Scheduling method for single processing path of wafer

Through the ROPN model and deadlock processing constraints, robot scheduling in semiconductor chip manufacturing is planned, deadlock and wafer processing path optimization problems are solved, and system production capacity and product quality are improved.

CN118983249AActive Publication Date: 2024-11-19XIANGNAN UNIV

Patent Information

Application Number
CN202411060489.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-05
Publication Date
2024-11-19
Estimated Expiration
2044-08-05

AI Technical Summary

Technical Problem

In the semiconductor chip manufacturing process, it is difficult for the prior art to effectively avoid deadlocks and optimize the wafer processing path, resulting in the failure of system production capacity to fully utilize and the quality of wafer products is affected.

Method used

The scheduling algorithm is constructed using the ROPN model, and the deadlock processing constraints of the main robot and the pick-and-place robot are planned, and the scheduling strategy of the process production area is ensured to cooperate with each other, avoid deadlocks and optimize the wafer processing path. .

Benefits of technology

The coordinated control of the robot is realized, the deadlock is avoided, the system production capacity is fully utilized, the impact of wafer product quality is reduced, and the processing efficiency and capacity utilization are improved.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118983249A_ABST
    Figure CN118983249A_ABST
Patent Text Reader

Abstract

The invention discloses a scheduling method for a single processing path of a wafer, which belongs to the technical field of semiconductor chip manufacturing, and comprises the following steps: constructing an ROPN model; establishing a constraint condition for deadlock processing of the main manipulator, and establishing a process production area scheduling algorithm for the main manipulator by utilizing an ROPN model according to the constraint condition for deadlock processing of the main manipulator; establishing a constraint condition for deadlock processing of the pick-and-place manipulator, and establishing a wafer dispatching strategy for the pick-and-place manipulator according to the constraint condition for deadlock processing of the pick-and-place manipulator by using an ROPN model; and establishing a wafer feeding strategy for the pick-and-place manipulator by using the ROPN model according to the constraint condition of the deadlock processing of the pick-and-place manipulator. The scheduling method of the single processing path of the wafer solves the problem that the coordination of the whole operation is difficult to guarantee by the existing scheduling strategy.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of semiconductor chip manufacturing, and in particular to a scheduling method for a single processing path of a wafer. Background Art

[0002] The essence of semiconductor chip manufacturing is the manufacturing of wafers. The complexity of the wafer flow determines the difficulty of the wafer in the manufacturing process. The wafer manufacturing system (SWFS) is classified as a third type of manufacturing system in the manufacturing field, different from the flow shop and the job shop. Kumar has conducted an in-depth study of the unique properties of SWFS and pointed out the complexity it faces in optimizing scheduling. Figure 1 As shown in the figure, the production layout can be divided into two areas, namely the material (wafer) merging area and the process production area. The material merging area has two main responsibilities, one is to control the flow of wafers in this area, and the other is to merge two wafers with the same wafer flow mode into a combined material; while the process production area is responsible for wafer processing according to the requirements of the wafer flow.

[0003] According to the above layout diagram, the functions of each module in the workshop are as follows:

[0004] The entrance station is the entrance for wafers to enter the system; the exit station is the exit for wafers to leave the system; O1 is the inspection station responsible for inspecting the wafers to be processed; O2 is the inspection station responsible for inspecting the processed wafers; the storage position is a module for storing unfinished wafers and finished wafers, with limited capacity; the rotating table is a module for rotating wafers, rotating the wafers to a reasonable angle; the merging station merges two wafers into a group of materials; the pick-and-place robot R1 is responsible for picking up and placing wafers, and can interact with O1, the storage position, the rotating table, and the pick-and-place robot R2; the pick-and-place robot R2 is responsible for picking up and placing wafers, and can interact with the rotating table, the merging station, and the pick-and-place robot R1; it is adjacent to the processing station P i-1 and P i , where i = 2, 4, ..., 2N, N∈N * , responsible for processing materials, a total of N units, with a capacity of 2 wafers, and a residence time requirement, that is, the materials need to be taken out within a specified time after completion, N * is a positive integer. The adjacent processing stations can be regarded as a processing unit. The order in which the materials pass must be the smaller numbered ones first and then the larger numbered ones. There will not be a process path that only passes through one of the adjacent processing stations. For example, the adjacent processing stations P1 and P2 can be regarded as a processing unit; the non-adjacent processing station P N+j , where j = 1, 2, ..., D, D ∈ N *, a module responsible for processing materials, with a total of D units and a capacity of 2 wafers. There is a residence time requirement, that is, the materials need to be taken out within the specified time after completion, and N is the upper limit value of the adjacent processing station; the main robot R3, the "main robot" of the processing area, is responsible for taking and placing materials, and can interact with the merging station, storage position, processing station, and cleaning station. The material transportation between different processing units is the responsibility of the only main robot R3; the sub-robot r, the "sub-robot" of the processing area, is responsible for the movement of materials between adjacent processing stations, with a capacity of 2 materials. Taking the sub-robot r1 as an example, when the material is completed at the processing station P1, the sub-robot r1 moves the material to the processing station P2; the cleaning station is responsible for cleaning the module of the main robot R3. When the main robot R3 enters two designated processing stations before and after, it needs to go to the cleaning station for cleaning before entering the next processing station.

[0005] At the initial stage of the production process, the wafers must first pass through the merging station. After the wafers to be processed enter the system, they are inspected at O1. Only qualified wafers can be taken out by the pick-and-place robot R1 and placed in the storage position, waiting for subsequent operation instructions. After receiving the start command, the pick-and-place robot R1 will take out two wafers of the same type of wafer flow and two unfinished wafers from the storage position and send them to the turntable in turn. The turntable rotates to a specified angle, and the pick-and-place robot R1 will then place the wafers and wafers in the merging station in turn to merge them into a combined material for subsequent processing.

[0006] Subsequently, the robot r sends the combined material to the corresponding processing station for processing according to its processing path. When the material completes the entire process flow, the auxiliary robot r will send the completed wafer back to the merging station, where the wafer is disassembled into two completed wafers and wafers. The two wafers are returned along the original route by the pick-and-place robot R2 and the turntable, and then the pick-and-place robot R2 puts them back to the storage position from the turntable. Finally, the pick-and-place robot R2 receives a recovery command and moves the completed wafer from the storage position to O2, and the wafer finally leaves the system.

[0007] The core link of semiconductor chip manufacturing lies in the production of wafers. The processing path of wafers is also called wafer flow pattern, also known as wafer flow. Wafer flow is usually divided into four modes: serial, parallel, re-entrant and hybrid. Figure 2 The wafer flow without reentry can be symbolically described as (m1, m2, ..., m n ), where n represents the number of process steps that the wafer needs to complete, i represents the i-th process step, and m i It means that there are m steps in the i-th process (1≤i≤n). i parallel processing stations. When m1=m2=…=m n =1, the wafer flow is in serial mode. The wafer flow is in parallel mode. Additionally, a reentrant wafer flow can generally be expressed as {m1,..., [(m i ) b ,..., m e ,..(m j ,..., m g ) k ,..., m n}, 1 < i < e < j ≤ g < n. Among them, if reentry occurs at a certain step, [(m i ) b ,... m e means that after the wafer completes the e-th step, the wafer will be moved by the robot to the processing module of the i-th step for b times of repeated processing, and then the e + 1-th step of processing will be carried out. If some consecutive steps are repeated, (m j ,..., m g ) k represents that after the wafer is completed at the (j - 1)-th step, the wafer will be repeatedly processed k times from the j-th step to the g-th step. After the repeated processing is completed, the robot will move the wafer to the (g + 1)-th step, and then the wafer will continue to be processed step by step.

[0008] The existing reentrant wafer flow pattern is as follows: [(m1, m2),..., (m i-1 , m i ) k ,...,(m N-1 , m N )... m N+j ,..., m N+g ,..., m N+M , where i = 2, 4,..., 2n, n ∈ N * , j = 1, 2,..., m, m ∈ N * , 1 ≤ i < N, 1 ≤ j ≤ g < M, that is, the wafer has N + M processing steps. There are N adjacent processing stations P i and M non-adjacent processing stations P N+j in the system. The reentry times b and k values are determined by the wafer processing technology.

[0009] At the beginning of system operation, wafers are put into the system according to the established feeding plan to ensure that the capacity limit of the storage position is not exceeded. During this process, the pick-and-place robot R1 and the pick-and-place robot R2 follow the preset dispatching strategy and are responsible for transporting the wafers in the storage position without causing system deadlock. At the same time, the main robot R3 ensures the smooth progress of the wafer processing flow through scheduling, and meets various constraints to effectively prevent deadlock during the processing. At the same time, the action of the main robot R3 will be affected by the dispatching strategy, and the control of the pick-and-place robot R1, the pick-and-place robot R2 and the main robot R3 needs to cooperate with each other to ensure the coordination of the overall operation. Therefore, the problem of how to effectively control the pick-and-place robot R1, the pick-and-place robot R2 and the main robot R3 arises in order to fully utilize the system's production capacity and minimize the impact on the quality of wafer products while avoiding deadlock and various constraints. Summary of the invention

[0010] In order to overcome the defects of the prior art, the present invention provides a scheduling method for a single processing path of a wafer to solve the above-mentioned problems.

[0011] The technical solution adopted by the present invention to solve the technical problem is: a scheduling method for a single processing path of a wafer, comprising the following steps:

[0012] S0: Construct the ROPN model using the inspection station, storage position, rotation station, merging station, pick-and-place robot R1, pick-and-place robot R2, processing station, main robot R3, auxiliary robot r and cleaning station;

[0013] S1: Establish the constraint conditions for the deadlock processing of the main robot, and use the ROPN model and establish the process production area scheduling algorithm for the main robot according to the constraint conditions for the deadlock processing of the main robot;

[0014] S2: Establish the constraint conditions for the deadlock processing of the pick-and-place robot, and use the ROPN model and the constraint conditions for the deadlock processing of the pick-and-place robot to establish the wafer dispatching strategy for the pick-and-place robot;

[0015] S3: Using the ROPN model and based on the constraints of the pick-and-place robot deadlock handling, a wafer feeding strategy for the pick-and-place robot is established.

[0016] The beneficial effect of the present invention is that in the scheduling method of the single processing path of the wafer, the constraint conditions of the deadlock processing of the main robot are used to plan the process production area scheduling algorithm for the main robot, and the constraint conditions of the deadlock processing of the pick-and-place robot are used to plan the wafer dispatching strategy for the pick-and-place robot and the wafer feeding strategy for the pick-and-place robot, which can ensure that the control needs of the pick-and-place robot R1, the pick-and-place robot R2 and the main robot R3 are coordinated with each other, so as to coordinate the overall operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 A workshop layout for multi-agent chip manufacturing;

[0018] Figure 2 Schematic diagram of three wafer flow modes;

[0019] Figure 3 It is a flow chart of a scheduling method for a single processing path of a wafer in one embodiment of the present invention;

[0020] Figure 4 A structural diagram of a ROPN model in one embodiment of the present invention;

[0021] Figure 5 The solution process of the algorithm PAFR in one embodiment of the present invention;

[0022] Figure 6 The following is a flow chart of solving the PAPR of an algorithm in one embodiment of the present invention;

[0023] Figure 7 The solution process of the algorithm PSFW in one embodiment of the present invention;

[0024] Figure 8 The solution process of the algorithm PSUW in one embodiment of the present invention;

[0025] Fig. 9 The present invention provides an overall framework for combining heuristic algorithms in one embodiment of the present invention. DETAILED DESCRIPTION

[0026] The specific embodiments of the present invention are further described below in conjunction with the accompanying drawings. It should be noted that the description of these embodiments is used to help understand the present invention, but does not constitute a limitation of the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0027] As a unique third-class manufacturing system, the wafer manufacturing system (SWFS) has the following main features:

[0028] (1) Complex manufacturing process: The wafer production process and its complexity are mainly reflected in two core aspects: one is the diversity of wafer flow patterns, and the other is that multiple different types of wafer flows are mixed together for processing.

[0029] (2) Reentrant processing: Semiconductor wafers are hierarchical structures. The processing methods between layers are similar in basic principles. Each layer will be processed in several designated areas, and the processing between layers is only adjusted in terms of processing accuracy or processing environment. At the same time, the processing equipment is expensive and it is impossible for companies to purchase it in large quantities. Therefore, the same equipment can be used to process different types of products with similar structures, thus creating a "reentrant" processing method. The reentrant property is the most essential feature of wafer manufacturing processing, which is an important feature that distinguishes it from other production systems. When products of multiple wafer flows are mixed for production, different types of wafers will compete for the same equipment resources. In addition, even for the same type of wafers, multiple process steps may be performed on the same equipment, which may lead to resource competition due to the similarity of the process.

[0030] (3) Strict residence time constraint: The residence time refers to the time it takes for the wafer to leave the processing station after the processing module completes the current processing. During wafer processing, the wafer must leave the processing module within a certain time range. For example, the low pressure chemical vapor deposition (LPCVD) process requires that the wafer must be unloaded from the processing module after completion within a very short time, otherwise the wafer will be damaged or even scrapped.

[0031] (4) Deadlock: In the wafer manufacturing process, system deadlock needs to be taken seriously. When resources such as buffers between reentering processing equipment and adjacent equipment are competed for by wafers, if there is no efficient scheduling and control, deadlock may occur in the allocation of limited resources. Once deadlock occurs, the operation of the system will be permanently restricted.

[0032] (5) Uncertain production conditions: There are many uncertain factors in wafer production, such as equipment failure and maintenance, changes in customer orders, fluctuations in processing time, automatic equipment cleaning, new product trials, temporary process changes, rework, and changes in customer demand, etc. These factors have caused wafer processing to be in an unstable production environment, affecting the normal operation of the system.

[0033] The implementation of the solution of the present invention is described based on the following premises:

[0034] 1) The capacity of one storage position is 1 wafer, one position can only store one wafer, and one wafer can only occupy one position; 2) The capacity of the rotating table is 1 wafer; 3) The capacity of the merging station is 2 wafers; 4) Robot R3 can operate one combined material, i.e. two wafers, at a time, and other robots can only operate one wafer at a time; 5) Each processing station can only process one combined material at a time, and one combined material can only be processed by one processing station; 6) None of the above equipment resources will fail; 7) The cleaning time of robot R3 is much less than the range of the residence time constraint; 8) Each robot will have a corresponding rotation time when operating, but the rotation time is relatively short, much shorter than the processing time of the equipment; 9) The distance between each processing station is a constant and will not exceed the movement range of the robot.

[0035] The ROPN model has elements such as transition, place, token and directed arc, which form a directed graph. Transition is represented by a rectangle “▍”, place is represented by a circle “○”, token is represented by “●”, directed arc is represented by A ROPN model can be defined by a six-tuple, represented by: Σ = (P, T, I, O, M, K), where: (1) The finite set of places is: P = {p1, p2, ..., p n}, n>0; (2) The finite set of transitions is: T = {t1, t2, ..., t m}, m>0; (3) I: Directed arc from transition to place; (4) O: Directed arc from place to transition; (5) M: Token number identifier of place, M0 is the initial identifier of the system, for example, M(p i ) = f, that is, place p i There are f tokens in it; (6) K:K(p i ) represents the library place p i The maximum number of tokens that can be accommodated.

[0036] In the ROPN model established in this paper, the process production area is firstly targeted, which is adjacent to the processing station P i-1 and P i , not adjacent to processing station P N+j , main robot R3, auxiliary robot r i-1 and cleaning station P w All belong to the library, among which K(Pi )=K(P i-1 )=m i , K(P N+j )=m N+j , K(R3)=K(r i-1 )=2; When there are two tokens in the library, it means that the robot is currently in an available state. The token can be understood as whether there is a wafer in the robot. When there is a token, it means that the robot is empty and can process wafers. When there is no token, it proves that the robot is processing wafers. The number of tokens represents the number of wafers that the robot can process in this processing. Cleaning station P w , because the wafer will not enter the cleaning station, only the main robot R3 will be cleaned here, so let K(P w )=1, that is, cleaning station P w There is only one state change, which is occupied by the main robot or idle (the idle state is generally ignored), K(P w )=1, it indicates that the main robot is performing a cleaning operation.

[0037] like Figure 4 As shown, the transition X0 indicates that the main robot R3 rotates to the merging station (with P h Indicated), using transition X w Indicates that the main robot R3 moves to the cleaning station (with P w Indicated), using transition X i-1 Indicates the auxiliary manipulator r i-1 Move to process step i-1, transition X i and Transition X N+j They represent the main robot R3 turning to process step i and process step N+j respectively. h The waiting time of the main robot R3 before unloading the wafer is expressed as W w Indicates that the main robot R3 has arrived at the cleaning station P w Waiting time before, using library W i-1 Indicates the auxiliary robot r before unloading the wafer from process step i-1 i-1 The waiting time of the library W i and W N+j represents the waiting time of the main robot R3 before unloading the wafer from the process step i and from the process step N+j. The transition U0 represents the waiting time of the main robot R3 from the merging station P h Unload the wafer using Transition U w Indicates that the main robot R3 enters the cleaning station P w Cleaning, using the U i-1 Indicates the auxiliary manipulator r i-1 Unload the wafer from process step i-1 and use transition U i and ChangeN+j Respectively represent the auxiliary manipulator r i-1 Unload wafers from step i and step N+j. Use library q0 to represent the main robot R3 from the merge station P. h The waiting time before moving to process step i is calculated using the library Q i-1 Indicates the auxiliary manipulator r i-1 The waiting time before moving from process step i-1 to process step i is expressed as Q i and Q N+j They represent the waiting time before the main robot R3 moves from process step i to process step i+1 and from process step N+j to process step N+j+1. 0(i-1) Indicates that the main robot R3 is from the merging station P h Move to process step i and use transition t (i-1)i Indicates the auxiliary manipulator r i-1 Move from process step i-1 to process step i, using transition t in Indicates that the main robot R3 moves from process step i to process step n, using transition t N(N+j) Indicates that the main robot R3 moves from process step N to process step N+j, using transition t (N+j)(N+m) Indicates that the main robot R3 moves from process step N+j to process step N+m, using transition t (N+M)0 Indicates that the main robot R3 moves from process step N+M to the merging station P h In case of reentry, use transition t N(N-1) Indicates that the main robot R3 moves from process step N to process step N-1, which can be extended to transition t ni , Figure 4 The dotted arcs are connected in the figure. Place Z0 represents the main robot R3 to the merging station P. h Waiting time before loading wafer in Z i Indicates the auxiliary manipulator r i Waiting time before loading a wafer into process step i, place Z i-1 and Z N+j The transition L0 represents the waiting time before the main robot R3 loads the wafer into the process step i-1 and the process step N+j. h Loading wafer, transition L w Indicates that the main robot R3 has been cleaned and the transition is L i Indicates the auxiliary manipulator manipulator r i Loading wafer into process step i, transition L i-1 and changes N+j The main robot R3 loads wafers into process step i-1 and step N+j respectively. The dotted line represents the possible re-entry.

[0038] For the material merging area, the places R1 and R2 represent the pick-and-place robot R1 and the pick-and-place robot R2 respectively, K(R1)=K(R2)=1, and when there is a token in the place, it means that the robot is currently available. The parameter symbols and meanings of the remaining transitions and places are as follows:

[0039]

[0040] The scheduling of wafer processing in the process production area mainly relies on the main robot R3 as a carrier, so the construction of the scheduling algorithm is closely related to the control of the main robot R3. The auxiliary robot only serves as a material handling tool between adjacent workstations, and the order in which materials pass through adjacent workstations must be first through the smaller number and then through the larger number, that is, it will only be sent from workstation to workstation, and will not pass through only one of the adjacent processing stations. Therefore, the auxiliary robot only serves as a handling tool and has little relevance to the scheduling of the overall area. Therefore, the focus is on the control of the main robot R3.

[0041] A scheduling method for a single processing path of a wafer comprises the following steps:

[0042] S0: Construct the ROPN model using the inspection station, storage position, rotation station, merging station, pick-and-place robot R1, pick-and-place robot R2, processing station, main robot R3, auxiliary robot r and cleaning station;

[0043] S1: Establish the constraint conditions for the deadlock processing of the main robot, and use the ROPN model and establish the process production area scheduling algorithm for the main robot according to the constraint conditions for the deadlock processing of the main robot;

[0044] S2: Establish the constraint conditions for the deadlock processing of the pick-and-place robot, and use the ROPN model and the constraint conditions for the deadlock processing of the pick-and-place robot to establish the wafer dispatching strategy for the pick-and-place robot;

[0045] S3: Using the ROPN model and based on the constraints of the pick-and-place robot deadlock handling, a wafer feeding strategy for the pick-and-place robot is established.

[0046] In the scheduling method for a single processing path of the wafer, the constraint conditions of the deadlock processing of the main robot are used to plan the process production area scheduling algorithm for the main robot, and the constraint conditions of the deadlock processing of the pick-and-place robot are used to plan the wafer dispatching strategy for the pick-and-place robot and the wafer feeding strategy for the pick-and-place robot, which can ensure that the control needs of the pick-and-place robot R1, the pick-and-place robot R2 and the main robot R3 cooperate with each other so as to coordinate the overall operation.

[0047] Deadlock processing for the main robot R3:

[0048] System deadlock is caused by many factors, which has a negative impact on the overall system operation. Therefore, solving the deadlock problem becomes the primary task of scheduling. During the wafer processing, the transfer of wafers between processing stations depends on the handling of the robot. Especially for the main robot R3, this dependence may lead to competition for the main robot R3 resources, thereby increasing the risk of system deadlock. Figure 4 In the example, since there is more than one output transition in place R3 (i.e., main manipulator R3), when M(R3)=2, i.e., there are 2 tokens in place R3, the choice of which output transition to trigger depends on the scheduling strategy. In order to effectively solve the deadlock problem, the following constraints are established:

[0049] For non-adjacent processing stations, j = 1, 2, ..., D, D∈N * , total number D, N * For positive integers build:

[0050] Condition 1-1: In the system identifier M(P N+j+1 )=m N+j+1 -1, change X N+j Set to a triggerable state; where M() represents the number of tokens in the library, P N+j+1 represents the N+j+1th non-adjacent processing station, m N+j+1 represents the number of parallel processing stations in the N+j+1th process step, and the transition X N+j Indicates that the main robot R3 rotates to process step N+j;

[0051] Condition 1-2: In the system identifier M(P N+j )=m N+j And M(P h )=2,M(P t )=M(R2)=1, the transition X N+M Set to triggerable state; where P N+j represents the N+jth non-adjacent processing station, m N+j represents the number of parallel processing stations in the N+jth process step, P h represents the merge station, P t Indicates the rotating table, represents the pick-and-place robot R2, X N+D Indicates that the main robot R3 rotates to process step N+D;

[0052] For adjacent processing stations, i = 2, 4, ..., 2n, n∈N * , the total number is N, which is divided into two cases, the first one is for non-reentrant wafer flow, and the second one is for reentrant wafer flow;

[0053] Under non-reentrant wafer flow, establish:

[0054] Condition 2-1: When 2 ≤ i ≤ N-2, and when the system marking M(P i+1 ) = m i+1 -1 or M(P i+2 ) = m i+2 -1 or M(r i+1 ) = 1, the transition X i is set to the triggerable state; where P i+1 represents the (i+1)-th adjacent processing station, m i+1 , P i+2 represents the (i+1)-th adjacent processing station, m i+2 table, r i+1 represents the (i+1)-th sub-manipulator, and the transition X i represents the main manipulator R3 rotating to the process step i;

[0055] Condition 2-2: When the system marking M(P1) = m1-1 or M(P2) = m2-1 or M(r1) = 1, the transition X0 is set to the triggerable state; where P1 represents the 1st adjacent processing station, m1 represents the number of parallel processing stations for the 1st process step, P2 represents the 2nd adjacent processing station, m2 represents the number of parallel processing stations for the 2nd process step, r1 represents the 1st sub-manipulator, and the transition X0 represents the main manipulator R3 rotating to the merging station;

[0056] Condition 2-3: When the system marking M(P i ) = m i , the transition X N is set to the triggerable state; where P i represents the i-th adjacent processing station, m i represents the number of parallel processing stations for the i-th process step, and the transition X N represents the main manipulator R3 rotating to the process step N;

[0057] Under the reentrant wafer flow, establish:

[0058] Condition 2-4: When 1 ≤ q < k, and when the system marking M(P i-1 ) = m i-1 -1 or M(r i-1 ) = 1, the transitions t i(i-1) and the transition X i are both set to the triggerable state; where P i-1 represents the (i-1)-th adjacent processing station, m i-1 represents the number of parallel processing stations for the (i-1)-th process step, r i-1 represents the (i-1)-th sub-manipulator, the transition t i(i-1) represents the main manipulator R3 moving from the process step i to the process step i-1, and the transition X i represents the main manipulator R3 rotating to the process step i;

[0059] Condition 2-5: When 2≤i≤N-2, and when the system identifier M(P i+1 )=m i+1 -1 or M(P i+2 )=m i+2 -1 or M(r i+1 )=1, the transition t (i-2)(i-1) And Transition X i-2 Set to a triggerable state; the transition t (i-2)(i-1) Indicates that the main robot R3 moves from process step i-2 to process step i-1, and transition X i-2 Indicates that the main robot R3 rotates to process step i-2;

[0060] Condition 2-6: When q = k, transition t i(i+1) And Transition X i Set to a triggerable state; the transition t i(i+1) Indicates that the main robot R3 moves from process step i to process step i+1, and the transition X i Indicates that the main robot R3 rotates to process step i.

[0061] Among them, re-entry wafer flow (m i-1 ,m i ) k Indicates that k is a reentrant value and 2≤i≤N-2; use W a (q) indicates the ath wafer entering the system and completing the last processing step m of the reentry process i , and completed q reentries on behalf of the wafer.

[0062] Construction of the process production area scheduling algorithm:

[0063] The construction of the process production area scheduling algorithm is divided into two cases. The first is the production area scheduling algorithm based on the first-come-first-served rule (ProcessArea SchedulingAlgorithmBased on FCFS Rule, referred to as PAFR), and the second is the production area scheduling algorithm based on the pull rule (ProcessArea Scheduling AlgorithmBased on PushRule, referred to as PAPR);

[0064] The meanings of the parameter symbols in the production area scheduling algorithm based on the first-come-first-served rule and the production area scheduling algorithm based on the pull rule are as follows:

[0065]

[0066] The production area scheduling algorithm based on the first-come-first-served rule can be understood as "the earlier the workpiece arrives, the earlier it can be processed." The PAFR algorithm comprehensively considers the conditions for controlling deadlock. When M(R3)=2, that is, there are 2 tokens in the library R3, the main robot R3 gives priority to processing wafers with an earlier estimated completion time. In short, the earlier the estimated completion time of the wafer, the higher its priority in scheduling, where the completion time Equal to arrival time Wafer processing time The sum of the ath wafer W at station P i The completion time of a task is equal to the sum of its arrival time and processing time.

[0067] like Figure 5 As shown, the production area scheduling algorithm based on the first-come-first-served rule includes the following steps:

[0068] S111: Initialize all parameters, set e=1, if C has no recorded value, set C=0, where e is the index standard of the wafer sequence, e starts from the initial value 1, that is, the search is established from the first wafer, and C is the station number when the main robot R3 places the wafer on the station; the station includes the merging station, the processing station and the cleaning station;

[0069] S112: When the wafer Enter platform P i or Platform P N+j When the wafer arrives, record the arrival time Platform P i As it is close to the processing station, platform P N+j Not adjacent to a processing station;

[0070] S113: Set the wafer At Platform P i or Platform P N+j The processing time is Calculate the wafer Completion time in

[0071] S114: For all the completion times obtained Arrange in ascending order to obtain the wafer sequence S to be transported W ;

[0072] S115: Get wafer sequence S W In S e The wafer at the position; where S e Indicates that in the wafer sequence S W The e-th wafer in ;

[0073] S116: Determine if it is in Se Location of wafer Whether the constraint conditions for the deadlock processing of the main manipulator are met;

[0074] S117: If the conditions are met, it indicates that there is no deadlock, and the process jumps to step S118; if the conditions are not met, it indicates that there is a deadlock, and e=e+1 is set, and the process jumps to step S116;

[0075] S118: Main robot R3 locks the wafer And record the wafer grab The platform number at the time is B;

[0076] S119: judging whether the main manipulator R3 needs to be cleaned according to the value changes of the station number B and the station number C; the station number B and the station number C are the stations corresponding to the grabbing and placing operations respectively. When the system is initialized, it will be set how many process steps the main manipulator R3 needs to be cleaned after the task is grabbed;

[0077] S1110: If cleaning is required, the main robot R3 is cleaned, and then the process jumps to step S1111; if cleaning is not required, the process jumps to step S1111;

[0078] S1111: The main robot R3 will be in S e Position of wafer The wafer is then moved to the next processing step station and then the main robot R3 Unlock;

[0079] S1112: Update the station number C for placing the wafer this time. e The wafer information of the position is deleted, and the S e The wafer information at the following positions is all moved up one position, for example, position S e+1 Rise to S e .

[0080] In the production area scheduling algorithm based on the pull rule, the pull rule is mostly used in the scheduling scenario of the single-arm modular equipment. In the present invention, the scheduling scenario of the processing area is very similar to it. The pull rule adopts the research idea of ​​"backward push". In the processing process of the same type of wafers, σ={A0,A1,...,A m} is the handling operation sequence of the main robot R3, A i (0≤i≤m) indicates that the i-th transport operation is to transport the wafer completed in processing module i to the next processing module. The main robot R3 completes all operations in A, indicating that the system has completed a production cycle. Then the system repeats σ and enters the next cycle, σ U ={A m , A m-1, ..., A1} is called Pull sequence. This algorithm is based on the idea of ​​pull rule, when M(R3)=2, that is, there are 2 tokens in R3, combined with the actual production environment and considering relevant production constraints.

[0081] like Figure 6 As shown in Figure 1, the production area scheduling algorithm based on pull rules includes the following steps:

[0082] S121: Initialize each parameter, set e=1, if C has no recorded value, then set C=0, where e is the index standard of the wafer sequence, e starts from the initial value 1, that is, the search is established from the first wafer, and C is the station number when the main robot R3 places the wafer on the station; the station includes the merging station, the processing station and the cleaning station;

[0083] S122: When the wafer Enter platform P i or Platform P N+j When the wafer arrives, record the arrival time Platform P i As it is close to the processing station, platform P N+j Not adjacent to a processing station;

[0084] S123: Set the wafer At Platform P i or Platform P N+j The processing time is Calculate the wafer Completion time in

[0085] S124: For all the completion times obtained Arrange in ascending order to obtain the wafer sequence S to be transported W ;

[0086] S125: Wafer determination Whether the station required for the next processing step is free;

[0087] S126: If yes, go to step S1210; if no, go to step S127;

[0088] S127: Retain wafer In the wafer sequence S W The information in the process is received and the station required for the next processing step is left idle.

[0089] S128: On the wafer During the waiting process, determine whether there are new completed wafers;

[0090] S129: If yes, jump to step S122; if no, jump to step S1210;

[0091] S1210: Get wafer sequence S W In position S e Wafers; where Se represents the wafer sequence S W The e-th wafer in ;

[0092] S1211: In S e Position of wafer Whether the constraint conditions for the deadlock processing of the main manipulator are met;

[0093] S1212: If it is in compliance, jump to step S1213; if it is not in compliance, set e=e+1 and jump to step S1211;

[0094] S1213: Main robot R3 locks the wafer And record the station number B when grabbing the wafer;

[0095] S1214: judging whether the main manipulator R3 needs to be cleaned according to the value changes of the station number B and the station number C; the station number B and the station number C are the stations corresponding to the grabbing and placing operations respectively, and when the system is initialized, it will be set how many process steps the main manipulator R3 needs to be cleaned after the task is grabbed;

[0096] S1215: If cleaning is required, the main robot R3 is cleaned and then the process jumps to step S1216; if cleaning is not required, the process jumps to step S1216.

[0097] S1216: The main robot R3 will be in S e Position of wafer The wafer is then moved to the next processing step station and then the main robot R3 Unlock;

[0098] S1217: Update the station number C for placing the wafer this time. e The wafer information of the position is deleted, and the S e The wafer information at the following positions is all moved up one position, for example, position S e+1 Rise to S e .

[0099] Construction of material merging area scheduling algorithm:

[0100] The scheduling algorithms for the material merging area are divided into wafer dispatching strategy and wafer material feeding strategy.

[0101] The construction of wafer dispatch strategy focuses on the control of pick-and-place robot R1 and pick-and-place robot R2. As components of the shared channel (NSP), the two robots have a decisive influence on the smoothness of the channel. Specifically, pick-and-place robot R1 can trigger Figure 4 Medium Change X c 、ChangeX s 、ChangeX s ' and changes X t ', and the pick-and-place robot R2 can trigger Figure 4 Medium Change X t and Transition X h '. Based on M(R1)=M(R2)=1, that is, there is 1 token in R1 and 1 token in R2, for the control of the two manipulators, it is necessary to focus on solving the potential deadlock problem caused by the shared channel (NSP) and reasonably handle the triggering timing of each transition to ensure the progress of the production process.

[0102] Deadlock handling for pick-and-place robot R1 and pick-and-place robot R2:

[0103] Condition 3-1: M(R1) = 1 and M(P s )≠0, the transition X c Set to a triggerable state; the transition X c It indicates that for unfinished wafers, the pick-and-place robot R1 rotates unloaded to the inspection table O1 responsible for inspecting the wafers to be processed; it indicates that if the pick-and-place robot R1 is in an available state and there is an available position in the storage bit, the pick-and-place robot R1 can rotate unloaded to the inspection table O1 to prepare for the handling operation;

[0104] Condition 3-2: M(R1) = 1, M(R2) = 1, M(P s )≠S、M(P t )=1 and M(P h )<2, transition X s Set to a triggerable state; the transition X s Indicates that for unfinished wafers, the pick-and-place robot R1 and the pick-and-place robot R2 rotate to the storage position without load; indicates that if the pick-and-place robot R1 is in an available state, the pick-and-place robot R2 is in an available state, there are wafers in the storage position, the rotating table is available, and there are less than 2 materials in the merging station, the pick-and-place robot R1 can rotate to the storage position without load to prepare for the handling operation;

[0105] Condition 3-3: M(R1)=1 and M(P s )≠S, transition X s 'Set to a triggerable state; where transition X s'Indicates that for the finished wafer, the pick-and-place robot R1 rotates to the storage position without load; indicates that if the pick-and-place robot R1 is in an available state and there is a wafer in the storage position, the pick-and-place robot R1 can rotate to the storage position without load to prepare for the handling operation;

[0106] Condition 3-4: M(R1)=1, M(P s )≠0, transition X t *Set to a triggerable state; where transition X t *Indicates that for the finished wafer, the pick-and-place robot R1 rotates to the rotating table without load; indicates that if the pick-and-place robot R1 is in an available state and there is an available position in the storage bit, the pick-and-place robot R1 can rotate to the rotating table without load to prepare for the handling operation;

[0107] Condition 3-5: M(R2)=1, M(P h )<2, change X t Set to a triggerable state; the transition X t Indicates that for unfinished wafers, the pick-and-place robot R2 rotates to the rotating table without load; indicates that if the pick-and-place robot R2 is in an available state and there are less than 2 materials in the merging station, the pick-and-place robot R2 can rotate to the rotating table without load to prepare for the handling operation;

[0108] Condition 3-6: M(R2)=1, M(P t )=1,M(P s )≠0, transition X h *Set to a triggerable state; where transition X h *Indicates that for the finished wafers, the pick-and-place robot R2 rotates to the merging station without load; indicates that if the pick-and-place robot R2 is in an available state, the state of the rotating table is available, and there is an available position in the storage position, the pick-and-place robot R2 can rotate to the merging station without load to prepare for the handling operation;

[0109] Where P s Represents a storage bit, which is used to store wafers. S represents the maximum capacity of the storage bit. P t represents the rotating stage, P h Indicates a merge station.

[0110] By controlling the conditions for the deadlock processing of the pick-and-place robot R1 and the pick-and-place robot R2, the deadlock problem caused by the chaos of the pick-and-place robot R1 and the pick-and-place robot R2 in the material merging area can be effectively solved. However, considering the existence of NSP, there is an inevitable conflict between the output of wafers in the processing area and the input of wafers in the merging area. To this end, combined with the traffic control mode, a state control mechanism similar to a "traffic light" is implemented for NSP. When there are wafers in the last sequence in the processing area, the Entrance is in a red light state and the Exit is in a green light state, that is, the finished wafers can be output, but the unprocessed wafers cannot be input. They will wait in the storage position. When the finished wafers are taken out, the states of Entrance and Exit can be interchanged, and the conflict of wafers can be eliminated in this way. Specifically, when there are wafers in the last process step in the processing area, the Entrance is set to red light status and the Exit is set to green light status, which means that the finished wafers can be exported, but the unprocessed wafers need to wait in the storage position and cannot be imported. When the finished wafers are taken away, the status of Entrance and Exit will be swapped, and the unprocessed wafers can be imported.

[0111] Through observation, it is found that during the actual execution of the algorithm, when there is a wafer in the last process step in the processing area, immediately starting the channel state transition may cause the wafer that should have enough time to enter the processing area to be unable to be smoothly added and processed. Therefore, the timing of the state transition can be studied together with the processing time of the last sequence of the wafer. In the present invention, three different strategies are used to solve the above problems. The first strategy is the Prioritized Strategy for Finished Wafers (PSFW for short), the second strategy is the Prioritized Strategy for Unprocessed Wafers (PSUW for short), and the third strategy is the Randomization Strategy (RS for short).

[0112] Set the wafer to enter the last sequence at time 0 and the waiting time to t w , Red is set to 1 until the finished wafer is sent to P s (storage bit), the rest of the time Red is set to 0. Waiting time before state transition t w The calculation formula is: The following table shows the parameter symbols and meanings of the dispatch strategy:

[0113]

[0114]

[0115] The strategy of finishing wafer first is based on M(R1)=1 and M(P s )<S, the scheduling of the wafer merging area is realized by controlling the robot R1, wherein, for the wafers in P s (storage position), the priority principle followed by this strategy is: first process the completed wafers in the storage position, and then consider putting the wafers to be processed into the processing area. Figure 7 As shown, the steps of the finished wafer first strategy are as follows:

[0116] S211: Determine M(P t ) is 0, if so (indicating that there is a wafer on the rotating table), step S212 is executed, if not (indicating that there is no wafer on the rotating table), step S214 is executed; wherein P t A rotating table;

[0117] S212: Determine if P t Whether the wafer (turntable) is completed and has not been P h (Merge station), if yes, execute step S213, if no, execute step S2112;

[0118] S213: Trigger Transition X t *;

[0119] S214: Determine whether there is a wafer inspection completed at the inspection station O1, if yes, execute step S215, if not, execute step S216;

[0120] S215: Trigger Transition X c ;

[0121] S216: Determine M(P s ) is not equal to S, if so, execute step S217, if not, execute step S2112; wherein P s For storage location;

[0122] S217: Determine P s Is there a completed wafer W in the storage location? f ', if yes, execute step S218, if no, execute step S219;

[0123] S218: Trigger Transition X s ';

[0124] S219: Determine whether Red is 1, if so, execute step S2112, if not, execute step S2110;

[0125] S2110: Determine Ps Are there more than two unprocessed wafers W in (storage location) f If yes, execute step S2111, if no, execute step S2112;

[0126] S2111: Trigger two transitions X consecutively s , two unprocessed wafers of the same type W f take out;

[0127] S2112: The pick-and-place robot R1 does not operate.

[0128] like Figure 8 As shown, the unprocessed wafer priority strategy is based on M(R1)=1 and M(P s )<S, for those in P s (storage position) wafers, take priority processing P s The wafers to be processed in the (storage position) are put into the processing area, and then the priority principle of the completed wafers is considered. The steps of the strategy of giving priority to unprocessed wafers are as follows:

[0129] S221: Determine M(P t ) is 0, if so (indicating that there is a wafer on the rotating table), step S222 is executed, if not (indicating that there is no wafer on the rotating table), step S224 is executed; wherein P t A rotating table;

[0130] S222: Determine if P t Wafer W on the rotating table f Whether it is completed and has not been P h (Merge station), if yes, execute step S223, if not, execute step S2212.

[0131] S223: Trigger Transition X t *;

[0132] S224: Determine whether there is a wafer inspected at the inspection station O1, if yes, execute step S225, if no, execute step S226;

[0133] S225: Trigger Transition X c ;

[0134] S226: Determine M(P s ) is not equal to S, if so, execute step S227, if not, execute step S222; wherein P s is the storage bit;

[0135] S227: Judge P s Are there more than two unprocessed wafers W in (storage location) fIf yes, execute step S228, if no, execute step S229;

[0136] S228: Determine whether Red is 1, if so, execute step S2212, if not, execute step S229;

[0137] S229: Determine P s Is there any unprocessed wafer W in (storage location) f If yes, execute step S2210, if no, execute step S2211;

[0138] S2210: Trigger two transitions X consecutively s , two unprocessed wafers of the same type W f take out;

[0139] S2211: Trigger transition X s ';

[0140] S2212: The pick-and-place robot R1 does not operate.

[0141] The random selection strategy means that there is no clear priority order for the input of wafers to be processed and the delivery of completed wafers. The pick-and-place robot R1 will randomly select wafers that meet the requirements for operation based on the current situation.

[0142] The construction of the wafer feeding strategy aims to solve the wafer flow type, the best time to feed, and the required feeding quantity of the wafer. After the wafer arrives at the entrance station, the process of feeding it into the system must follow certain rules to maximize the production capacity of the production line. In this embodiment, there are two types of wafer feeding strategies, the first is the process area fixed work-in-progress strategy, and the second is the fixed time interval strategy.

[0143] In the process area fixed work-in-process strategy, the number of work-in-process represents the number of workpieces being processed. The expected level of work-in-process quantity in the system is calculated according to the little law. As the WIP level value set by the system, its formula is: Where V i represents the output rate of product i, D i Represents the processing time of product i. Through calculation, it provides a reference for setting the appropriate number of WIP in the processing area. When formulating the feeding plan, in addition to the above calculation results, the storage capacity limit of the wafer merging area will also be considered, and the production efficiency and storage capacity will be comprehensively considered; the steps of the process area fixed WIP strategy are as follows:

[0144] S311: Initialize the arriving wafer set Q and Set the serial number b in the wafer set Q to 1, indicating that the process will start from the first wafer;

[0145] S312: All available wafers are listed in the wafer set Q one by one according to the order of their arrival time;

[0146] S313: Determine whether the conditions M(R1)=1 and M(P s )≠0; R1 is the pick-and-place robot R1, P s is the storage bit;

[0147] S314: If the condition M(R1)=1 and M(P s )≠0, wait until the condition is met and then determine whether WIP0 is less than

[0148] S315: If yes, then the wafer Placed into the system, placed wafers Remove from wafer set Q, update wafer set Q and then jump to step S312; if not, directly jump to step S312;

[0149] Where WIP0 is the WIP level value of the current system, The WIP level value set for the system, is the bth wafer W arriving at the entrance station with wafer flow pattern j.

[0150] The fixed time interval strategy is to put a corresponding number of wafers into the entire system at a fixed time interval. The time interval is τ, and τ follows an exponential distribution with a mean of μ. The steps are as follows:

[0151] S321: Initialize the incoming wafer set Q, set the serial number b in the wafer set Q to 1, indicating that the process will start from the first wafer, and set the wafer The time of arrival at the entrance station is the initial time τ0;

[0152] S322: All available wafers are listed in the wafer set Q one by one according to the order of their arrival time;

[0153] S323: Determine the current time τ n Is it τ0+τ?

[0154] S324: If yes, determine whether the conditions M(R1)=1 and M(P s )≠0; otherwise wait until τ n =τ0+τ, and then determine whether the conditions M(R1)=1 and M(P s )≠0;

[0155] S325: If the condition is not met, wait until the condition is met and then go to step S326; if the condition is met, go directly to step S326;

[0156] S326: Wafer Placed into the system, placed wafers Remove from wafer set Q, update wafer set Q, and go to step S322.

[0157] In the present invention, if Fig. 9 As shown, a combined heuristic algorithm is constructed by combining two scheduling algorithms in process production areas (a production area scheduling algorithm based on a first-come-first-served rule and a production area scheduling algorithm based on a pull rule), three wafer dispatching strategies (a finished wafer priority strategy, an unprocessed wafer priority strategy, and a random selection strategy) and two wafer feeding strategies (a process area fixed work-in-progress strategy and a fixed time interval strategy).

[0158] The embodiments of the present invention are described in detail above with reference to the accompanying drawings, but the present invention is not limited to the described embodiments. For those skilled in the art, various changes, modifications, substitutions and variations of these embodiments are made without departing from the principles and spirit of the present invention, and still fall within the scope of protection of the present invention.

Claims

1. A scheduling method for a single processing path of a wafer, It is characterized in that The following steps are involved: S0: Using inspection table, storage position, rotating table, merging station, pick and place robot R 1 , Pick and place robot R 2 , processing station, main robot R 3 , auxiliary manipulator r and cleaning station to build ROPN model; S1: Establish the constraint conditions for the deadlock processing of the main robot, and use the ROPN model and establish the process production area scheduling algorithm for the main robot according to the constraint conditions for the deadlock processing of the main robot; S2: Establish the constraint conditions for the deadlock processing of the pick-and-place robot, and use the ROPN model and the constraint conditions for the deadlock processing of the pick-and-place robot to establish the wafer dispatching strategy for the pick-and-place robot; S3: Using the ROPN model and based on the constraints of the pick-and-place robot deadlock handling, a wafer feeding strategy for the pick-and-place robot is established.

2. The method for scheduling a single processing path of a wafer according to claim 1, Features: In the step S0, a directed graph of the ROPN model is constructed using places, transitions, directed arcs from transitions to places, directed arcs from places to transitions, token quantity identifiers of places, and the maximum number of tokens that can be accommodated by the places; Among them, the inspection table, storage position, rotating table, merging station, pick-and-place robot R 1 , Pick and place robot R 2 , processing station, main robot R 3 , auxiliary robot r, cleaning station and the waiting time before each robot action all belong to the library P in the ROPN model, and the transition is the inspection station, storage position, rotating table, merging station, pick-and-place robot R 1 , Pick and place robot R 2 , processing station, main robot R 3 , the execution actions of the auxiliary robot r and the cleaning station.

3. The method for scheduling a single processing path of a wafer according to claim 2, Features: The relationship between the processing stations is divided into adjacent processing stations P i-1 and P i The relationship between the non-adjacent processing station P N+j The relationship between i=2, 4, ..., 2N, N∈N * , a total of N adjacent processing stations, j = 1, 2, ..., D, D ∈ N * , a total of D non-adjacent processing stations; In step S1, the constraints for the main manipulator deadlock processing include: For non-adjacent processing stations P N+j ,j=1,2,...,D,D∈N * , total number D, N * For positive integers build: Condition 1-1: In M(P N+j+1 )=m N+j+1 -1, change X N+j Set to a triggerable state; where M() represents the number of tokens in the library, P N+j+1 represents the N+j+1th non-adjacent processing station, m N+j+1 represents the number of parallel processing stations in the N+j+1th process step, and the transition X N+j Indicates the main robot R 3 Move to process step N+j; Condition 1-2: In M(P N+j )=m N+j And M(P h )=2,M(P t )=M(R 2 )=1, the transition X N+M Set to triggerable state; where P N+j represents the N+jth non-adjacent processing station, m N+j represents the number of parallel processing stations in the N+jth process step, P h represents the merge station, P t Represents a rotating table, represents a pick-and-place robot R 2 , X N+D Indicates the main robot R 3 Move to process step N+D; For the adjacent processing station P i-1 and P i , i=2, 4,..., 2N, N∈N * , the total number is N, under non-reentrant wafer flow, establish: Condition 2-1: When 2≤i≤N-2, and when M(P i+1 )=m i+1 -1 or M(P i+2 )=m i+2 -1 or M(r i+1 )=1, the transition X i Set to triggerable state; where P i+1 represents the i+1th adjacent processing station, m i+1 , P i+2 represents the i+1th adjacent processing station, m i+2 Table, r i+1 Indicates the i+1th auxiliary manipulator, transition X i Indicates the main robot R 3 Move to process step i; Condition 2-2: When M(P 1 )=m 1 -1 or M(P 2 )=m 2 -1 or M(r 1 )=1, the transition X 0 Set to triggerable state; where P 1 Indicates the first adjacent processing station, m 1 represents the number of parallel processing stations in the first process step, P 2 Indicates the second adjacent processing station, m 2 represents the number of parallel processing stations in the second process step, r 1 Indicates the first auxiliary robot, transition X 0 Indicates the main robot R 3 Turn to the merge station; Condition 2-3: When M(P i )=m i When, change X N Set to triggerable state; where P i represents the i-th adjacent processing station, m i represents the number of parallel processing stations in the i-th process step, and the transition X N Indicates the main robot R 3 Move to process step N; Re-entry wafer flow (m i-1 ,m i ) k Indicates that k is a reentrant value and 2≤i≤N-2; use W a (q) indicates the ath wafer entering the system and completing the last processing step m of the reentry process i , and represents that the wafer has completed q reentry times. Under the reentry wafer flow, establish: Condition 2-4: When 1 ≤ q < k and when M(P i-1 ) = m i-1 -1 or M(r i-1 ) = 1, the transitions t i(i-1) and the transition X i are both set to the triggerable state; where P i-1 represents the (i - 1)-th adjacent processing station, m i-1 represents the number of parallel processing stations in the (i - 1)-th process step, r i-1 represents the (i - 1)-th sub-manipulator, the transition t i(i-1) represents the main manipulator R 3 moving from the process step i to the process step i - 1, and the transition X i represents the main manipulator R 3 rotating to the process step i; Condition 2-5: When 2≤i≤N-2, and when the system identifier M(P i+1 )=m i+1 -1 or M(P i+2 )=m i+2 -1 or M(r i+1 )=1, the transition t (i-2)(i-1) And Transition X i-2 Set to a triggerable state; the transition t (i-2)(i-1) Indicates the main robot R 3 Move from process step i-2 to process step i-1, transition X i-2 Indicates the main robot R 3 Move to process step i-2; Condition 2-6: When q = k, transition t i(i+1) And Transition X i Set to a triggerable state; the transition t i(i+1) Indicates the main robot R 3 Move from process step i to process step i+1, transition X i Indicates the main robot R 3 Go to process step i.

4. The method for scheduling a single processing path of a wafer according to claim 3, Features: In step S1, the process production area scheduling algorithm for the main robot includes a production area scheduling algorithm based on a first-come, first-served rule, and the production area scheduling algorithm based on a first-come, first-served rule includes the following sub-steps: S111: Initialize all parameters, set e = 1, if C has no recorded value, set C = 0, where e is the index standard of the wafer sequence, e starts from the initial value 1, C is the main robot R 3 The station number when the wafer is placed on the station; stations include merge stations, processing stations, and cleaning stations; S112: When the wafer Enter the adjacent processing station P i Or not adjacent to processing station P N+j When the wafer arrives time S113: Set the wafer In the immediate vicinity of processing station P i Or not adjacent to processing station P N+j The processing time is Calculate the wafer Completion time in S114: For all the completion times obtained Arrange in ascending order to obtain the wafer sequence S to be transported W ; S115: Get wafer sequence S W In S e Position of the wafer; where S e Indicates that in the wafer sequence S W The e-th wafer in ; S116: Determine if it is in S e Location of wafer Whether the constraint conditions for the deadlock processing of the main manipulator are met; S117: If yes, jump to step S118; if no, set e=e+1 and jump to step S116; S118: Main manipulator R 3 Lock the wafer And record the wafer grab The platform number at the time is B; S119: According to the value changes of station number B and station number C, determine the main robot R 3 Whether cleaning is required; S1110: If cleaning is required, clean the main robot R 3 Perform a cleaning operation, and then jump to step S1111; If cleaning is not required, jump to step S1111; S1111: Through the main robot R 3 Will be in S e Location of wafer The robot is then moved to the next processing station and then the main robot R 3 For this wafer Unlock; S1112: Update the station number C for placing the wafer this time. e The wafer information of the position is deleted, and the S e All wafer information at the subsequent positions is moved up one position.

5. The method for scheduling a single processing path of a wafer according to claim 4, Features: In the step S1, the process production area scheduling algorithm for the main robot also includes a production area scheduling algorithm based on a pull rule, and the production area scheduling algorithm based on a pull rule includes the following sub-steps: S121: Initialize all parameters, set e = 1, if C has no recorded value, set C = 0, where e is the index standard of the wafer sequence, e starts from the initial value 1, C is the main robot R 3 The station number when the wafer is placed on the station; stations include merge stations, processing stations, and cleaning stations; S122: When the wafer Enter the adjacent processing station P i Or not adjacent to processing station P N+j When the wafer arrives, record the arrival time S123: Set the wafer In the immediate vicinity of processing station P i Or not adjacent to processing station P N+j The processing time is Calculate the wafer Completion time in S124: For all the completion times obtained Arrange in ascending order to obtain the wafer sequence S to be transported W ; S125: Wafer determination Whether the station required for the next processing step is free; S126: If yes, go to step S1210; if no, go to step S127; S127: Retain wafer In the wafer sequence S W The information in the process is received and the station required for the next processing step is left idle. S128: On the wafer During the waiting process, determine whether there are new completed wafers; S129: If yes, jump to step S122; if no, jump to step S1210; S1210: Get wafer sequence S W In position S e Wafers; where Se represents the wafer sequence S W The e-th wafer in ; S1211: In S e Location of wafer Whether the constraint conditions for the deadlock processing of the main manipulator are met; S1212: If it is in compliance, jump to step S1213; if it is not in compliance, set e=e+1 and jump to step S1211; S1213: Main manipulator R 3 Lock the wafer And record the station number B when grabbing the wafer; S1214: According to the value changes of station number B and station number C, determine the main robot R 3 Whether cleaning is required; S1215: If cleaning is required, the main robot R 3 Perform a cleaning operation and then jump to step S1216; if cleaning is not required, jump to step S1216. S1216: Through the main robot R 3 Will be in S e Location of wafer The robot is then moved to the next processing station and then the main robot R 3 For this wafer Unlock; S1217: Update the station number C for placing the wafer this time. e The wafer information of the position is deleted, and the S e All wafer information at the subsequent positions is moved up one position.

6. The method for scheduling a single processing path of a wafer according to claim 5, Features: In step S2, for the pick-and-place robot R 1 and pick and place robot R 2 ,The constraints for deadlock handling of the pick-and-place robot include: Condition 3-1: M(R 1 )=1 and M(P s )≠0, the transition X c Set to a triggerable state; the transition X c Indicates that for unfinished wafers, the pick-and-place robot R 1 The machine rotates unloaded to the inspection station O1 responsible for inspecting the wafers to be processed; Condition 3-2: M(R 1 )=1、M(R 2 )=1、M(P s )≠S、M(P t )=1 and M(P h )<2, transition X s Set to a triggerable state; the transition X s Indicates that for unfinished wafers, the pick-and-place robot R 1 and pick and place robot R 2 No-load rotation to storage position; Condition 3-3: M(R 1 )=1 and M(P s )≠S, transition X s 'Set to a triggerable state; where transition X s ' indicates that for the finished wafer, the pick-and-place robot R 1 No-load rotation to storage position; Condition 3-4: M(R 1 )=1,M(P s )≠0, transition X t *Set to a triggerable state; where transition X t *Indicates that for finished wafers, the pick-and-place robot R 1 Rotate to the rotating table without load; Condition 3-5: M(R 2 )=1,M(P h )<2, change X t Set to a triggerable state; the transition X t Indicates that for unfinished wafers, the pick-and-place robot R 2 Rotate to the rotating table without load; Condition 3-6: M(R 2 )=1,M(P t )=1,M(P s )≠0, transition X h *Set to a triggerable state; where transition X h *Indicates that for finished wafers, the pick-and-place robot R 2 Rotate empty to the merge station; Where P s Represents the storage bit, S represents the maximum capacity of the storage bit, P t represents the rotating stage, P h Indicates a merge station.

7. The method for scheduling a single processing path of a wafer according to claim 6, Features: In step S2, the wafer dispatching strategy for the pick-and-place robot includes a finished wafer priority strategy, and the finished wafer priority strategy includes the following sub-steps: S211: Determine M(P t ) is 0, if yes, execute step S212, if no, execute step S214; wherein P t A rotating table; S212: Determine whether the wafer on the rotating table is completed and has not passed through the merging station. If so, execute step S213; otherwise, execute step S2112; S213: Trigger Transition X t *; where the transition X t *Indicates that for finished wafers, the pick-and-place robot R 1 Rotate to the rotating table without load; S214: Determine whether there is a wafer inspection completed at the inspection station O1, if yes, execute step S215, if not, execute step S216; S215: Trigger Transition X c ; Among them, the transition X c Indicates that for unfinished wafers, the pick-and-place robot R 1 The machine rotates unloaded to the inspection station O1 responsible for inspecting the wafers to be processed; S216: Determine M(P s ) is not equal to S L If yes, then execute step S217, if no, then execute step S2112; where P s For storage location; S L Indicates the last processing step of the workpiece; S217: Determine whether there is a completed wafer W in the storage location f ', if yes, execute step S218, if no, execute step S219; S218: Trigger Transition X s '; Among them, the transition X s ' indicates that for the finished wafer, the pick-and-place robot R 1 No-load rotation to storage position; S219: Determine whether Red is 1, if so, execute step S2112, if not, execute step S2110; when Red is 0, the entrance is in a passable green light state, and the exit is in a passable red light state; when Red is 1, the entrance is in a passable red light state, and the exit is in a passable green light state; S2110: Determine whether there are more than two unprocessed wafers W in the storage location f If yes, execute step S2111, if no, execute step S2112; S2111: Trigger two transitions X consecutively s , two unprocessed wafers of the same type W f Take out; the transition means that for unfinished wafers, the pick-and-place robot and the empty-load rotation are rotated to the storage position; S2112: Pick and place robot R 1 No operation is performed.

8. The method for scheduling a single processing path of a wafer according to claim 7, Features: In step S2, the wafer dispatching strategy for the pick-and-place robot also includes a strategy of giving priority to unprocessed wafers, and the strategy of giving priority to unprocessed wafers includes the following sub-steps: S221: Determine M(P t ) is 0, if yes, execute step S222, if no, execute step S224; wherein P t A rotating table; S222: Determine the wafer W on the rotating table f Whether it is completed and has not passed the merging station, if so, execute step S223, if not, execute step S2212. S223: Trigger Transition X t *; where the transition X t *Indicates that for finished wafers, the pick-and-place robot R 1 Rotate to the rotating table without load; S224: Determine whether there is a wafer inspected at the inspection station O1, if yes, execute step S225, if no, execute step S226; S225: Trigger Transition X c ; Among them, the transition X c Indicates that for unfinished wafers, the pick-and-place robot R 1 The machine rotates unloaded to the inspection station O1 responsible for inspecting the wafers to be processed; S226: Determine M(P s ) is not equal to S, if so, execute step S227, if not, execute step S222; wherein P s is the storage bit; S227: Determine whether there are more than two unprocessed wafers W in the storage location. f If yes, execute step S228, if no, execute step S229; S228: Determine whether Red is 1, if so, execute step S2212, if not, execute step S229; when Red is 0, the entrance is in a passable green light state, and the exit is in a passable red light state; when Red is 1, the entrance is in a passable red light state, and the exit is in a passable green light state; S229: Determine whether there is an unprocessed wafer W in the storage location f If yes, execute step S2210, if no, execute step S2211; S2210: Trigger two transitions X consecutively s , two unprocessed wafers of the same type W f Take out; the transition means that for unfinished wafers, the pick-and-place robot and the empty-load rotation are rotated to the storage position; S2211: Trigger transition X s '; Among them, the transition X s ' indicates that for the finished wafer, the pick-and-place robot R 1 No-load rotation to storage position; S2212: Pick and place robot R 1 No operation is performed.

9. The method for scheduling a single processing path of a wafer according to claim 8, Features: In step S3, the wafer feeding strategy for the pick-and-place robot includes a process area fixed work-in-progress strategy, and the process area fixed work-in-progress strategy includes the following sub-steps: S311: Initialize the arriving wafer set Q and Set the serial number b in the wafer set Q to 1, indicating that the process will start from the first wafer next; S312: All available wafers are listed in the wafer set Q one by one according to the order of their arrival time; S313: Determine whether the condition M(R) is met 1 )=1 and M(P s )≠0; where R 1 For the pick and place robot R 1 , P s is the storage bit; S314: If the condition M(R 1 )=1 and M(P s )≠0, wait until the conditions are met and determine WIP 0 Is it less than S315: If yes, then the wafer Placed into the system, placed wafers Remove from wafer set Q, update wafer set Q and then jump to step S312; if not, directly jump to step S312; WIP 0 is the WIP level value of the current system, The WIP level value set for the system, is the bth wafer W arriving at the entrance station with wafer flow pattern j.

10. The method for scheduling a single processing path of a wafer according to claim 9, Features: In step S3, the wafer feeding strategy for the pick-and-place robot includes a fixed time interval strategy, and the fixed time interval strategy includes the following sub-steps: S321: Initialize the incoming wafer set Q, set the serial number b in the wafer set Q to 1, to indicate that the next step is to start from the first wafer, and set the wafer The time to arrive at the entrance station is the initial time τ 0 ; S322: All available wafers are listed one by one in the wafer set Q according to the order of their arrival time; S323: Determine the current time τ n Is it τ 0 +τ; where τ is the fixed time interval from wafer input to the entire system; S324: If yes, determine whether the condition M(R 1 )=1 and M(P s )≠0; otherwise wait until τ n =τ 0 +τ, and then judge whether it meets the condition M(R 1 )=1 and M(P s )≠0; where R 1 For the pick and place robot R 1 , P s is the storage bit; S325: If the condition M(R 1 )=1 and M(P s )≠0, wait until the condition is met and then go to step S326; if the condition M(R 1 )=1 and M(P s )≠0, go to step S326; S326: Wafer Placed into the system, placed wafers Remove from wafer set Q, update wafer set Q, and go to step S322; in is the bth wafer W arriving at the entrance station with wafer flow pattern j.

Citation Information

Patent Citations

  • Steady-state scheduling method for dual-combination equipment with full-parallel processing module

    CN117174618A

  • Flow shop scheduling method and device considering combined buffering and unmanned transportation

    CN117742260A

  • Method and system for auto dispatching wafer

    US6470231B1

Cited By

  • ROPN-based vacuum lock in-out control method, device and equipment and storage medium

    CN120335375A