Vacuum coating automatic control system based on multi-parameter cooperation

Abstract

The invention discloses a vacuum coating automatic control system based on multi-parameter collaboration, relates to the technical field of vacuum coating, and is characterized in that environmental data and film thickness data are acquired through a plurality of sensors, and abnormal value detection, denoising and standardization processing are carried out on the data, so that the reliability and uniformity of the data are ensured. Through the film layer quality detection module and the system, the film layer quality can be detected and calculated, different optimization strategies can be implemented according to detection results, and it is ensured that when it is detected that the film layer quality is lower than a qualified standard, environmental parameters can be optimized and adjusted immediately, so that the purpose of adjusting the film layer quality is achieved, and the stability of the film coating quality is ensured. Finally, through effective combination of a multi-parameter collaborative optimization module and an iterative optimization module, adaptive collaborative optimization adjustment is carried out on multiple parameters of the environment data, and it is ensured that the film layer quality is always close to a preset target value. The yield is obviously improved, and the overall stability and reliability of the system are enhanced.

Description

Technical Field

[0001] The present invention relates to the field of vacuum coating technology, and in particular to a vacuum coating automation control system based on multi-parameter collaboration. Background Art

[0002] Vacuum coating technology involves depositing materials onto substrate surfaces through physical or chemical methods in a vacuum environment to form a thin film. This technology is widely used in various fields, including optics, electronics, decoration, packaging, and energy, aiming to enhance the functionality and aesthetics of substrates by coating them with a thin film. With technological advancements, coating process requirements are becoming increasingly stringent, such as uniformity, adhesion, purity, and thickness accuracy. Improving the control accuracy and efficiency of the coating process in these complex application scenarios has become an unresolved issue.

[0003] In the traditional vacuum coating process, the temperature, air pressure, airflow and other parameters in the coating area often fluctuate due to some external factors, resulting in uneven thickness of the film layer in different areas. This not only affects the quality of the product, but also reduces production efficiency. Existing control systems often rely on manual adjustment or automatic adjustment of a single parameter, making it difficult to achieve comprehensive and accurate automated multi-parameter coordinated adjustment, and thus difficult to ensure that the uniformity and stability of the film layer meet the qualified standards. Especially in the face of vacuum systems, when the thickness and uniformity of the film layer deviate from the system's preset standards due to parameter fluctuations, traditional control methods are unable to make real-time adaptive changes based on the current thickness and uniformity of the film layer, and achieve effective coordination and optimization between multiple parameters during the production stage, resulting in the production of unqualified defective products, which in turn results in high financial and time costs for rework. Summary of the Invention

[0004] In view of the deficiencies in the prior art, the present invention provides a vacuum coating automation control system based on multi-parameter collaboration, which solves the problems mentioned in the background technology.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: a vacuum coating automation control system based on multi-parameter collaboration, including a coating parameter acquisition module, a coating parameter preprocessing module, a film quality detection module, a multi-parameter collaborative optimization module and an iterative optimization module;

[0006] The coating parameter acquisition module is used to collect environmental data and film thickness data of multiple regional monitoring points in the coating equipment, combine the environmental data of multiple regional monitoring points to form an original environmental data set RD, and combine the film thickness data of multiple regional monitoring points to form a film thickness data set H;

[0007] The coating parameter preprocessing module performs outlier detection on the original environmental data set RD. After the detection is completed, the original environmental data set RD is preprocessed to obtain the preprocessed environmental data feature vector set FA;

[0008] The film quality detection module performs quality detection on the film thickness data set H at time t, and outputs the quality detection result MQ, and performs optimization decision-making based on the quality detection result MQ;

[0009] The multi-parameter collaborative optimization module uses the optimization adjustment algorithm on the environmental data feature vector set FA at time t to calculate the environmental data adjustment amount set FB of each area, and uses the optimization control algorithm on the environmental data adjustment amount set FB to obtain the optimized target environmental data feature vector set FC;

[0010] The iterative optimization module detects the optimized data and outputs the film quality optimization result OC (j, t), and iteratively optimizes the optimization adjustment algorithm based on the film quality optimization result OC (j, t).

[0011] Preferably, the coating parameter acquisition module includes an environmental data acquisition unit and a film thickness acquisition unit;

[0012] The environmental data acquisition unit sets up multiple environmental sensors in different areas of the coating equipment, and each environmental sensor collects various environmental parameter data of different areas in the equipment in real time;

[0013] Environmental sensors include temperature sensors, air pressure sensors, air flow rate sensors, and power sensors;

[0014] Environmental parameter data include temperature C, air pressure P, air flow rate V and power W. The environmental parameter data are marked based on the region j and time t where the environmental parameter data are obtained. The marked data are temperature C(j, t), air pressure P(j, t), gas flow rate V(j, t) and power W(j, t). The marked environmental parameter data are integrated to form the original environmental data set RD.

[0015] Preferably, the film thickness acquisition unit sets multiple distance sensors in different areas within the coating equipment to collect the distance from the plated parts in different areas to the distance sensors in real time, obtains the film thickness data of different areas by calculating the difference between the distance value at time t and the initial distance value, and integrates the film thickness data to form a film thickness data set H.

[0016] Preferably, the coating parameter preprocessing module includes an abnormal value detection unit and a preprocessing unit;

[0017] The outlier detection unit performs traversal detection on the original environmental data set RD;

[0018] If the parameter of item i in the original environmental data set RD meets the preset standard range itarget, the coating operation is continued and data preprocessing is performed;

[0019] If the i-th parameter data in the original environmental data set RD deviates from the preset standard range itarget, the environmental sensor that collects the i-th parameter is self-tested to obtain the sensor self-test result;

[0020] If the sensor self-test result is a sensor failure, causing the i-th parameter data in the original environmental data set RD to deviate from the preset standard range itarget, the coating operation is stopped and the system is exited, and an alarm prompt is issued;

[0021] If the sensor self-test result shows that the sensor is normal, it is determined that due to external factors, including power W(j, t) fluctuations, the i-th parameter data in the original environmental data set RD deviates from the preset standard range itarget, then the coating operation is continued and data preprocessing is performed.

[0022] Preferably, the preprocessing unit preprocesses the original environmental data set RD, including data denoising preprocessing and data standardization preprocessing, to obtain a preprocessed environmental data feature vector set FA;

[0023] Data denoising preprocessing is to denoise the original environmental data set RD by using a digital signal processor to perform Kalman filtering algorithm;

[0024] The data standardization preprocessing performs Z-Score standardization on the original environmental data set RD after denoising preprocessing to obtain the environmental data feature vector FA after data standardization preprocessing.

[0025] Preferably, the film quality detection module includes a quality detection unit and a decision unit;

[0026] The quality inspection unit performs quality inspection on the film thickness data set H at time t and calculates the film thickness variance σ at time t 2 Ht and film thickness standard deviation σH(j, t);

[0027] By the preset film thickness variance threshold θσ 2 Ht and the standard deviation threshold θσH(j, t) of the film thickness in region j and the film thickness variance σ 2 Compare Ht with the standard deviation of film thickness σH(j, t), and output the quality detection result MQt at time t based on the comparison result;

[0028] .

[0029] Preferably, the decision unit makes an optimization decision based on the quality test result MQt;

[0030] If the quality inspection result MQt=1, the current film quality inspection result is determined to be qualified, and the environmental data feature vector set FA(j, t) of region j at time t is recorded. The environmental data feature vector set FA(j, t) of region j at time t is used as a multi-parameter coordinated reasonable solution and included in the optimization reference opinion. An optimization report is generated, and the optimization work at time t is ended. The optimization work at time t+1 begins. The optimization work includes constructing an adjustment algorithm, calculating the environmental data adjustment amount set FB, constructing a control algorithm based on the environmental data adjustment amount set FB, calculating the optimized target environmental data feature vector set FC, using the target environmental data feature vector set FC to replace the environmental data feature vector set FA at time t+1 and applying it, constructing an iterative optimization algorithm, and optimizing the adjustment algorithm in real time.

[0031] If the quality test result MQt=0, the current film quality test result is judged to be unqualified and the process enters the secondary decision-making stage;

[0032] During the secondary decision-making process, if the quality inspection result MQt-1 at time t-1 is 1, it is determined that the coating process at time t is not in a continuous optimization state, and the multi-parameter collaborative optimization module is executed;

[0033] During the secondary decision-making, if the quality inspection result MQt-1 at time t-1 is 0, it is determined that the coating work at time t is in a continuous optimization state, and the iterative optimization module is executed.

[0034] Preferably, the multi-parameter collaborative optimization module includes an adjustment unit and a control unit;

[0035] The adjustment unit constructs an adjustment algorithm based on the environmental data feature vector set FA and the film thickness standard deviation σH (j, t) at time t to obtain the environmental data adjustment amount set FB;

[0036] The adjustment algorithm formula is as follows:

[0037] ;

[0038] ;

[0039] Wherein, itargetMED is the median of the preset standard range itarget, itargetMAX is the maximum value of the preset standard range itarget, itargetMIN is the minimum value of the preset standard range itarget, ki(j, t) is the adjustment coefficient of the i-th parameter in region j at time t, FAi(j, t) is the i-th parameter of the environmental data feature vector set FA in region j at time t, and sig is the Sigmoid activation function.

[0040] Preferably, the control unit constructs a control algorithm based on the environmental data feature vector set FA and the environmental data adjustment amount set FB at time t, obtains an optimized target environmental data feature vector set FC, uses the optimized target environmental data feature vector set FC to replace the environmental data feature vector set FA at time t+1, performs denormalization processing on the environmental data feature vector set FA at time t+1, and applies the obtained result to replace the original environmental parameter set RD at time t+1;

[0041] The control algorithm formula is as follows:

[0042] ;

[0043] Where FBi(j, t) is the i-th parameter of the environmental data adjustment set FB in region j at time t.

[0044] Preferably, the iterative optimization module includes an iterative optimization unit;

[0045] The iterative optimization unit performs film quality optimization detection calculation based on the film thickness standard deviation σH(j, t) at time t, the film thickness average value μHt at time t, and the film thickness standard deviation σH(j, t-1) at time t-1, and obtains the film quality optimization result OC(j, t) after the optimization work at time t-1, and adaptively optimizes the adjustment coefficient ki(j, t) according to the film quality optimization result OC(j, t);

[0046] Among them, the film quality optimization detection formula is:

[0047] ;

[0048] Where μHt represents the average thickness of the film layer at time t, △T represents the time step, w1 represents the weight coefficient of the ratio of the standard deviation of the film thickness σH(j, t) to the average thickness μHt of the film layer, and w2 represents the weight coefficient of the rate of change of the standard deviation of the film thickness σH(j, t);

[0049] The adaptive optimization adjustment coefficient formula is:

[0050] ;

[0051] Where ki(j, t-1) represents the adjustment coefficient at time t-1, and ∂ represents the sign of the partial derivative;

[0052] The adaptively optimized adjustment coefficient ki(j, t) is input into the adjustment unit to replace the adjustment coefficient ki(j, t-1) at time t-1, and the environmental data adjustment amount set FB at time t is calculated based on the replaced adjustment coefficient ki(j, t).

[0053] The present invention provides a vacuum coating automation control system based on multi-parameter collaboration, which has the following beneficial effects:

[0054] (1) When the system is running, by configuring multiple sensors in different areas, it monitors and collects the environmental parameters and film thickness data of each area in real time and integrates them into the original environmental data set RD and film thickness data H, ensuring the comprehensiveness and accuracy of the data source and providing a reliable basis for subsequent data processing. Through the coating parameter preprocessing module, the system first detects abnormal values in the original environmental parameter set RD. When abnormal values appear, it determines whether the abnormality of the detection data is caused by sensor failure, so as to avoid the data abnormality caused by sensor failure affecting the accuracy of subsequent collected data, so that the coating quality cannot meet expectations. After eliminating the cause of the sensor failure, the original environmental data set RD is denoised and normalized to generate the preprocessed environmental data feature vector set FA. These high-quality data will provide a reliable data basis for the multi-parameter collaborative adjustment module, enhance the overall stability and reliability of the system, and thus improve the stability of the film quality.

[0055] (2) By performing a detailed quality inspection on the film thickness data set H at the current moment, different optimization strategies are implemented according to the inspection results to ensure that when the film thickness quality is detected to be below the unqualified level, the environmental parameters can be optimized and adjusted immediately, thereby achieving the purpose of adjusting the film quality and ensuring the stability of the coating quality. Furthermore, when the coating quality is detected to be unqualified, the system module to be executed next is analyzed based on the real-time status, so that the system can dynamically respond to the adjustment changes of the multi-parameter collaborative optimization module until the film quality inspection results return to qualified. At this time, the optimization adjustment state is exited and the parameters in the optimization adjustment process are recorded as historical reference data.

[0056] (3) Through the effective combination of the multi-parameter collaborative optimization module and the iterative optimization module, the coating process is optimized and continuously adaptively optimized and feedback controlled in real time. The multi-parameter collaborative optimization module uses the optimization adjustment algorithm to calculate the environmental data adjustment required for each area based on the current environmental data feature vector set and the standard deviation of the film thickness. The algorithm comprehensively considers the initial value preset by the historical data, the current environmental data feature vector and its standard deviation, and adjusts the change amplitude of each parameter through a specific function to ensure that the adjustment process is smooth and efficient. The optimization control algorithm further processes these adjustments to generate the final optimized environmental data feature vector set, thereby accurately controlling the various environmental parameters in the coating equipment. On this basis, the iterative optimization module performs a comprehensive quality assessment by calculating the standard deviation, average value and standard deviation of the film thickness at the current moment, and the film thickness standard deviation at the previous moment, and adaptively adjusts the optimization coefficient based on the assessment results. This closed-loop feedback mechanism can not only monitor the changes in the film thickness in real time, but also make adjustments quickly to ensure that the film quality is always close to the preset target value. In this way, the system can effectively coordinate and optimize multiple parameters during the production process, adaptively optimize the film quality in real time, and ensure the high precision and stability of the film quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 This is a schematic diagram of a block diagram of a vacuum coating automation control system based on multi-parameter collaboration according to the present invention;

[0058] Figure 2 This is a schematic diagram of the vacuum coating process flow chart of a vacuum coating automation control system based on multi-parameter collaboration in the present invention. DETAILED DESCRIPTION

[0059] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0060] Example 1

[0061] The present invention provides a vacuum coating automation control system based on multi-parameter coordination, please refer to Figure 1 , including coating parameter acquisition module, coating parameter preprocessing module, film quality detection module, multi-parameter collaborative optimization module and iterative optimization module;

[0062] The coating parameter acquisition module is used to collect environmental data and film thickness data of multiple regional monitoring points in the coating equipment, combine the environmental data of multiple regional monitoring points to form an original environmental data set RD, and combine the film thickness data of multiple regional monitoring points to form a film thickness data set H;

[0063] The coating parameter preprocessing module performs outlier detection on the original environmental data set RD. After the detection is completed, the original environmental data set RD is preprocessed to obtain the preprocessed environmental data feature vector set FA;

[0064] The film quality detection module performs quality detection on the film thickness data set H at time t, and outputs the quality detection result MQ, and performs optimization decision-making based on the quality detection result MQ;

[0065] The multi-parameter collaborative optimization module uses the optimization adjustment algorithm on the environmental data feature vector set FA at time t to calculate the environmental data adjustment amount set FB of each area, and uses the optimization control algorithm on the environmental data adjustment amount set FB to obtain the optimized target environmental data feature vector set FC;

[0066] The iterative optimization module detects the optimized data and outputs the film quality optimization result OC (j, t), and iteratively optimizes the optimization adjustment algorithm based on the film quality optimization result OC (j, t).

[0067] In this embodiment, by setting up multiple types of sensors in multiple areas to detect and collect environmental parameters and film thickness data of each area in real time, first, the original environmental data set RD is detected for outliers, and the Kalman filter algorithm and Z-Score standardization are used to generate a pre-processed environmental data feature vector set FA, thereby achieving data denoising and normalization, ensuring the accuracy and consistency of the data, and thus avoiding the inability to obtain correct environmental parameters due to sensor failure, which affects the coating quality. Secondly, by performing a detailed quality inspection on the film thickness data set H at time t, the film quality inspection result MQ is obtained. Based on the obtained film quality inspection result MQ, the system can promptly identify whether the film quality is qualified. When an unqualified result occurs, the optimization strategy is automatically executed according to the inspection result MQ, thereby achieving the purpose of adjusting the film quality and ensuring the stability of the coating quality. Subsequently, the multi-parameter collaborative optimization module uses an optimization adjustment algorithm to adjust the environmental data feature vector set FA during the coating process, deriving the environmental data adjustment amount set FB for each region. Based on the obtained environmental data adjustment amount set FB, a control algorithm is constructed to obtain the optimized target environmental data feature vector set FC, achieving precise control of the coating environmental conditions and thus optimizing the film quality. Combined with the iterative optimization module, the optimized coating process is further continuously monitored and dynamically adjusted. Based on the film quality optimization result OC(j, t), an adaptive real-time improvement optimization strategy is implemented to ensure that the optimization efficiency of the film quality can be steadily improved. Compared with traditional vacuum coating technology, this system reduces product quality issues caused by parameter fluctuations, achieves adaptive real-time collaborative optimization between multiple parameters, improves the yield of coated products, and brings higher economic benefits and technological advantages to enterprises.

[0068] Example 2

[0069] This embodiment is explained in Example 1, please refer to Figure 1 and Figure 2 ,Specifically: the coating parameter acquisition module includes an ,environmental data acquisition unit and a film thickness acquisition unit;

[0070] The environmental data acquisition unit sets up multiple environmental sensors in different areas of the coating equipment, and each environmental sensor collects various environmental parameter data of different areas in the equipment in real time;

[0071] Environmental sensors include temperature sensors, air pressure sensors, air flow rate sensors, and power sensors;

[0072] Environmental parameter data include temperature C, air pressure P, air flow rate V and power W. The environmental parameter data are marked based on the region j and time t where the environmental parameter data are obtained. The marked data are temperature C(j, t), air pressure P(j, t), gas flow rate V(j, t) and power W(j, t). The marked environmental parameter data are integrated to form the original environmental data set RD.

[0073] The film thickness acquisition unit sets multiple distance sensors in different areas of the coating equipment to collect the distance from the plated parts in different areas to the distance sensors in real time. The film thickness data of different areas are obtained by calculating the difference between the distance value at time t and the initial distance value. The film thickness data are integrated to form a film thickness data set H.

[0074] The coating parameter preprocessing module includes an abnormal value detection unit and a preprocessing unit;

[0075] The outlier detection unit performs traversal detection on the original environmental data set RD;

[0076] If the parameter of item i in the original environmental data set RD meets the preset standard range itarget, the coating operation is continued and data preprocessing is performed;

[0077] The preset standard range itarge is specifically the range value of various environmental data input based on historical data according to the selection of film materials and plated parts by the staff;

[0078] If the i-th parameter data in the original environmental data set RD deviates from the preset standard range itarget, the environmental sensor that collects the i-th parameter is self-tested to obtain the sensor self-test result;

[0079] If the sensor self-test result is a sensor failure, causing the i-th parameter data in the original environmental data set RD to deviate from the preset standard range itarget, the coating operation is stopped and the system is exited, and an alarm prompt is issued;

[0080] If the sensor self-test result shows that the sensor is normal, it is determined that due to external factors, including power W(j, t) fluctuations, the i-th parameter data in the original environmental data set RD deviates from the preset standard range itarget, then the coating operation is continued and data preprocessing is performed.

[0081] The preprocessing unit preprocesses the original environmental data set RD, including data denoising preprocessing and data standardization preprocessing, to obtain a preprocessed environmental data feature vector set FA;

[0082] Data denoising preprocessing is to denoise the original environmental data set RD by using a digital signal processor to perform Kalman filtering algorithm;

[0083] The data standardization preprocessing performs Z-Score standardization on the original environmental data set RD after denoising preprocessing to obtain the environmental data feature vector FA after data standardization preprocessing.

[0084] In this embodiment, by deploying temperature, pressure, airflow rate, and power sensors in multiple zones, accurate real-time acquisition and dynamic labeling of environmental parameters across different regions within the coating equipment are achieved. These parameters include temperature C(j, t), pressure P(j, t), air flow rate V(j, t), and power W(j, t). Combined with the multi-zone synchronous detection of the distance sensors in the film thickness acquisition unit, this ensures the spatial resolution and temporal continuity of the environmental data set RD and the film thickness data set H. The outlier detection unit effectively distinguishes the source of data anomalies through item-by-item comparison within a preset standard range itarget and a sensor self-checking mechanism. If a sensor failure causes a deviation from the preset standard range itarget, the operation is immediately terminated and an alarm is issued, preventing erroneous data from interfering with subsequent coating operations. If an anomaly is caused by external fluctuations, the operation continues and preprocessing is performed, improving data reliability. The preprocessing unit uses a Kalman filter algorithm to denoise the raw environmental data set RD, eliminating random interference. Combined with Z-Score normalization, it generates a highly consistent set of environmental data feature vectors FA, providing low-noise, high-precision input data for the subsequent multi-parameter collaborative optimization module. This refined data collection and preprocessing mechanism not only enhances the time accuracy of parameter tracking, but also enhances the overall stability and reliability of the system by quickly locating abnormal sources and optimizing data quality, thereby improving the stability of film quality.

[0085] Example 3

[0086] This embodiment is explained in Example 2, please refer to Figure 1 and Figure 2 ,Specifically: the film quality detection module includes a quality detection unit and a ,decision-making unit;

[0087] The quality inspection unit performs quality inspection on the film thickness data set H at time t and calculates the film thickness variance σ at time t 2 Ht and film thickness standard deviation σH(j, t);

[0088] By the preset film thickness variance threshold θσ 2 Ht and the standard deviation threshold θσH(j, t) of the film thickness in region j and the film thickness variance σ 2Ht is compared with the standard deviation of the film thickness σH(j, t), and the quality detection result MQt at time t is output according to the comparison result, as shown in Table 1;

[0089] .

[0090] Specific examples:

[0091] Preset film thickness variance threshold θσ 2 Ht: 0.05;

[0092] The preset film thickness standard deviation threshold θσH(j, t): 0.02;

[0093] Array 1: Film thickness variance σ 2 Ht: 0.05, film thickness standard deviation σH (j, t): 0.015;

[0094] Array 2: Film thickness variance σ 2 Ht: 0.04, film thickness standard deviation σH (j, t): 0.025;

[0095] Array 3: Film thickness variance σ 2 Ht: 0.06, film thickness standard deviation σH (j, t): 0.018;

[0096] Array 4: Film thickness variance σ 2 Ht: 0.07, film thickness standard deviation σH (j, t): 0.025;

[0097] Table 1:

[0098] Array <![CDATA[σ 2 Ht]]> <![CDATA[θσ 2 Ht]]> σH(j, t) θσH(j,t) MQ Array 1 0.05 0.05 0.015 0.02 1 Array 2 0.04 0.05 0.025 0.02 0 Array 3 0.06 0.05 0.018 0.02 0 Array 4 0.07 0.05 0.025 0.02 0

[0099] The decision-making unit makes optimization decisions based on the quality test results MQt;

[0100] If the quality inspection result MQt=1, the current film quality inspection result is determined to be qualified, and the environmental data feature vector set FA(j, t) of region j at time t is recorded. The environmental data feature vector set FA(j, t) of region j at time t is used as a multi-parameter coordinated reasonable solution and included in the optimization reference opinion. An optimization report is generated, and the optimization work at time t is ended. The optimization work at time t+1 begins. The optimization work includes constructing an adjustment algorithm, calculating the environmental data adjustment amount set FB, constructing a control algorithm based on the environmental data adjustment amount set FB, calculating the optimized target environmental data feature vector set FC, using the target environmental data feature vector set FC to replace the environmental data feature vector set FA at time t+1 and applying it, constructing an iterative optimization algorithm, and optimizing the adjustment algorithm in real time.

[0101] If the quality test result MQt=0, the current film quality test result is judged to be unqualified and the process enters the secondary decision-making stage;

[0102] During the secondary decision-making process, if the quality inspection result MQt-1 at time t-1 is 1, it is determined that the coating process at time t is not in a continuous optimization state, and the multi-parameter collaborative optimization module is executed;

[0103] During the secondary decision-making, if the quality inspection result MQt-1 at time t-1 is 0, it is determined that the coating work at time t is in a continuous optimization state, and the iterative optimization module is executed.

[0104] In this embodiment, the quality detection unit in the film quality detection module is linked with the decision unit to achieve dynamic quality assessment and real-time optimization decision-making of the film thickness during the coating process. The quality detection unit accurately calculates the film thickness variance σ based on the film thickness data set H at time t. 2 Ht and the standard deviation of the film thickness in each region σH(j, t), through the preset variance threshold θσ 2 The comparison between Ht and the standard deviation threshold value θσH(j, t) of the film thickness in region j generates the binary quality detection result MQ. 2 The dual-layer threshold judgment of Ht and the standard deviation of the film thickness in each region σH(j, t) ensures the comprehensiveness of the quality assessment, focusing on both overall uniformity and regional consistency, thereby accurately identifying global deviations and local defects in complex coating scenarios.

[0105] The decision-making unit executes the differentiation strategy based on the value of the quality test result MQ;

[0106] If the quality inspection result MQt=1, the set of environmental data feature vectors FA(j, t) for region j at time t is recorded and used as a reference for optimization. If the quality inspection result MQt=0, a secondary decision-making mechanism is triggered. By tracing back the quality inspection result MQt-1 at time t-1, it is dynamically determined whether it is in a continuous optimization state. When the quality inspection result MQt-1 at time t-1 is 1, the multi-parameter collaborative optimization module is executed, and when the quality inspection result MQt-1 at time t-1 is 0, the iterative optimization module is executed. This hierarchical decision-making mechanism not only provides real-time closed-loop feedback on film quality, but also determines the next execution strategy based on historical state associations. By dynamically switching optimization modules, it avoids redundant calculations in qualified states while ensuring rapid response in unqualified states. Ultimately, while ensuring the stability of coating quality, it optimizes system resource utilization and improves the accuracy and fault tolerance of the optimization strategy.

[0107] Example 4

[0108] This embodiment is explained in Example 3, please refer to Figure 1 ,Specifically: the multi-parameter collaborative optimization module includes an adjustment unit and a control unit;

[0109] The adjustment unit constructs an adjustment algorithm based on the environmental data feature vector set FA and the film thickness standard deviation σH (j, t) at time t to obtain the environmental data adjustment amount set FB;

[0110] The adjustment algorithm formula is as follows:

[0111] ;

[0112] ;

[0113] Wherein, itargetMED is the median of the preset standard range itarget, itargetMAX is the maximum value of the preset standard range itarget, itargetMIN is the minimum value of the preset standard range itarget, ki(j, t) is the adjustment coefficient of the i-th parameter in region j at time t, specifically an initial value pre-set according to historical data, FAi(j, t) is the i-th parameter of the environmental data feature vector set FA in region j at time t, sig is the Sigmoid activation function, which is used to map the adjustment amount to a fixed range to ensure that the adjustment amount is not too large or too small, and sig(-σH(j, t)) represents the influence of the film thickness standard deviation σH(j, t) on the environmental parameter adjustment amount set FB.

[0114] The purpose of this formula is to ensure that the adjustment direction is consistent with the preset target by taking the median itargetMED of the preset standard range itarget as the benchmark and combining the deviation amplitude of the dynamic balance parameter of the adjustment coefficient ki(j, t). The denominator introduces sig(-σH(j, t)), and uses the Sigmaoid function to use the film thickness standard deviation σH(j, t) as a nonlinear weight. The adjustment weight decays smoothly as the parameter stability improves. The smaller the value of the film thickness standard deviation σH(j, t), the better the film quality uniformity. Only slight adjustments need to be made to the environmental parameter adjustment amount set FB to avoid excessive adjustments caused by local fluctuations.

[0115] The control unit constructs a control algorithm based on the environmental data feature vector set FA and the environmental data adjustment amount set FB at time t, obtains an optimized target environmental data feature vector set FC, replaces the environmental data feature vector set FA at time t+1 with the optimized target environmental data feature vector set FC, performs denormalization processing on the environmental data feature vector set FA at time t+1, and replaces the obtained result with the original environmental parameter set RD at time t+1;

[0116] The control algorithm formula is as follows:

[0117] ;

[0118] Where FBi(j, t) is the i-th parameter of the environmental data adjustment set FB in region j at time t.

[0119] The purpose of this formula is to use the Sigmoid function to perform nonlinear compression on the i-th parameter FBi(j, t) of the environmental data adjustment amount, bind the control process with the parameter convergence characteristics, limit its contribution ratio when the i-th parameter FBi(j, t) of the environmental parameter adjustment amount set is large, and gradually converge when approaching the target value, thereby avoiding the impact of step mutations on the system and ensuring the asymptotic stability of the optimization process.

[0120] In this embodiment, the system achieves precise control of environmental parameters during the coating process through the effective integration of the adjustment unit and control unit of the multi-parameter collaborative optimization module. The control unit constructs an adjustment algorithm based on the environmental data feature vector set FA at time t and the film thickness standard deviation σH(j, t), generating a set of environmental data adjustment variables FB. This process not only considers the median value itargetMED of the preset standard range of historical data but also utilizes the Sigmoid activation function to ensure the smoothness and rationality of the adjustment variables, preventing over-adjustment from affecting film quality. The control unit further uses the i-th parameter FBi(j, t) of the environmental data adjustment variable set, combined with the environmental data feature vector set FA, to construct an optimization control algorithm. By introducing the Sigmoid activation function sig, the adjustment variables are nonlinearly compressed, effectively preventing over-adjustment caused by local fluctuations. The optimized target environmental data feature vector FC is generated and applied to the subsequent coating process. These design features enable the system to adaptively adjust coating conditions in a dynamically changing environment, ensuring uniformity and stability of film thickness.

[0121] Example 5

[0122] This embodiment is explained in Example 4. Please refer to Figure 1 and Figure 2 ,Specifically: the iterative optimization module includes an iterative optimization unit;

[0123] The iterative optimization unit performs film quality optimization detection calculation based on the film thickness standard deviation σH(j, t) at time t, the film thickness average value μHt at time t, and the film thickness standard deviation σH(j, t-1) at time t-1, and obtains the film quality optimization result OC(j, t) after the optimization work at time t-1, and adaptively optimizes the adjustment coefficient ki(j, t) according to the film quality optimization result OC(j, t);

[0124] Among them, the film quality optimization detection formula is:

[0125] ;

[0126] Where μHt represents the average thickness of the film at time t, △T represents the time step, w1 represents the weight coefficient of the ratio of the standard deviation of the film thickness σH(j, t) to the average film thickness μHt, w2 represents the weight coefficient of the rate of change of the standard deviation of the film thickness σH(j, t), Indicates the rate of change of the standard deviation of film thickness σH(j, t);

[0127] The adaptive optimization adjustment coefficient formula is:

[0128] ;

[0129] Where ki(j, t-1) represents the adjustment coefficient at time t-1, ∂ represents the sign of the partial derivative, Indicates the sensitivity of the i-th parameter FBi(j, t) of the environmental data adjustment set to the film quality optimization result OC(j, t). The larger the value, the greater the positive influence of the i-th parameter FBi(j, t) of the current environmental data adjustment set on the film quality optimization result OC(j, t). Conversely, the smaller the positive influence.

[0130] The adaptively optimized adjustment coefficient ki(j, t) is input into the adjustment unit to replace the adjustment coefficient ki(j, t-1) at time t-1, and the environmental data adjustment amount set FB at time t is calculated based on the replaced adjustment coefficient ki(j, t).

[0131] In this embodiment, the iterative optimization module achieves continuous monitoring and dynamic adjustment of the coating process through an iterative optimization unit. The iterative optimization unit performs film quality optimization detection and calculation based on the film thickness standard deviation σH(j, t) at time t, the average film thickness μHt at time t, and the film thickness standard deviation σH(j, t-1) at time t-1, obtaining the film quality optimization result OC(j, t). Based on the film quality optimization result OC(j, t), the system can adaptively optimize the adjustment coefficient ki(j, t) to more accurately control the coating environment parameters. This mechanism not only ensures continuous improvement of film quality during the coating process, but also achieves detailed capture and response to changing trends in coating quality by considering the film thickness standard deviation σH(j, t-1) and the average film thickness μHt and the rate of change of the film thickness standard deviation, respectively, through weight coefficients w1 and w2. By using adaptive optimization adjustment coefficients ki(j, t), the system can fine-tune the adjustment of various environmental parameters in each iteration to adapt to subtle changes in the coating process, greatly improving the stability and consistency of film quality. The iterative optimization unit enables real-time feedback and self-learning during the coating process, ensuring that film quality always approaches the preset target value, while reducing quality fluctuations between batches and improving production efficiency and yield.

[0132] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A vacuum coating automation control system based on multi-parameter collaboration, characterized by: It includes coating parameter acquisition module, coating parameter preprocessing module, film quality detection module, multi-parameter collaborative optimization module and iterative optimization module; The coating parameter acquisition module is used to collect environmental data and film thickness data of multiple regional monitoring points in the coating equipment, combine the environmental data of multiple regional monitoring points to form an original environmental data set RD, and combine the film thickness data of multiple regional monitoring points to form a film thickness data set H; The coating parameter preprocessing module performs outlier detection on the original environmental data set RD. After the detection is completed, the original environmental data set RD is preprocessed to obtain the preprocessed environmental data feature vector set FA; The film quality detection module performs quality detection on the film thickness data set H at time t, and outputs the quality detection result MQ, and performs optimization decision-making based on the quality detection result MQ; The multi-parameter collaborative optimization module uses the optimization adjustment algorithm on the environmental data feature vector set FA at time t to calculate the environmental data adjustment amount set FB of each area, and uses the optimization control algorithm on the environmental data adjustment amount set FB to obtain the optimized target environmental data feature vector set FC; The iterative optimization module detects the optimized data and outputs the film quality optimization result OC (j, t), and iteratively optimizes the optimization adjustment algorithm based on the film quality optimization result OC (j, t).

2. The vacuum coating automation control system based on multi-parameter collaboration according to claim 1, characterized in that: The coating parameter acquisition module includes an environmental data acquisition unit and a film thickness acquisition unit; The environmental data acquisition unit sets up multiple environmental sensors in different areas of the coating equipment, and each environmental sensor collects various environmental parameter data of different areas in the equipment in real time; Environmental sensors include temperature sensors, air pressure sensors, air flow rate sensors, and power sensors; Environmental parameter data include temperature C, air pressure P, air flow rate V and power W. The environmental parameter data are marked based on the region j and time t where the environmental parameter data are obtained. The marked data are temperature C(j, t), air pressure P(j, t), gas flow rate V(j, t) and power W(j, t). The marked environmental parameter data are integrated to form the original environmental data set RD.

3. The vacuum coating automation control system based on multi-parameter collaboration according to claim 2, characterized in that: The film thickness acquisition unit sets multiple distance sensors in different areas of the coating equipment to collect the distance from the plated parts in different areas to the distance sensors in real time. The film thickness data of different areas are obtained by calculating the difference between the distance value at time t and the initial distance value. The film thickness data are integrated to form a film thickness data set H.

4. The vacuum coating automation control system based on multi-parameter collaboration according to claim 3, characterized in that: The coating parameter preprocessing module includes an abnormal value detection unit and a preprocessing unit; The outlier detection unit performs traversal detection on the original environmental data set RD; If the parameter of item i in the original environmental data set RD meets the preset standard range itarget, the coating operation is continued and data preprocessing is performed; If the i-th parameter data in the original environmental data set RD deviates from the preset standard range itarget, the environmental sensor that collects the i-th parameter is self-tested to obtain the sensor self-test result; If the sensor self-test result is a sensor failure, causing the i-th parameter data in the original environmental data set RD to deviate from the preset standard range itarget, the coating operation is stopped and the system is exited, and an alarm prompt is issued; If the sensor self-test result shows that the sensor is normal, it is determined that due to external factors, including power W(j, t) fluctuations, the i-th parameter data in the original environmental data set RD deviates from the preset standard range itarget, then the coating operation is continued and data preprocessing is performed.

5. The vacuum coating automation control system based on multi-parameter collaboration according to claim 4, characterized in that: The preprocessing unit preprocesses the original environmental data set RD, including data denoising preprocessing and data standardization preprocessing, to obtain a preprocessed environmental data feature vector set FA; Data denoising preprocessing is to denoise the original environmental data set RD by using a digital signal processor to perform Kalman filtering algorithm; The data standardization preprocessing performs Z-Score standardization on the original environmental data set RD after denoising preprocessing to obtain the environmental data feature vector FA after data standardization preprocessing.

6. The vacuum coating automation control system based on multi-parameter collaboration according to claim 5, characterized in that: The film quality detection module includes a quality detection unit and a decision-making unit; The quality detection unit performs quality detection on the film thickness data set H at time t and calculates the film thickness variance σ at time t 2 Ht and film thickness standard deviation σH(j, t); By the preset film thickness variance threshold θσ 2 Ht and the standard deviation threshold θσH(j, t) of the film thickness in region j and the film thickness variance σ 2 Compare Ht with the standard deviation of film thickness σH(j, t), and output the quality inspection result MQt at time t according to the comparison result; 。 7. The vacuum coating automation control system based on multi-parameter collaboration according to claim 6, characterized in that: The decision-making unit makes optimization decisions based on the quality test results MQt; If the quality inspection result MQt=1, the current film quality inspection result is determined to be qualified, and the environmental data feature vector set FA(j, t) of region j at time t is recorded. The environmental data feature vector set FA(j, t) of region j at time t is used as a multi-parameter coordinated reasonable solution and included in the optimization reference opinion. An optimization report is generated, and the optimization work at time t is ended. The optimization work at time t+1 begins. The optimization work includes constructing an adjustment algorithm, calculating the environmental data adjustment amount set FB, constructing a control algorithm based on the environmental data adjustment amount set FB, calculating the optimized target environmental data feature vector set FC, using the target environmental data feature vector set FC to replace the environmental data feature vector set FA at time t+1 and applying it, constructing an iterative optimization algorithm, and optimizing the adjustment algorithm in real time. If the quality test result MQt=0, the current film quality test result is judged to be unqualified and the process enters the secondary decision-making stage; During the secondary decision-making process, if the quality inspection result MQt-1 at time t-1 is 1, it is determined that the coating process at time t is not in a continuous optimization state, and the multi-parameter collaborative optimization module is executed; During the secondary decision-making, if the quality inspection result MQt-1 at time t-1 is 0, it is determined that the coating work at time t is in a continuous optimization state, and the iterative optimization module is executed.

8. The vacuum coating automation control system based on multi-parameter collaboration according to claim 7, characterized in that: The multi-parameter collaborative optimization module includes an adjustment unit and a control unit; The adjustment unit constructs an adjustment algorithm based on the environmental data feature vector set FA and the film thickness standard deviation σH (j, t) at time t to obtain the environmental data adjustment amount set FB; The adjustment algorithm formula is as follows: ； ； Wherein, itargetMED is the median of the preset standard range itarget, itargetMAX is the maximum value of the preset standard range itarget, itargetMIN is the minimum value of the preset standard range itarget, ki(j, t) is the adjustment coefficient of the i-th parameter in region j at time t, FAi(j, t) is the i-th parameter of the environmental data feature vector set FA in region j at time t, and sig is the Sigmoid activation function.

9. The vacuum coating automation control system based on multi-parameter collaboration according to claim 8, characterized in that: The control unit constructs a control algorithm based on the environmental data feature vector set FA and the environmental data adjustment amount set FB at time t, obtains an optimized target environmental data feature vector set FC, replaces the environmental data feature vector set FA at time t+1 with the optimized target environmental data feature vector set FC, performs denormalization processing on the environmental data feature vector set FA at time t+1, and replaces the obtained result with the original environmental parameter set RD at time t+1; The control algorithm formula is as follows: ； Where FBi(j, t) is the i-th parameter of the environmental data adjustment set FB in region j at time t.

10. The vacuum coating automation control system based on multi-parameter collaboration according to claim 7, characterized in that: The iterative optimization module includes an iterative optimization unit; The iterative optimization unit performs film quality optimization detection calculation based on the film thickness standard deviation σH(j, t) at time t, the film thickness average value μHt at time t, and the film thickness standard deviation σH(j, t-1) at time t-1, and obtains the film quality optimization result OC(j, t) after the optimization work at time t-1, and adaptively optimizes the adjustment coefficient ki(j, t) according to the film quality optimization result OC(j, t); Among them, the film quality optimization detection formula is: ； Where μHt represents the average thickness of the film layer at time t, △T represents the time step, w1 represents the weight coefficient of the ratio of the standard deviation of the film thickness σH(j, t) to the average thickness μHt of the film layer, and w2 represents the weight coefficient of the rate of change of the standard deviation of the film thickness σH(j, t); The adaptive optimization adjustment coefficient formula is: ； Where ki(j, t-1) represents the adjustment coefficient at time t-1, and ∂ represents the sign of the partial derivative; The adaptively optimized adjustment coefficient ki(j, t) is input into the adjustment unit to replace the adjustment coefficient ki(j, t-1) at time t-1, and the environmental data adjustment amount set FB at time t is calculated based on the replaced adjustment coefficient ki(j, t).

Images