Rapid feeding and discharging method for semiconductor factory

Abstract

The invention discloses a rapid loading and unloading method for a semiconductor factory, which belongs to the technical field of automation, and comprises the following steps: detecting an unloading signal of a machine A in real time through a material management server, acquiring processing material information and determining the position of a next machine B; the server generates loading and unloading tasks and dispatches the AGV, and the AGV executes efficient transportation and loading and unloading operation of the materials from the machine A to the machine B according to the priority order and the electronic map, so that the technical problem of reducing the on-site machine waiting time is solved, the priority of the materials can be efficiently managed, it is ensured that the high-priority tasks can be executed preferentially, and the working efficiency is improved. The complexity of task allocation is reduced, conflicts between AGVs on the route are avoided, the safety and efficiency of the whole scheduling process are improved, shutdown and waiting of a machine table caused by asynchronous feeding and discharging are avoided, the speed of material feeding and discharging is increased, the idle time of the machine table is shortened, and the overall production efficiency is improved.

Description

Technical Field

[0001] The invention belongs to the field of automation technology, and in particular relates to a method for quickly loading and unloading materials in a semiconductor factory. Background Art

[0002] In modern industrial production, production machines are produced by extraction, that is, the machine extracts materials from the front opening unified pod (FOUP) to the inside of the machine for production, and puts them back into the original FOUP after production is completed. This results in the FOUP remaining in the original position after the machine production is completed. On-site automated transportation is carried out in the following way: when the production machine is idle, select AGV (Automated Guided Vehicle) to transport the FOUP to the machine. After the machine production is completed, operate the AGV to take out the FOUP and transport it to the idle machine of the next process or the shelf in the temporary storage area. After the AGV takes down the material, select a new FOUP to transport it to the current machine.

[0003] Deficiencies of existing technology:

[0004] 1. During the material transportation and unloading process on the machine, it is impossible to load the material, which will cause the machine on site to wait for materials for a long time.

[0005] 2. The AGV path and task allocation are not optimal, resulting in low transportation efficiency. Summary of the invention

[0006] The purpose of the present invention is to provide a method for fast loading and unloading of semiconductor factories, which solves the technical problem of reducing the waiting time of machines on site.

[0007] To achieve the above object, the present invention adopts the following technical solution:

[0008] A method for rapid loading and unloading of semiconductor factories comprises the following steps:

[0009] Step 1: After machine A completes production, a material unloading signal is generated. After the material management server detects and obtains the material unloading signal, the following steps 2 to 8 are executed;

[0010] Step 2: Obtain information of processed materials from machine A;

[0011] Step 3: Determine the next station machine B based on the information of the processed materials;

[0012] Step 4: Obtain site location information from the electronic map, the site location information includes the location information of machine A, the location information of material storage location C of machine A, and the location information of material storage location D of the next station machine B;

[0013] Step 5: Generate unloading tasks, including information of processed materials and site location information;

[0014] Step 6: Obtain the information of the unprocessed materials in the material storage location C, calculate the priority scores according to the process attributes, sort them, select the unprocessed materials with the highest priority as the materials to be processed, and record the priority scores of the materials to be processed;

[0015] Step 7: Generate a material loading task, including the information of the material to be processed and the site location information, and mark the priority score of the material to be processed as the priority score of the material loading task;

[0016] Step 8: Pack the loading task and unloading task into the same batch of task packages, and mark the priority score of the loading task as the priority score of the task package;

[0017] Send the task packages to the task scheduling server in batches, and the task scheduling server executes the following steps 9 to 10;

[0018] Step 9: Add the task packages to the task list and sort them according to the priority scores; select the task package ranked first in the task list to be sent to the AGV, and formulate the execution route of the AGV according to the site location information in the task package, and the execution route first goes to the material storage C;

[0019] Step 10: All AGVs upload their location information to the task scheduling server in real time, and the task scheduling server generates scheduling information;

[0020] Step 11: The task scheduling server randomly selects an idle AGV, namely AGV_C, and sends the highest priority task package and scheduling information in the task list to AGV_C. AGV_C executes the task according to the execution route and scheduling information in the task package.

[0021] Preferably, when executing step 4, the electronic map is a factory electronic map preset in the material management server, each machine is configured with a unique material storage location, and the geographical location of each machine and the geographical location of each material storage location are marked in the electronic map.

[0022] Preferably, when executing step 6, the following steps are specifically included:

[0023] Step 6-1: The material management server obtains information of all unprocessed materials from the material storage location C;

[0024] Step 6-2: Obtain the process attributes of each unprocessed material, including the placement time T i 、Emergency i and priority O i ;

[0025] Step 6-3: Calculate the priority score for each unprocessed material using the following priority formula:

[0026]

[0027] Among them, P i is the priority score of the i-th unprocessed material; E i is the urgency of the i-th unprocessed material; i is the production sequence of unprocessed materials. The lower the value, the higher the production sequence priority. i is the placement time of the unprocessed material. The larger the value, the longer the placement time. w1, w2, and w3 are all weight coefficients.

[0028] Step 6-4: Assign priority scores P to all unprocessed materials i Sort by

[0029] Step 6-5: Select the unprocessed material with the highest priority score as the material to be processed;

[0030] Step 6-6: Record the priority score of the material to be processed and mark it as the priority of the loading task.

[0031] Preferably, when executing step 9, specifically, a task list is established, and after obtaining a batch of task packages, the task packages of the batch are added to the task list and sorted according to the priority scores of the task packages of the batch;

[0032] Select the task package ranked first, and formulate the execution route of the AGV according to the site location information in the task package. The execution route is to go to the material storage C first, then to the machine A, and finally to the material storage D.

[0033] Preferably, when executing step 10, the following steps are specifically included:

[0034] Step 10-1: All AGVs upload their current location information to the task scheduling server in real time, and the system records the current location and motion status of each AGV;

[0035] Step 10-2: The task scheduling server calculates the estimated time T for each AGV to enter the next path segment based on the AGV's location information. i,k ;

[0036] Step 10-3: According to the length d of the path segment k and the AGV's travel speed v i , calculate the travel time T of each AGV in each path segment i,k ;

[0037] Step 10-4: Assign a time window to each AGV on each path:

[0038] TW i =[T i,start ,T i,end ];

[0039] Make sure the time windows of all AGVs do not overlap:

[0040]

[0041] Step 10-5: According to the arrangement of the time window, adjust the travel time of each AGV and generate the scheduling information of each AGV.

[0042] Preferably, when executing step 11, the following steps are specifically included:

[0043] Step 11-1: AGV_C goes to material storage C and takes out the materials to be processed;

[0044] Step 11-2: AGV_C goes to machine A, takes out the processed material and puts the material to be processed;

[0045] Step 11-3: AGV_C goes to material storage D to store the processed materials.

[0046] The semiconductor factory fast loading and unloading method described in the present invention solves the technical problem of reducing the waiting time of on-site machines. The present invention can efficiently manage the material priority by calculating the priority of each material. The loading task and the unloading task are combined and packaged into the task package of the same batch. The priority score of the task package is reasonably set to ensure that the high-priority task can be executed first. At the same time, the AGV scheduling generates the optimal route based on the site location information of the task package, reducing the complexity of task allocation, and dynamically scheduling through the real-time location information of the AGV. The travel time of each AGV on the path segment is calculated in real time, and an independent time window is allocated to each AGV to avoid conflicts between AGVs on the route, thereby improving the safety and efficiency of the entire scheduling process. When completing the material exchange task, the machine shutdown and waiting caused by the unsynchronized loading and unloading is avoided. Through reasonable task planning and AGV scheduling, the speed of material loading and unloading is accelerated, thereby reducing the idle time of the machine and improving the overall production efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 It is the main flow chart of the present invention;

[0048] Figure 2 A flow chart of the route when the AGV of the present invention performs a task;

[0049] Figure 3Schematic diagram of the path segments of the AGV of the present invention under two schemes. DETAILED DESCRIPTION

[0050] like Figure 1-Figure 3 A semiconductor factory rapid loading and unloading method is shown, comprising the following steps:

[0051] Step 1: After machine A completes production, a material unloading signal is generated. After the material management server detects and obtains the material unloading signal, the following steps 2 to 8 are executed;

[0052] Step 2: Obtain information of processed materials from machine A;

[0053] Step 3: Determine the next station machine B based on the information of the processed materials;

[0054] Step 4: Obtain site location information from the electronic map, the site location information includes the location information of machine A, the location information of material storage location C of machine A, and the location information of material storage location D of the next station machine B;

[0055] The electronic map is a factory electronic map preset in the material management server. Each machine is configured with a unique material storage location. The geographical location of each machine and the geographical location of each material storage location are marked in the electronic map.

[0056] In this embodiment, the factory electronic map records the location of each machine and each material storage location, and marks the distance between each device.

[0057] Step 5: Generate unloading tasks, including information of processed materials and site location information;

[0058] Step 6: Obtain the information of the unprocessed materials in the material storage location C, calculate the priority scores according to the process attributes, sort them, select the unprocessed materials with the highest priority as the materials to be processed, and record the priority scores of the materials to be processed;

[0059] The specific steps include:

[0060] Step 6-1: The material management server obtains information of all unprocessed materials from the material storage location C;

[0061] Step 6-2: Obtain the process attributes of each unprocessed material, including the placement time T i 、Emergency i and priority O i ;

[0062] In this embodiment, the process attributes of the material are set according to the order information of the material, including the placement time T of the material in the storage place. iThis parameter is generated by the material management server: when the material is stored in the material storage C, the material management server starts timing and records the storage time of the material in the material storage C. When the material is taken away from the material storage C, the material management server interrupts the timing.

[0063] Placement time T i The larger the value, the higher the contribution to the priority score.

[0064] Urgency E i Set according to order information;

[0065] Priority O i A sorting number that represents the production order of the material, which is sorted on a first-come, first-served basis.

[0066] Step 6-3: Calculate the priority score for each unprocessed material using the following priority formula:

[0067]

[0068] Among them, P i is the priority score of the i-th unprocessed material; E i is the urgency of the i-th unprocessed material; i is the production sequence of unprocessed materials. The lower the value, the higher the production sequence priority. i is the placement time of the unprocessed material. The larger the value, the longer the placement time. w1, w2, and w3 are all weight coefficients.

[0069] Step 6-4: Assign priority scores P to all unprocessed materials i Sort by

[0070] Step 6-5: Select the unprocessed material with the highest priority score as the material to be processed;

[0071] Step 6-6: Record the priority score of the material to be processed and mark it as the priority of the loading task.

[0072] Step 7: Generate a material loading task, including the information of the material to be processed and the site location information, and mark the priority score of the material to be processed as the priority score of the material loading task;

[0073] Step 8: Pack the loading task and unloading task into the same batch of task packages, and mark the priority score of the loading task as the priority score of the task package;

[0074] Send the task packages to the task scheduling server in batches, and the task scheduling server executes the following steps 9 to 10;

[0075] Step 9: Add the task packages to the task list and sort them according to the priority scores; select the task package ranked first in the task list to be sent to the AGV, and formulate the execution route of the AGV according to the site location information in the task package, and the execution route first goes to the material storage C;

[0076] When executing step 9, specifically, a task list is established. After obtaining a batch of task packages, the task packages of the batch are added to the task list and sorted according to the priority scores of the task packages of the batch;

[0077] Select the task package ranked first, and formulate the execution route of the AGV according to the site location information in the task package. The execution route is to go to the material storage C first, then to the machine A, and finally to the material storage D.

[0078] like Figure 2 Two driving route maps of a single AGV are shown, among which path segments L1, path segment L2, and path segment L3 are routes formulated in this embodiment, recorded as exchange material transportation routes, and path segments L4, path segment L5, path segment L6, and path segment L7 are conventional transportation routes in the prior art, recorded as conventional routes. It can be seen that in the conventional route, machine A will have a long standby time, during which processing cannot be carried out, while the exchange material transportation route in this embodiment avoids the entire standby time, so that machine A can load materials immediately after unloading, and its loading time can achieve a fixed value, thereby speeding up the production efficiency of machine A.

[0079] Tables 1 to 6 below are data tables for three application scenarios in this embodiment when the location distances between the machine A, the material storage location C, and the material storage location D are different:

[0080]

[0081]

[0082] Table 1

[0083]

[0084] Table 2

[0085]

[0086] Table 3

[0087]

[0088] Table 4

[0089]

[0090] Table 5

[0091]

[0092] Table 6

[0093] It can be seen from Tables 1 to 6 that in scenario 1, when the AGV transport time and material loading time are calculated and each segment of the normal transport distance is higher than the exchange material transport distance, the overall transport time is saved by 30s and the material loading time is saved by 50s.

[0094] In scenario 2, the AGV transport time and material loading time are calculated. When the data of each segment of the normal transport distance is the same as the exchange material transport distance, the overall transport time is saved by 15s and the material loading time is saved by 38s.

[0095] In scenario 3, the AGV transport time and material loading time are calculated. When each segment of the normal transport distance is lower than the exchange material transport distance, the overall transport time is saved by 0s and the material loading time is saved by 23s.

[0096] The total time for the materials to be processed to be put on the machine for the material exchange transport route adopted in this embodiment is fixed and is less than the total time for the materials to be processed to be put on the machine for the conventional transport route. This greatly improves production efficiency and reduces the time that machine A is in standby mode to a minimum.

[0097] Step 10: All AGVs upload their location information to the task scheduling server in real time, and the task scheduling server generates scheduling information;

[0098] The specific steps include:

[0099] Step 10-1: All AGVs upload their current location information to the task scheduling server in real time, and the system records the current location and motion status of each AGV;

[0100] Step 10-2: The task scheduling server calculates the estimated time T for each AGV to enter the next path segment based on the AGV's location information. i,k ;

[0101] Step 10-3: According to the length d of the path segment k and the AGV's travel speed v i , calculate the travel time T of each AGV in each path segment i,k ;

[0102] In this embodiment, the AGV travel speed v i It is a fixed speed. The driving speed of each AGV may be different. The driving time T is calculated by the speed formula i,k :

[0103]

[0104] Step 10-4: Assign a time window to each AGV on each path:

[0105] TW i =[T i,start ,T i,end ];

[0106] In this embodiment, TW i Represents the time window of the i-th AGV in a certain path segment;

[0107] T i,start Indicates the time when the i-th AGV starts to enter the path segment;

[0108] T i,end Indicates the time when the i-th AGV leaves the path segment;

[0109] The path segment represents the route of the AGV from one station to another. The distance d of each route segment is k Pre-marked on the electronic map.

[0110] Make sure the time windows of all AGVs do not overlap:

[0111]

[0112] TW j It represents the time window of the j-th AGV in a certain path segment. The time windows of the i-th AGV and the j-th AGV cannot have intersections, ensuring that they do not occupy the same path in the same time period.

[0113] Step 10-5: According to the arrangement of the time window, adjust the travel time of each AGV and generate the scheduling information of each AGV.

[0114] In this embodiment, the time taken by each AGV on each path segment is first calculated using the speed formula:

[0115] That is, using the formula Calculate the travel time T of the i-th AGV i,k , and then determine the time window of each AGV on each path segment.

[0116] Then schedule according to the time window of each AGV. Take two AGVs as an example:

[0117] AGV1 and AGV2 need to process tasks at the same time, and both will pass through the path segment L1 between the material storage C and the machine A, with a path distance of 10 meters; the speed of AGV1 is 1 meter / second, and the speed of AGV2 is 0.8 meter / second.

[0118] According to the speed formula, it can be calculated that AGV1 needs 10 seconds to pass through path segment L1, and AGV2 needs 12 seconds to pass through path segment L1.

[0119] According to the near time window, if AGV1 enters path segment L1 at time t=0, its time window is [0,10], which means that AGV1 occupies this path segment between t=0 and t=10.

[0120] AGV2 can only enter after AGV1 leaves. Therefore, the time window of AGV2 is [10, 22.5], ensuring that the two will not conflict.

[0121] Step 11: The task scheduling server randomly selects an idle AGV, namely AGV_C, and sends the task package with the highest priority in the task list and the scheduling information to AGV_C. AGV_C executes the task according to the execution route and scheduling information in the task package.

[0122] The specific steps include:

[0123] Step 11-1: AGV_C goes to material storage C and takes out the materials to be processed;

[0124] Step 11-2: AGV_C goes to machine A, takes out the processed material and puts the material to be processed;

[0125] Step 11-3: AGV_C goes to material storage D to store the processed materials.

[0126] The semiconductor factory fast loading and unloading method described in the present invention solves the technical problem of reducing the waiting time of on-site machines. The present invention can efficiently manage the material priority by calculating the priority of each material. The loading task and the unloading task are combined and packaged into the task package of the same batch. The priority score of the task package is reasonably set to ensure that the high-priority task can be executed first. At the same time, the AGV scheduling generates the optimal route based on the site location information of the task package, reducing the complexity of task allocation, and dynamically scheduling through the real-time location information of the AGV. The travel time of each AGV on the path segment is calculated in real time, and an independent time window is allocated to each AGV to avoid conflicts between AGVs on the route, thereby improving the safety and efficiency of the entire scheduling process. When completing the material exchange task, the machine shutdown and waiting caused by the unsynchronized loading and unloading is avoided. Through reasonable task planning and AGV scheduling, the speed of material loading and unloading is accelerated, thereby reducing the idle time of the machine and improving the overall production efficiency.

Claims

1. A method for rapid loading and unloading of semiconductor factories, characterized in that: The steps include: Step 1: After machine A completes production, a material unloading signal is generated. After the material management server detects and obtains the material unloading signal, the following steps 2 to 8 are executed; Step 2: Obtain information of processed materials from machine A; Step 3: Determine the next station machine B based on the information of the processed materials; Step 4: Obtain site location information from the electronic map, the site location information includes the location information of machine A, the location information of material storage location C of machine A, and the location information of material storage location D of the next station machine B; Step 5: Generate unloading tasks, including information of processed materials and site location information; Step 6: Obtain the information of the unprocessed materials in the material storage location C, calculate the priority scores according to the process attributes, sort them, select the unprocessed materials with the highest priority as the materials to be processed, and record the priority scores of the materials to be processed; Step 7: Generate a material loading task, including the information of the material to be processed and the site location information, and mark the priority score of the material to be processed as the priority score of the material loading task; Step 8: Pack the loading task and unloading task into the same batch of task packages, and mark the priority score of the loading task as the priority score of the task package; Send the task packages to the task scheduling server in batches, and the task scheduling server executes the following steps 9 to 10; Step 9: Add the task packages to the task list and sort them according to the priority score; select the first task package in the task list to be sent to the AGV, and formulate the execution route of the AGV according to the site location information in the task package, and the execution route first goes to the material storage C; Step 10: All AGVs upload their location information to the task scheduling server in real time, and the task scheduling server generates scheduling information; Step 11: The task scheduling server randomly selects an idle AGV, namely AGV_C, and sends the highest priority task package and scheduling information in the task list to AGV_C. AGV_C executes the task according to the execution route and scheduling information in the task package.

2. A semiconductor factory rapid loading and unloading method as claimed in claim 1, characterized in that: When executing step 4, the electronic map is a factory electronic map preset in the material management server. Each machine is configured with a unique material storage location, and the geographical location of each machine and the geographical location of each material storage location are marked in the electronic map.

3. A semiconductor factory rapid loading and unloading method as claimed in claim 1, characterized in that: When executing step 6, the specific steps include: Step 6-1: The material management server obtains information of all unprocessed materials from the material storage location C; Step 6-2: Obtain the process attributes of each unprocessed material, including the placement time T i 、Emergency i and priority O i ; Step 6-3: Calculate the priority score for each unprocessed material using the following priority formula: Among them, P i is the priority score of the i-th unprocessed material; E i is the urgency of the i-th unprocessed material; i is the production sequence of unprocessed materials. The lower the value, the higher the production sequence priority. i is the placement time of the unprocessed material. The larger the value, the longer the placement time. w1, w2, and w3 are all weight coefficients. Step 6-4: Assign priority scores P to all unprocessed materials i Sort by Step 6-5: Select the unprocessed material with the highest priority score as the material to be processed; Step 6-6: Record the priority score of the material to be processed and mark it as the priority of the loading task.

4. A semiconductor factory rapid loading and unloading method as claimed in claim 1, characterized in that: When executing step 9, specifically, a task list is established. After obtaining a batch of task packages, the task packages of the batch are added to the task list and sorted according to the priority scores of the task packages of the batch; Select the task package ranked first, and formulate the execution route of the AGV according to the site location information in the task package. The execution route is to go to the material storage C first, then to the machine A, and finally to the material storage D.

5. The semiconductor factory rapid loading and unloading method according to claim 1, characterized in that: When executing step 10, the following steps are specifically included: Step 10-1: All AGVs upload their current location information to the task scheduling server in real time, and the system records the current location and motion status of each AGV; Step 10-2: The task scheduling server calculates the estimated time T for each AGV to enter the next path segment based on the AGV's location information. i,k ; Step 10-3: According to the length d of the path segment k and the AGV's travel speed v i , calculate the travel time T of each AGV in each path segment i,k ; Step 10-4: Assign a time window to each AGV on each path: TW i =[T i,start ,T i,end ]; Make sure the time windows of all AGVs do not overlap: Step 10-5: According to the arrangement of the time window, adjust the travel time of each AGV and generate the scheduling information of each AGV.

6. A semiconductor factory rapid loading and unloading method as claimed in claim 1, characterized in that: When executing step 11, the following steps are specifically included: Step 11-1: AGV_C goes to material storage C and takes out the materials to be processed; Step 11-2: AGV_C goes to machine A, takes out the processed material and puts the material to be processed; Step 11-3: AGV_C goes to material storage D to store the processed materials.

Images