Product sampling method and apparatus, and storage medium

Abstract

The application provides a product sampling method and device and a storage medium. The method comprises the following steps: acquiring a current process performance index (PPK) value corresponding to a target quality parameter of a target product in a current production process; determining a statistical risk parameter associated with the target quality parameter; querying a target sampling sample size matched with the current PPK value and the statistical risk parameter based on a pre-constructed mapping relationship table; the mapping relationship table comprises a negative correlation between a preset PPK value and a sampling sample size under the constraint of different statistical risk parameter combinations; and outputting the target sampling sample size to indicate that sampling detection is performed on the target quality parameter of the target product. The application can more reasonably determine the sampling sample size based on the product process capability, and effectively alleviates the problem that the statistical risk is uncontrollable due to the disconnection between the traditional static sampling and the process capability.

Description

Technical Field

[0001] This application relates to the field of industrial process control technology, and in particular to a product sampling method, apparatus and storage medium. Background Technology

[0002] In the product manufacturing process, online sampling and testing of key quality characteristics is an important means to ensure product quality stability. Currently, the industry commonly uses fixed-rule sampling methods based on engineering experience, such as fixed-quantity sampling based on batch size, sampling at fixed percentages (e.g., 1%, 5%), or using interval sampling strategies (e.g., sampling one unit out of every three produced). These methods are widely used in various manufacturing enterprises due to their simplicity and ease of implementation.

[0003] However, the formulation of the above-mentioned traditional sampling schemes relies on engineering experience or fixed rules, without taking into account the process capability that reflects the actual quality level of the process. In practical applications, this can lead to uncontrollable statistical risks: for production lines with high process capability, the fixed sampling size may be too high, resulting in waste of inspection resources and increased costs; while for production lines with low process capability, the same sampling size may be too low, making it difficult to detect process abnormalities, thereby increasing the risk of missing nonconforming products. Summary of the Invention

[0004] In view of this, embodiments of this application provide a product sampling method, apparatus, and storage medium to solve at least one problem existing in the background art.

[0005] In a first aspect, embodiments of this application provide a product sampling method, the method comprising: Obtain the current process performance index (PPK) value corresponding to the target quality parameters of the target product in the current production process; Determine the statistical risk parameters associated with the target quality parameters; Based on a pre-built mapping table, the target sampling size that matches the current PPK value and the statistical risk parameter is queried; the mapping table contains a pre-defined negative correlation between the PPK value and the sampling size under the constraints of different combinations of statistical risk parameters. Output the target sampling size to indicate that sampling inspection will be performed on the target quality parameter of the target product.

[0006] In conjunction with the first aspect, in an alternative implementation, the mapping table is constructed in the following manner: For each of the multiple preset PPK values, the process standard deviation, which is inversely proportional to the preset PPK value, is calculated based on the preset PPK value and the standardized specification tolerance of the target quality parameter. For each of the multiple preset statistical risk parameter combinations, the statistical risk parameter combination and the equivalent value of the process standard deviation which is inversely proportional to the preset PPK value are input into the sample size calculation model to calculate the sampling sample size corresponding to the preset PPK value under the constraint of the statistical risk parameter combination. Based on the preset PPK value, the combination of statistical risk parameters, and the calculated sampling size, the mapping relationship table is constructed and stored.

[0007] In conjunction with the first aspect, in an optional implementation, the step of calculating the process standard deviation, which is inversely proportional to the preset PPK value, for each of a plurality of preset PPK values, based on the preset PPK value and the standardized specification tolerance of the target quality parameter, includes: For each of the multiple preset PPK values, the specification tolerance of the target quality parameter after standardization is divided by the product of the preset PPK value and the preset multiplier factor to obtain the process standard deviation and other values. The preset multiplier factor is used to characterize the range of process fluctuation distribution on which the process capability evaluation is based. The converted value of the process standard deviation, obtained by converting it through the specification tolerance of the target quality parameter, satisfies the following relationship with the actual process standard deviation: When the process center of production coincides with the specification center, the conversion value is equal to the actual standard deviation of the process. When the process center and specification center of the production process are inconsistent, the conversion value is greater than the actual standard deviation of the process, so that the calculated sample size is increased accordingly.

[0008] In conjunction with the first aspect, in an optional implementation, the sample size calculation model includes a sample size calculation formula based on a normal distribution and a sample size calculation formula based on a t-distribution; The step of inputting the combination of statistical risk parameters and the estimated process standard deviation, which is inversely proportional to the preset PPK value, into the sample size calculation model to calculate the sampling sample size corresponding to the preset PPK value under the constraints of the combination of statistical risk parameters includes: Substituting the combination of statistical risk parameters and the estimated process standard deviation into the sample size calculation formula based on the normal distribution, we obtain the initial sample size estimate. Using the initial sample size estimate as the current sample size, iteratively execute the following steps until the sample size estimates obtained from two adjacent iterations are the same after rounding: determine the degrees of freedom of the t-distribution based on the current sample size, substitute the combination of statistical risk parameters and the estimated process standard deviation into the sample size calculation formula based on the t-distribution to obtain a new sample size estimate, and update the current sample size with the new sample size estimate; The rounded sample size estimate is used as the sampling sample size corresponding to the combination of statistical risk parameters and the preset PPK value.

[0009] In conjunction with the first aspect, in an optional implementation, obtaining the current process performance index (PPK) value corresponding to the target quality parameters of the target product in the current production process includes: Statistical analysis is performed on the historical measurement data of the target quality parameter in the current production process to calculate the current PPK value; Alternatively, a PPK recommended value can be generated based on the product characteristic parameters input by the user and confirmed by the user, and then the PPK recommended value can be determined as the current PPK value. Alternatively, the PPK setting value can be received from an external system or input by the user as the current PPK value.

[0010] In conjunction with the first aspect, in an optional implementation, the current PPK value is calculated based on historical measurement data of the target quality parameter during the current production process, and the method further includes: The current PPK value is periodically calculated and updated to obtain the updated PPK value; When the updated PPK value changes effectively relative to the PPK value obtained in the previous calculation period, the query step of the target sampling sample size is triggered to update the target sampling sample size.

[0011] In conjunction with the first aspect, in an optional implementation, the updated PPK value has undergone an effective change relative to the PPK value obtained in the previous calculation period, including at least one of the following: The fluctuation range of the updated PPK value relative to the PPK value obtained in the previous calculation period exceeds the preset change threshold. The updated PPK value exceeds the preset process capability threshold; The updated PPK value shows a continuous decrease or increase over multiple consecutive calculation cycles, and the cumulative decrease or increase exceeds a preset threshold.

[0012] In conjunction with the first aspect, in an optional implementation, determining the statistical risk parameter associated with the target quality parameter includes: The system receives product characteristic levels input by the user through a visual configuration interface. Based on a preset correspondence between characteristic levels and risk parameters, it matches and obtains statistical risk parameters corresponding to the product characteristic levels, which are then used as statistical risk parameters associated with the target quality parameters; or, Receive statistical risk parameters associated with the target quality parameters input by the user.

[0013] Secondly, embodiments of this application provide a product sampling device, the device comprising: The acquisition module is used to acquire the current process performance index (PPK) value corresponding to the target quality parameters of the target product in the current production process. A determination module is used to determine statistical risk parameters associated with the target quality parameters; The query module is used to query the target sampling size that matches the current PPK value and the statistical risk parameter based on a pre-built mapping relationship table; the mapping relationship table contains a preset negative correlation between the PPK value and the sampling size under the constraints of different combinations of statistical risk parameters. The output module is used to output the target sampling sample size to indicate that sampling inspection should be performed on the target quality parameter of the target product.

[0014] Thirdly, embodiments of this application provide an electronic device including a processor, a memory, and a computer program stored in the memory and executable by the processor, wherein the processor executes the steps of the product sampling method provided in any optional embodiment of the first aspect when running the computer program.

[0015] Fourthly, embodiments of this application provide a storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the product sampling method provided in any optional embodiment of the first aspect.

[0016] This application provides a product sampling method, apparatus, and storage medium. By obtaining the current PPK value corresponding to the target quality parameter in the current production process of the target product and determining the statistical risk parameter associated with the target quality parameter, a target sampling sample size matching the PPK value and statistical risk parameter is quickly obtained based on a pre-built mapping table. This sampling sample size is then output to instruct sampling inspection of the target quality parameter of the target product. Thus, by utilizing the negative correlation between PPK value and sampling sample size, the sampling size is automatically reduced when the PPK value is high to decrease unnecessary inspections, and automatically increased when the PPK value is low to improve the detection capability of process anomalies. Compared to traditional fixed-rule sampling, this application associates the product sampling size with the PPK, which reflects process capability, alleviating the problem of uncontrollable statistical risk caused by the decoupling of traditional static sampling from process capability. Furthermore, the mapping table can be pre-built and embedded in the system; in practical applications, only a table lookup is needed to obtain the target sampling sample size, eliminating the need for complex calculations. This makes it easy to integrate into existing quality management systems and has good industrial applicability. Attached Figure Description

[0017] Figure 1A schematic flowchart illustrating a product sampling method provided in an embodiment of this application; Figure 2 A flowchart illustrating a method for constructing a mapping relationship table according to an embodiment of this application; Figure 3 for Figure 2 The flowchart of step S202 is shown below; Figure 4 This is a schematic diagram of the structure of a product sampling device provided in an embodiment of this application; Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0018] To make the technical solution and beneficial effects of this application more apparent and understandable, a detailed description is provided below by listing specific embodiments. The accompanying drawings are not necessarily drawn to scale, and local features may be enlarged or reduced to more clearly show the details of the local features; unless otherwise defined, the technical and scientific terms used herein have the same meanings as those in the technical field to which this application pertains.

[0019] The embodiments described in this application are not exhaustive, but merely illustrative of some embodiments, and are not intended to limit the scope of protection of this application. Unless otherwise specified, each step in a particular embodiment can be implemented as an independent embodiment, and the steps can be arbitrarily combined. For example, a solution after removing some steps in a particular embodiment can also be implemented as an independent embodiment, and the order of the steps in a particular embodiment can be arbitrarily interchanged. Furthermore, the optional implementation methods in a particular embodiment can be arbitrarily combined; moreover, the embodiments can be arbitrarily combined with each other. For example, some or all steps of different embodiments can be arbitrarily combined, and a particular embodiment can be arbitrarily combined with the optional implementation methods of other embodiments.

[0020] In each embodiment of this application, unless otherwise specified or in case of logical conflict, the terminology and / or descriptions of the embodiments are consistent and can be referenced by each other. Technical features in different embodiments can be combined to form new embodiments based on their inherent logical relationships.

[0021] The terminology used in the embodiments of this application is for the purpose of describing specific embodiments only and is not intended to limit the scope of this application.

[0022] In some embodiments, the notation "at least one of A and B", "A and / or B", "A in one case, B in another", "A in one case, B in another", etc., may include the following technical solutions depending on the situation: in some embodiments, A (A is executed regardless of B); in some embodiments, B (B is executed regardless of A); in some embodiments, execution is selected from A and B (A and B are selectively executed); in some embodiments, both A and B are executed. The same applies when there are more branches such as A, B, C, etc.

[0023] In some embodiments, the notation "A or B" may include the following technical solutions, depending on the situation: in some embodiments, A (execution of A regardless of B); in some embodiments, B (execution of B regardless of A); in some embodiments, selective execution from A and B (A and B are selectively executed). The same applies when there are more branches such as A, B, and C.

[0024] The inventors discovered that, under the same producer risk α and consumer risk β requirements, the sample size required to detect the same magnitude of process mean shift Δ varies depending on the process performance index (PPK). For example, when α is 0.05, β is 0.1, and Δ is 0.1, a process with PPK 1 requires approximately 60 samples, while a process with PPK 3 requires only 8. Conversely, if the same fixed sample size is applied to all processes, the actual statistical risks differ. For instance, with a fixed sample size of 60 samples and Δ of 0.1, the α for a PPK 1 process is 0.05 and β is 0.1, while the α and β for a PPK 3 process both drop to the order of 0.00001. This indicates that excessively stringent testing standards are imposed on high-quality processes, while low-quality processes risk insufficient testing. This fluctuation in statistical risk due to differences in process capability makes it difficult for traditional empirical sampling methods to balance statistical risk control and testing efficiency.

[0025] Therefore, this application provides a product sampling method applicable to scenarios where continuous quality parameters of semi-finished or finished products are sampled and inspected during online production in the manufacturing industry. The sampled products can be semiconductor products such as wafers, or various industrial products with continuous measurement characteristics, such as machined parts, electronic components, chemical materials, or food packaging. This method can dynamically determine the sample size based on the PPK value during the product manufacturing stage when the process is controlled, the standard deviation is known or estimable, and the distribution is normal or approximately normal, thus achieving sampling optimization under controllable statistical risk.

[0026] Figure 1This is a schematic flowchart of a product sampling method provided in an embodiment of this application. This product sampling method can be applied to a manufacturing execution system (MES), a statistical process control (SPC) system, or a quality monitoring system.

[0027] like Figure 1 As shown, the method includes steps S101 to S104.

[0028] S101: Obtain the current PPK value corresponding to the target quality parameters of the target product in the current production process.

[0029] Here, the target product refers to the product whose quality characteristics need to be monitored during the production process. The target quality parameter is continuous measurement data, a measurable indicator that characterizes the key quality characteristics of the target product in a specific process. For example, the target quality parameter can be a key quality parameter of a wafer, such as film thickness, groove width, roughness, or groove depth. Another example is the aperture of a mechanical part or the resistance value of an electronic component. This application does not limit the specific type of the target quality parameter, as long as it belongs to the measurable continuous data of the target product.

[0030] Target quality parameters can be acquired in real time through online measurement devices (such as laser diameter gauges, spectrometers, and sensors) or through offline sampling inspections, and stored in the Manufacturing Execution System (MES) or data warehouse for subsequent analysis.

[0031] PPK (Process Performance Scale) is a statistical indicator that measures the ability of a production process to meet specifications. It comprehensively considers process fluctuations and the deviation of the process center from the specification center. In this application, the value of PPK can be any real number. In practical applications, PPK values ​​can be specific values ​​such as 1.0, 1.33, 1.67, 2.0, or 3.0. The larger the PPK value, the better the process capability and the more stable the product quality; conversely, the smaller the PPK value, the more likely there is significant process variation or center deviation.

[0032] In some examples, the current PPK value can be obtained in multiple ways, including any of the following methods: Method 1: Statistical analysis is performed on historical measurement data of the target quality parameters during the current production process to calculate the current PPK value. The calculation of this current PPK value can be automatically completed by the system according to the calculation formula of the process performance index to reflect the actual performance of the current process, and can be updated according to a preset cycle (such as hourly, daily, or per batch).

[0033] For bilateral specifications, the calculation formula is: For unilateral specifications, the calculation formula is: or Where USL and LSL are the upper and lower limits of the target quality parameter, respectively. The mean of the process. The process standard deviation is used to characterize the fluctuation or dispersion of product quality characteristic values.

[0034] Method 2: A recommended PPK value is generated based on user-inputted product characteristic parameters. After user confirmation, the recommended PPK value is determined as the current PPK value corresponding to the target quality parameter. Product characteristic parameters can be input by the user through a visual configuration interface; these parameters can be product grade or quality requirement level. The system maps the product characteristic parameter to the corresponding recommended PPK value according to preset mapping rules. This recommended PPK value is displayed to the user through a visual interface, and in response to the user's confirmation command for the recommended value, it is determined as the current PPK value corresponding to the target quality parameter. The preset mapping rules can be a pre-established correspondence based on historical process performance data of similar products or processes. For example, based on different product types, process complexity, or precision levels, corresponding initial PPK estimates can be pre-set. When a new product is introduced or historical measurement data for the current process is lacking, the system can match the corresponding recommended PPK value according to the product characteristic parameters as the initial basis for obtaining the target sampling sample size.

[0035] Method 3: Receive the PPK setting value issued by an external system or input by the user, and use it as the current PPK value corresponding to the target quality parameter. The external system can be an upstream process quality assessment system or an Enterprise Resource Planning (ERP) system, etc.

[0036] It is worth noting that regardless of whether the current PPK value is obtained by calculation from actual data, determined after confirmation based on recommended values ​​generated from product characteristics, or directly input by an external system or user, it can be used as an input parameter reflecting the process capability level in this application for dynamic adjustment of the subsequent sampling sample size.

[0037] Furthermore, it's understandable that the current PPK value is not static. For example, if changes occur in the production process or if process performance indicators are detected to exceed the preset stable range, it's necessary to obtain an updated PPK value. This updated PPK value can be used to query the target sampling size, thus ensuring that the sampling strategy dynamically matches the current process performance.

[0038] S102: Determine the statistical risk parameters associated with the target quality parameters.

[0039] Statistical risk parameters may include α risk parameters, β risk parameters, and the minimum process mean offset Δ that needs to be detected.

[0040] The α risk parameter corresponds to producer risk, also known as Type I error or false negative error, representing the probability that a process is actually acceptable but is mistakenly judged as unacceptable (i.e., false alarm risk). The β risk parameter corresponds to user risk (or consumer risk), also known as Type II error or false negative error, representing the probability that a process deviation has actually occurred but has not been detected (i.e., missed alarm risk). Δ represents the minimum deviation of the process mean that must be identified. At a standardized tolerance of 1, the deviation Δ ranges from 0 to 1, and can be set to 0.10, 0.15, or 0.20 based on the stringency of quality requirements, corresponding to the ability to identify mean deviations within 10%, 15%, or 20% of the tolerance range, respectively.

[0041] In this embodiment, the statistical risk parameters can be set according to the quality requirements of the target product, industry standards, or customer agreements. For example, for critical characteristics involving safety or function, lower α and β values ​​and a smaller offset Δ can be set, such as α=0.05, β=0.10, corresponding to a confidence level of 95% and a test power of 90%, with the offset Δ set to a 10% tolerance. For general characteristics, the risk requirements and a larger offset Δ can be appropriately relaxed, for example, α=0.10, β=0.20, corresponding to a confidence level of 90% and a test power of 0.8, with the offset Δ set to a 15% or 20% tolerance.

[0042] For ease of use, the system can preset multiple product characteristic levels, such as critical characteristics, important characteristics, and general characteristics. Each level corresponds to a set of default statistical risk parameters. Users only need to select the corresponding product characteristic level, and the system will automatically match the corresponding statistical risk parameters, eliminating the need for manual input of complex values. Table 1 shows some product characteristics of wafers and their corresponding statistical risk parameters.

[0043] Table 1:

[0044] Confidence Level represents the confidence level (or confidence degree), which is the probability that the confidence interval contains the true parameter in repeated sampling. It is related to α and is set to 1-α. Power represents the test power, which is the probability of correctly rejecting the null hypothesis when the null hypothesis is actually false. The higher the power, the lower the risk of committing a β error. It is related to β and is set to 1-β.

[0045] In some examples, the implementation of step S102 above may include any of the following methods: Method 1: Receive statistical risk parameters associated with the target quality parameter input by the user. The system provides a user configuration interface, allowing operators to directly input the specific values ​​of α, β, and Δ, and store them in association with the target quality parameter.

[0046] Method 2: The system receives product characteristic levels input by the user through a visual configuration interface. Based on the preset correspondence between characteristic levels and risk parameters, it matches and obtains the statistical risk parameters corresponding to the product characteristic level, which are then used as the statistical risk parameters associated with the target quality parameter. The system can preset multiple product characteristic levels, such as critical characteristics, important characteristics, and general characteristics, each level corresponding to a set of default statistical risk parameters. The user only needs to select the corresponding product characteristic level, and the system can automatically match the corresponding combination of statistical risk parameters, thereby simplifying the operation process.

[0047] S103: Based on the pre-built mapping relationship table, query the target sampling sample size that matches the current PPK value and statistical risk parameters; the mapping relationship table contains the mapping relationship between the preset PPK value and the sampling sample size under the constraints of different combinations of statistical risk parameters.

[0048] To quickly and accurately obtain the sampling size appropriate to the current process capability, this embodiment pre-constructs a mapping table. This table provides the sampling size that meets risk control requirements for various preset PPK values ​​(e.g., 0.5, 0.8, 1.0, 1.33, 1.67, 2.0, 2.5, 3.0, etc.) under different combinations of statistical risk parameters (α, β, Δ). For the same combination of statistical risk parameters, the PPK value is correlated with the mapped sample size; that is, the higher the PPK value, the smaller the mapped sample size; and the lower the PPK value, the larger the mapped sample size. This negative correlation means that when the process capability is better, a smaller sample size can be used for monitoring, thereby reducing testing costs; conversely, when the process capability is poor, a larger sample size is needed to improve the ability to detect anomalies.

[0049] Specifically, the current PPK value obtained in step S101 and the statistical risk parameter determined in step S102 can be used as query conditions to find the corresponding target sampling size from the pre-built mapping table. If there is an entry in the mapping table that completely matches the current PPK value, the corresponding sample size is directly read as the target sample size.

[0050] In some examples, if the current PPK value is between two adjacent preset PPK values, linear interpolation or nearest neighbor interpolation can be used to estimate the sample size in order to obtain the target sample size.

[0051] Specifically, assume that two adjacent preset PPK values are PPK1 and PPK2 (PPK1 < PPK2), and the corresponding sampling sample sizes in the mapping relationship table are n1 and n2 respectively. For the current PPK value x (PPK1 < x < PPK2), the linear interpolation formula can be used to calculate the corresponding sample size estimate n as the target sample size, where n = n1 + (n2 - n1) × (x - PPK1) / (PPK2 - PPK1). Additionally, for simplicity of calculation, the nearest neighbor interpolation method can also be used, that is, taking the sample size corresponding to the preset PPK value closer to the current PPK value as the target sample size. For example, if x is closer to PPK1, then take n = n1; if x is closer to PPK2, then take n = n2. Specifically, the appropriate interpolation method can be selected according to actual application requirements.

[0052] As shown in Table 2, this mapping relationship table is indexed by PPK values and statistical risk parameters, and pre-stores the sampling sample sizes corresponding to different PPK values under different combinations of statistical risk parameters.

[0053] Table 2:

[0054] It can be understood that the example shown in Table 2 is only a partial example of this mapping relationship table. In actual applications, the range and density of preset PPK values can be expanded as needed, or more combinations of statistical risk parameters can be added. This application does not make specific limitations in this regard.

[0055] S104: Output the target sampling sample size to indicate the sampling inspection of the target quality parameter of the target product.

[0056] In this embodiment, the target sampling sample size is the number of units to be detected. According to different detection scenarios, the detection unit can be a single product or a single measurement point on the product. For example, in wafer manufacturing, if the detection strategy is based on the wafer as the basic unit, the detection unit is a single wafer, and the target sampling sample size can be set as the number of wafer pieces; if the detection strategy is based on the measurement points on the wafer as the basic unit, the detection unit is a single measurement point, and the target sampling sample size can be set as the number of single-wafer measurement points.

[0057] In some examples, the system can output a control command carrying the target sampling sample size, which is transmitted to the sampling actuator on the production line. For instance, in an automated production line, this control command can be transmitted to a robotic arm controller, driving the robotic arm to randomly grab samples from the production batch according to the target sampling sample size and send them to the online inspection station; or the control command can be transmitted to a sorting device, controlling the sorting device to divert products that meet the target sampling sample size to the offline inspection area. In other examples, the system can display the target sampling sample size on the human-machine interface, prompting quality inspectors to perform manual sampling inspection according to this quantity.

[0058] The system can obtain measurement results after measuring the target quality parameters of the target product based on the target sampling sample size. These measurement results can be used for subsequent process performance evaluation and PPK value update to form a closed-loop control from sampling size determination, detection execution to process performance feedback.

[0059] In some embodiments, after step S101, the changes in process standard deviation can be monitored based on the measurement data of the target quality parameters collected and stored in real time. For example, when the system detects that the actual process standard deviation has changed effectively relative to the previous calculation period (e.g., the fluctuation amplitude exceeds the preset standard deviation change threshold, exceeds the preset fluctuation level threshold, or shows a unidirectional increasing trend in multiple consecutive calculation periods and the cumulative increase exceeds the preset cumulative threshold), step S101 is automatically re-executed to obtain the updated PPK value, thereby updating the target sampling sample size. In this way, more timely adjustments to the sampling strategy can be achieved by monitoring the process standard deviation.

[0060] Compared to traditional fixed-rule sampling, the product sampling method provided in this embodiment utilizes the negative correlation between PPK value and sample size. This allows the sampling size to be automatically reduced when the PPK value is high to decrease unnecessary inspections, and automatically increased when the PPK value is low to improve the detection capability of process anomalies. This alleviates the uncontrollable statistical risk problem caused by the decoupling of traditional static sampling from process capability. Furthermore, the mapping table can be pre-built and embedded in the system. In practical applications, the target sample size can be obtained simply by looking up the table, without complex calculations. This method is easily integrated into existing quality management systems and has good industrial applicability.

[0061] In some embodiments, a method for constructing a mapping table is provided, and the mapping table constructed by this method can be used to implement... Figure 1 The product sampling method shown is as follows: Figure 2 As shown, the method may include the following steps: S201: For each of the multiple preset PPK values, calculate the process standard deviation, etc., which are inversely proportional to the preset PPK value, based on the preset PPK value and the specification tolerance of the target quality parameter after standardization.

[0062] The preset PPK value can be set to multiple discrete values ​​according to actual needs, such as 0.8, 1.0, 1.33, 1.67, 2.0, 2.5, 3.0, etc., to cover process capability levels from low to high. To construct a universal mapping table, the specification tolerances of the target quality parameters are first standardized by setting the tolerance value to 1. Based on this, for each preset PPK value, the process standard deviation, which is inversely proportional to the preset PPK value, can be derived using the definition formula of the process performance index.

[0063] In some examples, for each of the multiple preset PPK values, the standardized specification tolerance of the target quality parameter can be divided by the product of the preset PPK value and the preset multiplier factor to obtain the process standard deviation, etc., where the preset multiplier factor is used to characterize the range of process fluctuation distribution on which the process capability evaluation is based.

[0064] Specifically, the process standard deviation equivalent value can be calculated using the following formula: ;in, This represents the tolerance, which is the specification tolerance of the quality characteristic to be measured. Its value is standardized to 1. The preset multiplier factor is set to 6, representing the value of the process standard deviation under a standardized scale.

[0065] The converted value obtained by converting the equivalent value of the process standard deviation to the specification tolerance of the target quality parameter satisfies the following relationship with the actual process standard deviation: when the process center of the production process coincides with the specification center, the converted value is equal to the actual process standard deviation; when the process center of the production process does not coincide with the specification center, the converted value is greater than the actual process standard deviation, so that the calculated sampling size is increased accordingly.

[0066] As shown in Table 3 below, taking the film thickness of a wafer as an example, when both the process center and the specification center are 12000 μm and PPK=1, the actual standard deviation of the process is 333 μm, and the equivalent value after specification tolerance conversion is also 333 μm; however, when the process center shifts to 12700 μm and PPK=0.3, the actual standard deviation of the process is still 333 μm, but the equivalent value after conversion is 1110 μm, which is greater than the actual standard deviation of the process. Therefore, the sampling size calculated based on this equivalent value is increased accordingly to ensure the quality of the actual process.

[0067] Table 3:

[0068] S202: For each of the multiple preset statistical risk parameter combinations, input the statistical risk parameter combination and the equivalent value of the process standard deviation which is inversely proportional to the preset PPK value into the sample size calculation model to calculate the sampling sample size corresponding to the preset PPK value under the constraint of the statistical risk parameter combination.

[0069] The combination of statistical risk parameters can have preset values ​​according to different quality characteristics, such as (α=0.05, β=0.10, Δ=0.1), (α=0.05, β=0.10, Δ=0.15), (α=0.10, β=0.20, Δ=0.1), etc., to cover different risk control needs.

[0070] For each preset PPK value and each combination of statistical risk parameters, the equivalent value of the process standard deviation calculated in step S201 based on the preset PPK value is substituted into the sample size calculation model along with the combination of statistical risk parameters to calculate the corresponding sampling size. This sample size calculation model can be based on the comparison of two sample means to determine the minimum sample size required to detect a specified shift in the process mean under a given combination of statistical risk parameters.

[0071] In some examples, the sample size calculation model can be a two-sample hypothesis testing model, which combines the normal distribution and the t-distribution to calculate the sample size. The quantiles of the t-distribution depend on the sample size itself, so an iterative algorithm can be used to calculate them until a convergent sample size estimate is obtained.

[0072] S203: Based on the preset PPK value, the combination of statistical risk parameters, and the calculated sampling size, construct a mapping relationship table and store it.

[0073] The mapping table can be stored in the system's non-volatile memory or set in the configuration file of the quality monitoring software.

[0074] In this embodiment, the system can quickly obtain the target sampling size that matches the current PPK value and statistical risk parameters by looking up a table, without having to repeat complex statistical calculations, thereby improving the efficiency and real-time performance of the sampling plan.

[0075] It is understandable that the mapping table is not static. It can be dynamically updated according to application needs. For example, as process data accumulates, steps S201 to S203 can be re-executed based on the updated preset PPK value range or newly added statistical risk parameter combinations to calculate a new sampling size, which is then added to the mapping table to expand its coverage. Furthermore, when certain statistical risk parameter combinations are no longer applicable, or when new data reveals that the original sample size calculation results no longer accurately reflect the current process characteristics, the corresponding entries in the mapping table can be deleted or replaced. This dynamic maintenance mechanism ensures that the mapping table always matches the production process and quality control requirements, further improving the accuracy and adaptability of the sampling plan.

[0076] In some embodiments, the sample size calculation model may include sample size calculation based on a normal distribution and sample size calculation based on a t-distribution; such as Figure 3 As shown, the implementation process of step S202 above may include the following steps: S301: Substitute the combination of statistical risk parameters and the estimated process standard deviation into the sample size calculation formula based on the normal distribution to obtain the initial sample size estimate; S302: Using the initial sample size estimate as the current sample size, iteratively execute the following steps until the sample size estimates obtained from two adjacent iterations are the same after rounding: Determine the degrees of freedom of the t-distribution based on the current sample size, substitute the combination of statistical risk parameters and the estimated process standard deviation into the sample size calculation formula based on the t-distribution to obtain a new sample size estimate, and update the current sample size with the new sample size estimate. S303: The rounded sample size estimate is used as the sampling sample size corresponding to the combination of statistical risk parameters and the preset PPK value.

[0077] Here, the formula for calculating the sample size based on the normal distribution is: (1) The formula for calculating the sample size based on the t-distribution is: (2) in, For sample size; This represents a standard normal distribution (mean 0, standard deviation 1). quantiles; This represents a standard normal distribution (mean 0, standard deviation 1). quantiles; This represents a t-distribution with n-2 degrees of freedom. Quantiles (two-tailed test); This represents a t-distribution with n-2 degrees of freedom. Quantiles (corresponding to test power); For process standard deviation, This is the offset.

[0078] First, substitute the statistical risk parameter combination and the equivalent value of the process standard deviation into the calculation formula (1) to obtain the initial sample size estimate n0; then, using this initial sample size estimate as the current sample size, iteratively execute the following steps: based on the current sample size n0 k The t-distribution has n degrees of freedom. k Under the condition of -2, the statistical risk parameter combination and process standard deviation are substituted into the calculation formula (2) to obtain a new sample size estimate n. k+1 and with n k+1 Update the current sample size. Repeat the above iterative process. When the sample size estimates obtained from two consecutive iterations are the same after rounding (e.g., both rounded up), stop the iteration process. The sample size estimate at this time is used as the sampling sample size corresponding to the combination of statistical risk parameters and the preset PPK value.

[0079] In this embodiment, since the quantiles of the t-distribution are greater than those of the normal distribution at the same confidence level, the sample size calculated directly based on the normal distribution may be underestimated and may not meet the preset statistical risk requirements. However, by gradually correcting the sample size through an iterative algorithm, the estimated sample size converges to the smallest integer value that meets the risk requirements, thereby ensuring that the final sample size can meet the preset statistical risk requirements even in small sample scenarios.

[0080] It is understood that, in other possible implementations, the combination of statistical risk parameters and the estimated process standard deviation are substituted into the sample size calculation model to calculate the sample size corresponding to the preset PPK value. The calculation methods may include, but are not limited to: single-sample Z-test sample size (population standard deviation known), single-sample Z-test sample size (population standard deviation unknown), two-sample Z-test sample size (population standard deviation known and equal), single-sample variance test, two-sample variance test, or small-sample T-test iterative method, etc.

[0081] The formula for the one-sample Z-test is as follows: ; The formula for the one-sample t-test is as follows: ; The two-sample Z-test is calculated using the following formula: .

[0082] For the relevant symbols in the above calculation formulas, please refer to the symbol explanations of formulas (1) and (2) above, which will not be repeated here.

[0083] In some embodiments, the current PPK value is calculated based on historical measurement data of the target quality parameter in the current production process, and the method may further include: Periodically calculate and update the current PPK value to obtain the updated PPK value; When the updated PPK value changes effectively relative to the PPK value obtained in the previous calculation period, the query step for the target sample size is triggered to update the target sample size.

[0084] The calculation cycle of the PPK value can be set according to the sampling cycle of the online measurement equipment, the data acquisition frequency, or the process stability requirements. For example, it can be updated once per hour, per batch, or when a preset number of product subgroups are accumulated.

[0085] In this embodiment, when the updated PPK value changes effectively relative to the PPK value obtained in the previous calculation period, the target sampling sample size query step is re-executed to obtain the target sampling sample size that matches the updated PPK value and the current statistical risk parameters, thereby realizing the dynamic adjustment of the sampling plan.

[0086] In some examples, the updated PPK value has changed effectively relative to the PPK value obtained in the previous calculation period, including at least one of the following: Scenario 1: The fluctuation range of the updated PPK value relative to the PPK value obtained in the previous calculation period exceeds the preset change threshold; the fluctuation range of the PPK value can be the absolute value of the difference between the PPK values ​​of two adjacent calculation periods, and the preset threshold can be set to 0.1, 0.2 or other suitable values ​​based on expert experience.

[0087] Scenario 2: The updated PPK value exceeds a preset process capability threshold. This process capability threshold can be a process capability level threshold, with different threshold ranges corresponding to different process capability levels. For example, when the PPK value drops from greater than 1.33 to below 1.33 (i.e., moving from a higher capability range to a lower capability range), it indicates that process fluctuations have increased relatively, and the original sampling size is insufficient to control the same level of statistical risk. In this case, based on the updated PPK value, the sampling size needs to be recalculated and increased. The process capability threshold here can be set to 1.0, 1.33, 1.67, etc.

[0088] Scenario 3: The updated PPK value shows a continuous decrease or increase over multiple consecutive calculation cycles, and the cumulative decrease or increase exceeds a preset threshold. Here, the cumulative decrease refers to the cumulative decrease from the first cycle to the current cycle, and the cumulative increase refers to the cumulative increase from the first cycle to the current cycle. In this scenario, it indicates a trend change in process performance, such as etching rate drift caused by equipment aging in wafer etching processes, or slow deterioration of film thickness uniformity due to wear of the polishing pad in chemical mechanical polishing. When such a trend change is detected, the query for the target sampling size is re-executed to maintain a dynamic correlation between the sampling size and the trend of process capability changes.

[0089] It should be noted that the above three scenarios can be used individually or in combination. For example, it is possible to simultaneously monitor fluctuation amplitude and cumulative trend to balance the ability to capture both sudden anomalies and gradual changes. By monitoring effective changes in the PPK value, the system can more accurately identify substantial changes in process capability, reduce frequent adjustments triggered by random fluctuations, and also respond promptly to real process deviations, improving the dynamic adaptability of the sampling plan.

[0090] Next, taking wafers as an example, the product sampling method provided in the embodiments of this application will be further explained.

[0091] First, obtain the current PPK value corresponding to the target quality parameter of the wafer in the current production process (e.g., film thickness after chemical mechanical polishing (CMP)). This PPK value can be calculated based on historical measurement data of the target quality parameter from a preset number of samples. This historical measurement data can come from sampling inspection (e.g., measuring only specific points on each wafer, or sampling only a portion of wafers from a batch) or from full inspection (i.e., measuring every wafer within a specified range).

[0092] Next, the statistical risk parameters associated with this target quality parameter are determined, including producer risk α, consumer risk β, and the minimum process mean offset Δ to be detected. For example, for membrane thickness, α = 0.05, β = 0.10, and Δ = 10% tolerance can be set.

[0093] Then, using the current PPK value and statistical risk parameters as query conditions, the system searches for the corresponding target sampling sample size in a pre-built mapping table. This mapping table is pre-calculated based on a sample size calculation model, which provides the sampling sample size that meets risk control requirements for different preset PPK values ​​(e.g., 0.5, 1.0, 1.33, 1.67, 2.0, etc.) and different combinations of statistical risk parameters. For example, as shown in Table 2 above, when the CMP process performance is stable (PPK ≥ 1.67), the system finds a target sampling sample size of 22, meaning that only a small number of wafers need to be sampled for thickness measurement to meet the monitoring requirements; when the process shows a performance degradation trend (e.g., the polishing pad wear causes the PPK to drop below 1.33), the system finds a target sampling sample size of 22, meaning that the sampling sample size is automatically increased to detect possible thickness anomalies with a larger sample size.

[0094] Finally, the system outputs the target sampling quantity and sends it as a control command to the sampling execution mechanism on the production line. For example, in an automated wafer transfer system, this command can control a robotic arm to randomly pick up a specified number of wafers from the current batch and send them to an online film thickness measurement device for inspection. The thickness data obtained from the measurement can be used to determine whether the current batch is qualified, and it can also be fed back to the process performance evaluation system to update the subsequent PPK value calculation, thus forming a closed-loop control from sampling quantity determination, inspection execution to process performance feedback.

[0095] The product sampling method provided in this application enables the sampling strategy in the wafer manufacturing process to be adaptively adjusted according to the dynamic changes in process performance. When the process is in a stable state, the system can reduce the sampling sample size accordingly to avoid over-inspection and waste of production capacity. When the process fluctuates, the system increases the sampling sample size to ensure that potential anomalies are detected in a timely manner.

[0096] Furthermore, this method can effectively address the incremental process changes commonly seen in wafer manufacturing. Taking the CMP process as an example, wear of the polishing pads can lead to a slow deterioration in film thickness uniformity. In the early stages of such changes, the process mean usually has not yet exceeded the specification limits, but the PPK value has already shown a continuous downward trend. By monitoring changes in the PPK value and adjusting the sampling strategy in a timely manner, the sampling frequency can be increased before the risk of missed detection increases, thus achieving early detection of process drift.

[0097] Compared to traditional fixed sampling methods, this application maintains a constant statistical risk while matching the investment of testing resources with the actual process requirements, thereby improving testing efficiency while ensuring product yield.

[0098] Based on the specific data shown in Table 4, a comparative analysis is conducted between the product sampling method of this application and the traditional fixed sampling method.

[0099] Table 4:

[0100] As shown in Table 4, a batch of products (e.g., wafers) consists of 25 units. A fixed percentage sampling of 10% is used, meaning 3 units are sampled, and 5 points are measured on each unit, for a total of 15 measurement points. When PPK = 1.0, the user risk β is as high as 0.65 with 15 measurement points, indicating a very high risk of missed detections and a severely insufficient sampling size. As PPK increases, the β risk gradually decreases. When PPK drops to 2.0, β = 0.113. However, when PPK is not lower than 2.33, the β ranges from 0.042 to 0.003. A β risk value that is too low indicates an excessive sampling size, leading to increased quality costs. Therefore, the fixed percentage method is difficult to maintain stable statistical risk at different process performance levels.

[0101] In contrast, the PPK-based sampling method of this application dynamically adjusts the sample size for different PPK values ​​through a pre-constructed mapping table, ensuring that a preset β risk (e.g., 0.1) is maintained at each PPK level. As shown in Table 4, the sample size is 60 when PPK=1.0, decreases to 34 when PPK=1.33, decreases to 22 when PPK=1.67, is 16 when PPK=2.0, is 12 when PPK=2.33, is 10 when PPK=2.67, and is 8 when PPK=3.0. As process performance improves, the sample size automatically decreases, reducing unnecessary testing while maintaining stable β risk; conversely, when process performance declines, the sample size automatically increases, ensuring sufficient testing power to detect process anomalies. Therefore, the method of this application alleviates the problems of insufficient testing for high-risk processes and excessive testing for low-risk processes in traditional sampling methods.

[0102] In summary, the product sampling method provided in this embodiment is applicable to online sampling scenarios for various continuous measurement data. In practical applications, the required sample size can be directly obtained by querying a pre-built mapping table based on the PPK value input by the engineer (or combined with statistical risk parameters, such as offset Δ). By constructing a negatively correlated mapping relationship between PPK and the sampling size, the sampling plan is based on the actual quality level of the process rather than empirical rules, and the sample size can be dynamically adjusted according to process capability. This ensures that statistical risk is controllable while making the sampling plan more reasonable and practical, and facilitates deployment in software systems.

[0103] Based on the product sampling method in the foregoing embodiments, this application also provides a product sampling device. The embodiments or implementation methods in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Similar or identical parts between embodiments can be referred to mutually. For example, optional implementation methods in the product sampling device embodiments can be found in the relevant content of the product sampling method in the foregoing embodiments, and will not be repeated here.

[0104] Figure 4 This is a schematic diagram of the structure of a product sampling device provided in one embodiment of this application. Figure 4 As shown, the product sampling device 100 includes: The acquisition module 101 is used to acquire the current process performance index PPK value corresponding to the target quality parameters of the target product in the current production process; The determination module 102 is used to determine the statistical risk parameters associated with the target quality parameters; The query module 103 is used to query the target sampling sample size that matches the current PPK value and the statistical risk parameter based on a pre-built mapping relationship table; the mapping relationship table contains a pre-defined mapping relationship between the PPK value and the sampling sample size under the constraints of different combinations of statistical risk parameters. The output module 104 is used to output the target sampling sample size to indicate that sampling inspection will be performed on the target quality parameter of the target product.

[0105] In some embodiments, the apparatus further includes a construction module, the construction module comprising: The first calculation unit is used to calculate, for each of a plurality of preset PPK values, the process standard deviation and other values ​​that are inversely proportional to the preset PPK value, based on the preset PPK value and the specification tolerance of the target quality parameter after standardization. The second calculation unit is used to input the statistical risk parameter combination and the equivalent value of the process standard deviation which is inversely proportional to the preset PPK value into the sample size calculation model for each of the multiple preset statistical risk parameter combinations, and calculate the sampling sample size corresponding to the preset PPK value under the constraint of the statistical risk parameter combination. The construction unit is used to construct and store the mapping relationship table based on the preset PPK value, the combination of statistical risk parameters, and the calculated sampling size.

[0106] In some embodiments, the first computing unit is used for: For each of the multiple preset PPK values, the specification tolerance of the target quality parameter after standardization is divided by the product of the preset PPK value and the preset multiplier factor to obtain the process standard deviation and other values. The preset multiplier factor is used to characterize the range of process fluctuation distribution on which the process capability evaluation is based. The converted value of the process standard deviation, obtained by converting it through the specification tolerance of the target quality parameter, satisfies the following relationship with the actual process standard deviation: When the process center of production coincides with the specification center, the conversion value is equal to the actual standard deviation of the process. When the process center and specification center of the production process are inconsistent, the conversion value is greater than the actual standard deviation of the process, so that the calculated sample size is increased accordingly.

[0107] In some embodiments, the sample size calculation model includes a sample size calculation formula based on a normal distribution and a sample size calculation formula based on a t-distribution; the second calculation unit is used for: Substituting the combination of statistical risk parameters and the estimated process standard deviation into the sample size calculation formula based on the normal distribution, we obtain the initial sample size estimate. Using the initial sample size estimate as the current sample size, iteratively execute the following steps until the sample size estimates obtained from two adjacent iterations are the same after rounding: determine the degrees of freedom of the t-distribution based on the current sample size, substitute the combination of statistical risk parameters and the estimated process standard deviation into the sample size calculation formula based on the t-distribution to obtain a new sample size estimate, and update the current sample size with the new sample size estimate; The rounded sample size estimate is used as the sampling sample size corresponding to the combination of statistical risk parameters and the preset PPK value.

[0108] In some embodiments, the acquisition module 101 is used to: Statistical analysis is performed on the historical measurement data of the target quality parameter in the current production process to calculate the current PPK value; Alternatively, a PPK recommended value can be generated based on the product characteristic parameters input by the user and confirmed by the user, and then the PPK recommended value can be determined as the current PPK value; Alternatively, the PPK setting value can be received from an external system or input by the user as the current PPK value.

[0109] In some embodiments, the current PPK value is calculated based on historical measurement data of the target quality parameter in the current production process, and the device further includes an update module; The update module is used to periodically calculate and update the current PPK value to obtain the updated PPK value; The query module is used to trigger the query step of the target sampling sample size when the updated PPK value has a valid change relative to the PPK value obtained in the previous calculation period, so as to update the target sampling sample size.

[0110] In some embodiments, the updated PPK value has a valid change relative to the PPK value obtained in the previous calculation period, including at least one of the following: The fluctuation range of the updated PPK value relative to the PPK value obtained in the previous calculation period exceeds the preset change threshold. The updated PPK value exceeds the preset process capability threshold; The updated PPK value shows a continuous decrease or increase over multiple consecutive calculation cycles, and the cumulative decrease or increase exceeds a preset threshold.

[0111] In some embodiments, the determining module 102 is configured to: The system receives product characteristic levels input by the user through a visual configuration interface. Based on a preset correspondence between characteristic levels and risk parameters, it matches and obtains statistical risk parameters corresponding to the product characteristic levels, which are then used as statistical risk parameters associated with the target quality parameters; or, Receive statistical risk parameters associated with the target quality parameters input by the user.

[0112] Figure 5 This application provides a schematic diagram of the structure of an electronic device, such as... Figure 5 As shown, this application embodiment also provides an electronic device, including a processor, a memory, and a computer program stored in the memory and executable by the processor. When the processor runs the computer program, it performs the steps of the product sampling method provided in any of the foregoing embodiments.

[0113] This application also provides a storage medium storing a computer program thereon. When the computer program is executed by a processor, it implements the steps of the product sampling method provided in any of the foregoing embodiments. Its implementation principle and technical effects are similar to those of the above embodiments, and will not be repeated here.

[0114] For ease of understanding, the following focuses on explaining the terminology used in this embodiment: In this application embodiment, the processor is a circuit with signal processing capabilities. In one implementation, the processor can be a circuit with instruction read and execute capabilities, such as a Central Processing Unit (CPU), a microprocessor, a Graphics Processing Unit (GPU) (which can be understood as a type of microprocessor), or a Digital Signal Processor (DSP). In another implementation, the processor can implement certain functions through the logical relationships of hardware circuits. The logical relationships of the aforementioned hardware circuits are fixed or reconfigurable. For example, the processor is a hardware circuit implemented using an Application-Specific Integrated Circuit (ASIC) or a Programmable Logic Device (PLD), such as an FPGA. In a reconfigurable hardware circuit, the process of the processor loading a configuration document and configuring the hardware circuit can be understood as the process of the processor loading instructions to implement the functions of some or all of the above units or modules. In addition, it can also be a hardware circuit designed for artificial intelligence, which can be understood as an ASIC, such as a Neural Network Processing Unit (NPU), a Tensor Processing Unit (TPU), a Deep Learning Processing Unit (DPU), etc.

[0115] The aforementioned computer-readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The readable storage medium can be any available medium accessible to a general-purpose or special-purpose computer.

[0116] An exemplary readable storage medium is coupled to a processor, enabling the processor to read information from and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can reside in an Application Specific Integrated Circuit (ASIC). Alternatively, the processor and the readable storage medium can exist as discrete components in an electronic device or a host device.

[0117] In the description of this specification, references to "one embodiment," "some embodiments," "illustrative embodiment," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with an embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0118] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

Claims

1. A product sampling method, characterized in that, The method includes: Obtain the current process performance index (PPK) value corresponding to the target quality parameters of the target product in the current production process; Determine the statistical risk parameters associated with the target quality parameters; Based on a pre-built mapping table, the target sampling size that matches the current PPK value and the statistical risk parameter is queried; the mapping table contains a pre-defined negative correlation between the PPK value and the sampling size under the constraints of different combinations of statistical risk parameters. Output the target sampling size to indicate that sampling inspection should be performed on the target quality parameter of the target product.

2. The method according to claim 1, characterized in that, The mapping table is constructed in the following way: For each of the multiple preset PPK values, the process standard deviation, which is inversely proportional to the preset PPK value, is calculated based on the preset PPK value and the standardized specification tolerance of the target quality parameter. For each of the multiple preset statistical risk parameter combinations, the statistical risk parameter combination and the equivalent value of the process standard deviation which is inversely proportional to the preset PPK value are input into the sample size calculation model to calculate the sampling sample size corresponding to the preset PPK value under the constraint of the statistical risk parameter combination. Based on the preset PPK value, the combination of statistical risk parameters, and the calculated sampling size, the mapping relationship table is constructed and stored.

3. The method according to claim 2, characterized in that, For each of the multiple preset PPK values, the process standard deviation, which is inversely proportional to the preset PPK value, is calculated based on the preset PPK value and the standardized specification tolerance of the target quality parameter. This includes: For each of the multiple preset PPK values, the specification tolerance of the target quality parameter after standardization is divided by the product of the preset PPK value and the preset multiplier factor to obtain the process standard deviation and other values. The preset multiplier factor is used to characterize the range of process fluctuation distribution on which the process capability evaluation is based. The converted value of the process standard deviation, obtained by converting it through the specification tolerance of the target quality parameter, satisfies the following relationship with the actual process standard deviation: When the process center of production coincides with the specification center, the conversion value is equal to the actual standard deviation of the process. When the process center and specification center of the production process are inconsistent, the conversion value is greater than the actual standard deviation of the process, so that the calculated sample size is increased accordingly.

4. The method according to claim 2, characterized in that, The sample size calculation model includes a sample size calculation formula based on a normal distribution and a sample size calculation formula based on a t-distribution; the step of inputting the combination of statistical risk parameters and the estimated process standard deviation, which is inversely proportional to the preset PPK value, into the sample size calculation model to calculate the sampling sample size corresponding to the preset PPK value under the constraints of the combination of statistical risk parameters includes: Substituting the combination of statistical risk parameters and the estimated process standard deviation into the sample size calculation formula based on the normal distribution, we obtain the initial sample size estimate. Using the initial sample size estimate as the current sample size, iteratively execute the following steps until the sample size estimates obtained from two adjacent iterations are the same after rounding: determine the degrees of freedom of the t-distribution based on the current sample size, substitute the combination of statistical risk parameters and the estimated process standard deviation into the sample size calculation formula based on the t-distribution to obtain a new sample size estimate, and update the current sample size with the new sample size estimate; The rounded sample size estimate is used as the sampling sample size corresponding to the combination of statistical risk parameters and the preset PPK value.

5. The method according to claim 1, characterized in that, The process of obtaining the current process performance index (PPK) value corresponding to the target quality parameters of the target product in the current production process includes: Statistical analysis is performed on the historical measurement data of the target quality parameter in the current production process to calculate the current PPK value; Alternatively, a PPK recommended value can be generated based on the product characteristic parameters input by the user and confirmed by the user, and then the PPK recommended value can be determined as the current PPK value. Alternatively, the PPK setting value can be received from an external system or input by the user as the current PPK value.

6. The method according to claim 1, characterized in that, The current PPK value is calculated based on historical measurement data of the target quality parameter in the current production process, and the method further includes: The current PPK value is periodically calculated and updated to obtain the updated PPK value; When the updated PPK value changes effectively relative to the PPK value obtained in the previous calculation period, the query step of the target sampling sample size is triggered to update the target sampling sample size.

7. The method according to claim 6, characterized in that, The updated PPK value has a valid change relative to the PPK value obtained in the previous calculation period, including at least one of the following: The fluctuation range of the updated PPK value relative to the PPK value obtained in the previous calculation period exceeds the preset change threshold. The updated PPK value exceeds the preset process capability threshold; The updated PPK value shows a continuous decrease or increase over multiple consecutive calculation cycles, and the cumulative decrease or increase exceeds a preset threshold.

8. The method according to any one of claims 1 to 7, characterized in that, The determination of the statistical risk parameter associated with the target quality parameter includes: The system receives product characteristic levels input by the user through a visual configuration interface. Based on a preset correspondence between characteristic levels and risk parameters, it matches and obtains statistical risk parameters corresponding to the product characteristic levels, which are then used as statistical risk parameters associated with the target quality parameters; or, Receive statistical risk parameters associated with the target quality parameters input by the user.

9. A product sampling device, characterized in that, The device includes: The acquisition module is used to acquire the current process performance index (PPK) value corresponding to the target quality parameters of the target product in the current production process. A determination module is used to determine the statistical risk parameters associated with the target quality parameters; The query module is used to query the target sampling size that matches the current PPK value and the statistical risk parameter based on a pre-built mapping relationship table; the mapping relationship table contains a preset negative correlation between the PPK value and the sampling size under the constraints of different combinations of statistical risk parameters. The output module is used to output the target sampling sample size to indicate that sampling inspection should be performed on the target quality parameter of the target product.

10. A storage medium, characterized in that, It stores a computer program that, when executed by a processor, implements the steps of the product sampling method as described in any one of claims 1 to 8.

Images