Method and device for decision of lithium-containing mixed salt solution evaporation crystallization path

Abstract

This invention relates to the field of metal recycling technology, specifically to a method and apparatus for determining the evaporation and crystallization path of a lithium-containing mixed salt solution. The method includes: acquiring the concentration and temperature values ​​of the mixed salt solution in the crystallizer at the current moment, and acquiring the location tag information of the target crystallization unit in the crystallization state during the current evaporation stage; inputting the concentration, temperature, and location tag information into a trained evaporation path prediction model, and obtaining the minimum number of evaporation paths for the target crystallization unit output by the evaporation path prediction model. This invention focuses on the state of the target crystallization unit itself and actively searches for related crystallization units upstream, to its upper left, and directly above it, acquiring the optimal evaporation parameters of these units and using them to expand path control information. This allows the final evaporation path to effectively integrate the operating conditions of upstream, downstream, and adjacent units, avoiding overall mismatch or uneven crystallization particle size distribution caused by independent optimization of a single unit.

Description

Technical Field

[0001] This invention relates to the field of metal recycling technology, specifically to a method and apparatus for determining the evaporation and crystallization path of a lithium-containing mixed salt solution. Background Technology

[0002] Currently, traditional methods lack global collaborative optimization based on the spatial relationship of crystallization units. In conventional multi-stage or gridded evaporation crystallization processes, each crystallization unit usually sets its own operating parameters independently, rarely actively querying the actual crystallization rate and optimal evaporation parameters of upstream, upper left, or directly above related units. This isolated control mode leads to the inability of downstream units to predict and compensate for the fluctuations in crystal slurry concentration generated by upstream units in a timely manner, which can easily form local supersaturation abnormal zones in the crystallizer network. Ultimately, this results in a wide distribution of crystallization particle size, with large and fine crystals mixed together, making it difficult to consistently achieve the required product purity.

[0003] Furthermore, in traditional methods, the gain values ​​of the response curves between feed flow rate and crystallization rate are mostly calibrated offline and remain fixed. When the raw material composition, temperature, or evaporation load changes, there will be a continuous positive or negative average deviation between the actual crystallization rate and the reference path. However, traditional control systems lack a closed-loop correction mechanism that automatically increases or decreases the gain based on the magnitude of the average deviation. This results in either excessively strong adjustment actions that cause crystallization rate oscillations and overshoot, or excessively weak adjustments that allow the deviation to persist for a long time. As a result, the actual crystallization rate frequently deviates from the optimal path, requiring secondary processing or even remelting and recrystallizing defective products. Summary of the Invention

[0004] To achieve the above objectives, the present invention provides the following technical solution: a method for determining the evaporation and crystallization path of a lithium-containing mixed salt solution, comprising: Obtain the concentration and temperature values ​​of the mixed salt solution in the crystallizer at the current moment, and obtain the location tag information of the target crystallization unit in the crystallization state during the current evaporation stage; The concentration value, temperature value, and location label information are input into the trained evaporation path prediction model to obtain the minimum number of evaporation paths for the target crystallization unit output by the evaporation path prediction model. Based on the location tag information of the target crystallization unit, determine whether there is a corresponding associated crystallization unit for the target crystallization unit in the current evaporation stage; When the target crystallization unit has a corresponding associated crystallization unit in the current evaporation stage, the optimal evaporation parameter corresponding to each associated crystallization unit is obtained, and an extension operation is performed on the optimal evaporation parameter corresponding to each associated crystallization unit in sequence. The extension operation includes determining whether the optimal evaporation parameter corresponding to the associated crystallization unit is within the parameter range corresponding to the minimum number of evaporation paths in the target crystallization path control information. If the optimal evaporation parameter corresponding to the associated crystallization unit is not within the parameter range corresponding to the minimum number of evaporation paths in the target crystallization path control information, the evaporation path corresponding to the optimal evaporation parameter of the associated crystallization unit is added to the target crystallization path control information to obtain extended crystallization path control information. The crystallization control parameters of all evaporation paths in the extended crystallization path control information are optimized to determine the optimal evaporation path of the target crystallization unit, and the crystallizer operation is controlled based on the optimal evaporation path.

[0005] Preferably, the crystallization control parameters of all evaporation paths in the extended crystallization path control information are optimized to determine the optimal evaporation path for the target crystallization unit, including: For each evaporation path in the extended crystallization path control information, the crystallization control parameter evaluation value of each evaporation path is calculated by combining the parameter range corresponding to the minimum number of evaporation paths in the target crystallization path control information. The evaporation paths in the extended crystallization path control information are sorted in descending order of the evaluation values, and a preset number of evaporation paths with the highest ranking are selected to form a candidate evaporation path set. The evaporation path with the highest evaluation value of crystallization control parameters is selected from the set of candidate evaporation paths as the optimal evaporation path for the target crystallization unit.

[0006] Preferably, the process of generating the target crystallization path control information includes: Acquire initial information for crystallization path control and path adjustment reference information; the path adjustment reference information includes the actual crystallization adjustment amount corresponding to each moment in the preset historical evaporation stage; different preset historical evaporation stages correspond to different preset concentration ranges. The preset historical evaporation stages in which the actual crystallization rate at each moment does not meet the preset deviation condition are selected and used as the target evaporation stages; Obtain the preset concentration range corresponding to the target evaporation stage, and use it as the target concentration range; Based on the initial information of crystallization path control and the reference information of path adjustment, the reference adjustment amount of crystallization path at each moment in the target evaporation stage is calculated. Based on the difference between the crystallization path reference adjustment and the actual crystallization adjustment at each moment in the target evaporation stage, the average deviation value corresponding to the target concentration range is calculated. Based on the average deviation value, the path control information of the target concentration segment in the initial information of crystallization path control is adjusted until the crystallizer is run using the adjusted path control information of the target concentration segment. At each moment, the actual crystallization rate and the target crystallization rate of the target concentration segment meet the preset deviation condition. The initial information of crystallization path control with the adjusted path control information of the target concentration segment is determined as the target crystallization path control information.

[0007] Preferably, based on the initial crystallization path control information and the path adjustment reference information, the crystallization path reference adjustment amount corresponding to each moment in the target evaporation stage is calculated, including: Based on the pre-generated crystallization path reference adjustment calculation formula, the response curves of evaporation rate and crystallization rate at each moment in the target evaporation stage, the response curves of feed flow rate and crystallization rate at each moment in the target evaporation stage, and the corresponding evaporation rate and feed flow rate at each moment in the target evaporation stage, the crystallization path reference adjustment amount corresponding to each moment in the target evaporation stage is calculated.

[0008] Preferably, based on the average deviation value, the path control information for the target concentration range in the initial information for crystallization path control is adjusted, including: Based on the average deviation value, adjust the response curve gain value in the response curve of the feed flow rate and crystallization rate of the target evaporation stage corresponding to the target concentration segment in the initial information of crystallization path control.

[0009] Preferably, based on the average deviation value, adjusting the response curve gain value in the response curve of the feed flow rate and crystallization rate for the target evaporation stage corresponding to the target concentration segment in the initial information for crystallization path control includes: If the average deviation value is greater than a preset threshold, the gain value of the response curve in the response curve of the feed flow rate and crystallization rate of the target evaporation stage corresponding to the target concentration segment in the initial information of crystallization path control is reduced. If the average deviation value is less than a preset threshold, increase the gain value of the response curve in the response curve of the feed flow rate and crystallization rate of the target evaporation stage corresponding to the target concentration segment in the initial information of crystallization path control.

[0010] Preferably, the method further includes: When the target crystallization unit does not have a corresponding associated crystallization unit in the current evaporation stage, the crystallization control parameters of all evaporation paths corresponding to the minimum number of evaporation paths in the target crystallization path control information are optimized to determine the optimal evaporation path of the target crystallization unit.

[0011] Preferably, obtaining the optimal evaporation parameters corresponding to each of the associated crystallization units includes: The location tag information and actual crystallization rate of each associated crystallization unit are input into the evaporation path prediction model to obtain the optimal evaporation parameters corresponding to each associated crystallization unit output by the evaporation path prediction model. The associated crystallization unit is a crystallization unit located upstream of the target crystallization unit, to the upper left of the target crystallization unit, or directly above the target crystallization unit during the pre-division process of the current evaporation stage.

[0012] Preferably, determining whether there is a corresponding associated crystallization unit for the target crystallization unit in the current evaporation stage based on the location tag information of the target crystallization unit includes: Based on the location tag information of the target crystallization unit, locate the crystallization unit located upstream of the target crystallization unit, the crystallization unit located to the upper left of the target crystallization unit, and the crystallization unit located directly above the target crystallization unit in the pre-divided crystallization unit set of the current evaporation stage; When the number of found crystal units is greater than zero, it is determined that the target crystal unit has a corresponding associated crystal unit in the current evaporation stage; when the number of found crystal units is equal to zero, it is determined that the target crystal unit does not have a corresponding associated crystal unit in the current evaporation stage.

[0013] A decision-making device for the evaporation and crystallization path of a lithium-containing mixed salt solution, applicable to the aforementioned decision-making method for the evaporation and crystallization path of a lithium-containing mixed salt solution, comprising: The data acquisition unit is configured to acquire the concentration and temperature values ​​of the mixed salt solution in the crystallizer at the current moment, and to acquire the location tag information of the target crystallization unit in the crystallization state during the current evaporation stage; The path prediction unit is configured to input the concentration value, temperature value, and location label information into the trained evaporation path prediction model to obtain the minimum number of evaporation paths for the target crystallization unit output by the evaporation path prediction model. The location determination unit is configured to determine whether there is a corresponding associated crystallization unit for the target crystallization unit in the current evaporation stage based on the location tag information of the target crystallization unit. An extended operation unit is configured to, when the target crystallization unit has a corresponding associated crystallization unit in the current evaporation stage, obtain the optimal evaporation parameter corresponding to each associated crystallization unit, and sequentially perform an extended operation on the optimal evaporation parameter corresponding to each associated crystallization unit. The extended operation includes determining whether the optimal evaporation parameter corresponding to the associated crystallization unit exists within the parameter range corresponding to the minimum number of evaporation paths in the target crystallization path control information. If the optimal evaporation parameter corresponding to the associated crystallization unit does not exist within the parameter range corresponding to the minimum number of evaporation paths in the target crystallization path control information, the evaporation path corresponding to the optimal evaporation parameter corresponding to the associated crystallization unit is added to the target crystallization path control information to obtain extended crystallization path control information. The control decision unit is configured to optimize the crystallization control parameters of all evaporation paths in the extended crystallization path control information to determine the optimal evaporation path of the target crystallization unit, and control the crystallizer operation based on the optimal evaporation path.

[0014] Compared with the prior art, the beneficial effects of the present invention are: This invention achieves global collaborative optimization by associating the spatial relationships of crystallization units, significantly improving the consistency of crystallization quality in multi-stage crystallizer networks. This method not only focuses on the state of the target crystallization unit itself, but also actively searches for associated crystallization units upstream, to the upper left, and directly above it, obtains the optimal evaporation parameters of these units, and uses them to expand path control information. This allows the final evaporation path to effectively integrate the operating conditions of upstream, downstream, and adjacent units, avoiding overall mismatch or uneven crystallization particle size distribution caused by independent optimization of a single unit. As a result, more uniform crystal growth and higher product purity are achieved during the continuous crystallization process of lithium mixed salt solutions. This invention calculates the average deviation by measuring the difference between the reference adjustment amount and the actual crystallization adjustment amount at each moment within the target concentration range, and corrects the gain value of the feed flow rate and crystallization rate response curve accordingly. When the deviation exceeds a preset threshold, the gain is automatically increased or decreased, enabling the control model to be calibrated online according to the characteristics of different concentration ranges and different evaporation stages. This reduces the deviation between the actual crystallization rate and the reference path, shortens the adjustment time, and ultimately ensures that the lithium-containing mixed salt solution can operate stably on the optimal crystallization path at different evaporation stages, significantly reducing waste or secondary processing caused by parameter mismatch. Attached Figure Description

[0015] Figure 1 This is a schematic flowchart of the overall method in one embodiment of the present invention; Figure 2 This is a schematic diagram of the overall device architecture in one embodiment of the present invention.

[0016] In the diagram: 1. Data acquisition unit; 2. Path prediction unit; 3. Location determination unit; 4. Extended operation unit; 5. Control decision unit. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Example 1, please refer to Figure 1 This invention provides a technical solution: a method for determining the evaporation and crystallization path of a lithium-containing mixed salt solution, comprising: S1. Obtain the concentration and temperature values ​​of the mixed salt solution in the crystallizer at the current moment, and obtain the location tag information of the target crystallization unit in the crystallization state during the current evaporation stage; S2. Input the concentration value, temperature value and location label information into the trained evaporation path prediction model to obtain the minimum number of evaporation paths for the target crystallization unit output by the evaporation path prediction model. S3. Determine whether there is a corresponding associated crystallization unit in the current evaporation stage based on the location tag information of the target crystallization unit; S4. When the target crystallization unit has a corresponding associated crystallization unit in the current evaporation stage, obtain the optimal evaporation parameter corresponding to each associated crystallization unit, and perform an extension operation on the optimal evaporation parameter corresponding to each associated crystallization unit in sequence. The extension operation includes determining whether the optimal evaporation parameter corresponding to the associated crystallization unit is within the parameter range corresponding to the minimum number of evaporation paths in the target crystallization path control information. If the optimal evaporation parameter corresponding to the associated crystallization unit is not within the parameter range corresponding to the minimum number of evaporation paths in the target crystallization path control information, then add the evaporation path corresponding to the optimal evaporation parameter of the associated crystallization unit to the target crystallization path control information to obtain extended crystallization path control information. S5. Optimize the crystallization control parameters of all evaporation paths in the extended crystallization path control information to determine the optimal evaporation path of the target crystallization unit, and control the crystallizer operation based on the optimal evaporation path.

[0019] It should be noted that the lithium-containing mixed salt solution is processed using an evaporation crystallization process; the crystallizer contains multiple crystallization units, each corresponding to a region; the concentration of the mixed salt solution in the crystallizer is monitored in real time, which is 120 g / L and the temperature is 85 degrees Celsius, and the location tag information of the target crystallization unit in the crystallization state during the current evaporation stage is obtained, for example, numbered C5; The concentration value of 120, the temperature value of 85, and the location label C5 are input into the trained evaporation path prediction model. This model is a neural network trained based on historical data, and it outputs that the minimum number of evaporation paths for the target crystallization unit C5 is 3. This means that at least 3 different combinations of evaporation parameters are required from the current state to the completion of crystallization. Based on the location label C5 of the target crystallization unit, determine whether there are associated crystallization units in the current evaporation stage; associated crystallization units refer to units adjacent to C5 that influence each other's crystallization process, such as C4, C6, and D5; upon investigation, associated crystallization units C4, C6, and D5 are found. The optimal evaporation parameters for each associated crystallization unit are obtained sequentially. For example, the optimal evaporation parameters for C4 are: evaporation temperature 90°C, vacuum degree 0.08 MPa, and stirring speed 150 rpm; for C6, 88°C, 0.09 MPa, and 140 rpm; and for D5, 92°C, 0.07 MPa, and 160 rpm. These parameters are checked to see if they exist within the parameter range corresponding to the minimum number of evaporation paths (3) in the target crystallization path control information. The target crystallization path control information is a preset feasible parameter domain, such as a temperature range of 85 to 95°C and a vacuum degree of 0.07 to 0.10 MPa. All parameters for C4 are within this range, so no new path is added. The parameters for C6 are also within this range. The vacuum degree of 0.07 in D5 is just on the boundary, so it is considered present. Therefore, all parameters are within the range, and no additional paths need to be added. If the parameters of an associated crystallization unit exceed the range, such as a temperature of 98°C, its corresponding evaporation path is added to the control information to obtain extended crystallization path control information. The crystallization control parameters of all evaporation paths in the extended crystallization path control information are optimized. For example, the particle swarm optimization algorithm is used to find the parameter combination that results in the fastest crystallization rate and the highest purity. The optimal evaporation path for the target crystallization unit C5 is determined as follows: first, it runs at a temperature of 92 degrees Celsius, a vacuum of 0.07 MPa, and a stirring speed of 160 rpm for 30 minutes, and then switches to a temperature of 88 degrees Celsius, a vacuum of 0.09 MPa, and a stirring speed of 140 rpm for 20 minutes. Based on this optimal evaporation path, the heating power, vacuum pump, and stirrer of the crystallizer are automatically adjusted to control the crystallizer to operate according to the plan, and finally, a lithium carbonate product with uniform particle size and qualified purity is obtained.

[0020] In the embodiments of this application, the specific structure of the evaporation path prediction model and the processing procedures between its components specifically include: The model adopts a multi-layer feedforward neural network structure. The input layer receives features in three dimensions: the concentration and temperature of the mixed salt solution in the crystallizer at the current moment, and the location label information of the target crystallization unit. The location label information is a discrete categorical variable, such as the unit number. Therefore, it first passes through an embedding layer to encode the high-dimensional sparse label into a low-dimensional dense vector. The concentration and temperature are continuous values, which are first standardized and then concatenated with the embedding vector to form a comprehensive feature vector. Subsequently, the integrated feature vector is input into several fully connected hidden layers, each followed by a batch normalization layer and a ReLU activation function to extract high-dimensional nonlinear combined features; finally, the output layer is a linear neuron that outputs a scalar value representing the minimum number of evaporation paths required by the target crystallization unit in the current state; since the number of paths is a positive integer, the output is rounded or truncated. The data processing flow between the various structures is as follows: the location labels in the original data are mapped into dense vectors through the embedding layer; the concentration and temperature are standardized and then concatenated with the embedding vector along the feature dimension; the concatenated vector is passed sequentially through the first fully connected layer, batch normalization, and ReLU, then through the second fully connected layer, batch normalization, and ReLU, and so on; finally, the predicted value is obtained through the output layer; during training, the mean squared error loss function is used, and the minimum number of actual paths in the historical evaporation and crystallization data is used as the supervision signal; the model can learn the joint influence of concentration, temperature, and spatial location on the complexity of evaporation paths, thus providing initial constraints for subsequent path expansion and optimization.

[0021] In an optional embodiment, the crystallization control parameters of all evaporation paths in the extended crystallization path control information are optimized to determine the optimal evaporation path for the target crystallization unit, including: For each evaporation path in the extended crystallization path control information, the evaluation value of the crystallization control parameter for each evaporation path is calculated by combining the parameter range corresponding to the minimum number of evaporation paths in the target crystallization path control information. The evaporation paths in the extended crystallization path control information are sorted in descending order of evaluation value, and a preset number of evaporation paths with the highest ranking are selected to form a candidate evaporation path set. The evaporation path with the highest evaluation value of crystallization control parameters is selected from the set of candidate evaporation paths as the optimal evaporation path for the target crystallization unit.

[0022] It should be noted that the extended crystallization path control information includes the following five evaporation paths: Path A, Path B, Path C, Path D, and Path E; the parameter range corresponding to the minimum number of evaporation paths in the target crystallization path control information is: temperature 85 to 95 degrees Celsius, vacuum degree 0.07 to 0.10 MPa, and stirring rate 120 to 180 rpm; a crystallization control parameter evaluation value is calculated for each path, which is based on a comprehensive assessment of the path's evaporation rate, energy consumption, crystal particle size distribution, and impurity precipitation tendency; for example, the evaluation value for path A is 92. Path B has a score of 85, Path C has a score of 88, Path D has a score of 78, and Path E has a score of 95. Paths are ranked from highest to lowest score as follows: E, A, C, B, D. The preset selection quantity is 3, therefore the top three paths E, A, and C are selected to form a candidate evaporation path set. Finally, Path E, with the highest score, is selected from the candidate set as the optimal evaporation path for the target crystallization unit. The evaporation crystallization operation is performed according to the evaporation parameters of Path E: temperature 92 degrees Celsius, vacuum degree 0.08 MPa, and stirring speed 160 rpm, to achieve the most efficient and high-quality crystallization effect.

[0023] In an optional embodiment, the process of generating target crystallization path control information includes: Acquire initial information for crystallization path control and path adjustment reference information; the path adjustment reference information includes the actual crystallization adjustment amount corresponding to each moment in the preset historical evaporation stage; different preset historical evaporation stages correspond to different preset concentration ranges. The preset historical evaporation stages in which the actual crystallization rate at each moment does not meet the preset deviation condition are selected and used as the target evaporation stages; Obtain the preset concentration range corresponding to the target evaporation stage and use it as the target concentration range; Based on the initial information of crystallization path control and the reference information of path adjustment, the reference adjustment amount of crystallization path at each moment in the target evaporation stage is calculated. Based on the difference between the crystallization path reference adjustment and the actual crystallization adjustment at each moment in the target evaporation stage, the average deviation value corresponding to the target concentration range is calculated. Based on the average deviation value, the path control information of the target concentration segment in the initial information of crystallization path control is adjusted until the crystallizer is run using the adjusted path control information of the target concentration segment. At each moment corresponding to the target concentration segment, the actual crystallization rate and the target crystallization rate meet the preset deviation condition. The initial information of crystallization path control with the adjusted path control information of the target concentration segment is determined as the target crystallization path control information.

[0024] It should be noted that the initial crystallization path control information is set based on theoretical calculations and experience. For example, when the concentration range is 100 to 120 g / L, the temperature is set to 90 degrees Celsius, the vacuum degree to 0.09 MPa, and the stirring rate to 150 rpm. The path adjustment reference information comes from the actual crystallization adjustments at various times during the historical evaporation stages, i.e., the correction values ​​made by the operator or the system during operation to the temperature, vacuum degree, etc. The historical evaporation stages are divided according to the concentration range, such as the low concentration range of 80 to 100 g / L, the medium concentration range of 100 to 120 g / L, and the high concentration range of 120 to 140 g / L. Historical evaporation stages where the actual crystallization rate at each moment does not meet the preset deviation condition are selected. The preset deviation condition is that the difference between the actual crystallization rate and the target crystallization rate is less than ±2%. Assuming that in the three historical evaporation stages, the actual crystallization rate at multiple moments in the medium concentration range of 100 to 120 g / L is 5% to 8% lower than the target crystallization rate, which does not meet the condition, the medium concentration range is determined as the target evaporation stage, and the corresponding target concentration range is 100 to 120 g / L. Based on the initial information of crystallization path control (theoretical parameters) and the path adjustment reference information (historical correction amount), the crystallization path reference adjustment amount corresponding to each moment in the target evaporation stage is calculated; for example, if the theoretical temperature is 90 degrees, and the actual temperature is increased to 92 degrees in the historical record at a concentration of 110 g / L, then the reference adjustment amount is +2 degrees; the average of multiple moments is taken to obtain the average reference adjustment amount within this concentration range. Obtain the actual crystallization adjustment amount at each moment within the target evaporation stage, i.e., the actual correction value each time; calculate the difference between the reference adjustment amount and the actual adjustment amount at each moment, and then take the average value to obtain the average deviation value of the target concentration range; for example, the average deviation value is temperature +1.5 degrees and vacuum degree -0.005 MPa. Based on the average deviation value, the path control information for the target concentration range in the initial information of crystallization path control is adjusted; for example, the temperature in the medium concentration range in the initial information is adjusted from 90 degrees to 91.5 degrees, and the vacuum degree is adjusted from 0.09 MPa to 0.085 MPa. Then, the crystallizer is controlled using the adjusted parameters, and the actual crystallization rate and the target crystallization rate are monitored. If, in multiple batches, the actual crystallization rate and the target crystallization rate at each moment in this concentration range meet the condition that the deviation is less than 2%, then the adjusted path control information is determined as the target crystallization path control information. If there is still a deviation, iterative adjustments are continued until the condition is met. The final target crystallization path control information contains optimized parameters for all concentration ranges and is used for subsequent decision-making.

[0025] In an optional embodiment, based on the initial crystallization path control information and the path adjustment reference information, the crystallization path reference adjustment amount corresponding to each moment in the target evaporation stage is calculated, including: Based on the pre-generated formula for calculating the crystallization path reference adjustment, the response curves of evaporation rate and crystallization rate at each moment in the target evaporation stage, the response curves of feed flow rate and crystallization rate at each moment in the target evaporation stage, and the corresponding evaporation rate and feed flow rate at each moment in the target evaporation stage, the crystallization path reference adjustment corresponding to each moment in the target evaporation stage is calculated.

[0026] It should be noted that, in a specific evaporation stage of the continuous evaporation crystallization process, the 120th second is used as an example; at this time, the measured evaporation rate is 2.8 kg per second, and the measured feed flow rate is 0.9 cubic meters per minute. Two response curves for this stage were obtained through prior experiments: the curve showing the relationship between evaporation rate and crystallization rate showed that when the evaporation rate was in the range of 2.5 kg / s to 3.0 kg / s, the rate of increase of crystallization rate per second increased linearly from 0.05 to 0.10; the curve showing the relationship between feed flow rate and crystallization rate showed that when the feed flow rate was in the range of 0.8 m³ / min to 1.0 m³ / min, the rate of decrease of crystallization rate per second increased linearly from 0.01 to 0.03. Based on these two curves, linear interpolation of the current measured values ​​yields the following: the crystallization rate increase rate corresponding to an evaporation rate of 2.8 kg / s is 0.09 / s, and the crystallization rate decrease rate corresponding to a feed flow rate of 0.9 m / s is 0.02 / s. Therefore, the actual net rate of change of the crystallization rate at the current moment is an increase of 0.07 / s; while the crystallization rate increase rate required by the preset reference crystallization path at the 120th second is 0.08 / s, and the difference between the two is a negative 0.01 / s. The pre-generated crystallization path reference adjustment calculation formula adopts a proportional-integral form, where the proportionality coefficient is 0.6, the integral coefficient is 0.1, and the cumulative deviation of the previous time step is negative 0.002 per second; the current proportional term is calculated as 0.6 multiplied by negative 0.01, which equals negative 0.006, and the integral term is calculated as 0.1 multiplied by negative 0.002, which equals negative 0.0002. The sum of the two terms yields a crystallization path reference adjustment of negative 0.0062. This adjustment amount indicates that the target crystallization rate of the reference crystallization path at the 120th second needs to be corrected downward by 0.0062 percentage points, so that the actual evaporation and feeding operations can run more accurately along the corrected path and reduce subsequent deviations. Similarly, at each other moment, the respective crystallization path reference adjustment amount is calculated based on the evaporation rate, feed flow rate, corresponding response curve and cumulative deviation at that moment, which is used to guide the dynamic control of the entire evaporation stage.

[0027] In an optional embodiment, the path control information for the target concentration range in the initial information of crystallization path control is adjusted based on the average deviation value, including: Based on the average deviation value, adjust the gain value of the response curve in the response curve of the feed flow rate and crystallization rate of the target evaporation stage corresponding to the target concentration segment in the initial information of crystallization path control.

[0028] It should be noted that in the target evaporation stage of the continuous evaporation crystallization process, the initial set gain value of the response curve between the feed flow rate and the crystallization rate is 0.03; this gain value means that for every 0.1 cubic meters per minute change in the feed flow rate, the corresponding rate of change in the crystallization rate is 0.003 percentage points per second; after completing one round of actual operation, the deviation between the measured crystallization rate and the reference crystallization path at each moment in the entire target evaporation stage is collected, and the average deviation value is calculated to be positive 0.005 percentage points per second; A positive value means that the actual crystallization rate is higher than the reference path, indicating that the gain value in the original response curve is too small, so that the predicted feed flow rate has insufficient effect on reducing the crystallization rate. According to the preset adjustment rule, every 0.001 percentage points per second of the absolute value of the average deviation corresponds to a gain adjustment of 0.001. The current absolute value of the average deviation is 0.005, so the adjustment is 0.005. Adding 0.005 to the original gain value of 0.03, we get a new gain value of 0.035. In the updated response curve, an increase of 0.1 cubic meters per minute in the feed flow rate will change the rate of decrease in crystallization rate to 0.0035 percentage points per second, thus improving the adjustment sensitivity. In the subsequent control stage, the new gain value is used to recalculate the relationship between the feed flow rate and the crystallization rate at each moment, thereby reducing the average deviation between the actual crystallization path and the reference path in the future and achieving more precise control.

[0029] In an optional embodiment, based on the average deviation value, the response curve gain value in the response curve of the feed flow rate and crystallization rate for the target evaporation stage corresponding to the target concentration segment in the initial information for crystallization path control is adjusted, including: If the average deviation value is greater than the preset threshold, reduce the gain value of the response curve in the response curve of the feed flow rate and crystallization rate of the target evaporation stage corresponding to the target concentration segment in the initial information of crystallization path control. If the average deviation is less than the preset threshold, increase the gain value of the response curve in the response curve of the feed flow rate and crystallization rate of the target evaporation stage corresponding to the target concentration segment in the initial information of crystallization path control.

[0030] It should be noted that in the target evaporation stage of the continuous evaporation crystallization process, the gain value of the response curve between the initial feed flow rate and the crystallization rate is set to 0.04; the preset threshold for the absolute value of the average deviation is 0.002 percentage points per second; after the first round of actual operation, the average deviation between the measured crystallization rate and the reference path during the entire target evaporation stage is calculated to be positive 0.003 percentage points per second, which is greater than the threshold of 0.002. A positive value indicates that the actual crystallization rate is higher than the reference path, suggesting that the current gain value of 0.04 is too large, causing the feed flow rate to have an excessively strong effect on reducing the crystallization rate, which in turn causes the crystallization rate to deviate from the upper limit. Therefore, the gain value needs to be reduced. According to the adjustment rules, for every 0.001 percentage point exceeding the threshold, the gain value should be reduced by 0.002 per second. The excess is 0.003 minus 0.002 equals 0.001, so the gain value is reduced by 0.002, resulting in a new gain value of 0.038. After the second run, the average deviation was recalculated to be -0.001 percentage points per second. Its absolute value of 0.001 was less than the threshold of 0.002. The negative value indicates that the actual crystallization rate is lower than the reference path, which means that the current gain value of 0.038 is too small and the feed flow rate is insufficient to reduce the crystallization rate. Therefore, the gain value needs to be increased. The part below the threshold is 0.002 minus 0.001, which equals 0.001. The gain value is increased by 0.002, resulting in a new gain value of 0.040. Through this gain adjustment based on the deviation direction and comparison with the threshold, the response curve can be adaptively corrected, making the subsequent crystallization path control more precise.

[0031] In an optional embodiment, the method further includes: When there is no associated crystallization unit corresponding to the target crystallization unit in the current evaporation stage, the crystallization control parameters of all evaporation paths corresponding to the minimum number of evaporation paths in the target crystallization path control information are optimized to determine the optimal evaporation path of the target crystallization unit.

[0032] It should be noted that in the fine chemical production, a sodium sulfate continuous evaporation crystallizer is used as the target crystallization unit and is currently in the crystal slurry concentration and evaporation stage. According to the process flow design, the crystallizer operates independently. There is no upstream crystallizer to provide it with seed crystals, nor is there a downstream crystallizer to receive its mother liquor. That is, there is no corresponding associated crystallization unit in the current evaporation stage. In the pre-established target crystallization path control information database, multiple feasible evaporation paths are recorded. Each path is characterized by the number of control periods required to complete the stage. Path A requires 4 periods, Path B requires 2 periods, Path C requires 5 periods, and Path D requires 3 periods. Among them, evaporation path B has the fewest periods, only 2 periods. Therefore, the minimum number of evaporation paths is 2, corresponding to path B. After the control engineer locked path B, the crystallization control parameters were optimized for the two evaporation periods in the path. The optimization variables included the evaporation temperature setpoint, vacuum setpoint, feed flow rate setpoint, and stirring speed for each period. The optimization goal was to achieve the most concentrated product crystallization particle size distribution, crystallization purity of no less than 99.5%, and minimum unit energy consumption. Using a crystallization kinetics surrogate model established based on historical data, and through iterative calculations using a genetic algorithm, the optimal parameter combination for path B was finally obtained: evaporation temperature of 55 degrees Celsius, vacuum degree of 80 kPa, feed flow rate of 0.5 cubic meters per minute, and stirring speed of 120 revolutions per minute in the first stage; evaporation temperature of 58 degrees Celsius, vacuum degree of 75 kPa, feed flow rate of 0.3 cubic meters per minute, and stirring speed of 150 revolutions per minute in the second stage. This parameter combination is the optimal evaporation path for the target crystallization unit in the current evaporation stage. This path was directly adopted in subsequent actual operation, achieving a comprehensive optimization of crystallization quality and energy consumption.

[0033] In an optional embodiment, obtaining the optimal evaporation parameters corresponding to each associated crystallization unit includes: The location tag information and actual crystallization rate of each associated crystallization unit are input into the evaporation path prediction model to obtain the optimal evaporation parameters corresponding to each associated crystallization unit output by the evaporation path prediction model. Among them, the associated crystallization unit is the crystallization unit located upstream of the target crystallization unit, the crystallization unit located to the upper left of the target crystallization unit, or the crystallization unit located directly above the target crystallization unit during the pre-division process in the current evaporation stage.

[0034] It should be noted that in the continuous evaporation crystallization plant, the target crystallization unit is the third-stage crystallizer, numbered C3; the current evaporation stage is the medium concentration range; according to the pre-divided spatial layout, there are 3 crystallization units associated with C3: the upstream crystallization unit is the second-stage crystallizer C2, the crystallization unit located to the upper left of C3 is the first-stage crystallizer C1, and the crystallization unit located directly above C3 is the fourth-stage crystallizer C4. However, in practice, "directly above" usually refers to the upper layer of the same column. Here, it is assumed that C4 is directly above C3. For each associated crystallization unit, its position label, such as upstream, upper left, and directly above, as well as the actual crystallization rate of that unit at the current moment, are collected; C2's upstream label value is 1, and its actual crystallization rate is 78%; C1's upper left label value is 2, and its actual crystallization rate is 65%; C4's directly above label value is 3, and its actual crystallization rate is 82%. These paired information are used as inputs and fed into a pre-trained evaporation path prediction model; this model is a deep learning model based on long short-term memory networks and attention mechanisms. After forward computation, the model outputs the optimal evaporation parameters for each associated crystallization unit: for the upstream unit C2, the optimal evaporation parameters are an evaporation temperature of 52 degrees Celsius, a vacuum of 70 kPa, and a residence time of 40 minutes; for the upper left unit C1, the optimal evaporation parameters are an evaporation temperature of 48 degrees Celsius, a vacuum of 68 kPa, and a residence time of 35 minutes; for the upper right unit C4, the optimal evaporation parameters are an evaporation temperature of 55 degrees Celsius, a vacuum of 75 kPa, and a residence time of 45 minutes. These parameters will be used to guide the operation of C2, C1, and C4 in their respective evaporation stages, thereby achieving synergistic optimization of the entire crystallization network.

[0035] In an optional embodiment, determining whether there is a corresponding associated crystallization unit for the target crystallization unit in the current evaporation stage based on the location tag information of the target crystallization unit includes: Based on the location tag information of the target crystallization unit, find the crystallization unit located upstream of the target crystallization unit, the crystallization unit located to the upper left of the target crystallization unit, and the crystallization unit located directly above the target crystallization unit in the pre-divided crystallization unit set of the current evaporation stage; When the number of crystallization units found is greater than zero, it is determined that the target crystallization unit has a corresponding associated crystallization unit in the current evaporation stage; when the number of crystallization units found is equal to zero, it is determined that the target crystallization unit does not have a corresponding associated crystallization unit in the current evaporation stage.

[0036] It should be noted that all crystallization units are pre-divided according to a two-dimensional grid layout, and each unit has its row and column position labels; the current evaporation stage is the high-concentration evaporation section, and the target crystallization unit is located in the 3rd row and 2nd column, with its position label being row 3 and column 2; the control system stores a set of all crystallization units in this stage, including a total of 9 units such as row 1 and column 1, row 1 and column 2, row 1 and column 3, row 2 and column 1, row 2 and column 2, row 2 and column 3, row 3 and column 1, row 3 and column 2, and row 3 and column 3; The upstream crystallization unit is the unit with a row number lower than the target unit's row number and the same column number, i.e., row 1, column 2 and row 2, column 2; the upper left crystallization unit is the unit with a row number lower than the target unit's row number and the same column number, i.e., row 1, column 1 and row 2, column 1; the directly above crystallization unit is the unit with a row number lower than the target unit's row number and the same column number, i.e., row 1, column 2 and row 2, column 2. However, note that the directly above and upstream units overlap in this definition, but are listed separately in practice. Searching the upstream unit yields rows 1, column 2 and rows 2, column 2, totaling 2 units; searching the upper left unit yields rows 1, column 1 and rows 2, column 1, totaling 2 units; searching the directly above unit yields rows 1, column 2 and rows 2, column 2, totaling 2 units. The total number of crystallization units found in these three directions is 2 + 2 + 2 = 6, which is greater than zero. Therefore, it is determined that the target crystallization unit has a corresponding associated crystallization unit in the current evaporation stage. If the target crystallization unit is located in row 1, column 1, then searching the upstream unit reveals no units with a row number less than 1, resulting in a count of 0 units; similarly, the upper left unit also reveals no units with a row number less than 1 and a column number less than 1, resulting in a count of 0 units; and the directly above unit also reveals no units with a row number less than 1, resulting in a count of 0 units. The total number of crystallization units found is zero, therefore it is determined that there is no corresponding associated crystallization unit for the target crystallization unit in the current evaporation stage; this judgment result is used to decide whether to call the multi-unit collaborative optimization strategy in the future.

[0037] Example 2, please refer to Figure 2 This invention provides a technical solution: a decision-making device for the evaporation and crystallization path of a lithium-containing mixed salt solution, applicable to the aforementioned decision-making method for the evaporation and crystallization path of a lithium-containing mixed salt solution, comprising: Data acquisition unit 1 is configured to acquire the concentration and temperature values ​​of the mixed salt solution in the crystallizer at the current moment, and to acquire the location tag information of the target crystallization unit in the crystallization state during the current evaporation stage; The path prediction unit 2 is configured to input the concentration value, temperature value and location label information into the trained evaporation path prediction model to obtain the minimum number of evaporation paths for the target crystallization unit output by the evaporation path prediction model. The location determination unit 3 is configured to determine whether there is a corresponding associated crystallization unit for the target crystallization unit in the current evaporation stage based on the location tag information of the target crystallization unit; The extended operation unit 4 is configured to, when the target crystallization unit has a corresponding associated crystallization unit in the current evaporation stage, obtain the optimal evaporation parameter corresponding to each associated crystallization unit, and sequentially perform an extended operation on the optimal evaporation parameter corresponding to each associated crystallization unit. The extended operation includes determining whether the optimal evaporation parameter corresponding to the associated crystallization unit is within the parameter range corresponding to the minimum number of evaporation paths in the target crystallization path control information. If the optimal evaporation parameter corresponding to the associated crystallization unit is not within the parameter range corresponding to the minimum number of evaporation paths in the target crystallization path control information, the evaporation path corresponding to the optimal evaporation parameter corresponding to the associated crystallization unit is added to the target crystallization path control information to obtain extended crystallization path control information. Control decision unit 5 is configured to optimize the crystallization control parameters of all evaporation paths in the extended crystallization path control information to determine the optimal evaporation path of the target crystallization unit, and control the crystallizer operation based on the optimal evaporation path.

[0038] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the present invention is not limited thereto. Various changes can be made within the scope of knowledge possessed by those skilled in the art without departing from the spirit of the present invention.

Claims

1. A method for decision making for the evaporation crystallization path of a lithium containing mixed salt solution, characterized in that, include: Obtain the concentration and temperature values ​​of the mixed salt solution in the crystallizer at the current moment, and obtain the location tag information of the target crystallization unit in the crystallization state during the current evaporation stage; The concentration value, temperature value, and location label information are input into the trained evaporation path prediction model to obtain the minimum number of evaporation paths for the target crystallization unit output by the evaporation path prediction model. Based on the location tag information of the target crystallization unit, determine whether there is a corresponding associated crystallization unit for the target crystallization unit in the current evaporation stage; When the target crystallization unit has a corresponding associated crystallization unit in the current evaporation stage, the optimal evaporation parameter corresponding to each associated crystallization unit is obtained, and an extension operation is performed on the optimal evaporation parameter corresponding to each associated crystallization unit in sequence. The extension operation includes determining whether the optimal evaporation parameter corresponding to the associated crystallization unit is within the parameter range corresponding to the minimum number of evaporation paths in the target crystallization path control information. If the optimal evaporation parameter corresponding to the associated crystallization unit is not within the parameter range corresponding to the minimum number of evaporation paths in the target crystallization path control information, the evaporation path corresponding to the optimal evaporation parameter of the associated crystallization unit is added to the target crystallization path control information to obtain extended crystallization path control information. The crystallization control parameters of all evaporation paths in the extended crystallization path control information are optimized to determine the optimal evaporation path of the target crystallization unit, and the crystallizer operation is controlled based on the optimal evaporation path.

2. The method for determining the evaporation and crystallization path of a lithium-containing mixed salt solution according to claim 1, characterized in that, Optimize the crystallization control parameters of all evaporation paths in the extended crystallization path control information to determine the optimal evaporation path for the target crystallization unit, including: For each evaporation path in the extended crystallization path control information, the crystallization control parameter evaluation value of each evaporation path is calculated by combining the parameter range corresponding to the minimum number of evaporation paths in the target crystallization path control information. The evaporation paths in the extended crystallization path control information are sorted in descending order of the evaluation values, and a preset number of evaporation paths with the highest ranking are selected to form a candidate evaporation path set. The evaporation path with the highest evaluation value of crystallization control parameters is selected from the set of candidate evaporation paths as the optimal evaporation path for the target crystallization unit.

3. The method of claim 2, wherein the method is further characterized by: The process of generating the target crystallization path control information includes: Acquire initial information for crystallization path control and path adjustment reference information; the path adjustment reference information includes the actual crystallization adjustment amount corresponding to each moment in the preset historical evaporation stage; different preset historical evaporation stages correspond to different preset concentration ranges. The preset historical evaporation stages in which the actual crystallization rate at each moment does not meet the preset deviation condition are selected and used as the target evaporation stages; Obtain the preset concentration range corresponding to the target evaporation stage, and use it as the target concentration range; Based on the initial information of crystallization path control and the reference information of path adjustment, the reference adjustment amount of crystallization path at each moment in the target evaporation stage is calculated. Based on the difference between the crystallization path reference adjustment and the actual crystallization adjustment at each moment in the target evaporation stage, the average deviation value corresponding to the target concentration range is calculated. Based on the average deviation value, the path control information of the target concentration segment in the initial information of crystallization path control is adjusted until the crystallizer is run using the adjusted path control information of the target concentration segment. At each moment, the actual crystallization rate and the target crystallization rate of the target concentration segment meet the preset deviation condition. The initial information of crystallization path control with the adjusted path control information of the target concentration segment is determined as the target crystallization path control information.

4. The method of claim 3, wherein the method is further characterized by: Based on the initial crystallization path control information and the path adjustment reference information, the crystallization path reference adjustment amount corresponding to each moment in the target evaporation stage is calculated, including: Based on the pre-generated crystallization path reference adjustment calculation formula, the response curves of evaporation rate and crystallization rate at each moment in the target evaporation stage, the response curves of feed flow rate and crystallization rate at each moment in the target evaporation stage, and the corresponding evaporation rate and feed flow rate at each moment in the target evaporation stage, the crystallization path reference adjustment amount corresponding to each moment in the target evaporation stage is calculated.

5. The method of claim 4, wherein the method further comprises: Based on the average deviation value, the path control information for the target concentration range in the initial information of the crystallization path control is adjusted, including: Based on the average deviation value, adjust the response curve gain value in the response curve of the feed flow rate and crystallization rate of the target evaporation stage corresponding to the target concentration segment in the initial information of crystallization path control.

6. The method of claim 5, wherein the method further comprises: Based on the average deviation value, adjust the response curve gain value in the response curve of the feed flow rate and crystallization rate of the target evaporation stage corresponding to the target concentration segment in the initial information of crystallization path control, including: If the average deviation value is greater than a preset threshold, the gain value of the response curve in the response curve of the feed flow rate and crystallization rate of the target evaporation stage corresponding to the target concentration segment in the initial information of crystallization path control is reduced. If the average deviation value is less than a preset threshold, increase the gain value of the response curve in the response curve of the feed flow rate and crystallization rate of the target evaporation stage corresponding to the target concentration segment in the initial information of crystallization path control.

7. The method of claim 6, wherein the method is further characterized by: The method further includes: When the target crystallization unit does not have a corresponding associated crystallization unit in the current evaporation stage, the crystallization control parameters of all evaporation paths corresponding to the minimum number of evaporation paths in the target crystallization path control information are optimized to determine the optimal evaporation path of the target crystallization unit.

8. The method of claim 7, wherein the method is further characterized by: Obtaining the optimal evaporation parameters for each of the associated crystallization units includes: The location tag information and actual crystallization rate of each associated crystallization unit are input into the evaporation path prediction model to obtain the optimal evaporation parameters corresponding to each associated crystallization unit output by the evaporation path prediction model. The associated crystallization unit is a crystallization unit located upstream of the target crystallization unit, to the upper left of the target crystallization unit, or directly above the target crystallization unit during the pre-division process of the current evaporation stage.

9. The method of claim 8, wherein the method further comprises: determining a concentration of lithium in the solution; and determining a concentration of the at least one salt in the solution. Determining whether a target crystallization unit has a corresponding associated crystallization unit in the current evaporation stage based on the location tag information of the target crystallization unit includes: Based on the location tag information of the target crystallization unit, locate the crystallization unit located upstream of the target crystallization unit, the crystallization unit located to the upper left of the target crystallization unit, and the crystallization unit located directly above the target crystallization unit in the pre-divided crystallization unit set of the current evaporation stage; When the number of found crystal units is greater than zero, it is determined that the target crystal unit has a corresponding associated crystal unit in the current evaporation stage; when the number of found crystal units is equal to zero, it is determined that the target crystal unit does not have a corresponding associated crystal unit in the current evaporation stage.

10. A device for determining the evaporation and crystallization path of a lithium-containing mixed salt solution, applicable to the method for determining the evaporation and crystallization path of a lithium-containing mixed salt solution as described in any one of claims 1-9, characterized in that, include: The data acquisition unit is configured to acquire the concentration and temperature values ​​of the mixed salt solution in the crystallizer at the current moment, and to acquire the location tag information of the target crystallization unit in the crystallization state during the current evaporation stage; The path prediction unit is configured to input the concentration value, temperature value, and location label information into the trained evaporation path prediction model to obtain the minimum number of evaporation paths for the target crystallization unit output by the evaporation path prediction model. The location determination unit is configured to determine whether there is a corresponding associated crystallization unit for the target crystallization unit in the current evaporation stage based on the location tag information of the target crystallization unit. An extended operation unit is configured to, when the target crystallization unit has a corresponding associated crystallization unit in the current evaporation stage, obtain the optimal evaporation parameter corresponding to each associated crystallization unit, and sequentially perform an extended operation on the optimal evaporation parameter corresponding to each associated crystallization unit. The extended operation includes determining whether the optimal evaporation parameter corresponding to the associated crystallization unit exists within the parameter range corresponding to the minimum number of evaporation paths in the target crystallization path control information. If the optimal evaporation parameter corresponding to the associated crystallization unit does not exist within the parameter range corresponding to the minimum number of evaporation paths in the target crystallization path control information, the evaporation path corresponding to the optimal evaporation parameter corresponding to the associated crystallization unit is added to the target crystallization path control information to obtain extended crystallization path control information. The control decision unit is configured to optimize the crystallization control parameters of all evaporation paths in the extended crystallization path control information to determine the optimal evaporation path of the target crystallization unit, and control the crystallizer operation based on the optimal evaporation path.

Images