Multi-station spoke steel plate feeding system and control method thereof

Abstract

The invention provides a multi-station spoke steel plate feeding system and a control method thereof, and relates to the field of security doors. The system is composed of a core feeding module, a multi-sensor sensing module, a distributed decision control module, a material flow management module and a quality closed-loop control module which perform data interaction through an industrial Ethernet. The system solves the problems that an operator needs to manually complete steel plate separation and carrying, the labor intensity is large, fatigue is prone to occurring, and the working efficiency is high. The production takt is slow and the labor cost is high. Meanwhile, workers need to be in close contact with heavy equipment and molds, and serious mechanical injury risks exist. In addition, the problems that repeated positioning accuracy of manual placement is poor, consistency is difficult to guarantee, product scrapping, die damage and even equipment shutdown are possibly caused by tiny deviation, and product quality and production line efficiency are seriously restricted are solved.

Description

Technical Field

[0001] This invention relates to the field of automobile wheel manufacturing technology, and in particular to a multi-station wheel spoke steel plate loading system and its control method. Background Technology

[0002] Wheel spokes are devices that protect the rims and spokes of a vehicle's wheels. They are characterized by a pair of circular plates, the diameter of which is close to the diameter of the rim. The center of each plate has a hole larger than the wheel's axle, and there are openings near the edges. The edges of the plates have annular wheel plates whose curved surfaces fit snugly against the rim. Based on the spoke structure, wheels are divided into spoked and wheel-spoked types. Most mainstream passenger cars use spoked wheel structures.

[0003] In the automotive wheel manufacturing industry, wheel spokes, as key load-bearing components, are typically made from high-strength steel plates through multiple processes such as stamping, spinning, or cutting. The first step in this manufacturing process is to remove the original round steel plate blank from the stack and precisely place it onto the mold of a CNC machine tool or press for processing.

[0004] However, regarding existing multi-station wheel spoke steel plate loading systems and their control methods, operators need to manually separate individual steel plates from the stack. This involves extremely high labor intensity from handling heavy steel plates, leading to worker fatigue, slow production cycle, and the need for multiple operators per production line, resulting in continuously rising labor costs. Furthermore, operators are in close contact with heavy equipment and molds, posing a high risk of mechanical injury. Manual placement also results in low repeatability and inconsistent positioning. Even minor placement deviations can lead to product scrap, mold damage, or even equipment downtime, severely impacting product quality and overall production line efficiency. Summary of the Invention

[0005] In view of this, the present invention provides a multi-station wheel spoke steel plate loading system, which is composed of functional modules that interact with each other via industrial Ethernet: a core loading module, a multi-sensor perception module, a distributed decision control module, a material flow management module, and a quality closed-loop control module.

[0006] The core loading module includes: a loading station equipped with a separation and lifting device, a six-degree-of-freedom articulated robot mounted on a linear guide rail, and an adaptive vacuum suction cup clamp mounted at the end of the robot.

[0007] The multi-sensor perception module includes: at least one 2D vision camera fixed to the robot base, a laser displacement sensor integrated on the suction cup gripper, and a six-dimensional force / torque sensor for detecting the contact state between the gripper and the steel plate.

[0008] The distributed decision control module includes: a programmable logic controller (PLC) as the master station and a robot controller, vision processor and machine tool CNC system as slave stations. The PLC is equipped with a real-time Ethernet communication module.

[0009] The material flow management module includes: an automated storage and retrieval system (AS / RS), an autonomous guided vehicle (AGV) for stacking and transferring materials, and a warehouse management server running Manufacturing Execution System (MES) client software;

[0010] The quality closed-loop control module includes: an in-machine measurement probe installed inside the machining station and an industrial control computer running an adaptive PID control algorithm.

[0011] Furthermore, the 2D vision camera in the multi-sensor perception module calculates the pose deviation of the steel plate in the robot's base coordinate system using the following hand-eye calibration model. ,in The fixed transformation matrix of the camera relative to the robot's base coordinate system is obtained through hand-eye calibration; The pose of the marker points relative to the camera, obtained through image recognition; The transformation matrix of the marker point relative to the tool coordinate system of the robot flange; The real-time pose of the flange is fed back by the robot controller.

[0012] Furthermore, the distributed decision control module employs a state-based finite automaton model for task scheduling, and its dynamic workstation selection strategy... Cost function Defined as:

[0013]

[0014] in, For workstation index, Indicates workstation Is it ready? (1 if yes, 0 otherwise) Move the robot to the workstation The estimated time, This is the length of the queue of tasks waiting to be processed at this workstation; , , These are weighting factors that can be adjusted online, and The system selects to make Minimum ready station As the target workstation.

[0015] Furthermore, the adaptive PID control algorithm of the quality closed-loop control module adjusts the machine tool offset. The adjustment follows the following incremental formula:

[0016]

[0017] in, For the first dimensional error of the second measurement , , For controller parameters; this module further includes a parameter self-tuning unit, which adjusts the parameters based on the error rate of change. Based on historical adjustment amounts, online fine-tuning is performed using fuzzy rules. , , To improve system response.

[0018] Furthermore, the warehouse management server of the material flow management module integrates an inventory optimization model, which sets a safety stock level. and reorder points The calculation model is as follows:

[0019]

[0020]

[0021] in, The security factor corresponding to a specific service level. The standard deviation of daily demand. Lead time for replenishment (days). This represents the average daily demand; the server communicates with the PLC via the OPCUA protocol, and when the PLC reports a line-side inventory level below [a certain threshold]... At that time, the server automatically sends library instructions to the AGV system.

[0022] Furthermore, the system also includes a predictive maintenance and digital twin module, which includes:

[0023] A real-time data acquisition and monitoring control (SCADA) system is used to collect equipment status data such as motor current, joint temperature, and vibration spectrum.

[0024] A discrete event simulation model that runs synchronously with the physical system, serving as a digital twin;

[0025] A fault prediction model based on a Long Short-Term Memory (LSTM) network, which uses the acquired time series data... As input, the output is the device's output within a specific time window in the future. Health index within And the confidence interval for remaining useful life (RUL).

[0026] A control method for a multi-station wheel spoke steel plate loading system, characterized in that the method, under the coordination of a distributed decision control module, executes the following steps:

[0027] Step 1: Perception-Decision-Execution Loop Step: The multi-sensor perception module continuously acquires environmental data, and the decision control module executes the dynamic workstation selection strategy described in claim 3 based on this data, and drives the core material feeding module to complete the material feeding task;

[0028] Step 2: Material Flow Coordination Step: According to the inventory optimization model described in claim 5, the material flow management module autonomously schedules AGVs to complete the replenishment of stacked materials from the automated warehouse to the loading station when the reorder point is triggered;

[0029] Step 3: Quality Closed-Loop Control Step: The quality closed-loop control module, based on the adaptive PID control algorithm described in claim 4, uses on-machine measurement data to perform real-time compensation for the processing.

[0030] Furthermore, step one includes a compliant placement sub-step based on force / position hybrid control: when the robot places the steel plate onto the mold, the system switches from pure position control to force / position hybrid control; force control based on feedback from a six-dimensional force sensor is used in the Z-axis direction, with the target contact force... To maintain a constant small force, position control is still used on the XY plane and the rotation axis to ensure precise positioning of the steel plate.

[0031] Furthermore, the control method also includes a dynamic path replanning step based on digital twins: when the digital twin model predicts a path conflict between the robot and an AGV that may enter the work area in the future, the decision control module will... In time, a collision-free path is replanned for the robot based on the following cost function.

[0032] Furthermore, the control method also includes a system energy efficiency optimization step: the decision control module monitors the idle status of each processing station, and when the idle time of any station exceeds a threshold... When a new machining task is assigned to the station, the module sends a command to the CNC system at that station to put it into a low-power "sleep mode"; when a new machining task is assigned to the station, a "wake-up" command is sent, with a wake-up lead time of [missing information]. It is incorporated into the production cycle calculation in advance.

[0033] Beneficial effects:

[0034] 1. Through the core architecture of "mobile robot + linear guide + dynamic scheduling algorithm", one robot can efficiently serve multiple processing stations. The scheduling algorithm ensures the maximization of the system's total capacity, avoids station waiting, and the equipment utilization rate (up to 92% or more) is far higher than the solution of equipping each station with a fixed robot.

[0035] 2. Through machine vision guidance and force / position hybrid control, the system no longer relies on fixed mechanical positioning. It can automatically compensate for the positional deviation of the steel plate, the minor wear of the mold, and the absolute positioning error of the robot, achieving a repeatability accuracy within ±0.05mm, fundamentally eliminating product defects caused by inaccurate material loading.

[0036] 3. This application acquires data through a multi-sensor perception module, performs intelligent scheduling and path planning by a distributed decision control module, executes precisely by a core material feeding module, and achieves continuous optimization through a quality closed-loop control module and a predictive maintenance module.

[0037] 4. The system can detect workpiece dimensions in real time during processing and automatically adjust machine tool parameters to compensate for tool wear. This transforms the traditional "processing-inspection-scrap / rework" model into a preventative quality control model of "processing-monitoring-adaptive adjustment," thereby stabilizing the process pass rate at nearly 100% and significantly reducing quality loss costs. Attached Figure Description

[0038] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings of the embodiments will be briefly described below.

[0039] The accompanying drawings described below are only related to some embodiments of the invention and are not intended to limit the invention.

[0040] In the attached diagram:

[0041] Figure 1 This is a flowchart of a multi-station wheel spoke steel plate loading system and its control method according to an embodiment of the present invention.

[0042] Figure 2 This is a schematic diagram of the material flow management module of the multi-station wheel spoke steel plate loading system and its control method according to an embodiment of the present invention.

[0043] Figure 3 This is a schematic diagram of the multi-sensor module of the multi-station wheel spoke steel plate loading system and its control method according to an embodiment of the present invention.

[0044] Figure 4 This is a schematic diagram of the predictive maintenance and digital twin module of the multi-station wheel spoke steel plate loading system and its control method according to an embodiment of the present invention.

[0045] Figure 5 This is a schematic diagram of the distributed decision control module of the multi-station wheel spoke steel plate loading system and its control method according to an embodiment of the present invention.

[0046] Figure 6This is a schematic diagram of the core feeding module of the multi-station wheel spoke steel plate feeding system and its control method according to an embodiment of the present invention.

[0047] Figure 7 This is a schematic diagram of the self-use PID control algorithm of the multi-station spoke steel plate loading system and its control method according to an embodiment of the present invention. Detailed Implementation

[0048] The embodiments of the present invention will be described in further detail below with reference to the accompanying drawings and examples.

[0049] Example: Please refer to Figures 1 to 7 As shown:

[0050] This invention provides a multi-station wheel spoke steel plate loading system, which is composed of functional modules that interact via industrial Ethernet: a core loading module, a multi-sensor sensing module, a distributed decision control module, a material flow management module, and a quality closed-loop control module; its key feature is:

[0051] The core loading module includes: a loading station with a separation and lifting device, a six-degree-of-freedom articulated robot mounted on a linear guide rail, and an adaptive vacuum suction cup clamp mounted at the end of the robot;

[0052] The multi-sensor perception module includes: at least one 2D vision camera fixed to the robot base, a laser displacement sensor integrated on the suction cup gripper, and a six-dimensional force / torque sensor for detecting the contact state between the gripper and the steel plate.

[0053] The distributed decision control module includes: a programmable logic controller (PLC) as the master station and a robot controller, vision processor and machine tool CNC system as slave stations. The PLC is equipped with a real-time Ethernet communication module.

[0054] The material flow management module includes: an automated storage and retrieval system (AS / RS), an autonomous guided vehicle (AGV) for stacking and transferring materials, and a warehouse management server that runs the Manufacturing Execution System (MES) client software;

[0055] The quality closed-loop control module includes: an in-machine measurement probe installed inside the machining station and an industrial computer running an adaptive PID control algorithm.

[0056] Among them, the 2D vision camera in the multi-sensor perception module calculates the pose deviation of the steel plate in the robot base coordinate system using the following hand-eye calibration model. ,in The fixed transformation matrix of the camera relative to the robot's base coordinate system is obtained through hand-eye calibration; The pose of the marker points relative to the camera, obtained through image recognition; The transformation matrix of the marker point relative to the tool coordinate system of the robot flange; The real-time pose of the flange is fed back by the robot controller.

[0057] The distributed decision control module employs a state-based finite automaton model for task scheduling, and its dynamic workstation selection strategy... Cost function Defined as:

[0058]

[0059] in, For workstation index, Indicates workstation Is it ready? (1 if yes, 0 otherwise) Move the robot to the workstation The estimated time, This is the length of the queue of tasks waiting to be processed at this workstation; , , These are weighting factors that can be adjusted online, and The system selects to make Minimum ready station As the target workstation.

[0060] Among them, the adaptive PID control algorithm of the quality closed-loop control module controls the machine tool offset. The adjustment follows the following incremental formula:

[0061]

[0062] in, For the first dimensional error of the second measurement , , For controller parameters; this module further includes a parameter self-tuning unit, which adjusts the parameters based on the error rate of change. Based on historical adjustment amounts, online fine-tuning is performed using fuzzy rules. , , To improve system response.

[0063] The warehouse management server in the material flow management module integrates an inventory optimization model, which sets a safety stock level. and reorder points The calculation model is as follows:

[0064]

[0065]

[0066] in, The security factor corresponding to a specific service level. The standard deviation of daily demand. Lead time for replenishment (days). This represents the average daily demand; the server communicates with the PLC via the OPCUA protocol, and when the PLC reports a line-side inventory level below [a certain threshold]... At that time, the server automatically sends library instructions to the AGV system.

[0067] The system also includes a predictive maintenance and digital twin module, which includes:

[0068] A real-time data acquisition and monitoring control (SCADA) system is used to collect equipment status data such as motor current, joint temperature, and vibration spectrum.

[0069] A discrete event simulation model that runs synchronously with the physical system, serving as a digital twin;

[0070] A fault prediction model based on a Long Short-Term Memory (LSTM) network, which uses the acquired time series data... As input, the output is the device's output within a specific time window in the future. Health index within And the confidence interval for remaining useful life (RUL).

[0071] A control method for a multi-station wheel spoke steel plate loading system, characterized in that the method, under the coordination of a distributed decision control module, executes the following steps:

[0072] Step 1: Perception-Decision-Execution Loop Step: The multi-sensor perception module continuously acquires environmental data, and the decision control module executes the dynamic workstation selection strategy of claim 3 based on this data, and drives the core material feeding module to complete the material feeding task;

[0073] Step 2: Material Flow Coordination Step: According to the inventory optimization model in claim 5, the material flow management module autonomously schedules AGVs to complete the replenishment of stacked materials from the automated warehouse to the loading station when the reorder point is triggered.

[0074] Step 3: Quality Closed-Loop Control Step: The quality closed-loop control module is based on the adaptive PID control algorithm of claim 4, and uses on-machine measurement data to perform real-time compensation for the processing.

[0075] Step one includes a compliant placement sub-step based on force / position hybrid control: when the robot places the steel plate onto the mold, the system switches from pure position control to force / position hybrid control; force control based on feedback from a six-dimensional force sensor is used in the Z-axis direction, with the target contact force... To maintain a constant small force, position control is still used on the XY plane and the rotation axis to ensure precise positioning of the steel plate.

[0076] The control method also includes a dynamic path replanning step based on digital twins: when the digital twin model predicts a path conflict between the robot and an AGV that may enter the work area in the future, the decision control module will... In time, a collision-free path is replanned for the robot based on the following cost function.

[0077] The control method also includes a system energy efficiency optimization step: the decision control module monitors the idle status of each processing station, and when the idle time of any station exceeds a threshold... When a new machining task is assigned to the station, the module sends a command to the CNC system at that station to put it into a low-power "sleep mode"; when a new machining task is assigned to the station, a "wake-up" command is sent, with a wake-up lead time of [missing information]. It is incorporated into the production cycle calculation in advance.

[0078] Comparative experimental data of multi-station spoke steel plate loading system

[0079] I. Experimental Objective

[0080] The system of this invention was verified to demonstrate its comprehensive performance in terms of production efficiency, feeding accuracy, product quality control, flexibility, and equipment operation and maintenance, and compared with four production modes with different levels of automation.

[0081] II. Experimental Conditions

[0082] Test product: 17-inch round steel plate for automotive wheel spokes, 6mm thick, 430mm in diameter.

[0083] System configuration: 1 loading station, 3 CNC machining centers, 1 six-axis robot, linear guide, 2D vision system, and in-machine measurement probe.

[0084] Comparison objects:

[0085] Control Group A: Manual material loading (one worker at each of the three workstations). [Basic Comparison]

[0086] Control Group B: Fixed robot feeding (one fixed robot per workstation). [Automation Comparison]

[0087] Control Group C: A mobile robot with a simple guide rail, but without visual guidance and dynamic scheduling (it only feeds materials in a fixed order). [Function Stripping Comparison]

[0088] Control group D: Same hardware as this system, but with quality closed-loop control and predictive maintenance modules disabled. [Intelligent Stripping Comparison]

[0089] Test duration: 8 hours of continuous production. To test flexibility, a product changeover (replacing with spoke steel plates of different sizes) was performed after 4 hours.

[0090] III. Experimental Data and Results

[0091]

[0092] in conclusion:

[0093] The system of this invention achieves extremely high equipment utilization and stable operation with almost zero downtime while maintaining maximum output.

[0094] Control group C demonstrates that without vision and intelligent scheduling, the efficiency advantage and stability of mobile robots are greatly reduced.

[0095] The system of this invention exhibits excellent flexibility, with minimal changeover time and almost no impact on production cycle.

[0096] 2. Feeding positioning accuracy, quality, and stability

[0097]

[0098] Conclusion: Visual guidance and force-controlled placement are key to ensuring high success rates and high precision. Control group C, lacking visual guidance, could not compensate for steel plate positional deviations, resulting in poor precision and frequent jamming.

[0099] 3. Effectiveness of closed-loop quality control (long-term stability test)

[0100] Test background: 500 parts were continuously machined to simulate normal tool wear. The center bore diameter of the wheel spokes was monitored (target value Φ100.00mm±0.05mm).

[0101] Key results:

[0102]

[0103] Conclusion: The quality closed-loop control module is the core of this system's achievement of 100% pass rate and excellent process stability. Without it, the system is no different from traditional automation and cannot cope with process fluctuations such as tool wear.

[0104] Overall conclusion:

[0105] 1. The system of this invention achieves the best balance in efficiency, accuracy, quality, flexibility and operation and maintenance costs. It is not outstanding in a single indicator, but rather comprehensively leading in overall performance.

[0106] 2. The experiment clearly distinguished between "automation" and "intelligence". Control group C demonstrated that mobile robots lacking perception and decision-making capabilities cannot handle complex working conditions; while control group D highlighted the decisive role of the two intelligent modules, quality closed-loop and predictive maintenance, in ensuring long-term stable production and reducing total costs.

[0107] Final conclusion: This invention is not just a feeding system, but also an intelligent manufacturing unit for wheel spokes that integrates sensing, decision-making, execution, and optimization capabilities. Its technological advancement and economic rationality are fully supported by experimental data.

[0108] The specific usage and function of this embodiment: First stage: Production preparation and intelligent material supply

[0109] System initialization and self-test: Upon system power-on, all module controllers start, and moving parts such as the loading robot and AGV perform a zero-return operation. The predictive maintenance module reports the initial health status of the equipment. The digital twin model and the physical system are synchronized.

[0110] Inventory monitoring and automatic replenishment:

[0111] The warehouse management server in the material flow management module continuously monitors the line-side inventory at the material loading station.

[0112] According to the economic order batch model described in claim 5, when the inventory level is lower than the reorder point (ROP), the server automatically issues a warehouse instruction to the AGV system.

[0113] The AGV retrieves full stacks of materials of the corresponding specifications from the automated material storage system, transports them to the loading station, and returns the empty pallets, thus achieving unmanned material supply.

[0114] Steel plate positioning: The separation and lifting device (such as a magnetic sheet separator in conjunction with a lifting mechanism) at the loading station is activated to reliably separate and lift the top single steel plate of the stack to a fixed and precise "position to be picked up".

[0115] Phase Two: Precise Feeding and Dynamic Scheduling Guided by Sensors

[0116] Visual recognition and pose compensation:

[0117] The loading robot moves to the identification point in front of the loading station.

[0118] The 2D vision camera in the multi-sensor perception module takes a picture of the steel plate at the "location to be photographed".

[0119] Based on the hand-eye calibration model described in claim 2, the vision system calculates the deviation value ΔP between the theoretical position and the actual position of the steel plate relative to the robot gripper.

[0120] Dynamic workstation selection:

[0121] Meanwhile, the dynamic scheduling algorithm in the decision control module calculates the priority of each processing station in real time based on the cost function C(i) described in claim 3.

[0122] The algorithm takes into account factors such as the idle status of the workstation, the robot's movement distance, and the length of the task queue to select an optimal "target processing workstation".

[0123] Robot gripping and handling: The loading robot receives the pose compensation value ΔP from the vision system and the target station instruction from the scheduling algorithm, drives the end suction cup gripper to accurately grip the steel plate, and then moves it along the linear guide to the target station.

[0124] Phase 3: Compliance Placement and Closed-Loop Quality Control

[0125] Force / position hybrid control placement: As described in claim 8, the robot switches to force / position hybrid control mode when placing the steel plate. A constant, small contact force is maintained via force sensor feedback along the Z-axis (vertical direction) to prevent impact; high-precision position control is still used in the XY plane and rotational axes to ensure the steel plate is perfectly positioned on the mold.

[0126] Processing Start-up and Cycle: After the robot is placed in position, it sends a signal to the system controller, which then instructs the processing station to start the processing program. Simultaneously, the robot immediately returns to the loading station to begin the next loading cycle.

[0127] In-machine detection and adaptive compensation:

[0128] After the workpiece is machined, the quality closed-loop control module is activated. A high-precision online measuring instrument (such as a probe) installed inside the machining center quickly detects the critical dimensions of the workpiece.

[0129] The measurement data is transmitted to the system controller in real time. As described in claim 4, the adaptive compensation unit compares the measured value with the target value. If a trend deviation in the dimension is found (such as due to tool wear), a tool offset compensation amount Δu(k) is automatically calculated according to the incremental PID algorithm and sent to the CNC system of the machining center.

[0130] This allows for real-time, online, and adaptive adjustment of processing parameters, ensuring long-term stability of processing quality.

[0131] Phase 4: Finished Product Processing and System-Level Operation and Maintenance

[0132] Intelligent feeding and sorting:

[0133] The completed workpieces are picked up by the unloading robot and transferred to the sorting area.

[0134] The finished product vision sorting system identifies the workpieces (such as by reading codes or visual inspection) and determines them as "qualified products", "products to be repaired" or "scrap products" according to preset rules.

[0135] Based on the judgment result, the unloading robot places the workpiece onto the corresponding diversion conveyor line to achieve lean logistics.

[0136] Predictive maintenance and system optimization:

[0137] Throughout the process, predictive maintenance and the digital twin module operate continuously. The SCADA system collects operational data (current, temperature, vibration, etc.) from all critical equipment.

[0138] As described in claim 6, the LSTM network model analyzes these time series data to predict the health index and remaining useful life of the equipment, generates early warnings before failures occur, and achieves predictive maintenance.

[0139] Digital twin models can be used for production cycle simulation and virtual program debugging of new tasks, ensuring the safety and efficiency of the physical system during execution.

Claims

1. A multi-station wheel spoke steel plate loading system, comprising: a core loading module, a multi-sensor sensing module, a distributed decision control module, a material flow management module, and a quality closed-loop control module, which interact with each other via industrial Ethernet; characterized in that: The core loading module includes: a loading station equipped with a separation and lifting device, a six-degree-of-freedom articulated robot mounted on a linear guide rail, and an adaptive vacuum suction cup clamp mounted at the end of the robot. The multi-sensor perception module includes: at least one 2D vision camera fixed to the robot base, a laser displacement sensor integrated on the suction cup gripper, and a six-dimensional force / torque sensor for detecting the contact state between the gripper and the steel plate. The distributed decision control module includes: a programmable logic controller (PLC) as the master station and a robot controller, vision processor and machine tool CNC system as slave stations. The PLC is equipped with a real-time Ethernet communication module. The material flow management module includes: an automated storage and retrieval system (AS / RS), an autonomous guided vehicle (AGV) for stacking and transferring materials, and a warehouse management server running Manufacturing Execution System (MES) client software; The quality closed-loop control module includes: an in-machine measurement probe installed inside the machining station and an industrial control computer running an adaptive PID control algorithm.

2. The multi-station wheel spoke steel plate loading system according to claim 1, characterized in that, The 2D vision camera in the multi-sensor perception module calculates the pose deviation of the steel plate in the robot base coordinate system using the following hand-eye calibration model. ,in The fixed transformation matrix of the camera relative to the robot's base coordinate system is obtained through hand-eye calibration; The pose of the marker points relative to the camera, obtained through image recognition; The transformation matrix of the marker point relative to the tool coordinate system of the robot flange; The real-time pose of the flange is fed back by the robot controller.

3. The multi-station wheel spoke steel plate loading system according to claim 1, characterized in that, The distributed decision control module uses a state-based finite automaton model for task scheduling, and its dynamic workstation selection strategy... Cost function Defined as: in, For workstation index, Indicates workstation Is it ready? (1 if yes, 0 otherwise) Move the robot to the workstation The estimated time, This is the length of the queue of tasks waiting to be processed at this workstation; , , These are weighting factors that can be adjusted online, and The system selects to make Minimum ready station As the target workstation.

4. The multi-station wheel spoke steel plate loading system according to claim 1, characterized in that, The adaptive PID control algorithm of the quality closed-loop control module adjusts the machine tool offset. The adjustment follows the following incremental formula: in, For the first dimensional error of the second measurement , , For controller parameters; this module further includes a parameter self-tuning unit, which adjusts the parameters based on the error rate of change. Based on historical adjustment amounts, online fine-tuning is performed using fuzzy rules. , , To improve system response.

5. The multi-station wheel spoke steel plate loading system according to claim 1, characterized in that, The warehouse management server of the material flow management module integrates an inventory optimization model with a safety stock level. and reorder points The calculation model is as follows: in, The security factor corresponding to a specific service level. The standard deviation of daily demand. Lead time for replenishment (days). This represents the average daily demand; the server communicates with the PLC via the OPCUA protocol, and when the PLC reports a line-side inventory level below [a certain threshold]... At that time, the server automatically sends library instructions to the AGV system.

6. The multi-station wheel spoke steel plate loading system according to claim 1, characterized in that, The system also includes a predictive maintenance and digital twin module, which comprises: A real-time data acquisition and monitoring control (SCADA) system is used to collect equipment status data such as motor current, joint temperature, and vibration spectrum. A discrete event simulation model that runs synchronously with the physical system, serving as a digital twin; A fault prediction model based on a Long Short-Term Memory (LSTM) network, which uses the acquired time series data... As input, the output is the device's output within a specific time window in the future. Health index within And the confidence interval for remaining useful life (RUL).

7. A control method for a multi-station wheel spoke steel plate loading system, characterized in that, Under the coordination of the distributed decision control module, the method performs the following steps: Step 1: Perception-Decision-Execution Loop Step: The multi-sensor perception module continuously acquires environmental data, and the decision control module executes the dynamic workstation selection strategy described in claim 3 based on this data, and drives the core material feeding module to complete the material feeding task; Step 2: Material Flow Coordination Step: According to the inventory optimization model described in claim 5, the material flow management module autonomously schedules AGVs to complete the replenishment of stacked materials from the automated warehouse to the loading station when the reorder point is triggered; Step 3: Quality Closed-Loop Control Step: The quality closed-loop control module, based on the adaptive PID control algorithm described in claim 4, uses on-machine measurement data to perform real-time compensation for the processing.

8. The control method for a multi-station wheel spoke steel plate loading system according to claim 7, characterized in that, Step one includes a compliant placement sub-step based on force / position hybrid control: when the robot places the steel plate onto the mold, the system switches from pure position control to force / position hybrid control; force control based on feedback from a six-dimensional force sensor is used in the Z-axis direction, with the target contact force... To maintain a constant small force, position control is still used on the XY plane and the rotation axis to ensure precise positioning of the steel plate.

9. The control method for a multi-station wheel spoke steel plate loading system according to claim 7, characterized in that, The control method also includes a dynamic path replanning step based on digital twins: when the digital twin model predicts a path conflict between the robot and an AGV that may enter the work area in the future, the decision control module will... In time, a collision-free path is replanned for the robot based on the following cost function: in, , , These are the weighting coefficients. This is a conflict risk function based on the minimum distance between two moving bodies.

10. The control method for a multi-station wheel spoke steel plate loading system according to claim 7, characterized in that, The control method further includes a system energy efficiency optimization step: the decision control module monitors the idle status of each processing station, and when the idle time of any station exceeds a threshold... When a new machining task is assigned to the station, the module sends a command to the CNC system at that station to put it into a low-power "sleep mode"; when a new machining task is assigned to the station, a "wake-up" command is sent, with a wake-up lead time of [missing information]. It is incorporated into the production cycle calculation in advance.

Images