A closed-loop optimization control system for waste lithium iron phosphate material recycling process

Abstract

The application relates to the technical field of general control or regulation systems, and discloses a closed-loop optimization control system for a waste lithium iron phosphate material recycling process, which comprises a reaction heat flow calculation module for generating a thermodynamic state trajectory representing a reaction progress, and a state consistency arbitration module for generating a rheological state trajectory representing a physical form of a slurry in parallel, and for performing real-time arbitration on the consistency of the two trajectories to determine the reliability of the thermodynamic state trajectory and to intervene in the operation of a regulation module accordingly. The application establishes a cross-verification mechanism originating from different physical dimensions, so that the control system can confirm the reliability of the basis for core control in real time before executing a regulation instruction, avoids making an incorrect regulation action when a single information source fails to operate, and thus improves the robustness and decision safety of the system.

Description

Technical Field

[0001] This invention relates to a closed-loop optimization control system for the recycling process of waste lithium iron phosphate materials, belonging to the general field of control or regulation system technology. Background Technology

[0002] Currently, in the wet recycling process of waste lithium iron phosphate materials, the process control system typically aims to maintain a constant temperature of the material inside the reactor. This isothermal reaction control method is based on the assumption that a stable temperature is a prerequisite for ensuring a uniform reaction rate and consistent final product quality. This method can be achieved using a mature jacketed temperature control system, which is straightforward and simple in engineering. However, in large-scale continuous production, the activity and physical morphology of the waste raw materials entering the reactor fluctuate between batches, causing the exothermic rate of the chemical reaction to change accordingly. In this case, to maintain a constant temperature inside the reactor, the control system must instruct the jacketed cooling system to intervene with varying power. The direct technical consequence is that the dynamic information of the reaction rate is masked by the compensatory adjustment action taken by the jacketed system to stabilize the temperature. As a result, the operating system cannot obtain the true chemical reaction process inside the reactor and can only passively discover deviations in quality or recovery rate at the end of the production process through product sampling and analysis.

[0003] To obtain the real-time status of the reaction process, one technical approach is to add a concentration sensor to the reactor to analyze the composition of the slurry. However, considering that wet recycling processes are usually in a strong acid and high solids content slurry environment, the probes of such online analytical instruments face engineering problems such as corrosion, scaling, and signal drift during long-term operation, making it difficult to meet the requirements of industrial sites in terms of maintenance costs and operational reliability. For example, Chinese invention patent CN107739830A discloses a method for recycling waste lithium iron phosphate battery cathode materials. Its core technology lies in precisely controlling the pH value at the end of acid leaching, so that iron precipitates in the form of iron phosphate, while lithium remains in the solution, thereby achieving the separation of the two. The control logic of this method is essentially an open-loop chemical condition control based on a preset formula (such as a fixed molar ratio and reaction time). This prescription-based control strategy cannot perceive and respond to the real reaction process in the reactor in real time. Especially when faced with differences in activity between different batches of waste materials, its fixed control parameters cannot guarantee that each batch can achieve the optimal separation effect, thus limiting the further improvement of recycling efficiency and product quality.

[0004] This presents a technical dilemma: precise process control requires real-time and accurate reaction rate information, but directly measuring this information presents engineering challenges. Furthermore, the existing isostatic control methods inherently prevent the acquisition of this crucial information. Therefore, the technical problem this invention aims to solve is to design a control method utilizing existing industrial measurement and control components to calculate dynamic signals that directly characterize the core reaction process from the system's own operating parameters, and to construct a closed-loop regulation based on these signals to achieve proactive and process-oriented optimization control of the chemical reaction process. Summary of the Invention

[0005] This invention provides a closed-loop optimization control system for the recycling process of waste lithium iron phosphate materials. Its main purpose is to solve the problem that the existing constant temperature control method will mask the core reaction process information when maintaining the apparent temperature stability, while the method of adding a dedicated sensor has difficulties in engineering application.

[0006] To achieve the above objectives, the present invention provides a closed-loop optimization control system for the recycling process of waste lithium iron phosphate materials, the system comprising:

[0007] A reaction heat flux calculation module is configured to calculate the reaction heat flux signal based on the operating parameters of the reactor jacket temperature control system associated with the recovery process, and to perform time integration on the reaction heat flux signal to generate a first state trajectory, i.e., a thermodynamic state trajectory, characterizing the extent to which the recovery process has been carried out.

[0008] A state consistency arbitration module is configured to generate a second state trajectory, namely a rheological state trajectory, in parallel based on the real-time load current of the stirring motor of the stirring device associated with the recycling process and according to a benchmark model that defines the relationship between the real-time load current and the solid content of the material in the recycling process. In each control cycle, the thermodynamic state trajectory and the rheological state trajectory are compared. When the deviation between the two trajectories exceeds a preset threshold, the first state trajectory is determined to be unreliable.

[0009] An adjustment module is configured to generate a control signal for the feed actuator based on a deviation signal generated by comparing the thermodynamic state trajectory with a preset optimal trajectory in an optimal trajectory matching module; and the operation of the adjustment module is independently constrained by the state consistency arbitration module, whose operation rule is that once a failure judgment is received, the control authorization based on the optimal trajectory is revoked and a preset conservative control mode is switched.

[0010] Preferably, the reaction heat flow calculation module is configured to calculate the operating parameters of the jacket temperature control system for calculating the reaction heat flow signal, including the inlet temperature of the coolant entering the jacket, the outlet temperature of the coolant exiting the jacket, and the real-time circulation flow rate of the coolant in the jacket; the reaction heat flow calculation module is further configured to calculate the reaction heat flow signal based on the first law of thermodynamics, using the inlet temperature, outlet temperature, and real-time circulation flow rate.

[0011] Preferably, the optimal trajectory matching module is configured to have a built-in trajectory library storing multiple optimal trajectories; and the optimal trajectory matching module is configured to select an optimal trajectory from the trajectory library as the preset optimal trajectory before the recycling process begins, based on the initial information related to the characteristics of the waste lithium iron phosphate material to be processed.

[0012] Preferably, the reaction heat flux calculation module is configured to calculate the reaction heat flux signal using the following formula. : ,in, For real-time circulating traffic, The preset specific heat capacity of the coolant. Inlet temperature, For the outlet temperature, and The heat dissipation power of the reactor to the environment is calculated based on a preset offline calibrated function relationship.

[0013] Preferably, the system further includes a heat flux texture analysis module, configured to continuously receive the reaction heat flux signal and process the reaction heat flux signal through a digital high-pass filter to separate the high-frequency fluctuation component characterizing the fluctuation characteristics of the reaction heat flux signal; calculate the statistical characteristics of the high-frequency fluctuation component within a preset time window to extract at least one texture feature parameter, the texture feature parameter including the variance of the high-frequency fluctuation component and the dominant frequency obtained by fast Fourier transform; and diagnose the real-time mixing state of the recycling process based on the texture feature parameter and a preset knowledge base that defines the correspondence between the texture feature parameter and the mixing state.

[0014] Preferably, the system further includes an adaptive stirring optimization module, configured to generate a control signal for adjusting the operating state of the stirring device based on the real-time mixing state diagnosed by the thermal flow texture analysis module; wherein, when the real-time mixing state is diagnosed as having a material deposition zone, the control signal is configured to instruct the stirring device to execute a pulse cleaning program, the pulse cleaning program including periodically increasing and decreasing the stirring speed of the stirring device within a preset time.

[0015] Preferably, the system is further configured to receive real-time redox potential signals from a redox potential sensor associated with the recovery process; the preset optimal trajectory is an optimal state path defined in a two-dimensional state space with reaction heat flux and redox potential as coordinate axes; and the adjustment module is further configured to generate and output a second control signal to an actuator for replenishing oxidant to the recovery process based on the real-time redox potential signal and the deviation of the real-time state of the recovery process from the optimal state path in the two-dimensional state space.

[0016] Preferably, the system further includes a dynamic endpoint determination module, configured to calculate the time derivative of the reaction heat flow signal, i.e., the reaction acceleration, in real time using a digital differential algorithm; and when the absolute value of the reaction acceleration is consistently lower than a preset threshold within a continuous time window, determine that the recovery process has reached the dynamic endpoint, and generate a termination signal accordingly to terminate the addition of materials to the recovery process.

[0017] Preferably, the system also includes a cross-batch learning module, which is configured to, based on the complete historical reaction heat flow signal data of a batch and using a recursive least squares algorithm, adjust one or more model parameters of a preset optimal trajectory associated with the material type of the batch after the recycling process of a batch is determined to be completed by the dynamic endpoint determination module.

[0018] Preferably, the system further includes a heat transfer efficiency self-calibration module, configured to: at a preset diagnostic time, control the jacket temperature control system to apply a standardized thermodynamic disturbance to the reactor; capture and analyze the dynamic response characteristics of the reaction heat flow signal output by the reaction heat flow calculation module in response to the thermodynamic disturbance, the dynamic response characteristics including the peak time of the response and the decay half-life of the signal; and based on the dynamic response characteristics, and according to a preset surrogate model describing the relationship between the total heat transfer coefficient and the dynamic response characteristics, deduce a true heat transfer efficiency parameter, and use the true heat transfer efficiency parameter to calibrate the reaction heat flow calculation module.

[0019] Compared with the prior art, the beneficial effects of the present invention are:

[0020] 1. By synchronously acquiring the parameters of the jacket temperature control system used to calculate the reaction heat flux and the operating parameters of the stirring device used to characterize the physical morphology of the slurry, two state evolution trajectories originating from different physical dimensions were established. During the control and regulation process, the system continuously compares the consistency between the thermodynamic trajectory and the rheological trajectory. This parallel cross-validation method enables the control system to arbitrate and confirm the reliability of the core control basis, namely the reaction heat flux signal itself, in real time before executing the regulation command. This avoids the control system making erroneous adjustment actions when a single information source experiences silent drift or failure.

[0021] 2. The adjustment method of this application is based on the deviation between the real-time heat flow signal and a preset optimal heat flow trajectory. In this process, the system also analyzes the time series of the real-time heat flow signal and separates the high-frequency part that characterizes the signal fluctuation characteristics. The system further identifies the characteristics of the high-frequency fluctuation part and generates an adjustment signal for the stirring device based on the correspondence between the characteristics and the mixing state of the space inside the reactor. This method of combining the process signal of the main control loop with the signal texture of the auxiliary diagnostic loop makes the adjustment of the reaction process no longer just follow a target rate of the whole reactor average, but can be carried out in a relatively uniform physical environment that has been diagnosed and optimized in real time, ensuring the execution effect of the optimal control command.

[0022] 3. While receiving the reaction heat flux signal, the system also receives real-time potential signals from the redox potential sensor. Its preset optimal trajectory is a path defined in a two-dimensional state space with reaction heat flux and redox potential as coordinate axes. During operation, the control module calculates the spatial deviation vector between the current real-time state point and this optimal path. Based on the tangential and normal components of this deviation vector, it generates control signals for the two actuators for material replenishment and oxidant replenishment, respectively. This state-space vector deviation-based control mechanism transforms the control behavior from one-dimensional rate adjustment to two-dimensional coordinated regulation of rate and reaction selectivity. The system can automatically determine the root cause of process imbalance based on the direction of deviation. It also performs more targeted adjustment actions; while adjusting based on the heat flux deviation signal, it also performs pattern recognition on the time series of the heat flux deviation signal to determine the systematic deviation pattern between the preset optimal heat flux trajectory and the actual progress of the recovery process; the system adjusts the model parameters used to generate the subsequent optimal heat flux trajectory online according to the identified deviation pattern; this mechanism transforms the deviation signal generated in the control process from an instantaneous quantity used only for feedback adjustment into an information byproduct that can be used to correct the control model, enabling the control model to self-optimize and adjust according to the actual response during the execution of a single batch of tasks, and giving the entire control system the ability to adapt to the process when facing materials with unknown characteristics. Attached Figure Description

[0023] Figure 1 This is a diagram of the closed-loop optimization control system architecture with cross-validation mechanism of the present invention.

[0024] Figure 2 This is a comparative test diagram of the robustness of the system of the present invention under simulated silent failure conditions;

[0025] Figure 3 This is a module interaction sequence diagram of the online self-calibration process for heat transfer efficiency of the present invention. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below. Obviously, the described embodiments are only some embodiments of this invention, not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0027] This invention provides a closed-loop optimization control system for the recycling process of waste lithium iron phosphate materials. The overall architecture of the system mainly consists of a reaction heat flow calculation module, a state consistency arbitration module, and a regulation module. The reaction heat flow calculation module generates a first state trajectory, i.e., a thermodynamic state trajectory, that characterizes the progress of the recycling process. The state consistency arbitration module generates a second state trajectory in parallel, i.e., a rheological state trajectory, and performs real-time comparison and arbitration of the consistency between the two trajectories. Upon receiving control authorization from the arbitration module, the regulation module generates control signals for the process actuators based on the deviation between the thermodynamic state trajectory and the preset optimal trajectory, thereby constituting a closed-loop optimization control system. A closed-loop optimization control system with cross-validation capability; In the wet recycling process of waste lithium iron phosphate materials, the reaction is usually carried out in a reactor equipped with a jacketed temperature control system. However, the highly acidic, high-solids-content slurry environment poses a severe challenge of corrosion and scaling to online analytical instruments that directly measure reactant concentrations, making it difficult for the control system to obtain real-time reaction progress information. To address this challenge, the reaction heat flow calculation module in this invention is configured to use the operating parameters of the existing jacketed temperature control system used to regulate the reactor temperature to calculate the reaction heat flow signal. The data input terminal of this module is coupled to three standard industrial sensors of the jacketed temperature control system, which are used to collect the inlet temperature of the coolant entering the jacket. The outlet temperature of the coolant flowing out of the jacket and the real-time circulation flow rate of coolant within the jacket. The module internally contains a calculation procedure based on the first law of thermodynamics, and its core logic is expressed by the following formula. Execution, among which, The final output of the module is the reaction heat flow signal, which represents the total heat release power of the chemical reaction in the reactor at the current moment, and its unit is kW; This is the preset specific heat capacity of the coolant; for coolants such as water, this value is a known physical constant. The heat dissipation power from the reactor to the environment is calculated using a pre-defined offline calibration function. The calibration procedure is as follows: an inert medium is added to the reactor and maintained at a series of different constant temperatures. At each temperature, a known, constant heating power is applied through the jacket heating system. After the system stabilizes, the power required by the jacket cooling system to maintain that temperature is recorded. The difference between the two is the heat dissipation power from the reactor to the environment at that temperature. By fitting the heat dissipation power data at multiple temperature points, the heat dissipation power can be obtained. The functional relationship with the temperature inside the vessel; for example, when the inlet temperature is collected. It is 68.5 outlet temperature It is 70.0 Real-time circulating flow 10 (Approximately 2.78 kg / s), specific heat capacity of coolant It is 4.2 kJ / (kg· Furthermore, the ambient heat dissipation power is obtained by looking up a table based on the current reactor temperature. At 5kW, the module calculates the current reaction heat flux. for kW; The reaction heat flux calculation module calculates the reaction heat flux signal. Perform time integration, i.e., calculate This generates a first state trajectory, or thermodynamic state trajectory, representing the cumulative amount of reacted material in the recovery process. In this way, the module reuses the jacket temperature control system from a temperature regulating actuator into a device for acquiring reaction rate information, providing the core process observation for the entire control system.

[0028] Given that the accuracy of calculating the reaction heat flux signal depends on the measurement precision of sensors such as temperature and flow rate, any quiescent drift of a single sensor or changes in the heat transfer coefficient due to pipe scaling can introduce systematic errors, posing a risk to the safety of control decisions. To establish a cross-validation mechanism based on different physical principles, this system includes a state consistency arbitration module that generates a second state trajectory in parallel to corroborate the reliability of the thermodynamic state trajectory. This module utilizes the inherent law that chemical reaction processes are accompanied by changes in the physical morphology of the slurry. As the solid-phase lithium iron phosphate material dissolves, the overall solid content of the slurry decreases deterministically, causing changes in its rheological properties, which are then reflected in the load of the stirring device used to maintain the mixing state of the slurry. Specifically, the data input terminal of this module is configured to receive the real-time load current from the stirring motor of the stirring device associated with the recycling process. The module internally uses a pre-defined benchmark model that defines the relationship between real-time load current and the solid content of materials in the recovery process. This benchmark model is established through a standardized offline calibration procedure. This procedure involves preparing a series of simulated slurries with different known solid contents, similar to the actual process system, in the reactor. Under a constant stirring speed, the stable stirring motor load current value corresponding to each solid content is recorded, thereby establishing a current-solid content lookup table or piecewise linear function model. During system operation, the module uses this model to process the real-time acquired load current signal. The apparent solid content of the material is calculated by inversion and converted into a second state trajectory, namely the rheological state trajectory, which characterizes the extent to which the recovery process has been carried out. In each control cycle, the state consistency arbitration module compares the thermodynamic state trajectory with the rheological state trajectory. For example, after normalizing the two trajectories, the absolute value of their deviation at the current moment is calculated. When the deviation continues to exceed a preset threshold, for example, if the deviation is greater than 5% for three consecutive control cycles, the first state trajectory is determined to be unreliable, and an intervention signal is sent to the regulation module. Through this comparison of information from two different physical dimensions, thermodynamics and rheology, this module provides credibility confirmation for the entire control system before decision-making.

[0029] To achieve optimized control of the recycling process, the system's adjustment module is configured to guide the entire reaction process along a preset optimal trajectory. Specifically, the system has a built-in optimal trajectory matching module, which stores a trajectory library containing multiple optimal trajectories. Each trajectory is a curve showing the change of reaction heat flux over time, corresponding to different production targets. Before the recycling process begins, based on the initial information related to the characteristics of the waste lithium iron phosphate material to be processed, an optimal trajectory is selected from the trajectory library as the control benchmark for the current batch. Within a control cycle, the optimal trajectory matching module compares the real-time heat flux value output by the reaction heat flux calculation module with the target value on the optimal trajectory at the current moment, generating a deviation signal. The adjustment module, based on the deviation signal, generates control signals for the feed actuators through a proportional-integral-derivative (PID) controller. It should be noted that the operation of the adjustment module is independently constrained by the state consistency arbitration module. Its operating rule is as follows: under normal operating conditions, i.e., upon receiving a perceived reliable authorization signal from the arbitration module, it performs closed-loop adjustment based on the optimal trajectory matching result; however, once a failure judgment is received, it revokes the control authorization based on the optimal trajectory and switches to a preset conservative control mode, for example, fixing the output of all feed actuators to a historically verified safe, lower constant value. This control architecture allows the system to perform optimized adjustments while also ensuring the safety of decision-making.

[0030] To further enhance the ability to control reaction selectivity, the system of this invention is also configured to synchronously receive real-time redox potential signals from an industrial-grade redox potential (ORP) sensor associated with the recovery process. In this configuration, the preset optimal trajectory in the optimal trajectory matching module is defined as an optimal state path in a two-dimensional state space with reaction heat flux as one axis and redox potential as the other. This path is established by monitoring a series of recovery reactions aimed at obtaining high-purity products under laboratory conditions, recording data points on the co-evolution of reaction heat flux and redox potential over time, and then fitting these data points. Correspondingly, the adjustment module is further configured as a two-dimensional vector controller, whose operating procedure within one control cycle is as follows: First, the real-time collected reaction... The heat flux signal and the redox potential signal are combined to form a current state point, and the deviation vector between this point and the nearest target point on the two-dimensional optimal state path is calculated. Subsequently, the projection components of this deviation vector onto the tangential and normal directions of the path at the target point are calculated, where the tangential component represents the deviation in the overall reaction rate, and the normal component represents the deviation in reaction selectivity. The regulation module integrates two independent PID controllers. The first PID controller takes the tangential component as input to generate a first control signal for adjusting the material replenishment rate, while the second PID controller takes the normal component as input to generate and output a second control signal to the actuator for replenishing oxidant in the recovery process. In this way, the control behavior is transformed from one-dimensional rate regulation to two-dimensional coordinated regulation of reaction rate and reaction selectivity.

[0031] Example 1: In an industrial production scenario for continuous wet recycling of waste lithium iron phosphate materials, the closed-loop optimization control system of this invention operates stably. The reactor jacket temperature control system, stirring device, and feeding actuator are all under the automatic adjustment of this system. At the initial stage of a certain production batch, the system selects an optimal trajectory from the internal trajectory library based on the characteristics of the material to be processed. The adjustment module then controls the acid replenishment rate based on the deviation between the real-time heat flow signal output by the reaction heat flow calculation module and the optimal trajectory, so that the thermodynamic state trajectory characterizing the reaction process tracks the preset target. When the production batch reaches the middle stage, a change occurs. Due to fluctuations in the cooling circulating water quality, a layer of scale begins to slowly form on the inner wall of the reactor jacket. This phenomenon causes a slight but continuous decrease in the total heat transfer efficiency of the reactor. For the reaction heat flow calculation module, which relies on heat balance for calculation, this change causes a systematic deviation in its calculation results. Even if the exothermic rate of the chemical reaction inside the reactor does not change, the heat carried away by the cooling water is reduced due to the decrease in heat transfer efficiency, resulting in a decrease in the temperature difference between the inlet and outlet of the jacket collected by the sensor. The temperature is lowered, and therefore, the reaction heat flux calculation module calculates the reaction heat flux signal based on this temperature difference. The reaction rate begins to systematically drop below the heat flux value inside the reactor. At this point, if the control system relies solely on this thermodynamic information source, it will determine that the current reaction rate is below the target value. The regulating module will then instruct the feed actuator to increase the acid replenishment rate. The consequence is that the reaction rate inside the reactor is excessively increased, which may lead to local overheating and impurity leaching due to reduced reaction selectivity.

[0032] In the control system architecture of this invention, the state consistency arbitration module operates in parallel. While this operating condition occurs, the module continuously collects the real-time load current of the stirring motor. The system generates a rheological state trajectory based on a preset benchmark model. Because the adjustment module increases the replenishment rate based on a lower reaction heat flux signal, the actual reaction process inside the reactor accelerates, leading to a higher dissolution rate of solid-phase lithium iron phosphate. This results in a faster decrease in the solid content of the slurry compared to normal conditions. This physical change is monitored by the load current of the stirring device, which shows a faster decreasing trend than under normal conditions. Therefore, the rheological state trajectory generated by the state consistency arbitration module reflects the acceleration of the reaction process. At this point, within this module, two state trajectories originating from different physical dimensions bifurcate. The thermodynamic state trajectory generated by the integrated heat flux shows a lag in the reaction process, while the rheological state trajectory generated by the motor current... The trajectory shows that the reaction process is ahead of schedule; when the deviation between these two trajectories continues to exceed a preset threshold, the state consistency arbitration module determines that the first state trajectory has failed to keep track; this failure determination signal is sent to the adjustment module, triggering a switch in its operating rules; the adjustment module revokes the control authorization based on the optimal trajectory and switches to a preset conservative control mode, reducing the output rate of the feed actuator to a known safety lower limit, while simultaneously sending an alarm to the central control room, prompting an inspection of the heat transfer system; through the collaborative work of the reaction heat flow calculation module and the state consistency arbitration module, when faced with unknown disturbances in the core sensing information, the system prevents adjustment actions based on erroneous information through a verification mechanism based on different physical principles.

[0033] Example 2: To objectively verify the robustness of the closed-loop optimization control system of the present invention when faced with silent failure of core sensor information, the following comparative experiment was designed and executed; the experiment used a 50L standard laboratory reactor equipped with a jacketed temperature control system with a temperature control accuracy of ±0.5℃. The system includes: an adjustable-speed stirring device with a motor controller capable of outputting real-time load current with an accuracy of 0.01A at a frequency of 10Hz; a programmable logic controller (PLC)-based control system; and a jacket temperature control system equipped with inlet and outlet temperature sensors with a measurement accuracy of ±0.1A. The experiment included an electromagnetic flowmeter with a measurement accuracy of ±2% of full scale; the test material was waste lithium iron phosphate cathode powder from the same batch, which was pre-mixed with acid to form a slurry with an initial solid content of 30% before being added to the reactor; the experiment included a control group and the sample group of the present invention, both of which used the same initial slurry and a reaction temperature set at 70°C. The experiment included a preset optimal heat flux trajectory to guide the reaction process; the control system of the control group was configured to perform closed-loop regulation solely based on the thermodynamic state trajectory output by the reaction heat flux calculation module; the experimental group of this invention employed a complete control system, including a reaction heat flux calculation module, a state consistency arbitration module, and an adjustment module constrained by the arbitration module; both groups of experiments ran stably for 30 minutes in the initial stage; at the 30-minute mark, to simulate the sensor silent failure scenario caused by heat transfer deterioration, the inlet temperature of the reaction heat flux calculation module in both groups of experiments was monitored via control software. The reading was artificially introduced to linearly increase to -0.5 within 5 minutes. And by maintaining a stable negative bias, this will affect the system's calculated reaction heat flux signal. The heat flow is systematically lower than the actual reaction heat flow inside the reactor.

[0034] During the experiment, the control system of the control group, upon receiving a reaction heat flux signal that was systematically lower due to the bias, determined that the actual reaction rate was below the target value of the optimal trajectory. Its adjustment module then continuously increased the acid replenishment rate of the feed actuator. At 60 minutes into the experiment, the temperature inside the reactor overshooted, and the gas escape rate accelerated. The control system of the sample group of this invention, whose reaction heat flux calculation module was also affected by this bias, generated a delayed thermodynamic state trajectory. However, the rheological state trajectory generated by its state consistency arbitration module, because its data originated from the real-time load current of the unaffected stirring motor, reflected the effects of excessive feed. The actual reaction process was accelerated by the degree of [unclear]. At the 38th minute, the system detected that the normalized deviation of the two state trajectories continuously exceeded the preset threshold of 5%. The state consistency arbitration module determined that the thermodynamic state trajectory had lost credibility and sent an intervention signal to the regulation module. The regulation module then revoked the control authorization based on the optimal trajectory and switched to conservative control mode, forcibly locking the acid replenishment rate at a lower safety value. The two groups of experiments were terminated after 90 minutes. The sampling and analysis of the slurry at the reaction endpoint showed that the lithium recovery rate of the control group was 4.7% lower than that of the sample group of this invention, and the concentration of impurity ions in the leachate was more than 30% higher.

[0035] Table 1: Comparison of key data nodes in the experiment.

[0036]

[0037] Experimental data (see Table 1) show that the control system of the present invention, by establishing a state cross-verification mechanism of two different physical dimensions of thermodynamics and rheology, can identify the inherent logical contradictions in the system state perception when the single information source serving as the core control basis fails silently, and execute the preset safety procedures, thereby avoiding making adjustment actions based on erroneous information.

[0038] Example 3: This example combines Figures 1 to 3 This document describes a closed-loop optimization control system for the recycling process of waste lithium iron phosphate materials, such as... Figure 1 As shown, this module generates a thermodynamic state trajectory characterizing the reaction process based on the first law of thermodynamics. This thermodynamic state trajectory is sent to the optimal trajectory matching module to be compared with the preset optimal trajectory to generate a deviation signal. On the other hand, it is cross-validated with a rheological state trajectory generated in parallel by a state consistency arbitration module. The rheological state trajectory originates from the real-time load current of the stirring motor collected by the stirring device. The state consistency arbitration module outputs a failure judgment or intervention signal to the adjustment module based on the comparison result of the two trajectories. After comprehensively considering the deviation signal and the arbitration result, the adjustment module generates a control command and sends it to the feeding actuator. The actuator receives the control signal to adjust the material feeding rate, thus forming a closed-loop control loop with a dual verification mechanism.

[0039] like Figure 2 As shown in the figure, the horizontal axis represents time (minutes), and the vertical axis represents the reaction heat flux (kW). The three curves represent the actual heat flux trajectory of the control group, the actual heat flux trajectory of the present invention's sample group, and the optimal trajectory target as the control benchmark, respectively. It can be seen from the figure that after the fault simulation occurred, the reaction heat flux of the control group deviated from and remained below the optimal trajectory target. However, the present invention's sample group, due to its internal state consistency arbitration mechanism, intervened promptly, allowing its reaction heat flux to quickly adjust and more closely track the preset optimal trajectory target after a brief deviation. Figure 3 As shown, the process is triggered by the self-calibration module at a preset diagnostic timing. First, the calibration program is activated and the jacket temperature control system is instructed to apply a standardized thermodynamic disturbance to the electric heating system, for example, heating continuously at a constant power of 2kW for 30 seconds. During this disturbance, the jacket temperature control system sends real-time temperature and flow data to the reaction heat flux calculation module. The latter calculates the reaction heat flux signal based on this and forwards it to the self-calibration module as dynamic response process data. The self-calibration module analyzes the morphological characteristics of the response curve, extracts key parameters such as peak time and half-life, and uses these parameters to query a preset surrogate model to deduce a real heat transfer efficiency parameter. Finally, the heat dissipation power calculation function in the reaction heat flux calculation module is updated with this parameter.

[0040] Example 4: Before the closed-loop optimization control system of this invention is put into operation, a standardized offline calibration and parameter setting procedure needs to be performed. First, the benchmark model in the state consistency arbitration module is calibrated. The function of this model is to establish a quantitative correspondence between the real-time load current of the stirring motor and the solid content of the slurry material in the reactor. The calibration process is carried out in a 50L reactor with the same geometry and stirrer configuration as the actual production equipment. The same liquid medium as the acid system used in the actual process is added to the reactor in advance. The initial material is standard waste lithium iron phosphate powder that has been dried and screened. The calibration steps are as follows: First, a slurry with an initial solid content of 30% is prepared in the reactor, the stirring device is started and its speed is stabilized at the production speed. After running at 150 rpm for 5 minutes, the real-time load current value of the stirring motor was recorded. Subsequently, acid medium was added quantitatively to the reactor to dilute the solid content of the slurry to 25%, and the constant speed stirring and current recording steps were repeated. In this way, load current data were collected for six concentration gradients of solid content: 20%, 15%, 10%, and 5%, thereby obtaining a set of solid content-current correspondence datasets. The dataset was fitted with a second-order polynomial using the least squares method to obtain a mathematical expression describing the functional relationship between solid content and motor load current in this specific system. This expression was then stored in the memory of the state consistency arbitration module as the benchmark model on which it operates.

[0041] After establishing the baseline model, a preset threshold needs to be set for the state consistency arbitration module to determine the first state trajectory failure. This threshold is used to distinguish between normal measurement noise fluctuations and systematic deviations caused by sensor drift or process anomalies. The setting procedure is as follows: using a batch of standard materials, the recovery process is run once under normal operating conditions without any human error. During this period, the control system of this invention is set to a data recording mode, that is, both the reaction heat flow calculation module and the state consistency arbitration module run normally to generate their respective state trajectories, but the intervention function of the adjustment module is disabled. The system continuously records and calculates the real-time percentage deviation between the normalized thermodynamic state trajectory and the rheological state trajectory at a frequency of 1Hz throughout the entire reaction cycle, forming a deviation time series database. After the reaction is completed, all deviation values ​​in the database are statistically analyzed, and their standard deviation is calculated. To balance detection sensitivity and false alarm rate, the preset threshold for determining dishonesty is set to [value missing]. For example, if the calculated standard deviation is 1.2%, the threshold is set to 3.6%. At the same time, the logic condition for triggering the judgment is set to the time when the absolute value of the deviation between the two trajectories exceeds 3.6% for three consecutive control cycles. This quantified judgment rule is written into the logic controller of the state consistency arbitration module.

[0042] To enable the control module to perform control based on the optimal trajectory, it is necessary to establish content for the trajectory library of the optimal trajectory matching module. Taking the establishment of a high-purity trajectory aimed at obtaining higher product purity as an example, the establishment process is completed through a series of comparative experiments. Using the same test platform and materials as the aforementioned calibration process, three independent recovery reaction experiments were designed and executed. In these three sets of experiments, the control system's control module was configured to respectively transfer the reaction heat flux signal... The power was maintained at three different levels: 15 kW for group 1, 20 kW for group 2, and 25 kW for group 3. All three experiments were run until the reaction naturally terminated, meaning the heat flux signal decayed to less than 5% of its initial peak value. After the experiments, the chemical composition of the leachate from the final products of the three groups was analyzed to detect the concentration of impurity ions. The analysis results showed that at a constant heat flux of 15 kW, the reaction time was longer but the product purity was higher; at a constant heat flux of 25 kW, the reaction time was shortest but the amount of impurities leached was the highest, and the product purity was the lowest; while at a constant heat flux of 20 kW, a balance was reached between reaction time and product purity. Therefore, a heat flux curve with 20 kW as the main constant target value was determined as an effective high-purity trajectory and stored in the trajectory library of the optimal trajectory matching module for subsequent production use.

[0043] Example 5: To maintain the stability of the measurement benchmark during long-term continuous operation, the closed-loop optimization control system of this invention includes a heat transfer efficiency self-calibration module. This module is configured to execute an online calibration procedure at preset diagnostic times, such as after processing 100 batches of materials or during production intervals. In one calibration scenario, after the reactor completes the discharge and cleaning of the previous batch, it is filled with acidic medium in a static and thermal equilibrium state for the next batch of production. At this time, the heat transfer efficiency self-calibration module is activated and controls the jacket temperature control system to apply a standardized thermodynamic disturbance to the reactor. This disturbance is a 30-second thermal pulse applied by the electric heating system in the jacket at a constant power of 2kW. During this period, the reaction heat flow calculation module continues to run, and its output reaction heat flow signal is the dynamic response process of the system to this external heat injection.

[0044] The heat transfer efficiency self-calibration module performs morphological feature analysis on the output curve of the dynamic response process and extracts two dynamic response features through algorithms: the time required for the response signal to rise from the initial point to the peak value. And the half-life required for the signal to decay from its peak value to half its value. The surrogate model is established through an offline calibration procedure. The calibration process is carried out in a clean reactor. First, a standardized thermodynamic perturbation is performed on the clean reactor, and its dynamic response characteristics are recorded. , Using this as a reference point, a simulated scale layer with a known thermal conductivity and thickness is uniformly coated onto the inner wall of the reactor jacket. This scale layer introduces a known additional thermal resistance. Under this condition, the above-mentioned thermodynamic disturbance and dynamic response feature extraction steps are repeated to obtain the corresponding ( ) under this thermal resistance. , Data pairs; by changing the thickness of the simulated scale layer, the above steps were repeated to obtain a series of different total heat transfer efficiency parameters. , The data points are organized into a lookup table and embedded in the heat transfer performance self-calibration module to form a surrogate model; during online calibration, the module will measure the ( , The current heat transfer efficiency parameter is obtained by comparing the parameter with the lookup table and interpolating the result. This parameter is then used to update the reaction heat flow calculation module used to calculate heat dissipation power. The function is used to complete the calibration of the heat flux calculation.

[0045] Example 6: When processing waste lithium iron phosphate materials with different characteristics, the optimal trajectory used by the adjustment module of the system of the present invention can be adaptively adjusted through a cross-batch learning module; in one application scenario, when the recycling process of a batch is determined to be completed by the dynamic endpoint determination module, the complete historical data of the reaction heat flow signal of that batch, i.e., from the start of the reaction to the dynamic endpoint time, is obtained. The curve is used as a learning sample and input into the cross-batch learning module. This module compares the actual reaction heat flow trajectory of this operation with the preset optimal trajectory selected for this batch of materials, and corrects one or more parameters of the optimal trajectory model associated with this material type based on the systematic deviation between the two.

[0046] The corrected procedure is as follows: the cross-batch learning module first calculates the residual sequence between the two trajectories. The sequence was statistically analyzed; if it was found that the actual reaction heat flux was consistently higher than the preset optimal trajectory in the middle and later stages of the reaction, resulting in a residual integral value... A negative value indicates that the actual reactivity of the material is higher than the model's expectation. Based on this bias pattern, the cross-batch learning module uses a recursive least squares algorithm to evaluate the total package reaction rate constant in the optimal trajectory model associated with this material type. Incremental adjustments are made; the algorithm's input is the current residual value and a preset forgetting factor, which is set between 0.95 and 0.99. The forgetting factor's function is to assign higher weight to recent batches of learning samples when updating parameters. The corrected model parameters are updated and stored. When the next batch of the same type of material enters the reactor, the optimal trajectory matching module will generate a new preset optimal trajectory that more closely matches the material's reaction characteristics based on the optimized new model parameters. Through this mechanism, in each production task, the system transforms the deviation signal generated by the control process from a quantity used only for instantaneous adjustment into information that can be used to correct its internal model, enabling the control system to learn and adjust itself in continuous production.

[0047] To further verify the necessity and non-obviousness of the cross-validation mechanism introduced in this invention from a reverse perspective, the following comparative examples are provided.

[0048] Comparative Example 1: This comparative example aims to verify the technical deficiencies of a system that relies solely on reaction heat flow signals for feedback control in the absence of the core state consistency arbitration module of this invention. The experiment used the same 50L standard laboratory reactor, jacketed temperature control system, adjustable speed stirring device, PLC control system, and the same batch of waste lithium iron phosphate cathode powder as the aforementioned embodiments. All initial conditions, including an initial solid content of 30% slurry and a reaction temperature setpoint of 70°C, were maintained. The preset optimal heat flow trajectory, which serves as the control benchmark, is completely consistent with the aforementioned embodiments. The control system in this comparative example has the same hardware configuration as the system in the aforementioned embodiments, but in terms of control logic, the state consistency arbitration module and its intervention function on the adjustment module are artificially shielded. The system is configured to compare the thermodynamic state trajectory output by the reaction heat flow calculation module with the preset optimal trajectory, and to perform closed-loop adjustment of the feed actuator based on the deviation between the two. This configuration represents an improved technical path in the art prior to this invention, attempting to optimize constant temperature control by introducing heat flow calculation. The experimental procedure is as follows: The system first runs stably for 30 minutes. During this period, the control system effectively tracks the preset optimal trajectory based on the calculated reaction heat flow. At the 30-minute mark, to simulate the real working condition in an industrial environment where slow scaling and deterioration of heat transfer efficiency occur on the inner wall of the jacket due to changes in cooling water quality, the inlet temperature is adjusted in the reaction heat flow calculation module through the control software. The reading was artificially introduced, showing a linear increase to -0.5 within 5 minutes. And maintain a stable negative bias, this operation makes the system calculate the reaction heat flux signal The system continuously and persistently lowered reaction exothermic power than the actual power inside the reactor constitutes a silent sensor failure event that is difficult to detect by conventional diagnostic procedures. The test was terminated after running for 90 minutes.

[0049] The experimental results are recorded in Table 2. After the simulated failure occurred (30 minutes later), due to the decreased heat transfer efficiency in the simulation, the calculated reaction heat flux signal (e.g., 18.2 kW at the 35th minute) was lower than the actual value and deviated from the optimal trajectory target. Based on this, the control system incorrectly judged that the reaction rate was too slow, and therefore continuously instructed the feed actuator to increase the acid replenishment rate in an attempt to pull the calculated heat flux back to the target trajectory. The direct consequence of this adjustment was that the actual chemical reaction rate inside the reactor was excessively pushed up. By the 60-minute mark of the experiment, abnormal gas escape acceleration could be observed inside the reactor, and the temperature inside the reactor exceeded the set point by 1.5 kW. The uncontrolled overshoot was observed. After the experiment, the slurry at the endpoint was sampled and analyzed. The results showed that the lithium recovery rate was 88.1%, which was lower than the expected recovery rate (about 95%) under normal working conditions. Moreover, the concentration of impurity ions such as iron and copper in the leachate was 32% higher than that of the test group using the method of the present invention in Example 2.

[0050] Table 2: Key data node record table for Comparative Example 1.

[0051]

[0052] The experimental results show that a system that relies solely on a single thermodynamic information source for closed-loop control cannot inherently identify the failure of the control basis (reaction heat flow signal) caused by common industrial problems such as changes in heat transfer efficiency. In such cases, the system will make incorrect adjustment actions, which will not only fail to achieve process optimization, but will also exacerbate the instability of the production process, resulting in a double decline in product recovery rate and quality.

[0053] Example 7: Offline calibration and parameter setting procedure for key system modules. To establish the correspondence between thermal flux texture characteristic parameters and the mixing state in the reactor, i.e., the aforementioned knowledge base, the calibration procedure was carried out in a 50L transparent jacketed reactor with similar geometry to the production reactor and observation windows on the side walls. First, by adjusting the stirring speed and supplementing it with particle image velocimetry (PIV) analysis, two typical mixing states were reproduced in the reactor: a uniform suspension state with no obvious circulation dead zone and solid-liquid concentration gradient, and a bottom deposition state where a clearly visible solid material deposition layer with a thickness of more than 2cm appeared at the bottom of the reactor after reducing the stirring speed. After maintaining stable operation in these two states for more than 10 minutes, the corresponding reaction heat flux signals were continuously collected. The acquired signal was processed by a fourth-order Butterworth digital high-pass filter with a cutoff frequency set to 0.1 Hz to filter out the slow temperature drift during the process. Subsequently, the statistical variance of the filtered high-frequency fluctuation signal was calculated over a sliding time window of 60 seconds. The dominant frequency was obtained through Fast Fourier Transform. The multiple (calculated under uniform suspension state) , The distribution area of ​​the data pairs is defined as the normal mixing zone, while the distribution area of ​​the data pairs calculated under the bottom deposition state, which usually exhibits higher variance and a specific low-frequency dominant frequency, is defined as the unstable mixing zone. The boundaries of these two zones are fixed in the form of lookup tables or decision trees and stored in the thermal flow texture analysis module as the basis for its online diagnosis.

[0054] In the optimization and adjustment of the stirring device, when the real-time mixing state is diagnosed as having a material deposition zone, the system instructs the stirring device to execute a pulse cleaning program. A specific set of operating parameters is set as follows: based on the current stirring speed, for example, 150 rpm, a cleaning cycle lasting 5 minutes is executed, with a cycle period of 60 seconds. During the first 15 seconds of each cycle, the stirring speed is linearly increased to 180% of the base speed, i.e., 270 rpm, to generate instantaneous strong shear force to peel off and roll up the deposited material layer. In the following 45 seconds, the speed is restored to the base value of 150 rpm to reduce system energy consumption while maintaining material suspension. Simultaneously, to improve the accuracy of the rheological state trajectory characterization throughout the reaction cycle, the state consistency arbitration module is used to define the real-time load current. The baseline model relating the relationship to the solid content of the material is optimized by introducing a correction term based on the cumulative heat of reaction, and its complete state trajectory is obtained. By function The calculation shows that the function here... An extended offline calibration procedure was established to collect steady-state load currents under different solid contents. In addition, by adding reactants in batches, different reaction depths, i.e. different cumulative heats of reaction, were simulated at each solid content level. The load current is calculated, thereby quantifying and incorporating the effect of viscosity changes caused by liquid phase composition variations on motor load into the model, in order to obtain a rheological state trajectory that is more adaptable to the entire reaction process.

[0055] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0056] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A closed-loop optimization control system for the recycling process of waste lithium iron phosphate materials, characterized in that, The system comprises: a reaction heat flow calculation module configured to calculate a reaction heat flow signal based on operating parameters of a jacket temperature control system associated with the recovery process, and to time-integrate the reaction heat flow signal to generate a first state trajectory representing a degree of progress of the recovery process, i.e. a thermodynamic state trajectory; a state consistency arbitration module configured to generate a second state trajectory representing the degree of progress of the recovery process, i.e. a rheological state trajectory, in parallel based on a real-time load current of an agitator motor associated with an agitator device of the recovery process, and according to a reference model defining a relationship between the real-time load current and a solid content of a material of the recovery process; and to compare the thermodynamic state trajectory and the rheological state trajectory in each control cycle, and to determine that the first state trajectory has lost consistency when a deviation between the two trajectories continuously exceeds a preset threshold; a regulation module configured to generate a control signal for a material feeding actuator based on a deviation signal generated by comparing the thermodynamic state trajectory with a preset optimal trajectory by an optimal trajectory matching module; and the operation of the regulation module is independently subject to a rule of the state consistency arbitration module that once a loss of consistency is determined, the control authorization based on the optimal trajectory is revoked and a preset conservative control mode is switched to; and the system is further configured to receive a real-time redox potential signal of a redox potential sensor associated with the recovery process; the preset optimal trajectory is an optimal state path defined in a two-dimensional state space with the reaction heat flow and the redox potential as coordinate axes; and the regulation module is further configured to generate and output a second control signal for an oxidant feeding actuator of the recovery process based on the real-time redox potential signal and a deviation of a real-time state of the recovery process from the optimal state path in the two-dimensional state space; the system further comprises a heat transfer effectiveness self-calibration module configured to: at a preset diagnosis time, control the jacket temperature control system to apply a standardized thermodynamic disturbance to the reactor; capture and analyze a dynamic response characteristic of the reaction heat flow signal output by the reaction heat flow calculation module in response to the thermodynamic disturbance, the dynamic response characteristic including a peak time of the response and a signal decay half-life; and based on the dynamic response characteristic and according to a preset proxy model describing a relationship between the overall heat transfer coefficient and the dynamic response characteristic, to inverse a real heat transfer effectiveness parameter, and to calibrate the reaction heat flow calculation module using the real heat transfer effectiveness parameter.

2. The closed loop optimized control system for the recycling process of spent lithium iron phosphate material according to claim 1, wherein, The operating parameters of the jacket temperature control system used by the reaction heat flow calculation module to calculate the reaction heat flow signal include an inlet temperature of the coolant into the jacket, an outlet temperature of the coolant out of the jacket, and a real-time circulation flow rate of the coolant in the jacket; the reaction heat flow calculation module is further configured to calculate the reaction heat flow signal based on the first law of thermodynamics using the inlet temperature, the outlet temperature and the real-time circulation flow rate.

3. The closed loop optimized control system for the recycling process of spent lithium iron phosphate material according to claim 1, wherein, The optimal trajectory matching module is configured to internally store a trajectory library containing a plurality of optimal trajectories; and the optimal trajectory matching module is configured to, before the recycling process starts, select an optimal trajectory from the trajectory library as a preset optimal trajectory according to input initial information related to characteristics of the waste lithium iron phosphate material to be processed.

4. The closed loop optimized control system for the recycling process of spent lithium iron phosphate material according to claim 2, wherein, The reaction heat flow calculation module is configured to calculate the reaction heat flow signal by the following formula : , wherein, is the real-time circulating flow, is the preset specific heat capacity of the coolant, is the inlet temperature, is the outlet temperature, and is the heat dissipation power of the reactor to the environment, which is calculated according to a preset offline calibration function relationship.

5. The closed loop optimized control system for the recycling process of spent lithium iron phosphate material according to claim 1, wherein, The system further comprises a heat flow texture analysis module configured to continuously receive the reaction heat flow signal and process the reaction heat flow signal through a digital high-pass filter to separate out a high-frequency fluctuation part representing fluctuation characteristics of the reaction heat flow signal; calculate statistical characteristics of the high-frequency fluctuation part within a preset time window to extract at least one texture feature parameter, the texture feature parameter including variance of the high-frequency fluctuation part and a dominant frequency obtained by fast Fourier transform; and diagnose the real-time mixing state of the recycling process based on the texture feature parameter and a preset knowledge base defining the correspondence between the texture feature parameter and the mixing state.

6. The closed loop optimized control system for the recycling process of spent lithium iron phosphate material according to claim 5, wherein, The system further comprises an adaptive stirring optimization module configured to generate a control signal for adjusting the operating state of the stirring device based on the real-time mixing state diagnosed by the heat flow texture analysis module; wherein when the real-time mixing state is diagnosed as having a material deposition zone, the control signal is configured to instruct the stirring device to perform a pulse cleaning program, the pulse cleaning program including periodically increasing and decreasing the stirring speed of the stirring device within a preset time.

7. The closed loop optimized control system for the recycling process of spent lithium iron phosphate material according to claim 1, wherein, The system further comprises a dynamic endpoint determination module configured to calculate the time derivative of the reaction heat flow signal, i.e. the reaction acceleration, in real time through a digital differential algorithm; and when the absolute value of the reaction acceleration is always lower than a preset threshold within a continuous time window, determine that the recycling process has reached a dynamic endpoint and generate a termination signal accordingly.

8. The closed loop optimized control system for the recycling process of spent lithium iron phosphate material according to claim 7, wherein, The system further comprises a cross-batch learning module configured to, after a batch of recycling process is determined to be ended by the dynamic endpoint determination module, adjust one or more model parameters of the preset optimal trajectory associated with the material type of the batch based on the complete reaction heat flow signal history data of the batch through a recursive least squares algorithm.

Images