Intelligent distribution system and distribution method for re-feeding of single crystal furnace

Abstract

The invention discloses an intelligent distribution system and method for re-feeding of a single crystal furnace, and belongs to the field of production automation of single crystal furnaces. The system comprises an MES system, an MCS system, AGV equipment and a middle cache region, the MES system and the MCS system are linked to push the required time of a coil base, the charging sequence is intelligently corrected in combination with historical transportation data, and materials are firstly conveyed to the nearby middle cache region and then distributed to the coil base. The distribution method comprises the steps of acquiring required time, calculating transportation time consumption, correcting a charging sequence, caching in different regions, distributing from the caching region to the coil base and the like. The problems of long-distance material waiting, transportation congestion and flow non-linkage in existing AGV distribution are solved, real-time linkage of demand and feeding is achieved, the feeding sequence is intelligently sorted through historical data, the congestion risk is isolated through a middle cache region, it is guaranteed that the coil base feeding time is controllable, the re-feeding distribution efficiency and accuracy are improved, the safety risk is reduced, and the product quality is improved. The method is suitable for large-scale single crystal furnace production scenes.

Description

Technical Field

[0001] The present invention relates to the technical field of single crystal silicon production automation, and specifically to an intelligent distribution system and distribution method for re-feeding a single crystal furnace, which is suitable for large-scale single crystal furnace production scenarios in the semiconductor and photovoltaic industries. Background Art

[0002] In the field of single crystal silicon material production, the continuous re-feeding process of the CZ method single crystal furnace places high real-time requirements on the logistics and distribution system. The re-feeding process requires maintaining the uniformity of the silicon liquid level and composition. However, delivery delays can lead to temperature imbalances within the furnace, causing process anomalies such as over-temperature boiling of the silicon liquid or instability of the crystallization interface. With the popularization of large-scale single crystal furnace clusters (more than 300 units) in the semiconductor and photovoltaic industries, the traditional logistics and distribution model faces the dual technical challenges of multi-furnace coordinated scheduling and long-distance route planning. The existing system has exposed significant lack of adaptability when responding to dynamic demand changes.

[0003] The existing technology relies on manual judgment in the distribution mode, which lacks real-time linkage with the furnace process parameters and cannot dynamically adjust the distribution timing based on the silicon liquid consumption rate, resulting in a natural time deviation between the feeding operation and the process requirements.

[0004] Manual handling cannot achieve accurate tracking of material locations, and redundancy in distribution paths is prone to occur in complex layouts with multiple furnaces. In addition, there is a lack of automatic obstacle avoidance mechanisms, and operational reliability is difficult to guarantee in high-temperature and high-dust environments.

[0005] The manpower-based distribution architecture is difficult to adapt to the linear growth of the number of furnaces. When the scale of furnaces exceeds the critical value, the complexity of manual scheduling increases exponentially, which cannot meet the scalability requirements of intelligent manufacturing for the logistics system.

[0006] ① Time and space matching defects of scheduling algorithms

[0007] Limitations of the static scheduling model: The scheduling strategy triggered by a unified signal does not take into account the differences in the spatial distribution of furnaces. Traditional path planning algorithms (such as the Dijkstra algorithm) only optimize the shortest distance and ignore the dynamic relationship between transportation time and demand urgency, resulting in a supply-demand time mismatch at remote furnaces.

[0008] Insufficient adaptation of the process window: The time accuracy required by the re-feeding process (error ≤ ±5 minutes) is an order of magnitude different from the traditional AGV scheduling cycle (10-15 seconds), which cannot meet the millisecond-level response requirements in high-temperature process environments.

[0009] ② Dynamic congestion problem in logistics network

[0010] Defects in the channel resource competition mechanism: Obstacle avoidance strategies based on fixed priorities (such as lidar + QR code navigation) cannot optimize paths in real time to avoid congestion points in multi-AGV collaborative scenarios, and often lead to deadlocks in logistics channels due to traffic conflicts at intersections.

[0011] Inherent defects of long-distance transportation: There is path redundancy in the direct transportation mode from the loading room to the furnace. When the transportation distance exceeds the critical value (usually more than 80 meters), the accumulated positioning error (±10mm / meter) will cause the delivery position deviation to exceed the process allowable range (±50mm).

[0012] ③ Information gap in business processes

[0013] Multi-system data interaction delay: In the link from the furnace demand signal being triggered by the sensor to the AGV execution, there is a data interaction delay between the MES and MCS systems (usually 8-12 seconds). Combined with the manual confirmation link, the overall response link is too long.

[0014] Lack of closed-loop status information: Traditional systems lack the ability to track materials throughout the entire process and are unable to obtain dynamic parameters such as buffer area capacity and AGV load status in real time. This results in a lack of complete information support for scheduling decisions and is prone to logical paradoxes in sequential scheduling (such as prioritizing the delivery of non-urgent materials). Summary of the Invention

[0015] The core objective of the present invention is to provide an intelligent distribution system and method for re-feeding a single crystal furnace, which solves the problems of long distance waiting for materials, transportation congestion and non-linked processes in existing AGV distribution, realizes real-time linkage between demand and feeding, intelligently sorts the feeding sequence through historical data, uses the intermediate buffer area to isolate the congestion risk, ensures the controllable loading time of the furnace, improves the efficiency and accuracy of re-feeding distribution, reduces safety risks, and is suitable for large-scale single crystal furnace production scenarios.

[0016] In a first aspect, an embodiment of the present invention provides a single crystal furnace recharging intelligent distribution system, comprising:

[0017] The MES system is used to receive the demand time of the terminal furnace and push it to the charging end, realizing real-time linkage between demand and charging;

[0018] The MCS system is used to receive MES signals to generate transportation tasks, record task time, and store the furnace area correspondence table;

[0019] AGV equipment is used to transport materials from the loading room to the intermediate buffer or from the buffer to the furnace;

[0020] The intermediate buffer area is divided into multiple areas according to the furnace layout and is used for temporary storage of materials to isolate the risk of transportation congestion.

[0021] In some embodiments of the present invention, the intermediate buffer area is divided into one area for every two rows of furnaces, and the area information is stored in the MCS system.

[0022] In some embodiments of the present invention, when the MCS system generates a task, it calls the furnace area correspondence table according to the furnace number and delivers the material to the corresponding area cache.

[0023] In some embodiments of the present invention, when the above-mentioned loading sequence is corrected, (t1-t) is calculated as the countdown. If (t1-t) is a negative value, emergency loading is triggered and the AGV is scheduled for priority delivery, where t1 is the current system time.

[0024] In some embodiments of the present invention, the above-mentioned AGV equipment adopts the Hikvision Q7 series and is equipped with an MCS system to realize task reception and time recording.

[0025] In some embodiments of the present invention, the MES system adopts Hualei Xuntuo Orbit-MOMR16, which is linked with the single crystal furnace centralized procurement system to collect demand signals.

[0026] In some embodiments of the present invention, the above system links historical transportation data with demand time and intelligently corrects the charging sequence to balance the waiting time of different furnaces.

[0027] In a second aspect, an embodiment of the present application provides a method for intelligently distributing re-feeding to a single crystal furnace, the steps of which include:

[0028] Collect furnace demand signals and transmit them to the MES system, which then pushes the demand time to the charging end;

[0029] Calculate the average time Δt of the furnace's latest 10 transportation tasks;

[0030] Calculate the corrected feeding time based on the formula t = t2 - Δt, and determine the feeding order by sorting by t, where t2 is the required delivery time;

[0031] The materials are first transported to the intermediate buffer location in the corresponding area and then delivered to the furnace.

[0032] In some embodiments of the present invention, when the above-mentioned loading sequence is corrected, (t1-t) is calculated as the countdown. If (t1-t) is a negative value, emergency loading is triggered and the AGV is scheduled for priority delivery, where t1 is the current system time.

[0033] In some embodiments of the present invention, the transportation time from the intermediate buffer area to the furnace is dynamically calibrated using the historical average time consumption Δt.

[0034] In some embodiments of the present invention, when distributing and caching materials in different regions, the MCS system is used to identify the corresponding regions of the materials, thereby reducing path waste.

[0035] The embodiments of the present invention have at least the following advantages or beneficial effects:

[0036] This invention establishes a millisecond-level response chain from furnace demand signal acquisition to AGV task generation through industrial protocol docking between the MES and MCS systems. Compared to the delays (8-12 seconds) caused by multi-link data interaction in traditional systems, this invention achieves real-time synchronization of demand information from sensors to the execution end, resolving the core contradiction of "demand transmission delay" and "process window mismatch" in the existing technology, making the distribution response mechanism highly compatible with the real-time requirements of the single crystal furnace high-temperature process.

[0037] Based on a dynamic calibration model (t = t² - Δt) based on historical transportation data, this method converts the spatial distance between furnaces into a scheduling parameter in the time dimension, breaking through the static "distance-first" scheduling limitations of traditional AGVs. By intelligently sorting the order of loading, it effectively balances the waiting time of different furnaces, resolving the "supply-demand time mismatch" problem at remote furnaces. The algorithm possesses self-learning capabilities and continuously optimizes the Δt parameter based on real-time transportation data, enabling scheduling strategies to adapt to changes in workshop layout or equipment performance degradation, achieving adaptive scheduling in dynamic scenarios.

[0038] This invention establishes a "segmented long-distance transport" logistics model by dividing the furnace cluster into logical areas and configuring nearby buffer zones. This architecture splits the long path from the charging room to the furnace into two shorter paths: "charging room-buffer zone" and "buffer zone-furnace zone." The buffer zone is used to isolate the risk of congestion around the charging room, improving the stability of the transport time from the buffer zone to the furnace to within ±5 minutes of the process requirements. Compared to the traditional direct transport model, this design completely solves the problem of "delivery delays caused by channel resource competition" and achieves efficient reuse of logistics channels.

[0039] This invention utilizes an emergency scheduling algorithm based on a (t1-t) countdown mechanism, automatically triggering task priority redistribution in unexpected operating conditions, such as abnormal furnace temperature. By interrupting non-urgent tasks and planning direct routes without avoidance, the algorithm improves the response speed of emergency material delivery by over 50%. This overcomes the inability of traditional fixed-priority scheduling to cope with dynamic operating conditions, providing a real-time safety mechanism for high-temperature processes in single crystal furnaces.

[0040] This invention utilizes a distributed zone partitioning and hierarchical scheduling algorithm, freeing system capacity from the linear growth constraints of the number of furnaces. When the furnace scale expands from 50 to 300, scheduling efficiency can be improved year-on-year simply by increasing the number of cache zones. This avoids the exponential increase in scheduling complexity associated with the increase in furnace numbers in traditional systems, meeting the stringent scalability requirements of gigafactories in the semiconductor and photovoltaic industries.

[0041] This invention, through the synergy of full-link time synchronization, dynamic scheduling, and congestion isolation, controls the arrival time error of re-feed materials within the process requirement of ±5 minutes, significantly reducing the risk of process anomalies such as silicon liquid boiling and crucible damage caused by delivery delays. Compared with traditional models, this technical solution fundamentally resolves the chain reaction of "delivery delay-process failure", providing stable logistics support for the single crystal silicon growth process and ensuring the process indicators of silicon rod equal diameter length and material utilization.

[0042] The "demand-sensing, data-driven, intelligent execution" logistics system constructed by this invention transcends the traditional AGV's single function of "command execution," achieving a technological leap from "passive response" to "active prediction." A scheduling model trained on historical data can predict potential congestion points in advance and dynamically adjust distribution strategies based on buffer capacity, enabling the workshop logistics system to make autonomous decisions and laying the underlying technical foundation for the unmanned, intelligent upgrade of single crystal furnace production. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be obtained based on these drawings without paying any creative work.

[0044] Figure 1 This is a schematic diagram of the overall process control of the present invention;

[0045] Figure 2 This is a schematic diagram of process control data of the present invention. DETAILED DESCRIPTION

[0046] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions of the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings herein can be arranged and designed in various different configurations.

[0047] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention as claimed, but rather merely represents selected embodiments of the present invention. All other embodiments derived by persons of ordinary skill in the art based on the embodiments of the present invention without creative effort are intended to fall within the scope of protection of the present invention.

[0048] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, it does not require further definition or explanation in subsequent drawings.

[0049] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprises" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device that includes a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further limitations, an element defined by the sentence "comprises a ..." does not exclude the presence of other identical elements in the process, method, article or device that includes the element.

[0050] In the description of the embodiments of the present invention, "a plurality of" means at least two.

[0051] In the description of the embodiments of the present invention, it should be noted that, unless otherwise expressly specified or limited, the terms "disposed," "installed," "connected," and "connected" should be understood in a broad sense. For example, they can refer to fixed connections, detachable connections, or integral connections; they can refer to mechanical connections or electrical connections; they can refer to direct connections or indirect connections through an intermediate medium; and they can refer to internal connections between two components. Those skilled in the art will understand the specific meanings of the above terms in the present invention based on specific circumstances.

[0052] like Figure 1-2 The embodiment of the present invention discloses an intelligent distribution system for re-feeding a single crystal furnace, which is characterized by comprising:

[0053] The MES system is used to receive the demand time of the terminal furnace and push it to the charging end, realizing real-time linkage between demand and charging;

[0054] The MCS system is used to receive MES signals to generate transportation tasks, record task time, and store the furnace area correspondence table;

[0055] AGV equipment is used to transport materials from the loading room to the intermediate buffer or from the buffer to the furnace;

[0056] The intermediate buffer area is divided into multiple areas according to the furnace layout and is used for temporary storage of materials to isolate the risk of transportation congestion.

[0057] On the other hand, an embodiment of the present invention discloses a method for intelligently distributing re-feeding to a single crystal furnace, which is characterized by comprising:

[0058] Collect furnace demand signals and transmit them to the MES system, which then pushes the demand time to the charging end;

[0059] Calculate the average time Δt of the furnace's latest 10 transportation tasks;

[0060] Calculate the corrected feeding time based on the formula t = t2 - Δt, and determine the feeding order by sorting by t, where t2 is the required delivery time;

[0061] The materials are first transported to the intermediate buffer location in the corresponding area and then delivered to the furnace.

[0062] Example 1:

[0063] This embodiment is based on a distributed cache distribution system divided into regions, wherein the intermediate cache area provided is divided into logical regions according to every two columns of furnaces, and the region information is stored in the furnace region mapping table of the MCS system;

[0064] AGV supports segmented transportation from "loading room to cache area" and "cache area to furnace", and achieves precise positioning of cache positions through RFID.

[0065] The MES system collects the furnace demand time t2 in real time, and the MCS system retrieves the average Δt of the corresponding furnace's 10 historical transportation times;

[0066] The order of adding materials is sorted by t=t2-Δt. When generating the transportation task, the area mapping table is matched according to the furnace number, and the materials are delivered to the corresponding buffer area first.

[0067] Comparative Example 1:

[0068] Traditional AGV direct delivery system

[0069] AGV transports materials directly from the loading room to the furnace without any intermediate buffer area, and uses LiDAR + QR code navigation;

[0070] After receiving the demand signal, the MES system directly generates a direct task and schedules it according to the fixed rule of "distance priority".

[0071] Technical defects:

[0072] Long-distance transportation causes the waiting time of the remote furnace to exceed the process critical value, with a time error of ±15 minutes;

[0073] Competition for channel resources is likely to cause congestion, and a single congestion during peak hours can cause subsequent tasks to be delayed by 10-25 minutes.

[0074] Example 2:

[0075] This embodiment is based on the intelligent scheduling method of the time correction model, in which the correction charging time t = t2-Δt, where Δt is the average of the latest 10 transportation times of the furnace;

[0076] The countdown parameter (t1-t) is calculated in real time. If it is a negative value, the emergency dispatch flag is triggered and the AGV task priority is increased to the highest level.

[0077] The task execution logic is as follows: the MCS system generates a task queue based on the t-value sorting and dynamically adjusts the AGV driving path; emergency tasks use a non-avoidance direct path algorithm and interrupt non-emergency tasks to release channel resources.

[0078] Comparative Example 2:

[0079] Semi-automated delivery mode with manual intervention

[0080] Manual transportation from the charging room to the temporary buffer area is carried out by humans, while AGV performs delivery from the buffer area to the furnace.

[0081] The MES system only records demand signals, and the order of adding materials relies on manual experience and judgment, and is not linked to AGV scheduling.

[0082] Technical defects:

[0083] Manual handling results in a delay of 15-25 seconds from demand signal to execution, which cannot match the real-time requirements of high-temperature processes.

[0084] The lack of a data-driven scheduling model makes it easy for the logical paradox of "urgent material post-distribution" to occur in multi-furnace scenarios.

[0085] Example 3:

[0086] This embodiment is based on a hierarchical scheduling system for a large-scale furnace cluster: 300 furnaces are divided into 15 cache areas, each area is configured with an independent sub-scheduling system;

[0087] It adopts a two-tier architecture of "regional scheduling-central coordination" to achieve cross-regional data interaction through industrial Ethernet.

[0088] Among them, a priority function is constructed based on three-dimensional parameters: demand urgency, AGV load rate, and buffer area capacity;

[0089] Distributed computing nodes are used to parallel process task scheduling and update area mapping tables and path planning in real time.

[0090] Comparative Example 3:

[0091] Traditional AGV scheduling system with centralized cache

[0092] A centralized buffer area is set up near the charging room, and AGV performs transportation from "charging room-centralized buffer-furnace";

[0093] Fixed area division is adopted, and cache allocation is not dynamically adjusted according to the furnace layout.

[0094] Technical defects:

[0095] The long-distance transportation from the centralized buffer area to the furnace still poses a risk of congestion, and the buffer area only serves as a temporary storage point;

[0096] Fixed area division leads to redundant paths for AGVs from the buffer area to the furnace, and the average transportation time increases by 8-12 minutes compared to the direct mode.

[0097] In summary, the embodiment of the present invention has the following advantages compared with the comparative example:

[0098] 1. The distributed cache architecture enables segmented transportation routes, isolating the risk of long-distance congestion;

[0099] 2. The time correction algorithm converts spatial distance into time-dimensional scheduling parameters to resolve supply and demand mismatches;

[0100] 3. The hierarchical scheduling model supports linear expansion of large-scale clusters, avoiding the exponential growth of complexity in traditional systems.

[0101] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. An intelligent distribution system for re-feeding a single crystal furnace, characterized in that: include: The MES system is used to receive the demand time of the terminal furnace and push it to the charging end, realizing real-time linkage between demand and charging; The MCS system is used to receive MES signals to generate transportation tasks, record task time, and store the furnace area correspondence table; AGV equipment is used to transport materials from the loading room to the intermediate buffer or from the buffer to the furnace; The intermediate buffer area is divided into multiple areas according to the furnace layout and is used for temporary storage of materials to isolate the risk of transportation congestion.

2. A single crystal furnace re-feeding intelligent distribution method, characterized in that: include: Collect furnace demand signals and transmit them to the MES system, which then pushes the demand time to the charging end; Calculate the average time Δt of the furnace's latest 10 transportation tasks; Calculate the corrected feeding time based on the formula t = t2 - Δt, and determine the feeding order by sorting by t, where t2 is the required delivery time; The materials are first transported to the intermediate buffer location in the corresponding area and then delivered to the furnace.

3. The intelligent distribution system for recharging single crystal furnace according to claim 1, characterized in that: The intermediate buffer area is divided into one area for every two rows of furnaces, and the area information is stored in the MCS system.

4. The intelligent distribution system for recharging single crystal furnace according to claim 1, characterized in that: When the MCS system generates a task, it calls the furnace area correspondence table according to the furnace number and delivers the materials to the corresponding area cache.

5. The method for intelligently distributing re-feeding to a single crystal furnace according to claim 2, characterized in that: When correcting the feeding sequence, (t1-t) is calculated as the countdown. If (t1-t) is a negative value, emergency feeding is triggered and the AGV is scheduled for priority delivery, where t1 is the current system time.

6. The intelligent distribution system for recharging single crystal furnace according to claim 1, characterized in that: The AGV equipment adopts Hikvision Q7 series and is equipped with MCS system to realize task reception and time recording.

7. The intelligent distribution system for recharging single crystal furnace according to claim 1, characterized in that: The MES system uses Hualei Xuntuo Orbit-MOMR16, which is linked with the single crystal furnace centralized procurement system to collect demand signals.

8. The method for intelligently distributing re-feeding to a single crystal furnace according to claim 2, characterized in that: The transportation time from the intermediate buffer area to the furnace is dynamically calibrated using the historical average time consumption Δt.

9. The method for intelligently distributing re-feeding to a single crystal furnace according to claim 2, characterized in that: When distributing and caching by area, the corresponding area of ​​the material is identified through the MCS system.

10. The intelligent distribution system for recharging single crystal furnace according to claim 1, characterized in that: The system links historical transportation data with demand time and intelligently corrects the charging sequence to balance the waiting time of different furnaces.

Images