A quality monitoring method and device, electronic equipment and storage medium

Abstract

The application provides a quality monitoring method and device, electronic equipment and storage medium, belonging to the technical field of quality monitoring. The method comprises the following steps: obtaining a detection equipment error of a target detection item, wherein the detection equipment error is used to represent a measurement fluctuation introduced by a self factor of a detection equipment of the target detection item; obtaining a plurality of material detection results of the target detection item, and obtaining an observation total fluctuation based on the plurality of material detection results; obtaining a material detection true fluctuation based on the detection equipment error and the observation total fluctuation; evaluating a true process capability of the target detection item according to the material detection true fluctuation and a material specification limit; and evaluating a material quality state corresponding to the target detection item based on the true process capability. The application separates the detection equipment error and the material detection true fluctuation from the observation total fluctuation, and evaluates the true process capability and the material quality state based on the material detection true fluctuation, thereby avoiding the interference of the detection equipment error on the quality evaluation.

Description

Technical Field

[0001] This application relates to the field of quality monitoring technology, and in particular to a quality monitoring method, device, electronic device and storage medium. Background Technology

[0002] In industrialized production processes, such as those in the food, pharmaceutical, chemical, and electronics industries, quality monitoring of materials and products is typically required to ensure product quality. A common quality monitoring method involves using rapid testing equipment or methods to sample and test materials to ensure their quality characteristics meet pre-defined specifications.

[0003] To ensure the reliability of test results, existing technologies typically perform periodic calibration of rapid testing equipment. This involves using standard samples for comparison to obtain accuracy or repeatability indicators, and then calibrating based on these indicators. After calibration, the direct test results from the rapid testing equipment are compared with preset product standard limits to determine whether the product is qualified.

[0004] However, existing quality inspection methods ignore the influence of the inspection equipment itself on the inspection results. The process capability index obtained based on such inspection results cannot reflect the true stability and quality level of the production process. Summary of the Invention

[0005] This application provides a quality monitoring method, apparatus, electronic device, and storage medium, aiming to solve the problem that existing quality inspection methods ignore the influence of the inspection equipment itself on the inspection results, resulting in the process capability index obtained based on such inspection results failing to reflect the true stability and quality level of the production process.

[0006] Firstly, this application provides a quality monitoring method, comprising: The detection equipment error of the target detection item is obtained, and the detection equipment error is used to characterize the measurement fluctuation introduced by the detection equipment itself of the target detection item; Obtain multiple material detection results for the target detection item, and obtain the total observed fluctuation based on the multiple material detection results; The true fluctuation of material detection is obtained based on the error of the detection equipment and the total observed fluctuation. The actual process capability of the target testing item is evaluated based on the actual fluctuations in material testing and the material specification limits. The material quality status corresponding to the target detection item is evaluated based on the actual process capability.

[0007] As one embodiment, the step of obtaining the detection equipment error of the target detection item includes: The target reference method and target detection method of the target detection project are used to perform pairwise detection on the materials within the target period to obtain multiple pairs of pairwise detection results, and the accuracy deviation is obtained based on the pairwise detection results; The target detection items are repeatedly tested on the material within the target period to obtain multiple repeated detection results, and repeatability deviation is determined based on the repeated detection results; The material within the target period is tested using multiple testing devices corresponding to the target testing item, and the inter-device deviation is determined based on the test results of different testing devices for the same material. The accuracy deviation, repeatability deviation, and inter-machine deviation are combined to obtain the error of the detection equipment.

[0008] As one embodiment, obtaining the observed total fluctuation based on the multiple material detection results includes: Batch fluctuations are determined based on the material test results corresponding to multiple materials within the same batch. Inter-batch fluctuations are determined based on the test results of multiple batches of materials; The total observed fluctuation is obtained by combining the intra-batch fluctuation and the inter-batch fluctuation.

[0009] As one embodiment, the evaluation of the material quality status corresponding to the target inspection item based on the actual process capability includes: Predict the target material qualification rate based on the actual process capability; The target sampling representativeness is determined based on the target comprehensive risk benchmark value and the target material qualification rate; the target comprehensive risk benchmark value is the product of the target material qualification rate and the target sampling representativeness. The material sampling plan for the target detection item is determined based on the representativeness of the target sampling.

[0010] As one example, the quality monitoring method further includes: Obtain the actual material qualification rate based on the material sampling plan; When the actual material qualification rate is lower than the target material qualification rate, the target comprehensive risk benchmark value is adjusted, and the target sampling representativeness and the material sampling plan are adjusted based on the adjusted target comprehensive risk benchmark value and the target material qualification rate.

[0011] As one embodiment, obtaining the true fluctuation of material detection based on the error of the detection equipment and the total observed fluctuation includes: Determine the difference between the square of the total observed fluctuation and the square of the error of the detection equipment; The square root of the difference is performed to obtain the true fluctuation of the material detection.

[0012] As one example, the quality monitoring method further includes: The trend of target deviation changes within multiple consecutive target periods is statistically analyzed, wherein the target deviation includes at least one of the accuracy deviation, the repeatability deviation, and the inter-station deviation; Based on the aforementioned trend of change, predict the value of the target deviation within the target time period; When the predicted value exceeds the maintenance threshold of the target deviation, a predictive maintenance work order is generated.

[0013] Secondly, this application also provides a quality monitoring device, comprising: The equipment error acquisition module is used to acquire the detection equipment error of the target detection item. The detection equipment error is used to characterize the measurement fluctuation introduced by the detection equipment itself of the target detection item. The total fluctuation observation acquisition module is used to acquire multiple material detection results of the target detection item and obtain the total fluctuation observation based on the multiple material detection results. The real fluctuation acquisition module is used to obtain the real fluctuation of material detection based on the error of the detection equipment and the total observed fluctuation. The real process capability assessment module is used to assess the real process capability of the target testing item based on the real fluctuation of the material testing and the material specification limit. The quality status assessment module is used to assess the material quality status corresponding to the target detection item based on the actual process capability.

[0014] Thirdly, this application also provides an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement any of the above-mentioned quality monitoring methods.

[0015] Fourthly, this application also provides a non-transitory computer-readable storage medium storing a computer program that, when executed by a processor, implements any of the aforementioned quality monitoring methods.

[0016] Fifthly, this application also provides a computer program product, which includes a computer program that, when executed by a processor, implements any of the above-described quality monitoring methods.

[0017] This application provides a quality monitoring method, apparatus, electronic device, and storage medium. By separating the detection equipment error and the actual fluctuation of material detection from the total observed fluctuation, and evaluating the actual process capability and material quality status based on the actual fluctuation of material detection, the method avoids the interference of detection equipment error on quality assessment. This allows the assessment results to more accurately reflect the true quality level of the material, preventing the risk of misjudging a stable process as abnormal and incurring unnecessary improvement costs, or misjudging an unstable process as normal and causing non-conforming products to flow out. This significantly improves the accuracy and reliability of quality evaluation. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is one of the flowcharts illustrating the quality monitoring method provided in this application; Figure 2 This is the second flowchart of the quality monitoring method provided in this application; Figure 3 This is a schematic diagram of the quality monitoring device provided in this application; Figure 4 This is a schematic diagram of the structure of the electronic device provided in this application. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0021] It should be noted that, in the description of this application, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0022] The terms "first," "second," etc., used in this application are used to distinguish similar objects and not to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein, and the objects distinguished by "first," "second," etc., are generally of the same class, without limiting the number of objects; for example, a first object can be one or more. Furthermore, "and / or" indicates at least one of the connected objects, and the character " / " generally indicates that the preceding and following objects have an "or" relationship.

[0023] In industrialized production processes, such as those in the food, pharmaceutical, chemical, and electronics industries, quality monitoring of materials and products is typically required to ensure product quality. A common quality monitoring method involves using rapid testing equipment or methods to sample and test materials to ensure their quality characteristics meet pre-defined specifications.

[0024] To ensure the reliability of test results, existing technologies typically perform periodic calibration of rapid testing equipment. This involves using standard samples for comparison to obtain accuracy or repeatability indicators, and then calibrating based on these indicators. After calibration, the direct test results from the rapid testing equipment are compared with preset product standard limits to determine whether the product is qualified.

[0025] However, existing quality inspection methods neglect the impact of the inspection equipment itself on the test results. The process capability index obtained based on such test results cannot reflect the true stability and quality level of the production process. If the inspection error is large, the value of the process capability index will be lowered, causing the company to invest unnecessary quality costs; conversely, if the inspection error is negative, it may mask the true risk of nonconformity.

[0026] The quality monitoring methods, devices, electronic equipment, and storage media provided in this application are described below with reference to the accompanying drawings.

[0027] It should be noted that the quality monitoring method provided in this application embodiment is based on a quality monitoring device. The quality monitoring method achieves a complete closed loop from equipment calibration to final monitoring strategy adjustment through a full-chain process of error identification, fluctuation decomposition, real capability assessment, dynamic sampling, and review and closure. This solves problems such as inaccurate on-site testing, inaccurate sampling, and unreasonable adjustment in industrial settings.

[0028] This application describes a quality monitoring method using a quality monitoring device as the executing entity as an example.

[0029] Figure 1 This is one of the flowcharts illustrating the quality control method provided in this application. For example... Figure 1 As shown, the quality monitoring method provided in this application includes: S110: Obtain the detection equipment error of the target detection item, wherein the detection equipment error is used to characterize the measurement fluctuation introduced by the detection equipment itself of the target detection item.

[0030] The target detection item can be any quality indicator that needs to be monitored during industrial production. For example, in pharmaceutical production, the target detection item could be the concentration of the main ingredient in a tablet; in food production, it could be the content of a certain additive. Detection equipment error is a comprehensive parameter used to quantitatively describe the uncertainty of a detection system. It reflects the degree to which the detection result deviates from the true value of the material. This deviation is caused by systematic factors such as the detection equipment itself, the operating environment, and the operators, rather than by fluctuations in the quality of the material itself.

[0031] There are several ways to obtain the error of the testing equipment. For example, the testing equipment error can be a system parameter calculated in advance through a series of calibration experiments within a preset calibration period (such as one week or one month). At the start of the quality monitoring process, the quality monitoring system can directly retrieve or read the testing equipment error value from an internal database, configuration file, or external calibration certificate. In other embodiments, a calibration process can also be triggered in real time to dynamically calculate the testing equipment error.

[0032] S120: Obtain multiple material detection results for the target detection item, and obtain the total observed fluctuation based on the multiple material detection results.

[0033] The material testing results are direct readings obtained by testing material samples on the production line using testing equipment (including rapid testing equipment) for the target testing items. For example, for 10 batches of tablets produced continuously, several samples are taken from each batch at preset time intervals or in preset quantities, and their principal component concentrations are measured to obtain a dataset containing hundreds of data points, which are the material testing results.

[0034] The dispersion of the multiple material detection results obtained above is quantified to obtain the total observed fluctuation. For example, the sample standard deviation of all obtained material detection results is calculated as the total observed fluctuation.

[0035] The observed total fluctuation reflects the overall fluctuation of all test data over a period of time, which includes both the actual quality fluctuation of the material itself and the measurement fluctuation caused by the error of the test equipment.

[0036] S130: Obtain the true fluctuation of material detection based on the error of the detection equipment and the total observed fluctuation.

[0037] After obtaining the detection equipment error, representing the uncertainty of the detection system, and the total observed fluctuation, representing mixed fluctuations, the influence of the detection equipment error is removed from the total observed fluctuations. This yields an index that purely reflects the variation in the material's own quality, namely, the true fluctuation of the material detection. Figure 2 As shown.

[0038] Subsequent quality assessments based on the actual fluctuations in material testing allow for assessments to be conducted based on real fluctuation data that is closer to reality, thus eliminating interference from testing equipment errors.

[0039] S140: Evaluate the actual process capability of the target testing item based on the actual fluctuations in material testing and the material specification limits.

[0040] The material specification limits are typically predefined by product standards or process requirements, including upper specification limit (USL) and lower specification limit (LSL), which define the acceptable range of material quality characteristics.

[0041] Real process capability is the ability to compare the actual fluctuations of materials with the upper limit of specifications to determine the production process's ability to meet specifications.

[0042] In one specific implementation, the actual process capability is characterized by the Process Performance Index (Ppk), calculated using the following formula: (1) in, This is the average of the test results for all materials. This is to detect the actual fluctuations in material properties.

[0043] A higher Ppk value indicates a stronger actual production process capability, a narrower distribution width of material characteristics relative to the specification range, and a higher quality level.

[0044] S150: Evaluate the material quality status corresponding to the target detection item based on the actual process capability.

[0045] Specifically, the actual process capability can be compared with industry-recognized or internally set benchmarks to determine the material quality status.

[0046] In one possible implementation, if Ppk ≥ 1.33, the current material quality is considered to be good and stable; if Ppk < 1.33, the current material quality is considered to be at risk and process adjustments may be necessary.

[0047] In another possible implementation, the product qualification rate is further predicted based on the actual process capability, and the product qualification rate is used as the final evaluation conclusion.

[0048] This application embodiment separates the detection equipment error and the actual fluctuation of material detection from the observed total fluctuation, and evaluates the actual process capability and material quality status based on the actual fluctuation of material detection. This avoids the interference of detection equipment error on quality assessment, so that the assessment results can more accurately reflect the actual quality level of the material. It prevents the risk of misjudging a stable process as abnormal and investing unnecessary improvement costs, or misjudging an unstable process as normal and causing non-conforming products to flow out, thereby significantly improving the accuracy and reliability of quality evaluation.

[0049] In one possible implementation, such as Figure 2 As shown, in S110, the step of obtaining the detection equipment error of the target detection item includes: S1110: The target reference method and target detection method of the target detection project are used to perform pairwise detection on the materials within the target period to obtain multiple pairs of detection results, so as to obtain the accuracy deviation based on the pairwise detection results.

[0050] Accuracy deviation characterizes the bias introduced by the detection method. In one possible implementation, the specific steps for obtaining the accuracy deviation are as follows: First, within a preset target period, such as a calibration cycle, prepare one or more sets of high-precision standards measured using a target reference method (e.g., national standard method or industry gold standard method). For comprehensive evaluation, this set of standards covers the range of routine measurements of the testing equipment. For example, a "five-point method" can be used, with each set of standards including five standards evenly distributed across the low, medium, and high ranges of measurement.

[0051] Subsequently, a target reference method and a target detection method requiring calibration (such as a rapid detection method) are used simultaneously to perform paired tests on one or more sets of standards. In this way, a reference value and a measured value are obtained for each standard.

[0052] Finally, statistical analysis is performed on the five pairs of test results in each group to obtain the accuracy bias. The average of the differences in each group can be calculated as the accuracy bias. Alternatively, Deming regression or Passing-Bablok regression analysis can be used to obtain the accuracy bias. These two regression methods consider the possibility of errors in both measurement methods, and the intercept and slope of the obtained regression equation can more robustly reflect the systematic and proportional differences between the two methods, thus obtaining an accuracy bias function or fixed value that is representative across the entire measurement range.

[0053] As an example, five bottles of standard solutions with known concentrations are selected. These standard solutions uniformly cover the actual result range of everyday solutions. Five measurements are obtained using a rapid detection device. The differences between these five measurements and the true values ​​are then calculated, and the average of these five differences is the accuracy deviation.

[0054] S1120: Repeatedly perform the target detection items on the material within the target period multiple times to obtain multiple repeated detection results, and determine the repeatability deviation based on the repeated detection results.

[0055] Repeatability bias characterizes the degree of consistency among measurement results when the same material is measured repeatedly under identical conditions. In one possible implementation, the specific steps for obtaining repeatability bias are as follows: Select one or more stable materials (which can be standards or homogeneous production samples) and, within the shortest possible time, have the same operator, using the same testing equipment, and following the exact same testing procedures, perform multiple measurements consecutively, such as five or ten times.

[0056] The repeated test results are then statistically analyzed to determine repeatability bias. Specifically, the standard deviation, relative standard deviation (RSD), or range of the data can be calculated, and these values ​​can serve as quantitative indicators of repeatability bias.

[0057] As an example, for the five standard solutions of known concentrations mentioned above, each bottle is measured five times using a rapid detection device. The standard deviation, relative standard deviation, or range of these five measurements is the repeatability deviation for that concentration. Finally, the repeatability deviations of these five concentrations are combined to determine an overall repeatability deviation, such as the average of the five ranges.

[0058] S1130: Use multiple detection devices corresponding to the target detection item to detect the material within the target period, and determine the inter-device deviation based on the detection results of different detection devices for the same material.

[0059] When multiple testing devices of the same model or function exist in a production environment, even if they have been calibrated separately, inherent systematic differences may still exist between them, known as inter-device bias. The specific steps for obtaining inter-device bias are as follows: Prepare homogeneous and stable samples, and repeatedly measure the same sample on multiple testing devices (e.g., 2 or 3) within the same laboratory. Then calculate the mean measurement of the same sample on each device, and calculate the difference between the means of different devices as the inter-device deviation. For example, when comparing two devices, the difference between their means can be used as a measure of inter-device deviation.

[0060] To increase reliability, multiple samples from different batches can be used to measure the individual inter-sample difference, and the average of these multiple inter-sample differences can be taken as the final inter-sample deviation. For example, the average of the individual inter-sample differences of the five solutions mentioned above can be taken as the inter-sample deviation.

[0061] In another possible implementation, a round-robin test scheme can be adopted, in which multiple blind samples are sent to multiple laboratories or operators using different instruments to perform the same target test, and then the test results of the blind samples are statistically analyzed in a centralized manner to obtain a more objective and universally applicable inter-sample bias.

[0062] By incorporating inter-device deviations into the error of the testing equipment, it helps to ensure the consistency of judgment results in scenarios where multiple devices are used in parallel. It can also guide companies to adjust their equipment calibration, elimination, or allocation strategies (such as using high-precision equipment for critical projects).

[0063] S1140: The accuracy deviation, repeatability deviation, and inter-machine deviation are combined to obtain the detection equipment error.

[0064] According to the law of error propagation, if the error sources are independent of each other, the variance of the total error is equal to the sum of the variances of the individual errors. Therefore, the accuracy deviation, repeatability deviation, and inter-machine deviation are first squared, then these variance values ​​are added together and the square root is taken to obtain the error of the testing equipment (expressed in the form of standard deviation).

[0065] This application embodiment calculates the accuracy deviation, repeatability deviation, and inter-device deviation of the testing equipment separately, and then fuses these deviations from different sources to obtain the testing equipment error. This significantly improves the scientific nature and reliability of the testing equipment error and has high practical value in industrial scenarios where multiple devices are used together.

[0066] In one possible implementation, such as Figure 2 As shown, in S120, obtaining the observed total fluctuation based on the multiple material detection results includes: S1210: Determine batch fluctuations based on the material testing results corresponding to multiple materials within the same batch.

[0067] The intra-batch fluctuations reflect the degree of non-uniformity in material quality characteristics within the same production batch, packaging machine stability, and other factors within a single batch. For example, in a powder mixing process, intra-batch fluctuations represent the uniformity of mixing.

[0068] In one possible implementation, the specific steps for determining intra-batch fluctuations are as follows: First, determine a representative sampling plan. For example, using a complete production batch as the unit, calculate its average production time T and the number of packages N. Then, based on historical data or relevant standards, determine the minimum sample size n required to achieve sufficient sample representativeness (e.g., 95%). From this, calculate the sampling time interval t = T / n.

[0069] In actual production, during the production of a single batch, sampling and testing are carried out at fixed time intervals t to obtain the test results of n materials within that batch.

[0070] Finally, calculate the standard deviation or variance of these n material test results as the intra-batch fluctuation of that batch.

[0071] In one possible implementation, control charts in Statistical Process Control (SPC), such as the mean-range chart (Xbar-R chart) or the individual-moving range chart (I-MR chart), can be used. Control charts can not only calculate the quantitative value of the fluctuation within a batch (such as the mean range R-bar), but also visually show whether the material quality within the batch is stable, whether there are outliers or special trends, providing richer information for process monitoring.

[0072] S1220: Determine batch-to-batch fluctuations based on the results of material testing from multiple batches.

[0073] Inter-batch fluctuations reflect changes or drifts in the average quality characteristics of materials across different production batches. Inter-batch fluctuations are typically related to factors such as changes in raw material batches, variations in environmental conditions, and long-term equipment wear and tear.

[0074] In one possible implementation, the specific steps for determining inter-batch fluctuations are as follows: The material testing results of multiple (e.g., 10) production batches are collected continuously, with each batch having a set of testing data.

[0075] Then, the average value of the test data for each batch was calculated.

[0076] Finally, the standard deviation of the average values ​​of multiple batches is used as the inter-batch fluctuation.

[0077] S1230: Combine the intra-batch fluctuations and inter-batch fluctuations to obtain the total observed fluctuations.

[0078] Specifically, the intra-batch fluctuation and inter-batch fluctuation are squared to obtain their respective variances, and then the two are added together to obtain the total variance; finally, the total variance is squared to obtain the observed total fluctuation (expressed in the form of standard deviation).

[0079] This application's embodiments determine the total observed fluctuation by fusing detection fluctuations from both intra-batch and inter-batch dimensions, ensuring that the total observed fluctuation can comprehensively and accurately reflect the overall measurement error. Furthermore, this decomposition mechanism helps to accurately pinpoint whether the variation originates from within or between batches, improving diagnostic accuracy.

[0080] In one possible implementation, S130 involves obtaining the true fluctuation of material detection based on the detection equipment error and the total observed fluctuation, including: S1310: Determine the difference between the square of the observed total fluctuation and the square of the error of the detection equipment.

[0081] S1320: Perform a square root operation on the difference to obtain the true fluctuation of the material detection.

[0082] The variance of the observed total fluctuation can be considered as a linear superposition of the variance of the actual fluctuation of the material and the variance of the detection equipment error. Therefore, by squaring the observed total fluctuation to obtain the variance of the observed total fluctuation, and squaring the detection equipment error to obtain the variance of the detection equipment error, the square root of the difference between the two variances can be obtained to obtain the actual fluctuation of the material detection. It is an indicator that is expressed in standard deviation, has eliminated measurement system noise, and purely reflects the dispersion of the material itself.

[0083] In practical calculations, due to sampling errors, the variance of the total observed fluctuation may be less than the variance of the detection equipment error. This results in a negative difference, making square root calculation impossible. In engineering practice, this means that the fluctuation of the detection system completely overshadows the fluctuation of the material itself, making measurement error the primary source of variation.

[0084] Based on this consideration, this embodiment also includes an exception handling mechanism: when the difference between the two variances is negative, the actual fluctuation of the material detection is directly set to zero, and a warning flag is generated in the system log at the same time, indicating that the measurement system is the dominant fluctuation, prompting managers to prioritize improving or replacing the detection equipment, rather than adjusting the production process.

[0085] This application embodiment uses the principle of linear superposition of variances to mathematically and accurately decompose the actual material fluctuations from the observed total fluctuations. It also uses the difference in variances to determine the contribution of the detection equipment error, thereby achieving anomaly handling and correctly dealing with extreme situations that may occur in actual production, thus increasing the robustness of engineering applications.

[0086] In one possible implementation, S150, the evaluation of the material quality status corresponding to the target detection item based on the actual process capability includes: S1510: Predict the target material qualification rate based on the actual process capability.

[0087] The quality monitoring system internally stores or links to a statistical model that describes the mapping relationship between the actual process capability index Ppk and the theoretical material yield rate under the assumption of normal distribution. Table 1 provides an example of a mapping table between the actual process capability index Ppk and the theoretical material yield rate. Table 1

[0088] Using the calculated actual process capability index Ppk as input, the predicted target material qualification rate can be obtained by querying a mapping table. For example, if the actual Ppk is 1.33, the predicted target material qualification rate A is 99.99%. This target material qualification rate represents the theoretically achievable quality level under the current actual process capability.

[0089] S1520: Determine the target sampling representativeness based on the target comprehensive risk benchmark value and the target material qualification rate. The target comprehensive risk benchmark value is the product of the target material qualification rate and the target sampling representativeness.

[0090] The target sampling representativeness B is used to quantify the sampling plan's ability to identify problems. Higher representativeness means a more rigorous sampling plan and a lower risk of missed detections. The target comprehensive risk benchmark C is the company's pre-set, acceptable comprehensive risk level. The target comprehensive risk benchmark C is a strategic parameter; a higher C value indicates a higher overall quality assurance level required by the company. For example, a company can set C=0.98 for critical control points and C=0.95 for general monitoring points.

[0091] The relationship between the target comprehensive risk benchmark C, the target material qualification rate A, and the target sampling representativeness B is as follows: C = A × B (2).

[0092] This step performs reverse calculation: given the target comprehensive risk benchmark value C and the target material qualification rate A, the target sampling representativeness B that needs to be achieved can be determined by division, i.e., B=C / A.

[0093] S1530: Determine the material sampling plan for the target detection item based on the representativeness of the target sampling.

[0094] The material sampling plan includes the sampling quantity n, sampling frequency, etc. The quality monitoring system has a built-in conversion model between sample representativeness and the sampling plan. This application's embodiment uses an existing conversion model; for example, the conversion model can be based on a specific standard or other sampling theory (such as the Dodge-Romeogue sampling table). The conversion model can also be a dynamic calculation model based on binomial distribution or hypergeometric distribution theory. The system inputs the batch size N and the target sample representativeness B, and the model can calculate the required minimum sample size n in real time.

[0095] For example, if the calculated target sampling representativeness B is 90%, the system queries the conversion model and determines that 80 samples are needed for the current batch. If subsequent process improvements significantly increase the pass rate A, reducing the calculated target sampling representativeness B to only 70%, the system queries the model again and may only need to sample 30 samples. Ultimately, the system outputs the determined sampling quantity n, sampling interval t, and other information to the human-machine interface or automated sampling equipment to guide actual operations. This achieves intelligent monitoring that is both efficient and rigorous where necessary.

[0096] This application embodiment dynamically determines the target material qualification rate based on the actual process capability, and determines the target sampling representativeness based on the preset target comprehensive risk benchmark value and the target material qualification rate, so as to further determine the material sampling plan, thereby dynamically linking the actual process capability with the rigor of monitoring: when the process capability is strong and the inherent risk is low, sampling is automatically relaxed to save testing costs; when the process capability weakens and the inherent risk increases, sampling is automatically tightened to ensure quality and safety. Compared with the fixed sampling plan of the prior art, this application embodiment realizes the optimized allocation of monitoring resources, so that quality monitoring maximizes cost-effectiveness under the premise of controllable risk.

[0097] Existing quality monitoring risk assessments are mostly static qualitative scoring, which cannot quantify whether the current monitoring strategy (such as sampling plan) matches the actual risk.

[0098] Based on this consideration, in one possible implementation, the assessment of the material quality status corresponding to the target detection item based on the actual process capability also includes a review operation of the target comprehensive risk benchmark value, specifically including: S1540: Obtain the actual material qualification rate based on the material sampling plan.

[0099] After the material sampling plan has been implemented for a period of time (e.g., one day, one week, or one production cycle), all actual sampling results during this period are collected. Based on these actual test data, the proportion of qualified samples to the total number of samples is calculated, thereby determining the actual material qualification rate.

[0100] S1550: When the actual material qualification rate is lower than the target material qualification rate, adjust the target comprehensive risk benchmark value, and adjust the target sampling representativeness and the material sampling plan based on the adjusted target comprehensive risk benchmark value and the target material qualification rate.

[0101] The actual material qualification rate is compared with the theoretically achievable target material qualification rate A.

[0102] When the actual material qualification rate is lower than the target material qualification rate, and this difference persists or exceeds a certain preset statistical threshold, it indicates that the real non-conformance risk is greater than predicted by the theoretical model. This means that a new source of variation has emerged that the model has not considered, or that the originally set risk tolerance (i.e., the target comprehensive risk benchmark value C) is too high and is no longer suitable for the current production reality.

[0103] At this point, an adjustment mechanism will be automatically or semi-automatically triggered to appropriately reduce the target comprehensive risk benchmark value C based on the difference between the target material qualification rate and the actual material qualification rate. For example, using PI control: Proportional (P) adjustment: The larger the difference, the greater the adjustment to the C value; Integral (I) adjustment: If the difference persists (even if it is small), the integral term will accumulate over time, eventually triggering an adjustment to the C value to eliminate steady-state error. For example, the original C=0.95 might be adjusted to C=0.93. Reducing the C value is equivalent to increasing the system's risk aversion and decreasing its tolerance for risk.

[0104] Once the target comprehensive risk benchmark value C is reduced, this new C value will immediately take effect on formula (2). Even if the target material qualification rate A and the actual process capability remain unchanged, a lower C value will result in a higher target sampling representativeness B, thereby determining a more stringent material sampling plan based on the transformation model (e.g., increasing the sampling quantity from 30 to 50), thus achieving dynamic adaptive optimization of the sampling plan based on the nonconformity risk.

[0105] This more stringent sampling plan will continue until the new actual material qualification rate consistently equals or exceeds the target material qualification rate. At this point, it indicates that the new risk control strategy is effective, and the adjusted target comprehensive risk benchmark value C can be solidified and applied continuously until the next deviation occurs. The solidified target comprehensive risk benchmark value C accurately reflects the company's risk tolerance for this type of material, and is auditable and traceable, meeting the requirements of the quality management system for risk thinking and continuous improvement.

[0106] This application embodiment dynamically adjusts the target comprehensive risk benchmark value based on the actual material qualification rate, thereby adjusting the target sampling representativeness and the material sampling plan. When the theoretical prediction does not match the objective reality, the unknown risks are dealt with by tightening monitoring until a stable and predictable state is reached again. In this way, a self-learning and continuously improving intelligent quality system is built through the feedback loop, which improves the resilience and adaptability to cope with complex and dynamic production environments.

[0107] In one possible implementation, the quality monitoring method also includes: Predictive maintenance information is generated based on the changing trend of the target deviation. Specifically, this includes: P10: Statistically analyze the changing trend of the target deviation within multiple consecutive target periods, wherein the target deviation includes at least one of the accuracy deviation, the repeatability deviation, and the inter-station deviation.

[0108] The quality monitoring system not only uses the target deviations (including accuracy deviation, repeatability deviation, and inter-station deviation) calculated for each target period (e.g., weekly) for the current quality assessment, but also stores them independently in the historical database along with the corresponding period timestamps to form a time series dataset.

[0109] By statistically analyzing the dynamic information of the target deviation over time, we can obtain the corresponding trend of change, such as stabilizing at a certain level, linear growth, exponential deterioration, or exhibiting some kind of periodic fluctuation.

[0110] P20: Based on the stated trend, predict the target deviation within the target time period.

[0111] Specifically, the built-in trend prediction model is invoked, and the historical time series data obtained in the previous step is used as input to predict the value (i.e., the predicted value) that the target deviation may reach in a future target time period (e.g., within the next four weeks) based on the changing trend.

[0112] The trend prediction model can be implemented in various ways to adapt to changing trends of varying complexity, for example: When the trend of target deviation shows a clear linear increase or decrease, a linear regression model can be used. The system fits a straight line that best represents the trend of historical data using the least squares method, and extends along this line into the future to obtain the predicted value.

[0113] When there are certain random fluctuations in the biased data, a moving average or exponential smoothing model can be used to smooth out noise and predict short-term trends by giving higher weight to recent data.

[0114] The Autoregressive Integrated Moving Average (ARIMA) model can also be applied, as it captures the complex autocorrelation, trends, and seasonality inherent in the data. By differencing historical data to make it stable and analyzing its autocorrelation and partial autocorrelation functions, the model order is determined, thereby establishing a predictive model that more accurately reflects the inherent dynamics of the data.

[0115] To address more complex nonlinear degradation patterns, advanced machine learning models can be employed. For example, a Long Short-Term Memory (LSTM) network or a Gated Recurrent Unit (GRU) can be trained. These deep learning models excel at learning temporal dependencies from long sequences of data, enabling more accurate modeling and prediction of complex performance degradation patterns caused by the coupling of multiple factors. When training this model, the input is a historical sequence of target biases, and the output is a prediction for one or more future time steps.

[0116] P30: When the predicted value exceeds the maintenance threshold of the target deviation, a predictive maintenance work order is generated.

[0117] The maintenance threshold is an unacceptable upper limit of deviation set according to quality standards, equipment specifications, or regulatory requirements. For example, for repeatability deviation, the maintenance threshold might be set to 1.0. Once the repeatability deviation of the equipment reaches 1.0, the reliability of its test results will be considered unacceptable for production, and action must be taken.

[0118] When the system determines that a predicted value (e.g., the predicted repeatability deviation for week N+4 will reach 1.05) exceeds the maintenance threshold (1.0), it will automatically generate a predictive maintenance work order. This work order is a structured message designed to clearly communicate the maintenance task. A detailed predictive maintenance work order may include: a unique equipment identifier, the target deviation item that triggers the warning (e.g., repeatability deviation), the current target deviation, the predicted value and maintenance threshold, the predicted time window for exceeding the threshold, historical data trend charts (as a basis for judgment), and a list of suggested maintenance operations (e.g., cleaning optical components, replacing reagent batches, contacting the supplier for in-depth calibration, etc.).

[0119] Predictive maintenance work orders can be delivered in various ways, such as being automatically pushed to the enterprise's computerized maintenance management system via API, or being sent to a designated team of maintenance engineers via email or SMS.

[0120] This application's embodiments predict the time window when the target deviation exceeds a threshold by analyzing the changing trend of the target deviation, thereby issuing a risk warning in advance. This transforms the quality monitoring system from post-analysis and passive response into an intelligent management platform for proactive problem prevention, helping enterprises to plan maintenance activities in advance and significantly reducing unplanned production downtime caused by sudden failure of testing equipment; at the same time, it optimizes the allocation of maintenance resources and reduces operating costs.

[0121] Based on the above, this application also provides a quality monitoring device. The quality monitoring device and the aforementioned quality monitoring method can be referred to in correspondence.

[0122] As an example, such as Figure 3 As shown, the quality monitoring device includes: The equipment error acquisition module 310 is used to acquire the detection equipment error of the target detection item, wherein the detection equipment error is used to characterize the measurement fluctuation introduced by the detection equipment itself of the target detection item; The total fluctuation observation acquisition module 320 is used to acquire multiple material detection results of the target detection item and obtain the total fluctuation observation based on the multiple material detection results. The real fluctuation acquisition module 330 is used to obtain the real fluctuation of material detection based on the error of the detection equipment and the total observed fluctuation. The real process capability assessment module 340 is used to assess the real process capability of the target testing item based on the real fluctuation of the material testing and the material specification limit. The quality status assessment module 350 is used to assess the material quality status corresponding to the target detection item based on the actual process capability.

[0123] This application embodiment separates the detection equipment error and the actual fluctuation of material detection from the observed total fluctuation, and evaluates the actual process capability and material quality status based on the actual fluctuation of material detection. This avoids the interference of detection equipment error on quality assessment, so that the assessment results can more accurately reflect the actual quality level of the material. It prevents the risk of misjudging a stable process as abnormal and investing unnecessary improvement costs, or misjudging an unstable process as normal and causing non-conforming products to flow out, thereby significantly improving the accuracy and reliability of quality evaluation.

[0124] Figure 4 This is a schematic diagram of the structure of the electronic device provided in this application, such as... Figure 4As shown, the electronic device may include: a processor 410, a communications interface 420, a memory 430, and a communication bus 440, wherein the processor 410, the communications interface 420, and the memory 430 communicate with each other via the communication bus 440. The processor 410 can call logical instructions in the memory 430 to execute a quality monitoring method, which includes: The detection equipment error of the target detection item is obtained, and the detection equipment error is used to characterize the measurement fluctuation introduced by the detection equipment itself of the target detection item; Obtain multiple material detection results for the target detection item, and obtain the total observed fluctuation based on the multiple material detection results; The true fluctuation of material detection is obtained based on the error of the detection equipment and the total observed fluctuation. The actual process capability of the target testing item is evaluated based on the actual fluctuations in material testing and the material specification limits. The material quality status corresponding to the target detection item is evaluated based on the actual process capability.

[0125] Furthermore, the logical instructions in the aforementioned memory 430 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0126] On the other hand, this application also provides a computer program product, which includes a computer program stored on a non-transitory computer-readable storage medium. The computer program includes program instructions, and when the program instructions are executed by a computer, the computer is able to execute the quality monitoring method provided in the above embodiments. The method includes: The detection equipment error of the target detection item is obtained, and the detection equipment error is used to characterize the measurement fluctuation introduced by the detection equipment itself of the target detection item; Obtain multiple material detection results for the target detection item, and obtain the total observed fluctuation based on the multiple material detection results; The true fluctuation of material detection is obtained based on the error of the detection equipment and the total observed fluctuation. The actual process capability of the target testing item is evaluated based on the actual fluctuations in material testing and the material specification limits. The material quality status corresponding to the target detection item is evaluated based on the actual process capability.

[0127] In another aspect, this application also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to perform the quality monitoring methods provided in the above embodiments, the method comprising: The detection equipment error of the target detection item is obtained, and the detection equipment error is used to characterize the measurement fluctuation introduced by the detection equipment itself of the target detection item; Obtain multiple material detection results for the target detection item, and obtain the total observed fluctuation based on the multiple material detection results; The true fluctuation of material detection is obtained based on the error of the detection equipment and the total observed fluctuation. The actual process capability of the target testing item is evaluated based on the actual fluctuations in material testing and the material specification limits. The material quality status corresponding to the target detection item is evaluated based on the actual process capability.

[0128] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0129] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0130] 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 of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A quality monitoring method, characterized in that, include: The detection equipment error of the target detection item is obtained, and the detection equipment error is used to characterize the measurement fluctuation introduced by the detection equipment itself of the target detection item; Obtain multiple material detection results for the target detection item, and obtain the total observed fluctuation based on the multiple material detection results; The true fluctuation of material detection is obtained based on the error of the detection equipment and the total observed fluctuation. The actual process capability of the target testing item is evaluated based on the actual fluctuations in material testing and the material specification limits. The material quality status corresponding to the target detection item is evaluated based on the actual process capability.

2. The quality monitoring method according to claim 1, characterized in that, The detection equipment error for obtaining the target detection item includes: The target reference method and target detection method of the target detection project are used to perform pairwise detection on the materials within the target period to obtain multiple pairs of pairwise detection results, and the accuracy deviation is obtained based on the pairwise detection results; The target detection items are repeatedly tested on the material within the target period to obtain multiple repeated detection results, and repeatability deviation is determined based on the repeated detection results; The material within the target period is tested using multiple testing devices corresponding to the target testing item, and the inter-device deviation is determined based on the test results of different testing devices for the same material. The accuracy deviation, repeatability deviation, and inter-machine deviation are combined to obtain the error of the detection equipment.

3. The quality monitoring method according to claim 1, characterized in that, The method of obtaining the observed total fluctuation based on the multiple material detection results includes: Batch fluctuations are determined based on the material test results corresponding to multiple materials within the same batch. Inter-batch fluctuations are determined based on the test results of multiple batches of materials; The total observed fluctuation is obtained by combining the intra-batch fluctuation and the inter-batch fluctuation.

4. The quality monitoring method according to claim 1, characterized in that, The evaluation of the material quality status corresponding to the target detection item based on the actual process capability includes: Predict the target material qualification rate based on the actual process capability; The target sampling representativeness is determined based on the target comprehensive risk benchmark value and the target material qualification rate; the target comprehensive risk benchmark value is the product of the target material qualification rate and the target sampling representativeness. The material sampling plan for the target detection item is determined based on the representativeness of the target sampling.

5. The quality monitoring method according to claim 4, characterized in that, Also includes: Obtain the actual material qualification rate based on the material sampling plan; When the actual material qualification rate is lower than the target material qualification rate, the target comprehensive risk benchmark value is adjusted, and the target sampling representativeness and the material sampling plan are adjusted based on the adjusted target comprehensive risk benchmark value and the target material qualification rate.

6. The quality monitoring method according to claim 1, characterized in that, The process of obtaining the true fluctuation of material detection based on the error of the detection equipment and the total observed fluctuation includes: Determine the difference between the square of the total observed fluctuation and the square of the error of the detection equipment; The square root of the difference is performed to obtain the true fluctuation of the material detection.

7. The quality monitoring method according to claim 2, characterized in that, Also includes: The trend of target deviation changes within multiple consecutive target periods is statistically analyzed, wherein the target deviation includes at least one of the accuracy deviation, the repeatability deviation, and the inter-station deviation; Based on the aforementioned trend of change, predict the value of the target deviation within the target time period; When the predicted value exceeds the maintenance threshold of the target deviation, a predictive maintenance work order is generated.

8. A quality monitoring device, characterized in that, include: The equipment error acquisition module is used to acquire the detection equipment error of the target detection item. The detection equipment error is used to characterize the measurement fluctuation introduced by the detection equipment itself of the target detection item. The total fluctuation observation acquisition module is used to acquire multiple material detection results of the target detection item and obtain the total fluctuation observation based on the multiple material detection results. The real fluctuation acquisition module is used to obtain the real fluctuation of material detection based on the error of the detection equipment and the total observed fluctuation. The real process capability assessment module is used to assess the real process capability of the target testing item based on the real fluctuation of the material testing and the material specification limit. The quality status assessment module is used to assess the material quality status corresponding to the target detection item based on the actual process capability.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the quality monitoring method as described in any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium, wherein a computer program is stored on the non-transitory computer-readable storage medium, characterized in that, When the computer program is executed by the processor, it implements the quality monitoring method as described in any one of claims 1 to 7.

Images