A method and device for responding to a failure of a dual-arm robot

Abstract

The application discloses a kind of combination equipment dual-arm manipulator fault response method and device, wafer is established dual-arm manipulator combination equipment process-oriented Petri net model according to its processing path;Establish the label Petri net model of combination equipment;Scheduling algorithm obtains wafer processing initial scheduling scheme when no fault occurs;The instantaneous state of dual-arm manipulator is detected when fault occurs, and target library is sought, and controllable fault processing module is added on target library, and instantaneous state conversion at fault time is carried out through controllable fault processing module;Scheduling algorithm is recalculated, and steady-state scheduling scheme after fault occurs is obtained;If fault recovery is detected, controllable fault processing module is converted again, scheduling algorithm is recalculated, and steady-state scheduling scheme after fault recovery is obtained.The application realizes the steady-state scheduling of wafer processing after the fault of combination equipment dual-arm manipulator, avoids the shutdown of combination equipment caused by the fault of combination equipment, and improves wafer processing efficiency.

Description

Technical Field

[0001] This invention belongs to the field of semiconductor wafer processing, and specifically relates to a fault response method and device for a dual-arm robotic arm of a combined equipment. Background Technology

[0002] Semiconductor chip manufacturing is located upstream in the semiconductor industry chain, and improving wafer processing efficiency is of significant strategic importance. Wafer manufacturing is one of the most complex and advanced manufacturing industries today. During wafer manufacturing, a large number of semiconductor assembly devices are used. A typical semiconductor assembly device usually consists of a wafer loading module (LoadPort, LP), a vacuum lock module (LoadLock, LL), several processing modules (PMs), an alignment module (Aligner, AL), a cooling buffer module (Cooling Buffer, CL), and a single / dual-arm transfer robot (Transfer Robot, TR). It is a highly automated device that is reconfigurable, highly flexible, has a short processing cycle, and high space utilization. It has been proven that the scheduling and control of assembly devices is an NP-hard problem.

[0003] When using dual-arm robotic arms in a modular equipment, malfunctions of the robotic arms are common, especially when one of the arms malfunctions. In order to improve wafer processing efficiency and minimize the impact of the malfunction on wafer scheduling, the modular equipment needs to continue wafer processing scheduling without stopping. How to achieve a smooth transition between the two steady-state scheduling before and after the malfunction without violating the processing conditions of other modules has become a difficult problem to be solved in actual processing and production.

[0004] Semiconductor manufacturing scheduling systems are specialized shop floor scheduling systems. Currently, there are three main steady-state scheduling methods for shop floor scheduling systems after a fault: complete rescheduling, right-shift rescheduling, and affected process rescheduling. Complete rescheduling disregards the initial scheduling plan and reschedules all processes, leading to conflicts in wafer processing scheduling before and after the fault. Right-shift rescheduling assumes the assembly equipment is stopped, and manual intervention restores the system to its original state. The algorithm recalculates based on the system's processing status at the time of the detected fault. However, this method maximizes the time loss caused by the fault, causing wafer processing to stall. It may also cause some wafers to exceed their dwell time constraints in the processing cavity, resulting in wafer damage. Affected process rescheduling minimizes the impact of rescheduling on the original scheduling plan. Processing processes unaffected by the fault continue according to the original schedule, while affected processes are rescheduled. This method reduces the impact of the fault on wafer processing to some extent. However, due to the high real-time computational requirements of the algorithm during wafer processing, and the significant difficulty in finding affected processes when the wafer processing path is complex, this method is rarely used in practical scenarios. Summary of the Invention

[0005] To address the shortcomings of existing technologies and improve wafer processing efficiency by ensuring that wafer processing can continue based on the initial scheduling scheme even when one arm of a dual-arm robotic arm malfunctions, the present invention adopts the following technical solution:

[0006] A fault response method for a dual-arm robotic arm in a combined device includes the following steps:

[0007] Step S1: A process-oriented Petri net model of the dual-arm robotic arm assembly equipment is established based on the wafer's processing path. The time-based operation is performed on the storage locations and transitions in the model based on the assembly equipment in the production process. The goal of the POPN model scheduling is for all wafers in the wafer loading module to complete processing according to the predetermined processing path. The scheduling goal is reflected in the Petri net model as the token in the initial storage location is transferred to the target storage location after a series of transitions. The token represents the resource in the storage location in the Petri net.

[0008] Step S2: Establish a Petri net model for combined equipment labels and label the library of process-oriented Petri net model sub-objects;

[0009] Step S3: The scheduling algorithm obtains the initial scheduling scheme for wafer processing when no faults occur;

[0010] Step S4: Detect the instantaneous state of the dual-arm manipulator at the moment of failure. When one arm of the dual-arm manipulator fails, the corresponding manipulator resource label in the Petri net model changes. The location number of the resource with this label at the moment of failure is the target location to be processed by the fault handling module. Based on the fact that the resources in the target location have labels at the moment of failure, a controllable fault handling module is added to the target location. The instantaneous state transition at the moment of failure is performed through the controllable fault handling module.

[0011] Step S5: The scheduling algorithm is recalculated to obtain the steady-state scheduling scheme after the failure occurs;

[0012] Step S6: If fault recovery is detected, the controllable fault handling module performs a state transition again, the scheduling algorithm is recalculated, and a steady-state scheduling scheme after fault recovery is obtained.

[0013] Further, in step S1, a process-oriented Petri net model of the dual-arm robotic arm combination device is established. This model primarily targets combination devices with parallel processing modules in real-world scenarios. The sub-objects of the process-oriented Petri net model corresponding to the combination device include: a wafer loading module, a single / dual-arm robotic arm, a processing module, a vacuum lock module, and a cooling module. The wafers to be processed in the wafer loading module are represented by several tokens in the wafer loading module library; the number of available resources for the robotic arm is represented by several tokens in the robotic arm library; the capacity limits of the processing module, vacuum lock module, and cooling module are determined by the number of tokens in their respective capacity libraries; the transitions in the model are categorized as robotic arm wafer picking, robotic arm wafer placement, and robotic arm wafer exchange. There are three types of wafers. A group of wafers with the same processing path is a group of Pjobs. The processing steps of a group of Pjobs are represented by the wafer processing flow WPF = {m, p, s}, where m represents the number of wafers processed in a group of Pjobs, p represents the number of processing cavities of the parallel wafer loading module PM in the processing path, and s represents the number of processing cavities of the wafer loading module PM in a parallel path. Based on the different wafer processing paths, a process-oriented modeling Petri model N = (S, T; F, W, M0) is established, where S and T represent two finite non-empty and disjoint sets, S is the set of places, T is the set of transitions, F represents the directed arc from a place to a transition or from a transition to a place, W represents the weight of the directed arc, and M represents the number of resources in the place.

[0014] Furthermore, a time-delay Petri net model of the dual-arm robotic arm combination equipment is established. The time-delay Petri net model is represented by a triple (N,D,Q), where N represents the process-oriented Petri net model, D represents the set of time delays for all locations, and Q represents the set of time delays for all transitions. The locations and transitions in the model are timed according to the time consumed by each module of the combination equipment to perform relevant actions in the actual production process.

[0015] The time-delay Petri net model includes PTPN (Place Delay Petri Net) and TTPN (Transition Delay Petri Net). The time-delay Petri net model of the dual-arm robot combination equipment adds time factors to both the place and transition. The specific time of the place and transition is the time consumed by the corresponding actions of each module on the processing path on the same Pjob wafer.

[0016] Further, in step S2, the combined device tag Petri net model is represented by a triple (N, Σ, L), where N represents a process-oriented Petri net; Σ represents the tag set; and L represents a tagging function that can assign a tag to any resource in any storage location. The tagging function's function is to tag the tokens in each storage location in the initial state of the POPN model. The tag set includes wafer tags, robot tags, vacuum lock state transition tags, module capacity control tags, etc. The tagging function's function is to tag the tokens in each storage location in the initial state of the POPN model.

[0017] Furthermore, the scheduling algorithm in step S3 satisfies the transition-launch logic constraints of the timed Petri net model. The scheduling objective is to minimize the maximum completion time, which refers to the time required to return to LoadPort(LP) after completing the final process. Minimizing the maximum completion time is expressed as:

[0018] f = min(max(C) j ))

[0019] in, C j ≤C max , This represents the cumulative time taken for wafer j to complete each process after leaving wafer loading module LP; s represents the number of non-repeating processing steps for each wafer, which can also be understood as the number of non-repeating processing modules a single wafer needs to pass through to complete processing; m ik S represents the number of repetitions required for wafer k to complete the i-th process. ik r represents the processing time required for wafer k to complete the i-th process. ik T represents the additional module dwell time required for wafer k to complete the i-th process. TRrotation This indicates the time required for the robotic arms (ATR, VTR) to complete the entire wafer fabrication process. Both ATR and VTR represent the transfer robot (TR), but their positions differ. max This indicates the maximum time required to complete the processing of each group of wafers under extreme conditions (the next wafer only begins processing after the previous wafer has completed all processing steps);

[0020] In step S4, the scheduling objective after the combined equipment failure is to minimize the completion time deviation: after the dual-arm manipulator fails, a rescheduling generates a rescheduling plan to obtain a new makespan. The difference between the new makespan and the maximum global completion time obtained by the initial scheduling when no failure occurs is the completion time deviation, which can be expressed as:

[0021]

[0022] where C i ′ is the global completion time of wafer i after rescheduling, where i ≤ j.

[0023] The intelligent scheduling algorithm satisfies the principle of the earliest firing with the minimum transition time. For the same token, the algorithm determines the firing transition number according to the size of the global processing time calculated in real time after the transition fires. After a failure occurs, the failure transition in the failure handling module has the highest priority for firing.

[0024] The scheduling algorithm obtains the initial scheduling plan for wafer processing, and the combined equipment starts wafer processing according to the scheduling plan.

[0025] Furthermore, in step S4, the instantaneous states of the dual-arm manipulator are as follows: when a failure occurs in one of the manipulators, the failed manipulator does not hold a wafer, the failed manipulator holds an unprocessed wafer picked up from the vacuum chuck, the failed manipulator holds a wafer that has completed part of the processing operations picked up from the processing module, and the failed manipulator holds a wafer that has completed all processing operations picked up from the processing module;

[0026] When a failure occurs in one of the arms of the dual-arm manipulator, the combined equipment reports a fault-related error message "error" and transfers it to the algorithm. The corresponding manipulator resource label in the Petri net model changes from to The algorithm detects the place number where the resource with this label is located at the failure moment. This place is the target place to be processed by the failure handling module;

[0027] At the failure moment, the label H carried by the resource in the target place is divided into the following categories:

[0028] If it means that the manipulator does not hold a wafer when the failure occurs;

[0029] If it means that manipulator i is holding a wafer numbered j when the failure occurs, and the wafer is about to start processing;

[0030] If where 0 < k < G, it means that manipulator i is holding a wafer numbered j when the failure occurs, and the wafer has completed the processing of the k-th operation;

[0031] like This indicates that when the malfunction occurred, robotic arm i was holding wafer number j, and that wafer had already completed all processing steps in the processing cavity;

[0032] In step S6, when fault recovery is detected, the combined device reports fault repair information, which is transmitted to the algorithm. The corresponding robotic arm resource label in the Petri net model is then... Become At the time of fault recovery, the faulty robotic arm tag corresponding to the token is located in the fault buffer library in the fault handling module.

[0033] Further, in step S4, the fault handling module includes a fault handling repository and fault handling transitions. The fault handling repository includes a target repository and a fault buffer repository. The fault handling transitions include fault transitions and fault recovery transitions. If a fault is detected in one arm of the dual-arm manipulator, the fault transition allows launch, while the fault recovery transition prohibits launch. After a fault transition is launched once, it is prohibited to launch that transition again. The system state transition is complete, and the scheduling algorithm is recalculated. The latency at the target repository is the same as the latency added to that repository in the original model.

[0034] Further, in step S4, the target location is determined based on the instantaneous state after the fault occurs. The instantaneous state is represented by a tagged identifier M. The rules for generating the tagged identifier are as follows: Under identifier M, if transition t can occur, a new identifier is obtained from identifier M by transition t. The generation of M' follows these rules:

[0035]

[0036] Among them, M ′ This represents the new identifier generated after the transition launch, where M represents the original identifier before the transition launch, and p(m) i ,h i Let represent the location affected by this transition, i represent the location number before the transition, j represent the location number after the transition, m represent the distribution of resources (tokens) in the location, h represent the tag set of all resources (tokens) in the location, and W(p,t) / W(t,p) represent the weight function of the preceding or following transition t for location p. · t represents the set of preceding transitions of a place. · Denotes the set of post-transitions of places. · t · This represents the total set of preceding and following changes of the repository.

[0037] Furthermore, in step S6, when fault recovery is detected, the scheduling algorithm receives the fault signal, and the tag state in the robot resource in the corresponding Petri net model is changed; the transition recovery transition in the fault handling module is allowed to be transmitted once, the fault transition is prohibited from being transmitted, and after the fault recovery transition is transmitted once, the transition is prohibited from being transmitted again. After the system state transition is completed, the scheduling algorithm is recalculated to obtain the wafer processing steady-state scheduling scheme after fault repair.

[0038] A fault response device for a dual-arm manipulator of a combined device includes a memory and one or more processors. The memory stores executable code, and when the one or more processors execute the executable code, they are used to implement the fault response method for the dual-arm manipulator of a combined device.

[0039] The advantages and beneficial effects of this invention are as follows:

[0040] The combined equipment labeling POPN model established in this invention addresses the stability scheduling after a dual-arm robot failure. Compared to labeled Petri nets used in fault diagnosis, it further solves the problem of continuous wafer processing without stopping the combined equipment when a dual-arm robot fails. Instead of adding labels to transitions or places, the method of labeling tokens proposed in this invention avoids long label strings representing system states, reducing the complexity of state detection. Compared to using linear programming to obtain a rescheduling scheme after a dual-arm robot failure, this invention is more suitable for large-scale wafer processing scheduling and can be better applied to real-world production scenarios. Attached Figure Description

[0041] Figure 1 This is a flowchart illustrating the implementation of the fault response method for the dual-arm manipulator of the combined equipment in this embodiment of the invention.

[0042] Figure 2 This is a POPN model established using WPF = {6,2,1} as an example in this embodiment of the invention.

[0043] Figure 3 This is a labeled POPN model established using WPF = {6,2,1} as an example in this embodiment of the invention.

[0044] Figure 4 This is a schematic diagram of the controllable fault handling module in an embodiment of the present invention.

[0045] Figure 5 This is a Gantt chart of wafer fabrication obtained using an intelligent algorithm when WPF={6,2,1} is fault-free in an embodiment of the present invention.

[0046] Figure 6This is a Gantt chart of wafer processing obtained using an intelligent algorithm after a failure of the dual-arm robotic arm with WPF = {6,2,1} in this embodiment of the invention.

[0047] Figure 7 This is a schematic diagram of the structure of the fault response device for the dual-arm robotic arm of the combined equipment in this embodiment of the invention. Detailed Implementation

[0048] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0049] like Figure 1 As shown, a fault response method for a dual-arm robotic arm of a combined device according to the present invention includes the following steps:

[0050] Step S1: Establish a process-oriented Petri net (POPN) model for the dual-arm robotic arm combination device;

[0051] Step S2: Establish a Petri net model for combined device tags based on the POPN model;

[0052] Step S3: The scheduling algorithm obtains the initial scheduling scheme for wafer processing when no faults occur;

[0053] Step S4: Detect the instantaneous state of the system at the moment the fault occurs, locate the target storage location, add a controllable fault handling module to the target storage location, and the controllable fault handling module performs system state transition.

[0054] Step S5: The scheduling algorithm is recalculated to obtain the steady-state scheduling scheme after the failure occurs;

[0055] Step S6: If fault recovery is detected, the controllable fault handling module performs system state transition again, the scheduling algorithm is recalculated, and a steady-state scheduling scheme after fault recovery is obtained.

[0056] In this embodiment, the combined device comprises a wafer loading module (LoadPort, LP), a vacuum lock module (LoadLock, LL), several processing modules (PMs), a calibration module (Aligner, AL), a cooling buffer module (CoolingBuffer, CL), and a single / dual-arm transfer robot (TR).

[0057] In LP, multiple wafers with the same processing path are called a group of Pjobs. The processing path of a wafer in a group of Pjobs in the combined equipment is generally: LP→ATR→AL→LL→VTR→PM→VTR→PM→VTR→LL→CL→LP. If the wafer needs to be processed multiple times in PM, the VTR→PM action can be repeated multiple times according to the processing path.

[0058] A process-oriented Petri net model is established based on a set of Pjob processing paths. The processing steps of a set of Pjobs can be represented by the wafer processing flow WPF = {m, p, s}, where m refers to the number of wafers processed in a set of Pjobs, p refers to the number of parallel PM processing cavities in the processing path, and s refers to the number of PM processing cavities in a parallel path.

[0059] Furthermore, a process-oriented Petri net model of the dual-arm manipulator combination device is established, the POPN model N = (S, T; F, W, M0), where S and T are finite, non-empty, and disjoint sets, S is the set of places, T is the set of transitions, F is a directed arc representing a path from a place to a transition or vice versa, W is the weight of the directed arc, M is the number of tokens in a place, and a token represents a resource within a place in the Petri net. The number of tokens is the number of resources in a place. One token in a place means that the place has one resource at that time, and n tokens mean that the place has n resources at that time. Figure 2 In the p1 repository, there are 6 tokens. 18 The warehouse holds 3 tokens, p 23 The vault holds two tokens, p3, p5, p8, and p. 11 p 13 p 15 p 22 One token is held in the storage, and no tokens are held in other storages.

[0060] The established Petri model sub-objects include: wafer loading module, single / dual-arm robot, processing module, vacuum lock module, and cooling module. The wafers to be processed in the wafer loading module are represented by several tokens in the wafer loading module library; the number of available resources for the robot is represented by several tokens in the robot library; the capacity limits of the processing module, vacuum lock module, and cooling module are determined by the number of tokens in their respective capacity libraries; the transitions in the model are divided into three categories: robot picking up wafers, robot placing wafers, and robot exchanging wafers.

[0061] like Figure 2 As shown, the POPN model established with WPF={6,2,1} as an example has a wafer processing path corresponding to Pjob as: LP→LL→PM1 / PM2→LL→LP.

[0062] The specific information regarding the model structure is as follows:

[0063]

[0064]

[0065] In the model, the weight of each directed arc is 1, p 24 p 25 These correspond to the single-arm robot ATR and the dual-arm robot VTR, respectively, with the number of robot resources represented by tokens in the warehouses. The goal of the POPN model scheduling is for all wafers in the wafer loading module to complete processing according to the predetermined processing path. The realization of the scheduling goal in the Petri net model is that the tokens in the initial warehouses are transferred to the target warehouses after a series of transition sequences.

[0066] Based on the established combined dual-arm robotic arm device POPN, a corresponding time-delay Petri net model is constructed. The time-delay Petri net model can be represented by a triple (N,D,Q), where: N=(S,T;F,W,M0) is a standard Petri net, D={d1,d2,…,d… n Let} be the set of time delays for all locations, Q = {q1, q2, ..., q} n} represents the set of delay times for all transitions. The time consumed by each module of the combined equipment in the actual production process is used to assign time to the locations and transitions in the model. See the table below:

[0067]

[0068] The combined equipment tag Petri net model can be represented by a triple (N, ∑, L), where: N = (S, T; F, W, M0) is a standard Petri net; ∑ = {a, b, ...} is a set of labels; L: T → ∑{λ} is a labeling function that can affix a label ∑ or an empty label λ to any token in any storage location in the system. The label set includes wafer labels, robot labels, vacuum lock state transition labels, module capacity control labels, etc. The labeling function affixes labels to the tokens in each storage location in the initial state of the POPN model. The robot's operating state is divided into no fault and fault, represented by {safe, fault}, and indicated by robot token labels.

[0069] The POPN modeling algorithm for dual-arm robotic arm combination equipment labels:

[0070] Input: POPN model of a dual-arm robotic arm combination device;

[0071] Output: POPN model of the dual-arm robotic arm combination device label;

[0072] Step S2.1: For the library p corresponding to LP lp The token in the middle is labeled with a wafer label. i ,i∈M(p lp );

[0073] Step S2.2: For the library p corresponding to LL ll The tokens in the container are labeled with LL capacity limits.

[0074] Step S2.3: For the place p corresponding to PM pm The Tokens in the middle are labeled with PM capacity limits;

[0075]

[0076] Step S2.4: For the library p corresponding to CL cl The token in the middle is labeled with a CL capacity limit label. i ,i∈M(p cl );

[0077] Step S2.5: For the single-arm robotic ATR, specify the location p. atr The Token in the middle is labeled with the status tag of the robotic arm. safe ;

[0078] Step S2.6: For the storage location p corresponding to the dual-arm robotic arm VTR vtr The Token in the middle is labeled with the status of the robotic arm.

[0079] Step S2.7: Exit the algorithm.

[0080] like Figure 3 As shown, taking WPF={6,2,1} as an example, the token tags in each place of its initial identifier are defined in the following table:

[0081]

[0082] The relevant scheduling algorithm satisfies the transition-launch logic constraints of the timed POPN model. The scheduling objective of the algorithm is to minimize the maximum completion time, which refers to the time required to return to LoadPort(LP) after completing the final operation. Minimizing the maximum completion time can be expressed as:

[0083] f = min(max(C) j C j This represents the cumulative time taken for wafer j to complete each process after leaving LP.

[0084] C j ≤C max ,

[0085] The scheduling objective after a combined equipment failure is to minimize the completion time deviation: After the dual-arm manipulator fails, a rescheduling process generates a new maximum completion time. The difference between this new maximum completion time and the maximum global completion time obtained from the initial scheduling before the failure is resolved is the completion time deviation, which can be expressed as: Where C i ′ Let be the global completion time of wafer i after rescheduling, where i ≤ j.

[0086] An intelligent algorithm that satisfies the launch logic of the timed POPN model is used to obtain the initial scheduling scheme for wafer fabrication.

[0087] Taking an intelligent algorithm as an example, the core steps of the algorithm are:

[0088] Input: Petri net structure information, time information, and termination identifier;

[0089] Output: Petri net transition sequence and global processing time;

[0090] Step 1: Determine the enabling status of each transition under the initial flag, select one transition to launch, obtain the global processing time and flag after the transition launch, and add the relevant information to the priority queue;

[0091] Step 2: Select the transition launch with the shortest global processing time in the priority queue under the same identifier;

[0092] Step 3: Repeat steps 1 and 2. If the same identifier is encountered, backtrack the reachability graph and select the identifier with the shortest global processing time for transition emission operation.

[0093] Step 4: If the generated identifier is the same as the termination identifier, the algorithm exits.

[0094] like Figure 5 As shown in the figure, taking WPF={6,2,1} as an example, the initial wafer processing schedule obtained using the intelligent algorithm when no faults occur is shown in the figure.

[0095] During the transition launch process, the tokens and their tagged components in the quorum undergo state transitions according to the basic rules of Petri nets. The rules for generating the tagged identifiers are as follows: If M[t>M ′ This indicates that under the tagged identifier M, the emission transition t obtains a new tagged identifier M. ′ M ′ The generation follows these rules:

[0096]

[0097] Detect the instantaneous state of the system after a fault occurs. The main detection object is the state of the dual-arm manipulator at the time of the fault. Its instantaneous state can be divided into the following situations: when a fault occurs in one of the manipulators, the faulty manipulator does not hold a wafer, the faulty manipulator holds an unprocessed wafer picked up from the vacuum chuck, the faulty manipulator holds a wafer that has completed part of the processing process picked up from the processing module, and the faulty manipulator holds a wafer that has completed all processing processes picked up from the processing module.

[0098] When a fault occurs in one of the arms of the dual-arm manipulator, the combined equipment reports fault-related error information and transmits it to the algorithm. The corresponding arm token label in the Petri net model changes from to The algorithm detects the place number where the token with this label is located at the fault moment. This place is the target place that the fault handling module will process. At the fault moment, the label H carried by the token in the target place is divided into the following categories:

[0099] If It means that the manipulator does not hold a wafer when the fault occurs.

[0100] If It means that the manipulator i is holding a wafer numbered j when the fault occurs, and the wafer is about to start processing.

[0101] If where 0 < k < G, it means that the manipulator i is holding a wafer numbered j when the fault occurs, and the wafer has completed the processing of the k-th process.

[0102] If It means that the manipulator i is holding a wafer numbered j when the fault occurs, and the wafer has completed all the processes in the processing chamber.

[0103] [[ID=3:3]]According to the situation of the token in the target place carrying a label at the fault moment, a controllable fault handling module is added to the target place. The controllable fault handling module performs the conversion of the instantaneous state of the system at the fault moment.

[0104] As Figure 4 shown, the fault handling module consists of two fault handling places and two fault handling transitions. The fault handling places include the target place and a fault buffer place. The fault handling transitions are divided into fault transitions and fault recovery transitions. The time delay on the target place is the same as the time delay added to this place in the original model. The time delay on the fault buffer place is 0 s, and its place capacity limit is 1. The time delays added to the fault transitions and fault recovery transitions are 0 s.

[0105] A malfunction is detected in one arm of the dual-arm robotic arm. The combined equipment receives a fault signal, and the scheduling algorithm recalculates. At the time of the fault, a fault handling module is added to the target library. At the time of the fault, the fault transition has the highest transmission priority over other transitions; the fault transition is allowed to be transmitted, while the fault recovery transition is prohibited from being transmitted. After a fault transition is transmitted once, it is prohibited from being transmitted again. Once the system state transition is complete, the scheduling algorithm recalculates to obtain a rescheduling scheme after a wafer processing fault occurs. Figure 6 As shown, taking WPF={6,2,1} as an example, this is the wafer processing Gantt chart obtained using an intelligent algorithm after a dual-arm robotic arm malfunctions.

[0106] The scheduling objective after a failure is to minimize the completion time deviation: After the dual-arm manipulator fails, a rescheduling process generates a new maximum completion time. The difference between this new maximum completion time and the maximum global completion time obtained from the initial scheduling before the failure occurs is the completion time deviation, which can be expressed as:

[0107]

[0108] Where C i ′ Let be the global completion time of wafer i after rescheduling, where i ≤ j.

[0109] When fault recovery is detected, the combined equipment reports fault repair information, which is transmitted to the algorithm. The corresponding robot arm token label in the Petri net model is then... Become At the time of fault recovery, the faulty robotic arm tag corresponding to the token is located in the fault buffer library in the fault handling module.

[0110] In the fault handling module of the faulty robotic arm, the recovery transition is allowed to be launched once, while the faulty transition is prohibited from being launched. After the faulty recovery transition is launched once, it is prohibited from being launched again. Once the system state transition is complete, the scheduling algorithm is recalculated to obtain the steady-state scheduling scheme after fault repair. The fault handling module is then removed from the model. If the robotic arm malfunctions again before all wafers in the group have been processed, the above fault response is repeated.

[0111] Corresponding to the aforementioned embodiment of a fault response method for a dual-arm manipulator in a combined device, the present invention also provides an embodiment of a fault response device for a dual-arm manipulator in a combined device.

[0112] See Figure 7 The present invention provides a fault response device for a dual-arm manipulator of a combined device, comprising a memory and one or more processors. The memory stores executable code, and when the one or more processors execute the executable code, they are used to implement a fault response method for a dual-arm manipulator of a combined device as described in the above embodiment.

[0113] An embodiment of the fault response device for a combined dual-arm robotic arm of the present invention can be applied to any device with data processing capabilities, such as a computer. The device embodiment can be implemented through software, hardware, or a combination of both. Taking software implementation as an example, as a logical device, it is formed by the processor of any data processing device loading the corresponding computer program instructions from non-volatile memory into memory for execution. From a hardware perspective, such as... Figure 7 The diagram shown is a hardware structure diagram of any device with data processing capabilities, including the fault response device for a dual-arm robotic arm of the present invention. (Except for...) Figure 7 In addition to the processor, memory, network interface, and non-volatile memory shown, any data processing device in the embodiment may also include other hardware depending on the actual function of the data processing device, which will not be described in detail here.

[0114] The specific implementation process of the functions and roles of each unit in the above device can be found in the implementation process of the corresponding steps in the above method, and will not be repeated here.

[0115] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of the present invention according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0116] This invention also provides a computer-readable storage medium storing a program thereon, which, when executed by a processor, implements a fault response method for a dual-arm robotic arm of a combined device as described in the above embodiments.

[0117] The computer-readable storage medium can be an internal storage unit of any data processing device as described in any of the foregoing embodiments, such as a hard disk or memory. The computer-readable storage medium can also be an external storage device of any data processing device, such as a plug-in hard disk, smart media card (SMC), SD card, flash card, etc., equipped on the device. Furthermore, the computer-readable storage medium can include both internal storage units and external storage devices of any data processing device. The computer-readable storage medium is used to store the computer program and other programs and data required by the data processing device, and can also be used to temporarily store data that has been output or will be output.

[0118] 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 the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A fault response method for a dual-arm robotic arm in a combined device, characterized in that... Includes the following steps: Step S1: The wafer establishes a process-oriented Petri net model of the dual-arm robotic arm combination equipment based on its processing path, and performs time-based operations on the locations and transitions in the model based on the combination equipment in the production process. Step S2: Establish a Petri net model for combined equipment labels and label the library of process-oriented Petri net model sub-objects; Step S3: The scheduling algorithm obtains the initial scheduling scheme for wafer processing when no faults occur; Step S4: Detect the instantaneous state of the dual-arm manipulator at the moment of failure. When one arm of the dual-arm manipulator fails, the corresponding manipulator resource label in the Petri net model changes. The location number of the resource with this label at the moment of failure is the target location to be processed by the fault handling module. Based on the fact that the resources in the target location have labels at the moment of failure, a controllable fault handling module is added to the target location. The instantaneous state transition at the moment of failure is performed through the controllable fault handling module. Step S5: The scheduling algorithm is recalculated to obtain the steady-state scheduling scheme after the failure occurs; Step S6: If fault recovery is detected, the controllable fault handling module performs a state transition again, the scheduling algorithm is recalculated, and a steady-state scheduling scheme after fault recovery is obtained.

2. The fault response method for a dual-arm robotic arm in a combined device according to claim 1, characterized in that: In step S1, a group of wafers with the same processing path is a group of Pjobs, and the processing steps of a group of Pjobs are represented by a wafer processing flow (WPF). Based on the different wafer processing paths, a process-oriented modeling Petri model N = (S, T; F, W, M0) is established, where S and T represent two finite, non-empty, and disjoint sets, S is the set of places, T is the set of transitions, F represents the directed arc from a place to a transition or from a transition to a place, W represents the weight of the directed arc, and M represents the number of resources in the place.

3. The fault response method for a dual-arm robotic arm in a combined device according to claim 1, characterized in that: A time-delay Petri net model of a dual-arm robotic arm combination device is established. The time-delay Petri net model is represented by a triple (N,D,Q), where N represents the process-oriented Petri net model, D represents the set of time delays for all locations, and Q represents the set of time delays for all transitions. The locations and transitions in the model are timed according to the time consumed by each module of the combination device to perform relevant actions in the actual production process.

4. The fault response method for a dual-arm manipulator in a combined device according to claim 1, characterized in that: In step S2, the combined device tag Petri net model is represented by a triple (N,∑,L), where N represents a process-oriented Petri net; ∑ represents a tag set; and L represents a tagging function that can assign a tag to any resource in any location.

5. A fault response method for a dual-arm robotic arm in a combined device according to claim 1, characterized in that: The scheduling algorithm in step S3 satisfies the transition-launch logic constraints of the timed Petri net model. The scheduling objective is to minimize the maximum completion time, which refers to the time required to return after completing the last process. Minimizing the maximum completion time is expressed as follows: f=min(max(C j )) in, This represents the cumulative time taken for wafer j to complete each process after leaving the wafer loading module LP, s represents the number of non-repeating processing steps for each wafer, and m represents the total time taken. ik S represents the number of repetitions required for wafer k to complete the i-th process. ik r represents the processing time required for wafer k to complete the i-th process. ik T represents the additional module dwell time required for wafer k to complete the i-th process. TRrotation C represents the time required for the robotic arm to complete the entire wafer fabrication process. max This indicates the maximum time required to complete the processing of each group of wafers under extreme conditions; In step S4, the scheduling objective after a combined equipment failure is to minimize the completion time deviation: after the dual-arm manipulator fails, a rescheduling is performed to generate a new maximum completion time. The difference between this new maximum completion time and the maximum global completion time obtained from the initial scheduling when no failure occurred is the completion time deviation, which can be expressed as: Among them, C i ′ Let be the global completion time of wafer i after rescheduling, where i ≤ j.

6. The fault response method for a dual-arm manipulator in a combined device according to claim 1, characterized in that: In step S4, the instantaneous states of the dual-arm robotic arms are divided into: when one robotic arm malfunctions, the malfunctioning robotic arm does not hold a wafer; the malfunctioning robotic arm holds an unprocessed wafer held by the vacuum lock; the malfunctioning robotic arm holds a wafer that has completed part of the processing steps held by the processing module; and the malfunctioning robotic arm holds a wafer that has completed all processing steps held by the processing module. When one arm of the dual-arm robotic arm malfunctions, the combined equipment reports a fault-related error message, which is then passed to the algorithm. The corresponding robotic arm resource label in the Petri net model is then... Become The algorithm detects the location number of the resource with the tag at the time of the fault, and this location is the target location that the fault handling module will process. At the time of failure, the tags H of the resources in the target library fall into the following categories: like This indicates that the robotic arm was not holding a wafer when the malfunction occurred; like This indicates that when the malfunction occurred, robotic arm i was holding wafer number j, and that wafer was the beginning of the processing phase. If where \(0 < k < G\), indicating that when the fault occurs, the robotic arm \(i\) is grasping the wafer numbered \(j\), and this wafer has completed the processing of the \(k\)-th process; like This indicates that when the malfunction occurred, robotic arm i was holding wafer number j, and that wafer had already completed all processing steps in the processing cavity; In step S6, when fault recovery is detected, the combined device reports fault repair information, which is transmitted to the algorithm. The corresponding robotic arm resource label in the Petri net model is then... Become 7. A fault response method for a dual-arm robotic arm in a combined device according to claim 1, characterized in that: In step S4, the fault handling module includes a fault handling repository and a fault handling transition. The fault handling repository includes a target repository and a fault buffer repository. The fault handling transition includes a fault transition and a fault recovery transition. If a fault is detected in one arm of the dual-arm manipulator, the fault transition allows the launch, and the fault recovery transition prohibits the launch. After the fault transition is launched once, it is prohibited to launch the transition again. The system state transition is completed, and the scheduling algorithm is recalculated.

8. A fault response method for a dual-arm robotic arm in a combined device according to claim 1, characterized in that: In step S4, the target location is determined based on the instantaneous state after the fault occurs. The instantaneous state is represented by a tagged identifier M. The rules for generating the tagged identifier are as follows: Under identifier M, if transition t can occur, a new identifier is obtained from identifier M after transition t. The generation of M' follows these rules: Where M′ represents the new identifier generated after the transition launch, M represents the original identifier before the transition launch, and p(m i ,h i Let represent the location affected by this transition launch, i represent the location number before the transition launch, j represent the location number after the transition launch, m represent the distribution of resources in the location, h represent the tag set of all resources in the location, and W(p,t) / W(t,p) represent the weight function of the preceding or following transition t for location p. · t represents the set of preceding transitions of a place. · Denotes the set of post-transitions of places. · t · This represents the total set of preceding and following changes of the repository.

9. A fault response method for a dual-arm robotic arm in a combined device according to claim 1, characterized in that: In step S6, when fault recovery is detected, the scheduling algorithm receives the fault signal, and the tag state in the robot resource in the corresponding Petri net model is changed; the transition recovery transition in the fault handling module is allowed to be transmitted once, the fault transition is prohibited from being transmitted, and after the fault recovery transition is transmitted once, the transition is prohibited from being transmitted again. After the system state transition is completed, the scheduling algorithm is recalculated to obtain the wafer processing steady-state scheduling scheme after fault repair.

10. A fault response device for a dual-arm robotic arm in a combined equipment, characterized in that, The device includes a memory and one or more processors, wherein the memory stores executable code, and the one or more processors execute the executable code to implement a fault response method for a dual-arm manipulator of a combined device according to any one of claims 1-9.

Images