A method, apparatus, medium, and device for predicting phosphorous tailings granulation performance

Abstract

The application discloses a method, device, medium and equipment for predicting phosphorus tailing granulation performance. The method comprises the following steps: taking a plurality of process parameters and a plurality of properties of a binder as explanatory variables, taking phosphorus tailing particle performance as a response variable, and constructing a database based on the explanatory variables and the response variable; sorting the importance of the explanatory variables in the database by using a random forest algorithm, and selecting the first n explanatory variables as final explanatory variables according to the sorting result; taking the final explanatory variables as input and the response variable as output, training the random forest algorithm, and obtaining a random forest model; and using the random forest model to quantitatively predict the performance of phosphorus tailing particles produced under target process parameters and properties of a target binder. Through the above scheme, the problems of difficult prediction of granulation effect, inaccurate process control caused by the diversity of binder types, the complexity of the process and the like can be solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of mechanical equipment working condition detection, and particularly relates to a method, device, medium and equipment for predicting phosphorus tailing granulation performance. BACKGROUND

[0002] With the continuous expansion of the phosphorus chemical industry, the accumulation of phosphorus tailings is increasing, and the annual output of phosphorus tailings has reached 7 million tons, but the utilization rate is only 20%. The storage of phosphorus tailings occupies a large amount of land and has a huge potential threat to the environment. At the same time, the state is actively promoting the full utilization of bulk by-products. Phosphorus tailings are rich in calcium and magnesium nutrients and have a certain alkalinity, which has the potential to prepare soil conditioners and medium-element fertilizers. Phosphorus tailings are usually in powder form. Although powder has better fertilizer efficiency due to easier nutrient release in agricultural applications, it is not suitable for the requirements of agricultural mechanization. In addition, if the granular product after granulation cannot disintegrate in the soil, it will also affect the fertilizer efficiency. Therefore, how to granulate the raw material and ensure high granule strength and rapid disintegration rate is an urgent problem to be solved.

[0003] However, as a typical sandy material, phosphorus tailings, like calcium-magnesium phosphate fertilizer, potassium chloride, and ammonium chloride, all have the characteristic of being difficult to granulate. Taking phosphorus tailings as an example, its main component is dolomite CaMg(CO3)2. The large charge between calcium, magnesium ions and carbonate particles means that the electrostatic attraction between positive and negative ions is stronger, resulting in higher ionic bond energy. And due to its hexagonal lens structure, ions are arranged closely in the crystal lattice, and the overall strength of the ionic bond is large, which makes the hydration energy not enough to overcome the strong lattice energy, further making the material insoluble in water and difficult to form liquid bridges in the granulation process. In addition, the low surface activity of phosphorus tailings leads to a decrease in the mutual attraction of particles during granulation, making it difficult for particles to naturally bond. In summary, due to the close ion structure and low adhesion, it is difficult for phosphorus tailings and other series of sandy materials to granulate and the granulation effect is poor. At present, high-viscosity materials are added to adjust the disc granulation process parameters to realize the granulation of phosphorus tailings and make them have high granule strength and high disintegration rate.

[0004] However, due to the variety of products in the current binder market, the effective components and functional effects vary greatly, and it is difficult to predict the effect of different binders on granulation. Some indicators such as viscosity and swelling are also lacking method support. In addition to binders, disc granulation process parameters also have a great impact on the results. However, the numerous process combinations make the final results complex, making the phosphorus tailings granule product granulation rate, granule strength, and disintegration rate fluctuate and difficult to meet product requirements.

[0005] In summary, the prior art cannot consider the influence of different binder characteristics and granulation processes on the final performance of the granular product, resulting in an inability to accurately predict the granulation performance of phosphorus tailings. SUMMARY

[0006] Therefore, it is necessary to provide a method, device, medium and equipment for predicting the granulation performance of phosphorus tailings in order to solve the technical problem that the prior art cannot accurately predict the granulation performance of phosphorus tailings.

[0007] The present application adopts the following technical solutions:

[0008] In a first aspect, the present application provides a method for predicting the granulation performance of phosphorus tailings, the method comprising:

[0009] Constructing a database based on the explanatory variables and the response variables;

[0010] Using a random forest algorithm to sort the importance of the explanatory variables in the database, selecting the top n explanatory variables as the final explanatory variables according to the sorting results; using the final explanatory variables as input and the response variables as output, training the random forest algorithm to obtain a random forest model;

[0011] Using the random forest model to quantitatively predict the performance of the phosphorus tailings granules produced under target process parameters and the properties of target binders.

[0012] Further, using a random forest algorithm to sort the importance of the explanatory variables in the database, specifically comprising:

[0013] Calculating the indication value of the increase in mean square error and the indication value of improving node purity for each category of explanatory variable;

[0014] Determining the importance index of each category of explanatory variable according to the indication value of the increase in mean square error and the indication value of improving node purity;

[0015] According to the importance index of each category of explanatory variable, the importance of each category of explanatory variable in the database is sorted.

[0016] Further, before using the random forest model to quantitatively predict the performance of the phosphorus tailings granules produced under target process parameters and the properties of target binders, the parameters of the random forest model are also optimized, specifically comprising:

[0017] Using a grid tuning method to optimize the number of best variables mtry for a binary tree in a specified node in the random forest model;

[0018] Among them, the number of optimal decision trees ntree contained in the specified random forest in the random forest uses a fixed value.

[0019] Further, after optimizing the parameters of the random forest model, further comprising:

[0020] The prediction accuracy of the optimized random forest model is evaluated by the root mean square error and the determination coefficient.

[0021] Further, the plurality of process parameters specifically include rotational speed, granulation temperature, stirring rate, granulation time and water addition amount; the plurality of types of binders specifically include carboxymethyl cellulose, sodium carboxymethyl cellulose, polyacrylamide and starch; the plurality of properties of the binders specifically include binder dosage, viscosity, molecular weight, solid content, swelling property, specific surface area and void ratio; and the phosphorus tailing particle performance specifically includes standing rate, large particle rate, small particle rate, particle strength and dispersibility.

[0022] Further, before sorting the importance of the explanatory variables in the database using the random forest algorithm, further comprising:

[0023] The data in the database is normalized and the normalized data is cleaned.

[0024] In a second aspect, the present application provides a device for predicting the granulation performance of phosphorus tailings, comprising:

[0025] A database construction module is configured to use a plurality of process parameters and a plurality of properties of binders as explanatory variables, use phosphorus tailing particle performance as a response variable, and construct a database based on the explanatory variables and the response variable.

[0026] A model construction module is configured to sort the importance of the explanatory variables in the database using a random forest algorithm, select the top n explanatory variables as final explanatory variables according to the sorting results, use the final explanatory variables as inputs and the response variable as outputs, train the random forest algorithm, and obtain a random forest model.

[0027] A performance prediction module is configured to use the random forest model to quantitatively predict the performance of phosphorus tailing particles produced under target process parameters and properties of target binders.

[0028] The at least one technical scheme adopted by the present application can achieve the following beneficial effects: the present application constructs a database by taking various process parameters and various properties of binders as explanatory variables and taking phosphorus tailing particle performance as a response variable; the importance of the explanatory variables in the database is sorted by using a random forest algorithm, and the top n explanatory variables are selected as final explanatory variables according to the sorting result; the final explanatory variables are taken as input, and the response variable is taken as output, and the random forest algorithm is trained to obtain a random forest model; the performance of the phosphorus tailing particles produced under target process parameters and properties of target binders is quantitatively predicted by using the random forest model, which can solve the problems of difficult prediction of granulation effect, inaccurate process control caused by various types of binders, various characteristics and complex process, and realize accurate prediction of the performance of phosphorus tailing particles manufactured under different process parameters and different binder properties. BRIEF DESCRIPTION OF DRAWINGS

[0029] The accompanying drawings, which are included to provide a further understanding of the present application and constitute a part of this application, illustrate embodiments of the present application and together with the description serve to explain the present application. In the drawings:

[0030] Figure 1 A method flowchart for predicting the performance of phosphorus tailing granulation is provided for the present application;

[0031] Figure 2 A framework flowchart for predicting the performance of phosphorus tailing granulation is provided for the present application;

[0032] Figure 3 Screening of the influence of binder characteristics and key process indicators on the performance of phosphorus tailing granulation is provided for the present application;

[0033] Figure 4 A correlation analysis diagram of predicted values and actual values of the performance of phosphorus tailing granulation in the granulation test in Example 1 is provided for the present application;

[0034] Figure 5 A device schematic diagram for predicting the performance of phosphorus tailing granulation is provided for the present application;

[0035] Figure 6 A computer device schematic diagram for realizing a method for predicting the performance of phosphorus tailing granulation is provided for the present application. DETAILED DESCRIPTION

[0036] In order to make the objects, technical solutions and advantages of the present application clearer, the technical solutions of the present application will be described clearly and completely below in combination with specific embodiments of the present application and corresponding drawings. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0037] The server mentioned in the present application can be a server arranged in a service platform or a device such as a desktop computer, a notebook computer and the like capable of executing the scheme of the present application. For the convenience of description, the server will be taken as an execution subject below. The technical solutions provided by the embodiments of the present application will be described in detail below in combination with the drawings.

[0038] Reference Figure 1 The method for predicting the granulation performance of phosphorus tailings in the present application specifically comprises the following steps:

[0039] S10: A plurality of process parameters and a plurality of properties of the binder are taken as explanatory variables, and the phosphorus tailings granulation performance is taken as a response variable, and a database is constructed based on the explanatory variables and the response variable.

[0040] In the present embodiment, the plurality of process parameters include but are not limited to rotational speed, granulation temperature, stirring rate, detachment angle, granulation time and water addition amount; the binder includes but is not limited to carboxymethyl cellulose, sodium carboxymethyl cellulose, polyacrylamide and starch; the properties of the binder include but are not limited to binder amount, viscosity, molecular weight, solid content, swelling property, specific surface area and void ratio; and the phosphorus tailings granulation performance includes but is not limited to establishment rate, large particle rate, small particle rate, particle strength and dispersibility.

[0041] The data source of the database is collected based on batch granulation tests, and the specific granulation test scheme is as follows: after specific process parameters are set in a disc granulator, the phosphorus tailings granulation raw materials are put into the granulation disc, a part of the binder is uniformly sprinkled on the raw material mixture under the condition that the granulation disc rotates, small particles are formed, then the remaining raw material mixture and the remaining binder are synchronously added to continue granulation until a predetermined granulation time is reached, and the phosphorus tailings granulation is obtained.

[0042] The specific culture test scheme is shown in Table 1, and there are a total of 113 test groups:

[0043] Table 1 Granulation test of different binder parameters and process parameters

[0044]

[0045] The method for testing the phosphorus tailings granulation performance is as follows:

[0046] Granulation rate: screen with 2-4mm square hole sieve, the particle size between 2-4.75mm of the granules is qualified product.

[0047] Granulation rate = m1 / m2*100%;

[0048] In the formula, m1 is the mass of phosphorus tailing granular product with particle size between 2-4mm, and m2 is the total mass of mixed powder before granulation.

[0049] Large and small particle rates refer to the particles with particle size greater than 4.75mm and less than 2mm respectively.

[0050] Particle strength: take 30 uniform-sized fertilizers, and use KQ-3 type particle strength instrument to measure.

[0051] Dissolution rate: take a certain amount of phosphorus tailing granular product, and mark the mass as m1. Soak the product in still water at room temperature for 10min, and weigh the mass of the product that does not pass the 1.00mm test sieve, and mark the mass as m2.

[0052] Dissolution rate = (m1-m2) / m1*100%.

[0053] In this embodiment, the performance of phosphorus tailing granules is taken as the response index, which is more conducive to guiding the granulation practice of phosphorus tailing sand-like materials in the factory.

[0054] S20: sort the importance of the explanatory variables in the database by using the random forest algorithm, select the top n explanatory variables as the final explanatory variables according to the sorting result, take the final explanatory variables as the input and the response variable as the output, train the random forest algorithm to obtain the random forest model.

[0055] In this embodiment, with reference to Figure 2 , after screening the importance of the explanatory variables by using the random forest algorithm, the response variable and the screened explanatory variables are used to train the random forest model.

[0056] Specifically, 70% of the data set (i.e. the database) is taken as the training set, and 30% of the data set is taken as the test set, the random forest algorithm is simulated and trained to obtain the random forest model for predicting the performance of phosphorus tailing granules.

[0057] In this embodiment, before sorting the importance of the explanatory variables in the database by using the random forest algorithm, the following steps are further included:

[0058] The data in the database is normalized, and the normalized data is cleaned.

[0059] Specifically, after normalizing the data, the data in the database is read and cleaned in the R language environment to delete abnormal values. The data cleaning method is:

[0060] Check the missing values of the database, and fill in the missing values in the variable indicators: use the clustering method to detect abnormal values, and delete the abnormal values to complete the data cleaning.

[0061] The missing value filling method adopts one of the average value, median simple filling, K-neighbor algorithm, random forest filling missing value and multiple imputation method.

[0062] In this embodiment, the random forest algorithm is used to sort the importance of the explanatory variables in the database, specifically including:

[0063] Calculate the indication value of the mean square error increase and the indication value of the node purity improvement of each category explanatory variable.

[0064] According to the indication value of the mean square error increase and the indication value of the node purity improvement, the importance index of each category explanatory variable is determined.

[0065] According to the importance index of each category explanatory variable, the importance of each category explanatory variable in the database is sorted.

[0066] Specifically, the importance of all explanatory variable indicators is sorted in the random forest, and the importance index of the prediction explanatory variable is determined by the two indication values of the mean square error increase and the node purity improvement. The greater the index value, the greater the importance of the explanatory variable. Then the four-fold cross-validation method is used to determine the explanatory variables for constructing the random forest model.

[0067] In this embodiment, n is a positive integer, and the reference Figure 3 The finally screened explanatory variables are 8, including: binder amount, viscosity, swelling property, specific surface area, porosity, rotation speed, granulation time, and water addition amount.

[0068] S30: Using the random forest model to quantitatively predict the performance of the phosphorus tailing particles produced under the target process parameters and the properties of the target binder.

[0069] In the embodiment, the target process parameter refers to the process parameter used for actually producing the phosphate tailing particles, and the target type of the binder refers to the binder used for actually producing the phosphate tailing particles. The above parameters are input into the random forest model in the R language environment to predict the performance of the granulated phosphate tailing particles. The embodiment relies on the combination of the type of the binder, the properties of the binder and the process to predict the phosphate tailing granulation performance, and can produce the phosphate tailing particle products at the lowest cost according to the requirements of the actual application scenarios. Moreover, the method covers all key influencing factors of granulation, and the predicted particle indexes include important particle indexes such as particle strength, granulation rate and dispersibility, which have important guiding significance for improving the powder granulation process

[0070] Based on Figure 1 According to the method for predicting the phosphate tailing granulation performance, the method constructs a database by taking various process parameters and various properties of the binder as explanatory variables and taking the performance of the phosphate tailing particles as a response variable. The method sorts the importance of the explanatory variables in the database by using the random forest algorithm, selects the first n explanatory variables as the final explanatory variables according to the sorting result, takes the final explanatory variables as inputs and takes the response variable as an output, trains the random forest algorithm, and obtains a random forest model. The method uses the random forest model to quantitatively predict the performance of the phosphate tailing particles produced under the target process parameters and the properties of the target binder, can solve the problems that the granulation effect is difficult to predict and the process is not accurately controlled due to the various types of binders and the complex process, and realizes accurate prediction of the performance of the phosphate tailing particles manufactured under different process parameters and different binder properties.

[0071] When the method for predicting the phosphate tailing granulation performance is applied, the steps can be executed in the order shown in Figure 1 The execution order of the specific steps can be determined as required, and the application does not limit this.

[0072] In addition, in one or more embodiments of the application, before the performance of the phosphate tailing particles produced under the target process parameters and the properties of the target binder is quantitatively predicted by using the random forest model, the method further includes optimizing the parameters of the random forest model, specifically including:

[0073] The grid tuning method is used to optimize and tune the number of best variables mtry for the binary tree in the specified node in the random forest model.

[0074] The number of best decision trees ntree contained in the specified random forest in the random forest uses a fixed value.

[0075] Specifically, 70% of the data in the database was selected as training samples, and the remaining 30% of the data was selected as test samples. The grid parameter tuning method was used for binary and multi-class classification to screen and optimize the optimal number of variables (mtry) in the specified node of the random forest for binary trees. The optimal parameter mtry was determined to be 3. The optimal number of decision trees (ntree) contained in the specified random forest was fixed at 500.

[0076] Furthermore, in one or more embodiments of the present invention, after optimizing the parameters of the random forest model, the method further includes:

[0077] The prediction accuracy of the optimized random forest model is evaluated using root mean square error and coefficient of determination.

[0078] Specifically, the accuracy of the random forest model is evaluated using the root mean square error (RMSE) and the coefficient of determination (R²). The closer the predicted values ​​are to the true values, the closer the RMSE is to 0, and the closer the R² is to the true values. 2 The closer the value is to 1, the better the model performs.

[0079] RMSE and R 2 The calculation formulas are as follows:

[0080] ;

[0081] ;

[0082] Where: c i This represents the measured value (e.g., pH change) of the effect of the i-th binder on the phosphorus tailings particles after treatment with the process parameters; c ii c represents the predicted value of the effect of the i-th binder and process parameters on the phosphorus tailings particles; m This represents the average value of the measured particle effect (e.g., particle strength).

[0083] refer to Figure 4 , Figure 4 (a) shows a comparison between the predicted and measured values ​​of the particle strength of phosphate tailings particles, R. 2 The RMSE of 0.8282 indicates that the model fits the particle strength prediction well, and the RMSE of 1.7073 indicates that the difference between the predicted and measured values ​​is very small, and the model's prediction accuracy is high. Figure 4 (b) shows a comparison between the predicted and measured values ​​of the dispersion rate of phosphate tailings particles, R. 2The R2 is 0.7948, indicating that the model has a good fitting effect on the prediction of the content of the phosphorus tailing, the RMSE is 0.1118, indicating that the gap between the predicted value and the measured value is very small, and the prediction accuracy of the model is higher. Figure 4 (c) in FIG. 6 shows the comparison between the predicted value and the measured value of the small particle rate of the phosphorus tailing particles, R 2 The R2 is 0.6896, indicating that the model has a good fitting effect on the prediction of the small particle rate, the RMSE is 0.1142, indicating that the gap between the predicted value and the measured value is very small, and the prediction accuracy of the model is higher. Figure 4 (d) in FIG. 6 shows the comparison between the predicted value and the measured value of the granulation rate of the phosphorus tailing particles, R 2 The R2 is 0.6547, indicating that the model has a good fitting effect on the prediction of the granulation rate, the RMSE is 0.1347, indicating that the gap between the predicted value and the measured value is very small, and the prediction accuracy of the model is higher. Figure 4 (e) in FIG. 6 shows the comparison between the predicted value and the measured value of the large particle rate of the phosphorus tailing particles, R 2 The R2 is 0.5033, indicating that the model has a good fitting effect on the prediction of the large particle rate, the RMSE is 0.1870, indicating that the gap between the predicted value and the measured value is very small, and the prediction accuracy of the model is higher.

[0084] The above is a method for predicting the granulation performance of the phosphorus tailing provided by one or more embodiments of the present application, based on the same idea, the present application also provides a corresponding device for predicting the granulation performance of the phosphorus tailing, as shown in FIG. 6, which comprises: Figure 5

[0085] The database construction module is configured to take a plurality of process parameters and a plurality of properties of the binder as explanatory variables, take the performance of the phosphorus tailing particles as a response variable, and construct a database based on the explanatory variables and the response variable.

[0086] The model construction module is configured to sort the explanatory variables in the database in terms of importance by using a random forest algorithm, select the first n explanatory variables as final explanatory variables according to the sorting result, take the final explanatory variables as inputs, take the response variable as an output, train the random forest algorithm, and obtain a random forest model.

[0087] The performance prediction module is configured to quantitatively predict the performance of the phosphorus tailing particles produced under the target process parameters and the properties of the target binder by using the random forest model.

[0088] The present application also provides a computer readable storage medium, which stores a computer program, and the computer program can be used to execute the method for predicting the granulation performance of the phosphorus tailing provided by the present application. Figure 1 The present application also provides a computer readable storage medium, which stores a computer program, and the computer program can be used to execute the method for predicting the granulation performance of the phosphorus tailing provided by the present application.

[0089] The present application also provides a computer readable storage medium, which stores a computer program, and the computer program can be used to execute the method for predicting the granulation performance of the phosphorus tailing provided by the present application. Figure 6 ​The structural diagram of the computer device is shown as Figure 6 At the hardware level, the computer device includes a processor, an internal bus, a network interface, a memory, and a non-volatile memory, and can also include other hardware required by the business. The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs to implement the above-mentioned Figure 1 A message passing method for a multi-commodity flow problem in a communication network is provided.

[0090] The specific limitations of the device for predicting the pelletizing performance of phosphate tailings can be referred to the limitations of the method for predicting the pelletizing performance of phosphate tailings described above, which will not be repeated here. Each module in the device for predicting the pelletizing performance of phosphate tailings can be realized by software, hardware, and their combinations in whole or in part. The modules can be embedded in the processor in the computer device in hardware form or independent of the processor, or stored in the memory in the computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0091] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiments can be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer readable storage medium, and when executed, can include the processes of the embodiments of the above-mentioned methods. Any reference to memory, storage, database, or other medium used in the embodiments provided by the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, or optical memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration but not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0092] The technical features of the above-mentioned embodiments can be combined in any way. In order to make the description concise, not all possible combinations of the technical features in the embodiments are described, but as long as the combinations of the technical features do not exist contradictions, they should be considered as the scope of the present application.

Claims

1. A method for predicting the granulation performance of phosphate tailings, characterized in that, include: Multiple process parameters and various properties of the binder are used as explanatory variables, and the performance of phosphorus tailings particles is used as a response variable. A database is constructed based on the explanatory variables and the response variables. The random forest algorithm is used to rank the explanatory variables in the database by importance, and the top n explanatory variables are selected as the final explanatory variables according to the ranking results; Using the final explanatory variable as input and the response variable as output, the random forest algorithm is trained to obtain a random forest model; The random forest model is used to quantitatively predict the performance of phosphorus tailings particles produced under target process parameters and target binder properties. The random forest algorithm is used to rank the explanatory variables in the database by importance, specifically including: Calculate the indicator values ​​for the increase in mean squared error and the indicator values ​​for the improvement in node purity for each category of explanatory variables; Based on the indicator values ​​of the increase in mean squared error and the indicator values ​​of the improvement in node purity, the importance indicators of each category of explanatory variables are determined; The importance of each category of explanatory variables in the database is ranked according to the importance index of each category of explanatory variables; The various process parameters specifically include rotation speed, granulation temperature, stirring rate, granulation time, and water addition amount; the binder specifically includes carboxymethyl cellulose, sodium carboxymethyl cellulose, polyacrylamide, and starch; the various properties of the binder specifically include binder dosage, viscosity, molecular weight, solid content, swelling property, specific surface area, and porosity; the performance of the phosphorus tailings particles specifically includes settling rate, large particle rate, small particle rate, particle strength, and dissolution rate.

2. The method for predicting the granulation performance of phosphate tailings as described in claim 1, characterized in that, Before using the random forest model to quantitatively predict the performance of phosphate tailings particles produced under target process parameters and target binder properties, the process also includes optimizing the parameters of the random forest model, specifically including: The grid tuning method was used to optimize the optimal number of variables, mtry, for the binary tree in a specified node of the random forest model. The optimal number of decision trees (ntree) in a random forest is specified as a fixed value.

3. The method for predicting the granulation performance of phosphorus tailings as described in claim 2, characterized in that, After optimizing the parameters of the random forest model, the following steps are also included: The prediction accuracy of the optimized random forest model is evaluated using root mean square error and coefficient of determination.

4. The method for predicting the granulation performance of phosphate tailings as described in claim 1, characterized in that, Before using the random forest algorithm to rank the explanatory variables in the database by importance, the following steps are also included: The data in the database is normalized, and the normalized data is then cleaned.

5. An apparatus for predicting the granulation performance of phosphate tailings, characterized in that, include: The database construction module is used to construct a database based on the explanatory variables and response variables, using various process parameters and multiple properties of the binder as explanatory variables and the performance of phosphorus tailings particles as response variables. The model building module is used to rank the explanatory variables in the database by importance using the random forest algorithm, and select the top n explanatory variables as the final explanatory variables according to the ranking results; Using the final explanatory variable as input and the response variable as output, the random forest algorithm is trained to obtain a random forest model; The performance prediction module is used to quantitatively predict the performance of phosphorus tailings particles produced under target process parameters and the properties of the target binder using the random forest model. The random forest algorithm is used to rank the explanatory variables in the database by importance, specifically including: Calculate the indicator values ​​for the increase in mean squared error and the indicator values ​​for the improvement in node purity for each category of explanatory variables; Based on the indicator values ​​of the increase in mean squared error and the indicator values ​​of the improvement in node purity, the importance indicators of each category of explanatory variables are determined; The importance of each category of explanatory variables in the database is ranked according to the importance index of each category of explanatory variables; The various process parameters specifically include rotation speed, granulation temperature, stirring rate, granulation time, and water addition amount; the binder specifically includes carboxymethyl cellulose, sodium carboxymethyl cellulose, polyacrylamide, and starch; the various properties of the binder specifically include binder dosage, viscosity, molecular weight, solid content, swelling property, specific surface area, and porosity; the performance of the phosphorus tailings particles specifically includes settling rate, large particle rate, small particle rate, particle strength, and dissolution rate.

6. A computer-readable storage medium, characterized in that, The storage medium stores a computer program, which, when executed by a processor, implements a method for predicting the granulation performance of phosphorus tailings as described in any one of claims 1 to 4.

7. A computer device, characterized in that, The method includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the program, implements a method for predicting the granulation performance of phosphorus tailings as described in any one of claims 1 to 4.

Images