A method for wet leaching reaction in lithium battery recycling

Abstract

The application provides a lithium battery recycling wet leaching reaction method, which accurately collects pH and temperature data through a sensor array, eliminates noise by combining data fusion technology to ensure data reliability, analyzes fluctuation patterns and predicts dissolution rate by using support vector machine and neural network algorithm, dynamically adjusts reaction parameters, stabilizes pH and temperature by using proportional-integral-derivative controller and heat exchange system, optimizes separation effect, and finally reduces the risk of secondary pollution and iteratively optimizes parameters through a filtering module and a feedback cycle mechanism. The application significantly improves the metal ion extraction efficiency while effectively controlling the pollution risk.

Description

Technical Field

[0001] This invention relates to the field of lithium battery recycling technology, and in particular to a wet leaching reaction method for lithium battery recycling. Background Technology

[0002] The lithium battery recycling sector plays an indispensable role in promoting resource recycling and environmental protection. With the rapid development of the new energy industry, the number of used lithium batteries has increased dramatically. How to efficiently extract valuable metals from them while reducing environmental pollution has become a critical issue that the industry urgently needs to address. Research in this field not only concerns the efficiency of resource reuse but also directly impacts the sustainable development of ecosystems.

[0003] However, current wet leaching methods in lithium battery recycling generally suffer from some insurmountable drawbacks. Many existing technologies lack precise control over environmental conditions during the reaction process, limiting the extraction efficiency of metal ions and potentially leading to secondary pollution risks due to improper handling of reaction byproducts. These problems are not simply due to insufficient technical means, but rather stem from a lack of comprehensive and dynamic management of the reaction process, making fluctuations in environmental factors directly impact the recycling effect.

[0004] A deeper technical challenge lies in achieving real-time control of reaction conditions. During wet leaching, subtle changes in key parameters such as pH and temperature significantly affect the dissolution rate and separation efficiency of metal ions. If these parameters cannot be adjusted in time, the reaction may become inefficient or lead to the accumulation of byproducts. For example, when leaching metals in acidic solutions, excessively high or low pH levels may result in incomplete dissolution of some metal ions, or even the formation of difficult-to-treat precipitates; unstable temperatures may cause the reaction rate to run away from control, thereby increasing energy consumption and pollution risks. This insufficient control over the dynamic equilibrium of reaction conditions has become a core obstacle to improving recovery efficiency and reducing environmental impact.

[0005] Therefore, how to accurately monitor and dynamically adjust key parameters such as pH and temperature during wet leaching to ensure efficient extraction of metal ions while effectively controlling the pollution risks from byproducts has become a critical issue that urgently needs to be addressed in the field of lithium battery recycling. Summary of the Invention

[0006] To address the technical problems mentioned in the background section, this invention provides a wet leaching reaction method for lithium battery recycling, comprising: S1, The sensor array collects pH and temperature data in the wet leaching reaction system in real time, integrates multi-source signals and filters noise to obtain a smooth reaction condition sequence; S2, analyze the fluctuation patterns in the reaction condition sequence, determine whether the fluctuation amplitude exceeds the preset fluctuation threshold, and obtain the fluctuation classification results; S3. If the fluctuation classification results show that the pH deviates from the target range, adjust the acid reagent injection rate and determine the corrected pH level. S4: Obtain the corrected pH level and current temperature data, predict the changing trend of metal ion dissolution rate, and obtain the dissolution optimization parameters; S5, determine whether temperature fluctuations lead to overreaction; if so, adjust the cooling medium flow rate to obtain a stable temperature distribution. S6. Extract ion separation interference indicators from stable temperature distribution, assess interference levels, and determine separation effect adjustment schemes; S7 extracts efficiency indicators and recovery quality data to determine the risk of secondary pollution. If the risk is high, the filtration module is activated to remove pollutants and obtain purified leachate. S8 uses the purified leachate to detect the concentration of metal ions, iteratively optimizes the dynamic control parameters, and obtains the final resource extraction sequence.

[0007] Furthermore, step S1 includes: Step S11: The sensor array includes an acid-base sensor and a temperature sensor. The acid-base sensor collects the pH value signal of the reaction system, and the temperature sensor collects the temperature signal of the reaction system. Step S12: The pH signal and the temperature signal are processed by Kalman filtering using data fusion technology to obtain a smooth reaction condition sequence.

[0008] Furthermore, the pH sensor is a glass electrode type sensor, and the temperature sensor is a platinum resistance type sensor.

[0009] Furthermore, step S2 includes: Step S21: Input the reaction condition sequence and extract the time-domain and frequency-domain features from the sequence; Step S22: Calculate the fluctuation amplitude based on the time domain features and frequency domain features, compare the fluctuation amplitude with the preset fluctuation threshold, and obtain the fluctuation classification result.

[0010] Furthermore, step S3 includes: Step S31: The proportional-integral-derivative controller obtains the pH deviation as the input deviation and calculates the proportional, integral, and derivative terms. Step S32: The proportional-integral-derivative controller outputs a control signal based on the proportional term, integral term, and derivative term, and the control signal adjusts the injection rate of the acidic reagent. Step S33: Monitor the pH of the reaction system to obtain the corrected pH level.

[0011] Furthermore, step S4 includes: Step S41: Use the corrected pH level and current temperature data as input features; Step S42: The neural network algorithm processes the input features through multiple network layers and outputs the changing trend of the metal ion dissolution rate. Step S43: Adjust the reaction parameters according to the changing trend to obtain the optimized dissolution parameters.

[0012] Furthermore, step S5 includes: Step S51: Based on the temperature target value in the dissolution optimization parameters, analyze whether there are significant fluctuations in the current temperature data; Step S52: If the judgment result indicates that temperature fluctuations lead to an overreaction risk, then adjust the cooling medium flow rate. Step S53: Continuously monitor the temperature change of the reaction system to obtain a stable temperature distribution.

[0013] Furthermore, step S6 includes: Step S61: Generate an ion concentration distribution map from the stable temperature distribution and extract ion separation interference indicators; Step S62: The decision tree algorithm uses the ion separation interference index as the node splitting condition and traverses the tree structure to evaluate the interference level. Step S63: Select the branch path based on the interference level and determine the separation effect adjustment scheme.

[0014] Furthermore, step S8 includes: Step S81: Use a spectrometer to measure the ion concentration in the purified leachate; Step S82: Compare the ion concentration with the target concentration to generate an error signal; Step S83: Input the error signal into the proportional-integral-derivative controller and the neural network algorithm, iteratively adjust the dynamic control parameters, and obtain the final resource extraction sequence.

[0015] Furthermore, the dynamic control parameters include target pH value, target temperature value, and stirring rate.

[0016] The technical solution provided by this invention has the following beneficial effects: This invention discloses a wet leaching reaction method for lithium battery recycling, addressing the challenges of low metal ion extraction efficiency and high secondary pollution risks in lithium battery recycling processes. By real-time monitoring and dynamic control of reaction conditions, it achieves the dual goals of high-efficiency extraction and environmental friendliness. The invention uses a sensor array to accurately collect pH and temperature data, combined with data fusion technology to eliminate noise and ensure data reliability. Support vector machines and neural network algorithms are employed to analyze fluctuation patterns and predict dissolution rates, dynamically adjusting reaction parameters. A proportional-integral-derivative controller and a heat exchange system are used to stabilize pH and temperature, optimizing separation performance. Finally, a filtration module and feedback loop mechanism reduce the risk of secondary pollution and iteratively optimize parameters. The core innovation of this invention lies in its intelligent end-to-end control, seamlessly integrating data acquisition, analysis, adjustment, and purification, significantly improving metal ion extraction efficiency while effectively controlling pollution risks, providing a sustainable and efficient technical solution for the lithium battery recycling industry. Attached Figure Description

[0017] Figure 1 This is a flowchart of a wet leaching reaction method for lithium battery recycling according to the present invention. Detailed Implementation

[0018] To further understand the content of this invention, a detailed description of the invention is provided in conjunction with the accompanying drawings and embodiments. The specific embodiments described herein are for illustrative purposes only and are not intended to limit the invention. It should also be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings.

[0019] This invention provides a wet leaching reaction method for lithium battery recycling, aiming to improve metal ion extraction efficiency and reduce the risk of secondary pollution by real-time monitoring and dynamic control of reaction conditions. The technical solution of this invention is described in detail below with reference to specific embodiments to make the objectives, technical solutions, and advantages of this invention clearer.

[0020] In one embodiment, the method provided by this invention primarily targets the optimized control of the wet leaching reaction in lithium battery recycling, covering the entire process from data acquisition to reaction parameter adjustment. Wet leaching, as a crucial step in lithium battery recycling, directly affects the dissolution efficiency of metal ions and the final quality of resource recovery. By precisely monitoring and controlling key parameters such as pH and temperature in the reaction system, this method can effectively address fluctuations during the reaction process and ensure stable reaction conditions. The specific implementation steps include multiple stages such as data acquisition, fluctuation analysis, and parameter adjustment, each focused on improving leaching efficiency and resource recovery rate.

[0021] like Figure 1As shown in this embodiment, a wet leaching reaction method for lithium battery recycling specifically includes: S1, the sensor array collects pH and temperature data in the wet leaching reaction system in real time, integrates multi-source signals and filters noise to obtain a smooth reaction condition sequence.

[0022] Specifically, wet leaching reaction systems typically involve complex chemical environments. Acidity / alkalinity and temperature, as core parameters affecting the dissolution rate of metal ions, require real-time monitoring using high-precision sensors. The sensor array deployment must cover multiple key points within the reaction vessel to ensure the comprehensiveness and representativeness of the data. The acquired raw data often contains noise interference, such as signal jumps caused by uneven stirring of the reaction liquid or fluctuations in external ambient temperature. Therefore, data fusion techniques are needed to process the signals and improve data reliability.

[0023] In one embodiment, the implementation of step S1 can be further refined into multiple sub-steps.

[0024] Step S11: The sensor array includes an acid-base sensor and a temperature sensor. The acid-base sensor collects the pH value signal of the reaction system, and the temperature sensor collects the temperature signal of the reaction system.

[0025] In one embodiment, the sensor array comprises multiple pH and temperature sensors distributed at various locations within the reaction vessel, such as the bottom, sidewalls, and near the liquid surface. The pH sensors primarily measure the acidity or alkalinity of the reaction system, typically outputting a pH value. The temperature sensors monitor real-time temperature changes in the reaction liquid, usually outputting a signal in degrees Celsius. Sensor selection must consider corrosion resistance and high sensitivity to withstand the strong acid or alkaline environment of wet leaching reactions. For example, pH sensors can be glass electrode type sensors, which operate stably in strong acid environments, while temperature sensors can be platinum resistance type sensors to ensure measurement accuracy under high or low temperature conditions.

[0026] Step S12: The pH signal and the temperature signal are processed by Kalman filtering using data fusion technology to obtain a smooth reaction condition sequence.

[0027] In one embodiment, data fusion technology is primarily applied to process multi-source signals acquired by sensors to eliminate noise and improve data smoothness. In one possible implementation, data fusion technology can optimize pH and temperature signals based on the Kalman filter method. Kalman filtering, through weighted calculations of historical data and current measurements, can effectively filter out noise caused by sensor errors or external interference. For example, in a reaction system, the operation of a stirring device may cause short-term fluctuations in the pH sensor signal; Kalman filtering can smooth these fluctuations through prediction and update mechanisms, generating a continuous and stable sequence of reaction conditions. Such sequence data not only reflects the true state of the reaction system but also provides a reliable basis for subsequent fluctuation analysis.

[0028] For example, in the wet leaching scenario of lithium battery recycling, suppose three pH sensors and three temperature sensors are installed inside the reaction vessel, located at the bottom, middle of the side wall, and near the top of the vessel, respectively. These sensors collect data every second. The pH values ​​output by the pH sensors range from 1.5 to 3.5, and the temperature values ​​output by the temperature sensors range from 25 to 60 degrees Celsius. For the collected data, Kalman filtering first builds a state model based on the historical data of each sensor to predict the theoretical value at the current moment. Then, it is corrected by combining the actual measured values, ultimately generating a smooth sequence of reaction conditions. This processing method can significantly reduce signal jumps caused by liquid turbulence or sensor vibration, laying the foundation for subsequent analysis.

[0029] In one embodiment, data fusion technology can be combined with other auxiliary methods to further improve data quality. For example, a weighted average method can be introduced to comprehensively process signals from sensors at different locations, assigning higher weights to sensors closer to the core reaction region to highlight changes in reaction conditions in key areas. Assuming that sensors at the bottom of the container experience greater data fluctuations due to their proximity to the stirring device, while the data from sensors in the middle of the sidewalls are relatively stable, a higher weighting can be assigned to the sidewall sensors (e.g., 60%), while the bottom sensors have a weight of 30% and the top sensors 10%. Through this weighted fusion method, the resulting reaction condition sequence can more accurately reflect the overall state of the reaction system.

[0030] It should be noted that the arrangement of the sensor array and the specific implementation of the data fusion technology can be adjusted according to the structure and process requirements of the reaction vessel. For example, in a small laboratory reaction vessel, only two pH sensors and one temperature sensor may be needed, while in a large industrial-grade reaction vessel, more than ten sensor points may be required to cover a wider reaction area. The parameter settings of the data fusion technology, such as the process noise and measurement noise covariance in Kalman filtering, can also be dynamically adjusted according to the actual reaction environment to adapt to the data characteristics under different scenarios.

[0031] S2, analyze the fluctuation patterns in the reaction condition sequence, determine whether the fluctuation amplitude exceeds the preset fluctuation threshold, and obtain the fluctuation classification results.

[0032] In one embodiment, the reaction condition sequence includes continuous data on the changes in pH and temperature over time. These data often exhibit certain fluctuation characteristics, such as periodic fluctuations or abrupt jumps. The fluctuation pattern reflects the stability of the reaction system; if the fluctuation amplitude is too large, it may mean that the reaction conditions deviate from the ideal state, thereby affecting the dissolution efficiency of metal ions. Therefore, it is necessary to use a support vector machine algorithm to quantitatively analyze the fluctuation pattern and determine whether control measures are needed based on the analysis results.

[0033] In one embodiment, step S2 can be further broken down into multiple sub-steps.

[0034] Step S21: Input the reaction condition sequence and extract the time-domain and frequency-domain features from the sequence.

[0035] Specifically, time-domain features mainly include indicators such as the mean, variance, and peak value of the sequence data, used to describe the basic pattern of fluctuations; frequency-domain features, on the other hand, extract the dominant frequency components and energy distribution of fluctuations through spectral analysis of the sequence data, used to reveal the periodicity of the fluctuations. For example, for pH sequence data, the support vector machine algorithm can calculate its variance over the past 5 minutes as a time-domain feature, and simultaneously extract its main fluctuation frequencies through Fourier transform as frequency-domain features.

[0036] Step S22: Calculate the fluctuation amplitude based on the time domain features and frequency domain features, compare the fluctuation amplitude with the preset fluctuation threshold, and obtain the fluctuation classification result.

[0037] In one embodiment, based on the extracted time-domain and frequency-domain features, the support vector machine algorithm further calculates the fluctuation amplitude and compares it with a preset fluctuation threshold to obtain a fluctuation classification result. The fluctuation amplitude calculation can be based on a comprehensive weighted average of feature values, such as weighting and summing the variance value in the time-domain features and the energy value in the frequency-domain features according to a certain ratio to obtain a comprehensive fluctuation index. The preset fluctuation threshold can be determined based on process requirements and historical experience data; for example, the pH fluctuation threshold can be set to 0.3 pH units, and the temperature fluctuation threshold can be set to 3 degrees Celsius. If the comprehensive fluctuation index exceeds the preset fluctuation threshold, the fluctuation classification result is abnormal and requires subsequent adjustment; if it does not exceed the fluctuation threshold, the classification result is normal, and the reaction conditions can be maintained as they are.

[0038] For example, in the wet leaching process of lithium battery recycling, suppose the reaction condition sequence shows that the pH fluctuated from 2.0 to 2.5 and then fell back to 2.1 over the past 10 minutes, while the temperature rose from 40 degrees Celsius to 45 degrees Celsius. In step S21, the support vector machine algorithm extracts the variance of the pH sequence as 0.08 and the variance of the temperature sequence as 2.5. Simultaneously, spectral analysis reveals that the dominant frequency of pH fluctuations is 0.2 times per minute, and the dominant frequency of temperature fluctuations is 0.1 times per minute. In step S22, the algorithm weights these feature values ​​to calculate a comprehensive pH fluctuation index of 0.35, exceeding the pH fluctuation threshold of 0.3, while the comprehensive temperature fluctuation index is 2.8, not exceeding the temperature fluctuation threshold of 3.0. Therefore, the fluctuation classification results indicate that the pH fluctuation is abnormal, while the temperature fluctuation is normal, and subsequent regulation should focus on adjusting the pH.

[0039] In one possible implementation, the training process of the Support Vector Machine (SVM) algorithm can be optimized based on historical reaction data to improve the accuracy of fluctuation classification. For example, acidity and temperature sequence data from different reaction batches over the past month can be collected, and samples with abnormal and normal fluctuations can be labeled. These samples can then be used to train the SVM model, enabling it to more accurately identify fluctuation patterns. During training, the kernel function type and penalty parameters of the SVM can be adjusted to adapt to the data distribution characteristics under different reaction conditions. This training method can significantly improve the algorithm's adaptability to complex fluctuation patterns and reduce the occurrence of misclassifications.

[0040] It should be noted that the output format of the fluctuation classification results can be adjusted according to actual needs. For example, in some scenarios, the fluctuation classification results can be more than just a binary classification of abnormal or normal; they can be further refined into multiple levels, such as slight abnormality, moderate abnormality, and severe abnormality, each corresponding to different regulatory priorities. For slight abnormalities, only minor adjustments to the reaction parameters may be needed, while for severe abnormalities, the reaction may need to be immediately paused and a comprehensive inspection may be required. This multi-level classification approach allows for more flexible handling of different fluctuation situations in the reaction system.

[0041] In one embodiment, fluctuation analysis can be further improved by combining other auxiliary indicators. For example, in addition to the fluctuation amplitudes of pH and temperature, changes in the stirring rate in the reaction system can be introduced as an auxiliary feature. It is assumed that a sudden increase in the stirring rate may cause short-term fluctuations in the pH signal, but such fluctuations do not necessarily indicate that the reaction conditions are out of control. The support vector machine algorithm can use the change in stirring rate as an additional input feature, combining it with the fluctuation features of pH and temperature for comprehensive analysis, thereby avoiding misjudging normal fluctuations caused by equipment operation as abnormal situations. This multi-feature fusion analysis method can more comprehensively reflect the operating status of the reaction system.

[0042] S3. If the fluctuation classification results show that the pH deviates from the target range, adjust the acid reagent injection rate and determine the corrected pH level.

[0043] pH is a core parameter in wet leaching reactions, directly affecting the dissolution rate of metal ions and the chemical equilibrium of the reaction system. If the pH deviates from the target range—for example, too low a pH may lead to a slow reaction rate, while too high a pH may cause side reactions or equipment corrosion—precise control is necessary. The proportional-integral-derivative (PID) controller is a classic closed-loop control method that dynamically adjusts the control output based on the magnitude and trend of the deviation, ensuring that the pH quickly returns to the target range.

[0044] Specifically, the implementation of step S3 can be broken down into multiple sub-steps.

[0045] Step S31: The proportional-integral-derivative controller obtains the pH deviation as the input deviation and calculates the proportional, integral, and derivative terms.

[0046] For example, assuming the target pH is 2.0 and the current pH is 2.4, the deviation is 0.4. The controller calculates this deviation by breaking it down into three parts: a proportional term, an integral term, and a derivative term. The proportional term directly outputs the control quantity based on the magnitude of the deviation; the larger the deviation, the larger the control quantity. The integral term accumulates historical deviations to eliminate steady-state error. The derivative term predicts the trend of deviation changes to prevent control overshoot.

[0047] In step S32, the proportional-integral-derivative controller outputs a control signal based on the proportional term, integral term, and derivative term, and the control signal adjusts the injection rate of the acidic reagent.

[0048] Specifically, the proportional-integral-derivative (PID) controller outputs a comprehensive control signal based on the calculated proportional, integral, and derivative terms. This control signal is used to adjust the injection rate of the acidic reagent. For example, if the pH is too high, the controller will output a positive control signal to increase the injection rate of the acidic reagent, such as sulfuric acid solution, to lower the pH of the reaction system. The injection rate can be adjusted by controlling the pump speed or valve opening; the specific adjustment range is determined by the magnitude of the control signal. Assuming a significant deviation in pH, the proportional term outputs a large positive control value, while the integral term, accumulating past deviations, further increases the control value. Ultimately, the control signal instructs the pump to increase the acidic reagent injection rate from 100 ml / min to 150 ml / min.

[0049] Step S33: Monitor the pH of the reaction system to obtain the corrected pH level.

[0050] After adjusting the acid reagent injection rate, the sensor array continues to monitor the pH changes of the reaction system to obtain the corrected pH level. For example, after increasing the acid reagent injection rate, the pH gradually decreases from 2.4 to 2.1, approaching the target value of 2.0. At this point, the corrected pH level can be considered to have basically met the process requirements. If the corrected pH still does not reach the target range, the proportional-integral-derivative (PID) controller will continue to iteratively calculate the deviation and adjust the injection rate until the pH stabilizes within the target range. This closed-loop control method enables rapid response and precise control of the pH.

[0051] For example, in the wet leaching reaction of lithium battery recycling, assuming the target pH range is 1.8 to 2.2, and the current pH is 2.5, exceeding the upper limit of the target range. In step S31, the proportional-integral-derivative (PID) controller calculates the deviation as 0.5. The proportional term outputs a basic control quantity based on the deviation magnitude. The integral term detects that the pH has been consistently high over the past 5 minutes and further increases the control quantity. The derivative term detects a slow rate of change in pH and outputs a small positive control quantity. In step S32, after combining the three control quantities, the controller outputs a signal instructing the acid reagent injection pump to increase the injection rate of sulfuric acid solution from 80 ml / min to 120 ml / min. In step S33, within 10 minutes after the adjustment, the pH sensor detects that the pH has dropped to 2.1, entering the target range, and the corrected pH level meets the requirements.

[0052] In one possible implementation, the parameters of the proportional-integral-derivative (PID) controller can be optimized based on the characteristics of the reaction system. For example, for systems with fast reaction rates, the weight of the proportional term can be increased to accelerate the response; while for systems with significant hysteresis, the weight of the integral term can be increased to eliminate long-term biases. Parameter optimization can be achieved through simulation testing using historical reaction data to find the most suitable parameter combination for the current process. For instance, in small-scale laboratory reactions, the proportional term weight can be set to 0.6, the integral term weight to 0.3, and the derivative term weight to 0.1, while in large-scale industrial reactions, the proportional term weight may need to be reduced to 0.4 to avoid overshoot.

[0053] It should be noted that the selection and injection method of the acidic reagent can be adjusted according to specific process requirements. For example, in some lithium battery recycling scenarios, sulfuric acid solution can be chosen as the acidic reagent due to its low cost and high reactivity; while in scenarios requiring high reaction purity, nitric acid solution can be chosen to reduce the introduction of impurities. Regarding the injection method, in addition to direct addition via an injection pump, a dripping device can be used to achieve more precise micro-control, avoiding drastic fluctuations in pH. This flexible adjustment method can adapt to the control requirements under different reaction conditions.

[0054] In one embodiment, pH adjustment can be further enhanced by incorporating a predictive mechanism. For example, before the proportional-integral-derivative (PID) controller outputs a control signal, a simple predictive model can be introduced to estimate the pH range within the next 5 minutes based on the current pH trend and historical data. If the prediction indicates that the pH may continue to deviate from the target range, the controller can increase the control input in advance to prevent further deviation. For instance, assuming the current pH is 2.4 and has been rising over the past 3 minutes, and the predictive model predicts that the pH may reach 2.6 within the next 5 minutes, the controller will output a larger control signal based on the prediction, increasing the acid reagent injection rate directly from 100 ml / min to 160 ml / min, thereby bringing the pH back to the target range more quickly.

[0055] S4: Obtain the corrected pH level and current temperature data, predict the changing trend of metal ion dissolution rate, and obtain the dissolution optimization parameters.

[0056] In the wet leaching reaction of lithium battery recycling, the dissolution rate of metal ions directly determines the efficiency of resource extraction. Acidity / alkalinity and temperature are key influencing factors, and their changes significantly affect the dissolution rate. By comprehensively analyzing the corrected acidity / alkalinity levels and current temperature data using a neural network algorithm, the future trend of the dissolution rate can be predicted, thus providing a basis for adjusting subsequent reaction parameters. This predictive method helps process engineers anticipate potential reaction rate fluctuations, ensuring efficient dissolution of metal ions.

[0057] Specifically, step S4 can be further broken down into multiple sub-steps.

[0058] Step S41: Use the corrected pH level and current temperature data as input features.

[0059] The corrected pH level is typically a stable value obtained from the adjustments made by the proportional-integral-derivative (PID) controller in the preceding steps, while the current temperature data is acquired in real-time by a sensor array. The neural network algorithm organizes this data into a multi-dimensional feature vector for subsequent processing. For example, assuming the corrected pH level is 2.1 and the current temperature is 42 degrees Celsius, the input feature can be represented as a data pair containing these two values, describing the current state of the reaction system.

[0060] Step S42: The neural network algorithm processes the input features through multiple network layers and outputs the changing trend of the metal ion dissolution rate.

[0061] Neural networks typically consist of an input layer, hidden layers, and an output layer. The input layer receives pH and temperature data, the hidden layers perform nonlinear transformations on the data through multiple layers of neurons to extract potential feature correlations, and the output layer generates a prediction of the dissolution rate's changing trend. For example, the output might indicate that under the current pH and temperature conditions, the dissolution rate will either increase or decrease over the next 10 minutes. Such predictions reflect the dynamic characteristics of the reaction system, providing guidance for subsequent optimization.

[0062] Step S43: Adjust the reaction parameters according to the changing trend to obtain the optimized dissolution parameters.

[0063] Dissolution optimization parameters mainly include target values ​​for pH and temperature, as well as possible suggestions for adjusting the stirring rate. For example, if the prediction results show that the dissolution rate will decrease, the target pH or temperature values ​​can be appropriately increased to stimulate the dissolution of metal ions; if the prediction results show that the dissolution rate is too fast and may lead to side reactions, the target temperature value can be appropriately decreased to slow down the reaction rate. In this way, dissolution optimization parameters can ensure that the reaction system achieves a balance between efficient dissolution and stable control.

[0064] For example, in a wet leaching scenario for lithium battery recycling, assuming the corrected pH level is 2.0 and the current temperature is 45 degrees Celsius. In step S41, these data are input into a neural network algorithm as feature vectors. In step S42, the neural network, through multi-layer hidden layer analysis, discovers that the dissolution rate may decrease by about 10% within the next 15 minutes under the current conditions because the high temperature may cause some metal ions to precipitate. In step S43, based on this prediction, the dissolution optimization parameters suggest adjusting the target temperature from 45 degrees Celsius to 40 degrees Celsius, while keeping the target pH unchanged, to avoid a further decrease in the dissolution rate. Through this adjustment, the dissolution efficiency of metal ions is maintained at a high level.

[0065] In one possible implementation, the training process of the neural network algorithm can be optimized based on historical reaction data to improve prediction accuracy. For example, data from multiple past batches of wet leaching reactions, including changes in pH, temperature, and corresponding metal ion dissolution rates, can be collected and labeled as training samples to train the neural network model. During training, the number of layers and neurons in the network can be adjusted to adapt to the complexity of different reaction conditions. For instance, a three-layer network structure can be used in small-scale laboratory reactions, while five or more layers may be needed in large-scale industrial reactions to capture more complex feature associations.

[0066] It should be noted that the output of the neural network algorithm can also be corrected by incorporating other auxiliary information. For example, in some scenarios, there may be undetected interfering factors in the reaction system, such as changes in impurity content or uneven stirring, which may cause deviations in the prediction results. Therefore, a correction mechanism can be introduced to dynamically adjust the prediction results based on monitoring data of the actual dissolution rate. For instance, if the prediction indicates an increase in the dissolution rate, but actual monitoring shows a decrease, the correction mechanism will reduce the weight of the prediction result and prioritize parameter adjustments based on the actual monitoring data. This correction method can improve the reliability of the prediction.

[0067] S5 determines whether temperature fluctuations lead to overreaction. If so, the flow rate of the cooling medium is adjusted through the heat exchange system to obtain a stable temperature distribution.

[0068] Temperature fluctuations are a common problem in wet leaching reactions. Excessively high temperatures can lead to excessively rapid reaction rates, triggering side reactions or metal ion precipitation, while excessively low temperatures can slow down the dissolution rate and reduce extraction efficiency. Therefore, it is necessary to assess whether current temperature fluctuations will adversely affect the reaction based on the target temperature value in the dissolution optimization parameters, and to dynamically regulate the temperature through the heat exchange system to maintain it within a suitable range.

[0069] Specifically, step S5 can be broken down into multiple sub-steps.

[0070] Step S51: Based on the temperature target value in the dissolution optimization parameters, analyze whether there are significant fluctuations in the current temperature data. For example, assuming the recommended temperature target value in the dissolution optimization parameters is 40 degrees Celsius, and the current temperature data fluctuates from 38 degrees Celsius to 43 degrees Celsius in the past 5 minutes, it can be considered that the temperature fluctuation is large and may lead to overreaction. The criteria for judging overreaction can be set according to process requirements; for example, a temperature fluctuation exceeding 3 degrees Celsius or lasting for more than 10 minutes is considered a risk of overreaction.

[0071] Step S52: If the judgment result indicates that temperature fluctuations lead to an over-reaction risk, then adjust the flow rate of the cooling medium to reduce the temperature of the reaction system.

[0072] A heat exchange system typically consists of cooling pipes, a cooling medium such as cooling water or cooling oil, and flow control valves. The flow rate of the cooling medium can be adjusted by controlling the valve opening. For example, if the current temperature is 5 degrees Celsius higher than the target value, the valve opening can be increased, raising the cooling medium flow rate from 10 liters per minute to 15 liters per minute, thereby accelerating heat dissipation and lowering the temperature of the reaction system. In this way, excessive reactions caused by overheating can be effectively suppressed.

[0073] Step S53: Continuously monitor the temperature change of the reaction system to obtain a stable temperature distribution.

[0074] A stable temperature distribution typically manifests as temperature data fluctuating within a small range around the target value; for example, with a target value of 40 degrees Celsius, the temperature stabilizes between 39 and 41 degrees Celsius. A sensor array can collect temperature data from multiple points in real time and generate an overall temperature distribution map using data fusion technology, which is used to evaluate the control effect of the heat exchange system. If the temperature distribution remains unstable, the cooling medium flow rate can be further adjusted until the expected target is achieved.

[0075] For example, in the wet leaching reaction of lithium battery recycling, assuming the recommended target temperature for the dissolution optimization parameters is 42 degrees Celsius, while current temperature data shows that the reaction system temperature has risen from 40 degrees Celsius to 46 degrees Celsius in the past 10 minutes. In step S51, analysis reveals that the temperature fluctuation exceeds 3 degrees Celsius and lasts for a long time, indicating a risk of over-reaction. In step S52, the heat exchange system is activated, the control valve opening is increased, and the cooling water flow rate is increased from 8 liters per minute to 12 liters per minute to accelerate heat dissipation. In step S53, within 15 minutes after the adjustment, the temperature sensor monitors that the reaction system temperature gradually decreases to 43 degrees Celsius and stabilizes between 42 and 43 degrees Celsius, forming a relatively ideal temperature distribution.

[0076] In one possible implementation, the design of the heat exchange system can be optimized based on the size of the reaction vessel and process requirements. For example, in small laboratory reaction vessels, the heat exchange system can employ a simple external cooling water jacket, achieving cooling through natural circulation; while in large industrial-scale reaction vessels, multiple sets of cooling pipes may be built-in, equipped with high-power pumping equipment to ensure that the cooling medium can quickly cover the entire reaction area. The choice of cooling medium can also be adjusted according to specific circumstances; for example, in low-temperature environments, antifreeze can be used as the cooling medium to prevent the pipes from freezing.

[0077] It should be noted that temperature control can be further improved by combining it with predictive mechanisms. For example, before adjusting the cooling medium flow rate, the temperature fluctuation range within the next 10 minutes can be estimated based on the current temperature trend and historical data. If the prediction indicates that the temperature may continue to rise, the cooling medium flow rate can be increased in advance to prevent the temperature from exceeding the target range further. This combination of prediction and control can significantly improve the response speed and stability of temperature control.

[0078] S6. Extract ion separation interference indicators from the stable temperature distribution, assess the interference level, and determine the separation effect adjustment plan.

[0079] In wet leaching reactions, the separation efficiency of metal ions directly affects the final quality of resource recovery. Uneven temperature distribution can lead to uneven ion concentration distribution, resulting in separation interference. Therefore, it is necessary to extract relevant indicators from the stable temperature distribution, analyze the interference level, and formulate corresponding adjustment schemes using decision tree algorithms to optimize ion separation.

[0080] Specifically, step S6 can be broken down into multiple sub-steps.

[0081] Step S61: Generate an ion concentration distribution map from the stable temperature distribution and extract ion separation interference indicators.

[0082] Ion concentration distribution maps can be generated by collecting ion concentration data at multiple points within the reaction vessel, and this data is typically closely related to temperature distribution. For example, regions with higher temperatures may exhibit higher ion concentrations due to faster reaction rates, while regions with lower temperatures may have lower ion concentrations. Ion separation interference indices can be defined as the variance or maximum concentration difference of the concentration distribution, used to quantify the degree of non-uniformity in ion distribution.

[0083] In step S62, the decision tree algorithm uses the ion separation interference index as the node splitting condition and traverses the tree structure to evaluate the interference level.

[0084] Decision tree algorithms construct a tree structure to compare interference indicators with other auxiliary parameters such as temperature distribution range and stirring rate at multiple levels, progressively classifying the interference level. For example, the first-level node of the decision tree might use whether the variance of the interference indicator exceeds a certain threshold as a condition; if it does, it enters the high-interference branch; if not, it enters the low-interference branch. Subsequently, the second-level node might further consider the temperature distribution range to make a judgment, ultimately outputting an evaluation result of the interference level, such as high interference, medium interference, or low interference.

[0085] Step S63: Select a branch path based on the interference level assessment results to determine the separation effect adjustment plan. The separation effect adjustment plan mainly includes adjusting the stirring rate, local heating or cooling strategies, etc. For example, if the interference level is assessed as high, the adjustment plan might suggest increasing the stirring rate from 100 rpm to 150 rpm to promote uniform ion distribution; if the interference level is medium, it might suggest local heating of the lower-temperature areas to balance the reaction rate. In this way, differentiated adjustment measures can be taken for different interference levels to improve the ion separation effect.

[0086] For example, in the wet leaching reaction of lithium battery recycling, assuming a stable temperature distribution shows the bottom temperature of the reaction vessel is 40 degrees Celsius and the top temperature is 43 degrees Celsius. In step S61, data is collected using an ion concentration sensor to generate a concentration distribution map, revealing a lower ion concentration at the bottom and a higher concentration at the top, with a calculated interference index variance of 0.5. In step S62, the decision tree algorithm uses the variance value of 0.5 as input, determines that it exceeds the variance threshold of 0.3, and enters the high interference branch. Subsequently, it further confirms the interference level as high interference by combining the temperature distribution range. In step S63, based on the high interference assessment result, the adjustment plan suggests increasing the stirring rate from 80 revolutions per minute to 120 revolutions per minute, while simultaneously locally heating the bottom region to raise the temperature to 42 degrees Celsius to balance the ion concentration distribution.

[0087] In one possible implementation, the decision tree algorithm can be optimized based on historical reaction data. For example, ion concentration and temperature distribution data from multiple past batches can be collected, and the adjustment effects under different levels of interference can be labeled. This data can then be used as training samples to build the decision tree model. During training, the depth and splitting conditions of the decision tree can be adjusted to adapt to the interference characteristics under different reaction conditions. For instance, in small reaction vessels, the decision tree can be designed as a three-layer structure, while in large reaction vessels, a five-layer or deeper structure may be needed to capture more complex interference patterns.

[0088] It should be noted that the definition of ion separation interference indicators can be adjusted according to process requirements. For example, in some scenarios, interference indicators can be calculated not only based on the variance of concentration distribution, but also by combining the concentration ratios between ion species. If lithium and nickel ions are present in the reaction system, and the concentration ratio of the two ions deviates from the ideal range, the weight of the interference indicator may be further increased to highlight the difficulty of separating specific ions. This multi-dimensional indicator design allows for a more comprehensive evaluation of the separation effect.

[0089] S7 extracts efficiency indicators and recovery quality data to determine the risk of secondary pollution. If the risk is high, the filtration module is activated to remove pollutants and obtain purified leachate.

[0090] In wet leaching reactions, the extraction efficiency and recovery quality of metal ions are important standards for evaluating the process effectiveness. However, impurities or byproducts may be generated during the reaction, leading to the risk of secondary pollution. Therefore, it is necessary to comprehensively analyze the extraction efficiency and recovery quality data, assess the pollution risk, and, if necessary, purify the leachate through a filtration module to ensure the purity of the leachate.

[0091] For example, in the wet leaching reaction of lithium battery recycling, assuming that after the separation effect adjustment scheme is implemented, the metal dissolution rate is 88%, but the purity of the leachate is only 82%, these data are extracted as efficiency indicators and quality data. Risk assessment model analysis found that the purity value is lower than the purity threshold of 85% and the dissolution rate is high, which may indicate the presence of many by-product impurities. The assessment result is high risk, and the filtration module is activated. The leachate first passes through a microfiltration membrane with a pore size of 0.2 micrometers to remove suspended particles, and then passes through an activated carbon adsorption layer to remove dissolved organic impurities. Finally, the purity value is increased to 92%, and the purified leachate is obtained.

[0092] In one embodiment, the filtration module typically includes a microfiltration membrane and an adsorbent material. The microfiltration membrane is used to remove suspended particles and larger impurities, while the adsorbent material, such as activated carbon, is used to adsorb dissolved organic matter or heavy metal impurities. For example, after the leachate passes through the microfiltration membrane, particulate impurities are effectively intercepted, and then residual organic matter is adsorbed by the activated carbon layer, ultimately resulting in a purified leachate with higher purity.

[0093] In one possible implementation, the design of the filtration module can be optimized based on the characteristics of the leachate. For example, in scenarios with large impurity particles, a microfiltration membrane with a larger pore size can be selected to improve filtration efficiency; while in scenarios with a high amount of organic impurities, the thickness of the activated carbon layer can be increased or materials with higher adsorption capacity can be selected. The operating parameters of the filtration module, such as flow rate and filtration time, can also be dynamically adjusted according to the pollution risk level. For example, under high-risk conditions, the flow rate can be reduced and the filtration time extended to ensure purification effect.

[0094] It should be noted that the secondary pollution risk assessment model can be optimized by incorporating other auxiliary data. For example, in addition to metal leaching rate and purity value, the color and turbidity data of the leachate can be used as assessment criteria. If the leachate has a darker color or higher turbidity, the pollution risk weight may be further increased, prompting the system to activate the filtration module earlier. This multi-dimensional assessment approach allows for more accurate identification of potential pollution problems.

[0095] S8 uses the purified leachate to detect the concentration of metal ions, iteratively optimizes the dynamic control parameters, and obtains the final resource extraction sequence.

[0096] The ultimate goal of wet leaching is to efficiently extract target metal ions from the leachate, and the accurate detection of metal ion concentration and the optimization of dynamic control parameters are key to achieving this goal. Through a feedback loop mechanism, reaction parameters can be continuously adjusted based on actual concentration data to ensure optimal resource extraction results.

[0097] Specifically, step S8 can be broken down into multiple sub-steps.

[0098] Step S81: Use a spectrometer to measure the ion concentration in the purified leachate.

[0099] A spectrometer determines the concentration of a target metal ion by detecting the absorption characteristics of a leachate at specific wavelengths of light. For example, for lithium ions, a specific wavelength range can be selected for detection to obtain its concentration, such as 2.5 grams of lithium ions per liter of solution. The spectrometer offers high measurement accuracy, providing reliable data support for subsequent control measures.

[0100] Step S82: Compare the ion concentration with the target concentration to generate an error signal.

[0101] The feedback loop mechanism compares the measured ion concentration with the target concentration, generating an error signal. The target concentration is typically set according to process requirements; for example, the target concentration for lithium ions might be 3.0 grams per liter. If the actual concentration is 2.5 grams per liter, the error signal is -0.5, indicating that the concentration is too low and the dissolution rate needs to be increased or the reaction time extended. If the actual concentration is higher than the target value, the error signal is positive, indicating that the concentration is too high and the reaction rate needs to be slowed down. Through this comparison, the error signal reflects the difference between the current reaction state and the ideal state.

[0102] Step S83: Input the error signal into the proportional-integral-derivative controller and the neural network algorithm, iteratively adjust the dynamic control parameters, and obtain the final resource extraction sequence.

[0103] Dynamic control parameters include target pH, target temperature, and stirring rate. The proportional-integral-derivative (PID) controller adjusts these parameters in real-time based on error signals, while a neural network algorithm predicts the long-term impact of the adjusted parameters on ion concentration. For example, if the error signal indicates a low concentration, the controller might adjust the target pH from 2.0 to 1.8 to accelerate metal ion dissolution, while the neural network algorithm predicts that this adjustment will bring the concentration close to the target value within 30 minutes. The final resource extraction sequence records the parameter adjustment trajectory and concentration changes throughout the reaction process, guiding subsequent batch reactions.

[0104] For example, in the wet leaching reaction of lithium battery recycling, assuming the purified leachate, as detected by a spectrometer, has a lithium-ion concentration of 2.3 g / L, while the target concentration is 3.0 g / L. In step S81, the detected data is recorded as the current concentration value. In step S82, the feedback loop mechanism calculates an error signal of -0.7, indicating that the concentration is too low. In step S83, the error signal is input to a proportional-integral-derivative (PID) controller, which adjusts the target pH value from 2.1 to 1.9 and simultaneously increases the target temperature value from 40 degrees Celsius to 43 degrees Celsius to accelerate the dissolution rate. The neural network algorithm predicts that this adjustment will increase the concentration to 2.9 g / L within 20 minutes, close to the target value. The final resource extraction sequence records the adjusted parameter combination and concentration change trend, providing a reference for subsequent process optimization.

[0105] In one possible implementation, the iterative optimization process of the feedback loop mechanism can be combined with multiple rounds of adjustments. For example, after the first round of adjustments, if the ion concentration still does not reach the target value, a second round of parameter adjustments can be performed based on new error signals until the concentration stabilizes within the target range. During the iteration process, the parameter weights of the controller can be dynamically adjusted; for example, increasing the weight of the proportional term in the initial stage to accelerate the response speed, and increasing the weight of the integral term in the later stage to eliminate residual errors. Through this multi-round iterative approach, the ideal reaction state can be gradually approximated.

[0106] It should be noted that the output format of resource extraction sequences can be adjusted according to actual needs. For example, in some scenarios, resource extraction sequences can record not only parameter adjustment trajectories and concentration changes, but also auxiliary information such as reaction time and reagent dosage, for a comprehensive evaluation of the process's economy and efficiency. For instance, if the reaction time is long but the concentration increase is limited, it may suggest that process engineers optimize reagent ratios or shorten the reaction cycle to reduce costs. This multi-dimensional data recording method can provide comprehensive support for subsequent process improvements.

[0107] In one embodiment, metal ion concentration detection can be optimized by combining multiple detection methods. For example, in addition to a spectrometer, electrochemical analysis can be introduced to help determine the ion concentration by measuring the conductivity or potential value of the leachate. Suppose the spectrometer shows a lithium ion concentration of 2.4 grams per liter, while the electrochemical analysis measures 2.6 grams per liter; the average of the two can be taken as the final concentration data to improve detection accuracy. This combination of multiple methods can effectively reduce the error of a single detection method.

[0108] If the technical solution of this application involves personal information, the product using this solution has clearly informed the user of the personal information processing rules and obtained the user's voluntary consent before processing the personal information. If sensitive personal information is involved, the user's separate consent has been obtained before processing, and the "express consent" requirement is met. For example, a clear sign is placed at the collection device such as a camera to inform the user that they have entered the collection area, and the user's voluntary entry is considered as consent; or the processing device clearly indicates the processing rules and obtains authorization through pop-up windows or by asking the user to upload information themselves. The personal information processing rules include the processor, the purpose of processing, the processing method, and the types of personal information.

[0109] Although the present invention has been described in detail above with general descriptions and specific embodiments, modifications or improvements can be made to it, which will be obvious to those skilled in the art. Therefore, all such modifications or improvements made without departing from the spirit of the present invention fall within the scope of protection claimed by the present invention.

Claims

1. A wet leaching reaction method for lithium battery recycling, characterized in that, include: S1, The sensor array collects pH and temperature data in the wet leaching reaction system in real time, integrates multi-source signals and filters noise to obtain a smooth reaction condition sequence; S2, analyze the fluctuation patterns in the reaction condition sequence, determine whether the fluctuation amplitude exceeds the preset fluctuation threshold, and obtain the fluctuation classification results; S3. If the fluctuation classification results show that the pH deviates from the target range, adjust the acid reagent injection rate and determine the corrected pH level. S4: Obtain the corrected pH level and current temperature data, predict the changing trend of metal ion dissolution rate, and obtain the dissolution optimization parameters; S5, determine whether temperature fluctuations lead to overreaction; if so, adjust the cooling medium flow rate to obtain a stable temperature distribution. S6. Extract ion separation interference indicators from stable temperature distribution, assess interference levels, and determine separation effect adjustment schemes; S7 extracts efficiency indicators and recovery quality data to determine the risk of secondary pollution. If the risk is high, the filtration module is activated to remove pollutants and obtain purified leachate. S8 uses the purified leachate to detect the concentration of metal ions, iteratively optimizes the dynamic control parameters, and obtains the final resource extraction sequence.

2. The method as described in claim 1, characterized in that, Step S1 includes: Step S11: The sensor array includes an acid-base sensor and a temperature sensor. The acid-base sensor collects the pH value signal of the reaction system, and the temperature sensor collects the temperature signal of the reaction system. Step S12: The pH signal and the temperature signal are processed by Kalman filtering using data fusion technology to obtain a smooth reaction condition sequence.

3. The method as described in claim 2, characterized in that, The pH sensor is a glass electrode type sensor, and the temperature sensor is a platinum resistance type sensor.

4. The method as described in claim 1, characterized in that, Step S2 includes: Step S21: Input the reaction condition sequence and extract the time-domain and frequency-domain features from the sequence; Step S22: Calculate the fluctuation amplitude based on the time domain features and frequency domain features, compare the fluctuation amplitude with the preset fluctuation threshold, and obtain the fluctuation classification result.

5. The method as described in claim 1, characterized in that, Step S3 includes: Step S31: The proportional-integral-derivative controller obtains the pH deviation as the input deviation and calculates the proportional, integral, and derivative terms. Step S32: The proportional-integral-derivative controller outputs a control signal based on the proportional term, integral term, and derivative term, and the control signal adjusts the injection rate of the acidic reagent. Step S33: Monitor the pH of the reaction system to obtain the corrected pH level.

6. The method as described in claim 1, characterized in that, Step S4 includes: Step S41: Use the corrected pH level and current temperature data as input features; Step S42: The neural network algorithm processes the input features through multiple network layers and outputs the changing trend of the metal ion dissolution rate. Step S43: Adjust the reaction parameters according to the changing trend to obtain the optimized dissolution parameters.

7. The method as described in claim 1, characterized in that, Step S5 includes: Step S51: Based on the temperature target value in the dissolution optimization parameters, analyze whether there are significant fluctuations in the current temperature data; Step S52: If the judgment result indicates that temperature fluctuations lead to an overreaction risk, then adjust the cooling medium flow rate. Step S53: Continuously monitor the temperature change of the reaction system to obtain a stable temperature distribution.

8. The method as described in claim 1, characterized in that, Step S6 includes: Step S61: Generate an ion concentration distribution map from the stable temperature distribution and extract ion separation interference indicators; Step S62: The decision tree algorithm uses the ion separation interference index as the node splitting condition and traverses the tree structure to evaluate the interference level. Step S63: Select the branch path based on the interference level and determine the separation effect adjustment scheme.

9. The method as described in claim 1, characterized in that, Step S8 includes: Step S81: Use a spectrometer to measure the ion concentration in the purified leachate; Step S82: Compare the ion concentration with the target concentration to generate an error signal; Step S83: Input the error signal into the proportional-integral-derivative controller and the neural network algorithm, iteratively adjust the dynamic control parameters, and obtain the final resource extraction sequence.

10. The method as described in claim 9, characterized in that, The dynamic control parameters include target pH value, target temperature value, and stirring rate.

Images