Wafer automatic production control system and method

Abstract

This invention discloses a wafer automated production control system and method, relating to the field of industrial automation control technology. The method includes: integrating and associating standardized production event flows to construct a real-time production state graph; calculating process window residual values ​​and performing resource matching analysis to generate a candidate execution tuple set; based on the candidate execution tuple set, performing priority evaluation and conflict resolution to generate a preferred execution sequence; matching and allocating target machines and transport paths within the preferred execution sequence to generate a joint scheduling result; arranging control instructions based on the joint scheduling result to generate a collaborative control instruction set; performing arrival verification and processing permission processing to generate verified execution results; and performing processing tracking, virtual measurement confidence analysis, and adaptive adjustment of sampling paths on the verified execution results to generate quality control results. This invention improves the response accuracy, quality identification targeting, and control closed-loop stability in the production control process.

Description

Technical Field

[0001] This invention relates to the field of industrial automation control technology, and in particular to a wafer automated production control system and method. Background Technology

[0002] As wafer manufacturing moves towards higher integration, faster turnaround times, and multi-process collaboration, production control is gradually evolving from single-equipment management to a comprehensive control system covering manufacturing execution, automated handling, equipment access, process monitoring, and quality feedback. Technologies such as real-time event stream processing, state mapping, joint scheduling, and virtual metrology have become important development directions for automated wafer production.

[0003] While existing technologies can achieve batch scheduling, equipment access, or quality monitoring separately, different control links often operate in a segmented manner, making it difficult to establish continuous and consistent state associations and feedback linkages around the same batch. This results in a lack of real-time closed-loop coupling between scheduling decisions, execution verification, and quality correction, making it difficult to support dynamic collaborative control in the wafer manufacturing process. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a wafer automation production control method to solve the problem that it is difficult to form a unified real-time closed-loop linkage control for multi-source control states during wafer production.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] In a first aspect, the present invention provides a wafer automated production control method, comprising: collecting production control data and performing unified time-stamped object mapping and event standardization processing to generate a standardized production event flow; integrating the standardized production event flow into a state association to construct a real-time production state graph, and performing process window residual value calculation and resource matching analysis to generate a candidate execution tuple set; based on the candidate execution tuple set, performing priority evaluation and conflict resolution to generate a preferred execution sequence, and matching and allocating target machines and transport paths in the preferred execution sequence to generate a joint scheduling result; arranging control instructions based on the joint scheduling result to generate a collaborative control instruction set, and performing arrival verification and processing permission processing to generate a verified execution result; performing processing tracking, virtual measurement confidence analysis, and sampling path adaptive adjustment on the verified execution result to generate a quality control result; and performing release judgment and recipe correction on the quality control result to obtain the release judgment result and recipe correction result, and writing them back to the real-time production state graph to generate a closed-loop control result.

[0008] As a preferred embodiment of the wafer automated production control method of the present invention, the specific steps for generating a standardized production event flow are as follows.

[0009] Collect order execution information, process execution information, equipment operation information, handling execution information, and quality feedback information, and aggregate and organize them according to a unified data interface to generate production control data;

[0010] The production control data is processed with unified time stamping, cross-source time sequence correction and object mapping to establish the correspondence between the same batch, the same equipment, the same handling task and the same quality record, and generate associated production control data.

[0011] The process involves standardizing related production control data, eliminating duplicate records, verifying conflict records, and connecting dependent event chains to generate a standardized production event flow.

[0012] As a preferred embodiment of the wafer automated production control method of the present invention, the specific steps for constructing the real-time production status map are as follows.

[0013] Order execution information and process execution information are extracted from the standardized production event flow to form a batch status event flow;

[0014] The batch status event stream is associated with the equipment operation information to form an equipment-associated status event stream;

[0015] Associate the equipment-related status event stream with the handling execution information to form a handling-related status event stream;

[0016] By linking and integrating the transport-related status event stream with quality feedback information, a real-time production status map can be constructed.

[0017] As a preferred embodiment of the wafer automated production control method of the present invention, the specific steps for generating the candidate execution tuple set are as follows.

[0018] Based on the real-time production status map, analyze the batch execution progress, equipment accessibility and handling connection status, calculate the process time window residual value, and generate the time window status map;

[0019] Perform target machine availability analysis, material handling path reachability analysis, and resource matching analysis on the time window status map to generate candidate resource matching results;

[0020] Based on the candidate resource matching results, associate the batch, target machine, and transport path to generate a set of candidate execution tuples.

[0021] As a preferred embodiment of the wafer automated production control method of the present invention, the specific steps for generating the joint scheduling result are as follows.

[0022] The candidate execution tuple set is evaluated based on delivery urgency, process window residual value, resource utilization, and quality risk level to generate a priority candidate execution tuple set.

[0023] The set of priority candidate execution tuples is subjected to resource conflict identification, execution timing conflict identification, conflict resolution, and elimination of infeasible combinations to generate the preferred execution sequence.

[0024] Based on the preferred execution sequence, target machines and transport paths are matched and allocated, and the matching and allocation results are organized according to the execution order to generate joint scheduling results.

[0025] As a preferred embodiment of the wafer automated production control method of the present invention, the specific steps for generating the verified execution result are as follows.

[0026] Extract batch scheduling relationships, target machine relationships, and transport path relationships from the joint scheduling results, perform scheduling logic parsing, and generate scheduling parsing results;

[0027] The scheduling and parsing results are used to arrange the logic for transport and execution, the logic for machine access, and the logic for processing preparation, and a set of collaborative control instructions is generated.

[0028] The collaborative control instruction set is subjected to arrival verification, object consistency verification, and processing permission processing to generate verified execution results.

[0029] As a preferred embodiment of the wafer automated production control method of the present invention, the specific steps for generating quality control results are as follows.

[0030] The process of processing verified execution results is tracked, and information on status changes and process trajectories during the execution phase is collected to generate process tracking execution results.

[0031] Process features are extracted from the process tracking execution results, and virtual measurement calculations and confidence analysis are performed to generate virtual measurement confidence results.

[0032] Based on the virtual measurement confidence results, the prediction reliability and quality fluctuation risk of the corresponding batch are identified, and the entity sampling ratio and sampling flow path are adjusted according to the preset sampling control rules. The adjusted entity sampling ratio and sampling flow path are integrated to generate the sampling path adjustment results.

[0033] The sampling path adjustment results and the virtual measurement confidence results are correlated and integrated to generate quality control results.

[0034] As a preferred embodiment of the wafer automated production control method of the present invention, the specific steps for obtaining the release determination result and the formula correction result are as follows.

[0035] The quality control results are used to identify the quality status and verify the release conditions, and release judgment results are generated.

[0036] Extract deviation information from quality control results and conduct attribution analysis to distinguish the sources of batch deviations, target machine deviations, and processing deviations, and generate formula correction analysis results.

[0037] Based on the formula correction analysis results, the formula correction parameters are calculated and a correspondence is established with the quality control results, real-time production status map, and relevant statuses in the time window status map to generate the formula correction results.

[0038] As a preferred embodiment of the wafer automated production control method of the present invention, the specific steps for generating the closed-loop control result are as follows.

[0039] The release judgment result and the formula modification result are linked and integrated to establish a linkage relationship between the batch release status and the subsequent process parameter adjustment status under the same batch object, and generate write-back control results.

[0040] The write-back control results are written back to the real-time production status map, and the batch release status, quality risk status and formula correction status are updated synchronously to generate a real-time updated production status map.

[0041] The real-time updated production status map is used as the status input for subsequent production control cycles. The release judgment results and formula correction results of the previous control cycle are incorporated into the process window residual value calculation, resource matching analysis and scheduling decision of the next control cycle to generate closed-loop control results.

[0042] Secondly, this invention provides a wafer automated production control system, comprising: an event flow generation module for collecting production control data and performing unified time-stamped object mapping and event standardization processing to generate a standardized production event flow; a state mapping module for integrating the standardized production event flow into a state association, constructing a real-time production state graph, and performing process window residual value calculation and resource matching analysis to generate a candidate execution tuple set; and a joint scheduling module for performing priority evaluation and conflict resolution based on the candidate execution tuple set, generating a preferred execution sequence, and scheduling the target machines in the preferred execution sequence. The system matches and allocates workstations and transport paths to generate joint scheduling results. The instruction verification and control module is used to arrange control instructions based on the joint scheduling results, generate a collaborative control instruction set, and perform arrival verification and processing permission processing to generate verified execution results. The quality control judgment and adjustment module is used to track the processing process, perform virtual measurement confidence analysis, and adaptively adjust the sampling path based on the verified execution results to generate quality control results. The closed-loop write-back module is used to make release judgments and formula corrections based on the quality control results, obtain release judgment results and formula correction results, and write them back to the real-time production status map to generate closed-loop control results.

[0043] The beneficial effects of this invention are as follows: by adjusting the entity sampling ratio and sampling flow path through virtual measurement confidence results, the characteristics of the processing process, the reliability of prediction, and the risk of quality fluctuations can be directly incorporated into the sampling control link, so that the sampling action changes dynamically with the batch quality status, realizing the synchronous coordination of sampling resource allocation and risk level, and maintaining the continuous connection between quality judgment results and subsequent release and formula modification processes, thereby improving the response accuracy, quality identification targeting, and control closed-loop stability in the production control process. Attached Figure Description

[0044] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0045] Figure 1 This is a flowchart of a wafer automated production control method.

[0046] Figure 2 This is a schematic diagram of an automated wafer production control system.

[0047] Figure 3 A flowchart for constructing a real-time production status map.

[0048] Figure 4 This is a flowchart for quality control and closed-loop feedback. Detailed Implementation

[0049] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0050] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0051] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0052] Reference Figures 1-4 As one embodiment of the present invention, this embodiment provides a wafer automated production control method, including the following steps:

[0053] S1. Collect production control data, perform unified time stamp object mapping and event standardization processing, and generate standardized production event flow.

[0054] S1.1 Collect order execution information, process execution information, equipment operation information, handling execution information, and quality feedback information, and aggregate and organize them according to a unified data interface to generate production control data.

[0055] Specifically, order execution information is collected from batch dispatch records, work order flow records, and delivery management records in the production execution platform, obtained by reading the task issuance time, current process progress, and delivery status of the corresponding batch; process execution information is collected from the process management platform and formula issuance records, obtained by extracting the process steps, formula parameters, and execution sequence corresponding to the current batch; equipment operation information is collected from equipment communication interfaces, equipment status reporting records, and equipment accessibility records, obtained by listening to equipment event reports and reading the equipment number, operating status, and accessibility status; handling execution information is collected from handling scheduling records, carrier transfer records, and arrival confirmation records, obtained by reading the handling task number, handling location, and arrival status; quality feedback information is collected from measurement records, anomaly handling records, and release review records, obtained by extracting measurement results, anomaly records, and release status; order execution information, process execution information, equipment operation information, handling execution information, and quality feedback information are aligned, formatted, and correlated according to a unified data interface to form corresponding data entries, and these corresponding data entries are aggregated into production control data.

[0056] S1.2 Perform unified time stamping, cross-source time sequence correction, and object mapping processing on production control data to establish the correspondence between the same batch, the same equipment, the same handling task, and the same quality record, and generate associated production control data.

[0057] Specifically, a unified time stamp is added to the order execution information, process execution information, equipment operation information, handling execution information, and quality feedback information in the production control data to form time-stamped production control data. Cross-source time-series correction is performed on records from different sources in the time-stamped production control data to eliminate the time deviation of collection and form time-consistent corrected production control data. Object mapping is performed on the batch number, equipment number, handling task number, and quality record number in the corrected production control data to establish a correspondence between the same batch, the same equipment, the same handling task, and the same quality record. The corrected production control data after establishing the correspondence is integrated into associated production control data.

[0058] S1.3 Perform event standardization, duplicate record elimination, conflict record verification, and sequential processing of dependent event chains on the associated production control data to generate a standardized production event flow.

[0059] Specifically, order execution information, process execution information, equipment operation information, handling execution information, and quality feedback information in the associated production control data are uniformly converted into corresponding event records. Following unified event description rules, the event occurrence time, event object identifier, event type, event status, and value fields are extracted for each of these categories. Information from different sources is mapped to event records with a unified field structure, forming standardized event-related production control data. Duplicate event records in this standardized data are merged and eliminated. Based on the correspondence of the same batch, the same equipment, the same handling task, or the same quality record, the event type, event occurrence time, and event status are compared. Duplicate event records with identical fields and overlapping times are merged. For event records with duplicate content but different sources, event records with more complete time and more continuous status are retained, forming... The deduplicated production control data is processed by verifying time conflicts, object conflicts, and state conflicts. The occurrence time of corresponding event records within the same batch is compared with the expected process sequence; event records with inconsistent sequences or overlapping time intervals are identified as time conflicts. The batch number, equipment number, handling task number, and quality record number in the event records are compared item by item with the object mapping relationship; event records with inconsistent object identifiers or non-unique mapping relationships are identified as object conflicts. The state values ​​in the event records are compared item by item with the allowed state change sequence in the process advancement rules, equipment succession rules, handling connection rules, and quality feedback rules; event records with state changes that do not conform to the corresponding rules are identified as state conflicts. Finally, the verified event records are chained together according to process advancement relationships, equipment succession rules, handling connection rules, and quality feedback relationships to generate a standardized production event flow.

[0060] S2. Perform state association and integration on the standardized production event flow, construct a real-time production state map, and perform process window residual value calculation and resource matching analysis to generate a set of candidate execution tuples.

[0061] S2.1 Extract order execution information and process execution information from the standardized production event flow to form a batch status event flow.

[0062] Specifically, event records corresponding to order execution information are selected from the standardized production event flow to form an order execution event set, and event records corresponding to process execution information are selected from the standardized production event flow to form a process execution event set. The order execution event set and the process execution event set are then associated according to the flow relationship of the same batch in the order execution event set and the process relationship in the process execution event set. Finally, the associated order execution event set and the process execution event set are integrated into a batch status event flow.

[0063] S2.2 Associate the batch status event stream with the equipment operation information to form an equipment-associated status event stream.

[0064] Specifically, the equipment number, operating status, and accessibility status corresponding to the batch status event stream are filtered from the equipment operation information to form a set of equipment operation information. According to the batch flow relationship and process correspondence in the batch status event stream, the set of equipment operation information is associated with the corresponding batch status event stream one by one to form a set of equipment associated events containing the correspondence between batch status information and equipment operation information. The set of equipment associated events is then integrated into a set of equipment associated status event streams.

[0065] S2.3. Associate the equipment-related status event stream with the handling execution information to form a handling-related status event stream.

[0066] Specifically, the handling tasks, handling locations, and arrival statuses corresponding to the equipment-associated status event streams are filtered from the handling execution information to form a handling execution information set. According to the batch flow relationship and equipment succession relationship in the equipment-associated status event streams, the handling execution information set is associated with the corresponding equipment-associated status event streams one by one to form a handling-associated event set containing batch status information, equipment operation information, and the correspondence between handling execution information. The handling-associated event set is then integrated into a handling-associated status event stream.

[0067] S2.4. Integrate and link the transport-related status event stream with quality feedback information to construct a real-time production status map.

[0068] Specifically, measurement results, anomaly records, and release statuses corresponding to the handling-related status event streams are filtered from the quality feedback information to form a quality feedback information set. According to the batch flow relationship, equipment connection relationship, and handling connection relationship in the handling-related status event streams, the quality feedback information set is associated with the corresponding handling-related status event streams one by one, forming a quality-related status set containing the correspondence between batch status information, equipment operation information, handling execution information, and quality feedback information. The quality-related status set is then integrated into a real-time production status map.

[0069] S2.5. Based on the real-time production status map, analyze the batch execution progress, equipment accessibility, and handling connection status, calculate the process time window residual value, and generate the time window status map.

[0070] Specifically, the process position, waiting status, and flow relationship corresponding to each batch are extracted from the real-time production status graph to form a batch execution progress set. The access status and occupancy status of each target device are extracted from the real-time production status graph to form a device accessibility set. The material handling task connection relationship and arrival status corresponding to each batch are extracted from the real-time production status graph to form a material handling connection set. For the first... For each batch, the residual value of the process window is calculated using the following expression:

[0071] ;

[0072] in, Indicates the first residual value of the process window for each batch. Indicates the first The allowed execution window for each batch, Indicates the first Waiting time for each batch, Indicates the first The access duration of each batch of devices Indicates the first The handling and connection time for each batch, Index indicating batch;

[0073] The residual value of the process time window is integrated with the corresponding batch execution progress set, equipment accessibility set, and handling connection set to generate a time window status map.

[0074] S2.6 Perform target machine availability analysis, transport path reachability analysis, and resource matching analysis on the time window status map to generate candidate resource matching results.

[0075] Specifically, the target machine candidate relationships, access times, and occupancy status for each batch are extracted from the time window status map to form the target machine availability analysis results. The transport start point, transport end point, and path connectivity relationships for each batch are also extracted from the time window status map to form the transport path reachability analysis results. The target machine availability analysis results and the transport path reachability analysis results are correlated according to the same batch and the same process. Furthermore, combined with the process time window residual value and resource occupancy relationships in the time window status map, the target machine candidate relationships and transport path candidate relationships for each batch are matched item by item. The estimated access time and estimated transport connection time for each matching relationship are calculated. For each batch, target machine, and transport path candidate relationship in the time window status map, calculate the total duration of the expected access time and the expected transport connection time, and compare it with the process time window residual value of the corresponding batch. When the total duration is less than or equal to the process time window residual value, it is determined that the current candidate relationship meets the process time window constraint. Verify whether the target machine occupancy relationship and the transport path occupancy relationship meet the execution timing requirements, retain the candidate relationships that meet the process time window constraint and execution timing requirements, form the resource matching analysis results, and integrate the target machine availability analysis results, transport path reachability analysis results, and resource matching analysis results to generate candidate resource matching results.

[0076] S2.7. Based on the candidate resource matching results, associate the batch, target machine and transport path to generate a set of candidate execution tuples.

[0077] Specifically, the candidate relationships of target machines and transport paths corresponding to each batch are extracted from the candidate resource matching results to form a batch-related candidate set. According to the process sequence corresponding to the same batch, the accessibility relationship of the target machine, and the reachability relationship of the transport path, the batch, target machine, and transport path in the batch-related candidate set are associated item by item to form an execution object association set. The association relationships in the execution object association set are checked for consistency and executability, and the association results that meet the requirements of the same batch flow, the target machine succession requirements, and the transport path connection requirements are retained to obtain the filtered execution object association set. The filtered execution object association set is then integrated into a candidate execution tuple set.

[0078] It should be noted that integrating state mapping, time window constraints, and resource matching into the candidate execution object generation stage in advance can identify and filter out infeasible combinations that do not meet process time window constraints, machine access conditions, or material handling access conditions before scheduling. This improves the executability of candidate tasks, enhances the coordination between batch flow and equipment access, and provides a unified and continuous state basis for subsequent priority evaluation and conflict resolution, making it easier to form a real-time closed-loop decision chain for wafer automation production control.

[0079] S3. Based on the set of candidate execution tuples, perform priority evaluation and conflict resolution to generate the preferred execution sequence, and match and allocate the target machines and transport paths in the preferred execution sequence to generate joint scheduling results.

[0080] S3.1. Evaluate the urgency of delivery, residual value of process window, resource utilization, and quality risk level of the candidate execution tuple set to generate a priority candidate execution tuple set.

[0081] Specifically, the delivery deadline, process window residual value, target machine occupancy relationship, material handling path occupancy relationship, and quality status relationship for each batch are extracted from the candidate execution tuple set to form an evaluation candidate set. The urgency of delivery is calculated based on the time difference between the delivery deadline and the expected completion time corresponding to the current process progress; the smaller the time difference, the higher the urgency of delivery. The time window constraint is calculated based on the difference between the process window residual value and the expected execution duration corresponding to the current candidate execution tuple; the smaller the difference, the higher the time window constraint. Resource consumption is calculated based on the proportion of the occupancy time corresponding to the target machine occupancy relationship and the occupancy time corresponding to the material handling path occupancy relationship within the current control cycle. The higher the resource occupancy rate, the higher the resource occupancy rate. The quality risk level is calculated based on the measurement deviation amplitude, the number of abnormal records, and the release status in the quality status relationship. The higher the measurement deviation amplitude, the more abnormal records, or the more abnormal the release status, the higher the quality risk level. The urgency of delivery, the degree of time window constraint, the degree of resource occupancy, and the quality risk level are integrated to form a multi-dimensional evaluation result. The multi-dimensional evaluation result is mapped to each candidate execution tuple in the candidate execution tuple set to form an associated candidate set with evaluation results. The priority relationship of each candidate execution tuple is integrated based on the associated candidate set with evaluation results to generate a priority candidate execution tuple set.

[0082] S3.2. Perform resource conflict identification, execution timing conflict identification, conflict resolution, and infeasible combination elimination on the priority candidate execution tuple set to generate the preferred execution sequence.

[0083] Specifically, the target machine occupancy relationship, transport path occupancy relationship, and execution order relationship corresponding to each candidate execution tuple are extracted from the priority candidate execution tuple set to form a conflict identification candidate set. Resource conflicts are identified according to the target machine occupancy overlap relationship and transport path occupancy overlap relationship within the same control cycle. Execution timing conflicts are identified according to the sequential connection relationship and execution time relationship between each candidate execution tuple to form a conflict marker candidate set. The conflict marker candidate set is conflict-resolved based on the priority relationship in the priority candidate execution tuple set. Candidate execution tuples that meet the batch flow requirements, target machine succession requirements, and transport path connection requirements are retained, while infeasible combinations with resource conflicts, execution timing conflicts, or invalid connection relationships are eliminated. The retained candidate execution tuples are arranged according to the execution order to generate the preferred execution sequence.

[0084] S3.3. Based on the preferred execution sequence, match and allocate the target machine and the transport path, and organize the matching and allocation results according to the execution order to generate a joint scheduling result.

[0085] Specifically, the batch order, target machine candidate relationship, and transport path candidate relationship corresponding to each candidate execution tuple are extracted from the preferred execution sequence to form a matching and allocation candidate set. According to the execution order in the preferred execution sequence, each batch is sequentially matched with the target machine and transport path that satisfy the succession relationship to form the matching and allocation result of the target machine and transport path. The matching and allocation result of the target machine and transport path is organized and integrated according to the batch execution order, the target machine access order, and the transport path connection order to generate a joint scheduling result.

[0086] S4. Arrange control instructions according to the joint scheduling results, generate a collaborative control instruction set, perform arrival verification and processing permission processing, and generate verified execution results.

[0087] S4.1 Extract batch scheduling relationships, target machine relationships, and transport path relationships from the joint scheduling results, and perform scheduling logic parsing to generate scheduling parsing results.

[0088] Specifically, the scheduling sequence, target machine allocation relationship, and transport path allocation relationship corresponding to each batch are selected from the joint scheduling results to form a scheduling relationship set. According to the connection relationship between the first and last batches in the process of process advancement, the succession relationship of target machines in the execution process, and the correspondence of transport paths in the flow process, the process sequence, target machine allocation sequence, and transport path connection sequence of each batch in the scheduling relationship set are analyzed in turn. The batch execution order, target machine access order, and transport path connection sequence obtained by analysis are integrated to form the scheduling analysis result.

[0089] S4.2. Arrange the transport execution logic, machine access logic, and processing preparation logic on the scheduling analysis results to generate a collaborative control instruction set.

[0090] Specifically, the transport execution logic is arranged according to the batch execution order and transport path connection order in the scheduling analysis results to form a transport execution relationship set. The machine access logic is arranged according to the target machine access order and batch execution order in the scheduling analysis results to form a machine access relationship set. The processing preparation logic is arranged according to the transport execution relationship set and the machine access relationship set to form a processing preparation relationship set. The transport execution relationship set, the machine access relationship set, and the processing preparation relationship set are integrated into a collaborative control instruction set.

[0091] It should also be noted that the material handling execution logic refers to determining the material handling sequence, material handling triggering conditions, and material handling connection relationships for each batch based on the batch execution order and material handling path connection order in the scheduling analysis results; the machine access logic refers to determining the machine access sequence, machine access conditions, and access correspondence between the batch and the target machine based on the target machine access order and batch execution order in the scheduling analysis results; and the processing preparation logic refers to determining the connection relationship, preparation condition relationship, and preparation correspondence before subsequent processing permission for each batch after completing material handling and machine access, based on the material handling execution relationship set and the machine access relationship set.

[0092] S4.3 Perform arrival verification, object consistency verification, and processing permission processing on the collaborative control instruction set, and generate verified execution results.

[0093] Specifically, based on the transport execution relationship and target machine access relationship in the collaborative control instruction set, the arrival status of each batch is verified to form an arrival verification result. The consistency relationship between corresponding objects is verified based on the batch execution relationship, target machine access relationship, and transport path connection relationship in the collaborative control instruction set to form an object consistency verification result. Following the same batch of objects, the same target machine, and the same execution sequence, the arrival verification result and object consistency verification result are matched one by one with the processing preparation relationship in the collaborative control instruction set. It is determined whether the arrival status meets the arrival requirements in the processing preparation relationship, whether the object consistency meets the continuation requirements in the processing preparation relationship, and whether the preceding preparation conditions in the processing preparation relationship have been met. When both the arrival verification result and the object consistency verification result meet the conditions corresponding to the processing preparation relationship, the corresponding batch is determined to have the processing permit conditions, and a permit pass record is formed. When any condition is not met, the corresponding batch is determined to not have the processing permit conditions, and a permit restriction record is formed. The permit pass record and permit restriction record are then integrated to form a processing permit result. Finally, the arrival verification result, object consistency verification result, and processing permit result are integrated into a verified execution result.

[0094] S5. Perform process tracking, virtual measurement confidence analysis, and adaptive adjustment of sampling paths on the verified execution results to generate quality control results.

[0095] S5.1. Track the processing of the verified execution results, collect status change information and process trajectory information during the execution phase, and generate process tracking execution results.

[0096] Specifically, the processing license relationship, target machine access relationship, and execution start and end relationship corresponding to each batch are extracted from the verified execution results to form an execution tracking base set. Based on the execution tracking base set, the status change information and process trajectory information corresponding to each batch during the execution stage are continuously collected to form a process tracking information set. According to the execution order of the same batch and the corresponding processing process of the same target machine, the process tracking information set is associated and integrated with the execution tracking base set to generate the process tracking execution result.

[0097] S5.2 Extract process features from the process tracking execution results, perform virtual measurement calculations and confidence analysis, and generate virtual measurement confidence results.

[0098] Specifically, process features reflecting state changes and process trajectory changes during the execution phase are extracted from the process tracking execution results to form a process feature set. Based on the state change features, trajectory fluctuation features, and execution stability features corresponding to each batch in the process feature set, feature values ​​corresponding to the target measurement parameters are extracted. The impact of each feature value on the target measurement parameters is calculated according to a preset virtual measurement mapping relationship. These impact values ​​are then summarized to obtain the predicted measurement results for each batch. The predicted measurement results for each batch are then integrated into virtual measurement results. Finally, the virtual measurement results are correlated with the execution stability relationships in the process tracking execution results for the same batch, the same execution phase, and the same target machine. Each virtual measurement result is compared with the trajectory continuity relationship to see if the trend of change is consistent with the state fluctuation trend in the execution stability relationship. The changes in the virtual measurement results are also checked to see if they match the direction and magnitude of change in the trajectory continuity relationship. When the virtual measurement results are consistent with the execution stability relationship and the trajectory continuity relationship, the prediction results of the corresponding batch are considered to have a high degree of reliability. When the virtual measurement results deviate from the execution stability relationship or the trajectory continuity relationship, the prediction results of the corresponding batch are considered to have a low degree of reliability. The judgment results of each batch are then integrated to form a reliability analysis result. The virtual measurement results and the reliability analysis results are then integrated to form a virtual measurement confidence result.

[0099] S5.3 Based on the virtual measurement confidence results, identify the prediction reliability and quality fluctuation risk of the corresponding batch, and adjust the entity sampling ratio and sampling flow path according to the preset sampling control rules. Integrate the adjusted entity sampling ratio and sampling flow path to generate the sampling path adjustment result.

[0100] Specifically, the predicted measurement results, credibility analysis results, and corresponding process characteristic relationships for each batch are read from the virtual measurement confidence results. The prediction reliability is extracted based on the consistency between the virtual measurement results and the execution stability and trajectory continuity relationships in the credibility analysis results. The higher the consistency between the predicted measurement results and the execution stability and trajectory continuity relationships, the higher the prediction reliability of the corresponding batch. Then, the quality fluctuation risk is extracted based on the variation range of the predicted measurement results between batches, the trajectory fluctuation degree of the same batch during the execution phase, and the deviation in the credibility analysis results. The greater the variation range of the predicted measurement results, the higher the trajectory fluctuation degree, or the more obvious the deviation, the higher the quality fluctuation risk of the corresponding batch. The prediction reliability and quality fluctuation risk corresponding to each batch are integrated to form a sampling decision set. Based on the prediction reliability and quality fluctuation risk in the sampling decision set and according to the preset sampling control rules, the entity sampling ratio and sampling flow path corresponding to each batch are adjusted to form a sampling adjustment set. The entity sampling ratio and sampling flow path in the sampling adjustment set are integrated according to the correspondence of the same batch to generate the sampling path adjustment result.

[0101] It should also be noted that the preset sampling control rules refer to the sampling judgment criteria and adjustment criteria set in advance during the wafer automated production control process. These criteria are used to control the physical sampling ratio and sampling flow path based on the virtual measurement confidence results.

[0102] S5.4. Link and integrate the sampling path adjustment results and the virtual measurement confidence results to generate quality control results.

[0103] Specifically, the sampling ratio and sampling flow path corresponding to each batch are extracted from the sampling path adjustment results to form a sampling execution information set. The corresponding predicted measurement results and confidence analysis results are read one by one from the virtual measurement confidence results according to the batch number. The predicted measurement results and confidence analysis results are paired and collected according to the correspondence of objects in the same batch. The predicted measurement results and confidence analysis results corresponding to each batch are organized into a set of measurement judgment items. All measurement judgment items are summarized to form a measurement judgment information set. The sampling execution information set and the measurement judgment information set are linked one by one according to the correspondence of objects in the same batch to form a quality correlation information set. The quality correlation information set is integrated into the quality control results.

[0104] It should be noted that directly coupling process tracking, virtual measurement confidence analysis, and sampling flow path adjustment into the same quality control chain transforms the virtual measurement confidence results from "predictive reliability indicators" into "sampling path control signals," further generating quality control results. This allows for dynamic adjustment of the physical sampling ratio and sampling flow path based on the predicted reliability and quality fluctuation risk of the corresponding batch, ensuring that sampling resource allocation is synchronized with batch quality risk and maintaining a continuous correlation between sampling results and subsequent quality judgment processes. This improves the real-time responsiveness of the quality control chain, the targeting of sampling allocation, and the stability of the connection between subsequent release and formula modification.

[0105] S6. Perform release judgment and formula correction on the quality control results, obtain release judgment results and formula correction results, and write them back to the real-time production status map to generate closed-loop control results.

[0106] S6.1. Identify the quality status and verify the release conditions of the quality control results, and generate the release judgment result.

[0107] Specifically, the predicted measurement results, reliability analysis results, entity sampling ratios, and sampling flow paths corresponding to each batch are extracted from the quality control results to form a quality judgment information set. Based on the correspondence between the predicted measurement results in the quality judgment information set and the preset quality thresholds, the quality range of each batch is determined. Then, combined with the reliability analysis results and entity sampling ratios, the normal quality status, warning quality status, or abnormal quality status of each batch is identified. The identification results corresponding to each batch are integrated to form a quality status identification result. The quality status identification result is verified against the sampling flow paths and corresponding batch execution requirements in the quality control results to form a release condition verification result. The quality status identification result and the release condition verification result are integrated to generate a release judgment result.

[0108] It should also be noted that the preset quality threshold is determined based on the target process requirements in the process execution information, the distribution of historical measurement results in the quality feedback information, and the tolerance standards of the corresponding process. For example, when the target value of the target measurement parameter is 100, 97 to 103 can be set as the quality qualified range, 95 to 97 and 103 to 105 can be set as the critical deviation range, and less than 95 or greater than 105 can be set as the abnormal deviation range.

[0109] S6.2 Extract deviation information from the quality control results and conduct attribution analysis to distinguish the sources of batch deviation, the sources of target machine deviation, and the sources of processing deviation, and generate formula correction analysis results.

[0110] Specifically, the predicted measurement results, reliability analysis results, entity sampling ratios, and sampling flow paths corresponding to each batch are extracted from the quality control results to form a deviation information set. The source of batch deviation is identified based on the quality change relationship between different batches in the deviation information set. The source of target machine deviation is identified based on the execution relationship between the deviation information set and the target machine. The source of processing deviation is identified based on the process characteristic relationship between the deviation information set and the process tracking execution results to form a deviation attribution information set. The source of batch deviation, target machine deviation, and processing deviation in the deviation attribution information set are integrated to generate formula correction analysis results.

[0111] S6.3 Calculate the formula correction parameters based on the formula correction analysis results, and establish a correspondence between them and the relevant states in the quality control results, real-time production status map, and time window status map to generate the formula correction results.

[0112] Specifically, the batch deviation sources, target machine deviation sources, and processing deviation sources corresponding to each batch are extracted from the formula correction analysis results to form a formula correction parameter set. Based on the formula correction parameter set, the formula correction parameters corresponding to each batch are calculated to form parameter calculation results. The parameter calculation results are then established with the relevant states in the quality control results, the relevant states in the real-time production status map, and the relevant states in the time window status map to form a parameter correspondence set. The parameter correspondence set is then integrated into the formula correction results.

[0113] S6.4. Link and integrate the release judgment result and the formula modification result, establish the linkage relationship between the batch release status and the subsequent process parameter adjustment status under the same batch object, and generate the write-back control result.

[0114] Specifically, the quality status and release status of each batch are extracted from the release judgment results to form a release information set. The formula correction parameters and subsequent process parameter adjustment status of each batch are extracted from the formula correction results to form a formula adjustment information set. The release information set and the formula adjustment information set are linked and integrated one by one according to the correspondence of the same batch objects to form a linkage control information set. The linkage control information set is then integrated into the write-back control result.

[0115] S6.5 Write the write-back control results back to the real-time production status map, and synchronously update the batch release status, quality risk status and formula correction status to generate a real-time updated production status map.

[0116] Specifically, release information, quality association information, and formula adjustment information corresponding to each batch are extracted from the write-back control results to form a status write-back information set. According to the correspondence of the same batch of objects in the real-time production status map, the status write-back information set is written into the real-time production status map one by one to form a status update set. Based on the status update set, the batch release status, quality risk status, and formula correction status corresponding to each batch are updated synchronously. The status update set after the status update is completed is integrated into a real-time updated production status map.

[0117] S6.6. Use the real-time updated production status map as the status input for subsequent production control cycles, and incorporate the release judgment result and formula correction result of the previous control cycle into the process window residual value calculation, resource matching analysis and scheduling decision of the next control cycle to generate closed-loop control results.

[0118] Specifically, the real-time updated production status map is incorporated into the subsequent production control cycle to form the status input set for the next control cycle. The release judgment results and formula correction results of the previous control cycle are correlated with the status input set of the next control cycle according to the same batch of objects to form a closed-loop correlation information set. Based on the closed-loop correlation information set, process window residual value calculation, resource matching analysis and scheduling decisions are carried out to form a closed-loop control information set. The closed-loop control information set is then integrated into the closed-loop control result.

[0119] This embodiment also provides a wafer automated production control system, including:

[0120] The event flow construction module collects production control data, performs unified time-stamped object mapping and event standardization processing, and generates a standardized production event flow. The state mapping module integrates the standardized production event flow into a state association, constructs a real-time production state map, calculates process window residual values ​​and performs resource matching analysis, and generates a set of candidate execution tuples. The joint scheduling module performs priority evaluation and conflict resolution based on the candidate execution tuple set, generates a preferred execution sequence, and matches and allocates target machines and transport paths in the preferred execution sequence to generate joint scheduling results. The instruction verification and control module arranges control instructions based on the joint scheduling results, generates a collaborative control instruction set, performs arrival verification and processing permission processing, and generates verified execution results. The quality control judgment and adjustment module tracks the processing process, performs virtual measurement confidence analysis and adaptive adjustment of sampling paths based on the verified execution results, and generates quality control results. The closed-loop write-back module performs release judgment and formula correction on the quality control results, obtains release judgment results and formula correction results, and writes them back to the real-time production state map to generate closed-loop control results.

[0121] In summary, this invention adjusts the entity sampling ratio and sampling flow path by using virtual measurement confidence results. This allows the characteristics of the processing process, the reliability of predictions, and the risk of quality fluctuations to be directly incorporated into the sampling control link. This enables sampling actions to dynamically change with the batch quality status, achieves synchronous coordination between sampling resource allocation and risk level, and maintains continuous connection between quality judgment results and subsequent release and formula modification processes. As a result, it improves the responsiveness, the specificity of quality identification, and the stability of the control loop in the production control process.

[0122] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for controlling automated wafer production, characterized in that: include, Collect production control data, perform unified time stamp object mapping and event standardization processing, and generate standardized production event streams; Standardized production event flows are correlated and integrated to construct a real-time production status graph. Process window residual values ​​are calculated and resource matching analysis is performed to generate a set of candidate execution tuples. Based on the candidate execution tuple set, priority evaluation and conflict resolution are carried out to generate the preferred execution sequence. The target machines and transport paths in the preferred execution sequence are matched and allocated to generate joint scheduling results. Based on the joint scheduling results, control instructions are arranged, a collaborative control instruction set is generated, and arrival verification and processing permission are performed to generate verified execution results. The system tracks the processing of verified execution results, performs virtual measurement confidence analysis, and adaptively adjusts the sampling path to generate quality control results. The quality control results are used to make release decisions and formula corrections, and the release decision results and formula correction results are written back to the real-time production status map to generate closed-loop control results.

2. The wafer automated production control method as described in claim 1, characterized in that: The specific steps for generating a standardized production event flow are as follows: Collect order execution information, process execution information, equipment operation information, handling execution information, and quality feedback information, and aggregate and organize them according to a unified data interface to generate production control data; The production control data is processed with unified time stamping, cross-source time sequence correction and object mapping to establish the correspondence between the same batch, the same equipment, the same handling task and the same quality record, and generate associated production control data. The process involves standardizing related production control data, eliminating duplicate records, verifying conflict records, and connecting dependent event chains to generate a standardized production event flow.

3. The wafer automated production control method as described in claim 2, characterized in that: The specific steps for constructing the real-time production status map are as follows. Order execution information and process execution information are extracted from the standardized production event flow to form a batch status event flow; The batch status event stream is associated with the equipment operation information to form an equipment-associated status event stream; Associate the equipment-related status event stream with the handling execution information to form a handling-related status event stream; By linking and integrating the transport-related status event stream with quality feedback information, a real-time production status map can be constructed.

4. The wafer automated production control method as described in claim 3, characterized in that: The specific steps for generating the candidate execution tuple set are as follows: Based on the real-time production status map, analyze the batch execution progress, equipment accessibility and handling connection status, calculate the process time window residual value, and generate the time window status map; Perform target machine availability analysis, material handling path reachability analysis, and resource matching analysis on the time window status map to generate candidate resource matching results; Based on the candidate resource matching results, associate the batch, target machine, and transport path to generate a set of candidate execution tuples.

5. The wafer automated production control method as described in claim 4, characterized in that: The specific steps for generating the joint scheduling result are as follows: The candidate execution tuple set is evaluated based on delivery urgency, process window residual value, resource utilization, and quality risk level to generate a priority candidate execution tuple set. The set of priority candidate execution tuples is subjected to resource conflict identification, execution timing conflict identification, conflict resolution, and elimination of infeasible combinations to generate the preferred execution sequence. Based on the preferred execution sequence, target machines and transport paths are matched and allocated, and the matching and allocation results are organized according to the execution order to generate joint scheduling results.

6. The wafer automated production control method as described in claim 5, characterized in that: The specific steps for generating the verified execution result are as follows. Extract batch scheduling relationships, target machine relationships, and transport path relationships from the joint scheduling results, perform scheduling logic parsing, and generate scheduling parsing results; The scheduling and parsing results are used to arrange the logic for transport and execution, the logic for machine access, and the logic for processing preparation, and a set of collaborative control instructions is generated. The collaborative control instruction set is subjected to arrival verification, object consistency verification, and processing permission processing to generate verified execution results.

7. The wafer automated production control method as described in claim 6, characterized in that: The specific steps for generating the quality control results are as follows: The process of processing verified execution results is tracked, and information on status changes and process trajectories during the execution phase is collected to generate process tracking execution results. Process features are extracted from the process tracking execution results, and virtual measurement calculations and confidence analysis are performed to generate virtual measurement confidence results. Based on the virtual measurement confidence results, the prediction reliability and quality fluctuation risk of the corresponding batch are identified, and the entity sampling ratio and sampling flow path are adjusted according to the preset sampling control rules. The adjusted entity sampling ratio and sampling flow path are integrated to generate the sampling path adjustment results. The sampling path adjustment results and the virtual measurement confidence results are correlated and integrated to generate quality control results.

8. The wafer automated production control method as described in claim 7, characterized in that: The specific steps for obtaining the release determination result and the formula correction result are as follows. The quality control results are used to identify the quality status and verify the release conditions, and release judgment results are generated. Extract deviation information from quality control results and conduct attribution analysis to distinguish the sources of batch deviations, target machine deviations, and processing deviations, and generate formula correction analysis results. Based on the formula correction analysis results, the formula correction parameters are calculated and a correspondence is established with the quality control results, real-time production status map, and relevant statuses in the time window status map to generate the formula correction results.

9. The wafer automated production control method as described in claim 8, characterized in that: The specific steps for generating the closed-loop control result are as follows. The release judgment result and the formula modification result are linked and integrated to establish a linkage relationship between the batch release status and the subsequent process parameter adjustment status under the same batch object, and generate write-back control results. The write-back control results are written back to the real-time production status map, and the batch release status, quality risk status and formula correction status are updated synchronously to generate a real-time updated production status map. The real-time updated production status map is used as the status input for subsequent production control cycles. The release judgment results and formula correction results of the previous control cycle are incorporated into the process window residual value calculation, resource matching analysis and scheduling decision of the next control cycle to generate closed-loop control results.

10. A wafer automated production control system, based on the wafer automated production control method according to any one of claims 1 to 9, characterized in that: This includes an event flow generation module, which collects production control data and performs unified time stamp object mapping and event standardization processing to generate standardized production event flows; The state mapping module is used to integrate the state associations of standardized production event flows, construct a real-time production state map, perform process window residual value calculation and resource matching analysis, and generate a set of candidate execution tuples. The joint scheduling module is used to perform priority evaluation and conflict resolution based on the candidate execution tuple set, generate the preferred execution sequence, and match and allocate the target machines and transport paths in the preferred execution sequence to generate the joint scheduling result. The instruction verification and control module is used to arrange control instructions according to the joint scheduling results, generate a collaborative control instruction set, perform arrival verification and processing permission processing, and generate verified execution results. The quality control and adjustment module is used to track the processing of verified execution results, perform virtual measurement confidence analysis, and adaptively adjust the sampling path to generate quality control results. The closed-loop write-back module is used to make release decisions and formula corrections based on the quality control results, obtain the release decision results and formula correction results, and write them back to the real-time production status map to generate closed-loop control results.

Images