Design and Dynamic Scheduling Method and System for Human-Machine Collaborative Dismantling Line of Waste Power Batteries

By assessing the condition of retired power batteries and using a multi-dimensional index system, combined with a dynamic scheduling method for human-machine collaborative dismantling lines, the problem of error reporting and machine jamming in automated battery dismantling lines when facing non-standard conditions has been solved, thus improving the efficiency and safety of the entire line.

CN123118064APending Publication Date: 2026-11-17CHANGAN UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611279260.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-08-21
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing automated battery dismantling lines are prone to errors and jams when faced with non-standard conditions of retired power batteries, and lack dynamic scheduling capabilities for human-machine collaboration, resulting in low overall line efficiency.

Method used

By collecting basic information about the batteries to be disassembled, assessing their structural integrity and classifying their states, establishing a multi-dimensional evaluation index system, using triangular fuzzy numbers to define the execution time, constructing a disassembly line balance optimization model, using an improved discrete optimization algorithm to generate an initial disassembly operation plan, and triggering a two-way dynamic task migration strategy in real time to achieve human-machine collaborative disassembly.

Benefits of technology

It improved the operating efficiency of the battery dismantling line, reduced the interference of non-standard operating conditions on the overall line cycle time, prevented stagnation caused by local anomalies, isolated the risks of high voltage and thermal runaway, and achieved self-optimization of the production line.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123118064A_ABST
    Figure CN123118064A_ABST
Patent Text Reader

Abstract

This invention relates to the field of battery recycling and processing, and discloses a design and dynamic scheduling method and system for a human-machine collaborative dismantling line for waste power batteries. The method includes: first, collecting basic information about the batteries to be dismantled to assess their structural integrity and classify their initial state; then, decomposing the dismantling operation into task units, calculating workstation adaptation values ​​using a multi-dimensional evaluation index system to allocate execution resources; subsequently, establishing an AND-OR relationship model expressing process priority, defining execution time using triangular fuzzy numbers, and employing an improved discrete optimization algorithm to solve the dismantling line balance optimization model to generate a work plan for execution; finally, during execution, real-time monitoring of workstation and personnel status, triggering a two-way dynamic task migration strategy when anomalies occur, and recording data to correct the task adaptation values ​​for similar batteries. This invention can effectively address the non-standard characteristics of waste batteries, reduce the interference of local anomalies on the overall cycle time, and improve the fault tolerance and operating efficiency of the production line.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of battery recycling and processing technology, specifically to the design and dynamic scheduling method and system for a human-machine collaborative dismantling line for waste power batteries. Background Technology

[0002] With the rapid development of the new energy vehicle industry, the number of retired power batteries is increasing year by year, making the recycling and dismantling of used power batteries a crucial link in resource recycling. Traditional battery dismantling mainly relies on manual labor, which is not only inefficient but also exposes workers to high safety risks such as electric shock and high-voltage fires. To improve this situation, the industry has gradually introduced automated dismantling equipment to build assembly line operation models, attempting to replace manual labor with mechanical equipment to perform high-risk dismantling actions.

[0003] Currently, conventional automated dismantling lines typically operate in reverse order of new battery assembly, with pre-set fixed process paths and program instructions. However, retired power batteries exhibit significant non-standard physical characteristics due to differences in service life, usage environment, and collision accidents. Common phenomena during dismantling include battery casing deformation, aging and adhesion of structural adhesives, and severe corrosion and stripping of fastening bolts. Because existing automated equipment lacks sufficient flexibility in handling these non-standard conditions, it is prone to visual recognition failures, end effector positioning deviations, or torque over-limit errors, leading to partial shutdowns at certain workstations.

[0004] Existing production line scheduling methods are mostly designed for standard product manufacturing processes and cannot effectively cope with the strong randomness and fluctuations in working hours during battery dismantling. When a workstation experiences a backlog of tasks due to equipment malfunction or human fatigue, existing systems struggle to flexibly transfer the obstructed tasks to other workstations or execution units without interrupting the main production line. This rigid scheduling and production planning not only amplifies the negative impact of non-standard anomalies on the efficiency of the entire dismantling line but also fails to fully leverage the high efficiency of industrial robots and the high flexibility of construction workers, resulting in the actual production capacity of the entire line being far lower than the theoretical design value. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a design and dynamic scheduling method and system for a human-machine collaborative dismantling line for waste power batteries. This solves the problems of existing battery dismantling lines being prone to equipment errors and jamming when faced with non-standard retired batteries, and the lack of dynamic scheduling capabilities for human-machine tasks under abnormal operating conditions, which leads to local congestion and reduces the overall line operating efficiency.

[0006] To achieve the above objectives, the present invention provides the following technical solution: The design and dynamic scheduling method of a human-machine collaborative dismantling line for used power batteries includes the following steps: Basic information about the battery to be disassembled is collected, and the structural integrity of the battery to be disassembled is assessed based on the basic information. The initial state of the battery to be disassembled is divided into healthy state, sub-healthy state and high-risk state. The disassembly operation is broken down into multiple independent task units, a multi-dimensional evaluation index system is established, and the adaptation values ​​of the task units to robot workstations, manual workstations and human-machine collaborative workstations are calculated based on the initial state and the multi-dimensional evaluation index system. The task units are then pre-allocated to the corresponding execution resources. Based on the structural characteristics of the power battery pack, an AND-OR relationship model is established to express the dismantling priority relationship under different process requirements. Triangular fuzzy numbers are used to define the execution time of the same task unit under different health states. A dismantling line balance optimization model is constructed, and an improved discrete optimization algorithm is used to solve the execution position and operation sequence of the task unit to generate an initial dismantling operation plan. According to the initial disassembly operation plan, the hardware layout is organized in a fishbone-shaped master-slave parallel architecture to control the movement of the batteries to be disassembled along the main conveyor line to the corresponding area and execute the operation. During the disassembly process, the operating status is acquired in real time, triggering a two-way dynamic task migration strategy, which enables the task unit to move between the robot station and the manual station, and records data and corrects the task adaptation values ​​of the same type of battery.

[0007] Preferably, the step of assessing the structural integrity of the battery to be disassembled based on basic information and classifying the initial state of the battery to be disassembled into a healthy state, a sub-healthy state, and a high-risk state specifically includes: Extract the physical and electrical performance characteristics from the basic information and construct a comprehensive health assessment function; A weighted calculation method is used to integrate electrical performance health, physical damage, and insulation safety to output a health index that characterizes the overall condition of used power batteries. Based on the calculated health index, the initial state of the batteries to be disassembled is divided into three levels: healthy state, sub-healthy state, and high-risk state.

[0008] Preferably, the initial state and the multi-dimensional evaluation index system calculate the adaptation values ​​of the task unit for the robot station, the manual station, and the human-machine collaborative station, specifically including: Establish a multi-dimensional evaluation index system that includes dismantling risk level, operational complexity, and human workload; Read the health index corresponding to the initial state level, and perform dimensionless and normalized processing on the specific parameters of the dismantling risk level, operation complexity and human workload; The suitability values ​​of the task unit for the robot station and the human station are calculated using a weighted evaluation model based on the health index.

[0009] Preferably, the step of pre-allocating the task unit to the corresponding execution resource specifically includes: Compare the robot workstation adaptation values ​​with the human workstation adaptation values, and divide tasks based on mandatory safety operation constraints; When the disassembly risk exceeds the system's preset safety threshold, and the robot station adaptation value is greater than the human station adaptation value, the task unit is assigned to the robot station. When it is necessary to use mechanical equipment to move large-mass components, and the operation is complex and requires personnel to confirm the status and assist in disassembly, the task unit is assigned to the human-machine collaborative workstation.

[0010] Preferably, the establishment of the AND / OR relationship model to express the disassembly priority relationship under different process requirements specifically includes: The disassembly process is converted into a directed acyclic topology graph, where nodes represent disassembly task units and directed edges represent physical constraints on the execution order between tasks. Introduce logical AND nodes and logical OR nodes into the topology graph; The logic or nodes generate automatic disassembly paths for robot workstations under normal conditions and manual intervention paths for workstations under abnormal conditions.

[0011] Preferably, the step of defining the execution time of the same task unit under different health states using triangular fuzzy numbers specifically includes: A triangular fuzzy number representing the time boundary of an operation is constructed by combining three time variables: ideal time, normal time, and abnormal time. The expected value statistical algorithm is used to transform fuzzy working hours into deterministic time parameters; The transformed expected execution time and AND / OR relational model data containing branch constraints are stored and used as boundary conditions input into the subsequent decomposition line balancing optimization process.

[0012] Preferably, the construction of the dismantling line balance optimization model, which uses an improved discrete optimization algorithm to solve for the execution positions and work sequences of the task units and generate an initial dismantling operation plan, specifically includes: A multi-objective evaluation function is established with production line load balancing, safety risk reduction, and operating cost control as optimization objectives. The dismantling line balance optimization model is constructed using the dismantling priority relationship topology constraints, human-machine operation boundary constraints, physical workstation space capacity constraints, and cycle time upper limit constraints as boundary conditions. An improved discrete sparrow search algorithm is used to solve the dismantling line balance optimization model, and the model is decoded in reverse and converted to output the corresponding process allocation details table to generate the initial dismantling operation plan.

[0013] Preferably, the step of organizing the hardware layout according to a fishbone-shaped master-slave parallel architecture to control the movement of the battery to be disassembled along the main conveyor line to the corresponding area specifically includes: The main conveyor line carries the batteries to be disassembled and transfers them, and multiple branch lines connected to both sides of the main conveyor line receive the separated and unloaded substructures. According to the control commands, drive the motor assemblies on the main conveyor line and the branch line to move the battery and substructure to be disassembled to the robot station, manual station and human-machine collaborative station deployed along the line; A bypass buffer zone and alternative workstations are set up at the physical boundary between the robot workstation and the manual workstation, and a two-way logistics channel constructed by a double-layer roller conveyor with forward and reverse drive motors and an automated guided vehicle guides the forward and reverse flow of the batteries to be disassembled.

[0014] Preferably, the triggering of the bidirectional dynamic task migration strategy, which enables the task unit to move between the robot station and the manual station, specifically includes: When human fatigue exceeds the system's preset physiological threshold or the backlog of human workstation tasks exceeds the set limit, a positive human-to-machine task migration strategy is triggered. When the robot station experiences visual recognition failure, end-effector positioning error, or torque feedback exceeding limits due to severe bolt corrosion, a reverse robot-to-human task migration strategy is triggered, an action pause command is issued, and a safety risk assessment is performed. When the assessment determines that there is no risk of high voltage exposure or thermal runaway, the transfer mechanism is activated to remove the abnormal object to be dismantled from the main conveyor line and send it to a physically adjacent bypass buffer for temporary storage. The process of recording data and correcting task adaptation values ​​for similar batteries specifically includes: After a single task migration is completed and the object to be disassembled leaves the temporary storage area, the task queue length and the remaining available capacity of the bypass buffer for the corresponding workstation in the underlying system are updated synchronously. The trigger code, the characteristic parameters of the task units involved, the migration time, and the final handling result of the migration process are summarized and packaged into a structured log. By using the number of successful feature task migrations and the total number of migration triggers in the historical feedback logs, and combining them with the set learning rate constant, the weight parameters for the adaptation numerical calculation are corrected.

[0015] The design and dynamic scheduling system of the waste power battery human-machine collaborative dismantling line includes an information acquisition module, a task allocation module, an algorithm optimization module, a scheduling control module, and a task allocation module. It includes an information acquisition module, a task allocation module, an algorithm optimization module, a scheduling and control module, and an execution unit module; The information collection module collects basic information about the battery to be disassembled, assesses the structural integrity of the battery based on the basic information, and classifies the initial state of the battery to be disassembled into a healthy state, a sub-healthy state, and a high-risk state. The task allocation module decomposes the disassembly operation into multiple independent task units, establishes a multi-dimensional evaluation index system, calculates the adaptation value of the task unit to the robot station, the manual station and the human-machine collaborative station based on the initial state and the multi-dimensional evaluation index system, and pre-allocates the task unit to the corresponding execution resources. The algorithm optimization module establishes an AND-OR relationship model based on the structural characteristics of the power battery pack to express the dismantling priority relationship under different process requirements. It uses triangular fuzzy numbers to define the execution time of the same task unit under different health states. At the same time, it constructs a dismantling line balance optimization model and uses an improved discrete optimization algorithm to solve the execution position and operation sequence of the task unit to generate an initial dismantling operation plan. The scheduling and control module sends the initial dismantling operation plan to the execution unit. The execution unit organizes the hardware layout according to the fishbone master-slave parallel architecture and controls the battery to be dismantled to move along the main conveyor line to the corresponding area. The execution unit module acquires the running status in real time during the disassembly operation according to the initial disassembly operation plan, triggers a two-way dynamic task migration strategy, so that the task unit moves between the robot station and the manual station, records the data and feeds it back to the task allocation module, and corrects the task adaptation value of the same type of battery.

[0016] This invention provides a design and dynamic scheduling method for a human-machine collaborative dismantling line for spent power batteries. It offers the following advantages: This invention calculates a comprehensive health index by extracting the physical damage and electrical performance data of the batteries to be disassembled, and establishes a multi-dimensional evaluation index system by combining disassembly risk, operational complexity, and human workload. This feature enables the system to quantitatively assess the actual condition of each retired battery, accurately assigning high-risk, standardized tasks to automated equipment, and entrusting complex tasks requiring flexible judgment to manual handling. This avoids frequent alarms and shutdowns of automated equipment that relies solely on preset programs due to changing operating conditions, thereby improving the basic operational efficiency of the entire production line.

[0017] This invention uses triangular fuzzy numbers to define the execution time of task units under different health states, and introduces a logical AND-OR relation model to express normal automatic disassembly paths and abnormal manual intervention paths. An improved discrete optimization algorithm is used to solve a multi-objective disassembly line balancing model, transforming fuzzy, uncertain working conditions into quantitative production scheduling sequences. This reduces the interference of non-standard working conditions on the overall production line cycle time and achieves effective load balancing among physical workstations.

[0018] This invention utilizes a scheduling and control module to monitor equipment error messages, worker fatigue, and task backlog in real time during operations. When a local anomaly occurs, a bidirectional dynamic task migration strategy is triggered. Combined with a bypass buffer and bidirectional logistics channels in the underlying physical space, the system can quickly remove obstructed objects from the main line, transferring tasks between robot and manual intervention stations. Feedback logs of the processing results are used to correct the initial task allocation weight parameters. This mechanism prevents line-wide shutdowns caused by local anomalies, isolates potential high-pressure and thermal runaway risks, and endows the production line with self-optimization capabilities. Attached Figure Description

[0019] Figure 1 This is a system architecture diagram of the present invention.

[0020] Figure 2 This is a flowchart of the method steps of the present invention. Detailed Implementation

[0021] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] Please see the appendix Figure 1 This invention provides a design and dynamic scheduling system for a human-machine collaborative dismantling line for waste power batteries, including an information acquisition module, a task allocation module, an algorithm optimization module, a scheduling control module, and an execution unit.

[0023] The information acquisition module is used to obtain basic physical information and health status data of the used power batteries entering the dismantling line.

[0024] The task allocation module is used to generate an initial task allocation plan based on the acquired data and to define the work boundaries between manual workstations and robot workstations.

[0025] The algorithm optimization module is used to construct a dismantling path priority relationship model that takes into account non-standard characteristics, and outputs a work plan after the dismantling line balance optimization.

[0026] The scheduling and control module is used to monitor the status of the execution unit in real time during the disassembly process and trigger dynamic task migration instructions according to preset conditions.

[0027] The execution unit receives the work instructions issued by the scheduling and control module.

[0028] The execution unit includes a main conveyor line, branch lines, robot workstations, manual workstations, human-robot collaborative workstations, and bypass buffer zones. Objects to be disassembled flow between these hardware devices via a two-way logistics channel.

[0029] Please see the appendix Figure 2 This invention provides a design and dynamic scheduling method for a human-machine collaborative dismantling line for waste power batteries, comprising the following steps: S1. The information acquisition module collects basic information about the battery to be disassembled. The basic information includes the external dimensions, connection method, battery management system data, and external damage. The information acquisition module assesses the structural integrity of the battery to be disassembled based on the basic information and classifies the initial state of the battery to be disassembled into a healthy state, a sub-healthy state, and a high-risk state. S2, using the task allocation module to decompose the disassembly operation into multiple independent task units, the task allocation module establishes a multi-dimensional evaluation index system including disassembly risk, operation complexity and human workload, calculates the adaptation value of the task unit to the robot station, human station and human-machine collaborative station based on the initial state and the multi-dimensional evaluation index system, and pre-allocates the task unit to the corresponding execution resources. S3 utilizes the algorithm optimization module to establish an AND-OR relationship model based on the structural characteristics of the power battery pack to express the disassembly priority relationship under different process requirements. It uses triangular fuzzy numbers to define the execution time of the same task unit under different health states, and transforms discrete abnormal conditions into mathematical model parameters. S4. Using the algorithm optimization module, with the optimization objectives of production line load balancing, safety risk reduction and operation cost control, a disassembly line balancing optimization model is constructed under the conditions of setting human-machine operation boundaries, workstation capacity and cycle time constraints. An improved discrete optimization algorithm is used to solve the execution workstation and operation sequence of the task unit and generate the initial disassembly operation plan. S5, the scheduling and control module sends the initial dismantling operation plan to the execution unit. The execution unit organizes the hardware layout according to the fishbone master-slave parallel architecture and controls the battery to be dismantled to move along the main conveyor line to the corresponding area. The robot station performs high-risk repetitive tasks, the human station performs flexible judgment tasks, and the human-machine collaborative station and bypass buffer support the temporary storage and transfer of the object to be dismantled in physical space. S6, the scheduling and control module acquires the real-time operating status of the robot station, the manual station, the object to be disassembled, and the bypass buffer during the disassembly operation. Based on the fatigue level of the workers, the equipment error information, or the backlog of tasks, the scheduling and control module triggers a two-way dynamic task migration strategy to adjust the flow path of the object to be disassembled, so that the task unit can be transferred between the robot station and the manual station. After the task migration action is completed, the station task queue and the occupancy status of the bypass buffer are updated. The triggering reason, processing method and processing result of the task migration are recorded, and the recorded data is fed back to the task allocation module to correct the task adaptation value and abnormal handling rules of the same type of battery.

[0030] The technical solution of the present invention will be described in detail below with reference to specific steps.

[0031] Step S1 primarily addresses the non-standard nature of spent power batteries due to varying service environments and lifespans. It uses an information acquisition module to quantify and extract the physical and data states of the batteries, providing boundary input for subsequent task scheduling. The information acquisition module's actions specifically include the following sub-steps: S101, the information acquisition module collects multi-dimensional basic information about the spent power batteries to be dismantled. After the spent power batteries are transported to the inspection station, the information acquisition module uses an industrial camera to acquire images of the batteries' appearance, extracting their dimensions, bolt connection methods, and areas covered by structural adhesive. The information acquisition module also uses image recognition algorithms to extract the extent of surface damage, quantifying the depth of shell deformation and the ratio of surface corrosion area.

[0032] The information acquisition module connects to the external interface of the used power battery via a communication cable, reads historical operating logs and underlying data from the battery management system, and extracts internal resistance values, historical maximum temperature, number of charge-discharge cycles, and individual cell voltage ranges. The information acquisition module also calls an insulation tester to detect the insulation resistance between the battery's high-voltage terminal and the casing.

[0033] For the extraction of features from industrial camera images and the parsing of communication protocols of battery management systems, those skilled in the art can use conventional machine vision algorithms and CAN bus communication technology. The specific parsing process is well-known in the field and will not be elaborated here.

[0034] S102, the information acquisition module assesses the structural integrity of the battery to be disassembled based on the collected basic information and classifies the initial state of the battery into levels. The information acquisition module extracts the physical and electrical performance characteristics obtained in the previous steps and constructs a comprehensive health assessment function. To prevent assessment failure due to missing feature data, the information acquisition module uses a weighted calculation method to integrate electrical performance health, physical damage, and insulation safety, and outputs a health index characterizing the overall state of the used power battery.

[0035] The specific calculation logic of the health index is as follows: ; In the formula, The health index of used power batteries; The electrical performance health score is calculated from the battery management system data, with a value ranging from 0 to 1. The physical damage level is determined by visual image recognition, with a value ranging from 0 to 1. The insulation safety factor is converted for insulation resistance detection. It is set to 1 when the insulation resistance is greater than the set safety threshold, and 0 otherwise. For electrical performance weighting coefficients; These are the physical state weighting coefficients; These are the insulation state weighting coefficients, and the sum of the three weighting coefficients is 1.

[0036] Based on the calculated health index, the information acquisition module classifies the initial state of the battery to be disassembled into three levels: healthy, sub-healthy, and high-risk. When the health index is in the highest set range, the insulation safety factor is 1, and no obvious leakage or severe deformation features are extracted from the appearance image, the information acquisition module marks it as healthy, indicating that the battery meets the requirements for standardized automated disassembly.

[0037] When the health index is within the medium set range, and the appearance image shows localized corrosion or slight aging characteristics, or the battery management system data is incomplete, the information acquisition module marks it as sub-healthy, indicating that the battery may have deviated from the process path during disassembly. When the health index is below the minimum set threshold, or the insulation safety factor is 0, or the appearance image identifies large-area leakage and severe structural distortion, the information acquisition module marks it as high-risk, indicating that the battery has a potential for thermal runaway or cannot be matched with conventional tools, and directly enters the manual intervention and special protection process.

[0038] The information collection module packages the determined initial state level into a state information table and sends it to the task allocation module.

[0039] Step S2 addresses the issue of defining the dismantling operation boundaries, changing the traditional approach that relies solely on electrical performance indicators to determine battery allocation. It utilizes a task allocation module to transform non-standard characteristics into quantifiable allocation criteria. The task allocation module specifically executes the following steps: S201, the task allocation module, based on the reverse engineering principle of power battery assembly process, decomposes the overall dismantling operation of waste power batteries into multiple independently executable task units. The task allocation module assigns inherent attribute tags to each task unit, including task number, preceding tasks, subsequent tasks, required dismantling tools, and three-dimensional work space constraints.

[0040] The task allocation module establishes a multi-dimensional evaluation index system for assessing task units. This system includes risk level, operational complexity, and human workload.

[0041] Disassembly risk level is used to characterize the probability and severity of consequences of high-voltage electric shock, fire, liquid leakage, and heavy object falling during the execution of a task unit.

[0042] Task complexity is used to characterize the degree of non-standardization involved in task units, such as wire harness tangling, fastener stripping and rusting, structural adhesive bonding, and machine vision positioning failure.

[0043] Human workload is used to characterize the weight a worker bears when performing a task, the frequency of repetitive work, and the risk of fatigue accumulation due to restricted posture.

[0044] In step S202, the task allocation module, based on the health index obtained in step S1, calculates the suitability values ​​of each task unit for the robot station and the human station. The task allocation module reads the health index corresponding to the initial state level of the waste power battery and normalizes the specific parameters of dismantling risk, operation complexity, and human workload.

[0045] For the dedimensionalization and normalization calculation of multi-attribute index data, those skilled in the art can use the extreme value processing method or the Z-score standardization method. The specific data conversion logic is a well-known technology in this field and will not be elaborated here.

[0046] The task allocation module uses a weighted evaluation model to calculate the suitability values ​​of the task unit for both robot and human workstations. The calculation formula is as follows: ; ; In the formula, The values ​​for adapting the task unit to the robot workstation; The values ​​for adapting the task unit to the human workstation; The normalized dismantling risk level; This represents the normalized job complexity. Normalized human workload; The health index of used power batteries; , , Weighting coefficients calculated for robot workstation adaptation; , , Weighting coefficients calculated for human workstation adaptation.

[0047] S203, the task allocation module formulates pre-allocation rules for human-machine operation boundaries based on the calculated adaptation values, and pre-assigns task units to corresponding execution resources. The task allocation module compares the robot workstation adaptation values ​​with the human workstation adaptation values, and divides tasks in conjunction with mandatory safety operation constraints.

[0048] When the disassembly risk exceeds the set safety threshold, or when the task unit involves high-voltage disconnection and heavy-duty handling operations, and the robot station's compatibility value is greater than the human station's compatibility value, the task allocation module assigns the task unit to the robot station. Tasks assigned to the robot station typically have a clear tool operation path and rely on a closed protective zone to isolate potential hazards.

[0049] When the task is highly complex, involving severe corrosion of fasteners, disassembly of non-standard irregular structures, and organization of soft wiring harnesses, and the human workload does not exceed the physiological safety limit, the task allocation module will assign the task unit to the human workstation.

[0050] Manual workstations rely on operators' visual judgment and flexible operation to handle non-standard conditions. When the disassembly risk of a task unit is high, requiring the use of mechanical equipment to move large components, and the operation is complex enough to require personnel to confirm the status and assist in disassembly, the task allocation module assigns the task unit to a human-machine collaborative workstation.

[0051] The allocation module outputs a mapping table of processes, workstations and execution resources based on the above judgment rules, and generates an initial task allocation plan.

[0052] Step S3 primarily addresses the issues of variable process paths and fluctuating operation times caused by differences in the condition of spent power batteries. It constructs a mathematical model compatible with non-standard characteristics through an algorithm optimization module, providing fundamental parameters for subsequent production line balancing and scheduling. The algorithm optimization module specifically executes the following steps: S301, the algorithm optimization module establishes an AND-OR relationship model based on the structural characteristics of the power battery pack to express the disassembly priority relationship under different process requirements. The algorithm optimization module converts the disassembly process into a directed acyclic topology graph, where nodes represent disassembly task units and directed edges represent physical constraints on the execution order between tasks.

[0053] The algorithm optimization module introduces logical "AND" nodes and logical "OR" nodes into the topology graph. Logical "AND" nodes indicate that all subordinate branch tasks must be executed in a strict sequential order before proceeding to the next level, constraining the basic process flow for disassembling standard structural components of the battery pack. Logical "OR" nodes indicate that subordinate branch tasks have parallel substitution relationships; executing only one associated path is sufficient to satisfy the process node requirements, representing flexible process branches arising from differences in health status.

[0054] When the connection status of external fasteners cannot be determined, the algorithm optimization module generates the automatic disassembly path of the robot station under normal conditions and the manual intervention path of the station under abnormal conditions through the logical "OR" node.

[0055] The aforementioned topology enables the disassembly model to extend its path to cover non-standard conditions such as rust, slippage, structural deformation, and excessive adhesion of structural adhesive.

[0056] S302, the algorithm optimization module uses triangular fuzzy numbers to define the execution time of the same task unit under different health states, transforming the time fluctuations caused by non-standard states into mathematical model parameters. The time consumption of a single task unit in actual operation fluctuates significantly due to the aging degree of the used power battery.

[0057] The algorithm optimization module uses three discrete time variables to set the working hours of each task unit. The three time variables include ideal time, normal time, and abnormal time.

[0058] Ideal time represents the shortest operation time with the battery structure intact and connections easily detachable.

[0059] The standard time represents the time required for processing used power batteries in a normal decommissioning state.

[0060] Abnormal time indicates the longest operation time under conditions of severe fastener corrosion, obvious shell deformation, or difficulty in machine vision recognition.

[0061] The algorithm optimization module combines the above three time variables to construct a triangular fuzzy number that represents the time boundary of the operation.

[0062] For the basic algebraic operations of triangular fuzzy numbers, those skilled in the art can use the centroid method or the integral mean method for processing. The specific data defuzzification logic is a well-known technology in this field and will not be elaborated here.

[0063] To support quantitative production scheduling for production line balancing, the algorithm optimization module uses an expected value statistical algorithm to transform fuzzy time hours into deterministic time parameters. The calculation formula is as follows: ; In the formula, Calculate the expected execution time for the task unit output; The ideal time for the task unit to be in an ideal state; This refers to the normal time when the mission unit is in a general decommissioning state; This refers to the abnormal time when the task unit is in a severely damaged state. The algorithm optimization module stores the transformed expected execution time and the AND / OR relation model data containing branch constraints, and uses it as boundary conditions input into the subsequent decomposition line balancing optimization process.

[0064] Step S4 primarily addresses the problem of rationally allocating uncertain tasks to specific hardware resources under multidimensional constraints. It involves constructing a mathematical optimization model through the algorithm optimization module and solving for the optimal production sequence to reduce fluctuations and interference during the allocation process. S401, the algorithm optimization module constructs a production line balancing optimization model. The algorithm optimization module establishes a multi-objective evaluation function covering three dimensions: production line load balancing, safety risk reduction, and operational cost control. Production line load balancing is used to reduce the deviation between the actual allocated work time at each workstation and the theoretical cycle time of the production line, avoiding task backlog or resource idleness at local workstations.

[0065] Safety risk reduction is used to concentrate high-risk tasks in enclosed robotic workstations with high protection levels, thereby reducing the overall level of hazard exposure.

[0066] The algorithm optimization module sets rigid constraints on the model. These constraints include the topological constraints of the dismantling priority relationship generated in step S3, the human-machine operation boundary constraints set by the pre-allocation results in step S2, the physical workstation space capacity constraints, and the cycle time limit constraints. The cycle time limit constraints strictly require that the total expected execution time of tasks assigned to a single workstation cannot exceed the set production line cycle time.

[0067] The formula for the multi-objective integrated fitness function constructed by the algorithm optimization module is as follows: ; In the formula, The value of the multi-objective comprehensive evaluation function; The total number of workstations to be planned and arranged; The theoretical cycle time for a single workstation on the production line is set. For the first The expected total working hours for tasks assigned to each workstation; The total number of task units to be assigned; For the first The risk level of dismantling each task unit; As a decision variable, when the first The task unit was assigned to the first The value is 1 when the workstation is executed, otherwise the value is 0. The estimated total operating cost corresponding to the allocation plan; The target weights for load balancing; Weighting of safety risk objectives; The target weight for activity cost.

[0068] S402, the algorithm optimization module uses an improved discrete sparrow search algorithm to solve the above-mentioned splitting line balance optimization model. The algorithm optimization module discretizes the traditional continuous intelligent search algorithm to adapt to the discrete scheduling requirements of splitting and production. The algorithm optimization module uses a chaotic mapping mechanism to initialize the algorithm's population position, ensuring that the initial task sorting sequence is evenly distributed in the solution space, increasing the diversity of global optimization starting points to avoid getting trapped in local dead zones.

[0069] The algorithm optimization module employs a two-layer discretization coding rule. One layer encodes the execution order identifier of the task unit, while the other layer encodes the workstation allocation identifier of the task unit, thus mapping the mathematical solution space to a physical decomposition logical sequence. The algorithm optimization module introduces a dynamic weight control strategy to adjust the position update step size. In the early stages of iterative optimization, this expands the search range of the solution space, while in the later stages, it reduces the step size to accelerate convergence to the optimal solution.

[0070] For the design of chaotic mapping functions and the specific crossover and mutation operations of continuous algorithm discretization, those skilled in the art can use Tent mapping and discrete crossover and mutation operators to implement them. The principle of discretization optimization calculation is a well-known technology in this field and will not be elaborated here.

[0071] S403, after reaching the maximum set number of iterations, the algorithm optimization module outputs the solution that satisfies all constraints and has the optimal comprehensive objective evaluation function value. The algorithm optimization module then performs reverse decoding on the optimal solution, converting it into a corresponding process allocation detail table. The algorithm optimization module generates an initial decomposition work plan that clearly defines the execution station, resource type, and execution order of each task unit. This initial decomposition work plan is subsequently sent to the scheduling control module as a baseline timetable to guide the execution units in performing physical actions.

[0072] Step S5 primarily addresses the physical execution of the initial dismantling operation plan issued by the scheduling and control module. It involves organizing production line hardware resources of a specific configuration through the execution unit to spatially support the logistics diversion and temporary storage needs caused by differences in health status. The scheduling and control module and the execution unit cooperate to execute the following sub-steps: S501, the scheduling and control module parses the received initial disassembly plan, converts the control commands, and sends them to the programmable logic controller (PLC) at the bottom layer of the execution unit. The execution unit organizes its underlying hardware layout according to a herringbone master-slave parallel architecture. The herringbone master-slave parallel architecture includes a main conveyor line running through the entire production line and multiple branch lines that are vertically or obliquely connected to both sides of the main conveyor line.

[0073] The main conveyor line is used to carry out the sequential transfer of complete waste power battery packs and battery boxes with large volume and weight.

[0074] Multiple branch lines are used to receive the lower-level substructures separated and unloaded from the main conveyor line, including the independent and detailed disassembly of battery modules, high-voltage wiring harness assemblies, and cooling pipeline components.

[0075] The execution unit drives the motor assemblies on the main conveyor line and branch line according to the control instructions, so that the battery and sub-components to be disassembled move to the target node area according to the set rhythm.

[0076] S502, the execution unit deploys different types of execution stations along the main conveyor line and branch lines according to the task hazard level and operational flexibility requirements. The execution unit sets up the robot station within a closed protective area equipped with physical protective fences and safety light curtains.

[0077] The robot workstation is equipped with an industrial six-axis robot and a vision guidance system, which is used to receive instructions from the scheduling and control module to perform tasks such as disconnecting high-voltage terminals, moving large-mass top covers, and removing non-rusted standard bolts.

[0078] The execution unit deploys manual workstations on the sides of the main conveyor line and in a dedicated manual intervention area. These workstations are equipped with manual and pneumatic tools and local lighting, relying on the operator's subjective judgment to perform tasks such as stripping flexible hoses, straightening and organizing wire harnesses, and destructive removal of severely rusted and stripped fasteners.

[0079] The execution unit deploys a human-machine collaborative workstation at the intersection of the main conveyor line and the branch line. Within this workstation area, an auxiliary robotic arm equipped with force feedback sensors and personnel work positions are deployed, with physical safety isolation and emergency stop devices ensuring personnel safety. The human-machine collaborative workstation is used for performing mixed-type dismantling tasks involving heavy-duty lifting and requiring manual hole positioning.

[0080] For the visual guidance trajectory planning of industrial robots and the hard-wired interlocking control of safety light curtains, those skilled in the art can use conventional robot control algorithms and industrial safety relay circuits to achieve this. The specific motion control principle is a well-known technology in this field and will not be elaborated here.

[0081] S503, the execution unit is equipped with a bypass buffer and alternative workstation at the physical boundary between the robot workstation and the human workstation, and connects each work area using a two-way logistics channel.

[0082] The bypass buffer is set at the side extension position outside the main track of the main conveyor line. It is used to temporarily store battery packs or modules that cannot be processed immediately due to disassembly abnormalities when a local work station is blocked.

[0083] The alternative workstations are general-purpose disassembly workbenches in their idle state, used to continue processing non-standard tasks temporarily transferred from the main production queue. The two-way logistics channel is constructed using double-layer roller conveyors with forward and reverse drive motors and automated guided vehicles.

[0084] When localized task backlogs or detours occur, the bidirectional logistics channel changes its transmission direction based on the underlying scheduling instructions of the scheduling control module. This allows the objects to be dismantled to flow forward and backward between the main conveyor line, branch lines, manual intervention stations, robot stations, and bypass buffer zones. This physical space configuration provides the hardware foundation for handling dynamic task migration under abnormal dismantling conditions.

[0085] Step S6 primarily addresses task blocking issues caused by random equipment malfunctions or excessive personnel workload during the dismantling production line operation. The scheduling and control module, based on the underlying hardware status, triggers a bidirectional dynamic task migration strategy, specifically executing the following sub-steps: The S601 scheduling and control module acquires real-time operational status data of each node on the production line via industrial Ethernet during the disassembly process. It captures the continuous working time of workers at manual workstations and the current number of tasks in the queue, calculating worker fatigue levels by combining this data with heart rate readings from external wearable devices. The module also reads the peak current of the servo motors at the robot workstations, the visual matching confidence score, and fault alarm codes from the internal controller.

[0086] The scheduling and control module acquires signals from photoelectric sensors to determine the physical space occupancy of the bypass buffer and alternative workstations. The module also continuously reads temperature sensor readings from the surface of the object to be disassembled, as well as high-voltage leakage status reported by the insulation monitoring module.

[0087] For data capture and analog-to-digital conversion of sensor signals from industrial Ethernet networks, those skilled in the art can use a low-level programmable logic controller (PLC) communication program. The specific network communication protocol parsing method is a well-known technology in the field and will not be elaborated here.

[0088] S602, based on the collected status data, the scheduling control module triggers a positive human-to-machine task migration strategy when human fatigue exceeds a preset physiological threshold or the backlog of tasks at the human workstation exceeds a set upper limit. The scheduling control module searches the current queue of tasks to be processed at the human workstation and selects standardized task units with a work complexity lower than the set benchmark and no obvious damage to their appearance.

[0089] The scheduling and control module queries the load status of adjacent robot workstations and the matching status of the required disassembly tools. If a robot workstation has a matching tool and its task queue is not full, the scheduling and control module sends a path change command to the execution unit. The execution unit drives the bidirectional logistics channel to move the corresponding object to be disassembled from the manual workstation and transfer it to the target robot workstation.

[0090] The robot station then performs tasks such as loosening fasteners or moving standard parts. This logic allows manual tasks to be promptly diverted to idle automated equipment, alleviating production line slowdowns caused by backlogs at single workstations.

[0091] S603: When a robot station experiences visual recognition failure, end-effector positioning errors, or torque feedback exceeding limits due to severe bolt corrosion, the scheduling and control module triggers a reverse robot-to-human task migration strategy. The scheduling and control module issues a pause command to the robot station experiencing the anomaly and extracts the real-time temperature and insulation status of the object to be disassembled for a safety risk assessment.

[0092] When the assessment determines that there is no risk of high-voltage exposure or thermal runaway, the scheduling control module marks the task unit that caused the execution interruption as a manual intervention task. The execution unit activates the transfer mechanism to remove the abnormal object to be dismantled from the main conveyor line and send it to a physically adjacent bypass buffer for temporary storage. Based on the bypass buffer capacity and the workload of the manual workstations, the execution unit uses a two-way logistics channel to guide the temporarily stored object to an idle manual intervention workstation or an alternative workstation.

[0093] Workers use manual destructive tools to remove stripped bolts or peel off aged structural adhesive. If the initial safety risk assessment results exceed the controllable range, the dispatch control module instructs the execution unit to push the object into an explosion-proof isolation chamber to await professional personnel to handle the hazard, prohibiting it from entering the regular dismantling queue.

[0094] S604: After a single task migration is completed and the object to be dismantled leaves the temporary storage area, the scheduling control module synchronously updates the task queue length and the remaining available capacity of the bypass buffer for the corresponding workstation in the underlying system. The scheduling control module summarizes the triggering reason code, the characteristic parameters of the task units involved, the migration time, and the final handling result during the migration process, and packages the above data into a structured log and feeds it back to the task allocation module.

[0095] The task allocation module uses historical feedback logs to correct the weight parameters for the adaptive numerical calculation in step S2, and updates the calculation formula as follows:

[0096] In the formula, For the updated version New weighting parameters for each evaluation indicator in the adaptive numerical calculation; For the first time before the update The original weight parameters of each evaluation indicator; The learning rate constant is set, and its value ranges from 0 to 1; This represents the number of times this type of feature task migration processing was successfully completed in the historical feedback log. The total number of migrations triggered for this type of feature task.

[0097] The task allocation module uses updated weight parameters to adjust the task pre-allocation rule boundary when subsequent similar waste power batteries enter the production line, forming a closed-loop mechanism that enables the dismantling system to gradually adapt to the physical property distribution characteristics of complex non-standard batteries.

[0098] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A design and dynamic scheduling method for a human-machine collaborative dismantling line for waste power batteries, characterized in that, Includes the following steps: Basic information about the battery to be disassembled is collected, and the structural integrity of the battery to be disassembled is assessed based on the basic information. The initial state of the battery to be disassembled is divided into healthy state, sub-healthy state and high-risk state. The disassembly operation is broken down into multiple independent task units, a multi-dimensional evaluation index system is established, and the adaptation values ​​of the task units to robot workstations, manual workstations and human-machine collaborative workstations are calculated based on the initial state and the multi-dimensional evaluation index system. The task units are then pre-allocated to the corresponding execution resources. Based on the structural characteristics of the power battery pack, an AND-OR relationship model is established to express the dismantling priority relationship under different process requirements. Triangular fuzzy numbers are used to define the execution time of the same task unit under different health states. A dismantling line balance optimization model is constructed, and an improved discrete optimization algorithm is used to solve the execution position and operation sequence of the task unit to generate an initial dismantling operation plan. According to the initial disassembly operation plan, the hardware layout is organized in a fishbone-shaped master-slave parallel architecture to control the movement of the batteries to be disassembled along the main conveyor line to the corresponding area and execute the operation. According to the initial disassembly operation plan, the operating status is acquired in real time during the disassembly operation, triggering a two-way dynamic task migration strategy, which enables the task unit to move between the robot station and the manual station, and records data and corrects the task adaptation values ​​of the same type of battery.

2. The design and dynamic scheduling method for a human-machine collaborative dismantling line for waste power batteries according to claim 1, characterized in that, The process involves assessing the structural integrity of the battery to be disassembled based on basic information, and classifying the initial state of the battery into healthy, sub-healthy, and high-risk states, specifically including: Extract the physical and electrical performance characteristics from the basic information and construct a comprehensive health assessment function; A weighted calculation method is used to integrate electrical performance health, physical damage, and insulation safety to output a health index that characterizes the overall condition of used power batteries. Based on the calculated health index, the initial state of the batteries to be disassembled is divided into three levels: healthy state, sub-healthy state, and high-risk state.

3. The design and dynamic scheduling method for a human-machine collaborative dismantling line for waste power batteries according to claim 1, characterized in that, The initial state and the multi-dimensional evaluation index system calculate the adaptation values ​​of the task unit for the robot station, the manual station, and the human-robot collaborative station, specifically including: Establish a multi-dimensional evaluation index system that includes dismantling risk level, operational complexity, and human workload; Read the health index corresponding to the initial state level, and perform dimensionless and normalized processing on the specific parameters of the dismantling risk level, operation complexity and human workload; The suitability values ​​of the task unit for the robot station and the human station are calculated using a weighted evaluation model based on the health index.

4. The design and dynamic scheduling method for a human-machine collaborative dismantling line for waste power batteries according to claim 3, characterized in that, The step of pre-allocating the task units to corresponding execution resources specifically includes: Compare the robot workstation adaptation values ​​with the human workstation adaptation values, and divide tasks based on mandatory safety operation constraints; When the disassembly risk exceeds the system's preset safety threshold, and the robot station adaptation value is greater than the human station adaptation value, the task unit is assigned to the robot station. When it is necessary to use mechanical equipment to move large-mass components, and the operation is complex and requires personnel to confirm the status and assist in disassembly, the task unit is assigned to the human-machine collaborative workstation.

5. The design and dynamic scheduling method for a human-machine collaborative dismantling line for waste power batteries according to claim 1, characterized in that, The established AND / OR relationship model expresses the disassembly priority relationship under different process requirements, specifically including: The disassembly process is converted into a directed acyclic topology graph, where nodes represent disassembly task units and directed edges represent physical constraints on the execution order between tasks. Introduce logical AND nodes and logical OR nodes into the topology graph; The logic or nodes generate automatic disassembly paths for robot workstations under normal conditions and manual intervention paths for workstations under abnormal conditions.

6. The design and dynamic scheduling method for a human-machine collaborative dismantling line for waste power batteries according to claim 1, characterized in that, The method of defining the execution time of the same task unit under different health states using triangular fuzzy numbers specifically includes: A triangular fuzzy number representing the time boundary of an operation is constructed by combining three time variables: ideal time, normal time, and abnormal time. The expected value statistical algorithm is used to transform fuzzy working hours into deterministic time parameters; The transformed expected execution time and AND / OR relational model data containing branch constraints are stored and used as boundary conditions input into the subsequent decomposition line balancing optimization process.

7. The design and dynamic scheduling method for a human-machine collaborative dismantling line for waste power batteries according to claim 1, characterized in that, The construction of the dismantling line balance optimization model, using an improved discrete optimization algorithm to solve for the execution positions and work sequences of the task units, and generating an initial dismantling operation plan, specifically includes: A multi-objective evaluation function is established with production line load balancing, safety risk reduction, and operating cost control as optimization objectives. The dismantling line balance optimization model is constructed using the dismantling priority relationship topology constraints, human-machine operation boundary constraints, physical workstation space capacity constraints, and cycle time upper limit constraints as boundary conditions. An improved discrete sparrow search algorithm is used to solve the dismantling line balance optimization model, and the model is decoded in reverse and converted to output the corresponding process allocation details table to generate the initial dismantling operation plan.

8. The design and dynamic scheduling method for a human-machine collaborative dismantling line for waste power batteries according to claim 1, characterized in that, The process of organizing the hardware layout according to a fishbone-shaped master-slave parallel architecture to control the movement of the batteries to be disassembled along the main conveyor line to the corresponding area specifically includes: The main conveyor line carries the batteries to be disassembled and transfers them, and multiple branch lines connected to both sides of the main conveyor line receive the separated and unloaded substructures. According to the control commands, drive the motor assemblies on the main conveyor line and the branch line to move the battery and substructure to be disassembled to the robot station, manual station and human-machine collaborative station deployed along the line; A bypass buffer zone and alternative workstations are set up at the physical boundary between the robot workstation and the manual workstation, and a two-way logistics channel constructed by a double-layer roller conveyor with forward and reverse drive motors and an automated guided vehicle guides the forward and reverse flow of the batteries to be disassembled.

9. The design and dynamic scheduling method for a human-machine collaborative dismantling line for waste power batteries according to claim 1, characterized in that, The triggering of the bidirectional dynamic task migration strategy enables the task unit to move between the robot workstation and the manual workstation, specifically including: When human fatigue exceeds the system's preset physiological threshold or the backlog of tasks at human workstations exceeds the set limit, a positive human-to-machine task migration strategy is triggered. When the robot station experiences visual recognition failure, end-effector positioning error, or torque feedback exceeding limits due to severe bolt corrosion, a reverse robot-to-human task migration strategy is triggered, an action pause command is issued, and a safety risk assessment is performed. When the assessment determines that there is no risk of high voltage exposure or thermal runaway, the transfer mechanism is activated to remove the abnormal object to be dismantled from the main conveyor line and send it to a physically adjacent bypass buffer for temporary storage. The process of recording data and correcting task adaptation values ​​for similar batteries specifically includes: After a single task migration is completed and the object to be disassembled leaves the temporary storage area, the task queue length and the remaining available capacity of the bypass buffer for the corresponding workstation in the underlying system are updated synchronously. The trigger code, the characteristic parameters of the task units involved, the migration time, and the final handling result of the migration process are summarized and packaged into a structured log. By using the number of successful feature task migrations and the total number of migration triggers in the historical feedback logs, and combining them with the set learning rate constant, the weight parameters for the adaptation numerical calculation are corrected.

10. A design and dynamic scheduling system for a human-machine collaborative dismantling line for waste power batteries, characterized in that, It includes an information acquisition module, a task allocation module, an algorithm optimization module, a scheduling and control module, and an execution unit module; The information collection module collects basic information about the battery to be disassembled, assesses the structural integrity of the battery based on the basic information, and classifies the initial state of the battery to be disassembled into a healthy state, a sub-healthy state, and a high-risk state. The task allocation module decomposes the disassembly operation into multiple independent task units, establishes a multi-dimensional evaluation index system, calculates the adaptation value of the task unit to the robot station, the manual station and the human-machine collaborative station based on the initial state and the multi-dimensional evaluation index system, and pre-allocates the task unit to the corresponding execution resources. The algorithm optimization module establishes an AND-OR relationship model based on the structural characteristics of the power battery pack to express the dismantling priority relationship under different process requirements. It uses triangular fuzzy numbers to define the execution time of the same task unit under different health states. At the same time, it constructs a dismantling line balance optimization model and uses an improved discrete optimization algorithm to solve the execution position and operation sequence of the task unit to generate an initial dismantling operation plan. The scheduling and control module sends the initial dismantling operation plan to the execution unit. The execution unit organizes the hardware layout according to the fishbone master-slave parallel architecture and controls the battery to be dismantled to move along the main conveyor line to the corresponding area. The execution unit module acquires the running status in real time during the disassembly operation according to the initial disassembly operation plan, triggers a two-way dynamic task migration strategy, so that the task unit moves between the robot station and the manual station, records the data and feeds it back to the task allocation module, and corrects the task adaptation value of the same type of battery.