Method and apparatus for controlling electrodialysis process

Abstract

Provided are a method and an apparatus for controlling an electrodialysis process, the method comprising the steps of: predicting, on the basis of time-series data of the electrodialysis process, an anomaly of an electrodialysis process by using a first AI model; generating, in response to the anomaly predicted in the electrodialysis process, an operating condition by using a second AI model; simulating the electrodialysis process by using the operating condition; and transmitting the operating condition to a controller of the electrodialysis process according to the simulation results.

Description

Method and apparatus for controlling an electrodialysis process

[0001] This description claims priority to Korean Patent Application No. 10-2024-0191918, and the contents of the said patent application specification are all incorporated into this specification.

[0002] This description relates to a method and apparatus for controlling an electrodialysis process.

[0003] Bipolar Electrodialysis (BPED) is an electro-separation process used to separate or concentrate specific ions from a solution using ion exchange materials. BPED is gaining prominence as a particularly important water separation and regeneration technology.

[0004] BPED operation data may include various data collected during the operation of the technology. BPED operation data may include, for example, current and voltage data, substance concentration data, temperature and pressure data, and time data.

[0005] In order to improve BPED operations based on experience and increase the productivity of lithium (Li), there is a need for a model that predicts the concentration of lithium produced based on operation data. The prediction results can be used for the control or optimization of the production process.

[0006] One embodiment provides a method for controlling an electrodialysis process.

[0007] Another embodiment provides a device for controlling an electrodialysis process.

[0008] According to one embodiment, a method for controlling an electrodialysis process is provided.

[0009] The method includes the steps of: predicting an abnormality in the electrodialysis process using a first AI model based on time-series data of the electrodialysis process; generating operating conditions using a second AI model in response to the predicted abnormality in the electrodialysis process; simulating the electrodialysis process using the operating conditions; and transmitting the operating conditions to a controller of the electrodialysis process according to the results of the simulation.

[0010] In the above method, the time series data may include at least one of the temperature, conductivity, stack voltage, circulation pressure, and rectifier rectification of the electrodialysis process.

[0011] In the above method, the step of predicting anomalies in the electrodialysis process using a first AI model based on time series data of the electrodialysis process may include the step of predicting anomalies in the electrodialysis process through a combination of a time series trend analysis model and a linear relationship model.

[0012] In the above method, the step of predicting an anomaly in an electrodialysis process using a first AI model based on time series data of the electrodialysis process may include the step of separating the time series data into a trend component and a residual component; the step of processing the trend component and the residual component using different linear layers, respectively; the step of determining a prediction output by combining the processing results of the different linear layers; and the step of inputting the prediction output into a perceptron and determining the inference result output from the perceptron as a prediction result regarding the occurrence of an anomaly.

[0013] In the above method, the step of transmitting operating conditions to a controller of an electrodialysis process according to the results of a simulation includes the step of predicting anomalies in a simulated process using a first AI model based on virtual time series data generated in the simulation.

[0014] If an anomaly is predicted in the simulated process, a new operating condition may be generated using a second AI model, or if no anomaly is predicted in the simulated process, the operating condition may be transmitted to a controller.

[0015] According to another embodiment, a device for controlling an electrodialysis process is provided. The device includes a processor and a memory, wherein the memory stores instructions configured to cause the processor to perform a process, and the process includes the steps of: collecting time-series data of an electrodialysis process; predicting an anomaly in the electrodialysis process based on the time-series data using a first AI model; generating operating conditions using a second AI model in response to the prediction of an anomaly in the electrodialysis process; simulating the electrodialysis process using the operating conditions; and transmitting the operating conditions to a controller of the electrodialysis process according to the result of the simulation.

[0016] The step of predicting an abnormality in the electrodialysis process based on time series data using a first AI model in the above device may include the step of predicting whether an abnormality will occur in the electrodialysis process within m future hours based on time series data collected over n past hours.

[0017] In the above device, the first AI model may include a combination of a time series trend analysis model and a linear relationship model.

[0018] In the above device, the second AI model may include a combination of a prediction model and an optimization model.

[0019] By predicting abnormalities in the electrodialysis process and creating and verifying new operating conditions that prevent such abnormalities from occurring, the stability of the electrodialysis process can be enhanced and the production of lithium can be maximized.

[0020] FIG. 1 shows an electrodialysis device according to one embodiment.

[0021] FIG. 2 shows one end of an electrodialysis device according to one embodiment.

[0022] FIG. 3 shows a control system for an electrodialysis process according to one embodiment.

[0023] FIG. 4 illustrates a method for controlling an electrodialysis device according to one embodiment.

[0024] FIG. 5 shows the structure of a first AI model according to one embodiment.

[0025] Figure 6 shows an abnormality that occurred in an electrodialysis process according to one embodiment.

[0026] FIGS. 7a and 7b illustrate a notification delivered to a user when an abnormality occurs in an electrodialysis process according to one embodiment.

[0027] FIGS. 8a and FIGS. 8b show a comparison between a predicted value and a measured value according to one embodiment.

[0028] FIG. 9 is a block diagram showing a control system according to another embodiment.

[0029] The embodiments of this description are described below with reference to the attached drawings so that those skilled in the art can easily implement them. However, this description may be implemented in various different forms and is not limited to the embodiments described herein. Furthermore, in order to clearly explain this description in the drawings, parts unrelated to the explanation have been omitted, and similar parts throughout the specification are denoted by similar reference numerals.

[0030] In this description, each of the phrases such as “A or B”, “at least one of A and B”, “at least one of A or B”, “A, B or C”, “at least one of A, B and C”, and “at least one of A, B, or C” may include any one of the items listed together in the corresponding phrase, or all possible combinations thereof.

[0031] In this description, when a part is described as "including" a certain component, it means that, unless specifically stated otherwise, it does not exclude other components but may include additional components.

[0032] Expressions written in the singular in this description may be interpreted as singular or plural unless explicit expressions such as "one" or "singular" are used.

[0033] In this description, "and / or" includes each of the mentioned components and all combinations of one or more.

[0034] In this description, terms including ordinal numbers, such as first, second, etc., may be used to describe various components, but said components are not limited by said terms. Such terms are used solely for the purpose of distinguishing one component from another. For example, without departing from the scope of the present disclosure, the first component may be named the second component, and similarly, the second component may be named the first component.

[0035] In the flowchart described herein with reference to the drawings, the order of operations may be changed, multiple operations may be merged or some operations may be divided, and certain operations may not be performed.

[0036] The Artificial Intelligence model (AI model) of the present disclosure is a machine learning model that learns at least one task and may be implemented as a computer program executed by a processor. The task learned by the AI ​​model may refer to a problem to be solved through machine learning or a task to be performed through machine learning. The AI ​​model may be implemented as a computer program executed on a computing device, downloaded via a network, or sold in the form of a product. Alternatively, the AI ​​model may be linked with various devices via a network.

[0037] FIG. 1 shows an electrodialysis device according to one embodiment, and FIG. 2 shows one end of an electrodialysis device according to one embodiment.

[0038] In FIGS. 1 and 2, an electrodialysis device according to one embodiment may be a bipolar electrodialysis (BPED) device and may convert an aqueous lithium sulfate (Li2SO4) solution into lithium hydroxide (LiOH) and sulfuric acid (H2SO4). In one embodiment, the electrodialysis device may be an aqueous solution treatment facility that simultaneously performs water splitting / ion separation using an electrodialysis membrane in an electric field.

[0039] Referring to FIGS. 1 and 2, each stage of the electrodialysis device includes a salt chamber (liquid chamber), an acid chamber, and a base chamber between two electrodes that supply an electric field. A bipolar membrane (BPM) is placed between the acid chamber and the positive electrode and between the base chamber and the negative electrode. A cation exchange membrane (CEM) is placed between the salt chamber and the base chamber, and an anion exchange membrane (AEM) may be placed between the acid chamber and the salt chamber.

[0040] Cation exchange membranes (CEMs) contain anionic groups internally, allowing cations (e.g., Li) to... + ) can pass through the cation exchange membrane. The anion exchange membrane (AEM) contains cation groups internally, allowing anions (e.g., SO4) to pass through. 2- ) can pass through the anion exchange membrane. In a bipolar membrane (BPM), a cation membrane and an anion membrane are overlapped with a water splitting catalyst in between. The bipolar membrane (BPM) splits water within an electric field to produce hydrogen ions (H + ) and hydroxide ions (OH - It can generate ). That is, an electrodialysis device uses an electrodialysis membrane (cation dialysis membrane, anion dialysis membrane, bipolar membrane) within an electric field to perform water splitting (decomposition into H+, OH-) / ion separation (Li + , SO4 2- Ion separation can be performed simultaneously.

[0041] Referring to FIG. 1, in an electrodialysis device (BPED), through a three-stage process (Press) including the first to third stages, the LS solution is Li + and SO4 2- Ions can be transferred to the LH solution and the sulfuric acid (H2SO4) solution, respectively. For example, in an electrodialysis device, deionized water (DI Water) can be converted into the LH solution and the sulfuric acid solution by coming into counter-flow contact with the LS solution. Here, the LS solution is an aqueous lithium sulfate (Li2SO4) solution, and the LH solution is an aqueous lithium hydroxide (LiOH) solution.

[0042] The salt room, acid room, and base room of the first to third stages of Fig. 1 can be connected to the salt tank, acid tank, and base tank, respectively.

[0043] In each stage, lithium sulfate (Li2SO4) is supplied to the salt room, and desalted water can be produced after the reaction. Deionized water (DI Water) is supplied to the acid room, and sulfuric acid (H2SO4) can be produced after the reaction. Deionized water (DI Water) is supplied to the base room, and lithium hydroxide (LiOH) can be produced after the reaction.

[0044] The salt tank stores the produced demineralized water. The acid tank stores the produced sulfuric acid. The base tank stores the produced lithium hydroxide.

[0045] The output of the electrodialysis device (BPED) can be determined by the discharge flow rates of the aqueous sulfuric acid solution and the aqueous lithium hydroxide solution. The output of the electrodialysis device can be determined by the control of the input flow rate of water (H2O), the input flow rate of lithium sulfate, the current and voltage of the rectifier, and the pH, conductivity, circulation flow rate, and circulation pressure within each room. The input flow rate of the solution, the current and voltage of the rectifier, pH, conductivity, circulation flow rate, and circulation pressure, which correspond to the control and management elements, can be detected through internal sensors installed in each room. The control and management elements may correspond to variables that determine the concentration of the lithium hydroxide and sulfuric acid produced.

[0046] FIG. 3 shows a control system for an electrodialysis process according to one embodiment, FIG. 4 shows a method for controlling an electrodialysis device according to one embodiment, and FIG. 5 shows the structure of a first AI model according to one embodiment.

[0047] Referring to FIG. 3, the control system (10) of the electrodialysis process may include an anomaly predictor (100), an operating condition generator (200), and a controller (300).

[0048] In one embodiment, the anomaly predictor (100) may include a data collector (110) and a first AI model (120). The data collector (110) may collect time-series data of control variables of an electrodialysis process. The anomaly predictor (100) may predict an anomaly in an electrodialysis process using the first AI model (120) based on the time-series data of the electrodialysis process collected by the data collector (110).

[0049] In one embodiment, the operating condition generator (200) may include a second AI model (210) and a simulator (220). The operating condition generator (200) may generate operating conditions using the second AI model (210) according to productivity indicators of the electrodialysis process. Here, productivity indicators may include the purity, production volume, and conversion rate of lithium hydroxide and sulfuric acid, current efficiency of the electrodialysis device, etc.

[0050] In one embodiment, the simulator (220) can simulate an electrodialysis process using generated operating conditions and generate virtual time series data. In one embodiment, the anomaly predictor (100) can verify the newly generated operating conditions by using the first AI model (120) to predict anomalies in the simulated electrodialysis process based on the virtual time series data generated in the simulation.

[0051] If no abnormalities are predicted in the simulation of the electrodialysis process, that is, if the generated operating conditions are verified, the operating condition generator (200) can transmit the generated operating conditions to the controller (300). However, if abnormalities are predicted again in the simulation of the electrodialysis process using the operating conditions, the operating condition generator (200) can infer the operating conditions again using the second AI model.

[0052] Referring to FIG. 4, an anomaly predictor (100) of a control system (10) for an electrodialysis process can collect time-series data of the electrodialysis process using a data collector (110) and predict anomalies using a first AI model (120) based on the collected time-series data (S110). The data collector (110) can collect time-series data such as temperature, conductivity, stack voltage, circulation pressure, and rectifier current of the electrodialysis process from various sensors included in the electrodialysis device.

[0053] In one embodiment, the anomaly predictor (100) can predict whether an anomaly will occur in the electrodialysis process within n future hours based on time series data collected over a predetermined past time. For example, the anomaly predictor (100) can predict whether an anomaly will occur in the electrodialysis process over the next 2 hours based on time series data collected over the past 2 hours at 1-minute intervals. The first AI model (120) of the anomaly predictor (100) can be implemented through a combination of a time series trend analysis model (e.g., Dlinear) and a linear relationship model (e.g., SimpleLinear).

[0054] Referring to FIG. 5, the first AI model (120) of the anomaly predictor (100) separates time series data during a lookback window into a trend component and a residual component, and can process the trend component and the residual component using different linear layers, respectively. The trend component may represent long-term changes or patterns in the time series data (e.g., moving average, linear change). The residual component may represent noise or irregular patterns that deviate from the trend. The anomaly predictor (100) can separate the trend and the residual component using a signal decomposition technique.

[0055] Referring to FIG. 5, the first AI model (120) can determine a forecasting output by combining trend components and residual components processed in different linear layers and input the determined forecasting output into an input layer. In one embodiment, the forecasting output corresponding to time series data can be input into a perceptron that outputs a single inference result.

[0056] In FIG. 5, the time series data, trend component, residual component, and prediction output within the lookback window may be in the form of a graph image. In one embodiment, the graph image of the prediction output is converted into a feature vector, and each element of the feature vector may be input to each node of the input layer. For example, the pixel value of each pixel of the graph image of the prediction output may be input to each node of the input layer.

[0057] Referring to FIG. 5, a single inference result can be output from the output layer through a linear operation between the value of each node of the input layer and the weights and biases. In one embodiment, the anomaly predictor (100) can determine the inference result of the output layer for time series data as a prediction result regarding anomaly occurrence. In one embodiment, parameters such as the linear layer, weights and biases of the input layer within the first AI model (120) can be updated through supervised learning during the learning phase for the first AI model (120).

[0058] Referring again to FIG. 4, when the anomaly predictor (100) predicts that an anomaly will occur in the electrodialysis process within a future predetermined time (S120), the operation condition generator (200) can generate new operation conditions using the second AI model (210) (S130).

[0059] FIG. 6 shows an abnormality that occurred in an electrodialysis process according to one embodiment, and FIG. 7a and FIG. 7b show a notification delivered to a user when an abnormality occurs in an electrodialysis process according to one embodiment.

[0060] Referring to Fig. 6, the real-time conductivity and predicted conductivity of each room (salt room, acid room, and base room) are shown. To the left of the dotted line representing the current time, the real-time conductivity is superimposed on the predicted conductivity, and to the right of the dotted line, since it represents a future time, only the predicted conductivity is displayed.

[0061] Referring to Fig. 6, it was predicted that no abnormalities would occur in the electrodialysis process for the next 2 hours at a time slightly past the current prediction reference time of 12 o'clock, but the abnormality predictor (100) predicted that an abnormality would occur in the electrodialysis process within the next 2 hours at a time slightly later than the above time.

[0062] If it is predicted that an abnormality will occur in the electrodialysis process within m hours in the future, the abnormality predictor (100) can send a notification to the user through various means. Referring to FIG. 7a, a notification regarding the predicted abnormality may be displayed on a screen such as a computer or tablet connected to the control system (10). Alternatively, referring to FIG. 7b, a notification regarding the predicted abnormality may be displayed as various messages on the user's mobile phone.

[0063] In one embodiment, the second AI model (210) of the operating condition generator (200) can infer optimal process conditions that maximize lithium production through a combination of a prediction model (e.g., XGBOOST, MLP, etc.) and an optimization model (e.g., PSO). The operating condition generator (200) can infer optimal operating conditions that maximize lithium production based on operating indicators (conversion rate, current efficiency, product purity, etc.) and operating variables (circulation flow rate of each stage, input flow rate of deionized water and LS(Li2SO4) stock solution, etc.).

[0064] In one embodiment, the operating conditions generated by the second AI model (210) of the operating condition generator (200) may include control values ​​such as the flow rate of water input to the electrodialysis device, the supply flow rate and conductivity of lithium sulfate, the lithium concentration in lithium sulfate, the current and voltage of the rectifier, and at least one of the pH, conductivity, circulation flow rate, and circulation pressure within each room. For example, the operating condition generator (200) may use the second AI model (210) to determine operating conditions regarding at least one of the circulation flow rate of each room of each stage, the input flow rate of deionized water input into the acid room and base room, and the input flow rate of lithium sulfate (LS stock solution).

[0065] Referring to FIG. 4, the operating condition generator (200) can simulate an electrodialysis process according to newly generated operating conditions using a simulator (220) (S140). And the anomaly predictor (100) can predict an anomaly using a first AI model (120) based on virtual time series data generated in the simulation.

[0066] In one embodiment, if an abnormality is predicted in the simulated electrodialysis process according to the generated operating conditions, the operating condition generator (200) can use the second AI model to regenerate new operating conditions.

[0067] However, if it is predicted that no abnormalities will occur in the simulated electrodialysis process according to the generated operating conditions, the operating condition generator (200) can transmit the operating conditions verified through simulation to the controller (300). The controller (300) can control the electrodialysis device according to the new operating conditions transmitted from the operating condition generator (200).

[0068] As described above, a control system (10) for an electrodialysis process according to one embodiment can increase the stability of the electrodialysis process and maximize the production of lithium at the same time by predicting an abnormality in the electrodialysis process and creating and verifying new operating conditions to prevent the predicted abnormality from occurring.

[0069] FIGS. 8a and FIGS. 8b show a comparison between a predicted value and a measured value according to one embodiment.

[0070] FIG. 8a shows the conductivity of the salt room, acid room, and base room, and FIG. 8b shows the stack voltage. In FIG. 8a and FIG. 8b, the Mean Absolute Percentage Error (MAPE) value represents the ratio of the difference between the measured value and the predicted value divided by the measured value. Since MAPE represents the relative error ratio between the measured value and the predicted value, the prediction result can be intuitively represented. As can be seen in FIG. 8a and FIG. 8b, a control system (10) according to one embodiment can increase the stability of the electrodialysis process and maximize the production of lithium at the same time by predicting anomalies in the electrodialysis process and creating and verifying new operating conditions to prevent the predicted anomalies from occurring.

[0071] FIG. 9 is a block diagram showing a control system according to another embodiment.

[0072] A control system according to another embodiment may be implemented as a computer system, for example, a computer-readable medium. Referring to FIG. 9, the computer system (900) may include at least one of a processor (910) communicating via a bus (970), a memory (930), an input interface device (950), an output interface device (960), and a storage device (940). The computer system (900) may also include a communication device (920) coupled to a network.

[0073] At least one processor (910) may be a central processing unit (CPU) or a semiconductor device that executes instructions stored in memory (930) or a storage device (940). The processor (910) may implement the function, process, or method proposed in the embodiment. The operation of the computer system (900) according to the embodiment may be implemented by the processor (910). At least one processor (910) may include at least one of a GPU, a CPU, and an NPU. When the operation of the computer system (900) is implemented by at least one processor (910), each task may be divided among at least one processor (910) according to the load. For example, when one processor is a CPU, the other processor may be any one of a GPU, an NPU, an FPGA, or a DSP.

[0074] The memory (930) and storage device (940) may include various forms of volatile or non-volatile storage media. For example, the memory may include ROM (read only memory) and RAM (random access memory). The memory (930) may be connected to the processor (910) and may store various information for driving the processor (910) or at least one program executed by the processor (910). Alternatively, the memory (930) may store instructions that cause the processor (910) to perform a plurality of steps included in the function, process, or method proposed in the embodiment.

[0075] In the embodiments of the present description, the memory may be located inside or outside the processor, and the memory may be connected to the processor through various known means. The memory is a volatile or non-volatile storage medium of various forms, and, for example, the memory may include read-only memory (ROM) or random access memory (RAM).

[0076] Accordingly, the embodiment may be implemented as a method implemented on a computer or as a non-transient computer-readable medium storing computer-executable instructions. In one embodiment, when executed by a processor, the computer-readable instructions may perform a method according to at least one aspect of the present description.

[0077] The communication device (920) can transmit or receive wired or wireless signals.

[0078] Meanwhile, the embodiments are not implemented solely through the devices and / or methods described so far, but may also be implemented through a program that realizes a function corresponding to the configuration of the embodiments or a recording medium on which such a program is recorded. Such implementation can be easily achieved by a person skilled in the art to which the present invention pertains, based on the description of the embodiments described above. Specifically, the method according to the embodiments (e.g., network management method, data transmission method, transmission schedule generation method, etc.) may be implemented in the form of program instructions that can be executed through various computer means and may be recorded on a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, etc., either individually or in combination. The program instructions recorded on the computer-readable medium may be specially designed and configured for the embodiments, or they may be known and available to a person skilled in the art of computer software. The computer-readable recording medium may include a hardware device configured to store and execute program instructions. For example, computer-readable recording media may be magnetic media such as hard disks, floppy disks, and magnetic tapes; optical recording media such as CD-ROMs and DVDs; magneto-optical media such as floptical disks; ROM; RAM; flash memory; etc. Program instructions may include machine code, such as that generated by a compiler, as well as high-level language code that can be executed by a computer through an interpreter, etc.

[0079] Although the embodiments have been described in detail above, the scope of the rights is not limited thereto, and various modifications and improvements by those skilled in the art using the basic concepts defined in the following claims also fall within the scope of the rights.

Claims

1. As a method for controlling the electrodialysis process, A step of predicting abnormalities in the electrodialysis process using a first artificial intelligence (AI) model based on time series data of the electrodialysis process, A step of generating operating conditions using a second AI model in response to the prediction of an abnormality in the above electrodialysis process, A step of simulating the electrodialysis process using the above operating conditions, and A step of transmitting the operating conditions to the controller of the electrodialysis process according to the results of the above simulation. A method including 2. In Paragraph 1, The above time series data includes at least one of the temperature, conductivity, stack voltage, circulation pressure, and rectifier rectification of the electrodialysis process.

3. In Paragraph 1, The step of predicting abnormalities in the electrodialysis process using a first AI model based on time series data of the electrodialysis process is, A step of predicting abnormalities in the electrodialysis process through a combination of a time series trend analysis model and a linear relationship model. A method including 4. In Paragraph 1, The step of predicting abnormalities in the electrodialysis process using a first AI model based on time series data of the electrodialysis process is, A step of separating the above time series data into a trend component and a residual component, A step of processing the above trend component and the above residual component using different linear layers, respectively. A step of determining a prediction output by combining the processing results of the above different linear layers, and A step of inputting the above prediction output into a perceptron and determining the inference result output from the perceptron as a prediction result regarding anomaly occurrence. A method including 5. In Paragraph 1, The step of transmitting the operating conditions to the controller of the electrodialysis process according to the results of the above simulation is A step of predicting the anomaly in the simulated process using the first AI model based on virtual time series data generated in the above simulation, If the above anomaly is predicted in the above-mentioned simulated process, new operating conditions are generated using the second AI model or If the above abnormality is not predicted in the above simulated process, the step of transmitting the operating conditions to the controller A method including 6. As a device for controlling the electrodialysis process, It includes a processor and memory, wherein the memory stores instructions configured to cause the processor to perform a process, and the process, A step of collecting time series data of the above electrodialysis process, A step of predicting abnormalities in the electrodialysis process based on the time series data using a first artificial intelligence (AI) model, A step of generating operating conditions using a second AI model in response to the prediction of an abnormality in the above electrodialysis process, A step of simulating the electrodialysis process using the above operating conditions, and A step of transmitting the operating conditions to the controller of the electrodialysis process according to the results of the above simulation. A device including 7. In Paragraph 6, The step of predicting abnormalities in the electrodialysis process based on the time series data using the first AI model is: A step of predicting whether an abnormality will occur in the electrodialysis process within m future hours based on the time series data collected over the past n hours. A device including 8. In Paragraph 6, The above first AI model is a device comprising a combination of a time series trend analysis model and a linear relationship model.

9. In Paragraph 6, The above second AI model is a device comprising a combination of a prediction model and an optimization model.

Images