Semiconductor production monitoring method, apparatus, electronic device, and storage medium

Abstract

The application provides a semiconductor production monitoring method and device, electronic equipment and storage medium, and relates to the field of semiconductor production monitoring.The method comprises the following steps: standardizing each received original production report to obtain a corresponding standard production report; based on the standard production reports of the same product batch that have been received, comparing whether the standard station name and the standard substation name in the two standard production reports of adjacent report dates change in turn to obtain a comparison result; and dividing the standard production reports of the same product batch into multiple stay segments according to the comparison result, generating life cycle track data corresponding to the product batch based on the multiple stay segments, the starting time, the ending time and the stay days corresponding to each stay segment.The application can convert discrete static snapshot data into continuous product batch life cycle track, and reduces the system maintenance cost and the difficulty of multi-source data integration by standardizing the original production report.

Description

Technical Field

[0001] This invention relates to the field of semiconductor production monitoring, and more specifically, to a semiconductor production monitoring method, apparatus, electronic device, and storage medium. Background Technology

[0002] In the semiconductor manufacturing and testing field, the production process typically spans multiple independent factories or workshops (such as wafer fabrication plants, packaging plants, and testing plants). Due to network isolation and data security policies between the factories, the production client often cannot access the contractor's Manufacturing Execution System (MES) in real time, and can only obtain production progress information through the daily work-in-process (WIP) reports sent by the contractor.

[0003] like Figure 1 As shown, existing technologies mainly rely on static comparison methods based on snapshots for production monitoring. That is, the client receives daily static reports (containing fields such as batch number, current site, status, and quantity) and calculates the increase or decrease of inventory at each site by manually or through scripts comparing the differences between the reports of two adjacent days.

[0004] However, this monitoring method based on discrete snapshots can only obtain the production situation of two adjacent days, and cannot form a continuous product batch life cycle trajectory, resulting in a break in the production process record and making it difficult to trace. In addition, the report formats provided by different contractors are not uniform, and existing technologies usually rely on manual maintenance of multiple sets of conversion rules, which increases the system maintenance cost and the difficulty of integrating multi-source data. Summary of the Invention

[0005] In view of this, the purpose of the present invention is to provide a semiconductor production monitoring method, apparatus, electronic device and storage medium that can transform discrete static snapshot data into continuous product batch lifecycle trajectories, and reduce system maintenance costs and the difficulty of multi-source data integration by standardizing the original production reports.

[0006] To achieve the above objectives, the technical solutions adopted in the embodiments of the present invention are as follows: In a first aspect, the present invention provides a semiconductor production monitoring method, the method comprising: Each received raw production report is standardized to obtain a corresponding standard production report; Based on the standard production reports of the same product batch that have been received, the standard site name and standard sub-site name in two standard production reports with adjacent report dates are compared to see if they have changed, and the comparison results are obtained. Based on the comparison results, the standard production report of the same product batch is divided into multiple dwell periods, and the start time, end time and number of dwell days corresponding to each dwell period are recorded. Based on the multiple dwell periods, the start time, end time, and number of days of each dwell period, lifecycle trajectory data corresponding to the product batch is generated.

[0007] In an optional implementation, the standardization process for each received raw production report to obtain a corresponding standard production report includes: Each received raw production report undergoes field identification and data cleaning to extract product batch identifier, report date, original site name, original sub-site name, and product quantity. The fields corresponding to the product batch identifier, report date, original site name, original sub-site name, and product quantity in the original production report are converted into corresponding preset standard fields; The original subsite name is used to search a preset mapping rule base; the mapping rule base records the mapping relationship between standard subsite names and standard site names; If a target mapping relationship matching the original subsite name exists in the mapping rule base, then the standard subsite name and standard site name corresponding to the original subsite name are determined based on the target mapping relationship; Based on the product batch identifier, the report date, the standard site name, the standard sub-site name, and the product quantity, a corresponding standard production report is generated.

[0008] In an optional implementation, the method further includes: If there is no target mapping relationship matching the original sub-site name in the mapping rule base, the similarity between the original sub-site name and each standard sub-site name in the mapping rule base is determined based on a preset similarity algorithm, and the standard sub-site names with similarity higher than the preset similarity threshold are determined as candidate standard sub-site names. The frequency of occurrence of the candidate standard site name corresponding to the candidate standard sub-site name is counted. The original sub-site name is used as the standard sub-site name, and the candidate standard site name with the highest frequency is used as the standard site name. A new mapping relationship is constructed and saved to the mapping rule base.

[0009] In an optional implementation, the step of comparing the standard site name and standard sub-site name in two standard production reports from adjacent report dates, based on the received standard production reports of the same product batch, to obtain a comparison result, includes: The standard production reports for the same product batch that have been received are sorted according to the report date to obtain the production report sequence corresponding to the product batch. The comparison results are obtained by sequentially comparing whether the standard site name and standard sub-site name in two adjacent standard production reports in the production report sequence have changed.

[0010] In an optional implementation, the step of dividing the standard production report of the same product batch into multiple dwell periods based on the comparison results, and recording the start time, end time, and number of dwell days for each dwell period, includes: Set an initial grouping identifier for the first standard production report in the production report sequence corresponding to the product batch; Starting from the second standard production report in the production report sequence, when the comparison result indicates that the standard site name and standard sub-site name in the current standard production report are the same as those in the adjacent previous standard production report, the grouping identifier corresponding to the current standard production report is set to the grouping identifier corresponding to the adjacent previous standard production report. When the comparison result indicates that the standard site name in the current standard production report has changed compared to the standard site name in the adjacent previous standard production report, or the standard sub-site name in the current standard production report has changed compared to the standard sub-site name in the adjacent previous standard production report, the grouping identifier corresponding to the current standard production report is increased by a preset value based on the grouping identifier corresponding to the adjacent previous standard production report. Standard production reports with the same grouping identifier in the same product batch are divided into one dwell segment, thus obtaining multiple dwell segments; The minimum report date corresponding to the standard production report in each of the said dwell periods is recorded as the start time, the maximum report date corresponding to the standard production report in each of the said dwell periods is recorded as the end time, and the number of standard production reports in each of the said dwell periods is recorded as the number of dwell days.

[0011] In an optional implementation, the method further includes: Based on the lifecycle trajectory data of each product batch within the preset statistical window, the total number of outgoing batches corresponding to each standard site is determined; the total number of outgoing batches includes the sum of the number of products outgoing from the standard site within the preset statistical window for each product batch. Calculate the average daily throughput of each standard site based on the total number of migration batches corresponding to each standard site and the preset statistical window; Based on the current inventory of all product batches at each of the standard sites and the average daily throughput of each of the standard sites, the estimated emptying time for each of the standard sites is obtained. Based on the preset process route corresponding to each product batch, determine the target standard station where the product batch is currently located and all the standard stations it needs to pass through in the future; The estimated clearance time of the target standard site and the estimated clearance time of each subsequent standard site are added together to obtain the estimated completion time of each product batch.

[0012] In an optional implementation, the method further includes: Anomaly detection is performed based on the lifecycle trajectory data of each product batch. When the number of days of stay corresponding to any product batch is greater than a preset threshold, it is determined that the product batch has timed outage. When a duplicate standard station is detected in the standard station sequence corresponding to any product batch, it is determined that the product batch has a return flow anomaly.

[0013] In a second aspect, the present invention provides a semiconductor production monitoring device, the device comprising: The data standardization module is used to standardize each received raw production report to obtain the corresponding standard production report; The difference detection module is used to compare the standard site name and standard sub-site name in two standard production reports with adjacent report dates, based on the standard production reports of the same product batch that have been received, and obtain the comparison results. The dwell segment generation module is used to divide the standard production report of the same product batch into multiple dwell segments according to the comparison results, and record the start time, end time and dwell days corresponding to each dwell segment; The trajectory generation module is used to generate lifecycle trajectory data corresponding to the product batch based on multiple dwell segments, the start time, end time and number of days of each dwell segment.

[0014] Thirdly, the present invention provides an electronic device including a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the steps of the semiconductor production monitoring method as described in any of the foregoing embodiments.

[0015] Fourthly, the present invention provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the semiconductor production monitoring method as described in any of the foregoing embodiments.

[0016] This invention provides a semiconductor production monitoring method, apparatus, electronic device, and storage medium. The method includes: standardizing each received raw production report to obtain a corresponding standard production report; comparing the standard site name and standard sub-site name in two standard production reports from adjacent report dates to see if they have changed, based on the received standard production reports for the same product batch, and obtaining a comparison result; dividing the standard production reports for the same product batch into multiple dwell periods according to the comparison result, and recording the start time, end time, and dwell days for each dwell period; and generating lifecycle trajectory data corresponding to the product batch based on the multiple dwell periods and the start time, end time, and dwell days for each dwell period. This invention can transform discrete static snapshot data into a continuous product batch lifecycle trajectory, and by standardizing the raw production reports, it reduces system maintenance costs and the difficulty of integrating multi-source data.

[0017] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0018] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This diagram illustrates an existing snapshot-based static comparison method for production monitoring. Figure 2 This invention provides a schematic flowchart of a semiconductor production monitoring method. Figure 3 This illustration shows another schematic flowchart of the semiconductor production monitoring method provided in an embodiment of the present invention; Figure 4 This illustration shows another schematic flowchart of the semiconductor production monitoring method provided in an embodiment of the present invention; Figure 5 This invention illustrates a functional block diagram of a semiconductor production monitoring device provided in an embodiment of the present invention. Figure 6 This invention illustrates another functional block diagram of the semiconductor production monitoring device provided in an embodiment of the present invention; Figure 7 A block diagram of an electronic device provided in an embodiment of the present invention is shown.

[0020] Icons: 100 - Electronic device; 110 - Memory; 120 - Processor; 130 - Communication module; 600 - Semiconductor production monitoring device; 610 - Data standardization module; 620 - Difference detection module; 630 - Dwell time generation module; 640 - Trajectory generation module; 650 - Capacity assessment module; 660 - Anomaly detection module. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0022] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0023] It should be noted that relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, 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 limitations, 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.

[0024] Existing semiconductor production monitoring methods can only obtain production information for two consecutive days, failing to form a continuous product batch lifecycle trajectory. This results in gaps in production process records, making traceability difficult. Furthermore, the report formats provided by different contractors are inconsistent, and existing technologies typically rely on manual maintenance of multiple sets of conversion rules, increasing system maintenance costs and the difficulty of integrating multi-source data.

[0025] To address the aforementioned problems, embodiments of the present invention provide a semiconductor production monitoring method, apparatus, electronic device, and storage medium. The method includes: standardizing each received raw production report to obtain a corresponding standard production report; comparing the standard site name and standard sub-site name in two standard production reports from adjacent report dates to determine if they have changed, based on the received standard production reports for the same product batch, and obtaining a comparison result; dividing the standard production reports for the same product batch into multiple dwell periods according to the comparison result, and recording the start time, end time, and dwell days for each dwell period; and generating lifecycle trajectory data corresponding to the product batch based on the multiple dwell periods and the start time, end time, and dwell days for each dwell period. This invention can transform discrete static snapshot data into a continuous product batch lifecycle trajectory, and by standardizing the raw production reports, it reduces system maintenance costs and the difficulty of integrating multi-source data.

[0026] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0027] Please refer to Figure 2 This is a schematic flowchart of a semiconductor production monitoring method provided in an embodiment of the present invention. It should be noted that the semiconductor production monitoring method of the present invention is not based on... Figure 2 The specific order described below is a limitation. It should be understood that in other embodiments, the order of some steps in the semiconductor production monitoring method of the present invention can be interchanged according to actual needs, or some steps can be omitted or deleted. This semiconductor production monitoring method can be applied to electronic devices such as mobile phones, tablets, and PCs (Personal Computers). The following will describe... Figure 2 The specific process shown will be explained in detail.

[0028] Step S201: Standardize each received raw production report to obtain the corresponding standard production report.

[0029] In this embodiment, by standardizing the received raw production reports, there is no need to develop dedicated interfaces for each external data source. This allows for flexible adaptation to changes in report formats from different vendors, effectively solving the problem of inconsistent multi-source data formats and reducing system maintenance costs and the difficulty of integrating multi-source data.

[0030] Step S202: Based on the received standard production reports for the same product batch, compare the standard site name and standard sub-site name in two standard production reports with adjacent report dates to see if they have changed, and obtain the comparison result.

[0031] In this embodiment, a sliding window algorithm is used to perform adjacent comparisons on discrete snapshot data (daily standard production reports) sorted by time, thereby determining whether the standard site name and standard sub-site name have changed, and obtaining the comparison results. In this embodiment, a subsite can be understood as a step within a site. For example, a site called Status can be further subdivided into multiple subsites such as Step1, Step2, and Step3. When a product batch moves from one site to the next, it signifies a site change event; when a product batch moves from one subsite within a site to the next, it signifies an intra-site status change event.

[0032] Step S203: Based on the comparison results, the standard production report of the same product batch is divided into multiple dwell periods, and the start time, end time and dwell days of each dwell period are recorded.

[0033] In this embodiment, based on the comparison results of the standard site name and standard sub-site name in two standard production reports with adjacent report dates, consecutive standard production reports with no change in standard site name and standard sub-site name can be aggregated into a dwell segment. For each dwell segment, the corresponding start time, end time and dwell days are recorded.

[0034] Step S204: Based on multiple dwell periods, the start time, end time, and number of dwell days for each dwell period, generate lifecycle trajectory data corresponding to the product batch.

[0035] In this embodiment, by connecting all dwell periods in chronological order, the lifecycle trajectory data of the product batch can be formed. This transforms discrete static data into a continuous dynamic trajectory, enabling the reconstruction of the entire production process of the product batch even when direct connection to the contractor's MES system is not possible and only discrete daily reports are available.

[0036] As can be seen, the semiconductor production monitoring method provided in this embodiment of the invention standardizes each received raw production report to obtain a corresponding standard production report; based on the received standard production reports of the same product batch, it sequentially compares whether the standard site name and standard sub-site name in two standard production reports of adjacent report dates have changed, obtaining a comparison result; according to the comparison result, the standard production reports of the same product batch are divided into multiple dwell periods, and the start time, end time, and dwell days corresponding to each dwell period are recorded; based on multiple dwell periods and the start time, end time, and dwell days corresponding to each dwell period, lifecycle trajectory data corresponding to the product batch is generated. This invention can transform discrete static snapshot data into continuous product batch lifecycle trajectories, and by standardizing the raw production reports, it reduces system maintenance costs and the difficulty of integrating multi-source data.

[0037] In one implementation, step S201 above involves standardizing each received raw production report to obtain a corresponding standard production report, specifically including: Each received raw production report undergoes field identification and data cleaning to extract product batch identifier, report date, original site name, original sub-site name, and product quantity. The fields corresponding to product batch identifier, report date, original site name, original sub-site name, and product quantity in the raw production report are converted into corresponding preset standard fields. A preset mapping rule base is searched based on the original sub-site name. The mapping rule base records the mapping relationship between standard sub-site names and standard site names.

[0038] In this embodiment, considering that the client cannot directly access the contractor's MES system and can only receive daily production reports from the contractor, the reports provided by different contractors are not entirely consistent in terms of field names, field formats, site / subsite coding rules, and macro-stage division methods. This makes it impossible to directly compare, reconstruct, and statistically analyze production reports from different dates and sources. Therefore, this invention mainly achieves standardization of original production reports through two aspects: field identification and data cleaning, and determining the corresponding standard site name based on the original subsite name.

[0039] First, each received raw production report undergoes field identification and data cleaning. This involves extracting information such as product batch identifier, report date, original site name, original sub-site name, and product quantity, along with their corresponding fields. These fields are then converted into standardized, predefined fields. For example, different reports might use "Lot" and "VTestLot" to represent batch numbers, and "Date" and "ReportDate" to represent report dates. Through the standardization process of this invention, these fields are uniformly converted into predefined standard fields; for instance, "Lot" is uniformly used to represent batch numbers, and "Date" is uniformly used to represent report dates.

[0040] Secondly, the original sub-site names in the original production reports are matched against a pre-defined mapping rule base. This base records the mapping relationships between standard sub-site names and standard site names. These relationships can be manually confirmed or generated and persistently stored by algorithms. If a target mapping relationship matching the original sub-site name exists in the rule base, the corresponding standard sub-site name and standard site name can be directly determined based on this relationship. Then, based on the product batch identifier, report date, standard site name, standard sub-site name, and product quantity, a corresponding standard production report is generated.

[0041] For example, when the mapping rule base records mapping relationships such as FT1T0→WIP, LS1T0→WIP, WPIT0→WIP, WBT→WBT, WAT→WAT, and the original subsite name is FT1T0, the corresponding standard subsite name is determined to be FT1T0 and the corresponding standard site name is WIP by searching the mapping rule base.

[0042] In one implementation, the standardization process for each received raw production report in step S201 to obtain a corresponding standard production report further includes: If there is no target mapping relationship matching the original sub-site name in the mapping rule base, the similarity between the original sub-site name and each standard sub-site name in the mapping rule base is determined based on the preset similarity algorithm. The standard sub-site names with similarity higher than the preset similarity threshold are determined as candidate standard sub-site names. The frequency of occurrence of the candidate standard sub-site names is counted. The original sub-site name is used as the standard sub-site name, and the candidate standard site name with the highest frequency is used as the standard site name. A new mapping relationship is constructed and saved to the mapping rule base.

[0043] For original new subsite names that do not exist in the mapping rule base, and for which there is no target mapping relationship that matches the original new subsite name, the standard subsite name and standard site name corresponding to the original new subsite can be found by combining subsite name prefixes, string similarity, and the mapping relationship between standard subsite names and standard site names recorded in the mapping rule base.

[0044] In this embodiment, the preset similarity algorithm can be the Levenshtein distance algorithm (an edit distance algorithm) used to measure the degree of difference between two strings. Its core idea is to calculate the minimum number of single-character edit operations required to transform one string into another. Allowed edit operations include insertion, deletion, and substitution; insertion adds a character, deletion removes a character, and substitution replaces one character with another. The smaller the edit distance, the more similar the two strings are; a distance of 0 indicates that the two strings are completely identical. For example, for the first appearance of the original new subsite name FT3T0, the mapping relationships recorded in the mapping rule base are FT1T0→WIP, FT2T0→WIP, LS1T0→TEST, and WPIT1→WAT. There is no target mapping relationship matching FT3T0 in the mapping rule base. At this time, the similarity between FT3T0 and FT1T0, FT2T0, LS1T0, and WPIT1 is measured by calculating the Levenshtein distances (1, 1, 3, and 4 respectively) between FT3T0 and FT1T0, FT2T0, LS1T0, and WPIT1. Assuming that a standard sub-site name with a Levenshtein distance ≤ 3 is considered a standard sub-site name with a similarity higher than a preset similarity threshold, then FT1T0, FT2T0, and LS1T0 can be identified as candidate standard sub-site names. Since the candidate standard site names WIP corresponding to FT1T0 and FT2T0 appear twice, and the candidate standard site name TEST corresponding to LS1T0 appears once, the candidate standard site name WIP has the highest frequency of occurrence. Therefore, FT3T0 is selected as the standard sub-site name, and WIP is identified as the standard site name corresponding to FT3T0. The mapping relationship FT3T0→WIP is constructed and saved to the mapping rule base. Based on the product batch identifier, report date, product quantity, and the standard site name and standard sub-site name in the new mapping relationship, the corresponding standard production report is generated.

[0045] As can be seen, the embodiments of the present invention standardize the original production reports by performing field identification, data cleaning, and determining the corresponding standard site name based on the original sub-site name. This enables the automatic identification of newly emerging sub-sites, achieving automated cleaning and unified management of multi-source heterogeneous data, and significantly reducing the development cost and maintenance difficulty of connecting with new vendors.

[0046] In one implementation, step S202 above, based on the received standard production reports for the same product batch, sequentially compares whether the standard site name and standard sub-site name in two standard production reports from adjacent report dates have changed, and obtains a comparison result, specifically including: The standard production reports of the same product batch that have been received are sorted according to the report date to obtain the production report sequence corresponding to the product batch; the standard site name and standard sub-site name in two adjacent standard production reports in the production report sequence are compared to see if they have changed, and the comparison results are obtained.

[0047] In this embodiment, for all standard production reports received from the same product batch, they can be sorted in ascending order based on their report dates to generate an ordered snapshot sequence (i.e., a production report sequence) corresponding to that product batch; then, a sequence covering two adjacent time points T is constructed. n and T n+1 The sliding window sequentially compares the standard station name and standard sub-station name in two adjacent standard production reports within the sliding window, i.e., compares Status(T) n ) and Status (T n+1 If the two are inconsistent, then it is determined that T is the correct answer. n ~T n+1 A site migration event occurred within the time interval; compare Step(T) n ) and Step (T) n+1 If the sites are the same but the subsites are different, the determination is made in T. n ~T n+1 A status change event occurred within the time interval; if Status(T) n ) and Status (T n+1 The same applies to Step(T), and Step(T) is also the same. n ) and Step (T) n+1 If the T batch is also the same, then the product batch is considered to be in T. n ~T n+1 It remains at the same station and in a stable state throughout the time interval.

[0048] In one implementation, step S203 above, which divides the standard production report of the same product batch into multiple dwell periods based on the comparison results and records the start time, end time, and number of dwell days for each dwell period, specifically includes: An initial grouping identifier is set for the first standard production report in the production report sequence corresponding to the product batch. Starting from the second standard production report in the production report sequence, if the comparison result indicates that the standard station name and standard sub-site name of the current standard production report are the same as those of the adjacent previous standard production report, the grouping identifier corresponding to the current standard production report is set to the grouping identifier corresponding to the adjacent previous standard production report. If the comparison result indicates that the standard station name in the current standard production report has changed compared to the standard station name in the adjacent previous standard production report, or that the standard sub-site name in the current standard production report has changed compared to the standard sub-site name in the adjacent previous standard production report, the grouping identifier corresponding to the current standard production report is increased by a preset value based on the grouping identifier corresponding to the adjacent previous standard production report. Standard production reports with the same grouping identifier in the same product batch are divided into a single dwell segment, thus obtaining multiple dwell segments. The minimum report date corresponding to the standard production report in each dwell segment is recorded as the start time, the maximum report date corresponding to the standard production report in each dwell segment is recorded as the end time, and the number of standard production reports in each dwell segment is recorded as the dwell days.

[0049] For example, for a batch of standard production reports, arranged by report date, an ordered sequence of production reports can be obtained (corresponding report dates are T1, T2, T3, T4, T5, T6, T7, T8). The first standard production report in the production report sequence is assigned an initial group identifier (Group ID), for example, 1. Starting from the second standard production report, it is compared with the adjacent previous standard production report in turn. It is found that no site change or status change event occurred in the time interval T1~T2, so the Group ID is not incremented. That is, the Group ID of the second standard production report inherits the Group ID of the first standard production report and is still 1. In the time interval T2~T3, a site change occurred, so the Group ID needs to be incremented. Therefore, the Group ID of the third standard production report is increased by a preset value (let's say 1) based on the Group ID of the second standard production report, so the Group ID of the third standard production report is 2; and so on. Finally, all standard production reports with the same Group ID are aggregated into a single dwell period. The start time of this dwell period is the smallest report date, the end time is the largest report date, and the dwell time is the number of standard production reports contained within it. For example, the first and second standard production reports both have a Group ID of 1, so they are aggregated into a single dwell period with a start time of T1, an end time of T2, and a dwell time of 2 days.

[0050] Understandably, if neither a site change event nor a status change event occurs, the product batch is considered to be still in the same stage, and the Group ID of the corresponding standard production report will not increment, indicating they belong to the same group. Once a site change event or a status change event occurs, the product batch is considered to have entered a new stage, and the Group ID of the standard production report will increment, creating a new group. Ultimately, standard production reports with the same Group ID belong to the same group and will be aggregated into the same dwell period.

[0051] It should be noted that, in the aforementioned time-series reconstruction algorithms, besides using the timestamp-based sliding window algorithm, a finite state machine (FSM) model can also be employed. For example, a predefined standard state diagram of wafer flow can be used, with daily snapshot data as input events to the state machine, and intermediate processes can be derived through state transition logic. Although this method has higher implementation complexity, it can achieve similar reconstruction results. When comparing standard site names and standard sub-site names in two standard production reports from adjacent report dates, in addition to field-by-field comparison, hash fingerprinting technology can also be introduced. For example, a hash value can be calculated by combining key fields of daily records, and changes in hash values ​​can be quickly located by comparing changes in hash values, thus enabling the identification of site change or state change events.

[0052] As can be seen, the embodiments of the present invention, through the above-described time-series reconstruction algorithm, can infer the site change and status change events of product batches from discrete daily data, and then transform them into continuous lifecycle trajectory data, making the production process of product batches traceable.

[0053] Furthermore, in existing technologies, when assessing the estimated completion time of a product batch, the theoretical capacity provided by the equipment manufacturer is typically used directly as the calculation benchmark, and linear extrapolation is performed based on the current site inventory. However, the theoretical capacity is usually a theoretical value measured under ideal conditions and does not take into account dynamic factors such as equipment downtime, line changeover and debugging, personnel scheduling, and process anomalies in actual production. Therefore, directly using the theoretical capacity to calculate the estimated completion time often leads to a significant deviation between the predicted result and the actual delivery time.

[0054] Based on this, please refer to Figure 3 To address the issue of significant bias in existing capacity assessment methods, the semiconductor production monitoring method provided in this embodiment of the invention may further include the following steps: Step S301: Based on the lifecycle trajectory data of each product batch within the preset statistical window, determine the total number of migration batches corresponding to each standard site; the total number of migration batches includes the sum of the number of products migrated from the standard site within the preset statistical window for each product batch.

[0055] Step S302: Calculate the average daily throughput of each standard site based on the total number of migration batches corresponding to each standard site and the preset statistical window.

[0056] Step S303: Based on the current inventory of all product batches at each standard site and the average daily throughput of each standard site, obtain the estimated emptying time for each standard site.

[0057] Step S304: Based on the preset process route corresponding to each product batch, determine the target standard station where the product batch is currently located and all the standard stations that it needs to pass through in the future.

[0058] Step S305: The estimated clearing time of the target standard site and the estimated clearing time of each subsequent standard site are added together to obtain the estimated completion time of each product batch.

[0059] In this embodiment, the preset statistical window can be set according to actual conditions, for example, it can be set to the past 7 days. By traversing the lifecycle trajectory data corresponding to each product batch within the preset statistical window, standard sites where site migration events have occurred can be filtered out. Then, the total number of migration batches N_out for each standard site within the preset statistical window is counted. Therefore, the average daily throughput TP_daily for each standard site is = N_out / preset statistical window.

[0060] After obtaining the average daily throughput TP_daily for each standard site, the current product inventory WIP_current for each standard site is obtained. The estimated clearing time for each standard site is then calculated as ETC_status = WIP_current / TP_daily. If TP_daily is 0, it is marked as a blocked state.

[0061] For each product batch, the estimated completion time can be obtained by summing the estimated clearing time (ETC_status) of the current target standard station and the subsequent standard stations it needs to pass through.

[0062] For example, suppose the preset process route for product batch Lot_1 is: Standard Site A → Standard Site B → Standard Site C. At the current moment, Lot_1 has just arrived at Standard Site A, and the current product inventory at Standard Site A is WIP_A = 1000 units. The average daily throughput of Standard Site A over the past 7 days (preset statistical window) is TP_A = 500 units / day. Then, the estimated emptying time of Standard Site A is ETC_A = 1000 / 500 = 2 days. Therefore, it is assumed that Lot_1 will take approximately 2 days to pass through Standard Site A.

[0063] The current product inventory of standard site B is WIP_B=600 units, and the average daily throughput is TP_B=200 units / day. Therefore, the estimated emptying time of standard site B is ETC_B=600 / 200=3 days. At this time, it is assumed that Lot_1 will take about 3 days to pass through standard site B in the future.

[0064] The current product inventory of standard site C is WIP_C=400 units, and the average daily throughput is TP_C=400 units / day. Therefore, the estimated emptying time of standard site C is ETC_C=400 / 400=1 day. At this time, it is assumed that Lot_1 will take about 1 day to pass through standard site C in the future.

[0065] Finally, the estimated completion time (ETC_Total) of product batch Lot_1 is calculated as ETC_A + ETC_B + ETC_C = 2 + 3 + 1 = 6 days.

[0066] It's important to note that in semiconductor manufacturing, each product batch has a predefined process route. Even if a product batch is currently at site A, the process route indicates that it will inevitably pass through sites B and C in the future. Furthermore, sites B and C currently have their own product inventory and average daily throughput, allowing for the calculation of their estimated emptying times in advance.

[0067] Furthermore, the aforementioned calculation of the estimated clearing time for each station, ETC_status = WIP_current / TP_daily, physically represents "the time required to clear all work-in-process at the current station." However, in this embodiment, a macroscopic queuing theory estimation model is used. To simplify the calculation, when performing macroscopic capacity assessment, the overall clearing time of the station (i.e., the macroscopic average cycle) is directly used as the estimated value of the expected dwell time (i.e., the estimated clearing time) for each product batch passing through the station.

[0068] For the calculation of the daily throughput TP_daily, in addition to the Simple Moving Average algorithm mentioned above, the Exponentially Weighted Moving Average (EWMA) or Kalman Filter algorithm can also be used. These methods give higher weight to recent data and can more sensitively reflect capacity fluctuations.

[0069] For calculating the expected clearance time, in addition to using rule-based linear extrapolation models, machine learning regression models (such as random forests and LSTMs) can also be used. Using historical product inventory levels, site information, and product types as input features, the model is trained to directly predict the completion time. All of these methods fall under the category of dynamic prediction driven by historical data.

[0070] It should also be noted that, in this embodiment, the implementation of the preset process route is not a hard-coded, strictly linear linked list (e.g., A must go to B, B must go to C), but a flexible, virtual route based on priority rules.

[0071] Specifically, this can be achieved through a front-end and back-end combination: At the macro level, the back-end maintains a mapping table that maps the diverse original sites of various contract manufacturers to several standard macro states (such as waiting for WBT before testing, WIP in process, and waiting for WAT after testing), establishing the overall framework of the process route. At the micro level, the front-end maintains a priority dictionary of site prefixes (e.g., FT priority 10, LM priority 20, LS priority 30). Thus, when the system needs to know the subsequent sites of a product batch currently at site D, it doesn't need to consult a rigid route table. It only needs to extract the prefix of the current site D and compare it with the priority dictionary. All sites with priority values ​​greater than the current site D are automatically assigned to the set of subsequent standard sites. This method of dynamically deriving the process route based on the priority dictionary greatly enhances the system's robustness. Even if an external contract manufacturer temporarily adds a test site or changes its name, as long as the prefix conforms to the specifications, the system can still accurately derive its subsequent route and calculate the estimated completion time without errors or crashes. For example, in a WIP site, there are many subsites (process steps), such as LM01 and LM02. Users cannot know how many LM steps the factory will perform. Through the above method, it can be automatically inferred and sorted to obtain the process route.

[0072] As can be seen, the embodiments of the present invention abandon the idealized nominal capacity parameters and use the reconstructed historical trajectory data to dynamically calculate the actual output (i.e., dynamic throughput index) within a specific time window. This index includes actual production factors such as equipment downtime, line changeover and debugging, personnel scheduling and process abnormalities, so that the calculated expected completion time is more in line with the actual production and the prediction accuracy is significantly better than that of traditional linear extrapolation.

[0073] In one implementation, please refer to Figure 4 The semiconductor production monitoring method provided in this embodiment of the invention may further include the following steps: Step S401: Based on the lifecycle trajectory data of each product batch, perform anomaly detection. When the number of days of stay corresponding to any product batch is greater than a preset threshold, it is determined that the product batch has timed outage. When a duplicate standard station is detected in the standard station sequence corresponding to any product batch, it is determined that the product batch has a return flow anomaly.

[0074] In this embodiment, by comparing the number of days of stay in the lifecycle trajectory data of each product batch with a preset threshold, product batches that have exceeded the time limit can be screened out, thereby realizing the detection of the stay of each product batch. By analyzing the standard station sequence corresponding to each product batch, when the same standard station appears repeatedly in the sequence, for example, when a path structure of standard station A → standard station B → ... → standard station A appears, it is determined that the product batch has experienced a return flow anomaly. To perform the corresponding steps in the above embodiments and various possible methods, an implementation of a semiconductor production monitoring device is given below. Please refer to... Figure 5 This is a functional block diagram of a semiconductor production monitoring device 600 provided in an embodiment of the present invention. It should be noted that the basic principle and technical effects of the semiconductor production monitoring device 600 provided in this embodiment are the same as those in the above embodiments. For the sake of brevity, any parts not mentioned in this embodiment can be referred to the corresponding content in the above embodiments. The semiconductor production monitoring device 600 includes: a data standardization module 610, a difference detection module 620, a dwell segment generation module 630, and a trajectory generation module 640.

[0075] The data standardization module 610 is used to standardize each received raw production report to obtain the corresponding standard production report.

[0076] The difference detection module 620 is used to compare the standard site name and standard sub-site name in two standard production reports with adjacent report dates, based on the standard production reports of the same product batch that have been received, to obtain the comparison results.

[0077] The dwell segment generation module 630 is used to divide the standard production report of the same product batch into multiple dwell segments based on the comparison results, and record the start time, end time and dwell days of each dwell segment.

[0078] The trajectory generation module 640 is used to generate lifecycle trajectory data for a product batch based on multiple dwell segments, the start time, end time, and number of days for each dwell segment.

[0079] Optionally, the data standardization module 610 is specifically used to perform field identification and data cleaning on each received original production report, extracting the product batch identifier, report date, original site name, original sub-site name, and product quantity; converting the fields corresponding to the product batch identifier, report date, original site name, original sub-site name, and product quantity in the original production report into corresponding preset standard fields; searching a preset mapping rule library based on the original sub-site name; the mapping rule library records the mapping relationship between standard sub-site names and standard site names; if a target mapping relationship matching the original sub-site name exists in the mapping rule library, then the standard sub-site name and standard site name corresponding to the original sub-site name are determined based on the target mapping relationship; and generating the corresponding standard production report based on the product batch identifier, report date, standard site name, standard sub-site name, and product quantity.

[0080] Optionally, the data standardization module 610 is further configured to, if there is no target mapping relationship matching the original sub-site name in the mapping rule base, determine the similarity between the original sub-site name and each standard sub-site name in the mapping rule base based on a preset similarity algorithm, determine the standard sub-site name with a similarity higher than a preset similarity threshold as a candidate standard sub-site name, count the frequency of occurrence of the candidate standard sub-site name corresponding to the candidate standard sub-site name, use the original sub-site name as the standard sub-site name, use the candidate standard sub-site name with the highest frequency as the standard site name, construct a new mapping relationship and save it to the mapping rule base.

[0081] Optionally, the difference detection module 620 is specifically used to sort the received standard production reports of the same product batch according to the report date to obtain the production report sequence corresponding to the product batch; and to compare whether the standard site name and standard sub-site name in two adjacent standard production reports in the production report sequence have changed to obtain the comparison result.

[0082] Optionally, the dwell segment generation module 630 is specifically used to set an initial grouping identifier for the first standard production report in the production report sequence corresponding to the product batch; starting from the second standard production report in the production report sequence, when the comparison result indicates that the standard station name and standard sub-station name in the current standard production report are the same as those in the adjacent previous standard production report, the grouping identifier corresponding to the current standard production report is set to the grouping identifier corresponding to the adjacent previous standard production report; when the comparison result indicates that the standard station name in the current standard production report has changed compared to the standard station name in the adjacent previous standard production report, or the standard sub-station name in the current standard production report has changed compared to the standard sub-station name in the adjacent previous standard production report, the grouping identifier corresponding to the current standard production report is increased by a preset value based on the grouping identifier corresponding to the adjacent previous standard production report; standard production reports with the same grouping identifier in the standard production reports of the same product batch are divided into a dwell segment, thereby obtaining multiple dwell segments; the minimum report date corresponding to the standard production report in each dwell segment is recorded as the start time, the maximum report date corresponding to the standard production report in each dwell segment is recorded as the end time, and the number of standard production reports in each dwell segment is recorded as the dwell days.

[0083] Alternatively, please refer to Figure 6 The semiconductor production monitoring device 600 may also include a capacity assessment module 650 and an anomaly detection module 660.

[0084] The capacity assessment module 650 is used to determine the total number of outgoing batches for each standard site based on the lifecycle trajectory data of each product batch within a preset statistical window. The total number of outgoing batches includes the sum of the number of products outgoing from the standard sites for each product batch within the preset statistical window. Based on the total number of outgoing batches for each standard site and the preset statistical window, the module calculates the average daily throughput for each standard site. Based on the current inventory of all product batches at each standard site and the average daily throughput for each standard site, the module obtains the estimated emptying time for each standard site. Based on the preset process route for each product batch, the module determines the target standard site where the product batch is currently located and all the standard sites it needs to pass through subsequently. The module adds up the estimated emptying time of the target standard site and the estimated emptying time of each subsequent standard site to obtain the estimated completion time for each product batch.

[0085] The anomaly detection module 660 is used to perform anomaly detection based on the lifecycle trajectory data of each product batch. When the number of days of stay corresponding to any product batch is greater than a preset threshold, it is determined that the product batch has timed outages. When a duplicate standard station is detected in the standard station sequence corresponding to any product batch, it is determined that the product batch has a return flow anomaly.

[0086] Please refer to Figure 7 This is a block diagram of an electronic device 100 provided in an embodiment of the present invention. The electronic device 100 includes a memory 110, a processor 120, and a communication module 130. The memory 110, processor 120, and communication module 130 are electrically connected to each other directly or indirectly to realize data transmission or interaction. For example, these components can be electrically connected to each other through one or more communication buses or signal lines.

[0087] The memory 110 is used to store programs or data. The memory 110 may be, but is not limited to, random access memory (RAM), read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), etc.

[0088] The processor 120 is used to read / write data or programs stored in the memory 110 and to perform corresponding functions. For example, when a computer program stored in the memory 110 is executed by the processor 120, the semiconductor production monitoring method disclosed in the above embodiments can be implemented.

[0089] The communication module 130 is used to establish a communication connection between the electronic device 100 and other devices via a network, and to send and receive data via the network.

[0090] It should be understood that, Figure 7 The structure shown is only a schematic diagram of the electronic device 100. The electronic device 100 may also include components that are larger than... Figure 7 The more or fewer components shown, or having the same Figure 7 The different configurations shown. Figure 7 The components shown can be implemented using hardware, software, or a combination thereof.

[0091] This invention also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor 120, implements the semiconductor production monitoring method disclosed in the above embodiments.

[0092] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can also be implemented in other ways. The apparatus embodiments described above are merely illustrative; for example, the flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of apparatus, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram and / or flowchart, and combinations of blocks in block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0093] In addition, the functional modules in the various embodiments of the present invention can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0094] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a portion 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 invention. 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.

[0095] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A semiconductor manufacturing monitoring method, characterized in that, The method includes: Each received raw production report is standardized to obtain a corresponding standard production report; Based on the standard production reports of the same product batch that have been received, the standard site name and standard sub-site name in two standard production reports with adjacent report dates are compared to see if they have changed, and the comparison results are obtained. Based on the comparison results, the standard production report of the same product batch is divided into multiple dwell periods, and the start time, end time and number of dwell days corresponding to each dwell period are recorded. Based on the multiple dwell periods, the start time, end time, and number of days of each dwell period, lifecycle trajectory data corresponding to the product batch is generated.

2. The semiconductor production monitoring method according to claim 1, characterized in that, The standardization process for each received raw production report to obtain a corresponding standard production report includes: Each received raw production report undergoes field identification and data cleaning to extract product batch identifier, report date, original site name, original sub-site name, and product quantity. The fields corresponding to the product batch identifier, report date, original site name, original sub-site name, and product quantity in the original production report are converted into corresponding preset standard fields; The original subsite name is used to search a preset mapping rule base; the mapping rule base records the mapping relationship between standard subsite names and standard site names; If a target mapping relationship matching the original subsite name exists in the mapping rule base, then the standard subsite name and standard site name corresponding to the original subsite name are determined based on the target mapping relationship; Based on the product batch identifier, the report date, the standard site name, the standard sub-site name, and the product quantity, a corresponding standard production report is generated.

3. The semiconductor production monitoring method according to claim 2, characterized in that, The method further includes: If there is no target mapping relationship matching the original sub-site name in the mapping rule base, the similarity between the original sub-site name and each standard sub-site name in the mapping rule base is determined based on a preset similarity algorithm, and the standard sub-site names with similarity higher than the preset similarity threshold are determined as candidate standard sub-site names. The frequency of occurrence of the candidate standard site name corresponding to the candidate standard sub-site name is counted. The original sub-site name is used as the standard sub-site name, and the candidate standard site name with the highest frequency is used as the standard site name. A new mapping relationship is constructed and saved to the mapping rule base.

4. The semiconductor production monitoring method according to claim 1, characterized in that, The method involves comparing the standard site name and standard sub-site name in two standard production reports from adjacent report dates, based on the received standard production reports for the same product batch, to obtain comparison results, including: The standard production reports for the same product batch that have been received are sorted according to the report date to obtain the production report sequence corresponding to the product batch. The comparison results are obtained by sequentially comparing whether the standard site name and standard sub-site name in two adjacent standard production reports in the production report sequence have changed.

5. The semiconductor production monitoring method according to claim 4, characterized in that, The step of dividing the standard production report of the same product batch into multiple dwell periods based on the comparison results, and recording the start time, end time, and number of dwell days for each dwell period, includes: Set an initial grouping identifier for the first standard production report in the production report sequence corresponding to the product batch; Starting from the second standard production report in the production report sequence, when the comparison result indicates that the standard site name and standard sub-site name in the current standard production report are the same as those in the adjacent previous standard production report, the grouping identifier corresponding to the current standard production report is set to the grouping identifier corresponding to the adjacent previous standard production report. When the comparison result indicates that the standard site name in the current standard production report has changed compared to the standard site name in the adjacent previous standard production report, or the standard sub-site name in the current standard production report has changed compared to the standard sub-site name in the adjacent previous standard production report, the grouping identifier corresponding to the current standard production report is increased by a preset value based on the grouping identifier corresponding to the adjacent previous standard production report. Standard production reports with the same grouping identifier in the same product batch are divided into one dwell segment, thus obtaining multiple dwell segments; The minimum report date corresponding to the standard production report in each of the said dwell periods is recorded as the start time, the maximum report date corresponding to the standard production report in each of the said dwell periods is recorded as the end time, and the number of standard production reports in each of the said dwell periods is recorded as the number of dwell days.

6. The semiconductor production monitoring method according to any one of claims 1-5, characterized in that, The method further includes: Based on the lifecycle trajectory data of each product batch within a preset statistical window, the total number of outgoing batches corresponding to each standard site is determined; the total number of outgoing batches includes the sum of the number of products outgoing from the standard site within the preset statistical window for each product batch. Calculate the average daily throughput of each standard site based on the total number of migration batches corresponding to each standard site and the preset statistical window; Based on the current inventory of all product batches at each of the standard sites and the average daily throughput of each of the standard sites, the estimated emptying time for each of the standard sites is obtained. Based on the preset process route corresponding to each product batch, determine the target standard station where the product batch is currently located and all the standard stations it needs to pass through in the future; The estimated clearance time of the target standard site and the estimated clearance time of each subsequent standard site are added together to obtain the estimated completion time of each product batch.

7. The semiconductor production monitoring method according to any one of claims 1-5, characterized in that, The method further includes: Anomaly detection is performed based on the lifecycle trajectory data of each product batch. When the number of days of stay corresponding to any product batch is greater than a preset threshold, it is determined that the product batch has timed outage. When a duplicate standard station is detected in the standard station sequence corresponding to any product batch, it is determined that the product batch has a return flow anomaly.

8. A semiconductor manufacturing monitoring device, characterized in that, The device includes: The data standardization module is used to standardize each received raw production report to obtain the corresponding standard production report; The difference detection module is used to compare the standard site name and standard sub-site name in two standard production reports with adjacent report dates, based on the standard production reports of the same product batch that have been received, and obtain the comparison results. The dwell segment generation module is used to divide the standard production report of the same product batch into multiple dwell segments according to the comparison results, and record the start time, end time and dwell days corresponding to each dwell segment; The trajectory generation module is used to generate lifecycle trajectory data corresponding to the product batch based on multiple dwell segments, the start time, end time and number of days of each dwell segment.

9. An electronic device, characterized in that, It includes a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the steps of the semiconductor production monitoring method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, A computer program is stored on the computer-readable storage medium, which, when executed by a processor, implements the steps of the semiconductor production monitoring method as described in any one of claims 1-7.

Images