnormal fluctuation characteristics of the current production line parameter time series; Step A63: Based on the extracted actual abnormal fluctuation characteristics, perform cosine similarity matching calculation in the thin film printing defect root cause feature database; Step A64: Filter out feature similarity matches that reach a preset threshold. The historical typical abnormal fluctuation state is directly identified as the real physical root cause of the current image defect, and the specific printing color group where the physical root cause occurred is accurately located.

[0014] According to the above technical solution, the operation method of the adaptive control module includes the following steps: Step B1: In continuous film printing operations, the gradient monitoring submodule opens a data stream containing continuous surface quality images acquired at the feeding end. A sliding time window for each substrate print sheet; Step B2: Continuously extract the current color difference value of each substrate in the same physical calibration area within the sliding time window. Or overprinting deviation vector; Step B3: Calculate the displacement derivative of the color difference value between consecutive substrates using the first-order backward difference algorithm to obtain the first-order color difference drift rate of the current batch of film printing quality. ; Step B4: Calculate the rate derivative of the first-order chromatic aberration drift rate using a second-order backward difference algorithm to obtain the second-order chromatic aberration decay acceleration characterizing the quality deterioration trend. ; Step B5: The prediction output submodule monitors the first-order chromatic aberration drift rate in real time. Acceleration of second-order color difference decay When the numerical changes of both values ​​are determined to exceed the static anti-shake dead zone, the feedforward prediction and dynamic compensation pipeline is activated.

[0015] According to the above technical solution, the specific execution steps for starting the feedforward prediction and dynamic compensation pipeline in step B5 are as follows: extract the current color difference value. Combined with the calculated first-order chromatic aberration drift rate Acceleration of second-order color difference decay Solve for the optimal feedforward mechanical compensation amount Its calculation expression is: ; in, The number of adjustment steps for the doctor blade pre-press pressure or tension roller output of the target printing color group calculated in real time; The basic control output is used to maintain the current printing status; The first-order drift rate weighting coefficient; The second-order decay acceleration weighting coefficient; This is the absolute color difference weighting coefficient; The starting drift rate that triggers dynamic compensation; and These are the maximum color difference drift rate and maximum decay acceleration limits set by the system, respectively. The target color difference value is the standard reference benchmark. This represents the lower limit extreme value of the national standard tolerance. The prediction output submodule is based on the calculation results Encapsulate and generate feedforward control instructions.

[0016] According to the above technical solution, the step of the actuator control submodule adjusting the underlying production line action according to the feedforward control command further includes: Step B6: Transmit the packaged feedforward mechanical compensation amount via industrial fieldbus. The defect is sent to the corresponding underlying actuator before it exceeds the national standard tolerance limit. Before becoming scrap, perform dynamic correction actions with tiny steps; Step B7: The system's underlying processor uses a kinematic approximation model to calculate the remaining safe print volume in real time. The calculation formula is: ; Step B8: When the calculated optimal feedforward mechanical compensation amount is detected... Exceeding the physical adjustment limit of the underlying actuator, or the calculated remaining safety margin. When the value is less than the preset shutdown dead zone threshold; Step B9: The actuator control submodule forcibly outputs an audible and visual alarm command and triggers the automatic speed reduction or emergency stop interlock protection action of the film printing production line to prevent the generation of batches of continuous waste products from the physical equipment level.

[0017] Compared with existing technologies, the beneficial effects achieved by this invention are as follows: This invention utilizes the global clock pulse generated by the printing press spindle encoder, combined with a hardware interrupt mechanism to improve the accuracy of heterogeneous data acquisition, and ensures the reliability of the quality tracking data base by performing microsecond-level hard synchronization latching of high-frequency two-dimensional visual images and underlying transient mechanical parameters. Addressing issues such as nonlinear speed fluctuations and material tension distortion that may occur in high-speed film printing environments, the system can accurately deduce the precise historical timestamp of defect occurrence using inverse integral equations, and precisely match the physical root causes of defects, such as doctor blade pre-press pressure and ink viscosity, from a massive cache, significantly improving pixel-level traceability accuracy while avoiding the inefficiency caused by traditional manual blind inspection. Furthermore, based on real-time extracted parameters such as continuous sheet color difference drift rate, decay acceleration, and remaining safe print weight, the system dynamically assesses and predicts the deterioration trend of film printing quality, rationally calculates and issues the optimal feedforward mechanical compensation amount, maximizing the production line yield and dynamic correction efficiency. By opening a sliding time window and extracting the differential decay gradient in the early stage, the system scientifically predicts the defective product outbreak point in the subsequent printing stage, ensuring that the adaptive adjustment and extreme value interlock shutdown actions of the screen printing production line are accurately completed. This improves the intelligence level and practical value of high-end home appliance panel printing management, significantly reduces the expensive material consumption costs caused by continuous waste, and improves the safety and stability of industrial printing production. Attached Figure Description

[0018] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a schematic diagram of the system module composition of the present invention. Detailed Implementation

[0019] 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.

[0020] It should be noted first that the core application scenario of the big data-based printed material quality traceability and defect prediction system provided in this invention is a multi-color group cascaded IMD / IML thin-film screen printing production line (such as the printing of intelligent control panels for high-end home appliances like washing machines and refrigerators). In traditional thin-film screen printing production lines, visual inspection equipment usually exists as an independent third-party plug-in, interacting with the main control PLC only through simple I / O signals for alarm / stop communication, lacking underlying data exchange between the two.

[0021] To break down these data silos, the main execution component of this system employs a high-performance edge computing gateway deployed on the production line (e.g., an industrial-grade gateway equipped with a multi-core ARM architecture and an FPGA coprocessor). In terms of hardware communication topology, this edge computing gateway physically connects to the servo drives of each silkscreen color group and the underlying PLC via a high-speed industrial real-time Ethernet connection, achieving sub-millisecond-level periodic data polling. Simultaneously, the gateway's FPGA coprocessor directly exposes a hardware interrupt pin, enabling nanosecond-level trigger binding with the high-resolution absolute encoder on the main drive shaft and the linear CCD camera at the unloading end.

[0022] Please see Figure 1 This invention provides a big data-based system for tracing the source of printed materials and predicting defects. The system includes a data acquisition module, a source analysis module, and an adaptive control module. The data acquisition module extracts the spindle clock and transmits a synchronous trigger signal, synchronously latching visual images with underlying mechanical parameters. The source analysis module analyzes the spatial coordinates of defects in thin-film printed materials and establishes a reverse mapping matrix, providing a basis for tracing the root causes of defects. The adaptive control module predicts and dynamically adjusts the mechanical compensation parameters of the corresponding color group in real time based on the color difference decay gradient of the continuous thin-film substrate. Both the data acquisition module and the adaptive control module are electrically connected to the source analysis module.

[0023] The data acquisition module includes a clock reference submodule, an image acquisition submodule, and a parameter latch submodule. The clock reference submodule further includes an incremental spindle encoder, a high-speed counter, and a clock divider circuit. The clock reference submodule is used to generate and transmit a global pulse signal characterizing the feed length. The image acquisition submodule further includes a linear CCD camera, a light source strobe controller, and an image acquisition card. The image acquisition submodule is used to receive trigger signals at the feeding end and capture two-dimensional surface quality images of the thin film substrate. The parameter latch submodule further includes an industrial Ethernet parser, a direct memory access controller, and a PLC underlying data buffer. The parameter latch submodule is used to synchronously capture the ink viscosity, screen spacing, doctor blade pre-press pressure, and oven temperature data of each printing color group under the same pulse timestamp.

[0024] The data acquisition module completely eliminates the software clock, which is susceptible to system scheduling issues. Its operation method involves hard synchronization interception of underlying hardware signals, and the specific execution steps include: Step S1: The incremental spindle encoder in the clock reference submodule acquires the high-frequency rotation pulses of the printing press main drive shaft in real time. After processing by the high-speed counter and clock frequency divider circuit, the pulses are generated to correspond to the physical feed length. A global clock pulse sequence that exhibits a strictly linear mapping relationship; Step S2: The global clock pulse sequence is transmitted in parallel to the image acquisition submodule and the parameter latch submodule. When the pulse counter reaches the preset single-area detection interval, a synchronization interrupt signal is triggered. Step S3: After the linear CCD camera of the image acquisition submodule captures the synchronization interrupt signal, it instantly exposes the image and, in conjunction with the light source strobe controller, completes the line frequency scan of the current image area to extract the absolute length coordinates of the current image area on the global feed axis. ; Step S4: Within the same hardware clock cycle, the parametric latch submodule bypasses the CPU through the direct memory access controller, forcibly reads and latches the ink viscosity of each printing color group temporarily stored in the programmable logic controller. Pre-printing pressure of the scraper Oven temperature and transient vehicle speed and compare it with the absolute length coordinates Bind them with the same hardware timestamp, encapsulate them into heterogeneous synchronous data frames and store them in the underlying cache; Through the above steps, a hardware-level, seamless latching mechanism between the visual image stream and multi-dimensional mechanical state parameters can be achieved during the continuous high-speed material feeding process of a thin-film screen printing production line. This is achieved by cleverly mounting the entire underlying physical sensing system onto the physical pulse axis of an absolute encoder. During the operation of the main drive shaft, the system monitors and collects various heterogeneous data sources from the chassis that have a decisive impact on traceability accuracy in real time and with high precision: one of these is the absolute length coordinates that are completely bound to the image on the substrate surface. The coordinates directly serve as the absolute index of the entire roll of film in spatial geometry. The frequency and accuracy of coordinate changes are directly controlled by a hardware clock divider circuit, completely eliminating the false shifts caused by network jitter in traditional Ethernet communication layer timestamps. Secondly, it addresses the transient speed of the printing press production line. This parameter directly reflects the actual feed rate of the film substrate per microsecond and is the core independent variable for smoothing acceleration and deceleration distortion in the later stages; thirdly, it reflects the fluid and mechanical characteristics generated within each printing color group during processing, mainly including ink viscosity. Pre-printing pressure of the scraper And screen spacing, these drastic fluctuations in process parameters directly constitute the physical causes of defects on the printed surface. Based on these rich and crucial hard-synchronization data frames acquired at the mechanical level, the system can construct a heterogeneous high-speed cache queue with deterministic temporal and spatial latching characteristics. For example, when the thin-film screen printing production line is in an emergency acceleration phase (such as the machine speed increasing from 0.5 km / h within 3 seconds...), Soaring to In high-dynamic operating conditions, the encoder pulse frequency received by the clock reference submodule synchronously surges, directly driving the line frequency exposure of the linear CCD camera to perform hardware-level adaptive tracking, ensuring that the captured image of the home appliance control panel does not produce any longitudinal stretching distortion. Simultaneously, the direct memory access controller, within nanoseconds without CPU interrupt intervention, forcibly applies the transient vehicle speed during this rapid speed change. And the ink viscosity of each color group fluctuates instantaneously due to high-speed friction. scraper pressure This data is fed into the tail data packet of the current image frame. The final output frame of heterogeneous synchronization data will serve as an absolutely reliable noise floor data for pixel-level root cause analysis in complex acceleration and deceleration environments, ensuring that no spatiotemporal phase errors occur at the data source.

[0025] The source analysis module includes a spacing analysis submodule, a reverse deduction submodule, a data addressing submodule, and a feature comparison submodule. The spacing analysis submodule is used to obtain the absolute physical material path from each printing color group to the visual camera at the unloading end. The reverse deduction submodule is used to calculate the specific timestamp of the current image feature point passing through the historical color group based on the current speed and physical material path. The data addressing submodule is used to use the historical timestamp to address and match the transient mechanical parameters that induce defects in the underlying database. The feature comparison submodule is used to match and analyze the extracted transient mechanical parameters with the cause threshold of historical typical defects.

[0026] The operation method of the source tracing analysis module includes the following steps: Step A1: During continuous film printing operations, input the basic topology parameters of the current production line, including parameters from the first... The precise physical feed length between the center of the screen printing line of each printing color group and the vertical projection point of the optical axis of the linear CCD camera lens at the feeding end. ,in This represents the total number of color groups on the current production line. Step A2: The image acquisition submodule performs pattern recognition on the acquired two-dimensional surface quality image. When a defect is detected in a certain area, the local coordinates of the defect image within the image area are extracted. And combined with the absolute length coordinates of the image area. The precise physical mileage calibration point at the location of the defect feature was calculated. ; Step A3: After calibrating the current defect mileage location, obtain the real-time time series of equipment speed recorded by the data acquisition module. Then, using the reverse derivation submodule, through integral equations... Calculate and obtain the calibration point of the defect feature at a historical moment when it crosses the first... Time difference during the printing of each color group's printing line ,in It is a velocity distribution function with respect to the material feed length; Step A4: Based on the current absolute time when the defect was detected. Combined with the calculated time difference Calculate the time the defective area experiences during the [number]th [period]. Exact historical transient timestamps during the printing of color units Repeat steps A3 and A4 until the set of historical timestamps of all associated color groups for the defective area is obtained. ; Step A5: The data addressing submodule uses a set of historical timestamps. The pointer address offset is precisely extracted from the heterogeneous synchronized data frames in the underlying cache. The set of transient mechanical parameters at the corresponding moment; Step A6: Identify the extracted set of transient mechanical parameters, analyze the abnormal characteristics of adjacent printing color groups after parameter changes such as doctor blade pre-printing pressure fluctuation, ink viscosity mutation or temperature step change, and match the real physical root cause that induces the image defect in the feature comparison submodule. Step A7: Use the source analysis module to output a complete quality source tracing report that includes the historical status of each color group, root cause mechanical parameters, and the current defect mapping.

[0027] Step A6 further includes: Step A61: Establish a database of root cause features of film printing defects. Record the abnormal fluctuations of mechanical parameters of the corresponding color group when various typical surface image defects occur in history into the feature database. Abnormal fluctuations include: the drop amplitude of doctor blade pre-printing pressure, the slope of ink viscosity decay, and the step variance of oven temperature. Step A62: Perform time-domain waveform analysis on the set of transient mechanical parameters extracted in step A5 to extract the actual abnormal fluctuation characteristics of the current production line parameter time series; Step A63: Based on the extracted actual abnormal fluctuation characteristics, perform cosine similarity matching calculation in the thin film printing defect root cause feature database; Step A64: Filter out feature similarity matches that reach a preset threshold. The historical typical abnormal fluctuation state is directly identified as the real physical root cause of the current image defect, and the specific printing color group where the physical root cause occurred is accurately located. Through the above steps, at the lag point where the visual camera at the material feeding end detects the defective product image, in order to achieve reverse causal locking from defects in the two-dimensional surface quality image to physical failures in the underlying mechanical components, the continuous time axis of the entire production line along the material feeding direction is cleverly decomposed into a time axis based on physical spacing. The spatial inverse path integral system. During the source mapping process, the source analysis module analyzes and calculates various key physical indicators that provide deterministic guidance for root cause localization in real time and accurately: First, the precise physical mileage calibration point at the location of the defect feature. First, the calibration point directly anchors the absolute physical location of the defect on the entire roll of substrate; second, it is the absolute physical feed length between the center of each screen printing color group's printing line and the lens optical axis. The topological spacing forms the physical span benchmark for reverse inference; third, it is the historical vehicle speed time series. Nonlinear integral value along the travel path This value directly characterizes the transient dwell time of the target defect point when passing through a specific color group screen within a precise time period in the past. The integral equation directly reduces the complex mechanical acceleration and deceleration motion to a linear historical time stamp. This is based on the set of historical timestamps derived from the inverse integration in the preceding stage. The system can scientifically and rationally reset the data block addressing pointer in memory to a specific physical moment several seconds ago. Combined with cosine similarity matching using a thin-film printing defect root cause feature database, it accurately diagnoses the specific process defect components that lead to the outbreak of quality degradation.

[0028] For example, when the image acquisition submodule detects a defect such as blurred color blocks or ink leakage at a specific button pattern on the control panel of the washing machine film being produced at the feeding end, the system instantly locks the physical mileage calibration point corresponding to the defective product location. The reverse engineering submodule extracts the historical vehicle speed sequence of the film at each stage, and through path integral equations, perfectly smooths out the disturbance caused by intermediate tension fluctuations, accurately calculating the location of the defect point. Seconds ago, it was passing through the second silver silkscreen color group, while Seconds ago, it was passing through the first white-background silkscreen color group. The data addressing submodule then used this timestamp set as the pointer offset to accurately retrieve it from the database. Historical data for color group 2 from 1 second ago. System comparison revealed the pre-printing pressure of the squeegee for color group 2 at that instant. An irreversible drop occurred. Cosine similarity calculation using high-dimensional feature vectors showed that the drop waveform had a 92% similarity to the "scraper cylinder micro-leakage" feature in the root cause feature database. This final traceability result will directly guide maintenance personnel to the source of the fault, eliminating the need for blindly shutting down the entire production line for troubleshooting.

[0029] The adaptive control module includes a gradient monitoring submodule, a prediction output submodule, and an actuator control submodule. The gradient monitoring submodule is used to monitor the color difference offset and the change in the overprinting deviation vector of the continuous film substrate in real time. The prediction output submodule is used to formulate the assessment result of the remaining safe print volume and the corresponding feedforward compensation strategy based on the change in the deviation vector. The actuator control submodule is used to control the operation of the production line by adjusting the tension torque of the servo motor and the execution step parameters of the pre-printing pressure of the bottom blade.

[0030] The operation of the adaptive control module includes the following steps: Step B1: In continuous film printing operations, the gradient monitoring submodule opens a data stream containing continuous surface quality images acquired at the feeding end. A sliding time window for each substrate print sheet; Step B2: Continuously extract the current color difference value of each substrate in the same physical calibration area within the sliding time window. Or overprinting deviation vector; Step B3: Calculate the displacement derivative of the color difference value between consecutive substrates using the first-order backward difference algorithm to obtain the first-order color difference drift rate of the current batch of film printing quality. ; Step B4: Calculate the rate derivative of the first-order chromatic aberration drift rate using the second-order backward difference algorithm to obtain the second-order chromatic aberration decay acceleration, which characterizes the trend of quality deterioration. ; Step B5: The prediction output submodule monitors the first-order chromatic aberration drift rate in real time. Acceleration of second-order color difference decay When the numerical changes of both values ​​are determined to exceed the static anti-shake dead zone, the feedforward prediction and dynamic compensation pipeline is activated.

[0031] In step B5, the specific execution steps for starting the feedforward prediction and dynamic compensation pipeline are as follows: extract the current color difference value. Combined with the calculated first-order chromatic aberration drift rate Acceleration of second-order color difference decay Solve for the optimal feedforward mechanical compensation amount Its calculation expression is: ; in, The number of adjustment steps for the doctor blade pre-press pressure or tension roller output of the target printing color group calculated in real time; The basic control output is used to maintain the current printing status; The first-order drift rate weighting coefficient; The second-order decay acceleration weighting coefficient; This is the absolute color difference weighting coefficient; The starting drift rate that triggers dynamic compensation; and These are the maximum color difference drift rate and maximum decay acceleration limits set by the system, respectively. The target color difference value is the standard reference benchmark. This represents the lower limit extreme value of the national standard tolerance. The prediction output submodule is based on the calculation results Encapsulate and generate feedforward control instructions.

[0032] The steps of the actuator control submodule in adjusting the underlying production line actions according to feedforward control commands further include: Step B6: Transmit the packaged feedforward mechanical compensation amount via industrial fieldbus. The defect is sent to the corresponding underlying actuator before it exceeds the national standard tolerance limit. Before becoming scrap, perform dynamic correction actions with tiny steps; Step B7: The system's underlying processor uses a kinematic approximation model to calculate the remaining safe print volume in real time. The calculation formula is: ; Step B8: When the calculated optimal feedforward mechanical compensation amount is detected... Exceeding the physical adjustment limit of the underlying actuator, or the calculated remaining safety margin. When the value is less than the preset shutdown dead zone threshold; Step B9: The actuator control submodule forcibly outputs an audible and visual alarm command and triggers the automatic speed reduction or emergency stop interlock protection action of the film printing production line to prevent the generation of batches of continuous waste products from the physical equipment level. Through the above steps, in continuous multi-batch film screen printing operations, a control paradigm shift from post-production waste interception to pre-production trend fine-tuning feedforward can be achieved. This transforms the quality deterioration trajectory of continuously printed substrates into a dynamic kinematic decay trend based on a sliding time window of N consecutive substrate sheets. During the continuous production process, the adaptive control module monitors and extrapolates various differential characteristic indicators that significantly warn of waste outbreaks in real time and accurately: primarily, the first-order color difference drift rate. This rate directly reflects the absolute speed at which printing quality moves towards the lower limit of the national standard tolerance as the number of printed sheets increases; secondly, it reflects the acceleration of second-order color difference decay. This parameter measures whether the quality deviation trend has the kinetic energy to "further deteriorate," and is the core basis for assessing nonlinear sudden defects caused by screen tension fatigue or severe squeegee wear; furthermore, it is the remaining safety margin calculated in real time. This indicator couples the microscopic color difference deviation rate with the deterioration acceleration to directly predict the remaining physical lifespan before a mass defective product is produced on the production line. Based on these high-confidence degradation gradient characteristics extracted within the current sliding time window, the adaptive control module can scientifically and rationally calculate the optimal corrective action to counteract the current deterioration trend using specific adaptive feedforward control compensation equations and a lifespan approximation model.

[0033] For example, when the production line is continuously printing the 5000th control panel film, the screen printing plate undergoes long-term mechanical stretching, leading to tension fatigue. This results in a color difference value between the most recent 50 film substrates within the sliding time window. It begins to show a slow, non-linear rise, with the first-order chromatic aberration drift rate... The system bypassed the anti-shake dead zone. Through a kinematic approximation model, it instantly calculated that without mechanical intervention, the production line would continue printing 350 more sheets (i.e., the remaining safe print run). After that, it completely broke through the lower limit of the national standard tolerance. It becomes scrap. At this point, the multi-parameter feedforward equation is immediately activated, performing a weighted dot product of the current deviation, drift velocity, and acceleration to directly calculate the downward fine-tuning compensation required for the servo scraper actuator of the current third color group. Step 1. This control command forcibly corrects insufficient ink application caused by screen fatigue by increasing the doctor blade pre-press pressure before waste is actually generated. Conversely, if the compensation calculated by the feedforward exceeds the physical adjustment limit of the actuator, or the calculated remaining safe print volume... If the system's dead zone drops to below 10 sheets, it will unconditionally issue an unmasked emergency stop command. This final control loop will serve as a crucial basis for ensuring production line yield, guaranteeing that the entire printing process can make reasonable and efficient adaptive adjustments based on actual mechanical fatigue conditions.

[0034] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0035] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0036] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0037] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims. All of these forms are within the protection scope of the present invention.

Claims

1. A big data-based system for tracing the source of printed materials and predicting defects, comprising a data acquisition module, a traceability analysis module, and an adaptive control module, characterized in that: The data acquisition module is used to extract the spindle clock and transmit a synchronous trigger signal to synchronously latch the visual image and the underlying mechanical parameters. The source analysis module is used to analyze the spatial coordinates of defects in the thin film printing and establish a reverse mapping matrix to provide a basis for tracing the root cause of defects. The adaptive control module is used to predict and dynamically adjust the mechanical compensation parameters of the corresponding color group in real time according to the color difference decay gradient of the continuous thin film substrate. Both the data acquisition module and the adaptive control module are electrically connected to the source analysis module.

2. The big data-based printed matter quality traceability and defect prediction system according to claim 1, characterized in that: The data acquisition module includes a clock reference submodule, an image acquisition submodule, and a parameter latch submodule. The clock reference submodule further includes an incremental spindle encoder, a high-speed counter, and a clock divider circuit. The clock reference submodule is used to generate and transmit a global pulse signal characterizing the feed length. The image acquisition submodule further includes a linear CCD camera, a light source strobe controller, and an image acquisition card. The image acquisition submodule is used to receive a trigger signal at the feeding end and capture a two-dimensional surface quality image of the thin film substrate. The parameter latch submodule further includes an industrial Ethernet parser, a direct memory access controller, and a PLC underlying data buffer. The parameter latch submodule is used to synchronously capture the ink viscosity, screen spacing, doctor blade pre-press pressure, and oven temperature data of each printing color group under the same pulse timestamp.

3. The system for tracing the quality of printed materials and predicting defects based on big data according to claim 1, characterized in that: The source tracing analysis module includes a spacing analysis submodule, a reverse deduction submodule, a data addressing submodule, and a feature comparison submodule. The spacing analysis submodule is used to obtain the absolute physical material path from each printing color group to the visual camera at the unloading end. The reverse deduction submodule is used to calculate the specific timestamp of the current image feature point passing through the historical color group based on the current speed and physical material path. The data addressing submodule is used to use the historical timestamp to address and match the transient mechanical parameters that induce defects in the underlying database. The feature comparison submodule is used to match and analyze the extracted transient mechanical parameters with the cause threshold of historical typical defects.

4. The system for tracing the quality of printed materials and predicting defects based on big data as described in claim 1, characterized in that: The adaptive control module includes a gradient monitoring submodule, a prediction output submodule, and an actuator control submodule. The gradient monitoring submodule is used to monitor the color difference offset and overprinting deviation vector change of the continuous film substrate in real time. The prediction output submodule is used to formulate the assessment result of the remaining safe print volume and the corresponding feedforward compensation strategy based on the deviation vector change. The actuator control submodule is used to control the operation of the production line by adjusting the tension torque of the servo motor and the execution step parameters of the pre-printing pressure of the bottom blade.

5. The system for tracing the quality of printed materials and predicting defects based on big data as described in claim 2, characterized in that: The operation method of the data acquisition module includes the following steps: Step S1: The incremental spindle encoder in the clock reference submodule acquires the high-frequency rotation pulses of the printing press main drive shaft in real time. After processing by the high-speed counter and clock frequency divider circuit, the pulses are generated to correspond to the physical feed length. A global clock pulse sequence that exhibits a strictly linear mapping relationship; Step S2: The global clock pulse sequence is transmitted in parallel to the image acquisition submodule and the parameter latch submodule. When the pulse counter reaches the preset single-area detection interval, a synchronization interrupt signal is triggered. Step S3: After the linear CCD camera of the image acquisition submodule captures the synchronization interrupt signal, it instantly exposes the image and, in conjunction with the light source strobe controller, completes the line frequency scan of the current image area to extract the absolute length coordinates of the current image area on the global feed axis. ; Step S4: Within the same hardware clock cycle, the parametric latch submodule bypasses the CPU through the direct memory access controller, forcibly reads and latches the ink viscosity of each printing color group temporarily stored in the programmable logic controller. Pre-printing pressure of the scraper Oven temperature and transient vehicle speed and compare it with the absolute length coordinates Bind them with the same hardware timestamp, encapsulate them into heterogeneous synchronous data frames and store them in the underlying cache.

6. The system for tracing the quality of printed materials and predicting defects based on big data according to claim 3, characterized in that: The operation method of the source tracing analysis module includes the following steps: Step A1: During continuous film printing operations, input the basic topology parameters of the current production line, including parameters from the first... The precise physical feed length between the center of the screen printing line of each printing color group and the vertical projection point of the optical axis of the linear CCD camera lens at the feeding end. ,in This represents the total number of color groups on the current production line. Step A2: The image acquisition submodule performs pattern recognition on the acquired two-dimensional surface quality image. When a defect is detected in a certain area, the local coordinates of the defect image within the image area are extracted. And combined with the absolute length coordinates of the image area. The precise physical mileage calibration point at the location of the defect feature was calculated. ; Step A3: After calibrating the current defect mileage location, obtain the real-time time series of equipment speed recorded by the data acquisition module. Then, using the reverse derivation submodule, through integral equations... Calculate and obtain the calibration point of the defect feature at a historical moment when it crosses the first... Time difference during the printing of each color group's printing line ,in It is a velocity distribution function with respect to the material feed length; Step A4: Based on the current absolute time when the defect was detected. Combined with the calculated time difference Calculate the time the defective area experiences during the [number]th [period]. Exact historical transient timestamps during the printing of color units Repeat steps A3 and A4 until the set of historical timestamps of all associated color groups for the defective area is obtained. ; Step A5: The data addressing submodule uses the historical timestamp set. The pointer address offset is precisely extracted from the heterogeneous synchronized data frames in the underlying cache. The set of transient mechanical parameters at the corresponding moment; Step A6: Identify the extracted set of transient mechanical parameters, analyze the abnormal characteristics of adjacent printing color groups after the occurrence of doctor blade pre-printing pressure fluctuations, ink viscosity abrupt changes, or temperature step parameter changes, and match the real physical root cause that induces the image defect in the feature comparison submodule. Step A7: Use the source analysis module to output a complete quality source tracing report that includes the historical status of each color group, root cause mechanical parameters, and the current defect mapping.

7. The system for tracing the quality of printed materials and predicting defects based on big data according to claim 6, characterized in that: Step A6 further includes: Step A61: Establish a database of root cause features of thin film printing defects. Record the abnormal fluctuations of mechanical parameters of the corresponding color groups when various typical surface image defects occur in history into the feature database. The abnormal fluctuations include: the drop amplitude of doctor blade pre-printing pressure, the slope of ink viscosity decay, and the step variance of oven temperature. Step A62: Perform time-domain waveform analysis on the set of transient mechanical parameters extracted in step A5 to extract the actual abnormal fluctuation characteristics of the current production line parameter time series; Step A63: Based on the extracted actual abnormal fluctuation characteristics, perform cosine similarity matching calculation in the thin film printing defect root cause feature database; Step A64: Filter out feature similarity matches that reach a preset threshold. The historical typical abnormal fluctuation state is directly identified as the real physical root cause of the current image defect, and the specific printing color group where the physical root cause occurred is accurately located.

8. The system for tracing the quality of printed materials and predicting defects based on big data as described in claim 4, characterized in that: The operation method of the adaptive control module includes the following steps: Step B1: In continuous film printing operations, the gradient monitoring submodule opens a data stream containing continuous surface quality images acquired at the feeding end. A sliding time window for each substrate print sheet; Step B2: Continuously extract the current color difference value of each substrate in the same physical calibration area within the sliding time window. Or overprinting deviation vector; Step B3: Calculate the displacement derivative of the color difference value between consecutive substrates using the first-order backward difference algorithm to obtain the first-order color difference drift rate of the current batch of film printing quality. ; Step B4: Calculate the rate derivative of the first-order chromatic aberration drift rate using a second-order backward difference algorithm to obtain the second-order chromatic aberration decay acceleration characterizing the quality deterioration trend. ; Step B5: The prediction output submodule monitors the first-order chromatic aberration drift rate in real time. Acceleration of second-order color difference decay When the numerical changes of both values ​​are determined to exceed the static anti-shake dead zone, the feedforward prediction and dynamic compensation pipeline is activated.

9. A big data-based system for tracing the quality of printed materials and predicting defects, as described in claim 8, is characterized in that: In step B5, the specific execution steps for starting the feedforward prediction and dynamic compensation pipeline are as follows: extract the current color difference value. Combined with the calculated first-order chromatic aberration drift rate Acceleration of second-order color difference decay Solve for the optimal feedforward mechanical compensation amount Its calculation expression is: ； in, The number of adjustment steps for the doctor blade pre-press pressure or tension roller output of the target printing color group calculated in real time; The basic control output is used to maintain the current printing status; The first-order drift rate weighting coefficient; The second-order decay acceleration weighting coefficient; This is the absolute color difference weighting coefficient; The starting drift rate that triggers dynamic compensation; and These are the maximum color difference drift rate and maximum decay acceleration limits set by the system, respectively. The target color difference value is the standard reference benchmark. This represents the lower limit extreme value of the national standard tolerance. The prediction output submodule is based on the calculation results Encapsulate and generate feedforward control instructions.

10. A big data-based system for tracing the quality of printed materials and predicting defects, as described in claim 9, is characterized in that: The step of the actuator control submodule adjusting the underlying production line action according to the feedforward control command further includes: Step B6: Transmit the packaged feedforward mechanical compensation amount via industrial fieldbus. The defect is sent to the corresponding underlying actuator before it exceeds the national standard tolerance limit. Before becoming scrap, perform dynamic correction actions with tiny steps; Step B7: The system's underlying processor uses a kinematic approximation model to calculate the remaining safe print volume in real time. The calculation formula is: ； Step B8: When the calculated optimal feedforward mechanical compensation amount is detected... Exceeding the physical adjustment limit of the underlying actuator, or the calculated remaining safety margin. When the value is less than the preset shutdown dead zone threshold; Step B9: The actuator control submodule forcibly outputs an audible and visual alarm command and triggers the automatic speed reduction or emergency stop interlock protection action of the film printing production line to prevent the generation of batches of continuous waste products from the physical equipment level.

Images