Intelligent control device and method for wet lithium battery diaphragm white oil extraction RO water treatment

Abstract

The application discloses a wet-process lithium battery diaphragm white oil extraction RO water treatment intelligent regulation and control device and method, and belongs to the technical field of lithium battery diaphragm production equipment. By adding a temperature sensor, a visual window with a layered interface scale, an industrial camera and an edge computing gateway with a built-in multi-AI model to the original liquid tank, online accurate calculation of the ion liquid concentration and full-automatic closed-loop regulation and control of the RO system are realized. The application applies AI visual recognition and deep learning of a dissolution curve to concentration detection in wet-process diaphragm production. Through automatic identification of a layered interface by a first AI model, real-time interpretation of a special dissolution curve by a second AI model and autonomous output of a regulation and control pressure by a third AI model, the automation level and regulation and control precision of the device are improved. Meanwhile, the device is compatible with the original equipment architecture, has an incremental learning ability and can adapt to working condition changes, thereby providing a feasible intelligent solution for efficient and stable operation of diaphragm production.

Description

Technical Field

[0001] This invention relates to the field of lithium battery separator production equipment technology, and more specifically to an intelligent control device and method for wet-process lithium battery separator white oil extraction RO water treatment. Background Technology

[0002] Currently, the integrated wet-process lithium battery separator white oil extraction, drying, and separation equipment based on invention patent CN119174929A has achieved full automation of the extraction, washing, drying, and separation process. Its accompanying RO water treatment device collects the overflow liquid (ionic liquid + water mixture) from the washing process through a first and second stock solution tank, and uses a magnetic level gauge to obtain the total liquid level of the mixture, providing basic data for subsequent concentration treatment. The ionic liquid extractant used in this equipment has a specific water-extractant dissolution curve. The concentration of the ionic liquid in the stock solution fluctuates non-linearly with temperature, and the mixture exhibits a stratified state of water and miscible layers.

[0003] However, existing technologies lack online identification methods for the interface height between the aqueous and miscible layers, making it impossible to accurately obtain the location of the layer interface and thus preventing the independent calculation of the miscible layer volume. Furthermore, the nonlinear characteristics of the ionic liquid dissolution curve require manual decomposition into formulas or table lookups, hindering real-time mapping of the temperature-concentration relationship. These deficiencies mean that the concentration of the ionic liquid in the stock solution can only be obtained through manual sampling and analysis, resulting in significant lag and failing to meet real-time control requirements. Ultimately, this leads to lag in RO high-pressure pump parameter adjustments, large fluctuations in the ionic liquid concentration of the concentrate, frequent manual intervention, and difficulty in achieving precise recovery and reuse of the ionic liquid extractant, severely restricting the improvement of equipment automation and intelligence.

[0004] Therefore, how to propose an intelligent control device and method for wet lithium battery separator white oil extraction RO water treatment, realize online accurate calculation of the concentration of raw liquid ionic liquid, automatically control RO system parameters, and stably control the concentrate at the target concentration, is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0005] In view of this, the present invention provides an intelligent control device and method for wet-process lithium battery separator white oil extraction and RO water treatment. Based on the integrated wet-process lithium battery separator white oil extraction, drying and separation equipment of invention patent CN119174929A, the device reuses the magnetic float level gauge built into the raw liquid tank of its RO water treatment device to obtain the total liquid level height. Combined with the dissolution curve characteristics of the ionic liquid extractant, the device uses an industrial camera to collect the layer interface scale and an AI200 gateway to integrate multiple AI models in a collaborative architecture to achieve online accurate calculation of the raw liquid ionic liquid concentration. The entire process does not require manual setting of the dissolution curve formula / lookup table rules.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: On one hand, this invention discloses an intelligent control device for wet-process lithium battery separator white oil extraction and RO water treatment, which is applied to the integrated wet-process lithium battery separator white oil extraction, drying, and separation equipment disclosed in invention patent CN119174929A. The device includes a first raw liquid tank, a second raw liquid tank, and a high-pressure water pump. The first and second raw liquid tanks are equipped with features for obtaining the total liquid level H of the mixture. 总 The magnetic level gauge, the device further includes: A temperature sensor, located inside the stock solution tank, is used to collect the temperature T of the mixed solution; A visualization window is located on the side of the stock solution tank, marked with a layer interface scale to indicate the location of the interface between the water layer and the ionic liquid interphase layer. An industrial camera is positioned in front of the visualization window to capture images of the layered interface scale. The edge computing gateway is communicatively connected to the PLC controllers of the magnetic level gauge, temperature sensor, industrial camera, and water purification high-pressure pump, respectively, and is used to calculate the total liquid level H of the mixed liquid. 总 The concentration of the original liquid ionic liquid is calculated based on the temperature T of the mixed solution and the scale image of the layered interface. The set pressure of the water purification high-pressure pump is output according to the concentration of the original liquid ionic liquid to stabilize the concentration of the concentrate at the target value.

[0007] Preferably, the edge computing gateway includes a first AI model, a second AI model, and a third AI model; The first AI model identifies the layered interface scale based on the layered interface scale image captured by the industrial camera, outputs the water layer height h2, and derives the intersoluble layer height h. 互 =H 总 -h2; The second AI model outputs the mass percentage ω(T) of the ionic liquid in the miscible solution at the current temperature based on the temperature T of the mixture. The third AI model determines the concentration factor based on the ratio of the original ionic liquid concentration to the target concentration, and outputs the corresponding set pressure of the water purification high-pressure pump.

[0008] Preferably, the first AI model is a U-Net deep learning visual recognition model; The second AI model is a lightweight neural network MLP model, which is pre-trained using water and extractant dissolution curve data specific to ionic liquid extractants as the training set; The third AI model is the XGBoost control model, which is pre-trained based on training samples covering the operating conditions and is used to establish the mapping relationship between the concentration factor and the set pressure of the water purification high-pressure pump.

[0009] Preferably, the edge computing gateway further includes a concentration calculation module, used to calculate the concentration based on the intersolid layer height h. 互 The concentration of the original ionic liquid is calculated from the mass percentage ω(T) of the ionic liquid in the miscible solution. The formula is as follows: =[ω(T)×h 互 ] / H 总 ×100%.

[0010] Preferably, the edge computing gateway further includes a dual-box cross-validation module, used to calculate the detection error of the concentration of the raw liquid ionic liquid in the first raw liquid box and the second raw liquid box, and when the detection error exceeds a preset threshold, to correct the set pressure of the water purification high-pressure pump with the average concentration of the raw liquid ionic liquid in the two boxes.

[0011] Preferably, the edge computing gateway further includes an incremental learning module, which is used to automatically update the parameters of the first AI model, the second AI model, and the third AI model after accumulating a predetermined amount of valid data.

[0012] On the other hand, the present invention also discloses an intelligent control method for RO water treatment in wet-process lithium battery separator white oil extraction, applied to the aforementioned intelligent control device for RO water treatment in wet-process lithium battery separator white oil extraction, comprising the following steps: S1. Obtain images of the total liquid level, temperature, and layer interface of the mixture; S2. Using the first AI model, the layer interface scale is identified based on the layer interface scale image to determine the water layer height, and the intersolution layer height is determined in combination with the total liquid level height of the mixed liquid. S3. Using the second AI model, determine the mass percentage of ionic liquid in the mixed solution at the current temperature based on the temperature of the mixed solution; S4. Determine the concentration of the original ionic liquid based on the mass percentage of the ionic liquid in the intermixed solution, the height of the intermixed solution layer, and the total liquid level of the mixture; S5. Using the third AI model, determine the concentration factor based on the concentration of the original ionic liquid and the target concentration, and determine the corresponding set pressure of the water purification high-pressure pump based on the concentration factor, and output the set pressure command; S6. Issue the set pressure command to regulate the operation of the RO system and stabilize the concentration of the concentrate at the target concentration.

[0013] Preferably, a method for intelligent control of RO water treatment in wet lithium-ion battery separator white oil extraction also includes a dual-compartment cross-validation step: The total liquid level, temperature, and layer interface scale images of the mixture in the first and second stock solution tanks were obtained respectively, and the concentration of the stock solution ionic liquid in the two tanks was determined respectively. The detection error of the concentration of the original ionic liquid in the two boxes is calculated. When the detection error exceeds a preset threshold, the set pressure of the water purification high-pressure pump is corrected by the average value of the concentrations in the two boxes.

[0014] Preferably, a method for intelligent control of RO water treatment in wet lithium-ion battery separator white oil extraction also includes an incremental learning step: After accumulating a predetermined amount of valid data, the parameters of the first AI model, the second AI model, and the third AI model are automatically updated.

[0015] As can be seen from the above technical solution, the present invention discloses an intelligent control device and method for wet-process lithium battery separator white oil extraction RO water treatment, which has the following advantages compared with the prior art: 1. Low modification cost and strong compatibility: Based on the invention patent CN119174929A, the device is upgraded and reuses the original magnetic float level gauge of the raw liquid tank to obtain the total height. Only a camera, thermometer and AI200 gateway are added. There is no need to change the core architecture of the original equipment. The hardware division of labor is clear, and the accuracy of raw liquid concentration calculation is greatly improved (prediction error MAE≤0.2%). It completely replaces the traditional manual sampling and testing method, eliminates sampling lag, and matches the precise use requirements of ionic liquid extractants. 2. High detection accuracy and strong stability: The U-Net visual recognition model is only for the recognition of layered interface scales, which simplifies the visual detection logic. It has strong resistance to interference from industrial scenarios such as oil stains and light fluctuations. The error of layered interface scale recognition is ≤0.1cm. Combined with the interpretation of the exclusive dissolution curve of ionic liquid extractant, it ensures the stability and specificity of the stock solution concentration calculation. 3. Precise control and high recovery efficiency: The AI200 gateway directly learns the nonlinear law of the dissolution curve corresponding to the patent of ionic liquid extractant, eliminating the need for manual formula fitting / table lookup and completely eliminating manual fitting errors. Combined with the accurate calculation results of the original solution concentration, it automatically adjusts the parameters of the RO high-pressure pump to stably control the concentration of ionic liquid in the concentrate at 50%±0.5%, improving the recovery and reuse efficiency of ionic liquid extractant and improving product quality consistency by more than 20%. 4. High degree of automation and improved production efficiency: The entire process is automated, eliminating manual sampling, manual curve fitting, and manual parameter adjustment. It can run continuously for 72 hours without accuracy loss, improving production efficiency by more than 30%, significantly reducing labor costs, and realizing the automation and intelligent upgrade of the equipment with invention patent CN119174929A. 5. Strong adaptability and stable long-term operation: The incremental learning mechanism can adaptively update all AI models in the gateway, adapting to the complex industrial scenarios of lithium battery separator production and the usage characteristics of ionic liquid extractants. The accuracy does not decrease over long-term operation, and the communication protocol is compatible with the industrial general standard and the original linkage logic of the device with invention patent CN119174929A. After the upgrade, the device has strong compatibility and adaptability. Attached Figure Description

[0016] 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 only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0017] Figure 1 This is a schematic diagram of an RO water treatment device; Figure 2 This is a schematic diagram of the dissolution curves of water and extractant. Detailed Implementation

[0018] 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.

[0019] This embodiment is based on the integrated wet-process lithium battery separator white oil extraction, drying, and separation equipment disclosed in invention patent CN119174929 A (see the specific implementation details in the patent specification), and is improved upon its RO water treatment device. Specifically, as follows: Figure 1 As shown, the RO water treatment device 9 includes a first raw liquid tank 901, a second raw liquid tank 902, a carbon filter tank 904, a pleated filter 905, an ultrafiltration device 906, a TOC degrader 907, an ultrafiltration water storage tank 908, a high-pressure water pump 909, an RO water purifier 9010, and a purified water storage tank 9011, which are connected in sequence by pipes. A raw liquid booster pump 903 is connected between the second raw liquid tank 902 and the carbon filter tank 904. The purified water storage tank 9011 is connected to a water washing device through a purified water transfer pump 9012. The top of the carbon filter tank 904, the ultrafiltration device 906, and the RO water purifier 9010 are also connected to a concentrated liquid storage tank 9013 by pipes. The concentrated liquid storage tank 9013 is connected to a short evaporation device through a concentrated liquid transfer pump 9014.

[0020] Based on this, this invention discloses an intelligent control device for RO water treatment in wet-process lithium battery separator white oil extraction. It utilizes the RO water treatment device based on invention patent CN119174929 A. The raw liquid tank is equipped with a magnetic level gauge (measuring range 0-100cm, accuracy ±0.1cm), which is connected to a PLC via an RS485 interface. The edge computing gateway (AI200 gateway) communicates with the PLC via its Modbus TCP communication port to control the total liquid level H. 总 The data upload configuration requires no modification to the original level gauge circuitry; Thermometer Installation: Platinum resistance thermometers (range -20℃ to 55℃, accuracy ±0.1℃, corrosion-resistant material) are installed in the first stock solution tank 901 and the second stock solution tank 902. They communicate with the AI200 edge computing gateway via the Modbus RTU protocol to upload the mixed solution temperature T in real time, providing data support for temperature compensation of the ionic liquid extractant dissolution curve. The installation location avoids areas of fluid disturbance within the stock solution tanks to ensure accurate temperature detection. Camera and window installation: A high borosilicate glass visualization window (10mm thick, pressure ≥0.6MPa) is opened on the side of the raw liquid tank of the device in invention patent CN119174929 A, in an unobstructed area with a good field of view. The layer interface scale is marked (the baseline is the bottom of the window, 0-100cm, accuracy 0.1cm) to indicate the interface position between the water layer and the ionic liquid intermixed layer. An industrial camera (1920×1080 resolution, 30fps frame rate, IP67 protection, with oil-proof lens) is installed on the outside of the window and fixed by a bracket. The lens is vertically aligned with the layer interface scale area of ​​the visualization window (to ensure clear imaging of the scale) to capture the layer interface scale image. The image is connected to the Ethernet port of the AI200 gateway via a network cable. Gateway and PLC linkage: The AI200 gateway is deployed in the RO control cabinet of the device with invention patent CN119174929 A. It establishes communication with the original RO water purification high pressure pump PLC controller (Siemens S7-1200) of the device through the Modbus TCP protocol to realize the real-time transmission of pressure parameters without modifying the original PLC control program.

[0021] The AI200 gateway communicates with the PLC controllers of the magnetic level gauge, temperature sensor, industrial camera, and high-pressure water pump, respectively, to determine the total liquid level H of the mixed liquid. 总 The concentration of the original liquid ionic liquid is calculated based on the temperature T of the mixed solution and the scale image of the layered interface. The set pressure of the water purification high-pressure pump is output according to the concentration of the original liquid ionic liquid to stabilize the concentration of the concentrate at the target value.

[0022] The first AI model (U-Net visual model), the second AI model (AI curve interpretation model (MLP)), and the third AI model (pressure regulation model (XGBoost)) are quantized into INT8 format (quantization error ≤ ±2%) and deployed to the AI200 gateway via SDK. The sampling frequency of all detection devices is preset to 1Hz, which matches the detection frequency of the original devices.

[0023] The first AI model identifies the layered interface scale based on images captured by an industrial camera, outputs the water layer height h2, and derives the intersoluble layer height h. 互 =H 总 -h2.

[0024] The second AI model outputs the mass percentage ω(T) of the ionic liquid in the miscible solution at the current temperature, based on the mixed solution temperature T. Specifically: The "water-extractant solubility curve" corresponding to the ionic liquid extractant (reference) Figure 2 The curve corresponds to temperature on the horizontal axis and the mass fraction of water in the miscible solution on the vertical axis. Complete data (temperature-solubility samples covering the entire temperature range from -20℃ to 55℃) is used as the training set, which can accurately match the solubility characteristics of the extractant without the need for manual fitting of formulas / extraction of discrete points. The second AI model uses a lightweight neural network model (MLP) deployed in the AI200 gateway. The input is "temperature T" and the output is "the mass percentage of ionic liquid in the mixed solution ω(T)". The training objective is that the error between the output value and the true value of the dissolution curve of the ionic liquid extractant is ≤0.1%. Any measured temperature T can be input and the corresponding ω(T) can be output directly. There are no manually set formulas / table lookup rules throughout the process.

[0025] The edge computing gateway also includes a concentration calculation module for calculating the concentration based on the intercomponent solution layer height h. 互 The concentration of the original ionic liquid is calculated from the mass percentage ω(T) of the ionic liquid in the miscible solution. The formula is as follows: =[ω(T)×h 互 ] / H 总 ×100%.

[0026] The third AI model determines the concentration factor K (K=C) based on the ratio of the original ionic liquid concentration to the target concentration. 目标 / C 原液 ), outputting the corresponding set pressure of the high-pressure water pump.

[0027] The third AI model is the XGBoost control model, deployed in the AI200 gateway. In this embodiment, based on 100,000 training samples (covering 2%-10% stock solution concentration, -20℃ to 55℃ temperature, adapting to the operating range of the device in invention patent CN119174929 A), a mapping relationship of "concentration factor K-high pressure pump pressure P" is established, and the pressure value is output (control error ±0.05MPa).

[0028] Preferably, the edge computing gateway also includes a dual-box cross-validation module, which is used to calculate the detection error of the concentration of the raw liquid ionic liquid in the first raw liquid box and the second raw liquid box, and when the detection error exceeds a preset threshold (0.3%), the set pressure of the water purification high-pressure pump is corrected by the average value of the concentration of the raw liquid ionic liquid in the two boxes.

[0029] Preferably, the edge computing gateway also includes an incremental learning module, which is used to automatically update the parameters of the first AI model, the second AI model and the third AI model after accumulating a predetermined number (1000) of valid data (concentration prediction error ≤1%), to adapt to the long-term operating condition drift of the device.

[0030] The fully automatic control process is as follows: 1. Basic data acquisition: The magnetic float level gauge in the raw liquid tank collects the total liquid level height H. 总 The thermometer collects T data and uploads it to the AI200 gateway, while the industrial camera collects the layered interface scale image and transmits it to the AI200 gateway. 2. Data Processing: The AI200 gateway uses the U-Net model to identify the layer interface scale, obtains the water layer height h2, and derives the intersoluble layer height h. 互 The AI ​​curve interpretation model transforms T into ω(T) that matches the properties of the ionic liquid extractant. 3. Concentration Calculation: The AI200 gateway accurately calculates the concentration of the stock solution ionic liquid (Cstock solution) using the formula. 4. Control parameter output: The AI200 gateway outputs the set pressure to the RO water purification high-pressure pump PLC controller of the device in invention patent CN119174929 A through Modbus TCP protocol according to the concentration ratio K; 5. Dual-box feedback calibration: The total H, T, and layer interface image data of the first and second stock solution tanks are synchronously uploaded to the AI200 gateway. The C stock solution of the two tanks is calculated separately. When the concentration detection error between the two tanks is >0.3%, the pressure parameter is corrected by the average value. 6. Closed-loop stable output: After the RO water purifier is running, the concentrated liquid concentration data is fed back to the AI200 gateway. If the deviation from the target concentration of 50% is greater than 0.5%, the gateway dynamically fine-tunes the high-pressure pump pressure parameters to ensure the stability of the ionic liquid concentration.

[0031] On the other hand, the present invention also discloses an intelligent control method for RO water treatment in wet-process lithium battery separator white oil extraction, applied to the aforementioned intelligent control device for RO water treatment in wet-process lithium battery separator white oil extraction, comprising the following steps: S1. Obtain images of the total liquid level, temperature, and layer interface of the mixture; S2. Using the first AI model, the layer interface scale is identified based on the layer interface scale image to determine the water layer height, and the intersol layer height is determined in combination with the total liquid level height of the mixed liquid. S3. Using the second AI model, determine the mass percentage of ionic liquid in the mixed solution at the current temperature based on the temperature of the mixed solution; S4. Determine the concentration of the original ionic liquid based on the mass percentage of the ionic liquid in the miscible solution, the height of the miscible solution layer, and the total liquid level of the mixture; S5. Using the third AI model, determine the concentration factor based on the concentration of the original ionic liquid and the target concentration, and determine the corresponding set pressure of the water purification high-pressure pump based on the concentration factor, and output the set pressure command; S6. Issue a pressure setting command to regulate the operation of the RO system and stabilize the concentration of the concentrate at the target concentration.

[0032] Preferably, a method for intelligent control of RO water treatment in wet lithium-ion battery separator white oil extraction also includes a dual-compartment cross-validation step: The total liquid level, temperature, and layer interface scale images of the mixture in the first and second stock solution tanks were obtained respectively, and the concentration of the stock solution ionic liquid in the two tanks was determined respectively. The detection error of the concentration of the original ionic liquid in the two boxes is calculated. When the detection error exceeds the preset threshold, the set pressure of the water purification high-pressure pump is corrected by the average value of the concentrations in the two boxes.

[0033] Preferably, a method for intelligent control of RO water treatment in wet lithium-ion battery separator white oil extraction also includes an incremental learning step: After accumulating a predetermined amount of valid data, the parameters of the first AI model, the second AI model, and the third AI model are automatically updated.

[0034] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to the method section.

[0035] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A smart control device for wet-process lithium-ion battery membrane white oil extraction and RO water treatment, applied to an integrated wet-process lithium-ion battery membrane white oil extraction, drying, and separation equipment including a first raw liquid tank, a second raw liquid tank, and a high-pressure water pump, wherein the first and second raw liquid tanks are equipped with devices for obtaining the total liquid level H of the mixture. 总 The magnetic float level gauge is characterized by, The device further includes: A temperature sensor, located inside the stock solution tank, is used to collect the temperature T of the mixed solution; A visualization window is located on the side of the stock solution tank, marked with a layer interface scale to indicate the location of the interface between the water layer and the ionic liquid interphase layer. An industrial camera is positioned in front of the visualization window to capture images of the layered interface scale. The edge computing gateway is communicatively connected to the PLC controllers of the magnetic level gauge, temperature sensor, industrial camera, and water purification high-pressure pump, respectively, and is used to calculate the total liquid level H of the mixed liquid. 总 The concentration of the original liquid ionic liquid is calculated based on the temperature T of the mixed solution and the scale image of the layered interface. The set pressure of the water purification high-pressure pump is output according to the concentration of the original liquid ionic liquid to stabilize the concentration of the concentrate at the target value.

2. The intelligent control device for wet-process lithium battery separator white oil extraction RO water treatment according to claim 1, characterized in that, The edge computing gateway includes a first AI model, a second AI model, and a third AI model; The first AI model identifies the layered interface scale based on the layered interface scale image captured by the industrial camera, outputs the water layer height h2, and derives the intersoluble layer height h. 互 =H 总 -h2; The second AI model outputs the mass percentage ω(T) of the ionic liquid in the miscible solution at the current temperature based on the temperature T of the mixture. The third AI model determines the concentration factor based on the ratio of the original ionic liquid concentration to the target concentration, and outputs the corresponding set pressure of the water purification high-pressure pump.

3. The intelligent control device for wet-process lithium battery separator white oil extraction RO water treatment according to claim 2, characterized in that, The first AI model is the U-Net deep learning visual recognition model; The second AI model is a lightweight neural network MLP model, which is pre-trained using water and extractant dissolution curve data specific to ionic liquid extractants as the training set; The third AI model is the XGBoost control model, which is pre-trained based on training samples covering the operating conditions and is used to establish the mapping relationship between the concentration factor and the set pressure of the water purification high-pressure pump.

4. The intelligent control device for wet-process lithium battery separator white oil extraction RO water treatment according to claim 2, characterized in that, The edge computing gateway also includes a concentration calculation module, used to calculate the concentration based on the intercomponent solution layer height h. 互 The concentration of the original ionic liquid is calculated from the mass percentage ω(T) of the ionic liquid in the miscible solution. The formula is as follows: =[ω(T)×h 互 ] / H 总 ×100%。 5. The intelligent control device for wet-process lithium battery separator white oil extraction RO water treatment according to claim 4, characterized in that, The edge computing gateway also includes a dual-box cross-validation module, which is used to calculate the detection error of the concentration of the raw liquid ionic liquid in the first raw liquid box and the second raw liquid box, and when the detection error exceeds a preset threshold, the set pressure of the water purification high-pressure pump is corrected by the average value of the concentration of the raw liquid ionic liquid in the two boxes.

6. The intelligent control device for wet-process lithium battery separator white oil extraction RO water treatment according to claim 2, characterized in that, The edge computing gateway also includes an incremental learning module, which is used to automatically update the parameters of the first AI model, the second AI model, and the third AI model after accumulating a predetermined amount of valid data.

7. A method for intelligent control of RO water treatment in wet-process lithium-ion battery membrane white oil extraction, applied to an intelligent control device for RO water treatment in wet-process lithium-ion battery membrane white oil extraction as described in any one of claims 1-6, characterized in that, Includes the following steps: S1. Obtain images of the total liquid level, temperature, and layer interface of the mixture; S2. Using the first AI model, the layer interface scale is identified based on the layer interface scale image to determine the water layer height, and the intersolution layer height is determined in combination with the total liquid level height of the mixed liquid. S3. Using the second AI model, determine the mass percentage of ionic liquid in the mixed solution at the current temperature based on the temperature of the mixed solution; S4. Determine the concentration of the original ionic liquid based on the mass percentage of the ionic liquid in the intermixed solution, the height of the intermixed solution layer, and the total liquid level of the mixture; S5. Using the third AI model, determine the concentration factor based on the concentration of the original ionic liquid and the target concentration, and determine the corresponding set pressure of the water purification high-pressure pump based on the concentration factor, and output the set pressure command; S6. Issue the set pressure command to regulate the operation of the RO system and stabilize the concentration of the concentrate at the target concentration.

8. The intelligent control method for RO water treatment in wet-process lithium battery separator white oil extraction according to claim 7, characterized in that, It also includes a two-box cross-validation step: The total liquid level, temperature, and layer interface scale images of the mixture in the first and second stock solution tanks were obtained respectively, and the concentration of the stock solution ionic liquid in the two tanks was determined respectively. The detection error of the concentration of the original ionic liquid in the two boxes is calculated. When the detection error exceeds a preset threshold, the set pressure of the water purification high-pressure pump is corrected by the average value of the concentrations in the two boxes.

9. The intelligent control method for RO water treatment in wet-process lithium battery separator white oil extraction according to claim 7, characterized in that, It also includes incremental learning steps: After accumulating a predetermined amount of valid data, the parameters of the first AI model, the second AI model, and the third AI model are automatically updated.

Images