High-precision PCB pressing control method based on machine learning

Abstract

The invention relates to the technical field of machine learning, and provides a high-precision PCB pressing control method based on machine learning, and the method comprises the steps: collecting the technological index parameters of target PCB pressing equipment in a historical period; analyzing the characterization value of the process index; based on a comparison result of the process index characterization value and a preset process index characterization threshold value, whether the operation process of the target PCB pressing equipment meets the standard or not is determined; in response to the condition that the operation process of the target PCB pressing equipment meets the standard, collecting quality index parameters of the target PCB pressing equipment in a historical period; analyzing the characterization value of the quality index; determining whether the target PCB pressing equipment meets a risk assessment standard or not based on a difference result between the quality index characterization value and a preset quality index characterization threshold value; and determining a processing strategy based on the quality index characterization value in response to the fact that the target PCB pressing equipment does not conform to the risk assessment standard. The rejection rate and the energy consumption are reduced through self-adaptive regulation and control.

Description

Technical Field

[0001] This application relates to the field of machine learning technology, and in particular to a high-precision PCB lamination control method based on machine learning. Background Technology

[0002] The multilayer lamination process of printed circuit boards (PCBs) is a critical step in bonding inner copper foil, insulating dielectric, and the core board together under high temperature and pressure. The lamination quality directly determines the board thickness tolerance, interlayer bonding strength, impedance consistency, and the reliability of the final product. Currently, this process commonly uses fixed process curves and single-variable PID control. Process parameters are largely based on experience, making it difficult to respond in real-time to changes in material properties and environmental conditions. Furthermore, it suffers from systemic bottlenecks such as multivariable coupling misalignment and dynamic response lag, resulting in insufficient precision in thickness and deformation control. Warpage rates typically exceed 0.7%, and interlayer dielectric thickness deviations are difficult to stably control within ±5%. New board type process debugging can take several hours, exhibiting poor process adaptability and easily leading to defects such as delamination and bubbles. This process fails to meet the stringent requirements of high-density, high-speed PCBs for impedance deviation ≤±3%, thickness uniformity, and warpage ≤0.3%, and also fails to meet the high standards of zero-defect reliability required in fields such as automotive electronics.

[0003] Chinese Patent Publication No. CN111405759A discloses a method for manufacturing a high-precision communication optical module printed circuit board. Specifically, by setting up a lamination control method, an electroplating control method, a surface copper control method, a circuit control method, an outer layer size control method, and a gold finger appearance control method in the existing circuit board manufacturing process, the manufacturing process is controlled. By combining the above methods, the quality and pass rate of the product are strictly controlled at each stage of the manufacturing process, improving the gold surface appearance yield and reducing scrap due to gold penetration.

[0004] Chinese Patent Publication No. CN114282413A discloses a simulation method and system for the lamination molding process of printed circuit boards (PCBs). The method involves: acquiring parameter data of the PCB; determining a partitioning scheme for each wiring layer based on the acquired parameter data; identifying the feature information of each partition based on the partitioning scheme; establishing a three-dimensional geometric model of the PCB with partition identification information based on the partitioning scheme; calculating the equivalent performance parameters of each partition based on the identified feature information; configuring the equivalent performance parameters calculated based on the feature information for each partition of the PCB three-dimensional geometric model based on the partition identification information; configuring the boundary conditions and loads for the simulation calculation; calling the PCB resin curing deformation module for solving the problem; ending the simulation calculation at the moment of lamination completion in the lamination molding process; and obtaining the simulation results. This invention achieves high-precision and efficient automatic modeling and simulation analysis of complex, large-size, and multi-layer PCB lamination molding processes under general computer hardware conditions.

[0005] Therefore, it is evident that the existing technology has the following problems: Existing lamination control technology mainly relies on static fixed process curves, resulting in insufficient control precision in the lamination process. It cannot respond in real time to material batch fluctuations, changes in ambient temperature, and equipment status drift, causing quality defects such as uneven plate thickness, warping deviations, and poor interlayer bonding, thus leading to increased scrap rate and energy consumption. Summary of the Invention

[0006] To address this, the present invention provides a high-precision PCB lamination control method based on machine learning, which overcomes the problem that existing lamination control mainly relies on static fixed process curves, resulting in insufficient control precision in the lamination process. This makes it impossible to respond in real time to material batch fluctuations, changes in ambient temperature, and equipment status drift, leading to quality defects such as uneven board thickness, warping deviations, and poor interlayer bonding. Consequently, it increases scrap rate and energy consumption, resulting in low PCB lamination control efficiency.

[0007] To achieve the above objectives, this invention provides a high-precision PCB lamination control method based on machine learning, comprising: Collect process parameters from the historical cycle of the target PCB lamination equipment; Analyze the characterization values ​​of the process indicators based on the aforementioned process index parameters; Based on the comparison results between the process index characterization values ​​and the predetermined process index characterization thresholds, it is determined whether the operating process of the target PCB lamination equipment meets the standards. In response to the fact that the operating process of the target PCB lamination equipment meets the standards, the quality index parameters of the target PCB lamination equipment within the historical cycle are collected. Analyze the characterization values ​​of the quality indicators based on the aforementioned quality indicator parameters; The difference between the quality index characterization value and the predetermined quality index characterization threshold is used to determine whether the target PCB lamination equipment meets the risk assessment criteria. In response to the target PCB lamination equipment failing to meet the risk assessment criteria, a processing strategy is determined based on the difference between the quality indicator characterization value and the predetermined quality indicator characterization threshold; and the adjustment range of the process indicator characterization threshold is determined. The process parameters include heating rate, pressure increase rate, and vacuum degree. The quality indicators include impedance error, thickness deviation, warpage, and interlayer bonding strength.

[0008] Furthermore, the process of analyzing the process index characterization values ​​includes: Collect the heating rate, pressure rate, and vacuum level of the target PCB lamination equipment during historical cycles; The ratio of the calculated heating rate to the predetermined heating rate threshold is determined as the first process limiting index parameter; The ratio of the calculated boost rate to the predetermined boost rate threshold is determined as the second process limiting parameter. The ratio of the calculated vacuum level to the predetermined vacuum level threshold is determined as the third process limiting parameter. The summation of the first process constraint parameter, the second process constraint parameter, and the third process constraint parameter is determined as the process indicator characterization value.

[0009] Furthermore, the process of determining whether the target PCB lamination equipment's operating process meets the standard by comparing the process index characterization value with the predetermined process index characterization threshold includes: Extract the comparison results of the process index characterization values ​​with the predetermined process index characterization thresholds; If the process index value is less than the predetermined process index threshold, then the operating process of the target PCB lamination equipment is determined to meet the standard.

[0010] Furthermore, the process of determining whether the operating process of the target PCB laminating equipment does not meet the standard by comparing the process index characterization value with the predetermined process index characterization threshold includes: Extract the comparison results of the process index characterization values ​​with the predetermined process index characterization thresholds; If the process index value is greater than or equal to the predetermined process index threshold, then the operating process of the target PCB lamination equipment is determined to be non-compliant with the standard.

[0011] Furthermore, the process of analyzing the quality index parameters to determine the quality index characterization value includes: Collect impedance error, thickness deviation, warpage, and interlayer bonding strength of the target PCB lamination equipment during historical cycles; The ratio of the calculated impedance error to the predetermined impedance error threshold is the first quality limit parameter. The ratio of the thickness deviation to the predetermined thickness deviation threshold is calculated as the second quality limit index parameter; The ratio of the warpage rate to the predetermined warpage rate threshold is calculated as the third quality control parameter. The predetermined interlayer bond strength threshold and the ratio of interlayer bond strength are calculated as the fourth quality limit index parameter. The summation of the first quality limit index parameter, the second quality limit index parameter, the third quality limit index parameter, and the fourth quality limit index parameter is determined as the quality index characterization value.

[0012] Furthermore, the process of determining whether the target PCB laminating equipment meets the risk assessment criteria based on the difference between the quality indicator characterization value and the predetermined quality indicator characterization threshold includes: Calculate the difference between the quality indicator value and the predetermined quality indicator threshold; If the difference between the quality indicator value and the predetermined quality indicator threshold is less than the predetermined difference threshold, then the target PCB lamination equipment is determined to meet the risk assessment criteria.

[0013] Furthermore, the process of determining whether the target PCB laminating equipment fails to meet the risk assessment criteria based on the difference between the quality indicator characterization value and the predetermined quality indicator characterization threshold includes: Calculate the difference between the quality indicator value and the predetermined quality indicator threshold; If the difference between the quality indicator value and the predetermined quality indicator threshold is greater than or equal to the predetermined difference threshold, then the target PCB lamination equipment is determined to be non-compliant with the risk assessment criteria.

[0014] Furthermore, the process of determining the corresponding handling strategy in response to the target PCB lamination equipment failing to meet the risk assessment criteria includes: Calculate the difference between the quality indicator value and the predetermined quality indicator threshold; The adjustment range of the process indicator characterization threshold is determined based on the difference between the quality indicator characterization value and the predetermined quality indicator characterization threshold.

[0015] Furthermore, the process of determining the adjustment range of the process indicator characterization threshold based on the difference between the quality indicator characterization value and the predetermined quality indicator characterization threshold includes: Calculate the difference between the quality indicator value and the predetermined quality indicator threshold; If the difference between the quality indicator value and the predetermined quality indicator threshold is greater than the predetermined difference threshold, then the process indicator threshold will be lowered.

[0016] Furthermore, the process of ensuring that the target PCB lamination equipment's operating procedures and risk assessments comply with standards includes: Extract the comparison results of the process index characterization values ​​with the predetermined process index characterization thresholds; Calculate the difference between the quality indicator value and the predetermined quality indicator threshold; If the process index value is less than the predetermined process index threshold and the difference between the quality index value and the predetermined quality index threshold is less than the predetermined difference threshold, then the operating process and risk assessment of the target PCB lamination equipment are determined to meet the standards.

[0017] Compared with existing technologies, the beneficial effects of this invention are that it provides a high-precision PCB lamination control method based on machine learning. By establishing a hierarchical and quantitative evaluation system for process and quality data, the compliance of equipment operation is determined at the process level based on parameters such as heating rate, pressure rate, and vacuum degree. After the process meets the standards, risk assessment is further conducted based on quality indicators such as impedance error, thickness deviation, and warpage, thereby realizing closed-loop management from process control to result early warning. This not only improves the controllability and standardization of the lamination process, but also enables early identification and strategic intervention of quality risks through deviation analysis. At the same time, it supports continuous optimization based on historical data, comprehensively enhancing the quality assurance capability and process stability of the PCB lamination manufacturing process.

[0018] In particular, this invention selects key process parameters—heating rate, pressurization rate, and vacuum level—that directly determine resin rheology, interlayer pressing, and venting effects. These parameters are then compared and summed with preset thresholds to construct a comprehensive process index characterization value that reflects the synergistic state of heat, force, and environment. This enables the quantification and comprehensive evaluation of the pressing process's health. By comparing this value with preset thresholds and establishing a clear rule that values ​​below the threshold indicate compliance, an automated and standardized process compliance judgment mechanism is formed. This provides reliable pre-trigger conditions for subsequent quality risk monitoring. It not only transforms traditional experience into calculable and monitorable quantitative indicators but also lays the foundation for subsequent multi-variable collaborative optimization control through multi-parameter coupling characterization, thereby improving the overall controllability, early warning capability, and closed-loop control logic of the process.

[0019] In particular, this invention selects impedance error, thickness deviation, warpage, and interlayer bonding strength—ultimate quality parameters that directly determine the electrical performance, structural accuracy, and long-term reliability of PCBs—and uses a ratio normalization and summation method to construct a comprehensive quality index characterization value, achieving an overall quantitative assessment of the lamination output quality. Furthermore, by calculating the difference between this characterization value and a target threshold, and establishing a tiered judgment based on the difference threshold, a refined and tiered risk control decision-making mechanism is established. This not only transforms the stringent industry standards of high-end PCBs into calculable and optimizable clear targets, shifting quality control from experience-based judgment to data-driven approaches, but also uses the quantitative results to pinpoint weak links in the process, providing key feedback signals for closed-loop optimization of preceding process parameters and adaptive learning of the system, thus forming a continuous improvement closed loop oriented towards final quality.

[0020] In particular, this invention constructs a complete quality-driven adaptive control closed loop. When the system determines a risk exceeding the limit based on the difference between the quality indicator value and the threshold, it can automatically adjust the amplitude of the process indicator threshold accordingly. This transforms the terminal quality deviation into an optimization signal for the process control standard in real time, realizing reverse feedback and dynamic calibration from quality results to process parameters. Simultaneously, by defining the simultaneous satisfaction of both process and quality indicators as the stable compliance state of the system, a clear benchmark is provided for maintaining optimal production. This entire mechanism enables the system not only to proactively prevent and self-correct risks when they occur, but also to maintain stable operation under optimal conditions, forming an intelligent control closed loop with continuous self-optimization capabilities. Attached Figure Description

[0021] Figure 1 This is a flowchart illustrating the steps of the high-precision PCB lamination control method based on machine learning in an embodiment of the present invention. Figure 2 This is a flowchart illustrating the steps involved in analyzing the characterization values ​​of process indicators in an embodiment of the present invention. Figure 3 This invention provides a logic diagram for determining whether the operating process of the target PCB lamination equipment conforms to the standard. Figure 4 This is a logic diagram for determining whether a target PCB lamination device meets the risk assessment criteria in an embodiment of the present invention. Detailed Implementation

[0022] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0023] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0024] Please see Figure 1 The diagram shows the steps of a high-precision PCB lamination control method based on machine learning, as described in an embodiment of the present invention. The present invention provides a high-precision PCB lamination control method based on machine learning, comprising: Step S1: Collect process parameters of the target PCB lamination equipment during the historical cycle. Step S2: Analyze the process index characterization values ​​based on the process index parameters; Step S3: Determine whether the operating process of the target PCB laminating equipment meets the standard based on the comparison result between the process index characterization value and the predetermined process index characterization threshold. Step S4: In response to the target PCB lamination equipment's operating process meeting the standard, collect the quality index parameters of the target PCB lamination equipment within the historical cycle; analyze the quality index characterization value based on the quality index parameters; Step S5: Determine whether the target PCB lamination equipment meets the risk assessment criteria based on the difference between the quality index characterization value and the predetermined quality index characterization threshold. Step S6: In response to the target PCB lamination equipment not meeting the risk assessment criteria, a processing strategy is determined based on the difference between the quality indicator characterization value and the predetermined quality indicator characterization threshold to determine the adjustment range of the process indicator characterization threshold. The process parameters include heating rate, pressure increase rate, and vacuum degree. The quality indicators include impedance error, thickness deviation, warpage, and interlayer bonding strength.

[0025] In this embodiment, a hierarchical and quantitative evaluation system for process and quality data is established. At the process level, the compliance of equipment operation is determined based on parameters such as heating rate, pressure rate, and vacuum degree. After the process meets the standards, risk assessment is further conducted based on quality indicators such as impedance error, thickness deviation, and warpage. This achieves closed-loop management from process control to result early warning. It not only improves the controllability and standardization of the lamination process, but also enables early identification and strategic intervention of quality risks through deviation analysis. At the same time, it supports continuous optimization based on historical data, comprehensively enhancing the quality assurance capability and process stability of the PCB lamination manufacturing process.

[0026] Please see Figure 2 The diagram shown is a flowchart illustrating the steps involved in analyzing the characteristic values ​​of process indicators according to an embodiment of the present invention. The process for analyzing the characteristic values ​​of process indicators according to the present invention includes: Step S21: Collect the heating rate, pressure rate, and vacuum level of the target PCB lamination equipment during the historical cycle. Step S22: Calculate the ratio of the heating rate to the predetermined heating rate threshold and determine it as the first process limiting index parameter; calculate the ratio of the pressure rate to the predetermined pressure rate threshold and determine it as the second process limiting index parameter; calculate the ratio of the vacuum degree to the predetermined vacuum degree threshold and determine it as the third process limiting index parameter. Step S23: The first process limiting index parameter, the second process limiting index parameter, and the third process limiting index parameter are summed to determine the process index characterization value.

[0027] In this embodiment, the heating rate is calculated by calculating the difference between the ending temperature and the starting temperature, and then dividing the temperature difference by the cycle time to determine the heating rate. In this embodiment, the pressure rise rate is calculated by calculating the difference between the final pressure value and the initial pressure value, and then dividing the difference by the cycle time to determine the pressure rise rate. In this embodiment, the predetermined heating rate threshold, pressure rate threshold, and vacuum degree threshold are all obtained in advance. The heating rate, pressure rate, and vacuum degree of the target PCB lamination equipment are collected within 3 months of stable operation, and their average values ​​are calculated as the heating rate threshold, pressure rate threshold, and vacuum degree threshold.

[0028] In the embodiments, the existing PCB lamination process has significant defects in the control of temperature, pressure and vacuum. The temperature control accuracy is insufficient and the dynamic response is lagging, resulting in uneven curing, excessive warpage and batch-to-batch deviations of up to ±3℃. The decoupling between pressure regulation and heating process can easily cause local resin loss, pressure loss or lamination slippage. The lack of linkage between vacuum control and other parameters makes it impossible to effectively suppress delamination and bubbles. The core problem is the loss of control over the strong coupling relationship between the three parameters. Traditional independent control schemes cannot coordinate their interaction, such as the impact of pressure changes on heat transfer, resulting in a narrow process window and poor adaptability, ultimately leading to quality fluctuations such as inconsistent thickness and impedance and glue overflow.

[0029] Please see Figure 3 As shown, this is a logic diagram for determining whether the operating process of the target PCB laminating equipment meets the standard according to an embodiment of the present invention. The process of determining whether the operating process of the target PCB laminating equipment meets the standard based on the comparison result of the process index characterization value and the predetermined process index characterization threshold includes: Extract the comparison results of the process index characterization values ​​with the predetermined process index characterization thresholds; If the process index characterization value is less than the predetermined process index characterization threshold, then the operating process of the target PCB lamination equipment is determined to meet the standard. If the process index value is greater than or equal to the predetermined process index threshold, then the operating process of the target PCB lamination equipment is determined to be non-compliant with the standard.

[0030] In this embodiment, the predetermined process index characterization threshold is obtained in advance. All process index characterization values ​​of the target PCB lamination equipment are collected within 3 months of stable operation, and their average value is calculated as the process index characterization threshold. The process index characterization threshold is selected within the range [3.05, 3.15], and is preferably 3.10 in this embodiment.

[0031] In this embodiment, key process parameters that directly determine resin rheology, interlayer pressing, and venting effects—heating rate, pressurization rate, and vacuum degree—are selected and compared and summed with preset thresholds to construct a comprehensive process index characterization value that reflects the synergistic state of heat, force, and environment. This enables the quantification and comprehensive evaluation of the health of the pressing process. By comparing this value with preset thresholds and establishing a clear judgment rule that values ​​below the threshold are compliant, an automated and standardized process compliance judgment mechanism is formed. This provides a reliable pre-trigger condition for subsequent quality risk monitoring. It not only transforms traditional experience into calculable and monitorable quantitative indicators but also lays the foundation for subsequent multi-variable collaborative optimization control through multi-parameter coupling characterization, thereby improving the overall controllability, early warning capability, and closed-loop control logic of the process.

[0032] Specifically, the process of analyzing the quality index parameters to determine the quality index characterization values ​​includes: Collect impedance error, thickness deviation, warpage, and interlayer bonding strength of the target PCB lamination equipment during historical cycles; The ratio of the calculated impedance error to the predetermined impedance error threshold is the first quality limit parameter. The ratio of the thickness deviation to the predetermined thickness deviation threshold is calculated as the second quality limit index parameter; The ratio of the warpage rate to the predetermined warpage rate threshold is calculated as the third quality control parameter. The predetermined interlayer bond strength threshold and the ratio of interlayer bond strength are calculated as the fourth quality limit index parameter. The summation of the first quality limit index parameter, the second quality limit index parameter, the third quality limit index parameter, and the fourth quality limit index parameter is determined as the quality index characterization value.

[0033] In this embodiment, the predetermined impedance error threshold, thickness deviation threshold, warpage threshold, and interlayer bonding strength threshold are all obtained in advance. All impedance errors, thickness deviations, warpage, and interlayer bonding strength of the target PCB lamination equipment are collected within 3 months of stable operation, and their average values ​​are calculated as the impedance error threshold, thickness deviation threshold, warpage threshold, and interlayer bonding strength threshold.

[0034] In this embodiment, the impedance error is calculated by calculating the difference between the actual measured impedance value and the predetermined target impedance value, and then dividing the difference by the predetermined target impedance value to determine the impedance difference. In this embodiment, the thickness deviation is calculated by calculating the difference between the actual measured thickness and the designed target thickness, and then dividing the difference by the predetermined target thickness to determine the thickness deviation.

[0035] In this embodiment, the warpage is calculated by dividing the vertical distance from the point of maximum curvature on the PCB board to the reference plane when the PCB board is freely laid flat on the reference plane by the effective diagonal length of the corresponding PCB board.

[0036] In this embodiment, the interlayer bonding strength is calculated by calculating the ratio of the total area of ​​detected delamination defects to the total predetermined area inside the sample, and the absolute value of the difference between the ratio and 1 is determined as the interlayer bonding strength.

[0037] In this embodiment, the thickness uniformity is defined by the range or standard deviation of the thickness measurement at nine points on the board surface; the interlayer bond strength is defined by the peel force per unit width measured by the standard peel test; all target impedance lines are measured using a time domain reflectometer, and the standard deviation of the time domain reflectometer relative to the design value is used to characterize the impedance consistency; the warpage rate is calculated by measuring the ratio of the maximum warpage height to the diagonal length according to the IPC-TM-650 standard.

[0038] In this embodiment, to verify the effectiveness of the invention in engineering applications, a complete testing and evaluation process needs to be designed. First, a representative control scheme is selected as a baseline, and N batches of the same PCB product are laminated using both the baseline and the method of this invention. Key indicators for each batch of finished boards are tested, including: board thickness uniformity, interlayer bonding strength, impedance consistency, and warpage. Then, the average value, fluctuation range, and defect rate of each indicator under the two methods are compared and statistically analyzed. It is expected that the present invention will significantly outperform traditional processes: for example, board thickness tolerance is reduced from ±10% to within ±5%, 50Ω characteristic impedance deviation is reduced from ±10Ω to within ±3Ω (error ≤ ±6%), finished product warpage is all below 0.3% with no delamination or bubbles, and the single-batch scrap rate is below 1%. Statistical analysis can use t-tests or ANOVA to verify the significance of the improvement, while calculating the process capability index Cpk to quantify the improvement in process capability. Finally, the stability of the method is evaluated in a longer period of continuous production, such as after 100 consecutive batches, the key indicators remain within the acceptable threshold without significant drift. The above evaluation proves that the machine learning lamination control method described in this invention can significantly improve the consistency and accuracy of PCB lamination, providing a reliable process guarantee for high-end PCB manufacturing.

[0039] Please see Figure 4 As shown, this is a logic diagram for determining whether a target PCB laminating equipment meets the risk assessment criteria according to an embodiment of the present invention. The process of determining whether a target PCB laminating equipment meets the risk assessment criteria based on the difference between the quality index characterization value and the predetermined quality index characterization threshold includes: Calculate the difference between the quality indicator value and the predetermined quality indicator threshold; If the difference between the quality indicator value and the predetermined quality indicator threshold is less than the predetermined difference threshold, then the target PCB lamination equipment is determined to meet the risk assessment criteria. If the difference between the quality indicator value and the predetermined quality indicator threshold is greater than or equal to the predetermined difference threshold, then the target PCB lamination equipment is determined to be non-compliant with the risk assessment criteria.

[0040] In this embodiment, the predetermined quality index characterization threshold is obtained in advance. All quality index characterization values ​​of the target PCB lamination equipment are collected within 3 months of stable operation, and their average value is calculated as the quality index characterization threshold. The quality index characterization threshold is selected within the range [4.05, 4.15], and is preferably 4.10 in this embodiment.

[0041] In this embodiment, the predetermined difference threshold is obtained in advance. The difference between all quality index characterization values ​​of the target PCB laminating equipment within 3 months of stable operation and the predetermined quality index characterization threshold is collected, and the average value is calculated as the difference threshold. The predetermined difference threshold is selected within the range [0.15, 0.35], and is preferably 0.20 in this embodiment.

[0042] In this embodiment, by selecting impedance error, thickness deviation, warpage, and interlayer bonding strength—ultimate quality parameters that directly determine the electrical performance, structural accuracy, and long-term reliability of the PCB—and using a ratio normalization and summation method, a comprehensive quality index characterization value is constructed, achieving an overall quantitative assessment of the lamination output quality. Furthermore, by calculating the difference between this characterization value and a target threshold, and establishing a difference threshold for tiered judgment, a refined and tiered risk control decision-making mechanism is established. This not only transforms the stringent industry standards of high-end PCBs into calculable and optimizable clear targets, shifting quality control from experience-based judgment to data-driven approaches, but also uses the quantitative results to pinpoint weak links in the process, providing key feedback signals for closed-loop optimization of preceding process parameters and adaptive learning of the system, thus forming a continuous improvement closed loop guided by final quality.

[0043] Specifically, the process of determining the corresponding handling strategy in response to the target PCB lamination equipment failing to meet the risk assessment criteria includes: Calculate the difference between the quality indicator value and the predetermined quality indicator threshold; The adjustment range of the process indicator characterization threshold is determined based on the difference between the quality indicator characterization value and the predetermined quality indicator characterization threshold.

[0044] In this embodiment, by establishing an intelligent closed loop of quality quantification feedback and process threshold self-adjustment, a fundamental shift from passive detection to proactive optimization of the PCB lamination process is achieved. Specifically, the system automatically calculates the precise adjustment range of the threshold values ​​of key upstream process parameters based on the quantified difference between the comprehensive characterization value of the finished product quality indicators and the target threshold. This not only transforms the traditional, lagging intervention relying on human experience into data-driven, forward-looking, and adaptive precise control, effectively preventing defect recurrence, but also enables the production process to continuously self-optimize as production data accumulates, thereby significantly improving product consistency, production yield, and overall intelligence level.

[0045] Specifically, the process of determining the adjustment range of the process indicator characterization threshold based on the difference between the quality indicator characterization value and the predetermined quality indicator characterization threshold includes: Calculate the difference between the quality indicator value and the predetermined quality indicator threshold; If the difference between the quality indicator value and the predetermined quality indicator threshold is greater than the predetermined difference threshold, then the process indicator threshold will be lowered.

[0046] In this embodiment, a closed-loop optimization of process control standards is achieved by establishing an automatic adjustment mechanism based on quantitative feedback and conditional triggering. When the overall deviation of quality indicators exceeds the preset tolerance, the system automatically lowers the threshold of process parameters to ensure that process adjustments are targeted and objective. Calibration is only initiated when the actual deviation reaches the warning line, avoiding ineffective intervention. Simultaneously, a feedforward risk prevention capability is constructed. By tightening process requirements in advance to suppress quality degradation trends, the system forms a stable and traceable autonomous rule that can respond consistently to quality fluctuations based on predefined logic. This drives continuous convergence of process levels without human intervention, ensuring long-term consistency of product performance.

[0047] Specifically, the process of ensuring that the target PCB lamination equipment's operating procedures and risk assessments comply with standards includes: Extract the comparison results of the process index characterization values ​​with the predetermined process index characterization thresholds; Calculate the difference between the quality indicator value and the predetermined quality indicator threshold; If the process index value is less than the predetermined process index threshold and the difference between the quality index value and the predetermined quality index threshold is less than the predetermined difference threshold, then the operating process and risk assessment of the target PCB lamination equipment are determined to meet the standards.

[0048] In this embodiment, by constructing a complete quality-driven adaptive control closed loop, when the system determines that a risk exceeds the limit based on the difference between the quality indicator value and the threshold, it can automatically adjust the amplitude of the process indicator threshold accordingly. This transforms the terminal quality deviation into an optimization signal for the process control standard in real time, realizing reverse feedback and dynamic calibration from quality results to process parameters. Simultaneously, by defining the simultaneous satisfaction of both process and quality indicators as the stable compliance state of the system, a clear benchmark is provided for maintaining optimal production. This entire mechanism enables the system not only to proactively prevent and self-correct risks when they occur, but also to maintain stable operation under optimal conditions, forming an intelligent control closed loop with continuous self-optimization capabilities.

[0049] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to the specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.

Claims

1. A high-precision PCB lamination control method based on machine learning, characterized in that, include: Collect process parameters from the historical cycle of the target PCB lamination equipment; Analyze the characterization values ​​of the process indicators based on the aforementioned process index parameters; Based on the comparison results between the process index characterization values ​​and the predetermined process index characterization thresholds, it is determined whether the operating process of the target PCB lamination equipment meets the standards. In response to the fact that the operating process of the target PCB lamination equipment meets the standards, the quality index parameters of the target PCB lamination equipment within the historical cycle are collected. Analyze the characterization values ​​of the quality indicators based on the aforementioned quality indicator parameters; The difference between the quality index characterization value and the predetermined quality index characterization threshold is used to determine whether the target PCB lamination equipment meets the risk assessment criteria. In response to the target PCB lamination equipment failing to meet the risk assessment criteria, a processing strategy is determined based on the difference between the quality indicator characterization value and the predetermined quality indicator characterization threshold; and the adjustment range of the process indicator characterization threshold is determined. The process parameters include heating rate, pressure increase rate, and vacuum degree. The quality indicators include impedance error, thickness deviation, warpage, and interlayer bonding strength.

2. The high-precision PCB lamination control method based on machine learning according to claim 1, characterized in that, The process of analyzing the characteristic values ​​of process indicators based on the aforementioned process indicator parameters includes: Collect the heating rate, pressure rate, and vacuum level of the target PCB lamination equipment during historical cycles; The ratio of the calculated heating rate to the predetermined heating rate threshold is determined as the first process limiting index parameter; The ratio of the calculated boost rate to the predetermined boost rate threshold is determined as the second process limiting parameter. The ratio of the calculated vacuum level to the predetermined vacuum level threshold is determined as the third process limiting parameter. The summation of the first process constraint parameter, the second process constraint parameter, and the third process constraint parameter is determined as the process indicator characterization value.

3. The high-precision PCB lamination control method based on machine learning according to claim 2, characterized in that, The process of determining whether the target PCB lamination equipment's operating process meets the standard based on the comparison results between the process index characterization values ​​and the predetermined process index characterization thresholds includes: Extract the comparison results of the process index characterization values ​​with the predetermined process index characterization thresholds; If the process index value is less than the predetermined process index threshold, then the operating process of the target PCB lamination equipment is determined to meet the standard.

4. The high-precision PCB lamination control method based on machine learning according to claim 3, characterized in that, The process of determining whether the operating process of the target PCB laminating equipment does not meet the standard based on the comparison result between the process index characterization value and the predetermined process index characterization threshold includes: Extract the comparison results of the process index characterization values ​​with the predetermined process index characterization thresholds; If the process index value is greater than or equal to the predetermined process index threshold, then the operating process of the target PCB lamination equipment is determined to be non-compliant with the standard.

5. The high-precision PCB lamination control method based on machine learning according to claim 4, characterized in that, The process of analyzing the quality indicator characterization value based on the aforementioned quality indicator parameters includes: Collect impedance error, thickness deviation, warpage, and interlayer bonding strength of the target PCB lamination equipment during historical cycles; The ratio of the calculated impedance error to the predetermined impedance error threshold is the first quality limit parameter. The ratio of the thickness deviation to the predetermined thickness deviation threshold is calculated as the second quality limit index parameter; The ratio of the warpage rate to the predetermined warpage rate threshold is calculated as the third quality control parameter. The predetermined interlayer bond strength threshold and the ratio of interlayer bond strength are calculated as the fourth quality limit index parameter. The summation of the first quality limit index parameter, the second quality limit index parameter, the third quality limit index parameter, and the fourth quality limit index parameter is determined as the quality index characterization value.

6. The high-precision PCB lamination control method based on machine learning according to claim 5, characterized in that, The process of determining whether the target PCB lamination equipment meets the risk assessment criteria based on the difference between the quality indicator characterization value and the predetermined quality indicator characterization threshold includes: Calculate the difference between the quality indicator value and the predetermined quality indicator threshold; If the difference between the quality indicator value and the predetermined quality indicator threshold is less than the predetermined difference threshold, then the target PCB lamination equipment is determined to meet the risk assessment criteria.

7. The high-precision PCB lamination control method based on machine learning according to claim 6, characterized in that, The process of determining whether the target PCB lamination equipment fails to meet the risk assessment criteria based on the difference between the quality indicator characterization value and the predetermined quality indicator characterization threshold includes: Calculate the difference between the quality indicator value and the predetermined quality indicator threshold; If the difference between the quality indicator value and the predetermined quality indicator threshold is greater than or equal to the predetermined difference threshold, then the target PCB lamination equipment is determined to be non-compliant with the risk assessment criteria.

8. The high-precision PCB lamination control method based on machine learning according to claim 7, characterized in that, The process of determining the corresponding handling strategy in response to the target PCB lamination equipment failing to meet the risk assessment criteria includes: Calculate the difference between the quality indicator value and the predetermined quality indicator threshold; The adjustment range of the process indicator characterization threshold is determined based on the difference between the quality indicator characterization value and the predetermined quality indicator characterization threshold.

9. The high-precision PCB lamination control method based on machine learning according to claim 8, characterized in that, The process of determining the adjustment range of the process indicator characterization threshold based on the difference between the quality indicator characterization value and the predetermined quality indicator characterization threshold includes: Calculate the difference between the quality indicator value and the predetermined quality indicator threshold; If the difference between the quality indicator value and the predetermined quality indicator threshold is greater than the predetermined difference threshold, then the process indicator threshold will be lowered.

10. The high-precision PCB lamination control method based on machine learning according to claim 9, characterized in that, The process of ensuring that the target PCB lamination equipment's operating procedures and risk assessments comply with standards includes: Extract the comparison results of the process index characterization values ​​with the predetermined process index characterization thresholds; Calculate the difference between the quality indicator value and the predetermined quality indicator threshold; If the process index value is less than the predetermined process index threshold and the difference between the quality index value and the predetermined quality index threshold is less than the predetermined difference threshold, then the operating process and risk assessment of the target PCB lamination equipment are determined to meet the standards.

Images