An intelligent buffer machine dynamic scheduling method and system for an SMT production line

Abstract

This invention relates to the field of intelligent scheduling technology for SMT production lines, and in particular to a dynamic scheduling method and system for intelligent buffer machines in SMT production lines. The method includes real-time acquisition of operating status data from each workstation, construction of a global load status vector, dynamic generation of a scheduling instruction set including buffer priority, release sequence, and transport path, and continuous optimization of the scheduling strategy through a buffer monitoring and feedback adjustment mechanism. This application enables comprehensive perception and dynamic response to the overall production line status, effectively alleviating problems such as cycle time imbalance, board congestion, and path conflicts in multi-model mixed-line production, and improving buffer resource utilization efficiency and production line collaborative operation capabilities.

Description

Technical Field

[0001] This invention relates to the field of electronic manufacturing automation technology, and more specifically, to a method and system for dynamic scheduling of intelligent buffer machines for SMT production lines. Background Technology

[0002] Against the backdrop of the electronics manufacturing industry's shift towards high-density, high-reliability, and flexible production, SMT production lines are placing higher demands on cycle time coordination and equipment collaborative control between various processes. To alleviate cycle time differences between processes such as placement, testing, and subsequent assembly, buffer machines are widely installed in production lines to perform temporary storage and transfer functions for boards. Their operating efficiency directly affects the continuity and capacity utilization level of the entire line.

[0003] Currently, the control methods for buffer machines in SMT production lines mostly adopt preset logic rules or fixed scheduling strategies, such as executing the buffering, release, or transfer operations of boards according to a predetermined order. This approach can maintain a basic production rhythm when production tasks are relatively simple and operating conditions are relatively stable. However, under complex operating conditions such as multi-model mixed-line production, frequent order changes, or fluctuations in the status of local equipment, the existing control strategies have limited responsiveness to real-time load changes at each workstation. When upstream placement equipment adjusts its cycle time, downstream testing or assembly equipment experiences short-term downtime, or when additional orders are inserted, the buffer machine, lacking a comprehensive perception and dynamic scheduling mechanism for the overall production status, may experience problems such as uneven buffer space allocation, excessive board dwell time, or repeated transport paths. This can lead to localized congestion, equipment idling, or the entire line waiting, affecting the overall operating efficiency and intelligence level of the SMT production line.

[0004] Therefore, there is a need for an intelligent cache machine control method and system that can integrate production status perception, dynamic allocation of cache resources and scheduling decision linkage mechanism to improve the collaborative operation capability and resource utilization efficiency of SMT production line under complex working conditions. Summary of the Invention

[0005] In view of this, the present invention proposes a dynamic scheduling method and system for intelligent buffer machines in SMT production lines. It aims to solve the problem that the existing buffer machine control strategies in the current technology rely on preset logic rules or fixed scheduling sequences. Under complex working conditions such as multi-model mixed production lines, frequent order switching, or fluctuations in the status of local equipment, they cannot comprehensively perceive and dynamically respond to the overall production line status, resulting in uneven buffer space allocation, excessive board dwell time, and repeated handling paths, which in turn cause local congestion, equipment idling, or waiting of the entire line.

[0006] This invention proposes a dynamic scheduling method for intelligent buffer machines in SMT production lines, comprising:

[0007] Real-time acquisition of operating status data of each station in the SMT production line, including the cycle output frequency of the upstream placement equipment, the operating status indicators of the downstream testing and assembly equipment, and the insertion signal of the insertion task;

[0008] Based on the operational status data, a global load status vector for the production line is constructed. The global load status vector includes the real-time load index, task queue length, and equipment availability flag for each workstation.

[0009] The dynamic scheduling instruction set of the cache machine is obtained based on the global load state vector. The dynamic scheduling instruction set includes board cache priority, release timing and handling path planning.

[0010] The dynamic scheduling instruction set is sent to the execution unit of the cache machine to control the receiving, temporary storage, release and transfer operations of the cache machine execution board;

[0011] During the operation of the cache machine, the dwell time and location distribution of each board in the cache area are continuously monitored to generate a cache resource occupancy status diagram;

[0012] Based on the matching degree between the cache resource occupancy state diagram and the global load state vector, the dynamic scheduling instruction set for the next scheduling cycle is adjusted.

[0013] Furthermore, when collecting real-time operational status data for each workstation in the SMT production line, this includes:

[0014] The status messages output by the device controller are obtained by the industrial communication interface modules deployed at each workstation at a predetermined sampling period.

[0015] Parse the status message to extract the device operating mode, task completion count, fault code, and number of pending tasks;

[0016] The extracted data is encapsulated into standardized status data packets and transmitted to the central dispatch server via industrial Ethernet.

[0017] Furthermore, when constructing the global load state vector for the production line, the following is included:

[0018] Assign a unique workstation identifier to each workstation;

[0019] Establish a workstation index table based on the workstation identifier;

[0020] The real-time load index of each workstation is defined as the ratio of the number of tasks to be processed to the rated processing capacity of the equipment.

[0021] Set the device availability flag to a binary value, where the first value indicates that the device is in an operational state and the second value indicates that the device is in a shutdown or maintenance state.

[0022] Arrange the real-time load index, task queue length, and equipment availability flags of all workstations in order according to the workstation index table to form a global load status vector.

[0023] Furthermore, when obtaining the dynamic scheduling instruction set of the cache machine, it includes:

[0024] Identify upstream high-load workstations and downstream low-load workstations based on global load state vectors;

[0025] When the real-time load index of the upstream high-load station is greater than the first preset threshold and the equipment availability flag of the downstream low-load station is the first value, a board caching instruction is generated to temporarily store the board in the specified cache area of ​​the cache machine.

[0026] When the real-time load index of the downstream workstation is less than the second preset threshold and its equipment availability flag is the first value, a board release command is generated, the corresponding board is extracted from the cache machine and sent to the downstream workstation.

[0027] When an order insertion signal is detected, the cache priority of all boards is reacquired. The priority is determined by the urgency level of the order to which the board belongs and the current retention time.

[0028] Furthermore, when generating the cache resource usage state graph, the following is included:

[0029] The cache machine is divided into multiple logical cache units, each with unique location coordinates.

[0030] Record the location coordinates, entry timestamp, and order identifier of each component in the cache machine;

[0031] The dwell time for each component is obtained as the difference between the current system time and the entry timestamp;

[0032] Map the occupancy status of all logical cache units, board dwell time, and order identifier to a two-dimensional grid to form a cache resource occupancy status diagram.

[0033] Furthermore, when adjusting the dynamic scheduling instruction set for the next scheduling cycle, the following are included:

[0034] Get the percentage of boards whose dwell time exceeds the third preset threshold in the cache resource usage status graph;

[0035] If the proportion is greater than the fourth preset threshold, the release priority of this type of board will be increased in the next scheduling cycle;

[0036] Obtain the number of workstations in the global load status vector whose load index changes more than the fifth preset threshold within a predetermined number of consecutive sampling periods.

[0037] If the quantity exceeds the sixth preset threshold, the path optimization algorithm will be activated in the next scheduling cycle to replan the board handling path and avoid repeatedly passing through high-fluctuation workstations.

[0038] Compared with existing technologies, the advantages of this invention are as follows: By collecting real-time operating status data of each workstation in the SMT production line and constructing a global load status vector, a comprehensive perception capability of the entire production line's status is established. This perception mechanism no longer relies on fixed scheduling rules, but dynamically generates a set of scheduling instructions for the buffer machine based on the cycle output frequency of the upstream placement equipment, the operating status indicators of the downstream inspection and assembly equipment, and the insertion signal of the order insertion task. In specific implementation, the system obtains the status messages of the equipment controller through the industrial communication interface module at a predetermined sampling period, and parses out the equipment operating mode, task completion count, and number of pending tasks, ensuring the real-time performance and accuracy of the data source. Based on the global load status vector, the scheduling decision module can identify upstream high-load workstations and downstream low-load workstations, and generate board buffer or release instructions accordingly, effectively mitigating the cycle time differences between processes. When an order insertion task is detected, the system re-acquires the board buffer priority, which is determined by the order urgency indicator and the current dwell time, thereby ensuring the timely processing of high-priority tasks. The cache monitoring module continuously tracks the location coordinates and dwell time of each board in the cache area using a visual recognition camera and RFID reader, generating a cache resource occupancy status diagram to provide data support for scheduling optimization. The feedback adjustment module dynamically adjusts the release priority and handling path in the next scheduling cycle based on the matching degree between the cache resource occupancy status diagram and the global load status vector, avoiding excessive board dwell time and duplicate handling paths.

[0039] On the other hand, this application also provides a dynamic scheduling system for intelligent buffer machines in an SMT production line, comprising:

[0040] The status acquisition module is used to collect real-time operating status data of each station in the SMT production line. The status acquisition module is connected to the equipment controller of each station through an industrial communication interface.

[0041] The load modeling module is electrically connected to the status acquisition module and is used to construct a global load status vector of the production line based on the operating status data.

[0042] The scheduling decision module is electrically connected to the load modeling module and is used to obtain the dynamic scheduling instruction set of the cache machine based on the global load state vector.

[0043] An execution control module, electrically connected to the scheduling decision module, is used to send dynamic scheduling instruction sets to the execution unit of the cache machine;

[0044] The cache monitoring module is electrically connected to the execution control module and is used to monitor the dwell time and location distribution of each board in the cache area and generate a cache resource occupancy status diagram.

[0045] The feedback adjustment module is electrically connected to the cache monitoring module and the load modeling module, respectively, and is used to adjust the dynamic scheduling instruction set of the next scheduling cycle based on the matching degree between the cache resource occupancy status diagram and the global load status vector.

[0046] Furthermore, the status acquisition module is configured to poll the controllers of each workstation at fixed time intervals and to encapsulate and transmit data using the OPC UA protocol.

[0047] Furthermore, the load modeling module is configured as a maintenance workstation index table, which stores the physical location, processing capacity parameters and communication address of each workstation, and is used to perform structured mapping on the collected raw status data.

[0048] Furthermore, the scheduling decision module has a built-in priority acquisition unit and a path planning unit. The priority acquisition unit obtains the cache priority based on the order urgency identifier and the board dwell time. The path planning unit uses A... ∗ The algorithm generates conflict-free transport paths.

[0049] Furthermore, the execution control module is connected to the servo driver, pneumatic gripper, and position sensor of the buffer machine via hardwiring. After receiving the dynamic scheduling instruction set, it generates corresponding motor control signals, gripper opening and closing instructions, and position verification requests.

[0050] Furthermore, the cache monitoring module obtains the board location coordinates and order identifiers through a visual recognition camera and an RFID reader installed inside the cache machine. The visual recognition camera collects images of the cache area at a preset frequency, and the RFID reader automatically reads the tag information when the board enters or leaves the cache area.

[0051] Furthermore, the feedback adjustment module is configured to obtain the retention anomaly rate in the cache resource occupancy status graph and the load volatility rate in the global load status vector at the end of each scheduling cycle, and decide whether to enable the advanced scheduling strategy based on whether the product of the two exceeds the seventh preset threshold.

[0052] It is understood that the intelligent buffer dynamic scheduling method and system for SMT production lines in the above embodiments of the present invention have the same beneficial effects, and will not be described in detail here. Attached Figure Description

[0053] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings:

[0054] Figure 1 A flowchart illustrating a dynamic scheduling method for an intelligent buffer machine in an SMT production line, provided by an embodiment of the present invention;

[0055] Figure 2 A flowchart illustrating a dynamic scheduling method for an intelligent buffer machine in an SMT production line, provided by an embodiment of the present invention;

[0056] Figure 3 This is a functional block diagram of an intelligent buffer dynamic scheduling system for an SMT production line, provided as an embodiment of the present invention. Detailed Implementation

[0057] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the disclosure to those skilled in the art. It should be noted that, unless otherwise specified, embodiments and features in the embodiments of the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0058] When introducing the intelligent buffer dynamic scheduling method and system for SMT production lines according to the present invention, it is important to note that the technical solution adopted in this application includes:

[0059] A buffer is an intermediate device placed between different processes on a production line. It is used to temporarily store, sort, and control the transfer of boards that are to be processed or have already been processed. Through the buffering and releasing functions of the buffer, waiting or congestion problems caused by inconsistent cycle times between upstream and downstream equipment can be alleviated. In SMT production lines, it is usually used as a key scheduling node for board flow.

[0060] A surface mount technology (SMT) production line refers to an automated production line used for surface mount technology. It typically includes multiple processes such as printing, placement, reflow soldering, inspection, and post-assembly. This production line is characterized by high cycle time, strong equipment interoperability, and high requirements for production scheduling and status control.

[0061] Binary values ​​are digital representations of information using two states, "0" and "1," and are commonly used to describe device status, signal triggering conditions, or logical judgment results. In automated control systems, binary values ​​are typically used to indicate device start / stop, status validity, or command execution results.

[0062] Upstream refers to the production unit located before the current equipment or process in the production flow, whose output directly serves as the input source for the current process. In an SMT production line, the operating status and output cycle time of upstream equipment directly affect the scheduling strategies of buffer machines and subsequent processes.

[0063] Downstream refers to the production unit located after the current equipment or process in the production flow. Its processing capacity and operating status determine the board release rhythm of the current process or buffer machine. When downstream equipment experiences congestion or shutdown, it is often necessary to adjust the board flow through buffering strategies.

[0064] The OPC UA protocol is an open communication protocol for industrial automation, used to enable data interaction and status sharing between different devices, systems, and platforms. This protocol supports standardized data modeling and secure communication, and is commonly used in SMT production lines to achieve equipment status acquisition, control command issuance, and system-level collaborative control.

[0065] A * The algorithm is an optimal path or optimal decision search algorithm based on heuristic search ideas. It evaluates multiple candidate solutions and prioritizes the optimal solution by comprehensively considering the current cost and the expected cost of the objective. In the field of automation control and scheduling, A... * The idea of ​​algorithms is often used in the decision-making process of optimizing path selection or scheduling order.

[0066] An RFID reader is an information acquisition device used in radio frequency identification systems to read or write information from RFID tags attached to items. In SMT production lines, RFID readers are typically used to automatically identify and track board identities, process flow status, or order information, providing data support for scheduling and control.

[0067] like Figure 1 As shown in some embodiments of this application, this embodiment provides a dynamic scheduling method for intelligent buffer machines in an SMT production line, including:

[0068] Step S100: Real-time acquisition of operating status data of each station in the SMT production line. The operating status data includes the cycle output frequency of the upstream placement equipment, the operating status indicators of the downstream testing equipment and assembly equipment, and the insertion signal of the insertion task.

[0069] Specifically, when collecting real-time operating status data of each workstation in the SMT production line, the process includes: acquiring status messages output by the equipment controller at a predetermined sampling period through industrial communication interface modules deployed at each workstation; parsing the status messages to extract equipment operating mode, task completion count, fault codes, and the number of tasks to be processed; encapsulating the extracted data into standardized status data packets and transmitting them to the central scheduling server via industrial Ethernet.

[0070] Understandably, by deploying industrial communication interface modules at key workstations on the SMT production line, the system polls the equipment controller at fixed time intervals to obtain status messages containing equipment operating modes, task completion counts, fault codes, and the number of tasks pending. This process uses the OPC UA protocol to encapsulate the raw data into structured status data packets, which are then transmitted to the central scheduling server via industrial Ethernet. This mechanism ensures the real-time performance, integrity, and format consistency of the operating status data, providing a reliable data input source for subsequent global load modeling.

[0071] In a specific embodiment of this application, the above steps are implemented as follows: On an SMT production line including a pick-and-place machine, an AOI inspection machine, a reflow oven, and a post-assembly machine, an industrial communication interface module is configured on the controller side of each device. Each device is polled with a sampling period of 500ms to obtain its current operating status. For example, the status message returned by the pick-and-place machine includes fields such as "Operating Mode: Continuous Operation," "Task Completion Count: 128," and "Number of Tasks to be Processed: 35"; the AOI inspection machine returns "Operating Mode: Idle," "Fault Code: 0x00," and "Number of Tasks to be Processed: 0." This data is parsed and encapsulated into a JSON-formatted status data packet, which is then sent to the central dispatch server via gigabit industrial Ethernet. This example demonstrates the complete acquisition and transmission link of operating status data from physical devices to the central dispatch system.

[0072] The above scenarios are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

[0073] Step S200: Construct a global load status vector for the production line based on the operating status data. The global load status vector includes the real-time load index, task queue length, and equipment availability flag for each workstation.

[0074] Specifically, constructing the global load state vector of the production line includes: assigning a unique workstation identifier to each workstation; establishing a workstation index table based on the workstation identifiers; defining the real-time load index of each workstation as the ratio of the number of tasks to be processed to the rated processing capacity of the equipment; setting the equipment availability flag to a binary value, where the first value indicates that the equipment is in an operational state and the second value indicates that the equipment is in a shutdown or maintenance state; and arranging the real-time load index, task queue length, and equipment availability flag of all workstations in the order of the workstation index table to form the global load state vector.

[0075] Understandably, after receiving the status data packets from each workstation, the central dispatch server performs structured mapping on the data according to a pre-defined workstation index table. The workstation index table stores the physical location, communication address, and rated processing capacity parameters of each workstation. The system obtains the real-time load index for each workstation by dividing the number of tasks to be processed by its rated processing capacity (unit: boards / minute), and sets the equipment availability flag to 1 (running) or 0 (downtime). Finally, the real-time load indexes, task queue lengths, and equipment availability flags of all workstations are concatenated according to the index table order to form a fixed-dimensional global load status vector, which serves as the unified input for scheduling decisions.

[0076] In a specific embodiment of this application, the above steps are implemented as follows: A certain SMT production line contains 5 workstations, whose workstation indexes are as follows: W1 (Pick-and-place machine, rated capacity 60 boards / hour), W2 (AOI, 90 boards / hour), W3 (Reflow soldering, 120 boards / hour), W4 (Functional testing, 75 boards / hour), and W5 (Assembly machine, 50 boards / hour). At a certain sampling time, the number of tasks to be processed at each workstation is 30, 10, 5, 20, and 25, respectively. The system obtains the real-time load index: W1 is 30 / (60 / 60)=30, W2 is 10 / (90 / 60)≈6.67, W3 is 5 / (120 / 60)=2.5, W4 is 20 / (75 / 60)=16, and W5 is 25 / (50 / 60)=30. Assuming that all equipment is fault-free, the equipment availability flag is 1 for all equipment. The final global load state vector is [30,30,1,6.67,10,1,2.5,5,1,16,20,1,30,25,1]. This vector fully describes the load distribution of the entire line at that moment.

[0077] The above scenarios are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

[0078] Step S300: Obtain the dynamic scheduling instruction set of the cache machine according to the global load state vector. The dynamic scheduling instruction set includes board cache priority, release timing and transportation path planning.

[0079] Specifically, when acquiring the dynamic scheduling instruction set of the cache machine, the process includes: identifying upstream high-load workstations and downstream low-load workstations based on the global load state vector; when the real-time load index of the upstream high-load workstation is greater than a first preset threshold and the equipment availability flag of the downstream low-load workstation is a first value, generating a board caching instruction to temporarily store the board in a designated cache area of ​​the cache machine; when the real-time load index of the downstream workstation is less than a second preset threshold and its equipment availability flag is a first value, generating a board release instruction to extract the corresponding board from the cache machine and transport it to the downstream workstation; when an order insertion signal is detected, re-acquiring the cache priority of all boards, the priority being jointly determined by the urgency flag of the order to which the board belongs and the current dwell time.

[0080] Understandably, after receiving the global load status vector, the scheduling decision module first identifies upstream high-load workstations (such as pick-and-place machines) and downstream low-load workstations (such as assembly machines). If the load index of the upstream workstation exceeds a threshold (e.g., 25) and the downstream workstation is available, a cache instruction is generated, instructing the cache unit to receive the board from the upstream and temporarily store it in a designated logical area. Conversely, if the load index of the downstream workstation is below a threshold (e.g., 10) and is available, a release instruction is generated, retrieving the corresponding board from the cache and sending it to the next station. When the system receives an order insertion signal, the priority acquisition unit obtains a new cache priority based on the order urgency level indicator (e.g., level 0-5) and the current dwell time of the board (in seconds), and the path planning unit then calls...

[0081] A ∗ The algorithm generates conflict-free transport paths.

[0082] In a specific embodiment of this application, the above steps are implemented as follows: In the aforementioned global load state vector, the load index of W1 is 30 > 25 (first threshold), the load index of W5 is 30 > 10, but the load index of W4 is 16 > 10, while the load index of W3 is 2.5 < 10 and its availability is 1. Therefore, the system determines that W3 is a low-load downstream workstation, generates a release instruction, and retrieves the board to be sent to W3 from the cache machine. At the same time, due to the high load of W1, a cache instruction is generated to receive newly produced boards. At this time, the system receives an insertion task with an urgency level of 5 (highest), and the corresponding board has been in the cache for 120 seconds. The priority acquisition unit obtains the priority according to the formula P = 0.7 × E + 0.3 × (T / Tmax), where E = 5, T = 120, Tmax = 300, resulting in P = 0.7 × 5 + 0.3 × 0.4 = 3.62. This board is assigned the highest cache priority, and the path planning unit plans the shortest conflict-free path for it.

[0083] The above scenarios are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

[0084] Step S400: The dynamic scheduling instruction set is sent to the execution unit of the cache machine to control the receiving, temporary storage, release and transfer operations of the cache machine execution board.

[0085] Specifically, the execution control module is connected to the servo driver, pneumatic gripper, and position sensor of the buffer unit via hardwiring. After receiving the dynamic scheduling instruction set, it generates corresponding motor control signals, gripper opening and closing instructions, and position verification requests.

[0086] Understandably, the execution control module connects to the underlying execution mechanism of the buffer unit via a hard-wired interface. Upon receiving a dynamic scheduling instruction set, the module parses the instruction content, generates pulse signals to control the movement of the X / Y / Z axis servo motors, outputs digital signals to control the opening and closing of the pneumatic gripper, and sends a verification request to the position sensor to confirm the board's position. This process ensures that the scheduling instructions are accurately translated into physical actions, achieving precise handling and positioning of the board within the buffer unit.

[0087] In a specific embodiment of this application, the above steps are implemented as follows: After receiving the instruction "receive the board from the inlet to the buffer area (3,2)", the execution control module sends 12,000 pulses (corresponding to a 300mm stroke) to the X-axis servo driver and 8,000 pulses (corresponding to a 200mm stroke) to the Y-axis. When the buffer machine's robotic arm moves above the target position, the module outputs a high-level signal to cause the pneumatic gripper to close and grasp the board. Subsequently, a verification request is sent to the photoelectric sensor installed at position (3,2) to confirm that the board has been correctly placed. The entire process is completed within 2.5 seconds, with an error of less than ±0.1mm.

[0088] The above scenarios are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

[0089] Step S500: During the operation of the cache machine, continuously monitor the dwell time and location distribution of each board in the cache area, and generate a cache resource occupancy status diagram.

[0090] Specifically, generating the cache resource occupancy status diagram includes: dividing the cache machine into multiple logical cache units, each with unique location coordinates; recording the location coordinates, entry timestamp, and order identifier of each board in the cache machine; obtaining the dwell time of each board as the difference between the current system time and the entry timestamp; and mapping the occupancy status of all logical cache units, board dwell time, and order identifier to a two-dimensional grid to form the cache resource occupancy status diagram.

[0091] Understandably, the cache monitoring module works in conjunction with an RFID reader / writer via a visual recognition camera installed inside the cache unit. The visual camera captures images of the cache area at a frequency of 1Hz and identifies the occupancy status of each logical cache unit through image segmentation; the RFID reader / writer automatically reads the order identifier and entry timestamp from the tag when a board enters or leaves. The system maps the occupancy status (empty / occupied), dwell time (seconds), and order ID of each unit to a two-dimensional array consistent with the physical layout, forming a cache resource occupancy status diagram.

[0092] In a specific embodiment of this application, the above steps are implemented as follows: A cache machine is internally divided into 48 logical cache units (6 rows × 8 columns), with coordinates from (1,1) to (6,8). At a certain moment, unit (3,2) is occupied. The RFID reader reads the order ID as ORD-20240515-087, with an entry timestamp of 1715769200 (Unix time). The current system time is 1715769320, and the dwell time is 120 seconds. The vision system confirms that a board exists in this unit. The system writes this information into the 3rd row and 2nd column of a two-dimensional array, marked as {occupied:true,dwell_time:120,order_id:"ORD-20240515-087"}. The remaining units are processed in the same way, ultimately generating a complete cache resource occupancy status diagram.

[0093] The above scenarios are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

[0094] Step S600: Based on the matching degree between the cache resource occupancy status diagram and the global load status vector, adjust the dynamic scheduling instruction set for the next scheduling cycle.

[0095] Specifically, when adjusting the dynamic scheduling instruction set for the next scheduling cycle, the following steps are taken: obtaining the percentage of boards whose dwell time exceeds a third preset threshold in the cache resource occupancy status graph; if the percentage is greater than a fourth preset threshold, then increasing the release priority of such boards in the next scheduling cycle; obtaining the number of workstations in the global load status vector whose load index changes more than a fifth preset threshold within a predetermined number of consecutive sampling periods; if the number is greater than a sixth preset threshold, then enabling the path optimization algorithm in the next scheduling cycle to replan the board handling path and avoid repeatedly passing through high-fluctuation workstations.

[0096] Understandably, at the end of each scheduling cycle, the feedback adjustment module synchronously receives the cache resource occupancy status diagram and the latest global load status vector. The module first calculates the percentage of time-out boards (e.g., >300 seconds). If this percentage exceeds 10% (the fourth threshold), such boards are prioritized for release in the next cycle. Simultaneously, the module analyzes the standard deviation of the load index for each workstation over the past five sampling cycles. If the number of workstations with a standard deviation >5 exceeds two (the sixth threshold), it indicates severe production line fluctuations. The module then activates a path optimization algorithm to avoid transport channels near highly volatile workstations, reducing path conflicts.

[0097] In a specific embodiment of this application, the above steps are implemented as follows: At the end of the current scheduling cycle, the cache resource occupancy status diagram shows that 8 boards out of 48 units have been delayed for more than 300 seconds, accounting for 16.7% > 10%. The feedback adjustment module generates an instruction to raise the release priority of these 8 boards to the highest level in the next cycle. At the same time, historical data from the global load status vector shows that the standard deviations of the load indices of W1 and W4 in the past 5 cycles are 6.2 and 5.8, respectively, both > 5, and the number is 2, which is equal to the sixth threshold (set to 2). The system determines that path optimization needs to be enabled, and the path planning unit re-obtains the transport path, avoiding the intersection area between the W1 exit and the W4 entrance, and uses the backup channel instead. The adjusted dynamic scheduling instruction set is sent to the execution control module.

[0098] The above scenarios are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

[0099] Step S700, repeat steps S100 to S600 to form a closed-loop scheduling control process.

[0100] Understandably, this method constitutes a closed-loop control process through six stages: status acquisition, load modeling, scheduling decision-making, execution control, cache monitoring, and feedback adjustment. Within each scheduling cycle (e.g., 1 second), the system completes a full data acquisition, analysis, decision-making, and execution process, and dynamically adjusts the strategy for the next cycle based on the execution results, achieving adaptive response to complex operating conditions on the SMT production line.

[0101] In the above embodiments, by collecting real-time operating status data of each workstation on the SMT production line and constructing a global load status vector, a comprehensive perception capability of the entire production line's status is established. This perception mechanism no longer relies on fixed scheduling rules, but dynamically generates a set of scheduling instructions for the buffer machine based on the cycle output frequency of the upstream placement equipment, the operating status identifiers of the downstream testing and assembly equipment, and the insertion signals of order insertion tasks. In specific implementation, the system obtains the status messages of the equipment controller through the industrial communication interface module at a predetermined sampling period, and parses out the equipment operating mode, task completion count, and number of pending tasks, ensuring the real-time performance and accuracy of the data source. Based on the global load status vector, the scheduling decision module can identify upstream high-load workstations and downstream low-load workstations, and generate board buffer or release instructions accordingly, effectively mitigating the cycle time differences between processes. When an order insertion task is detected, the system re-acquires the board buffer priority, which is determined by the order urgency identifier and the current dwell time, thereby ensuring the timely processing of high-priority tasks. The cache monitoring module continuously tracks the location coordinates and dwell time of each board in the cache area using a visual recognition camera and RFID reader, generating a cache resource occupancy status diagram to provide data support for scheduling optimization. The feedback adjustment module dynamically adjusts the release priority and handling path in the next scheduling cycle based on the matching degree between the cache resource occupancy status diagram and the global load status vector, avoiding excessive board dwell time and duplicate handling paths.

[0102] In another preferred embodiment based on the above embodiments, such as Figure 2 As shown, this embodiment provides an intelligent buffer dynamic scheduling system for SMT production lines, including: a status acquisition module, a load modeling module, a scheduling decision module, an execution control module, a buffer monitoring module, and a feedback adjustment module.

[0103] Specifically, the system includes: a status acquisition module for real-time acquisition of operating status data from each workstation in the SMT production line, connected to the controllers of each workstation via an industrial communication interface; a load modeling module electrically connected to the status acquisition module for constructing a global load status vector for the production line based on the operating status data; a scheduling decision module electrically connected to the load modeling module for obtaining a dynamic scheduling instruction set for the cache machine based on the global load status vector; an execution control module electrically connected to the scheduling decision module for sending the dynamic scheduling instruction set to the execution unit of the cache machine; a cache monitoring module electrically connected to the execution control module for monitoring the dwell time and location distribution of each board in the cache area and generating a cache resource occupancy status diagram; and a feedback adjustment module electrically connected to both the cache monitoring module and the load modeling module for adjusting the dynamic scheduling instruction set for the next scheduling cycle based on the matching degree between the cache resource occupancy status diagram and the global load status vector.

[0104] Specifically, the status acquisition module is configured to poll the controllers of each workstation at fixed time intervals and use the OPC UA protocol for data encapsulation and transmission.

[0105] Furthermore, the load modeling module is configured as a maintenance workstation index table, which stores the physical location, processing capacity parameters, and communication addresses of each workstation, used for structured mapping of the collected raw state data. The scheduling decision module has a built-in priority acquisition unit and a path planning unit. The priority acquisition unit obtains the cache priority based on the order urgency level identifier and the board dwell time. The path planning unit uses the A* algorithm to generate conflict-free handling paths. The execution control module is connected to the servo driver, pneumatic gripper, and position sensor of the cache machine via hardwiring. After receiving the dynamic scheduling instruction set, it generates corresponding motor control signals, gripper opening and closing instructions, and position verification requests. The cache monitoring module obtains the board position coordinates and order identifiers through a visual recognition camera and RFID reader installed inside the cache machine. The visual recognition camera collects images of the cache area at a preset frequency, and the RFID reader automatically reads the tag information when the board enters or leaves the cache area. The feedback adjustment module is configured to, at the end of each scheduling cycle, obtain the retention anomaly rate in the cache resource occupancy state graph and the load volatility rate in the global load state vector, and determine whether to enable the advanced scheduling strategy based on whether the product of the two exceeds a seventh preset threshold.

[0106] To enable those skilled in the art to fully understand and implement this invention, the specific implementation principles of this invention are further supplemented below with a specific application scenario.

[0107] In a flexible SMT production line that includes a high-speed pick and place machine (station W1), a 3D-AOI inspection device (station W2), and a back-end automated assembly machine (station W3), an intelligent buffer machine is set between W1 and W2 to smooth out the cycle time fluctuations between processes.

[0108] Step 1: Real-time Modeling of Global Production Status. The status acquisition module polls the controllers of W1, W2, and W3 every 500ms via the industrial communication interface to obtain status messages containing the current placement frequency, equipment fault codes, and the number of boards to be processed. After receiving the status messages, the load modeling module extracts the real-time output rate of W1 (60 boards / minute) and the rated processing capacity of W2 (45 boards / minute) from the workstation index table. The system executes the acquisition logic, dividing the number of tasks to be processed for W1 by its rated processing capacity to obtain the real-time load index of W1. Subsequently, the load modeling module concatenates the real-time load index, task queue length, and equipment availability flag of each workstation in a predetermined order to generate the global load status vector of the production line at the current moment.

[0109] Step 2: Generation and Priority Recalculation of Dynamic Scheduling Instructions. The scheduling decision module receives the global load status vector. When it determines that the real-time load index of W1 exceeds the first preset threshold (e.g., 30) and W2 is in a high-load state (real-time load index greater than 15), the scheduling decision module identifies a local congestion risk on the production line. At this time, if the system receives an order insertion signal from the MES system, the priority acquisition unit immediately retrieves the order urgency level identifier (e.g., level 5) corresponding to the insertion task and the current board's dwell time in the buffer area. The system performs weighted acquisition according to the formula P=0.7×E+0.3×(T / Tmax) to generate a new board cache priority. The path planning unit synchronously calls A. ∗ The algorithm, based on the physical layout structure inside the cache machine, avoids the already occupied logical cache units and generates a conflict-free transport path from the feed port to the specified cache coordinates (4,5), which is finally summarized into a dynamic scheduling instruction set.

[0110] Step 3: Physical Motion Control of the Execution Unit. After receiving the dynamic scheduling instruction set, the execution control module parses the instructions into electrical signals recognizable by the underlying mechanism. The module sends a specific frequency pulse sequence to the X-axis and Y-axis servo drivers, driving the robotic arm of the buffer unit to move to the feed port position. When the robotic arm reaches the target coordinates, the execution control module outputs a high-level signal to the solenoid valve of the pneumatic gripper, controlling the gripper to close and grasp the sheet metal. Subsequently, the robotic arm moves to coordinates (4,5) according to the path planning instructions and sends a verification request to the position sensor at that location. When the sensor returns a position signal, the execution control module controls the gripper to release, completing the temporary storage operation of the sheet metal.

[0111] Step 4: Monitoring Cache Resource Occupancy. During cache machine operation, the cache monitoring module uses a top-mounted visual recognition camera to acquire images of the cache area at a frequency of 1Hz. Image processing algorithms segment the acquired images, identifying the brightness and darkness features or contour information of each logical cache unit to determine the occupancy status of each unit. Simultaneously, an RFID reader reads the attached electronic tag when a board enters the cache area, recording the entry timestamp and order identifier. The system subtracts the entry timestamp from the current system time to obtain the real-time dwell time of each board. The cache monitoring module maps the above location, status, duration, and order information onto a 6×8 two-dimensional grid to generate a cache resource occupancy status map.

[0112] Step 5: Closed-loop feedback adjustment of the scheduling strategy. At the end of each scheduling cycle (e.g., 1 second), the feedback adjustment module retrieves the cache resource occupancy status graph for data statistics. When the number of boards with a dwell time exceeding 300 seconds accounts for more than 10% of the total occupied units, the feedback adjustment module determines that the current release efficiency is insufficient. Simultaneously, the module analyzes the fluctuation of the global load state vector over the past 5 cycles. If the standard deviation of the load index of downstream workstation W2 is found to be greater than 5, it determines that there is operational fluctuation downstream. The feedback adjustment module generates adjustment parameters based on the product of the dwell time anomaly rate and the load fluctuation rate, and modifies the weight coefficient of the scheduling decision module in the next scheduling cycle, increasing the release priority of boards with longer dwell times and replanning the handling path to avoid workstation interfaces with drastic fluctuations.

[0113] Step Six: Cyclic Execution and Adaptive Optimization. The system continuously repeats the above steps of state acquisition, modeling, decision-making, execution, monitoring, and feedback. Through closed-loop adjustment within each cycle, the buffer machine switches its operating strategy in real time according to the dynamic changes in the overall line load. When W1 is producing at high speed, the buffer depth is increased, and when W2 returns to idle, the release is accelerated. When an order arrives, the system ensures that urgent boards pass through the buffer area first through dynamic priority rearrangement, achieving real-time response to complex operating conditions on the SMT production line.

[0114] All content not described in detail in this specification is prior art known to those skilled in the art, and the communication protocols and hardware parameters of each module are not specifically limited; conventional equipment can be used. Electrical control components not mentioned in this technical solution are not shown in detail in the figures because they are prior art.

[0115] It is understood that the intelligent buffer dynamic scheduling method and system for SMT production lines in the above embodiments of the present invention have the same beneficial effects, and will not be described in detail here.

[0116] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program goods. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program goods embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0117] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program goods according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0118] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0119] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the protection scope of the claims of the present invention.

Claims

1. A dynamic scheduling method for intelligent buffer machines in an SMT production line, characterized in that, include: Real-time acquisition of operating status data for each station in the SMT production line. The operating status data includes the cycle output frequency of the upstream placement equipment, the operating status indicators of the downstream testing and assembly equipment, and the insertion signal of the order insertion task. A global load status vector for the production line is constructed based on the operational status data. The global load status vector includes the real-time load index, task queue length, and equipment availability flag for each workstation. The real-time load index is the ratio of the number of tasks to be processed to the rated processing capacity of the equipment. The dynamic scheduling instruction set of the cache machine is obtained based on the global load state vector. The dynamic scheduling instruction set includes board cache priority, release timing and transportation path planning. The dynamic scheduling instruction set is sent to the execution unit of the cache machine to control the receiving, temporary storage, release and transfer operations of the cache machine execution board; During the operation of the cache machine, the position coordinates, entry timestamps and order identifiers of each board are recorded to obtain the board's dwell time and generate a cache resource occupancy status diagram. Based on the matching degree between the cache resource occupancy state graph and the global load state vector, the dynamic scheduling instruction set for the next scheduling cycle is adjusted by obtaining the abnormal percentage of dwell time or load volatility.

2. The intelligent buffer dynamic scheduling method for SMT production lines as described in claim 1, characterized in that, When collecting real-time operational status data, it includes: The status messages output by the equipment controller are obtained at a predetermined sampling period based on the industrial communication interface modules deployed at each workstation, and the status messages output by the equipment controller are parsed. Based on the OPC UA protocol, the parsed status messages are encapsulated into standardized status data packets and transmitted to the central dispatch server via industrial Ethernet.

3. The intelligent buffer dynamic scheduling method for SMT production lines as described in claim 2, characterized in that, The industrial communication interface module polls the controllers of each workstation device with a sampling period of 500 ms. The extracted status messages include the device operating mode, task completion count, fault code, and number of pending tasks.

4. The intelligent buffer dynamic scheduling method for SMT production lines as described in claim 1, characterized in that, When constructing the global load state vector, the following are included: Assign a unique workstation identifier to each workstation, and arrange the real-time load index, task queue length and equipment availability flag of each workstation in order according to the workstation index table. The device availability flag is a binary value, where 1 indicates an operational status and 0 indicates a shutdown or maintenance status.

5. The intelligent buffer dynamic scheduling method for SMT production lines as described in claim 4, characterized in that, When obtaining the real-time load index for each workstation, the following should be included: Based on a comparison of the number of tasks to be processed at each workstation and the rated processing capacity of the equipment at that workstation, the real-time load index of each workstation is determined, wherein: The more tasks to be processed and the lower the processing capacity of the equipment per unit time, the higher the real-time load index of the corresponding workstation.

6. The intelligent buffer dynamic scheduling method for SMT production lines as described in claim 1, characterized in that, When obtaining the dynamic scheduling instruction set of the cache machine based on the global load state vector, it includes: Based on the relationship between the real-time load index of the upstream or downstream workstation and the first or second preset load index, the dynamic scheduling instruction set of the buffer machine is obtained; where: When the real-time load index of the upstream station is greater than the first preset load index and the equipment availability flag of the downstream station is 1, a board caching instruction is generated. When the real-time load index of the upstream station is less than or equal to the first preset load index, and the real-time load index of the downstream station is less than the second preset load index and its equipment availability flag is 1, a board release command is generated.

7. The intelligent buffer dynamic scheduling method for SMT production lines as described in claim 6, characterized in that, Also includes: When an insertion signal for a task is detected, the cache priority of each board in the cache is retrieved again, where: The cache priority is determined based on the urgency level of the order and the current dwell time of the board, and is comprehensively evaluated based on the relationship between the dwell time and the maximum allowable dwell time. Among them, when the urgency weight of an order is higher than the retention time weight, the board corresponding to the inserted order task will be processed with higher priority in the cache scheduling than other boards.

8. The intelligent buffer dynamic scheduling method for SMT production lines as described in claim 7, characterized in that, When generating a cache resource usage state graph, the following is included: The cache machine is divided into multiple logical cache units, and each logical cache unit has unique location coordinates. The location coordinates, entry timestamp, and order identifier of the board are obtained by using a visual recognition camera and an RFID reader in collaboration. The occupancy status, retention time, and order identifier of each logical cache unit are mapped to a two-dimensional grid array to form a cache resource occupancy status diagram; Among them, the visual recognition camera acquires images from the buffer area at a frequency of 1Hz; The RFID reader automatically reads the tag information when the board enters or leaves the buffer area.

9. The intelligent buffer dynamic scheduling method for SMT production lines as described in claim 8, characterized in that, When adjusting the dynamic scheduling instruction set for the next scheduling cycle, the following are included: If the proportion of boards whose retention time in the cache resource occupation status graph exceeds the third preset threshold is greater than the fourth preset threshold, then the release priority of such boards will be increased. If the number of workstations whose load index change amplitudes exceed the fifth preset threshold in the global load state vector in consecutive five sampling periods is greater than a sixth preset threshold, the A * The path optimization strategy of the algorithm re-plans the plate carrying path.

10. A dynamic scheduling system for intelligent buffer machines in an SMT production line, employing the dynamic scheduling method for intelligent buffer machines in an SMT production line as described in any one of claims 1-9, characterized in that, include: The status acquisition module is used to collect real-time operating status data of each station in the SMT production line. The status acquisition module is connected to the equipment controller of each station through an industrial communication interface. The load modeling module is electrically connected to the status acquisition module and is used to construct a global load status vector of the production line based on the operating status data. The scheduling decision module, which is electrically connected to the load modeling module, is used to obtain the dynamic scheduling instruction set of the cache machine based on the global load state vector. An execution control module, electrically connected to the scheduling decision module, is used to send dynamic scheduling instruction sets to the execution unit of the cache machine; The cache monitoring module is electrically connected to the execution control module and is used to monitor the dwell time and location distribution of each board in the cache area and generate a cache resource occupancy status diagram. The feedback adjustment module is electrically connected to the cache monitoring module and the load modeling module, respectively, and is used to adjust the dynamic scheduling instruction set of the next scheduling cycle based on the matching degree between the cache resource occupancy status diagram and the global load status vector.

Images