CN-122007118-A - New energy battery pack disassembling process

CN122007118ACN 122007118 ACN122007118 ACN 122007118ACN-122007118-A

Abstract

The invention relates to the technical field of battery recovery, and provides a new energy battery pack disassembling process which is implemented in a closed automatic workstation, performs battery pack entrance risk diagnosis and pretreatment through multi-source perception fusion, and autonomously generates and executes a fine disassembling action sequence according to a real-time three-dimensional state model and multi-sensor feedback based on an intelligent reinforcement learning decision module so as to realize self-adaptive and flexible separation of a shell and an internal component. The invention can improve the safety, the working condition adaptability and the nondestructive recovery rate of the core component in the disassembly process, and continuously optimize the disassembly process through a data closed-loop driving process, thereby providing a reliable intelligent solution for battery echelon utilization and material efficient recovery.

Inventors

CHEN ZEDING
HU LUNJUN
RAN FUQIANG
CHEN ZEHUI
Zhang Zecan

Assignees

宁德久鼎科技有限公司

Dates

Publication Date: 20260512
Application Date: 20260414

Claims (10)

1. The new energy battery pack disassembling process is characterized by being implemented in a closed automatic disassembling workstation with negative pressure suction and inert gas protection functions, and comprises the following steps of: the entrance diagnosis and safety preprocessing step comprises the steps of obtaining a reference digital file of a battery pack, performing non-contact scanning to identify a safety risk area, and then performing electric isolation and grading discharge; The shell self-adaptive releasing step is that a shell model is built through high-precision three-dimensional scanning, the connection characteristics are identified by combining machine vision, a disassembly sequence is generated, and the shell connection is self-adaptively released based on real-time force feedback by utilizing an intelligent actuator integrating multiple sensors and a tool library; The method comprises the steps of removing a shell, initializing the state of the shell, controlling the operation of an actuator based on a reinforcement learning decision module, wherein the reinforcement learning decision module adopts a framework combining simulation learning initialization and on-line reinforcement learning fine adjustment, and the real-time system state is a high-dimensional vector formed by fusion coding of local visual point cloud characteristics, normalized readings of six-dimensional force/moment sensors, tool head identifiers, end pose of the actuator and embedded representation of a historical action sequence; and the material processing and data optimizing step is to classify and recycle the disassembled parts, and upload the whole process data associated battery pack identity codes to a cloud digital twin platform for optimizing the disassembling strategy.
2. The new energy battery pack disassembling process according to claim 1, wherein in the entrance diagnosis and safety pretreatment step, non-contact scanning comprises the steps of adopting a thermal infrared imager to scan, marking an area with the temperature higher than 15 ℃ of the environment as a primary risk area, and adopting millimeter wave radar to scan, marking an area with the problems of structural collapse, module displacement or effusion as a secondary risk area.
3. The process for disassembling the new energy battery pack according to claim 1, wherein the step of discharging is to discharge the battery pack at a current of not more than 0.05 ℃ through a high-voltage maintenance interface of the battery pack until the total voltage of the battery pack is reduced to below 60V, wherein C is the rated capacity of the battery cell, and the rate of temperature rise of the surface of the battery pack is monitored in real time during the discharging process, and when the rate is greater than 1 ℃ per minute, the discharging is stopped and a cooling program is started.
4. A new energy battery pack disassembling process according to claim 3 is characterized in that for a battery pack marked as a primary risk area or having collision and wading histories, special pretreatment is performed after discharging, wherein the battery pack is placed in a negative pressure pretreatment cabin, inert gas is introduced and a reactive polymer polymerizing agent is injected to form a solid polymer in situ at a potential electrolyte leakage point of the battery pack for fixation.
5. A new energy battery pack disassembling process according to claim 1, wherein in the case self-adaptive disassembling step, the disassembling sequence is generated by comparing a three-dimensional scanning model with a reference digital file, marking the actual position of a case joint with a nominal position deviation of a file record as a nonstandard variation point when the deviation is greater than 5mm, and generating a disassembling action sequence comprising operation priority and risk remarks according to the principle of firstly removing electrical connection, then removing mechanical connection, firstly removing peripheral connection and then removing core connection.
6. The process for disassembling the new energy battery pack according to claim 5, wherein when the intelligent actuator is disconnected from the shell: For screw connection, the preset torque range is disassembled, when the actual torque reaches the preset upper limit and the screw is not screwed, the actuator firstly applies micro-vibration assistance with the frequency of 100Hz and the amplitude of 0.1mm in the axial direction; For the sealant connection, a constant temperature hot air knife is adopted to heat along a joint for 60 to 90 seconds at the temperature of 200+/-10 ℃ to soften the joint, and then a flexible shovel knife with force feedback is used to apply a separation force of 5 to 10N for stripping.
7. The new energy battery pack disassembling process according to claim 1, wherein the reinforcement learning decision module adopts a framework combining simulation learning initialization and online reinforcement learning fine adjustment, the real-time system state is a high-dimensional vector formed by fusion coding of local visual point cloud characteristics, six-dimensional force/moment sensor normalized readings, tool head identifications, end pose of an actuator and historical action sequence embedded representation, and the composite action instruction comprises millimeter displacement increment, newton-level acting force components and discrete commands for tool holding or switching under a tool coordinate system.
8. The process for disassembling a new energy battery pack according to claim 7, wherein the optimization objective of the reinforcement learning decision module is defined by a reward function, the reward function comprising a task completion reward for successful disassembly of the objective component, an efficiency reward for smooth execution of the action, a safety penalty for occurrence of force exceeding a threshold or unexpected deformation of the component, and a progress reward for improvement of tool alignment accuracy or increase of connection-disconnection ratio.
9. The process for disassembling the new energy battery pack according to claim 1, wherein in the step of intelligently separating the internal components: for the battery module fixed by the structural adhesive, cutting off the adhesive layer by adopting a linear ultrasonic knife with the frequency matched with the characteristics of the adhesive layer; After the module frame is removed, a self-adaptive vacuum adsorption array gripper with an independent air pressure control and micro-displacement sensor is adopted to extract the single battery cell at a speed of 5 to 10 mm/s; and (3) scanning type cutting is carried out on the laser welding points of the aluminum busbar by adopting a pulse fiber laser with the pulse energy of 20J and the pulse frequency of 100Hz, and the damage depth of the surface layer of the pole is controlled to be less than 0.1mm.
10. The process for disassembling the new energy battery pack according to claim 1, wherein in the material processing and data optimizing step, the cloud digital twin platform analyzes the uploaded whole process data by using a machine learning algorithm, digs optimized disassembling strategies under different fault or damage modes, and issues the updated strategies to each disassembling workstation to realize continuous self-adaptive optimization of a disassembling system.

Description

New energy battery pack disassembling process Technical Field The invention relates to the technical field of battery recovery, in particular to a new energy battery pack disassembling process. Background With the rapid development of the new energy automobile industry, the large-scale retirement of power battery packs has become an unavoidable reality problem. The efficient, safe and environment-friendly battery pack disassembly and recovery are key links for realizing resource recycling and industrial sustainable development. However, the current main stream disassembly process still faces a serious technical bottleneck, namely firstly, the traditional disassembly mode relying on a fixed program or manual experience is poor in adaptability and low in efficiency and is difficult to cope with unknown internal risks due to the fact that the brand, model, structure and damage state of the retired battery pack are quite different, secondly, the disassembly process involves multiple potential safety hazards such as high pressure, high temperature and electrolyte leakage, the prior art lacks systematic advanced risk sensing and self-adaptive disposal capability and is high in accident risk, thirdly, the high-value components such as the battery core, the module and the copper bar are extremely easily damaged due to rough mechanical disassembly, and the gradient utilization rate and economic value of materials are seriously reduced. Therefore, the disassembly process of the new energy battery pack is provided, and the multi-dimensional nondestructive evaluation and risk classification of the battery pack entrance state are realized through infrared thermal imaging, millimeter wave radar, electric diagnosis and cloud data fusion, and active safety discharge and targeted pretreatment are executed. And then, establishing an accurate operation map by utilizing high-precision three-dimensional scanning and machine vision, and automatically generating and executing a refined disassembly action sequence which is suitable for the current component state (such as a bolt, colloid and welding spots) through an intelligent decision module fused with a reinforcement learning algorithm and fed back by a plurality of sensors in real time. The execution process adopts flexible processes such as force-position mixed control, constant temperature softening, ultrasonic and laser cutting, and the like, so that the integrity of the component is ensured to the greatest extent. Finally, the whole process data is synchronized to a cloud digital twin platform, and continuous iteration of the disassembly strategy and integral optimization of the system are driven. Therefore, on the premise of ensuring the operation safety, the disassembly efficiency and the resource recovery value are greatly improved. Disclosure of Invention Aiming at the problems of poor adaptability, high safety risk and easy damage to high-value components of the existing disassembly method, the invention provides a disassembly process of a new energy battery pack. In order to achieve the above-mentioned purpose, the invention provides a new energy battery pack disassembling process implemented in a closed automatic disassembling workstation with negative pressure suction and inert gas protection functions, the method comprises the following steps: The entrance diagnosis and safety preprocessing step is to acquire a reference digital file of the battery pack, perform non-contact scanning to identify a safety risk area, and then execute electrical isolation and grading discharge. And a shell self-adaptive releasing step, namely constructing a shell model through high-precision three-dimensional scanning, combining machine vision to identify connection characteristics and generating a disassembly sequence, and utilizing an intelligent actuator integrating multiple sensors and a tool library to self-adaptively release the shell connection based on real-time force feedback. And the reinforcement learning decision module outputs a compound action instruction for controlling the displacement, acting force and tool switching of the actuator according to the real-time system state integrating visual characteristics, force sense feedback and process context, so as to realize self-adaptive fine separation of the internal components and simultaneously monitor and intervene in real time by an independent security daemon. And the material processing and data optimizing step is to classify and recycle the disassembled parts, and upload the whole process data associated battery pack identity codes to a cloud digital twin platform for optimizing the disassembling strategy. Further, in the entrance diagnosis and safety pretreatment steps, non-contact scanning comprises the steps of scanning by an infrared thermal imager, marking an area with the temperature higher than 15 ℃ of the environment as a primary risk area, and marking an area with the problems of structural collapse, module