PVD (Physical Vapor Deposition) processing monitoring method, PVD processing prediction model training method and related equipment

Abstract

The invention relates to the field of monitoring, and provides a PVD processing monitoring method, a PVD processing prediction model training method and related equipment. The PVD processing monitoring method comprises the following steps: acquiring PVD processing data when processing equipment processes a product; performing feature processing on the PVD processing data to obtain PVD processing features, the PVD processing data including a plurality of PVD processing factors, and correspondingly, the PVD processing features including a plurality of PVD feature factors; the multiple PVD characteristic factors are input into the PVD processing prediction model to obtain the product demolding value, the method can accurately predict the demolding condition of the product so as to analyze the PVD processing factors which have important influences on the product demolding value, and therefore processing equipment or the PVD processing prediction model is dynamically adjusted to monitor the important PVD processing factors in real time.

Description

Technical Field

[0001] The present application relates to the field of monitoring, and particularly to a monitoring method for PVD processing, a method for training a PVD processing prediction model, and related devices. Background Art

[0002] Currently, in the process of processing products using equipment such as Physical Vapor Deposition (PVD), if components in the PVD equipment malfunction, it will cause the product to have a film peeling situation. To avoid the product having a film peeling situation, the PVD equipment is usually comprehensively maintained regularly. However, the comprehensive maintenance of the PVD equipment will cause some components to be unnecessarily maintained, which will lead to a waste of operation and maintenance resources. Summary of the Invention

[0003] The present application provides a monitoring method for PVD processing, a method for training a PVD processing prediction model, and related devices to solve the technical problem of being unable to accurately predict the film peeling situation of products.

[0004] In a first aspect of an embodiment of the present application, a monitoring method for PVD processing is provided. The method includes: obtaining PVD processing data of a processing device when processing a product; performing feature processing on the PVD processing data to obtain PVD processing features, where the PVD processing data includes multiple PVD processing factors, and correspondingly, the PVD processing features include multiple PVD feature factors; inputting multiple PVD feature factors into a PVD processing prediction model to obtain a product film peeling value.

[0005] In a second aspect of an embodiment of the present application, a method for training a PVD processing prediction model is provided. The method includes: obtaining multiple groups of historical PVD processing data and the historical product film peeling values corresponding to each product respectively; performing feature processing on the multiple groups of historical PVD processing data to obtain historical PVD processing features, where each group of PVD processing data includes multiple historical PVD processing factors, and correspondingly, the historical PVD processing features include multiple historical PVD feature factors; performing feature processing on the historical product film peeling values to obtain historical product film peeling feature values; training an initial prediction model based on multiple historical PVD feature factors, the historical product film peeling values, and the historical product film peeling feature values to obtain a PVD processing prediction model.

[0006] In a third aspect of an embodiment of the present application, an electronic device is provided. The electronic device includes: a memory storing computer-readable instructions; and a processor executing the computer-readable instructions stored in the memory to implement the above-mentioned monitoring method for PVD processing and the method for training a PVD processing prediction model.

[0007] A fourth aspect of the embodiments of the present application provides a computer-readable storage medium, in which computer-readable instructions are stored, and the computer-readable instructions are executed by a processor in an electronic device to implement the monitoring method for PVD processing and the training method for the PVD processing prediction model.

[0008] In multiple embodiments of the present application, by performing feature processing on the PVD processing data, PVD processing features can be obtained, which can assist the PVD processing prediction model in understanding various PVD processing factors. Through the analysis of various PVD feature factors by the PVD processing prediction model, an accurate product demolding value can be obtained, so as to analyze the PVD processing factors that have an important impact on the product demolding value, and thereby dynamically adjust the processing equipment or the PVD processing prediction model to monitor important PVD processing factors in real time. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Figure 1 It is an application scenario diagram of the monitoring method for PVD processing and the training method for the PVD processing prediction model provided by the embodiments of the present application.

[0010] Figure 2 It is a flowchart of the monitoring method for PVD processing provided by the embodiments of the present application.

[0011] Figure 3 It is a schematic diagram of the curve of the etching temperature provided by the embodiments of the present application.

[0012] Figure 4 It is a schematic diagram of the curve of the etching current provided by the embodiments of the present application.

[0013] Figure 5 It is a flowchart of the monitoring method for PVD processing provided by another embodiment of the present application.

[0014] Figure 6 It is a schematic diagram showing the key PVD feature factors provided by the embodiments of the present application.

[0015] Figure 7 It is a flowchart of the training method for the PVD processing prediction model provided by the embodiments of the present application.

[0016] Figure 8 It is a schematic diagram of the product in the demolding state provided by the embodiments of the present application.

[0017] Figure 9 It is a specific flowchart of training the initial prediction model provided by the embodiments of the present application.

[0018] Figure 10 It is a flowchart of the method for judging whether the multiple historical product demolding values of each product are independent provided by the embodiments of the present application.

[0019] Figure 11 It is a flowchart of a method for judging whether the data distributions of multiple historical product demolding values of each product provided in the embodiments of the present application are the same.

[0020] Figure 12 It is a schematic structural diagram of an electronic device according to an embodiment of a method for monitoring PVD processing or a method for training a PVD processing prediction model implemented in the present application. Detailed implementation manners

[0021] To make the objectives, technical solutions, and advantages of the present application clearer, the present application will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0022] It should be noted that in the present application, "at least one" means one or more, and "a plurality" means two or more than two. "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B may represent: A exists alone, A and B exist simultaneously, and B exists alone, where A and B may be singular or plural. Terms such as "first", "second", "third", "fourth", etc. (if any) in the specification, claims, and drawings of the present application are used to distinguish similar objects, rather than to describe a specific order or sequence.

[0023] In the embodiments of the present application, words such as "exemplary" or "for example" are used to represent examples, illustrations, or explanations. Any embodiment or design solution described as "exemplary" or "for example" in the embodiments of the present application should not be construed as being more preferred or having more advantages than other embodiments or design solutions. Rather, the use of words such as "exemplary" or "for example" is intended to present related concepts in a specific manner. Without conflict, the following embodiments and the features in the embodiments may be combined with each other.

[0024] As Figure 1 shown, it is an application scenario diagram of a method for monitoring PVD processing and a method for training a PVD processing prediction model provided in the embodiments of the present application.

[0025] In the embodiments of the present application, the method for monitoring PVD processing and the method for training a PVD processing prediction model can be applied to an electronic device, and the electronic device may include various types of processing devices 100, servers 200, computer terminals, etc.

[0026] The processing device 100 can be a Physical Vapor Deposition (PVD) device. The processing device 100 is used for coating the surface of a product, and can also be used for film plating on the surface of the product. The specific type of the processing device 100 is not particularly limited in the embodiments of the present application. The server 200 can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers. It can also be a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, Content Delivery Network (CDN), and big data and artificial intelligence platforms, but is not limited thereto.

[0027] It should be noted that the method in the embodiments of the present application can be executed independently by the processing device 100, a computer terminal, or the server 200, or can be jointly executed by any two or more of the server 200, the computer terminal, and the processing device 100. When jointly executed by the server 200 and the processing device 100, the server 200 can train the initial prediction model and then deploy the trained PVD processing prediction model to the processing device 100, and the processing device 100 can implement the monitoring process of PVD processing. Alternatively, part of the model training or application process can be implemented by the processing device 100, and part can be implemented by the server 200, and the two cooperate to implement the model training or application process. Specific configurations can be made according to the situation in actual applications, and no specific limitations are provided here.

[0028] In the embodiments of the present application, the processing device 100 and the server 200 can be directly or indirectly communicatively connected through one or more networks. The network can be a wired network or a wireless network. For example, the wireless network can be a mobile cellular network or a Wireless-Fidelity (WIFI) network. Of course, it can also be other possible networks, and the embodiments of the present application do not limit this. It should be noted that Figure 1 The above is only an example. In fact, the number of processing devices and servers is not limited, and no specific limitations are provided in the embodiments of the present application.

[0029] Next, in combination with the above-described application scenarios, the PVD processing monitoring method and the PVD processing prediction model training method provided by the embodiments of the present application will be described with reference to the accompanying drawings. It should be noted that the above application scenarios are only shown for the convenience of understanding the spirit and principle of the embodiments of the present application, and the embodiments of the present application are not limited in this regard.

[0030] As Figure 2As shown in the figure, it is a flowchart of a monitoring method for PVD processing provided by an embodiment of the present application. The monitoring method for PVD processing can be applied to an electronic device. For example, Figure 1 the processing device 100 shown in the figure. According to different requirements, the order of steps in this flowchart can be changed, and some steps can be omitted.

[0031] S201, obtain PVD processing data of the processing device when processing products.

[0032] In at least one embodiment of the present application, the processing device can process multiple products per furnace. For example, the processing device can coat a preset number (such as 1500) of products simultaneously. The coating of products by the processing device includes the following stages: air extraction stage, pressure holding stage, etching stage, coating stage, etc.

[0033] In at least one embodiment of the present application, the electronic device can obtain that the data formed by the processing device when processing each furnace of products can be one or more groups. Each group of PVD processing data includes multiple PVD processing factors. The multiple PVD processing factors can include but are not limited to: air extraction pressure, etching temperature, workpiece voltage in the etching stage, coating temperature, and workpiece voltage in the coating stage, etc. For example, in the air extraction stage, the electronic device can obtain the first time series of the processing device. Each time point in the first time series carries the PVD air extraction pressure. For example, the first time series is: [10:00, 0.01 pa],

[0034] [10:05, 0.009 pa], ……, [10:35, 0.008 pa]. In the etching stage, the electronic device can obtain the second time series and the third time series of the processing device. Each time point in the second time series carries the etching temperature, and each time point in the third time series carries the workpiece voltage in the etching stage. In the coating stage, the electronic device can obtain the fourth time series and the fifth time series of the processing device. Each time point in the fourth time series carries the coating temperature, and each time point in the fifth time series carries the workpiece voltage in the coating stage.

[0035] S202, perform feature processing on the PVD processing data to obtain PVD processing features.

[0036] S203, input multiple PVD feature factors into the PVD processing prediction model to obtain the product film peeling value.

[0037] In at least one embodiment of the present application, the network architecture of the PVD processing prediction model can be the XGBoost architecture, and the network architecture of the PVD processing prediction model can also be other deep neural networks. The training method of the PVD processing prediction model can refer to the following text Figure 7Detailed description. The PVD processing prediction model can be used to predict the product demolding value when the processing equipment processes the product. The product demolding value can be the demolding area of the product at any demolding point, or the total demolding area of the product at all demolding points, or the average demolding area of the product at all demolding points. For example, the PVD processing prediction model predicts the demolding value of the processing equipment at n specified positions (i.e., demolding points) of the product. n is a positive integer greater than 0. The product demolding value can be the demolding area of any one demolding point, or the total area of n demolding points, or the average area of n demolding points.

[0038] In at least one embodiment of the present application, the PVD processing features include multiple PVD feature factors, and the multiple PVD

[0039] feature factors include but are not limited to: etching time feature, average etching temperature, PVD first pumping pressure feature, PVD second pumping pressure feature, PVD first temperature feature, PVD second temperature feature, highest workpiece voltage, lowest workpiece voltage, average coating temperature, etc. The etching time feature can represent the time from the start of etching to the highest etching temperature. The average etching temperature can represent the average temperature of the processing equipment during the etching stage. The PVD first pumping pressure feature and the PVD second pumping pressure feature can respectively represent different pumping pressures of the processing equipment during the pumping stage. The PVD first temperature feature and the PVD second temperature feature can respectively represent different temperatures of the processing equipment during the pumping stage. The highest workpiece voltage and the lowest workpiece voltage can respectively represent the workpiece voltage of the processing equipment during the etching stage. The average coating temperature can represent the average temperature of the processing equipment during the coating stage.

[0040] In at least one embodiment of the present application, the PVD processing data includes a PVD time series, and each time point in the PVD time series carries a PVD temperature, and the PVD temperature includes the etching temperature. The obtaining method of the etching time feature is as follows:

[0041] S2021a, based on the etching temperature carried by each time point in the PVD time series, identify the highest etching temperature and the corresponding time point.

[0042] The user can set a preset temperature according to actual needs. The preset temperature can be used to indicate the temperature at the start of etching and can be adjusted according to the actual needs of the user.

[0043] S2022a, based on the etching temperature and the preset temperature carried by each time point in the PVD time series, determine the start temperature of etching and the corresponding time point.

[0044] S2023a, determine the etching time feature based on the time point corresponding to the highest etching temperature and the time point corresponding to the etching start temperature.

[0045] The etching time feature can be the difference between the time point corresponding to the highest etching temperature and the time point corresponding to the etching start temperature. For example Figure 3 as shown Figure 3 is a schematic diagram of the etching temperature curve provided by the embodiment of the present application. For example, if the time point corresponding to the etching start temperature is 0 minutes and the time point corresponding to the highest etching temperature is 200 minutes, the determined etching time feature is 200 minutes. In this embodiment, through the preset temperature, the etching start temperature and the corresponding time point can be quickly determined. Through the time point corresponding to the highest etching temperature and the time point corresponding to the etching start temperature, the etching time feature can be accurately determined.

[0046] Specifically, the electronic device determines the etching start temperature and the corresponding time point based on the etching temperature and the preset temperature carried by each time point in the PVD time series, including: if the preset temperature is included in the PVD time series, the electronic device determines the preset temperature as the etching start temperature and determines the time point corresponding to the preset temperature in the PVD time series as the time point corresponding to the etching start temperature. If the preset temperature is not included in the PVD time series, determine the temperature difference between each etching temperature and the preset temperature in the PVD time series, and determine the etching temperature with the smallest absolute value of the temperature difference as the etching start temperature, and determine the time point corresponding to the etching start temperature in the PVD time series as the time point corresponding to the etching start temperature. In this embodiment, when the preset temperature is included in the PVD time series, the etching start temperature and the corresponding time point can be directly determined through the preset temperature. When the preset temperature is not included in the PVD time series, by determining the temperature difference between each etching temperature and the preset temperature in the PVD time series, the etching start temperature and the corresponding time point can be identified from the PVD time series, improving the identification accuracy.

[0047] In at least one embodiment of the present application, the electronic device calculates the average etching temperature based on the etching temperature carried by each time point in the PVD time series.

[0048] In at least one embodiment of the present application, the PVD processing data includes another PVD time series, and each time point in the PVD time series carries an etching current. The electronic device calculates the average current amplitude based on the etching current carried by each time point in the PVD time series. For example Figure 4 as shown Figure 4 is a schematic diagram of the etching current curve provided by the embodiment of the present application. Figure 4 The average current amplitude in

[0049] In at least one embodiment of the present application, the PVD processing data includes another PVD time series, and each time point in the PVD time series carries the PVD pumping pressure. The user can set a first preset pumping pressure and a second preset pumping pressure according to actual needs. The first preset pumping pressure is different from the second preset pumping pressure, and the first preset pumping pressure and the second preset pumping pressure can be adjusted according to the actual needs of the user. The obtaining methods of the PVD first pumping pressure feature and the PVD second pumping pressure feature are as follows:

[0050] S2021b, based on the PVD pumping pressure carried by each time point in the PVD time series, extract the PVD pumping pressures and the corresponding time points that are the same as the first preset pumping pressure and the second preset pumping pressure respectively.

[0051] S2022b, determine the PVD pumping pressures that are the same as the first preset pumping pressure and the second preset pumping pressure as the PVD first pumping pressure feature and the PVD second pumping pressure feature respectively.

[0052] For example, the electronic device determines the PVD pumping pressure that is the same as the first preset pumping pressure as the PVD first pumping pressure feature, and determines the PVD pumping pressure that is the same as the second preset pumping pressure as the PVD second pumping pressure feature.

[0053] Furthermore, the obtaining methods of the PVD first temperature feature and the PVD second temperature feature are as follows:

[0054] S2021c, based on the time points corresponding to the PVD first pumping pressure feature and the time points corresponding to the PVD second pumping pressure feature, extract the PVD temperatures corresponding to the two time points respectively from the PVD temperatures carried by each time point in the PVD time series.

[0055] S2022c, determine the PVD temperatures corresponding to the two time points as the PVD first temperature feature and the PVD second temperature feature respectively.

[0056] Specifically, if the electronic device defines the time point corresponding to the PVD first pumping pressure feature as the first time point, the PVD temperature corresponding to the first time point in the PVD time series can be used as the PVD first temperature feature at this time. For example, the PVD first pumping pressure feature is 0.010 Pa, and the corresponding time point is the 5th second. If the pumping pressure is 0.010 Pa at the 5th second, extract the temperature at the 5th second as the PVD first temperature feature. The electronic device defines the time point corresponding to the PVD second pumping pressure feature as the second time point, and uses the PVD temperature corresponding to the second time point in the PVD time series as the PVD second temperature feature.

[0057] In at least one embodiment of the present application, the PVD processing data includes another PVD time series, and each time point in the PVD time series carries a workpiece voltage. The electronic device determines the highest workpiece voltage and the lowest workpiece voltage based on the workpiece voltage carried by each time point in the PVD time series.

[0058] In at least one embodiment of the present application, the PVD processing data includes another PVD time series, and each time point in the PVD time series carries a PVD temperature, and the PVD temperature further includes a coating temperature. The electronic device calculates the average coating temperature based on the coating temperature carried by each time point in the PVD time series.

[0059] In at least one embodiment of the present application, the monitoring method of PVD processing further includes:

[0060] S204, dividing a plurality of PVD characteristic factors into a first set and a second set, where the number of types of PVD characteristic factors in the first set is less than the number of types of PVD characteristic factors in the second set.

[0061] S205, inputting the PVD characteristic factors in the first set into the PVD processing prediction model to obtain a first film removal value.

[0062] The first film removal value is the product film removal value predicted by the PVD processing prediction model through the PVD characteristic factors in the first set.

[0063] S206, inputting the PVD characteristic factors in the second set into the PVD processing prediction model to obtain a second film removal value.

[0064] The second film removal value is the product film removal value predicted by the PVD processing prediction model through the PVD characteristic factors in the second set. In this embodiment, through the PVD processing prediction model, the first film removal value corresponding to the PVD characteristic factors in the first set can be quickly obtained, and the second film removal value corresponding to the PVD characteristic factors in the second set can be quickly obtained.

[0065] S207, calculating the contribution degree corresponding to each PVD characteristic factor based on the number of types of PVD characteristic factors in the first set, the number of types of a plurality of PVD characteristic factors, the first film removal value, and the second film removal value.

[0066] The calculation formula for the contribution degree corresponding to each PVD characteristic factor can be expressed as:

[0067] where φ i can represent the contribution degree corresponding to the i-th type of PVD characteristic factor, S can represent the number of types of PVD characteristic factors in the first set, M can represent the number of types of a plurality of PVD characteristic factors, f x(S) may represent the first demolding value corresponding to the first set, f x (S ∪ {i}) may represent the second demolding value corresponding to the second set, where the second set includes the PVD characteristic factors in the first set and the i-th PVD characteristic factor.

[0068] For example, if multiple PVD characteristic factors include X1, X2, and X3, and the contribution degree of the PVD characteristic factor X1 is calculated, then when the first set is an empty set, the second set includes the PVD characteristic factor X1; when the first set includes the PVD characteristic factor X2, the second set includes the PVD characteristic factor X2 and the PVD characteristic factor X1; when the first set includes the PVD characteristic factor X3, the second set includes the PVD characteristic factor X3 and the PVD characteristic factor X1; when the first set includes the PVD characteristic factors X2 and X3, the second set includes the PVD characteristic factors X2, X3, and the PVD characteristic factor X1. The contribution degree of the PVD characteristic factor X1 can be:

[0069]

[0070] Among them, f(X1) represents the second demolding value corresponding to the second set including the PVD characteristic factor X1, f(x1,x2) represents the second demolding value corresponding to the second set including the PVD characteristic factors X1 - X2, f(x2) represents the first demolding value corresponding to the first set including the PVD characteristic factor X2, f(x1,x3) represents the second demolding value corresponding to the second set including the PVD characteristic factors X1, X3, f(x3) represents the first demolding value corresponding to the first set including the PVD characteristic factor X3, f(X1,X2,x3) represents the second demolding value corresponding to the second set including the PVD characteristic factors X1 - X3, and F(X2,X3) represents the first demolding value corresponding to the first set including the PVD characteristic factors X2 - X3.

[0071] In this embodiment, through the number of types of PVD characteristic factors in the first set, the number of types of multiple PVD characteristic factors, the first demolding value, and the second demolding value, the contribution degree of the i-th PVD characteristic factor in the second set can be quantitatively obtained in sequence.

[0072] In multiple embodiments of the present application, by performing feature processing on PVD processing data, PVD processing features can be obtained, which can assist the PVD processing prediction model in understanding multiple PVD processing factors. Through the analysis of multiple PVD characteristic factors by the PVD processing prediction model, an accurate product demolding value can be obtained, realizing effective monitoring of the processing equipment.

[0073] In the above step S204, dividing multiple PVD characteristic factors into a first set and a second set, the method specifically includes:

[0074] S2041. Randomly select N PVD characteristic factors from a variety of PVD characteristic factors, where N is a positive integer greater than or equal to 1.

[0075] S2042. Based on the N PVD characteristic factors, construct a first set, where the first set includes the N PVD characteristic factors.

[0076] S2043. Based on the remaining PVD characteristic factors after randomly selecting N PVD characteristic factors from the variety of PVD characteristic factors, further extract Y PVD characteristic factors.

[0077] N > Y, and Y is a positive integer greater than or equal to 1. The electronic device constructs a second set based on the N PVD characteristic factors and the Y PVD characteristic factors, where the second set includes the N PVD characteristic factors and the Y PVD characteristic factors.

[0078] In this embodiment, by randomly selecting N PVD characteristic factors from a variety of PVD characteristic factors, the randomness of constructing the first set can be improved. By further extracting Y PVD characteristic factors from the remaining PVD characteristic factors after randomly selecting N PVD characteristic factors from the variety of PVD characteristic factors, it is possible to avoid having the same PVD characteristic factors between the Y PVD characteristic factors and the N PVD characteristic factors, thereby ensuring that the number of types of PVD characteristic factors in the first set is less than the number of types of PVD characteristic factors in the second set.

[0079] In one embodiment, the electronic device can construct an initial set based on a variety of PVD characteristic factors. The electronic device randomly selects N PVD characteristic factors from the initial set and constructs a first set based on the N PVD characteristic factors. The electronic device randomly selects Y PVD characteristic factors other than the N PVD characteristic factors from the initial set, and the electronic device constructs a second set based on the N PVD characteristic factors and the Y PVD characteristic factors. Usually, Y can be 1.

[0080] In another embodiment, the electronic device can construct an initial set based on a variety of PVD characteristic factors. The electronic device randomly selects N PVD characteristic factors from the initial set and constructs a first set based on the N PVD characteristic factors. The electronic device constructs an intermediate set based on the PVD characteristic factors other than the N PVD characteristic factors in the initial set. The electronic device randomly selects Y PVD characteristic factors from the intermediate set, and the electronic device constructs a second set based on the N PVD characteristic factors and the Y PVD characteristic factors. Usually, Y can be 1.

[0081] As Figure 5 shown, it is a flowchart of a monitoring method for PVD processing provided by another embodiment of the present application, including the following processes:

[0082] S501. Obtain multiple contribution degrees of each PVD characteristic factor within a preset time period.

[0083] In at least one embodiment of the present application, the preset time period can be set and adjusted according to actual needs. For example, the preset time period can be set to one week, one month, etc. The electronic device can obtain multiple groups of PVD processing data when the processing device processes products within the preset time period. Each group of PVD processing data includes multiple PVD processing factors. The multiple PVD processing factors can include, but are not limited to, pumping pressure, etching temperature, workpiece voltage in the etching stage, coating temperature, and workpiece voltage in the coating stage, etc.

[0084] In at least one embodiment of the present application, the electronic device performs feature processing on each group of PVD processing data to obtain PVD processing features. Each group of PVD processing features includes multiple PVD characteristic factors. The multiple PVD characteristic factors include, but are not limited to, any one or more of etching time feature, average etching temperature, PVD first pumping pressure feature, PVD second pumping pressure feature, PVD first temperature feature, PVD second temperature feature, highest workpiece voltage, lowest workpiece voltage, and average coating temperature.

[0085] In at least one embodiment of the present application, the preset time period includes multiple groups of PVD processing features. The electronic device can determine the contribution degree of each PVD characteristic factor according to the multiple PVD characteristic factors in each group of PVD processing features. The determination method of the contribution degree of each PVD characteristic factor can refer to the detailed description of step S203 above, and will not be repeated here. The electronic device can obtain multiple contribution degrees of each PVD characteristic factor within the preset time period. Figure 2 The detailed description of step S203 above is referred to here and will not be repeated. The electronic device can obtain multiple contribution degrees of each PVD characteristic factor within the preset time period.

[0086] S502. Calculate the average contribution degree of each PVD characteristic factor based on the multiple contribution degrees of each PVD characteristic factor.

[0087] In at least one embodiment of the present application, for each PVD characteristic factor, the electronic device can calculate the average value of the corresponding multiple contribution degrees to obtain the average contribution degree of each PVD characteristic factor.

[0088] In another embodiment, for each PVD characteristic factor, the electronic device can calculate the weighted sum of the corresponding multiple contribution degrees according to the average value of the multiple contribution degrees corresponding to each PVD characteristic factor to obtain the average contribution degree of each PVD characteristic factor. Among them, the weight corresponding to each contribution degree can be determined according to the contribution degree deviation between each contribution degree and the average value of the corresponding multiple contribution degrees. The weight corresponding to each contribution degree can be inversely proportional to the contribution degree deviation. For example, the greater the contribution degree deviation, the smaller the corresponding weight.

[0089] S503. Sort all types of PVD feature factors based on the average contribution degree of each PVD feature factor to obtain a queue.

[0090] In at least one embodiment of the present application, the electronic device can sort all types of PVD feature factors in descending order according to the average contribution degree of each PVD feature factor to obtain a queue.

[0091] In another embodiment, the electronic device can sort all types of PVD feature factors in ascending order according to the average contribution degree of each PVD feature factor to obtain a queue.

[0092] S504. Select M PVD feature factors from the queue as key PVD feature factors based on a preset rule.

[0093] In at least one embodiment of the present application, the electronic device selects the M PVD feature factors with the largest average contribution degree from the queue as key PVD feature factors. M is a positive integer greater than or equal to 1, and can be 3, 5, 10, etc., which can be set according to actual needs.

[0094] In one embodiment, if the queue is sorted in descending order according to the average contribution degree, the preset rule can be set to select the first M PVD feature factors in the queue. The electronic device selects the first M PVD feature factors in the queue as key PVD feature factors.

[0095] In another embodiment, if the queue is sorted in ascending order according to the average contribution degree, the preset rule can be set to select the last M PVD feature factors in the queue. The electronic device selects the last M PVD feature factors in the queue as key PVD feature factors.

[0096] In at least one embodiment of the present application, the electronic device displays the key PVD feature factors according to the average contribution degree of the key PVD feature factors. Each key PVD feature factor corresponds to a contribution degree threshold, and the contribution degree threshold can be set and adjusted according to requirements. The electronic device determines the key PVD feature factors whose average contribution degree is greater than or equal to the corresponding contribution degree threshold as the key PVD feature factors to be identified. Send the identified key PVD feature factors to the originating terminal to remind the user to pay attention to the PVD processing factors corresponding to the key PVD feature factors, so as to make corresponding PVD processing data adjustments to the processing equipment according to the PVD processing factors corresponding to the key PVD feature factors, or equipment maintenance, etc. It can also dynamically adjust the PVD processing prediction model and monitor the key PVD feature factors and the corresponding PVD processing factors in real time.

[0097] In at least one embodiment of the present application, the electronic device uses a preset identifier to identify the key PVD characteristic factors to be identified. Among them, the preset identifier can be a text identifier, and the preset identifier can also be a color identifier. The embodiments of the present application do not make specific limitations on the preset identifier. Refer to Figure 6 as shown in Figure 6 FIG. Figure 6 shown is a schematic diagram showing the key PVD characteristic factors provided by the embodiments of the present application. As

[0098] In at least one embodiment of the present application, the monitoring method for PVD processing further includes:

[0099] S505, determining the components to be maintained in the processing equipment corresponding to the key PVD characteristic factors.

[0100] For example, if the key PVD characteristic factor is the PVD first pumping pressure characteristic or the PVD second pumping pressure characteristic, the components to be maintained can be components such as a vacuum pump and a vacuum gauge. The electronic device obtains the maintenance strategy corresponding to the component to be maintained to recommend the maintenance strategy to the user. Among them, the maintenance strategy can be a strategy corresponding to the component to be maintained. For example, if the component to be maintained is a vacuum pump, the corresponding maintenance strategy can be a strategy related to the sealing performance of the vacuum pump. Through the key PVD characteristic factors in this embodiment, the corresponding components to be maintained can be determined, and the corresponding maintenance strategy can be obtained according to the components to be maintained, which can avoid maintaining the entire processing equipment, not only improving the maintenance effectiveness of the processing equipment, but also avoiding the waste of operation and maintenance resources.

[0101] In multiple embodiments of the present application, through the multiple contribution degrees of each PVD characteristic factor within a preset time period, the average contribution degree of each PVD characteristic factor can be accurately determined, so that the key PVD characteristic factors affecting the product yield can be accurately determined, which is beneficial to the targeted monitoring and maintenance of the processing equipment, not only improving the maintenance effectiveness of the processing equipment, but also avoiding the waste of operation and maintenance resources.

[0102] As Figure 7 shown is a flowchart of a method for training a PVD processing prediction model provided by an embodiment of the present application. The method for training a PVD processing prediction model can be applied to an electronic device, such as a server. According to different requirements, the order of the steps in this flowchart can be changed, and some steps can be omitted.

[0103] S701, obtaining multiple groups of historical PVD processing data and the historical product demolding values of each corresponding product.

[0104] In at least one embodiment of the present application, multiple sets of historical PVD processing data can be obtained from a single processing device processing a furnace of products, or from multiple processing devices processing only a furnace of products, or from multiple processing devices processing multiple furnaces. Each furnace of the processing device can process multiple products. For example, the processing device can simultaneously coat a preset number (such as 1500) of products. The electronic device can obtain one or more sets of historical PVD processing data formed by the processing device when processing each furnace of products, and each set of historical PVD processing data includes multiple historical PVD processing factors. The multiple historical PVD processing factors can include, but are not limited to, pumping pressure, etching temperature, workpiece voltage during the etching stage, coating temperature, and workpiece voltage during the coating stage, etc.

[0105] In at least one embodiment of the present application, the historical PVD processing data includes a historical PVD time series, and each time point in the historical PVD time series carries a historical PVD temperature, and the historical PVD temperature includes a historical etching temperature and a historical coating temperature. Each time point in the historical PVD time series also carries a historical PVD pumping pressure. Each time point in the historical PVD time series also carries a historical workpiece voltage.

[0106] In at least one embodiment of the present application, a set of historical PVD processing data formed by the processing device processing a furnace of products corresponds to the historical product demolding values of at least one product. After the processing device processes multiple products, any one or more products are randomly selected from the multiple products in each furnace for destructive testing. For any product subjected to destructive testing, the product has multiple demolding points, and then the demolding values of the demolding points are detected optically to obtain the historical product demolding values. For example Figure 8 As shown, the product has eight designated positions, that is, demolding points, after destructive testing. The electronic device determines the area of each demolding point in the products subjected to destructive testing to obtain the historical product demolding values of each product. For example Figure 8 As shown, the processing device punches holes at any eight positions of the product to obtain eight historical product demolding values of the product. The historical product demolding value can be the area corresponding to each demolding point. Of course, in other embodiments, destructive testing may not be performed, and the demolding area of each of the eight designated positions (i.e., demolding points) after processing by the processing device is directly detected visually to obtain the historical product demolding values of each product. It is also possible to obtain the demolding area of each of the eight designated positions through manual analysis, and then obtain the historical product demolding values of each product.

[0107] S702. Perform feature processing on multiple sets of historical PVD processing data to obtain historical PVD processing features.

[0108] S703. Perform feature processing on the historical product demolding values to obtain historical product demolding feature values.

[0109] In at least one embodiment of the present application, the historical product demolding value of each product includes the historical product demolding values of multiple demolding points. The historical product demolding characteristic value may be the average value of the historical product demolding values of multiple demolding points in each product. The historical product demolding characteristic value may also be the weighted sum of the historical product demolding values of multiple demolding points in each product.

[0110] In at least one embodiment of the present application, the electronic device calculates the average value of the historical product demolding values of each product based on the historical product demolding values of multiple demolding points of each product. The electronic device determines the average value of the historical product demolding values of each product as the historical product demolding characteristic value. For example, the historical product demolding values of eight demolding points are 500 square microns, 900 square microns, 24000 square microns, 8000 square microns, 3000 square microns, 2000 square microns, 5000 square microns, and 20 square microns respectively. The historical product demolding characteristic value may be the average value of the historical product demolding values of multiple demolding points, for example: 5427.5 square microns.

[0111] In another embodiment, the electronic device calculates the weighted value of the historical product demolding values of each product based on the historical product demolding values of multiple demolding points of each product. The weight corresponding to each demolding point may be determined according to the historical product demolding value corresponding to each demolding point. For example, the weight corresponding to each demolding point is inversely proportional to the historical product demolding value corresponding to the demolding point. The electronic device determines the weighted value of the historical product demolding values of each product as the historical product demolding characteristic value.

[0112] S704, based on a variety of historical PVD characteristic factors, historical product demolding values, and historical product demolding characteristic values, train the initial prediction model to obtain a PVD processing prediction model.

[0113] In at least one embodiment of the present application, the network architecture of the initial prediction model may be an XGBoost architecture, and the network architecture of the initial prediction model may also be other deep neural networks.

[0114] In at least one embodiment of the present application, the historical PVD processing characteristics include a variety of historical PVD characteristic factors, and the variety of historical PVD characteristic factors include but are not limited to: historical etching time characteristics, PVD first historical pumping pressure characteristics, PVD second historical pumping pressure characteristics, PVD first historical temperature characteristics, PVD second historical temperature characteristics, historical average etching temperature, historical maximum workpiece voltage, historical minimum workpiece voltage, historical average coating temperature, etc.

[0115] In at least one embodiment of the present application, the historical PVD processing data includes a historical PVD time series, and each time point in the historical PVD time series carries a historical PVD temperature, and the historical PVD temperature includes a historical etching temperature. The obtaining method of the historical etching time feature is as follows:

[0116] S7021a, based on the historical etching temperature carried by each time point in the historical PVD time series, identify the highest historical etching temperature and the corresponding time point.

[0117] The user can set a historical preset temperature according to actual needs. The historical preset temperature can be used to indicate the temperature at the start of etching, and the historical preset temperature can be adjusted according to the actual needs of the user.

[0118] S7022a, based on the historical etching temperature and the historical preset temperature carried by each time point in the historical PVD time series, determine the historical etching start temperature and the corresponding time point.

[0119] S7023a, based on the time point corresponding to the highest historical etching temperature and the time point corresponding to the historical etching start temperature, determine the historical etching time feature.

[0120] The historical etching time feature can be the difference between the time point corresponding to the highest historical etching temperature and the time point corresponding to the historical etching start temperature. In this embodiment, through the historical preset temperature, the historical etching start temperature and the corresponding time point can be quickly determined. Through the time point corresponding to the highest historical etching temperature and the time point corresponding to the historical etching start temperature, the historical etching time feature can be accurately determined.

[0121] Specifically, if the historical preset temperature is included in the historical PVD time series, the electronic device determines the historical preset temperature as the historical etching start temperature, and determines the time point corresponding to the historical preset temperature in the historical PVD time series as the time point corresponding to the historical etching start temperature. If the historical preset temperature is not included in the historical PVD time series, determine the historical temperature difference between each historical etching temperature in the historical PVD time series and the historical preset temperature, and determine the historical etching temperature with the smallest absolute value of the historical temperature difference as the historical etching start temperature, and determine the time point corresponding to the historical etching start temperature in the historical PVD time series as the time point corresponding to the historical etching start temperature. In this embodiment, when the historical preset temperature is included in the historical PVD time series, the historical etching start temperature and the corresponding time point can be directly determined through the historical preset temperature. When the historical preset temperature is not included in the historical PVD time series, by determining the temperature difference between each historical etching temperature in the historical PVD time series and the historical preset temperature, the historical etching start temperature and the corresponding time point can be identified from the historical PVD time series, improving the identification accuracy.

[0122] In at least one embodiment of the present application, the electronic device calculates the historical average etching temperature based on the historical etching temperature carried by each time point in the historical PVD time series.

[0123] In at least one embodiment of the present application, the historical PVD processing data includes another historical PVD time series, and each time point in the historical PVD time series carries the historical etching current. The electronic device calculates the historical average current amplitude based on the historical etching current carried by each time point in the historical PVD time series.

[0124] In at least one embodiment of the present application, the historical PVD processing data includes another historical PVD time series, and each time point in the historical PVD time series also carries the historical PVD pumping pressure. The user can set the first preset historical pumping pressure and the second preset historical pumping pressure according to actual needs. The first preset historical pumping pressure is different from the second preset historical pumping pressure, and the first preset historical pumping pressure and the second preset historical pumping pressure can be adjusted according to the actual needs of the user. The obtaining methods of the PVD first historical pumping pressure feature and the PVD second historical pumping pressure feature are as follows:

[0125] S7021b, based on the historical PVD pumping pressure carried by each time point in the historical PVD time series, respectively extract the historical PVD pumping pressure and the corresponding time points that are the same as the first preset historical pumping pressure and the second preset historical pumping pressure.

[0126] S7022b, determine the historical PVD pumping pressure that is the same as the first preset historical pumping pressure and the second preset historical pumping pressure as the PVD first historical pumping pressure feature and the PVD second historical pumping pressure feature, respectively.

[0127] Further, the obtaining methods of the PVD first historical temperature feature and the PVD second historical temperature feature are as follows:

[0128] S7021c, based on the time points corresponding to the PVD first historical pumping pressure feature and the time points corresponding to the PVD second historical pumping pressure feature, respectively extract the corresponding historical PVD temperature from the historical PVD temperature carried by each time point in the historical PVD time series.

[0129] S7022c, determine the extracted historical PVD temperature as the PVD first historical temperature feature and the PVD second historical temperature feature, respectively.

[0130] Specifically, if the electronic device defines the time point corresponding to the first historical pumping pressure feature of PVD as the first historical time point, the historical PVD temperature corresponding to the first historical time point in the historical PVD time series can be used as the first historical temperature feature of PVD. For example, if the first historical pumping pressure feature of PVD is 0.010 Pa and the corresponding time point is the 5th second, when the pumping pressure is 0.010 Pa at the 5th second, the temperature at the 5th second is extracted as the first temperature feature of PVD. The electronic device defines the time point corresponding to the second historical pumping pressure feature of PVD as the second historical time point, and uses the PVD historical temperature corresponding to the second historical time point in the historical PVD time series as the second historical temperature feature of PVD.

[0131] In at least one embodiment of the present application, the historical PVD processing data includes another historical PVD time series, and each time point in the historical PVD time series also carries the historical workpiece voltage. The electronic device determines the historical maximum workpiece voltage and the historical minimum workpiece voltage based on the historical workpiece voltage carried by each time point in the historical PVD time series.

[0132] In at least one embodiment of the present application, the historical PVD processing data includes another historical PVD time series, and each time point in the historical PVD time series also carries the historical coating temperature. The electronic device calculates the historical average coating temperature based on the historical coating temperature carried by each time point in the historical PVD time series.

[0133] In at least one embodiment of the present application, the electronic device uses multiple sets of historical PVD processing data and the historical product demolding values of each corresponding product as a data set. Based on a preset ratio, the data set is randomly divided into a training set and a test set. The preset ratio can be set according to actual needs. Usually, the data ratio of the training set is greater than that of the test set. For example, the preset ratio can be set as training set: test set = 8:2. The electronic device trains an initial prediction model using various historical PVD feature factors in the training set. The specific process of training the initial prediction model can refer to Figure 9 the shown process, which will be described below in conjunction with Figure 9 This is explained. The PVD processing prediction model training method further includes:

[0134] S7041, input various historical PVD feature factors into the initial prediction model to obtain the first predicted demolding value and the contribution degree of each historical PVD feature factor.

[0135] In at least one embodiment of the present application, the initial prediction model predicts the first predicted demolding value of the processing equipment when processing a product through a variety of historical PVD characteristic factors. The first predicted demolding value can be the demolding area of the product at any demolding point, or the total demolding area of the product at all demolding points, or the average demolding area of the product at multiple demolding points.

[0136] S7042. Determine the accuracy of the initial prediction model based on the first predicted demolding value, the historical product demolding value, and the historical product demolding characteristic value.

[0137] S7043. Detect whether the accuracy is within a preset interval.

[0138] In at least one embodiment of the present application, the preset interval can be set according to actual needs. For example, the preset interval can be set to [0.8, 1]. If the accuracy is not within the preset interval, execute step S7044. If the accuracy is within the preset interval, execute step S7045.

[0139] S7044. Obtain new historical PVD processing data to train the initial prediction model.

[0140] In at least one embodiment of the present application, the manner in which the electronic device uses the new historical PVD processing data to train the initial prediction model can refer to Figure 7 the process shown, and the present application will not repeat the description here.

[0141] S7045. According to the contribution degree of each historical PVD characteristic factor, screen out the key historical PVD characteristic factors from a variety of historical PVD characteristic factors according to a preset screening rule.

[0142] In at least one embodiment of the present application, the electronic device sorts all types of historical PVD characteristic factors according to the contribution degree of each historical PVD characteristic factor to obtain a list. The electronic device selects K historical PVD characteristic factors from the list as the key historical PVD characteristic factors based on a preset rule, where K is a positive integer greater than 1.

[0143] In at least one embodiment of the present application, the electronic device can sort all types of historical PVD characteristic factors in descending order according to the contribution degree of each historical PVD characteristic factor to obtain a list.

[0144] In another embodiment, the electronic device can sort all types of historical PVD characteristic factors in ascending order according to the contribution degree of each historical PVD characteristic factor to obtain a list.

[0145] In at least one embodiment of the present application, the electronic device selects the K historical PVD feature factors with the greatest contribution from the list as the key historical PVD feature factors.

[0146] S7046, input the key historical PVD feature factors into the initial prediction model to obtain a second predicted demolding value.

[0147] In at least one embodiment of the present application, the initial prediction model predicts the corresponding second predicted demolding value of the processing device when processing a product through the key historical PVD feature factors. The second predicted demolding value may be the demolding area of the product at any demolding point, the second predicted demolding value may also be the total demolding area of the product at all demolding points, and the second predicted demolding value may further be the average demolding area of the product at multiple demolding points.

[0148] S7047, determine the first loss value of the initial prediction model based on the first predicted demolding value and the historical product demolding feature value.

[0149] In at least one embodiment of the present application, the calculation formula of the first loss value can be expressed as:

[0150] where RMSE can represent the first loss value, can represent the historical product demolding feature value of the i-th product, can represent the first predicted demolding value of the i-th product, and N can represent the total number of products.

[0151] S7048, determine the second loss value of the initial prediction model according to the second predicted demolding value and the historical product demolding feature value.

[0152] In at least one embodiment of the present application, the calculation method of the second loss value is similar to that of the first loss value, and the present application will not repeat the description here.

[0153] S7049, based on the first loss value, the second loss value, and the number of types of key historical PVD feature factors, determine whether to use the trained initial prediction model as the PVD processing prediction model.

[0154] In multiple embodiments of the present application, when the accuracy is within a preset range, by the contribution degree of each historical PVD characteristic factor, the key historical PVD characteristic factors are screened and input into the initial prediction model for prediction, and a second predicted demolding value can be obtained. By the first predicted demolding value and the historical product demolding characteristic values, the first loss value of the initial prediction model is quantified, and by the second predicted demolding value and the historical product demolding characteristic values, the second loss value of the initial prediction model is quantified. Furthermore, in combination with the first loss value, the second loss value, and the number of types of key historical PVD characteristic factors, a PVD processing prediction model is determined, which can ensure the accuracy of the PVD processing prediction model while also guaranteeing the prediction efficiency of the PVD processing prediction model.

[0155] In one embodiment, step S7049 includes:

[0156] S70491, determining the absolute value of the difference between the first loss value and the second loss value.

[0157] The first preset value can be set and adjusted according to user requirements.

[0158] S70492, if the absolute value of the difference is less than the first preset value, based on the key historical PVD characteristic factors and their corresponding historical product demolding values and historical product demolding characteristic values, step S704 is cyclically executed until the number of types of key historical PVD characteristic factors is less than or equal to the second preset value, so as to execute the step of "using the trained initial prediction model as the PVD processing prediction model".

[0159] Among them, the second preset value can be set and adjusted according to user requirements. In this embodiment, when the absolute value of the difference is less than the first preset value, the initial prediction model is trained by the key historical PVD characteristic factors, which can realize the reduction of the key historical PVD characteristic factors, thereby improving the prediction efficiency of the PVD processing prediction model and ensuring the accuracy of the PVD processing prediction model.

[0160] In another embodiment, step S7049 further includes:

[0161] S70493, if the absolute value of the difference between the first loss value and the second loss value is greater than or equal to the first preset value, based on the contribution degree of each historical PVD processing factor, key historical PVD characteristic factors are re-screened from various historical PVD processing factors according to a preset screening rule;

[0162] S70494, based on the re-screened key historical PVD characteristic factors, step S7046 is executed.

[0163] When the absolute value of the difference between the first loss value and the second loss value is greater than or equal to the first preset value, this embodiment can ensure the accuracy of the PVD processing prediction model by re-screening key historical PVD characteristic factors.

[0164] In at least one embodiment of the present application, the electronic device uses an optimization function to optimize the initial prediction model. The optimization function can be expressed as: where l() represents the loss value of the initial prediction model, Tree t () represents the output of the initial prediction model in the t-th iteration, Ω() represents the complexity of the initial prediction model, and constant can represent a preset value. The electronic device iteratively optimizes the initial prediction model until the PVD processing prediction model is obtained.

[0165] where h t = Tree t (S). This embodiment optimizes the initial prediction model through the optimization function, improving the optimization rationality of the PVD processing prediction model. In at least one embodiment of the present application, the historical product demolding value of each product includes the historical product demolding values of multiple demolding points. The historical product demolding characteristic value can be the average value or weighted sum of the historical product demolding values of multiple demolding points in each product. For example, in at least one embodiment of the present application, step S7042 determines the accuracy of the initial prediction model based on the first predicted demolding value, the historical product demolding value, and the historical product demolding characteristic value, including:

[0166] S70421, based on the historical product demolding values of multiple demolding points of each product, determine the final standard deviation of the historical product demolding value of each demolding point of each product;

[0167] S70422, according to the historical product demolding characteristic value, the final standard deviation, the first predicted demolding value, and the third preset value of each product, determine the accuracy.

[0168] Specifically, the calculation formula for accuracy can be expressed as: where can represent the accuracy of the initial prediction model, Q can represent the first predicted demolding value, the first predicted demolding value can represent the average demolding area of the product at multiple demolding points, P can represent the historical product demolding characteristic value of each product, can represent the third preset value, and the third preset value can be adjusted according to actual needs, can represent the final standard deviation. The closer the accuracy is to the preset standard value (such as 1), the more accurate it is.

[0169] In some embodiments, the PVD processing prediction model training method further includes:

[0170] S70420a, determine whether the historical product demolding values of multiple demolding points of each product are independent of each other.

[0171] S70420b, if the historical product demolding values of multiple demolding points of each product are independent of each other, determine whether the data distributions of the historical product demolding values of the demolding points of the product are the same.

[0172] Specifically, determine whether the data distributions of the historical product demolding values of the demolding points at the same position of all products are the same.

[0173] S70420c, if the data distributions of the historical product demolding values of the demolding points of the product are the same, execute step S70421, and based on the historical product demolding values of multiple demolding points of each product, determine the final standard deviation of the historical product demolding value of each demolding point of each product.

[0174] Specifically, the calculation formula of the final standard deviation can be expressed as: Where, can represent the final standard deviation corresponding to the situation where the historical product demolding values of multiple demolding points of each product are independent of each other and the data distributions of the historical product demolding values of the demolding points at the same position of each product are the same, A i can represent the historical product demolding value of the i-th demolding point of each product, A can represent the historical product demolding characteristic value (i.e., the average value) of each product, and n can represent the number of demolding points of each product.

[0175] In at least one embodiment of the present application, the PVD processing prediction model training method further includes:

[0176] S70420d, if the data distributions of the historical product demolding values of the demolding points of the product are different, obtain the historical product demolding values of the demolding points of the new product to form reference product demolding values.

[0177] Specifically, when the data distributions of the historical product demolding values of the demolding points at the same position of multiple products are different, obtain the historical product demolding values of multiple demolding points of another batch of multiple products and execute step S70420e.

[0178] S70420e, calculate the intermediate value standard deviation corresponding to the historical product demolding value of each demolding point according to the historical product demolding value and the reference product demolding value.

[0179] Specifically, the calculation method of the intermediate value standard deviation can be expressed by the formula: Where, σ i can represent the intermediate value standard deviation corresponding to the historical product demolding value of the i-th demolding point, A iThe historical product demolding value of the i-th demolding point of each product can be represented as A, the average value of the reference product demolding values of each product can be represented as B, and the number of demolding points of each product can be represented as n.

[0180] S70420f, based on the standard deviation of the intermediate values corresponding to the historical product demolding values of each demolding point, obtains the final standard deviation of the historical product demolding values of each demolding point.

[0181] Specifically, the calculation formula of the final standard deviation can be expressed as: Specifically, σ ′ can represent the final standard deviation corresponding to the historical product demolding values of multiple demolding points of each product being independent of each other and the data distributions of the historical product demolding values of the demolding points at the same position of each product being different, and σ i can represent the standard deviation of the intermediate value corresponding to the historical product demolding value of the i-th demolding point.

[0182] In this embodiment, when the historical product demolding values of multiple demolding points of each product are independent of each other and the data distributions of the historical product demolding values of the demolding points at the same position of each product are different, by combining the reference product demolding values of the demolding points of other products to calculate the final standard deviation, the accuracy of the final standard deviation is improved.

[0183] As Figure 10 shown, it is a flowchart of a method for judging whether the historical product demolding values of multiple products provided by an embodiment of the present application are independent of each other. Step S70420a includes the following processes:

[0184] S1001, obtain the demolding coordinates of multiple demolding points in each product.

[0185] In at least one embodiment of the present application, the electronic device constructs a coordinate system with the center position of the product as the origin and determines the demolding coordinates of multiple demolding points on the coordinate system. In another embodiment, the electronic device can also construct a coordinate system with any corner or other position of the product as the origin, and can also construct a coordinate system with a position outside the product as the origin and determine the demolding coordinates of multiple demolding points on the coordinate system.

[0186] S1002, calculate the demolding coefficient of each demolding point according to the demolding coordinates of the demolding point corresponding to the historical product demolding value of each demolding point.

[0187] In at least one embodiment of the present application, the demoulding coefficient can be any one of the Pearson correlation coefficient, the Spearman rank correlation coefficient, and the Kendall rank correlation coefficient. The electronic device calculates the Pearson correlation coefficient as the demoulding coefficient according to the historical product demoulding values of each demoulding point and the demoulding coordinates of the corresponding demoulding point. The electronic device can also calculate the Spearman rank correlation coefficient as the demoulding coefficient according to the historical product demoulding values of each demoulding point and the demoulding coordinates of the corresponding demoulding point. The electronic device can also calculate the Kendall rank correlation coefficient as the demoulding coefficient according to the historical product demoulding values of each demoulding point and the demoulding coordinates of the corresponding demoulding point.

[0188] S1003, determine whether the absolute value of the demoulding coefficient of each demoulding point is less than a first preset threshold.

[0189] In at least one embodiment of the present application, the first preset threshold can be set and adjusted according to actual needs. If the absolute value of the demoulding coefficient of each demoulding point is less than the first preset threshold, execute step S1004; if the absolute value of the demoulding coefficient of each demoulding point is greater than or equal to the first preset threshold, execute step S1005.

[0190] S1004, determine that the historical product demoulding values of multiple demoulding points in each product are independent of each other, so as to execute step S70420b.

[0191] S1005, determine that the historical product demoulding values of multiple demoulding points in each product are not independent of each other.

[0192] Through the demoulding coordinates corresponding to the historical product demoulding values of each demoulding point in the embodiments of the present application, the demoulding coefficient of each demoulding point can be quantitatively obtained. By comparing the absolute value of the demoulding coefficient of each demoulding point with the first preset threshold, it is possible to determine whether the historical product demoulding values of multiple demoulding points in each product are independent of each other, improving the determination accuracy.

[0193] As Figure 11 shown, it is a flowchart of a method for judging whether the data distributions of multiple historical product demoulding values of each product provided by the embodiments of the present application are the same, including the following processes:

[0194] S1101, respectively form multiple demoulding curves according to the historical product demoulding values of multiple demoulding points.

[0195] In at least one embodiment of the present application, each demoulding curve can be obtained according to the multiple historical product demoulding values of each product. The multiple demoulding curves represent the demoulding curves corresponding to multiple products.

[0196] S1102, determine the curve similarity between each demoulding curve and a preset distribution curve.

[0197] In at least one embodiment of the present application, the preset distribution curve may be a normal distribution curve, and the preset distribution curve may also be other distribution curves. The present application does not limit this. The electronic device may use the dynamic time warping algorithm to determine the curve similarity between each demoulding curve and the preset distribution curve.

[0198] S1103. Determine whether the curve similarity is less than the second preset threshold.

[0199] In at least one embodiment of the present application, the second preset threshold may be set and adjusted according to actual needs. If the curve similarity is less than the second preset threshold, step S1104 is executed; if the curve similarity is greater than or equal to the second preset threshold, step S1105 is executed.

[0200] S1104. Determine that the data distributions of the historical product demoulding values of multiple demoulding points are different, and execute step S70420d.

[0201] S1105. Determine that the data distributions of the historical product demoulding values of multiple demoulding points are the same, and execute step S70421.

[0202] In the embodiment of the present application, the corresponding demoulding curve can be obtained through the multiple historical product demoulding values of each product. Through the curve similarity between the demoulding curve and the preset distribution curve, the data distribution degree of the historical product demoulding values of multiple demoulding points can be quantified, so as to accurately determine whether the data distributions of the historical product demoulding values of multiple demoulding points are the same.

[0203] In one embodiment, S7041. Input multiple historical PVD characteristic factors into the initial prediction model to obtain the first predicted demoulding value and the contribution degree of each historical PVD characteristic factor, including the following steps:

[0204] S70411. Divide multiple historical PVD characteristic factors into a first historical set and a second historical set. The number of historical PVD characteristic factors in the first historical set is less than the number of historical PVD characteristic factors in the second historical set.

[0205] For example, the first historical set includes N historical PVD characteristic factors, where N is a positive integer greater than or equal to 1. The second historical set includes N historical PVD characteristic factors and Y historical PVD characteristic factors, where Y is a positive integer greater than or equal to 1.

[0206] S70412. Input the historical PVD characteristic factors in the first historical set into the initial prediction model to obtain the first historical predicted demoulding value.

[0207] The first historical predicted demoulding value is the demoulding value predicted by the initial prediction model through the historical PVD characteristic factors in the first historical set.

[0208] S70413, input the historical PVD characteristic factors in the second historical set into the initial prediction model to obtain the second historical predicted film removal value.

[0209] The second historical predicted film removal value is the film removal value predicted by the initial prediction model through the historical PVD characteristic factors in the second historical set.

[0210] S70413, calculate the contribution degree corresponding to each historical PVD characteristic factor based on the number of types of historical PVD characteristic factors in the first historical set, the number of types of multiple historical PVD characteristic factors, the first historical predicted film removal value, and the second historical predicted film removal value.

[0211] Specifically, the calculation formula for the contribution degree corresponding to each historical PVD characteristic factor can be expressed as:

[0212] Among them, φ i can represent the contribution degree corresponding to the i-th type of historical PVD characteristic factor, S can represent the number of types of historical PVD characteristic factors in the first historical set, M can represent the number of types of multiple historical PVD characteristic factors, f x (S) can represent the first historical predicted film removal value corresponding to the first historical set, f x (S ∪ {i}) can represent the second historical predicted film removal value corresponding to the second historical set, and the second historical set includes the historical PVD characteristic factors in the first historical set and the i-th type of historical PVD characteristic factor.

[0213] As Figure 12 shown, it is a schematic structural diagram of an electronic device of a preferred embodiment for implementing the monitoring method of PVD processing or the training method of the PVD processing prediction model in the present application. In an embodiment of the present application, the electronic device 10 includes, but is not limited to, a memory 1402, a processor 1401, and a computer program stored in the memory 1402 and executable on the processor 1401, such as a calibration program.

[0214] Those skilled in the art can understand that the schematic diagram is only an example of the electronic device 10, and does not constitute a limitation on the electronic device 10. It may include more or fewer components than shown, or combine certain components, or different components. For example, the electronic device 10 may also include input / output devices, network access devices, buses, etc.

[0215] The processor 1401 can be a Central Processing Unit (CPU), or can also be other general-purpose processors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field-Programmable Gate Arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc. The processor 1401 is the computing core and control center of the electronic device 10, connecting all parts of the entire electronic device 10 through various interfaces and lines, and obtaining the operating system of the electronic device 10 and various installed application programs, program codes, etc.

[0216] The processor 1401 obtains the operating system of the electronic device 10 and various installed application programs. The processor 1401 obtains the application programs to implement the steps in the embodiments of the above-mentioned various monitoring methods for PVD processing, such as Figure 2 、 Figure 5 the steps shown; the processor 1401 obtains the application programs to implement the steps in the embodiments of the above-mentioned various training methods for PVD processing prediction models, such as Figure 7 、 Figures 9 - 11 the steps shown.

[0217] Exemplarily, the computer program can be divided into one or more modules / units, and one or more modules / units are stored in the memory 1402 and obtained by the processor 1401 to complete this application. One or more modules / units can be a series of computer program instruction segments capable of performing specific functions, and these instruction segments are used to describe the obtaining process of the computer program in the electronic device 10.

[0218] The memory 1402 can be used to store computer programs and / or modules. By running or obtaining the computer programs and / or modules stored in the memory 1402, and invoking the data stored in the memory 1402, the processor 1401 realizes various functions of the electronic device 10. The memory 1402 mainly includes a program storage area and a data storage area. Among them, the program storage area can store an operating system, application programs required for at least one function (such as a sound playback function, an image playback function, etc.); the data storage area can store data created according to the use of the electronic device. In addition, the memory 1402 can include non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) card, a Flash Card, at least one magnetic disk storage device, a flash memory device, or other non-volatile solid-state storage devices.

[0219] The memory 1402 can be an external memory and / or an internal memory of the electronic device 10. Further, the memory 1402 can be a memory in a physical form, such as a memory stick, a TF card (Trans-flash Card), etc.

[0220] If the modules / units integrated in the electronic device 10 are implemented in the form of software functional units and sold or used as independent artifacts, they can be stored in a computer-readable storage medium. Based on such an understanding, to implement all or part of the processes in the above-mentioned embodiment methods of the present application, it can also be completed by instructing relevant hardware through computer-readable instructions. The computer-readable instructions can be stored in a computer-readable storage medium. When the computer-readable instructions are executed by a processor, the steps of the above-mentioned various method embodiments can be implemented.

[0221] Among them, the computer-readable instructions include computer-readable instruction codes, and the computer-readable instruction codes can be in the form of source code, object code, executable files, or some intermediate forms, etc. The computer-readable medium can include: any entity or device capable of carrying the computer-readable instruction codes, a recording medium, a USB flash drive, a mobile hard disk, a magnetic disk, an optical disc, a computer memory, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory).

[0222] The memory 1402 can be used to store computer-readable instructions and / or modules. By running or executing the computer-readable instructions and / or modules stored in the memory 1402, and by invoking the data stored in the memory 1402, the processor 1401 realizes various functions of the electronic device 10. The memory 1402 mainly includes a program storage area and a data storage area. Among them, the program storage area can store an operating system, application programs required for at least one function (such as a sound playback function, an image playback function, etc.); the data storage area can store data created according to the use of the electronic device. The memory 1402 can include non-volatile and volatile memories, such as: hard disks, memory, plug-in hard disks, smart media cards (SMC), secure digital (SD) cards, flash cards, at least one magnetic disk storage device, flash memory device, or other storage devices.

[0223] Exemplarily, the computer-readable instructions can be divided into one or more modules / units. One or more modules / units are stored in the memory 1402 and executed by the processor 1401 to complete this application. One or more modules / units can be a series of computer-readable instruction segments capable of completing specific functions, and the computer-readable instruction segments are used to describe the execution process of the computer-readable instructions in the electronic device 10.

[0224] For the detailed content of the functions of each module / unit, reference can be made to the above Figure 2 、 Figure 5 、 Figure 7 、 Figures 9 - 11 detailed description, and it will not be repeated here.

[0225] In several embodiments provided in this application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of modules is only a logical function division, and there can be other division methods in actual implementation.

[0226] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical units, that is, they can be located in one place, or they can be distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0227] In addition, in each embodiment of the present application, each functional module can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above integrated unit can be implemented in the form of hardware, or in the form of a combination of hardware and software functional modules.

[0228] Therefore, from any perspective, the embodiments should be regarded as exemplary and non-restrictive. The scope of the present application is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be encompassed within the present application. Any reference signs in the claims should not be construed as limiting the claims involved.

[0229] Furthermore, it is obvious that the term "including" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or devices can also be implemented by one unit or device through software or hardware. The terms such as "first" and "second" are used to represent names and do not indicate any specific order.

[0230] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application and not to limit them. Although the present application has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present application can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present application.

Claims

1. A monitoring method for PVD processing, characterized in that, The method includes: Obtaining PVD processing data of a processing device when processing a product; Performing feature processing on the PVD processing data to obtain PVD processing features, where the PVD processing data includes multiple PVD processing factors, and correspondingly, the PVD processing features include multiple PVD feature factors; Inputting multiple PVD feature factors into a PVD processing prediction model to obtain a product demolding value.

2. The monitoring method for PVD processing according to claim 1, characterized in that, It further includes: Dividing multiple PVD feature factors into a first set and a second set, where the number of types of PVD feature factors in the first set is less than the number of types of PVD feature factors in the second set; Inputting the PVD feature factors in the first set into the PVD processing prediction model to obtain a first demolding value; Inputting the PVD feature factors in the second set into the PVD processing prediction model to obtain a second demolding value; Calculating the contribution degree corresponding to each PVD feature factor based on the number of types of PVD feature factors in the first set, the number of types of multiple PVD feature factors, the first demolding value, and the second demolding value.

3. The monitoring method for PVD processing according to claim 2, wherein, The step of dividing multiple PVD feature factors into a first set and a second set includes: Randomly extracting N types of PVD feature factors from multiple PVD feature factors; Constructing the first set based on the N types of PVD feature factors; Further extracting Y types of PVD feature factors based on the PVD feature factors remaining after randomly extracting the N types of PVD feature factors from multiple PVD feature factors, where N > Y, N is a positive integer greater than or equal to 1, and Y is a positive integer greater than or equal to 1; Constructing the second set based on the N types of PVD feature factors and the Y types of PVD feature factors.

4. The monitoring method for PVD processing according to claim 1, wherein The PVD processing data includes a PVD time series, each time point in the PVD time series carries a PVD temperature, the PVD temperature includes an etching temperature, the PVD feature factor includes an etching time feature, and the step of performing feature processing on the PVD processing data to obtain PVD processing features includes: Identifying the highest etching temperature and the corresponding time point based on the etching temperature carried by each time point in the PVD time series; Determining the etching start temperature and the corresponding time point based on the etching temperature carried by each time point in the PVD time series and a preset temperature; Determining the etching time feature based on the time point corresponding to the highest etching temperature and the time point corresponding to the etching start temperature.

5. The monitoring method for PVD processing according to claim 4, wherein Each time point in the PVD time series further carries a PVD pumping pressure, the PVD feature factor further includes a PVD first pumping pressure feature and a PVD second pumping pressure feature, and the step of performing feature processing on the PVD processing data to obtain PVD processing features further includes: Respectively extracting the PVD pumping pressure and the corresponding time point that are the same as a first preset pumping pressure and a second preset pumping pressure based on the PVD pumping pressure carried by each time point in the PVD time series; The PVD pumping pressures that are the same as the first preset pumping pressure and the second preset pumping pressure are respectively determined as the PVD first pumping pressure feature and the PVD second pumping pressure feature.

6. The monitoring method for PVD processing according to claim 5, characterized in that, The PVD feature factors further include a PVD first temperature feature and a PVD second temperature feature. The step of performing feature processing on the PVD processing data to obtain PVD processing features further includes: Based on the time points corresponding to the PVD first pumping pressure feature and the time points corresponding to the PVD second pumping pressure feature, two PVD temperatures corresponding to the two time points are respectively extracted from the PVD temperatures carried by each time point in the PVD time series. The PVD temperatures corresponding to the two time points are respectively determined as the PVD first temperature feature and the PVD second temperature feature.

7. The monitoring method for PVD processing according to claim 4, characterized in that, Each time point in the PVD time series also carries a workpiece voltage, and the PVD temperature further includes a coating temperature; the PVD feature factors further include any one or more of an average etching temperature, a maximum workpiece voltage, a minimum workpiece voltage, and an average coating temperature; the step of performing feature processing on the PVD processing data to obtain PVD processing features further includes any one or more of the following steps: Calculating the average etching temperature based on the etching temperature carried by each time point in the PVD time series. Determining the maximum workpiece voltage and the minimum workpiece voltage based on the workpiece voltage carried by each time point in the PVD time series. Calculating the average coating temperature based on the coating temperature carried by each time point in the PVD time series.

8. The monitoring method for PVD processing according to claim 1, wherein The method further includes: Obtaining multiple contribution degrees of each PVD feature factor within a preset time period. Calculating the average contribution degree of each PVD feature factor based on the multiple contribution degrees of each PVD feature factor. Sorting all types of PVD feature factors based on the average contribution degree of each PVD feature factor to obtain a queue. Selecting M PVD feature factors as key PVD feature factors from the queue based on a preset rule, where M is a positive integer greater than or equal to 1.

9. The monitoring method for PVD processing according to claim 8, wherein, The method further includes: Determining the components to be maintained in the processing equipment corresponding to the key PVD feature factors. Obtaining the maintenance strategy corresponding to the components to be maintained to recommend the maintenance strategy to the user.

10. A method for training a PVD processing prediction model, characterized in that, The method includes: Obtaining multiple groups of historical PVD processing data and the historical product demolding values of each corresponding product. Performing feature processing on the multiple groups of historical PVD processing data to obtain historical PVD processing features. Each group of PVD processing data includes multiple historical PVD processing factors. Correspondingly, the historical PVD processing features include multiple historical PVD feature factors. Performing feature processing on the historical product demolding values to obtain historical product demolding feature values. Training an initial prediction model based on the multiple historical PVD feature factors, the historical product demolding values, and the historical product demolding feature values to obtain a PVD processing prediction model.

11. The PVD processing prediction model training method according to claim 10, wherein Training an initial prediction model based on the multiple historical PVD characteristic factors, the historical product demolding values, and the historical product demolding characteristic values to obtain a PVD processing prediction model, including: Inputting the multiple historical PVD characteristic factors into the initial prediction model to obtain a first predicted demolding value and the contribution degree of each historical PVD characteristic factor; Determining the accuracy of the initial prediction model based on the first predicted demolding value, the historical product demolding value, and the historical product demolding characteristic value; If the accuracy is within a preset range, screening out the key historical PVD characteristic factors from the multiple historical PVD characteristic factors according to a preset screening rule based on the contribution degree of each historical PVD characteristic factor; Inputting the key historical PVD characteristic factors into the initial prediction model to obtain a second predicted demolding value; Determining a first loss value of the initial prediction model based on the first predicted demolding value and the historical product demolding characteristic value; Determining a second loss value of the initial prediction model according to the second predicted demolding value and the historical product demolding characteristic value; Determining whether to use the trained initial prediction model as the PVD processing prediction model based on the first loss value, the second loss value, and the number of types of the key historical PVD characteristic factors; 12. The PVD processing prediction model training method according to claim 11, wherein, The step of determining whether to use the trained initial prediction model as the PVD processing prediction model based on the first loss value, the second loss value, and the number of types of the key historical PVD characteristic factors includes: Determining the absolute value of the difference between the first loss value and the second loss value; If the absolute value of the difference is less than a first preset value, repeatedly execute the step of "training the initial prediction model based on the multiple historical PVD characteristic factors, the historical product demolding value, and the historical product demolding characteristic value" based on the key historical PVD characteristic factors and their corresponding historical product demolding values and historical product demolding characteristic values until the number of types of the key historical PVD characteristic factors is less than or equal to a second preset value, so as to execute the step of "using the trained initial prediction model as the PVD processing prediction model"; 13. The PVD processing prediction model training method according to claim 11, wherein, The step of determining whether to use the trained initial prediction model as the PVD processing prediction model based on the first loss value, the second loss value, and the number of types of the key historical PVD characteristic factors further includes: If the absolute value of the difference between the first loss value and the second loss value is greater than or equal to the first preset value, re-screening the key historical PVD characteristic factors from the multiple historical PVD processing factors according to the preset screening rule based on the contribution degree of each historical PVD processing factor; Based on the re-screened key historical PVD characteristic factors, execute the step of "inputting the PVD characteristic factors corresponding to the key historical PVD characteristic factors into the initial prediction model to obtain a second predicted demolding value"; 14. The PVD processing prediction model training method according to claim 11, wherein, The method further includes: If the accuracy is not within the preset range, obtain new historical PVD processing data to train the initial prediction model.

15. The PVD processing prediction model training method according to claim 11, wherein, The historical product demolding values of each product include the historical product demolding values of multiple demolding points. Determining the accuracy of the initial prediction model based on the first predicted demolding value, the historical product demolding values, and the historical product demolding characteristic values includes: Based on the historical product demolding values of multiple demolding points of each product, determine the final standard deviation of the historical product demolding values of each demolding point of each product; According to the historical product demolding characteristic values, the final standard deviation, the first predicted demolding value, and a third preset value of each product, determine the accuracy.

16. The PVD processing prediction model training method according to claim 15, wherein It further includes: Judge whether the historical product demolding values of multiple demolding points of each product are independent of each other; If the historical product demolding values of multiple demolding points of each product are independent of each other, judge whether the data distributions of the historical product demolding values of the demolding points of the product are the same; If the data distributions of the historical product demolding values of the demolding points of the product are the same, execute the step of "Based on the historical product demolding values of multiple demolding points of each product, determine the final standard deviation of the historical product demolding values of each demolding point of each product".

17. The PVD processing prediction model training method according to claim 16, wherein, It further includes: If the data distributions of the historical product demolding values of the demolding points of the product are different, obtain the historical product demolding values of the demolding points of the new product to form reference product demolding values; According to the historical product demolding values and the reference product demolding values, calculate the intermediate value standard deviation corresponding to the historical product demolding values of each demolding point; Based on the intermediate value standard deviation corresponding to the historical product demolding values of each demolding point, obtain the final standard deviation of the historical product demolding values of each demolding point.

18. The PVD processing prediction model training method according to claim 16, wherein The judgment of whether the historical product demolding values of multiple demolding points of multiple products are independent of each other includes: Obtain the demolding coordinates of multiple demolding points in each product; According to the demolding coordinates of the demolding points corresponding to the historical product demolding values of each demolding point, calculate the demolding coefficient of each demolding point; If the absolute value of the demolding coefficient of each demolding point is less than a first preset threshold, determine that the historical product demolding values of multiple demolding points in each product are independent of each other; If the absolute value of the demolding coefficient of each demolding point is greater than or equal to the first preset threshold, determine that the historical product demolding values of multiple demolding points in each product are not independent of each other.

19. The model training method according to claim 16, wherein The judgment of whether the data distributions of the historical product demolding values of multiple demolding points are the same includes: According to the historical product demolding values of multiple demolding points, respectively form multiple demolding curves; Determine the curve similarity between each demolding curve and a preset distribution curve; If the curve similarity is less than a second preset threshold, determine that the data distributions of the historical product demolding values of multiple demolding points are different; If the curve similarity is greater than or equal to the second preset threshold, determine that the data distributions of the historical product demolding values of multiple demolding points are the same.

20. An electronic device, characterized in that, Including a memory, a processor, and computer-readable instructions stored in the memory and executable on the processor, the computer-readable instructions, when executed by the processor, implement the monitoring method for PVD processing according to any one of claims 1 to 9 or the training method for the PVD processing prediction model according to any one of claims 10 to 19.

21. A computer-readable storage medium, characterized in that: Computer-readable instructions are stored in the computer-readable storage medium, and the computer-readable instructions are executed by a processor in an electronic device to implement the monitoring method for PVD processing according to any one of claims 1 to 9 or the training method for the PVD processing prediction model according to any one of claims 10 to 19.

Images