A negative electrode slurry mixing process intelligent closed-loop control method and system

Abstract

The application discloses a kind of negative electrode slurry mixing process intelligent closed-loop control method and system, belong to lithium ion battery intelligent manufacturing field.The method includes: S1 system initialization and establishes healthy acoustic baseline;S2 real-time signal acquisition;S3 is diagnosed by machine learning model exception;S4 generates and executes customized dredging instruction according to exception type and location;S5 verifies dredging effect and self-optimization.The application realizes early, on-line, non-invasive monitoring and automatic dredging to the blockage such as graphite agglomeration, binder wall sticking in negative electrode slurry mixing, and forms intelligent closed-loop control.

Description

Technical Field

[0001] This invention relates to the field of pipeline blockage monitoring, and in particular to an intelligent closed-loop control method and system for the negative electrode grouting process based on acoustic detection and ultrasonic active unblocking. Background Technology

[0002] Negative electrode slurry preparation is a crucial step in lithium-ion battery production. The slurry primarily consists of graphite, conductive agent (SP), thickener (CMC), binder (SBR), and deionized water. This process faces two unique challenges:

[0003] First, SBR binder is extremely prone to adhering to the tank wall, agitator and pipes, forming a gel-like substance that is difficult to remove;

[0004] Secondly, graphite particles are prone to agglomeration, resulting in uneven dispersion of the slurry, and they are also prone to settling and accumulating in low-flow-rate areas during transportation.

[0005] These characteristics lead to frequent blockages in the negative electrode slurry mixing system. Ultrasonic vibration is effective in clearing viscous blockages in SBRs and blockages caused by the accumulation of soft materials such as graphite.

[0006] Traditional pipeline inspection methods include pressure gauge testing and manual sampling at pipeline openings. Installing pressure gauges at locations prone to blockages can indirectly reveal the extent of internal blockages, but this method is time-sensitive: a pressure gauge reading anomaly indicates a blockage has already formed, meaning the pressure gauge cannot detect the blockage formation process and cannot prevent it from worsening. Manual sampling at pipeline openings typically involves shutting down the system for inspection and cleaning, and removing introduced metal foreign objects. This can lead to moisture control failures and disruption of the internal vacuum environment, significantly impacting the quality of the negative electrode product. Therefore, there is an urgent need for an intelligent solution that enables online, real-time, non-invasive early monitoring tailored to the characteristics of negative electrode materials, and can automatically and online perform unblocking operations. Summary of the Invention

[0007] To address the existing problems, this invention provides an intelligent closed-loop control method and system for the negative electrode slurry mixing process, the specific solution of which is as follows:

[0008] A smart closed-loop control method for the negative electrode slurry preparation process is applied to the preparation stage of lithium-ion battery negative electrode slurry. The method includes the following steps:

[0009] S1: System initialization, when the negative electrode mixing machine is running standard formula slurry, learns and establishes a healthy acoustic baseline for the formula and equipment;

[0010] S2: Acoustic sensor arrays distributed in the negative electrode mixing tank and pipelines are used to collect sound wave and vibration signals at each monitoring point in real time.

[0011] S3: Process the real-time acquired sound wave and vibration signals, extract features, and input the features into a machine learning model trained with negative electrode slurry production process data for diagnosis, and identify the abnormal state, abnormal type and location caused by the characteristics of negative electrode materials.

[0012] S4: Based on the diagnosed abnormality type and location, the decision controller automatically generates customized unblocking instructions and drives the ultrasonic actuators at specific locations to actively unblock the blockage.

[0013] S5: After the unblocking operation is performed, the signal is collected again and the effect is verified. If the verification fails, the event data is automatically recorded and used to optimize the unblocking strategy parameters of the specific negative electrode formula.

[0014] This invention protects a complete intelligent closed-loop control process from "baseline learning" to "effect verification and self-optimization." Its core lies in the first-ever integration of non-invasive acoustic early monitoring, machine learning-based intelligent diagnosis, and targeted proactive unblocking and effect verification into a complete automated closed loop in the negative electrode slurry mixing process. This fundamentally changes the traditional passive, lagging, and manual-dependent processing mode, achieving full-process, online, and intelligent management of the unique clogging problems of negative electrode slurry (such as graphite agglomeration and SBR wall adhesion), significantly improving production continuity, stability, and product quality consistency.

[0015] Preferably, the machine learning model in step S3 can distinguish between abnormal slurry viscosity signals caused by uneven dispersion of graphite or conductive agent and abnormal equipment status signals caused by adhesive adhesion.

[0016] This solution further defines the specific diagnostic capabilities of the machine learning model, significantly improving the accuracy and practicality of diagnosis. By distinguishing between two different root causes of faults—"abnormal slurry viscosity" and "abnormal equipment adhesion"—the system can provide a clearer judgment on the cause of the anomaly. This makes subsequent unblocking or process adjustment commands more precise and effective. For example, for slurry-related problems, process parameters may need to be adjusted, while for equipment adhesion, physical unblocking can be initiated immediately, avoiding misoperation and improving the system's intelligence level and problem-solving efficiency.

[0017] Preferably, the machine learning model in step S3 can also identify the unique acoustic characteristics of uneven slurry drying caused by improper solvent addition rate, and issue a warning for process parameter adjustment.

[0018] This solution endows the machine learning model with early warning capabilities for process parameters, extending the system's functionality from "post-event correction" to "pre-event prevention." By identifying uneven wetting and drying characteristics caused by improper solvent addition, the system can provide early warnings to operators to adjust process parameters such as the feeding rate before the slurry properties deteriorate severely or blockages occur. This demonstrates a higher level of intelligent control, helping to reduce the occurrence of abnormal states from the source and improving the stability and controllability of the production process.

[0019] Preferably, in step S4, the customized unblocking command provides multiple unblocking strategies for the negative electrode slurry characteristics: for the initial adhesion of the adhesive on the tank wall, the ultrasonic actuator is controlled to adopt a high-frequency low-power mode; for the accumulation of graphite slurry in the pipeline, the ultrasonic actuator is controlled to adopt a medium-frequency high-power mode.

[0020] This solution defines differentiated unblocking strategies based on the type of anomaly, enabling precise and optimized delivery and optimization of unblocking energy and methods. For viscous initial adhesion, a "high-frequency, low-power" approach effectively disrupts the surface of the adhesion layer without disturbing the overall slurry. For accumulated sediment, a "medium-frequency, high-power" approach breaks up the sediment with stronger cavitation effects and shear forces. This targeted strategy is more efficient, energy-efficient, and less disruptive to the slurry's condition than traditional single-frequency / power ultrasonic applications.

[0021] Preferably, in step S4, when the decision controller starts the unblocking mode, it communicates with the PLC controller of the negative electrode mixing machine to adjust the mixing speed or the lifting frequency of the mixing blade in conjunction with the ultrasonic waves to destroy the agglomerated or adhered structure.

[0022] This solution incorporates integrated control with the main control system of the slurry mixing machine, achieving a synergistic effect between the "ultrasonic energy field" and the "mechanical agitation flow field," resulting in a "1+1>2" unblocking effect. Mechanical agitation alters the flow state and stress distribution near the slurry or adhered material, allowing ultrasonic energy to be more effectively transmitted and applied to the target area, thereby more thoroughly breaking down agglomerates or peeling off the adhered layer. This synergistic control strategy significantly enhances the effectiveness of single unblocking methods.

[0023] This invention also discloses an intelligent closed-loop control system for implementing any of the methods described above in the negative electrode slurry mixing process, the system comprising:

[0024] Data acquisition module: includes an acoustic sensor group distributed in the negative electrode mixing tank, discharge pipe and valve, used to collect sound wave and vibration signals directly related to the dispersion state of graphite and binder and the adhesion of equipment;

[0025] Signal processing and intelligent diagnosis module: connected to the data acquisition module, with a built-in machine learning algorithm model trained on the negative electrode slurry production process data, used to process signals, extract features, and compare with the pre-stored acoustic baseline of the negative electrode slurry health, and output diagnostic results for the characteristics of the negative electrode material.

[0026] Closed-loop control and execution module: connected to the signal processing and intelligent diagnosis module, including a decision controller and an ultrasonic actuator; the decision controller is used to generate unblocking instructions suitable for the characteristics of the negative electrode slurry adhesion based on the diagnosis results; the ultrasonic actuator is installed on the outer wall of the equipment to receive the unblocking instructions and generate high-frequency mechanical vibration;

[0027] Human-computer interaction and alarm module: used for system status display and alarm prompts.

[0028] This solution protects the hardware and software system architecture for implementing the aforementioned methods. Its advantage lies in providing a highly integrated and modular physical solution. It integrates disparate sensors, intelligent analysis units, actuators, and human-machine interfaces into a cohesive whole, clearly defining the functions and data flow of each module. This systematic design allows the intelligent closed-loop control solution to be easily integrated into existing negative electrode slurry production lines as a standardized functional unit or subsystem, demonstrating good engineering feasibility and promotional value.

[0029] Preferably, the acoustic sensor group is specifically designed to collect early characteristic signals of graphite particle agglomeration and binder adhesion in the negative electrode slurry, and its installation locations include the tank wall below the agitator bearing seat, the lower half of the tank side wall, the discharge valve flange, and the outer wall of the conveying pipeline.

[0030] This scheme specifically defines the preferred installation locations for acoustic sensors, ensuring that the monitoring network can most effectively capture early fault signals. These locations (such as agitator bearing housings, the lower half of the tank, and the outer wall of pipes) are physically sensitive and prone areas for graphite agglomeration, sedimentation, or binder adhesion. Targeted placement optimizes the signal-to-noise ratio, improves monitoring sensitivity and reliability, and lays a solid hardware data foundation for subsequent accurate diagnosis.

[0031] Preferably, the decision controller is configured to select and drive an ultrasonic actuator at the corresponding location based on the diagnosed anomaly type and location. The ultrasonic actuator is installed on the outer wall of the dispersion tank, the outer wall of the critical pipeline, the outer wall of the electromagnetic separator cone, and the screw pump body.

[0032] This solution specifically defines the preferred installation location for the ultrasonic actuators, ensuring that the unblocking energy is precisely delivered to the fault point where it is most needed. By deploying the actuators at corresponding locations (such as tank walls, pipes, and pump bodies) corresponding to the areas monitored by the sensors, a rapid local closed loop of "monitoring-execution" can be formed, minimizing energy attenuation and diffusion within the equipment structure and improving the targeting and effectiveness of unblocking operations.

[0033] Preferably, after the effect verification fails, the signal processing and intelligent diagnosis module records the abnormal signal characteristics, unblocking command parameters and process parameters of this event for the optimization and iteration of the machine learning algorithm model.

[0034] This solution clearly defines the system's self-learning and optimization mechanisms, endowing the system with the ability to continuously evolve. By recording and analyzing data from "verification failure" cases, the system can continuously revise and optimize its diagnostic model (machine learning algorithm) and strategy library (unblocking instruction parameters). This enables the system to better adapt to new challenges brought about by different negative electrode formulations, equipment aging, or process changes. Its intelligence level and application effect will continuously improve with the increase of usage time, possessing long-term life cycle value.

[0035] The present invention also discloses a computer-readable storage medium storing a computer program, which, when executed, performs the method described in any of the above-mentioned embodiments.

[0036] The beneficial effects of this invention are as follows:

[0037] 1. Early and accurate diagnosis: Through a distributed acoustic sensor network and a machine learning model specifically for negative electrode slurry, it can capture early characteristic signals of graphite agglomeration and initial adhesion of binder, so as to achieve early and accurate identification and classification of blockage risks, overcoming the lag problem of traditional pressure gauge monitoring.

[0038] 2. Intelligent and proactive unblocking: Based on the diagnosed abnormality type and location, it automatically matches and executes customized ultrasonic unblocking strategies (such as different frequency power combinations for adhesion and siltation), realizing intelligent closed-loop control from "sensing" to "execution", and the unblocking action is highly targeted and timely.

[0039] 3. Non-invasive and online: All sensors and actuators are installed on the outer wall of the equipment, without the need to modify the existing production process and equipment or shut down for insertion. This achieves true online, real-time, and non-invasive monitoring and maintenance, avoiding the pollution and quality risks caused by manual inspection.

[0040] 4. Collaborative Optimization and Self-Learning: The unblocking command can be linked with the main control system of the mixing machine to form a synergistic effect of mechanical stirring and ultrasonic vibration, improving unblocking efficiency. The system has effect verification and failure case learning functions, which can continuously optimize the diagnostic model and unblocking strategy, and adapt to different negative electrode formulas. Attached Figure Description

[0041] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0042] Figure 1 A schematic diagram showing the specific installation location of the acoustic wave monitoring device;

[0043] Figure 2 This is a schematic diagram showing the installation location of the ultrasonic vibration device. Detailed Implementation

[0044] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of 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.

[0045] The core of this invention lies in achieving early and accurate diagnosis of typical blockage types in the negative electrode slurry mixing process through acoustic monitoring, and intelligently driving an ultrasonic actuator to actively clear blockages based on the diagnostic results, forming an online, real-time, non-invasive intelligent closed-loop control system.

[0046] I. System Hardware Configuration and Installation

[0047] As attached Figure 1 and attached Figure 2 As shown, adhesive acoustic emission sensors are installed at the following locations: 1. at the bottom of the agitator bearing seat at the bottom of the dispersion tank; 2. at the pipe from the dispersion tank outlet to the scraper filter; 3. at the pipe between the electromagnetic separator and the screw pump; and 4. at the pipe between the screw pump and the transfer tank. Ultrasonic cleaners are also installed below the slurry surface in the lower part of the dispersion tank; 5. on the side wall or conical part of the electromagnetic separator's iron discharge chute; 6. on the guide plate below the magnetic system; 7. on the surface of the non-magnetic stainless steel protective plate; 8. at the inlet and outlet of the screw pump; and 9. on the pump body.

[0048] The hardware implementation shown in the diagram mainly includes the installation and integration of two major parts: the acoustic monitoring network and the ultrasonic unblocking network.

[0049] Construction of an acoustic monitoring network:

[0050] To achieve full-process status monitoring of the negative electrode slurry mixing, multiple acoustic sensors (such as high-frequency accelerometers or acoustic emission sensors) are deployed on key equipment and pipelines of the slurry mixing system, forming a distributed monitoring network. Preferred installation locations include:

[0051] Dispersion tank area: the tank wall below the agitator bearing housing at the bottom of the dispersion tank ( Figure 1 A sensor is installed at position 1 to monitor the operating status of the agitator and the initial characteristics of graphite agglomeration or SBR wall adhesion near the agitator blades; a sensor is installed on the side wall below the slurry surface in the lower part of the dispersion tank. Figure 1 (Position 5) Install a sensor to capture acoustic signals generated by viscosity changes and agglomeration within the slurry body.

[0052] Piping area: Piping from the dispersion tank outlet to the scraper filter ( Figure 1 Location 2), the pipeline between the electromagnetic separator and the screw pump ( Figure 1 (Location 3), and the pipeline between the screw pump and the transfer tank ( Figure 1 Sensors are installed on the outer wall at location 4). These locations are key points where slurry flow rate changes or sedimentation and accumulation are likely to occur, and can effectively monitor abnormalities in the flow state of graphite slurry within the pipeline.

[0053] Electromagnetic separator area: the side wall or conical hopper section of the electromagnetic separator's unloading chute ( Figure 1 Position 6), the guide plate below the magnetic system ( Figure 1 (Position 7) and non-magnetic stainless steel guard plate surface ( Figure 1 (Location 8) Install a sensor to monitor the adhesion of the binder in the slurry in this area.

[0054] All sensors are industrial-grade adhesive mounts to ensure good coupling with the equipment wall and are moisture-proof and corrosion-resistant to adapt to the working environment of slurry preparation.

[0055] Construction of an ultrasonic unblocking network:

[0056] Ultrasonic actuators (or ultrasonic cleaners) are installed at appropriate locations on the outer wall of equipment or pipes, depending on the target of the unblocking process. (See attached image.) Figure 2 As shown, preferred installation locations include:

[0057] The outer wall of the dispersion tank corresponds to the slurry area in the lower part of the tank and is used to break the SBR binder adhesion layer on the tank wall and destroy the graphite agglomeration structure in the near-wall area.

[0058] Key pipe outer walls: especially elbows, valves, and easily clogged pipe sections detected by the aforementioned acoustic sensors, used to break up the graphite slurry accumulated inside the pipe.

[0059] The outer wall of the conical hopper of the electromagnetic separator and the pump body of the screw pump are designed to prevent slurry from adhering and accumulating inside these components.

[0060] II. System Software Flow and Closed-Loop Control Method

[0061] The intelligent closed-loop control method of this system is implemented according to the following steps:

[0062] S1: System Initialization and Healthy Acoustic Baseline Learning

[0063] When the mixing machine begins producing a new negative electrode formulation (such as a specific ratio of graphite, SP, CMC, SBR, and deionized water), it first runs several batches under known process parameters (such as stirring speed and solvent addition rate). During this period, the data acquisition module continuously collects acoustic and vibration signals from each monitoring point. The signal processing and intelligent diagnostic module learns from these signals, extracts the acoustic characteristics of the formulation and equipment in a "healthy" state (such as energy values ​​in specific frequency bands, signal amplitude statistical characteristics, spectral envelope, etc.), and establishes and stores them as the "healthy acoustic baseline" for the formulation.

[0064] S2: Real-time data acquisition

[0065] During normal production, the distributed acoustic sensor array continuously and synchronously collects sound wave and vibration signals from each monitoring point and transmits the real-time data stream to the signal processing and intelligent diagnostic module.

[0066] S3: Intelligent Signal Diagnosis and Anomaly Detection

[0067] The signal processing and intelligent diagnostic module preprocesses real-time signals (e.g., filtering, noise reduction) and extracts features. These features are then input into a machine learning algorithm model (e.g., a deep convolutional neural network or support vector machine model) pre-trained with a large amount of negative electrode slurry production process data. This model compares and analyzes the real-time features against a pre-stored "healthy acoustic baseline," not only determining the presence of anomalies but also further distinguishing the types of anomalies, specifically including:

[0068] Identify the unique acoustic characteristics of slurry viscosity abnormalities caused by uneven dispersion of graphite conductive agents (such as SP).

[0069] Identify specific vibration signals generated by the initial adhesion of SBR / CMC adhesive to the equipment wall.

[0070] As described in claim 7, the model can also identify the unique acoustic characteristics of uneven slurry drying caused by improper deionized water addition rate, and issue early warnings for process parameter adjustment to operators at this stage, achieving more advanced control.

[0071] S4: Intelligent Decision-Making and Proactive Execution

[0072] The decision controller in the closed-loop control and execution module automatically generates customized unblocking instructions based on the diagnostic results of step S3 (including anomaly type, location, and severity). The instructions are optimized for the negative electrode slurry characteristics, for example:

[0073] When initial adhesion of SBR binder is detected on the tank wall, the decision controller drives the ultrasonic actuator at the corresponding position to operate in a high-frequency, low-power mode (e.g., 40kHz, 200W). This mode can effectively break the adhesion of the sticky substance without excessively impacting the slurry.

[0074] When graphite slurry buildup is diagnosed in the pipe, a medium-frequency high-power mode (e.g., 28kHz, 500W) is used for unblocking to provide stronger energy to break up the buildup.

[0075] Meanwhile, the decision controller is linked with the PLC controller of the mixing machine through an industrial communication interface (such as Profinet, EtherCAT). When the ultrasonic cleaning is started, the mixing speed of the mixing machine or the lifting frequency of the mixing paddle is adjusted appropriately to form a synergistic effect of mechanical mixing and ultrasonic vibration, which more effectively destroys the agglomerated or adhered structure.

[0076] S5: Effect Verification and System Self-Optimization

[0077] After a clearing command is executed, the system does not immediately close the closed loop. Instead, it restarts the acoustic monitoring network (repeating steps S2-S3) to collect equipment status signals after clearing and verify the effect. If the signal characteristics return to the normal range, the clearing is considered successful, and production continues. If the verification fails, it indicates that the clearing strategy did not completely solve the problem. The system will automatically record all relevant data for this event (including abnormal signal characteristics, the clearing command parameters used, and the process parameters at the time). This data will be used to optimize the clearing strategy parameters for this specific negative electrode formula, enabling continuous self-learning and optimization of the system.

[0078] III. Human-computer interaction and alarm

[0079] The human-machine interaction and alarm module provides a graphical interface that displays the status of each monitoring point, diagnostic results, unblocking operation records, and system alarm information in real time. When an anomaly is diagnosed or unblocking verification fails, the module will trigger an on-site audible and visual alarm and push detailed information to the upper-level production management system (MES) to remind staff to intervene.

[0080] The prior description of this disclosure is provided to enable any person skilled in the art to make or use this disclosure. Various modifications to this disclosure will be apparent to those skilled in the art, and the general principles defined herein may be applied to other variations without departing from the spirit or scope of this disclosure. Therefore, this disclosure is not intended to be limited to the examples and designs described herein, but should be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0081] Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for intelligent closed-loop control of the negative electrode slurry mixing process, characterized in that, The method, applied in the preparation of negative electrode slurry for lithium-ion batteries, includes the following steps: S1: System initialization, when the negative electrode mixing machine is running standard formula slurry, learns and establishes a healthy acoustic baseline for the formula and equipment; S2: Acoustic sensor arrays distributed in the negative electrode mixing tank and pipelines are used to collect sound wave and vibration signals at each monitoring point in real time. S3: Process the real-time acquired sound wave and vibration signals, extract features, and input the features into a machine learning model trained with negative electrode slurry production process data for diagnosis, and identify the abnormal state, abnormal type and location caused by the characteristics of negative electrode materials. S4: Based on the diagnosed abnormality type and location, the decision controller automatically generates customized unblocking instructions and drives the ultrasonic actuators at specific locations to actively unblock the blockage. S5: After the unblocking operation is performed, the signal is collected again and the effect is verified. If the verification fails, the event data is automatically recorded and used to optimize the unblocking strategy parameters of the specific negative electrode formula.

2. The intelligent closed-loop control method according to claim 1, characterized in that: The machine learning model in step S3 can distinguish between abnormal slurry viscosity signals caused by uneven dispersion of graphite or conductive agent and abnormal equipment status signals caused by adhesive adhesion.

3. The intelligent closed-loop control method according to claim 1, characterized in that: The machine learning model in step S3 can also identify the unique acoustic characteristics of uneven slurry drying caused by improper solvent addition rate and issue a warning for process parameter adjustment.

4. The intelligent closed-loop control method according to claim 1, characterized in that, In step S4, the customized unblocking command provides a variety of unblocking strategies for the negative electrode slurry characteristics: for the initial adhesion of the adhesive on the tank wall, the ultrasonic actuator is controlled to adopt a high frequency and low power mode; for the accumulation of graphite slurry in the pipeline, the ultrasonic actuator is controlled to adopt a medium frequency and high power mode.

5. The intelligent closed-loop control method according to claim 1, characterized in that: In step S4, when the decision controller starts the unblocking mode, it communicates with the PLC controller of the negative electrode mixing machine to adjust the mixing speed or the stirring blade lifting frequency of the mixing machine in conjunction with the ultrasonic waves to destroy the agglomerated or adhered structure.

6. A smart closed-loop control system for implementing the method of any one of claims 1-5 in the negative electrode slurry mixing process, characterized in that, The system includes: Data acquisition module: includes an acoustic sensor group distributed in the negative electrode mixing tank, discharge pipe and valve, used to collect sound wave and vibration signals directly related to the dispersion state of graphite and binder and the adhesion of equipment; Signal processing and intelligent diagnosis module: connected to the data acquisition module, with a built-in machine learning algorithm model trained on the negative electrode slurry production process data, used to process signals, extract features, and compare with the pre-stored acoustic baseline of the negative electrode slurry health, and output diagnostic results for the characteristics of the negative electrode material. Closed-loop control and execution module: connected to the signal processing and intelligent diagnosis module, including a decision controller and an ultrasonic actuator; the decision controller is used to generate unblocking instructions suitable for the characteristics of the negative electrode slurry adhesion based on the diagnosis results; the ultrasonic actuator is installed on the outer wall of the equipment to receive the unblocking instructions and generate high-frequency mechanical vibration; Human-computer interaction and alarm module: used for system status display and alarm prompts.

7. The system according to claim 6, characterized in that: The acoustic sensor array is specifically designed to collect early characteristic signals of graphite particle agglomeration and binder adhesion to the wall in the negative electrode slurry. Its installation locations include the tank wall below the agitator bearing seat, the lower half of the tank side wall, the discharge valve flange, and the outer wall of the conveying pipeline.

8. The system according to claim 6, characterized in that: The decision controller is configured to select and drive an ultrasonic actuator at the corresponding location based on the diagnosed anomaly type and location. The ultrasonic actuator is installed on the outer wall of the dispersion tank, the outer wall of the critical pipeline, the outer wall of the electromagnetic separator cone, and the screw pump body.

9. The system according to claim 6, characterized in that: After the effect verification fails, the signal processing and intelligent diagnosis module records the abnormal signal characteristics, unblocking command parameters and process parameters of this event for the optimization and iteration of the machine learning algorithm model.

10. A computer-readable storage medium, characterized in that: The medium contains a computer program, which, when run, performs the method as described in any one of claims 1 to 5.

Images