Battery positive electrode material recycling purification method and system based on automatic control

Abstract

The application provides a battery positive electrode material cyclic extraction and purification method and system based on automatic control, relates to the battery material recycling technical field, and comprises the following steps: obtaining a battery positive electrode material to be treated and initial component data thereof, determining an optimal separation sequence of metal components through a multi-objective optimization algorithm and performing staged dissolution separation, obtaining real-time component data after the completion of each separation stage and comparing the real-time component data with expected component concentration to obtain a comprehensive deviation vector; calculating a process state deviation metric value through a state evaluation algorithm and determining a predicted minimum cycle number, calculating an operation parameter correction value and an actual cycle execution number based on a correlation relationship and generating a cycle control instruction; executing the cycle control instruction and collecting product purity data, identifying a key separation stage through a cause-effect inference algorithm and updating the cycle control instruction if the product purity data does not meet the standard, and repeatedly executing until high-purity metal material is obtained.

Description

Technical Field

[0001] This invention relates to the field of battery material recycling technology, and in particular to a method and system for the cyclic extraction and purification of battery cathode materials based on automated control. Background Technology

[0002] With the rapid development of new energy vehicles and the energy storage industry, the use of lithium-ion batteries has experienced explosive growth. Waste lithium-ion batteries contain a large amount of valuable metals such as cobalt, nickel, manganese, and lithium, possessing significant economic and strategic value. The recycling and reuse of battery cathode materials can not only alleviate the shortage of raw materials but also effectively reduce environmental pollution.

[0003] Traditional methods for recycling battery cathode materials mainly fall into two categories: pyrometallurgy and hydrometallurgy. Pyrometallurgy has a simple process flow but is energy-intensive, polluting, and has a low metal recovery rate. Hydrometallurgy, due to its environmental friendliness and high metal recovery rate, has become the mainstream research direction, mainly using chemical dissolution, extraction, and precipitation to separate and purify metals.

[0004] However, existing battery cathode material cyclic extraction and purification technologies still suffer from problems such as failure to dynamically adjust the separation sequence according to the actual component characteristics of different battery cathode materials, resulting in some metal components being difficult to separate effectively, and lack of real-time process status monitoring and parameter dynamic adjustment mechanisms, leading to problems such as response lag and insufficient adjustment accuracy. Summary of the Invention

[0005] This invention provides a method and system for the cyclic extraction and purification of battery cathode materials based on automated control, which can at least solve some of the problems existing in the prior art.

[0006] A first aspect of this invention provides a method for the cyclic extraction and purification of battery cathode materials based on automated control, comprising:

[0007] The cathode material of the battery to be processed and its corresponding initial component data are obtained. The optimal separation sequence of the metal components in the initial component data is determined by a multi-objective optimization algorithm, and a staged dissolution and separation operation is performed. After each separation stage is completed, the real-time component data of the intermediate solution is obtained and compared with the corresponding expected component concentration to obtain a comprehensive deviation vector.

[0008] The process state deviation metric between the comprehensive deviation vector and the expected component concentration is calculated by the state evaluation algorithm, and the predicted minimum number of cycles is determined. Based on the process state deviation metric and the predicted minimum number of cycles, the correlation between the combination of operating parameters corresponding to each separation stage and the comprehensive deviation vector is determined. Based on the correlation, the operating parameter correction value and the actual number of cycle executions are calculated, and the corresponding cycle control command is generated.

[0009] The cyclic control command is executed, and after each cycle, product purity data is collected and compared with a preset purity threshold. If the product purity data is less than the purity threshold, the process state deviation metric is decomposed into the contribution component of each separation stage through a causal inference algorithm, and the key separation stage is identified. Based on the combination of operating parameters corresponding to the key separation stage, the parameter adjustment amount is calculated and the cyclic control command is updated. The execution is repeated until the product purity data is not less than the purity threshold, and high-purity metal material is obtained.

[0010] In one alternative implementation,

[0011] The process involves acquiring the cathode material to be processed and its corresponding initial component data, determining the optimal separation sequence of metal components in the initial component data using a multi-objective optimization algorithm, and performing a staged dissolution and separation operation, including:

[0012] The composition of the battery cathode material to be processed is analyzed to obtain the mass percentage of metal elements and the distribution characteristics of impurity elements to obtain the initial composition data. The standard electrode potential difference and solubility product constant of each metal element in the initial composition data are extracted and the separation feasibility coefficient between different metal elements is calculated. A separation feasibility matrix is ​​constructed based on the separation feasibility coefficient.

[0013] Traverse the separation feasibility matrix, take the metal component pair with the maximum separation feasibility coefficient as the priority separation combination, take the position index of the priority separation combination in the separation feasibility matrix as the starting point, mark the metal component corresponding to the priority separation combination as the assigned metal and determine the unmarked metal component, and calculate the cumulative value of the mutual interference between the unmarked metal component and the assigned metal.

[0014] With the optimization objectives of maximizing metal recovery rate and minimizing the number of separation stages, the cumulative value of mutual interference is used as a constraint. Multiple candidate extraction sequences that satisfy the optimization objectives are calculated to obtain a set of candidate schemes. The candidate scheme that minimizes the cumulative value of mutual interference is selected and the corresponding optimal separation sequence is determined. Based on the optimal separation sequence, the type of dissolution medium corresponding to each separation stage is matched, the dissolution temperature range and dissolution time range of each separation stage are determined, and the dissolution separation operation is performed.

[0015] In one alternative implementation,

[0016] After each separation stage, real-time component data of the intermediate solution is acquired and compared with the corresponding expected component concentrations to obtain a comprehensive deviation vector, including:

[0017] After each separation stage is completed, the measured concentration values ​​of each metal element and the measured concentration values ​​of impurity elements in the intermediate solution are obtained to obtain the real-time component data;

[0018] Based on the target metal element in the current separation stage of the optimal separation sequence and the corresponding dissolution temperature range and dissolution time range, the expected concentration values ​​of the target metal element and impurity element in the current separation stage are calculated to obtain the expected component concentration. The measured concentration values ​​of each metal element in the real-time component data are compared with the expected concentration values ​​of the corresponding metal element in the expected component concentration to obtain the concentration deviation value of each metal element.

[0019] According to the arrangement order of different metal elements in the optimal separation sequence, the concentration deviation values ​​of each metal element are combined to construct a concentration deviation vector. The difference between the measured concentration value of the impurity element in the real-time component data and the expected concentration value of the impurity element in the expected component concentration is calculated to obtain the concentration deviation value of the impurity element.

[0020] The impurity residue index for the current separation stage is calculated based on the concentration deviation value of the impurity element, and the comprehensive deviation vector is obtained by combining the concentration deviation vector with the impurity residue index.

[0021] In one alternative implementation,

[0022] The process state deviation metric between the comprehensive deviation vector and the expected component concentration is calculated using a state assessment algorithm, and the minimum predicted cycle number is determined, including:

[0023] A multidimensional feature space is constructed by extracting a concentration deviation vector and an impurity residue index from the comprehensive deviation vector. The zero-deviation state corresponding to the expected concentration values ​​of each metal element and the expected concentration values ​​of impurities in the expected component concentration is taken as the target reference point in the multidimensional feature space. The Mahalanobis distance from the concentration deviation vector to the target reference point is calculated as the concentration deviation sub-value. The impurity deviation sub-value is calculated based on the distance from the impurity residue index to the impurity dimension corresponding to the target reference point. The concentration deviation sub-value and the impurity deviation sub-value are mapped to a preset deviation metric space. A dynamic covariance matrix is ​​constructed by combining the rate of change in the continuous separation stage. The process state deviation metric value is obtained by coupling and transforming the concentration deviation sub-value and the impurity deviation sub-value through the dynamic covariance matrix.

[0024] Obtain the process state deviation metric and operating parameter combination in the historical separation stage and determine the influence relationship between the process state deviation metric and the operating parameter combination and the number of cycles. Substitute the process state deviation metric and the current operating parameter combination corresponding to the current separation stage into the influence relationship, and calculate the numerical sequence of the process state deviation metric converging to zero under different number of cycles in combination with the real-time component data. Identify the cycle number node that makes the process state deviation metric first decrease to the preset convergence threshold in the numerical sequence and determine the predicted minimum number of cycles.

[0025] In one alternative implementation,

[0026] Based on the process state deviation metric and the predicted minimum number of cycles, the correlation between the combination of operating parameters for each separation stage and the comprehensive deviation vector is determined. Based on the correlation, the operating parameter correction value and the actual number of cycle executions are calculated, and the corresponding cycle control instructions are generated, including:

[0027] Extract the combination of operating parameters and the comprehensive deviation vector of each separation stage from the historical separation data, and determine the degree of influence of the combination of operating parameters on each component of the comprehensive deviation vector. Quantify the degree of influence to obtain the parameter sensitivity vector. Construct a cycle efficiency index based on the numerical difference between the process state deviation metric and the predicted minimum number of cycles in the current separation stage. Construct a response tensor based on the parameter sensitivity vector and the cycle efficiency index, and identify the operating parameter adjustment path that maximizes the convergence rate of the comprehensive deviation vector to zero. Use the operating parameter adjustment path as the correlation relationship.

[0028] Substitute the comprehensive deviation vector of the current separation stage into the correlation relationship, combine the real-time component data to calculate the operation parameter adjustment amount that reduces the comprehensive deviation vector to a preset deviation threshold, and add it to the operation parameter combination of the current separation stage to obtain the operation parameter correction value.

[0029] The operation parameter correction value and the deviation metric value of the current separation stage are substituted into the correlation relationship. The decay rate of the current process state deviation metric value converging to zero is calculated by combining the real-time component data and matched with the preset deviation threshold to obtain the correction cycle number. The correction cycle number is compared with the predicted minimum cycle number, and the cycle number with the smaller value is taken as the actual cycle execution number. The operation parameter correction value and the actual cycle execution number are encapsulated to obtain the cycle control instruction.

[0030] In one alternative implementation,

[0031] The process state deviation metric is decomposed into the contribution components of each separation stage using a causal inference algorithm, and key separation stages are identified, including:

[0032] Each separation stage is taken as a causal node. Based on the transmission relationship between the real-time component data and each separation stage, the directed causal edges between different causal nodes are identified. The common influence of environmental conditions and raw material batches on multiple separation stages is obtained and potential confounding factor nodes are set. The causal strength is calculated for each causal edge and the causal coefficient is assigned. Based on the deviation metric of the process state, the causal effect value of different separation stages is calculated by counterfactual inference.

[0033] The current process state deviation metric is used as the overall deviation to perform virtual intervention operations on each causal node. Under the condition of fixing the value of the potential confounding factor node, the pure causal contribution of each separation stage to the overall deviation is calculated. Based on the causal coefficient and the causal effect value, the pure causal contribution is decomposed to each causal node to obtain the contribution component.

[0034] A comprehensive impact index is calculated based on the contribution components and the causal effect value. Based on the deviation between the comprehensive impact index and the product purity data, the direct and indirect causal responsibility for the purity non-compliance of each separation stage is calculated and the total causal responsibility score is obtained through nonlinear combination. The separation stages in which the total causal responsibility score exceeds a preset responsibility threshold and the causal effect value exceeds a preset effect threshold are identified as key separation stages.

[0035] In one alternative implementation,

[0036] Based on the combination of operating parameters corresponding to the key separation stage, the parameter adjustment amount is calculated and the loop control command is updated. This process is repeated until the product purity data is not less than the purity threshold, resulting in high-purity metallic materials, including:

[0037] Extract the current combination of operating parameters corresponding to the key separation stage, substitute the contribution component of the key separation stage and the pre-acquired causal effect value into the preset parameter response relationship, and calculate the range of change in operating parameters required to reduce the contribution component to the preset target contribution threshold to obtain the parameter adjustment amount of the key separation stage.

[0038] The parameter adjustment amount is superimposed on the current combination of operating parameters corresponding to the critical separation stage to obtain the updated combination of operating parameters. The combination of operating parameters for non-critical separation stages remains unchanged and is combined and encapsulated to generate an updated loop control instruction.

[0039] The loop control instruction is executed. After execution, the purity data of the current product is collected and compared with the purity threshold. If the purity data of the current product is not less than the purity threshold, the loop is terminated and the current product is output to obtain the high-purity metal material. Otherwise, the loop control instruction is executed repeatedly until the purity of the current product is not less than the purity threshold.

[0040] A second aspect of this invention provides a battery cathode material cyclic extraction and purification system based on automated control, comprising:

[0041] The separation analysis module is used to acquire the cathode material of the battery to be processed and the corresponding initial component data. It determines the optimal separation sequence of the metal components in the initial component data through a multi-objective optimization algorithm and performs a staged dissolution and separation operation. After each separation stage is completed, it acquires the real-time component data of the intermediate solution and compares it with the corresponding expected component concentration to obtain a comprehensive deviation vector.

[0042] The evaluation and optimization module is used to calculate the process state deviation metric between the comprehensive deviation vector and the expected component concentration through a state evaluation algorithm and determine the predicted minimum number of cycles. Based on the process state deviation metric and the predicted minimum number of cycles, it determines the correlation between the combination of operating parameters corresponding to each separation stage and the comprehensive deviation vector. Based on the correlation, it calculates the operating parameter correction value and the actual number of cycles and generates the corresponding cycle control command.

[0043] The cyclic purification module is used to execute the cyclic control command. After each cycle, it collects product purity data and compares it with a preset purity threshold. If the product purity data is less than the purity threshold, it uses a causal inference algorithm to decompose the deviation of the process state into the contribution component of each separation stage and identifies the key separation stage. Based on the combination of operating parameters corresponding to the key separation stage, it calculates the parameter adjustment amount and updates the cyclic control command. This process is repeated until the product purity data is not less than the purity threshold, thus obtaining high-purity metal material.

[0044] A third aspect of the present invention provides an electronic device, comprising:

[0045] A processor and a memory for storing processor-executable instructions, wherein the processor is configured to invoke instructions stored in the memory to perform the aforementioned method.

[0046] A fourth aspect of the present invention provides a computer-readable storage medium having stored thereon computer program instructions that, when executed by a processor, implement the aforementioned method.

[0047] In this invention, a multi-objective optimization algorithm is used to determine the optimal separation sequence of metal components and execute a staged dissolution and separation operation. This enables the formulation of personalized separation strategies based on the initial component characteristics of different battery cathode materials, avoiding resource waste and inefficiency caused by fixed processes. This significantly improves the overall efficiency and economy of metal separation. A state evaluation algorithm is used to calculate the process state deviation metric in real time and predict the minimum number of cycles. A dynamic correlation between the combination of operating parameters and the comprehensive deviation vector is established, enabling precise control and adaptive adjustment of the separation and purification process. This effectively shortens the processing time required to reach the target purity and reduces production costs. A causal inference algorithm is used to decompose the process state deviation metric into the contribution components of each separation stage and identify key separation stages. This allows for precise location of the core links affecting product purity and targeted parameter adjustments, avoiding process fluctuations caused by blindly adjusting all parameters. This ensures the high purity and quality stability of the final product and improves the overall level of battery material recycling and resource recycling efficiency. Attached Figure Description

[0048] Figure 1 This is a schematic flowchart of an automated control-based cyclic extraction and purification method for battery cathode materials according to an embodiment of the present invention.

[0049] Figure 2 This is a flowchart illustrating the optimization of purification operation parameters for a cyclic extraction and purification method for battery cathode materials based on automated control, as described in an embodiment of the present invention. Detailed Implementation

[0050] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0051] The technical solution of the present invention will be described in detail below with reference to specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

[0052] Figure 1 This is a schematic flowchart of an automated control-based cyclic extraction and purification method for battery cathode materials according to an embodiment of the present invention. Figure 1 As shown, the method includes:

[0053] The cathode material of the battery to be processed and its corresponding initial component data are obtained. The optimal separation sequence of the metal components in the initial component data is determined by a multi-objective optimization algorithm, and a staged dissolution and separation operation is performed. After each separation stage is completed, the real-time component data of the intermediate solution is obtained and compared with the corresponding expected component concentration to obtain a comprehensive deviation vector.

[0054] The process state deviation metric between the comprehensive deviation vector and the expected component concentration is calculated by the state evaluation algorithm, and the predicted minimum number of cycles is determined. Based on the process state deviation metric and the predicted minimum number of cycles, the correlation between the combination of operating parameters corresponding to each separation stage and the comprehensive deviation vector is determined. Based on the correlation, the operating parameter correction value and the actual number of cycle executions are calculated, and the corresponding cycle control command is generated.

[0055] The cyclic control command is executed, and after each cycle, product purity data is collected and compared with a preset purity threshold. If the product purity data is less than the purity threshold, the process state deviation metric is decomposed into the contribution component of each separation stage through a causal inference algorithm, and the key separation stage is identified. Based on the combination of operating parameters corresponding to the key separation stage, the parameter adjustment amount is calculated and the cyclic control command is updated. The execution is repeated until the product purity data is not less than the purity threshold, and high-purity metal material is obtained.

[0056] In one alternative implementation,

[0057] The process involves acquiring the cathode material to be processed and its corresponding initial component data, determining the optimal separation sequence of metal components in the initial component data using a multi-objective optimization algorithm, and performing a staged dissolution and separation operation, including:

[0058] The composition of the battery cathode material to be processed is analyzed to obtain the mass percentage of metal elements and the distribution characteristics of impurity elements to obtain the initial composition data. The standard electrode potential difference and solubility product constant of each metal element in the initial composition data are extracted and the separation feasibility coefficient between different metal elements is calculated. A separation feasibility matrix is ​​constructed based on the separation feasibility coefficient.

[0059] Traverse the separation feasibility matrix, take the metal component pair with the maximum separation feasibility coefficient as the priority separation combination, take the position index of the priority separation combination in the separation feasibility matrix as the starting point, mark the metal component corresponding to the priority separation combination as the assigned metal and determine the unmarked metal component, and calculate the cumulative value of the mutual interference between the unmarked metal component and the assigned metal.

[0060] With the optimization objectives of maximizing metal recovery rate and minimizing the number of separation stages, the cumulative value of mutual interference is used as a constraint. Multiple candidate extraction sequences that satisfy the optimization objectives are calculated to obtain a set of candidate schemes. The candidate scheme that minimizes the cumulative value of mutual interference is selected and the corresponding optimal separation sequence is determined. Based on the optimal separation sequence, the type of dissolution medium corresponding to each separation stage is matched, the dissolution temperature range and dissolution time range of each separation stage are determined, and the dissolution separation operation is performed.

[0061] The composition of the cathode material to be processed was analyzed to obtain the mass percentage of metal elements and the distribution characteristics of impurity elements, thus obtaining initial composition data. Taking a ternary lithium battery cathode material as an example, inductively coupled plasma mass spectrometry (ICP-MS) analysis revealed that the content of nickel was 32.5%, cobalt was 15.7%, manganese was 13.2%, lithium was 7.1%, and aluminum was 0.8%. Trace impurities of iron (0.3%) and copper (0.2%) were also detected.

[0062] The standard electrode potentials and solubility product constants of each metal element were extracted from the initial composition data in the periodic table and physicochemical databases. For the detected elements nickel, cobalt, manganese, lithium, and aluminum, their standard electrode potentials were recorded as -0.25V, -0.28V, -1.18V, -3.04V, and -1.66V, respectively. The standard electrode potential differences between different metal elements were calculated; for example, the electrode potential difference between nickel and cobalt was 0.03V, and the electrode potential difference between nickel and manganese was 0.93V. The solubility product constants of the elements in different solution environments were recorded; for example, the solubility product constant of nickel in sulfuric acid solution was 4.0 × 10⁻⁶. -16 Cobalt is 4.0 × 10 -20 .

[0063] The feasibility coefficients for separating different metallic elements were calculated based on the standard electrode potential difference and the solubility product constant. The calculation of the feasibility coefficients considers the ratio of the electrode potential difference to the solubility product constant; a larger difference and a more significant difference in the solubility product constant indicate a higher feasibility of separation between the elements. The calculated feasibility coefficients for separating nickel and cobalt were 0.32, for nickel and manganese 0.78, for nickel and lithium 0.93, and for cobalt and manganese 0.65.

[0064] The calculated separation feasibility coefficients between all metal element pairs are used to construct a separation feasibility matrix. In this embodiment, a 5×5 matrix is ​​constructed, containing the separation feasibility coefficients between the five main metal elements: nickel, cobalt, manganese, lithium, and aluminum. The diagonal elements in the matrix are zero, indicating that elements of the same type do not require separation.

[0065] The separation feasibility matrix is ​​traversed to identify the metal component pairs with the highest separation feasibility coefficient as priority separation combinations. In this embodiment, the separation feasibility coefficient of nickel and lithium, 0.93, is the maximum value in the matrix, thus nickel and lithium are determined as the priority separation combination. Starting from the position index of the priority separation combination in the separation feasibility matrix, nickel and lithium are marked as assigned metals, while cobalt, manganese, and aluminum are identified as unmarked metal components.

[0066] The cumulative mutual interference between unlabeled and assigned metals was calculated. This cumulative mutual interference took into account the chemical similarity, co-precipitation tendency, and potential cross-contamination during the separation process. The calculated cumulative mutual interference values ​​were 0.41 for cobalt and assigned metals, 0.56 for manganese, and 0.25 for aluminum.

[0067] With the optimization objectives of maximizing metal recovery and minimizing the number of separation stages, and using the cumulative value of mutual interference as a constraint, multiple candidate extraction sequences satisfying the optimization objectives are calculated, resulting in a set of candidate schemes. A traversal algorithm generates several possible separation sequence schemes, such as Scheme 1: Nickel-Lithium-Cobalt-Manganese-Aluminum; Scheme 2: Nickel-Lithium-Aluminum-Cobalt-Manganese; Scheme 3: Lithium-Nickel-Aluminum-Manganese-Cobalt. For each scheme, the total recovery rate and the required number of separation stages are calculated. For example, Scheme 1 has an expected total recovery rate of 93.2% and requires 4 separation stages; Scheme 2 has an expected total recovery rate of 94.1% and requires 4 separation stages; Scheme 3 has an expected total recovery rate of 91.8% and requires 5 separation stages.

[0068] Candidate schemes that minimize the cumulative mutual interference were screened, and the corresponding optimal separation sequence was determined. After calculation and comparison, Scheme 2 had the smallest cumulative mutual interference value of 0.28 and the highest total recovery rate. Therefore, the optimal separation sequence was determined to be nickel-lithium-aluminum-cobalt-manganese.

[0069] The optimal separation sequence was used to match the type of dissolving medium for each separation stage. In the first stage, sulfuric acid solution was used to preferentially dissolve nickel, with the pH controlled between 3.5 and 4.0. In the second stage, sodium bicarbonate solution was used to extract lithium, with the pH controlled between 8.0 and 9.0. In the third stage, sodium hydroxide solution was used to separate aluminum, with the pH controlled between 10.5 and 11.0. In the fourth stage, ammonia solution was used to separate cobalt, with the pH controlled between 9.5 and 10.0. In the final stage, ferrous sulfate solution was used to recover manganese.

[0070] The dissolution temperature and time ranges for each separation stage were determined. The dissolution temperature for nickel was set at 65 to 75°C, and the dissolution time at 45 to 60 minutes; the dissolution temperature for lithium was set at 45 to 55°C, and the dissolution time at 30 to 40 minutes; the dissolution temperature for aluminum was set at 55 to 60°C, and the dissolution time at 20 to 30 minutes; the dissolution temperature for cobalt was set at 60 to 65°C, and the dissolution time at 40 to 50 minutes; and the dissolution temperature for manganese was set at 50 to 55°C, and the dissolution time at 35 to 45 minutes.

[0071] In this embodiment, by performing a detailed analysis of the initial composition data of the battery cathode material and constructing a separation feasibility matrix by combining the standard electrode potential difference and solubility product constant of the metal elements, the inherent separability of different metal components can be quantitatively characterized during the separation path planning stage. By introducing the cumulative value of mutual interference as a constraint, and combining the dual optimization objectives of maximizing metal recovery rate and minimizing the number of separation stages, the generation of candidate extraction sequences takes into account both separation efficiency and process simplification. Through the synergistic mechanism of separation feasibility matrix and mutual interference evaluation, the intelligent and scientific planning of separation sequence is realized, significantly reducing the impact of mutual interference between metals on the overall separation quality.

[0072] In one alternative implementation,

[0073] After each separation stage, real-time component data of the intermediate solution is acquired and compared with the corresponding expected component concentrations to obtain a comprehensive deviation vector, including:

[0074] After each separation stage is completed, the measured concentration values ​​of each metal element and the measured concentration values ​​of impurity elements in the intermediate solution are obtained to obtain the real-time component data;

[0075] Based on the target metal element in the current separation stage of the optimal separation sequence and the corresponding dissolution temperature range and dissolution time range, the expected concentration values ​​of the target metal element and impurity element in the current separation stage are calculated to obtain the expected component concentration. The measured concentration values ​​of each metal element in the real-time component data are compared with the expected concentration values ​​of the corresponding metal element in the expected component concentration to obtain the concentration deviation value of each metal element.

[0076] According to the arrangement order of different metal elements in the optimal separation sequence, the concentration deviation values ​​of each metal element are combined to construct a concentration deviation vector. The difference between the measured concentration value of the impurity element in the real-time component data and the expected concentration value of the impurity element in the expected component concentration is calculated to obtain the concentration deviation value of the impurity element.

[0077] The impurity residue index for the current separation stage is calculated based on the concentration deviation value of the impurity element, and the comprehensive deviation vector is obtained by combining the concentration deviation vector with the impurity residue index.

[0078] After each separation stage, inductively coupled plasma mass spectrometry (ICP-MS) was used to obtain the measured concentrations of each metal element and impurity element in the intermediate solution, yielding real-time component data. Taking nickel extraction in the first separation stage as an example, the measured concentration of nickel in the intermediate solution was 5.63 g / L. The measured concentrations of cobalt (0.42 g / L), manganese (0.28 g / L), lithium (0.13 g / L), and aluminum (0.05 g / L) were also detected. Regarding impurity elements, the measured concentrations of iron (0.07 g / L) and copper (0.04 g / L) were also observed. These data constitute the real-time component dataset for the first stage.

[0079] Based on the target metal element in the current separation stage of the optimal separation sequence, along with the corresponding dissolution temperature and time ranges, the expected concentrations of the target metal element and impurity elements in the current separation stage are calculated to obtain the expected component concentrations. The calculation method is based on known material composition, separation temperature, dissolution time, and dissolution medium characteristics, combined with a metal element dissolution kinetic model for prediction. For the first separation stage, under the conditions of dissolution temperature 70℃ and dissolution time 50 minutes in sulfuric acid solution, the expected concentration of nickel is 5.85 g / L. Simultaneously, the expected concentrations are 0.31 g / L for cobalt, 0.22 g / L for manganese, 0.08 g / L for lithium, and 0.03 g / L for aluminum. Regarding impurity elements, the expected concentrations are 0.05 g / L for iron and 0.03 g / L for copper.

[0080] The concentration deviation of each metal element is calculated by comparing the measured concentration of each metal element in the real-time component data with the expected concentration of the corresponding metal element in the expected component concentration. In the first separation stage, the calculated concentration deviation of nickel is -0.22 g / L, indicating that the actual dissolved nickel concentration is lower than expected; the concentration deviation of cobalt is 0.11 g / L, manganese is 0.06 g / L, lithium is 0.05 g / L, and aluminum is 0.02 g / L. The deviation values ​​indicate the degree of difference between the actual dissolution behavior of each metal element during the separation process and the expected situation.

[0081] According to the arrangement order of different metal elements in the optimal separation sequence, the concentration deviation values ​​of each metal element are combined to construct a concentration deviation vector. Based on the aforementioned determined optimal separation sequence nickel-lithium-aluminum-cobalt-manganese, the constructed concentration deviation vector is [-0.22, 0.05, 0.02, 0.11, 0.06], with units of g / L.

[0082] The concentration deviation of impurity elements is obtained by calculating the difference between the measured concentration of impurity elements in the real-time component data and the expected concentration of impurity elements in the expected component concentration. In the first separation stage, the concentration deviation of iron is 0.02 g / L and the concentration deviation of copper is 0.01 g / L.

[0083] The residual impurity index for the current separation stage was calculated based on the concentration deviation values ​​of impurity elements. The calculation of the residual impurity index considered the toxicity coefficient of each impurity element, its impact on subsequent separation processes, and the absolute value of the deviation, and was obtained through a weighted average. The toxicity coefficients of iron and copper in the first separation stage were 1.2 and 1.5, respectively, and their impact coefficients on subsequent separation processes were 0.8 and 1.1, respectively. The overall calculated residual impurity index for the first separation stage was 0.046.

[0084] The concentration deviation vector is combined with the impurity residue index to obtain the comprehensive deviation vector. In the first separation stage, the comprehensive deviation vector is [-0.22, 0.05, 0.02, 0.11, 0.06, 0.046], in g / L, with the last digit being the dimensionless impurity residue index.

[0085] In this embodiment, by calculating the concentration deviation and quantifying the residual impurities in real time, not only can the visualization and controllability of the separation process be enhanced, but also an accurate basis can be provided for subsequent separation stage adjustments. This avoids the reduction in separation efficiency caused by impurity accumulation or insufficient metal dissolution. The introduction of the comprehensive deviation vector enables the system to fully reflect the overall deviation status of the separation stage, improve the ability to identify abnormal behavior, and thus improve the stability, consistency and final metal recovery quality of the entire separation process.

[0086] In one alternative implementation,

[0087] The process state deviation metric between the comprehensive deviation vector and the expected component concentration is calculated using a state assessment algorithm, and the minimum predicted cycle number is determined, including:

[0088] A multidimensional feature space is constructed by extracting a concentration deviation vector and an impurity residue index from the comprehensive deviation vector. The zero-deviation state corresponding to the expected concentration values ​​of each metal element and the expected concentration values ​​of impurities in the expected component concentration is taken as the target reference point in the multidimensional feature space. The Mahalanobis distance from the concentration deviation vector to the target reference point is calculated as the concentration deviation sub-value. The impurity deviation sub-value is calculated based on the distance from the impurity residue index to the impurity dimension corresponding to the target reference point. The concentration deviation sub-value and the impurity deviation sub-value are mapped to a preset deviation metric space. A dynamic covariance matrix is ​​constructed by combining the rate of change in the continuous separation stage. The process state deviation metric value is obtained by coupling and transforming the concentration deviation sub-value and the impurity deviation sub-value through the dynamic covariance matrix.

[0089] Obtain the process state deviation metric and operating parameter combination in the historical separation stage and determine the influence relationship between the process state deviation metric and the operating parameter combination and the number of cycles. Substitute the process state deviation metric and the current operating parameter combination corresponding to the current separation stage into the influence relationship, and calculate the numerical sequence of the process state deviation metric converging to zero under different number of cycles in combination with the real-time component data. Identify the cycle number node that makes the process state deviation metric first decrease to the preset convergence threshold in the numerical sequence and determine the predicted minimum number of cycles.

[0090] Concentration deviation vector and impurity residue index are separated from the comprehensive deviation vector to construct a multi-dimensional feature space characterizing the separation state. Taking nickel extraction in the first separation stage as an example, the comprehensive deviation vector is [-0.22, 0.05, 0.02, 0.11, 0.06, 0.046]. The first five elements [-0.22, 0.05, 0.02, 0.11, 0.06] constitute the concentration deviation vector, representing the difference between the measured and expected concentrations of five metal elements (nickel, lithium, aluminum, cobalt, and manganese), in g / L. The last element, 0.046, is the impurity residue index, representing the comprehensive residue level of impurity elements such as iron and copper. Based on these data, a six-dimensional feature space is constructed, where the first five dimensions represent the concentration deviation of each metal element, and the sixth dimension represents the impurity residue level.

[0091] A target reference point is determined in a multidimensional feature space as a benchmark for evaluating the degree of deviation. Ideally, the concentration deviation is zero and there are no impurity residues, so the coordinates of the target reference point are [0, 0, 0, 0, 0, 0]. A covariance matrix is ​​constructed to calculate the Mahalanobis distance from the concentration deviation vector to the target reference point. Based on historical data analysis, the covariance matrix of the first separation stage is obtained, which reflects the correlation between the deviations of different metal elements. The Mahalanobis distance of the concentration deviation vector [-0.22, 0.05, 0.02, 0.11, 0.06] is calculated through covariance matrix transformation, yielding a concentration deviation value of 0.37.

[0092] The impurity residue index was processed using a nonlinear transformation to calculate the impurity deviation value. The effect of impurity residue on separation purity exhibits a nonlinear relationship, with trace impurities having a more significant impact on subsequent separation. An exponential nonlinear transformation function was used to process the impurity residue index to 0.046, yielding an impurity deviation value of 0.29. The impurity deviation value, along with the concentration deviation value, constitutes a comprehensive evaluation index of the separation status.

[0093] Concentration deviation and impurity deviation values ​​are mapped to a two-dimensional polar coordinate system to form a unified deviation measurement space. In this space, the vector length represents the overall degree of deviation, and the vector angle represents the proportional relationship between concentration deviation and impurity deviation. The calculated polar coordinates are (0.47, 0.66), where 0.47 is the radius value representing the total degree of deviation, and 0.66 is the radian value representing the structural characteristics of concentration and impurity deviations.

[0094] Data from three consecutive separation stages were analyzed, and the rate of change of each element's concentration was calculated to construct a dynamic covariance matrix. From the first to the second separation stage, the rate of change of nickel concentration was -0.15, lithium concentration was 0.42, aluminum concentration was 0.08, cobalt concentration was -0.09, and manganese concentration was -0.12. The rate of change data constituted a 5×5 dynamic covariance matrix, which captured the correlation patterns of the concentration changes of each element during the separation process.

[0095] A coupled transformation of the polar coordinate representation using a dynamic covariance matrix is ​​employed to calculate the process state deviation metric. This coupled transformation considers the interactions and dynamic characteristics between different elements, converting the polar coordinates (0.47, 0.66) into a scalar form process state deviation metric of 0.52. The process state deviation metric comprehensively reflects the degree of deviation between the current separation stage and the ideal state; a smaller value indicates that the separation effect is closer to expectations.

[0096] Historical data from nearly 30 separation operations in the first separation stage were collected, including process deviation metrics and corresponding combinations of operating parameters. Operating parameters included key process variables such as dissolution temperature, dissolution time, solution pH, and solvent concentration. Through multiple regression analysis, the nonlinear relationship between process deviation metrics, operating parameter combinations, and the number of cycles was established. This step first concatenates the process deviation metrics and operating parameter combinations into a whole, and then determines the nonlinear relationship between this whole and the number of cycles.

[0097] Substituting the current process deviation metric of 0.52 and the combination of operating parameters into the influence relationship, the sequence of changes in the process deviation metric under different cycle numbers was calculated. The current combination of operating parameters is: dissolution temperature 70℃, dissolution time 50 minutes, solution pH 3.5, and solvent concentration 2.2 mol / L. Through iterative calculation, the sequence of changes in the process deviation metric with the number of cycles was obtained: [0.52, 0.43, 0.36, 0.30, 0.25, 0.21, 0.18, 0.15, 0.13, 0.11, 0.09].

[0098] A convergence threshold of 0.10 was set, and the numerical sequence was analyzed to determine the minimum number of predicted cycles. By comparing the relationship between each element in the numerical sequence and the convergence threshold, the number of cycles that first fell below the threshold was identified. In the sequence [0.52, 0.43, 0.36, 0.30, 0.25, 0.21, 0.18, 0.15, 0.13, 0.11, 0.09], the process state deviation metric corresponding to the 10th cycle was 0.11, which was close to the threshold for the first time, and the value corresponding to the 11th cycle was 0.09, which was below the threshold for the first time. Therefore, the minimum number of predicted cycles was determined to be 10.

[0099] In this embodiment, by introducing a multi-dimensional feature space corresponding to the comprehensive deviation vector, concentration deviation and impurity residue are incorporated into a unified deviation measurement system. The zero-deviation state is used as the target reference point, enabling a precise and unified quantitative description of the degree of deviation in the process state. By calculating the Mahalanobis distance of the concentration deviation and the impurity deviation sub-value, and combining the rate of change in the continuous separation stage to construct a dynamic covariance matrix, accurate modeling of the correlation, stage fluctuations, and dynamic coupling relationships between deviation features is achieved. This allows the process state deviation measurement value to more realistically reflect the overall operating state of the current separation process. Mapping deviation features to the deviation measurement space and coupling transformation through dynamic covariance effectively identifies abnormal trends under complex operating conditions. Combining historical cycle data to establish the influence relationship between the process state deviation measurement value, the combination of operating parameters, and the number of cycles, the minimum number of cycles required for the process state deviation measurement value to converge to zero is predicted. This allows the system to estimate the shortest time cost to achieve the target separation effect in advance, significantly reducing unnecessary process repetition, improving process execution efficiency, reducing energy consumption and material waste, and enhancing the stability, convergence speed, and overall economy of the separation process.

[0100] In one alternative implementation,

[0101] Based on the process state deviation metric and the predicted minimum number of cycles, the correlation between the combination of operating parameters for each separation stage and the comprehensive deviation vector is determined. Based on the correlation, the operating parameter correction value and the actual number of cycle executions are calculated, and the corresponding cycle control instructions are generated, including:

[0102] Extract the combination of operating parameters and the comprehensive deviation vector of each separation stage from the historical separation data, and determine the degree of influence of the combination of operating parameters on each component of the comprehensive deviation vector. Quantify the degree of influence to obtain the parameter sensitivity vector. Construct a cycle efficiency index based on the numerical difference between the process state deviation metric and the predicted minimum number of cycles in the current separation stage. Construct a response tensor based on the parameter sensitivity vector and the cycle efficiency index, and identify the operating parameter adjustment path that maximizes the convergence rate of the comprehensive deviation vector to zero. Use the operating parameter adjustment path as the correlation relationship.

[0103] Substitute the comprehensive deviation vector of the current separation stage into the correlation relationship, combine the real-time component data to calculate the operation parameter adjustment amount that reduces the comprehensive deviation vector to a preset deviation threshold, and add it to the operation parameter combination of the current separation stage to obtain the operation parameter correction value.

[0104] The operation parameter correction value and the deviation metric value of the current separation stage are substituted into the correlation relationship. The decay rate of the current process state deviation metric value converging to zero is calculated by combining the real-time component data and matched with the preset deviation threshold to obtain the correction cycle number. The correction cycle number is compared with the predicted minimum cycle number, and the cycle number with the smaller value is taken as the actual cycle execution number. The operation parameter correction value and the actual cycle execution number are encapsulated to obtain the cycle control instruction.

[0105] The combination of operating parameters and the comprehensive deviation vector for each separation stage were extracted from historical separation data to determine the degree of influence of the operating parameter combinations on each component of the comprehensive deviation vector. Taking the separation stage of nickel in the recovery of lithium-ion battery cathode materials as an example, the combination of operating parameters, including dissolution temperature, dissolution time, solution pH, solvent concentration, and stirring speed, was extracted from nearly 30 historical separation data. The corresponding comprehensive deviation vector includes the concentration deviations of five metal elements (nickel, lithium, aluminum, cobalt, and manganese) and the residual impurity index. By changing the operating parameters one by one and recording the change range of each component of the comprehensive deviation vector, the degree of influence was quantified to obtain the parameter sensitivity vector. For example, the sensitivity vector of dissolution temperature to the concentration deviation of each element is [0.18, 0.12, 0.07, 0.15, 0.10, 0.21], representing the change value of each component when the temperature changes by 1℃. Among them, the influence on the nickel element concentration deviation is 0.18, and the influence on the residual impurity index is 0.21. Similarly, the sensitivity vectors for dissolution time are [0.10, 0.08, 0.05, 0.12, 0.07, 0.13], for solution pH are [0.22, 0.15, 0.11, 0.19, 0.14, 0.25], for solvent concentration are [0.25, 0.18, 0.13, 0.21, 0.16, 0.19], and for stirring speed are [0.06, 0.04, 0.03, 0.05, 0.03, 0.08].

[0106] A cycle efficiency index is constructed based on the numerical difference between the current process state deviation metric and the predicted minimum number of cycles. The current process state deviation metric is 0.52, and the predicted minimum number of cycles is 10. Historical data analysis shows that the ideal number of cycles required for complete convergence is 4, thus indicating a numerical difference of 6. The cycle efficiency index is constructed by weighting the process state deviation metric and the difference in the number of cycles. Using a non-linear weighting method, the process state deviation metric is assigned a weight of 0.6, and the difference in the number of cycles is assigned a weight of 0.4, resulting in a cycle efficiency index value of 0.63. A higher cycle efficiency index value indicates lower current separation efficiency and greater optimization potential.

[0107] A response tensor is constructed based on the parameter sensitivity vector and the cycle performance index to identify the operational parameter adjustment path that maximizes the convergence rate of the comprehensive deviation vector towards zero. The response tensor contains the influence functions of five operational parameters on the cycle performance index, and its construction is based on the historical correspondence between the parameter sensitivity vector and the cycle performance index. By analyzing the changing trends of the cycle performance index under different combinations of operational parameter adjustments, the adjustment direction and step size for each parameter are calculated. For dissolution temperature (current value 70℃), the optimal adjustment direction is to increase, with a step size of 2℃ and an adjustment magnitude of 2.86% of the current value. For dissolution time (current value 50 minutes), the optimal adjustment direction is to increase, with a step size of 5 minutes and an adjustment magnitude of 10% of the current value. For solution pH (current value 3.5), the optimal adjustment direction is to increase, with a step size of 0.3 and an adjustment magnitude of 8.57% of the current value. For solvent concentration (current value 2.2 mol / L), the optimal adjustment direction is to increase, with a step size of 0.3 mol / L and an adjustment magnitude of 13.64% of the current value. For the stirring speed, the current value is 600 revolutions per minute. The optimal adjustment direction is to decrease it, with a step size of 50 revolutions per minute and an adjustment range of 8.33% of the current value. The combined adjustment parameters form an operating parameter adjustment path, which serves as the correlation between the operating parameters and the deviation metric of the process state.

[0108] Substituting the comprehensive deviation vector of the current separation stage into the correlation relationship, and combining it with real-time component data, the adjustment amount of the operating parameters that reduces the comprehensive deviation vector to the preset deviation threshold is calculated and added to the combination of operating parameters for the current separation stage to obtain the corrected operating parameter values. The comprehensive deviation vector of the current separation stage is [-0.22, 0.05, 0.02, 0.11, 0.06, 0.046], and the preset deviation threshold is 0.1. The adjustment amounts for each operating parameter are calculated based on the correlation relationship: dissolution temperature adjustment is +2℃, dissolution time adjustment is +5 minutes, solution pH adjustment is +0.3, solvent concentration adjustment is +0.3 mol / L, and stirring speed adjustment is -50 rpm. These adjustments are then added to the current combination of operating parameters to obtain the corrected operating parameter values: dissolution temperature is 72℃, dissolution time is 55 minutes, solution pH is 3.8, solvent concentration is 2.5 mol / L, and stirring speed is 550 rpm.

[0109] The operating parameter correction value and the process state deviation metric value of the current separation stage are substituted into the correlation relationship. Combined with real-time component data, the decay rate at which the current process state deviation metric value converges to zero is calculated and matched with a preset deviation threshold to obtain the number of correction cycles. The decay rate calculation considers the changing trend of the process state deviation metric value with the number of cycles. For the current process state deviation metric value of 0.52 and the operating parameter correction value, the calculated decay sequence is [0.52, 0.39, 0.29, 0.22, 0.16, 0.12, 0.09, 0.07, 0.05]. Analysis of the decay sequence shows that the process state deviation metric value first drops below the preset deviation threshold of 0.1 after the 6th cycle; therefore, the number of correction cycles is 6.

[0110] The corrected number of cycles is compared with the predicted minimum number of cycles, and the cycle number with the smaller value is taken as the actual number of cycles executed. The corrected operating parameter value and the actual number of cycles executed are encapsulated to obtain the cycle control command. The corrected number of cycles is 6, and the predicted minimum number of cycles is 10. The smaller value of 6 is taken as the actual number of cycles executed. The corrected operating parameter value (dissolution temperature 72℃, dissolution time 55 minutes, solution pH 3.8, solvent concentration 2.5 mol / L, stirring speed 550 rpm) and the actual number of cycles executed (6) are encapsulated into a cycle control command, which is then transmitted to the separation system controller to execute subsequent separation operations.

[0111] In this embodiment, by analyzing the correlation between the combination of operating parameters and the comprehensive deviation vector in historical separation data, the influence of each parameter on the deviation component is quantified and a parameter sensitivity vector is formed. This clarifies the strength of the effect of different operating parameters on the separation effect. By combining the deviation metric of the current process state with the predicted minimum number of cycles, a cycle efficiency index is constructed, and a response tensor is further formed to identify the operating parameter adjustment path that maximizes the deviation convergence rate. This makes the parameter adjustment directional and optimal, achieving accurate analysis of the coupling effect of multiple parameters in a complex separation process. The comprehensive deviation vector is substituted into the correlation relationship to calculate the parameter adjustment amount, and the convergence decay rate is evaluated again based on the state after parameter correction. The corrected number of cycles is determined and a cycle control command is formed so that the process rhythm and parameter adjustment can be dynamically optimized simultaneously.

[0112] In one alternative implementation,

[0113] The process state deviation metric is decomposed into the contribution components of each separation stage using a causal inference algorithm, and key separation stages are identified, including:

[0114] Each separation stage is taken as a causal node. Based on the transmission relationship between the real-time component data and each separation stage, the directed causal edges between different causal nodes are identified. The common influence of environmental conditions and raw material batches on multiple separation stages is obtained and potential confounding factor nodes are set. The causal strength is calculated for each causal edge and the causal coefficient is assigned. Based on the deviation metric of the process state, the causal effect value of different separation stages is calculated by counterfactual inference.

[0115] The current process state deviation metric is used as the overall deviation to perform virtual intervention operations on each causal node. Under the condition of fixing the value of the potential confounding factor node, the pure causal contribution of each separation stage to the overall deviation is calculated. Based on the causal coefficient and the causal effect value, the pure causal contribution is decomposed to each causal node to obtain the contribution component.

[0116] A comprehensive impact index is calculated based on the contribution components and the causal effect value. Based on the deviation between the comprehensive impact index and the product purity data, the direct and indirect causal responsibility for the purity non-compliance of each separation stage is calculated and the total causal responsibility score is obtained through nonlinear combination. The separation stages in which the total causal responsibility score exceeds a preset responsibility threshold and the causal effect value exceeds a preset effect threshold are identified as key separation stages.

[0117] Each separation stage was treated as a causal node, and directed causal edges between different causal nodes were identified based on the transmission relationships of real-time component data between the separation stages. In the four-stage separation process for lithium battery cathode material recycling, acid leaching extraction, precipitation separation, solvent extraction, and crystallization purification were set as nodes in a causal network. Analysis of real-time component data from 30 batches identified the causal transmission relationships between the separation stages. The causal edge strength from acid leaching extraction to precipitation separation was 0.78, indicating that state changes in the acid leaching extraction stage affected 78% of the precipitation separation stage; the causal edge strength from precipitation separation to solvent extraction was 0.65, indicating that state changes in the precipitation separation stage affected 65% of the solvent extraction stage; the causal edge strength from solvent extraction to crystallization purification was 0.72, indicating that state changes in the solvent extraction stage affected 72% of the crystallization purification stage; and the causal edge strength from acid leaching extraction directly to crystallization purification was 0.28, indicating that state changes in the acid leaching extraction stage bypassed intermediate steps and directly affected 28% of the crystallization purification stage.

[0118] The combined effects of environmental conditions and raw material batches on multiple separation stages were obtained, and potential contamination factor nodes were established. Ambient temperature varied between 15℃ and 28℃, and humidity varied between 35% and 68%. These environmental factors collectively affected each separation stage, constituting a contamination effect. Comparative analysis of raw material batches revealed that the elemental composition deviation between different batches fluctuated between 5% and 12%, and the impurity content deviation fluctuated between 3% and 8%. Based on these common influencing factors, three potential contamination factor nodes were established: ambient temperature and humidity nodes, raw material composition nodes, and raw material impurity nodes. The influence coefficients of the ambient temperature and humidity nodes on the four stages of acid leaching extraction, precipitation separation, solvent extraction, and crystallization purification were 0.15, 0.12, 0.08, and 0.06, respectively; the influence coefficients of the raw material composition node were 0.22, 0.17, 0.13, and 0.10, respectively; and the influence coefficients of the raw material impurity node were 0.18, 0.14, 0.11, and 0.09, respectively.

[0119] For each causal edge, the causal strength was calculated and a causal coefficient was assigned. Time lag correlation analysis was used to extract the temporal dependencies between different separation stages from real-time component data. For the causal edge from acid leaching to precipitation separation, the causal coefficients for nickel were 0.83, lithium 0.76, aluminum 0.65, cobalt 0.80, and manganese 0.72. For the causal edge from precipitation separation to solvent extraction, the corresponding causal coefficients were 0.75, 0.70, 0.58, 0.73, and 0.67, respectively. For the causal edge from solvent extraction to crystallization purification, the corresponding causal coefficients were 0.82, 0.75, 0.63, 0.78, and 0.74, respectively. For the causal edge directly from acid leaching to crystallization purification, the corresponding causal coefficients were 0.32, 0.25, 0.18, 0.30, and 0.27, respectively.

[0120] Based on the process state deviation metric, the causal effect values ​​of different separation stages were calculated using counterfactual inference. In the current batch, the process state deviation metric is 0.52. By creating a counterfactual scenario, fixing the process parameters of other stages, and adjusting the parameters of a single separation stage, the change in the process state deviation metric was calculated. The causal effect value of the acid leaching extraction stage is 0.28, indicating that optimizing the acid leaching extraction stage could reduce the process state deviation metric by 0.28; the causal effect value of the precipitation separation stage is 0.15, the causal effect value of the solvent extraction stage is 0.22, and the causal effect value of the crystallization purification stage is 0.12. The sum of the causal effect values ​​is 0.77, which is greater than the current process state deviation metric of 0.52, indicating that there are interactive effects between the stages, and further analysis of the pure causal contribution of each stage is needed.

[0121] The current process deviation metric is used as the overall deviation. Virtual intervention is performed on each causal node, and the pure causal contribution of each separation stage to the overall deviation is calculated under the condition of fixed potential contaminant node values. The ambient temperature is fixed at 22℃, humidity at 50%, raw material composition deviation at 8%, and raw material impurity deviation at 5%. The pure causal contribution of the acid leaching extraction stage is 0.23, the pure causal contribution of the precipitation separation stage is 0.12, the pure causal contribution of the solvent extraction stage is 0.18, and the pure causal contribution of the crystallization purification stage is 0.09. The sum of the pure causal contributions of the four stages is 0.62, which differs from the process deviation metric of 0.52 by 0.10. This difference is attributed to the direct influence of contaminants.

[0122] Based on the causal coefficient and causal effect value, the pure causal contribution was decomposed to each causal node to obtain the contribution components. For the acid leaching extraction stage, the pure causal contribution was 0.23. Decomposition according to the causal coefficient of each element yielded contribution components of 0.067, 0.062, 0.053, 0.065, and 0.058 for nickel, lithium, aluminum, cobalt, and manganese, respectively. For the precipitation separation stage, the pure causal contribution was 0.12. The decomposed contribution components were 0.032, 0.030, 0.025, 0.031, and 0.029. For the solvent extraction stage, the pure causal contribution was 0.18. The decomposed contribution components were 0.051, 0.047, 0.039, 0.049, and 0.046. For the crystallization purification stage, the pure causal contribution was 0.09. The decomposed contribution components were 0.025, 0.023, 0.019, 0.024, and 0.022.

[0123] A comprehensive impact index was calculated based on contribution components and causal effect values. The direct and indirect causal responsibility for purity non-compliance at each separation stage was calculated based on the deviation between the comprehensive impact index and product purity data. A total causal responsibility score was obtained through nonlinear combination. The target product purity was 99.2%, and the actual purity was 98.7%, resulting in a deviation of 0.5%. The direct causal responsibility for the acid leaching extraction stage was 0.35, the indirect causal responsibility was 0.18, and the total causal responsibility score was 0.43; for the precipitation separation stage, the direct causal responsibility was 0.20, the indirect causal responsibility was 0.12, and the total causal responsibility score was 0.26; for the solvent extraction stage, the direct causal responsibility was 0.28, the indirect causal responsibility was 0.10, and the total causal responsibility score was 0.32; and for the crystallization purification stage, the direct causal responsibility was 0.15, the indirect causal responsibility was 0.05, and the total causal responsibility score was 0.17.

[0124] Key separation stages were identified where the total causal responsibility score exceeded a preset responsibility threshold and the causal effect value exceeded a preset effect threshold. The preset responsibility threshold was set to 0.25, and the preset effect threshold to 0.20. The acid leaching extraction stage had a total causal responsibility score of 0.43, exceeding the preset responsibility threshold of 0.25, and a causal effect value of 0.28, exceeding the preset effect threshold of 0.20; therefore, the acid leaching extraction stage was a key separation stage. The solvent extraction stage had a total causal responsibility score of 0.32, exceeding the preset responsibility threshold of 0.25, and a causal effect value of 0.22, exceeding the preset effect threshold of 0.20; therefore, the solvent extraction stage was also a key separation stage. The precipitation separation stage had a total causal responsibility score of 0.26, slightly exceeding the preset responsibility threshold of 0.25, but a causal effect value of 0.15, not exceeding the preset effect threshold of 0.20; therefore, the precipitation separation stage was not a key separation stage. The crystallization purification stage had a total causal responsibility score of 0.17, below the preset responsibility threshold of 0.25, and a causal effect value of 0.12, below the preset effect threshold of 0.20; therefore, the crystallization purification stage was not a key separation stage.

[0125] In this embodiment, by constructing each separation stage as a causal node and identifying the transmission relationship between stages based on real-time component data, a causal structure between different process steps in the separation process is effectively established. Environmental conditions and raw material batches are introduced as potential contamination factors, enabling causal analysis to cover common external interference factors in real-world operations. By performing virtual intervention on each causal node, the pure causal contribution of each separation stage to the overall deviation under fixed contamination factors is identified. The contribution is then combined with the causal coefficient and causal effect value to decompose it into explainable contribution components, enhancing the transparency and traceability of the deviation source. By integrating the contribution components, causal effect value, and product purity deviation information to construct a comprehensive influence index, and calculating the total causal responsibility score based on direct and indirect influences, the key separation stages leading to substandard purity can be accurately identified.

[0126] In one alternative implementation,

[0127] Based on the combination of operating parameters corresponding to the key separation stage, the parameter adjustment amount is calculated and the loop control command is updated. This process is repeated until the product purity data is not less than the purity threshold, resulting in high-purity metallic materials, including:

[0128] Extract the current combination of operating parameters corresponding to the key separation stage, substitute the contribution component of the key separation stage and the pre-acquired causal effect value into the preset parameter response relationship, and calculate the range of change in operating parameters required to reduce the contribution component to the preset target contribution threshold to obtain the parameter adjustment amount of the key separation stage.

[0129] The parameter adjustment amount is superimposed on the current combination of operating parameters corresponding to the critical separation stage to obtain the updated combination of operating parameters. The combination of operating parameters for non-critical separation stages remains unchanged and is combined and encapsulated to generate an updated loop control instruction.

[0130] The loop control instruction is executed. After execution, the purity data of the current product is collected and compared with the purity threshold. If the purity data of the current product is not less than the purity threshold, the loop is terminated and the current product is output to obtain the high-purity metal material. Otherwise, the loop control instruction is executed repeatedly until the purity of the current product is not less than the purity threshold.

[0131] The current operating parameter combinations corresponding to the key separation stages are extracted. The contribution components of the key separation stages and the pre-acquired causal effect values ​​are substituted into the preset parameter response relationship to calculate the required change in operating parameters to reduce the contribution component to the preset target contribution threshold, thus obtaining the parameter adjustment amount for the key separation stages. For the identified key separation stages of acid leaching extraction and solvent extraction, the current operating parameter combinations are extracted. The current operating parameter combinations for the acid leaching extraction stage include: dissolution temperature 72℃, dissolution time 55 min, solution pH 3.8, solvent concentration 2.5 mol / L, and stirring speed 550 rpm; the current operating parameter combinations for the solvent extraction stage include: extraction temperature 65℃, extraction time 40 min, ratio 1:1.2, extractant concentration 0.8 mol / L, and oscillation frequency 120 times per minute. The contribution components of the acid leaching extraction stage were 0.067, 0.062, 0.053, 0.065, and 0.058, corresponding to the five elements nickel, lithium, aluminum, cobalt, and manganese, with a causal effect value of 0.28. The contribution components of the solvent extraction stage were 0.051, 0.047, 0.039, 0.049, and 0.046, with a causal effect value of 0.22. The preset target contribution threshold was 50% of the contribution components of each element, i.e., the target contribution components of the acid leaching extraction stage were 0.034, 0.031, 0.027, 0.033, and 0.029, and the target contribution components of the solvent extraction stage were 0.026, 0.024, 0.020, 0.025, and 0.023.

[0132] Based on historical separation data and experimental design, a pre-defined parameter response relationship was established to describe the relationship between changes in operating parameters and changes in contribution components. During the acid leaching extraction stage, for every 1°C increase in dissolution temperature, the contribution components of nickel, lithium, aluminum, cobalt, and manganese decreased by 0.008, 0.007, 0.006, and 0.008 respectively; for every 5 minutes increase in dissolution time, the contribution components of nickel, lithium, aluminum, cobalt, and manganese decreased by 0.012, 0.011, 0.009, 0.012, and 0.010 respectively; for every 0.2 increase in solution pH, the contribution components of nickel, lithium, and aluminum decreased by 0.010, 0.009, and 0.009 respectively. The contribution of nickel decreases by 0.008, the contribution of cobalt decreases by 0.010, and the contribution of manganese decreases by 0.009 for every 0.2 mol / L increase in solvent concentration; the contribution of nickel decreases by 0.015, the contribution of lithium decreases by 0.014, the contribution of aluminum decreases by 0.012, the contribution of cobalt decreases by 0.015, and the contribution of manganese decreases by 0.013 for every 50 rpm increase in stirring speed; the contribution of nickel decreases by 0.005, the contribution of lithium decreases by 0.004, the contribution of aluminum decreases by 0.003, the contribution of cobalt decreases by 0.005, and the contribution of manganese decreases by 0.004 for every 50 rpm increase in stirring speed.

[0133] Similarly, the parameter response relationships in the solvent extraction stage are as follows: for every 1°C increase in extraction temperature, the contribution of nickel decreases by 0.006, lithium by 0.005, aluminum by 0.004, cobalt by 0.006, and manganese by 0.005; for every 5 minutes increase in extraction time, the contribution of nickel decreases by 0.010, lithium by 0.009, aluminum by 0.007, cobalt by 0.010, and manganese by 0.008; compared to every 0.1 increase, the contribution of nickel decreases by 0.008, and the contribution of lithium by 0.00. 7. The contribution of aluminum decreases by 0.006, the contribution of cobalt decreases by 0.008, and the contribution of manganese decreases by 0.007; for every 0.1 mol / L increase in extractant concentration, the contribution of nickel decreases by 0.012, the contribution of lithium decreases by 0.011, the contribution of aluminum decreases by 0.009, the contribution of cobalt decreases by 0.012, and the contribution of manganese decreases by 0.010; for every 10 oscillations per minute increase in oscillation frequency, the contribution of nickel decreases by 0.004, the contribution of lithium decreases by 0.003, the contribution of aluminum decreases by 0.002, the contribution of cobalt decreases by 0.004, and the contribution of manganese decreases by 0.003.

[0134] Based on the aforementioned parameter response relationship, the required change in operating parameters to reduce the contribution component to the preset target contribution threshold was calculated. For the acid leaching extraction stage, to reduce the nickel element contribution component from 0.067 to 0.034, the required parameter combination is: increase the dissolution temperature by 4°C, increase the dissolution time by 10 minutes, increase the solution pH by 0.4, increase the solvent concentration by 0.3 mol / L, and increase the stirring speed by 50 rpm. For the solvent extraction stage, to reduce the nickel element contribution component from 0.051 to 0.026, the required parameter combination is: increase the extraction temperature by 4°C, increase the extraction time by 10 minutes, increase the pH by 0.2, increase the extractant concentration by 0.2 mol / L, and increase the oscillation frequency by 20 times per minute.

[0135] Taking into account the limitations of adjusting operating parameters and the requirements for process stability, the parameter adjustment amounts for the key separation stages were determined. The parameter adjustment amounts for the acid leaching extraction stage were: dissolution temperature increased by 3℃, dissolution time increased by 10 minutes, solution pH increased by 0.3, solvent concentration increased by 0.3 mol / L, and stirring speed increased by 50 rpm. The parameter adjustment amounts for the solvent extraction stage were: extraction temperature increased by 3℃, extraction time increased by 10 minutes, pH increased by 0.2, extractant concentration increased by 0.2 mol / L, and oscillation frequency increased by 20 times per minute.

[0136] The adjusted parameters are superimposed onto the current operating parameter combinations corresponding to the critical separation stages to obtain updated operating parameter combinations. The operating parameter combinations for non-critical separation stages remain unchanged, and these combinations are packaged to generate updated cyclic control commands. The updated operating parameter combinations for the acid leaching extraction stage are: dissolution temperature 75℃, dissolution time 65 minutes, solution pH 4.1, solvent concentration 2.8 mol / L, and stirring speed 600 rpm. The updated operating parameter combinations for the solvent extraction stage are: extraction temperature 68℃, extraction time 50 minutes, ratio 1:1.4, extractant concentration 1.0 mol / L, and oscillation frequency 140 times per minute. The operating parameters for the non-critical separation stages of precipitation separation and crystallization purification remain unchanged. The operating parameters for the precipitation separation stage are: precipitation temperature 60℃, precipitation time 30 minutes, precipitant concentration 1.5 mol / L, pH adjustment 5.0, and stirring speed 500 rpm. The operating parameters for the crystallization purification stage are: crystallization temperature 55℃, crystallization time 120 minutes, cooling rate 2℃ per minute, seed crystal addition 5%, and stirring speed 400 rpm. The updated four separate-stage operation parameters are combined and encapsulated to generate update loop control instructions.

[0137] The process executes a cyclic control command. After execution, the current product purity data is collected and compared with a purity threshold. If the current product purity data is not less than the purity threshold, the cycle terminates and the current product is output, yielding a high-purity metallic material. Otherwise, the cyclic control command is repeated until the current product purity is not less than the purity threshold. For example, after executing the first update cyclic control command, the measured purity of nickel is 99.1%, and the purity threshold is 99.2%. Since the current product purity is less than the purity threshold, the cyclic control command needs to be executed again. After repeating the second cyclic control command, the measured purity of nickel is 99.3%, which is higher than the purity threshold of 99.2%. The cycle terminates and the current product is output, yielding a high-purity nickel material. A total of two cyclic control commands were executed, with a total of eight cycles, which is two fewer than the predicted minimum of ten cycles, thus improving separation efficiency.

[0138] In this embodiment, by substituting the contribution component and causal effect value of the key separation stage into a preset parameter response relationship, the parameter adjustment amount required to reduce the contribution component to the target contribution threshold is calculated. This allows parameter updates to be directly optimized around the core cause of purity deviation, precisely focusing parameter adjustments on the key separation stage and avoiding unnecessary disturbances to non-key stages. This improves the targeting and effectiveness of the adjustments. By superimposing the calculated parameter adjustment amount onto the current operating parameter combination of the key separation stage while keeping the operating parameters of other stages stable, maximum deviation improvement can be achieved with minimal process changes. By encapsulating the updated parameter combination into a loop control command and collecting product purity data in real time after execution, and dynamically comparing it with the purity threshold, closed-loop adaptive control oriented towards the final quality target can be achieved.

[0139] Figure 2 This is a flowchart illustrating the optimization of purification operation parameters for a cyclic extraction and purification method for battery cathode materials based on automated control, as described in an embodiment of the present invention.

[0140] A second aspect of this invention provides a battery cathode material cyclic extraction and purification system based on automated control, comprising:

[0141] The separation analysis module is used to acquire the cathode material of the battery to be processed and the corresponding initial component data. It determines the optimal separation sequence of the metal components in the initial component data through a multi-objective optimization algorithm and performs a staged dissolution and separation operation. After each separation stage is completed, it acquires the real-time component data of the intermediate solution and compares it with the corresponding expected component concentration to obtain a comprehensive deviation vector.

[0142] The evaluation and optimization module is used to calculate the process state deviation metric between the comprehensive deviation vector and the expected component concentration through a state evaluation algorithm and determine the predicted minimum number of cycles. Based on the process state deviation metric and the predicted minimum number of cycles, it determines the correlation between the combination of operating parameters corresponding to each separation stage and the comprehensive deviation vector. Based on the correlation, it calculates the operating parameter correction value and the actual number of cycles and generates the corresponding cycle control command.

[0143] The cyclic purification module is used to execute the cyclic control command. After each cycle, it collects product purity data and compares it with a preset purity threshold. If the product purity data is less than the purity threshold, it uses a causal inference algorithm to decompose the deviation of the process state into the contribution component of each separation stage and identifies the key separation stage. Based on the combination of operating parameters corresponding to the key separation stage, it calculates the parameter adjustment amount and updates the cyclic control command. This process is repeated until the product purity data is not less than the purity threshold, thus obtaining high-purity metal material.

[0144] A third aspect of the present invention provides an electronic device, comprising:

[0145] A processor and a memory for storing processor-executable instructions, wherein the processor is configured to invoke instructions stored in the memory to perform the aforementioned method.

[0146] A fourth aspect of the present invention provides a computer-readable storage medium having stored thereon computer program instructions that, when executed by a processor, implement the aforementioned method.

[0147] This invention can be a method, apparatus, system, and / or computer program product. The computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for performing various aspects of the invention.

[0148] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for cyclic extraction and purification of battery cathode materials based on automated control, characterized in that, include: The cathode material of the battery to be processed and its corresponding initial component data are obtained. The optimal separation sequence of the metal components in the initial component data is determined by a multi-objective optimization algorithm, and a staged dissolution and separation operation is performed. After each separation stage is completed, the real-time component data of the intermediate solution is obtained and compared with the corresponding expected component concentration to obtain a comprehensive deviation vector. The process state deviation metric between the comprehensive deviation vector and the expected component concentration is calculated by the state evaluation algorithm, and the predicted minimum number of cycles is determined. Based on the process state deviation metric and the predicted minimum number of cycles, the correlation between the combination of operating parameters corresponding to each separation stage and the comprehensive deviation vector is determined. Based on the correlation, the operating parameter correction value and the actual number of cycle executions are calculated, and the corresponding cycle control command is generated. The cyclic control command is executed, and after each cycle, product purity data is collected and compared with a preset purity threshold. If the product purity data is less than the purity threshold, the process state deviation metric is decomposed into the contribution component of each separation stage through a causal inference algorithm, and the key separation stage is identified. Based on the combination of operating parameters corresponding to the key separation stage, the parameter adjustment amount is calculated and the cyclic control command is updated. The execution is repeated until the product purity data is not less than the purity threshold, and high-purity metal material is obtained.

2. The method according to claim 1, characterized in that, Acquiring the cathode material to be processed and its corresponding initial component data, determining the optimal separation sequence of metal components in the initial component data using a multi-objective optimization algorithm, and performing a staged dissolution and separation operation includes: The composition of the battery cathode material to be processed is analyzed to obtain the mass percentage of metal elements and the distribution characteristics of impurity elements to obtain the initial composition data. The standard electrode potential difference and solubility product constant of each metal element in the initial composition data are extracted and the separation feasibility coefficient between different metal elements is calculated. A separation feasibility matrix is ​​constructed based on the separation feasibility coefficient. Traverse the separation feasibility matrix, take the metal component pair with the maximum separation feasibility coefficient as the priority separation combination, take the position index of the priority separation combination in the separation feasibility matrix as the starting point, mark the metal component corresponding to the priority separation combination as the assigned metal and determine the unmarked metal component, and calculate the cumulative value of the mutual interference between the unmarked metal component and the assigned metal. With the optimization objectives of maximizing metal recovery rate and minimizing the number of separation stages, the cumulative value of mutual interference is used as a constraint. Multiple candidate extraction sequences that satisfy the optimization objectives are calculated to obtain a set of candidate schemes. The candidate scheme that minimizes the cumulative value of mutual interference is selected and the corresponding optimal separation sequence is determined. Based on the optimal separation sequence, the type of dissolution medium corresponding to each separation stage is matched, the dissolution temperature range and dissolution time range of each separation stage are determined, and the dissolution separation operation is performed.

3. The method according to claim 1, characterized in that, After each separation stage, real-time component data of the intermediate solution is acquired and compared with the corresponding expected component concentrations to obtain a comprehensive deviation vector, including: After each separation stage is completed, the measured concentration values ​​of each metal element and the measured concentration values ​​of impurity elements in the intermediate solution are obtained to obtain the real-time component data; Based on the target metal element in the current separation stage of the optimal separation sequence and the corresponding dissolution temperature range and dissolution time range, the expected concentration values ​​of the target metal element and impurity element in the current separation stage are calculated to obtain the expected component concentration. The measured concentration values ​​of each metal element in the real-time component data are compared with the expected concentration values ​​of the corresponding metal element in the expected component concentration to obtain the concentration deviation value of each metal element. According to the arrangement order of different metal elements in the optimal separation sequence, the concentration deviation values ​​of each metal element are combined to construct a concentration deviation vector. The difference between the measured concentration value of the impurity element in the real-time component data and the expected concentration value of the impurity element in the expected component concentration is calculated to obtain the concentration deviation value of the impurity element. The impurity residue index for the current separation stage is calculated based on the concentration deviation value of the impurity element, and the comprehensive deviation vector is obtained by combining the concentration deviation vector with the impurity residue index.

4. The method according to claim 1, characterized in that, The process state deviation metric between the comprehensive deviation vector and the expected component concentration is calculated using a state assessment algorithm, and the minimum predicted cycle number is determined, including: A multidimensional feature space is constructed by extracting a concentration deviation vector and an impurity residue index from the comprehensive deviation vector. The zero-deviation state corresponding to the expected concentration values ​​of each metal element and the expected concentration values ​​of impurities in the expected component concentration is taken as the target reference point in the multidimensional feature space. The Mahalanobis distance from the concentration deviation vector to the target reference point is calculated as the concentration deviation sub-value. The impurity deviation sub-value is calculated based on the distance from the impurity residue index to the impurity dimension corresponding to the target reference point. The concentration deviation sub-value and the impurity deviation sub-value are mapped to a preset deviation metric space. A dynamic covariance matrix is ​​constructed by combining the rate of change in the continuous separation stage. The process state deviation metric value is obtained by coupling and transforming the concentration deviation sub-value and the impurity deviation sub-value through the dynamic covariance matrix. Obtain the process state deviation metric and operating parameter combination in the historical separation stage and determine the influence relationship between the process state deviation metric and the operating parameter combination and the number of cycles. Substitute the process state deviation metric and the current operating parameter combination corresponding to the current separation stage into the influence relationship, and calculate the numerical sequence of the process state deviation metric converging to zero under different number of cycles in combination with the real-time component data. Identify the cycle number node that makes the process state deviation metric first decrease to the preset convergence threshold in the numerical sequence and determine the predicted minimum number of cycles.

5. The method according to claim 1, characterized in that, Based on the process state deviation metric and the predicted minimum number of cycles, the correlation between the combination of operating parameters for each separation stage and the comprehensive deviation vector is determined. Based on the correlation, the operating parameter correction value and the actual number of cycle executions are calculated, and the corresponding cycle control instructions are generated, including: Extract the combination of operating parameters and the comprehensive deviation vector of each separation stage from the historical separation data, and determine the degree of influence of the combination of operating parameters on each component of the comprehensive deviation vector. Quantify the degree of influence to obtain the parameter sensitivity vector. Construct a cycle efficiency index based on the numerical difference between the process state deviation metric and the predicted minimum number of cycles in the current separation stage. Construct a response tensor based on the parameter sensitivity vector and the cycle efficiency index, and identify the operating parameter adjustment path that maximizes the convergence rate of the comprehensive deviation vector to zero. Use the operating parameter adjustment path as the correlation relationship. Substitute the comprehensive deviation vector of the current separation stage into the correlation relationship, combine the real-time component data to calculate the operation parameter adjustment amount that reduces the comprehensive deviation vector to a preset deviation threshold, and add it to the operation parameter combination of the current separation stage to obtain the operation parameter correction value. The operation parameter correction value and the deviation metric value of the current separation stage are substituted into the correlation relationship. The decay rate of the current process state deviation metric value converging to zero is calculated by combining the real-time component data and matched with the preset deviation threshold to obtain the correction cycle number. The correction cycle number is compared with the predicted minimum cycle number, and the cycle number with the smaller value is taken as the actual cycle execution number. The operation parameter correction value and the actual cycle execution number are encapsulated to obtain the cycle control instruction.

6. The method according to claim 1, characterized in that, The process state deviation metric is decomposed into the contribution components of each separation stage using a causal inference algorithm, and key separation stages are identified, including: Each separation stage is taken as a causal node. Based on the transmission relationship between the real-time component data and each separation stage, the directed causal edges between different causal nodes are identified. The common influence of environmental conditions and raw material batches on multiple separation stages is obtained and potential confounding factor nodes are set. The causal strength is calculated for each causal edge and the causal coefficient is assigned. Based on the deviation metric of the process state, the causal effect value of different separation stages is calculated by counterfactual inference. The current process state deviation metric is used as the overall deviation to perform virtual intervention operations on each causal node. Under the condition of fixing the value of the potential confounding factor node, the pure causal contribution of each separation stage to the overall deviation is calculated. Based on the causal coefficient and the causal effect value, the pure causal contribution is decomposed to each causal node to obtain the contribution component. A comprehensive impact index is calculated based on the contribution components and the causal effect value. Based on the deviation between the comprehensive impact index and the product purity data, the direct and indirect causal responsibility for the purity non-compliance of each separation stage is calculated and the total causal responsibility score is obtained through nonlinear combination. The separation stages in which the total causal responsibility score exceeds a preset responsibility threshold and the causal effect value exceeds a preset effect threshold are identified as key separation stages.

7. The method according to claim 1, characterized in that, Based on the combination of operating parameters corresponding to the key separation stage, the parameter adjustment amount is calculated and the loop control command is updated. This process is repeated until the product purity data is not less than the purity threshold, resulting in high-purity metallic materials, including: Extract the current combination of operating parameters corresponding to the key separation stage, substitute the contribution component of the key separation stage and the pre-acquired causal effect value into the preset parameter response relationship, and calculate the range of change in operating parameters required to reduce the contribution component to the preset target contribution threshold to obtain the parameter adjustment amount of the key separation stage. The parameter adjustment amount is superimposed on the current combination of operating parameters corresponding to the critical separation stage to obtain the updated combination of operating parameters. The combination of operating parameters for non-critical separation stages remains unchanged and is combined and encapsulated to generate an updated loop control instruction. The loop control instruction is executed. After execution, the purity data of the current product is collected and compared with the purity threshold. If the purity data of the current product is not less than the purity threshold, the loop is terminated and the current product is output to obtain the high-purity metal material. Otherwise, the loop control instruction is executed repeatedly until the purity of the current product is not less than the purity threshold.

8. A battery cathode material cyclic extraction and purification system based on automated control, used to implement the method of any one of claims 1-7, characterized in that, include: The separation analysis module is used to acquire the cathode material of the battery to be processed and the corresponding initial component data. It determines the optimal separation sequence of the metal components in the initial component data through a multi-objective optimization algorithm and performs a staged dissolution and separation operation. After each separation stage is completed, it acquires the real-time component data of the intermediate solution and compares it with the corresponding expected component concentration to obtain a comprehensive deviation vector. The evaluation and optimization module is used to calculate the process state deviation metric between the comprehensive deviation vector and the expected component concentration through a state evaluation algorithm and determine the predicted minimum number of cycles. Based on the process state deviation metric and the predicted minimum number of cycles, it determines the correlation between the combination of operating parameters corresponding to each separation stage and the comprehensive deviation vector. Based on the correlation, it calculates the operating parameter correction value and the actual number of cycles and generates the corresponding cycle control command. The cyclic purification module is used to execute the cyclic control command. After each cycle, it collects product purity data and compares it with a preset purity threshold. If the product purity data is less than the purity threshold, it uses a causal inference algorithm to decompose the deviation of the process state into the contribution component of each separation stage and identifies the key separation stage. Based on the combination of operating parameters corresponding to the key separation stage, it calculates the parameter adjustment amount and updates the cyclic control command. This process is repeated until the product purity data is not less than the purity threshold, thus obtaining high-purity metal material.

9. An electronic device, characterized in that, include: processor; Memory used to store processor-executable instructions; The processor is configured to invoke instructions stored in the memory to execute the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having computer program instructions stored thereon, characterized in that, When the computer program instructions are executed by the processor, they implement the method described in any one of claims 1 to 7.

Images