A progressive optimization design method for spiral chute structural parameters

Abstract

The present invention provides a method for progressive optimization design of spiral chute structural parameters, which relates to the technical field of spiral chute optimization design. The method establishes a numerical simulation model of a gas-liquid-solid three-phase flow field, verifies and optimizes the numerical simulation model to achieve a predetermined standard. Confirm the first-level optimized structural parameters, randomly generate multiple groups of first-level optimized structural parameters as sample solutions, and use the RBF neural network training model to predict the sorting efficiency and flow uniformity, and finally calculate the fitness value and perform genetic algorithm optimization. After optimization, further correction is performed to obtain the optimal parameter value as the first group of optimized structural parameters. Then, the cross-section of the spiral chute is subdivided into straight segments and curved segments to confirm the second-level optimized structural parameters, and fitness maximization is continued to be selected as the second optimization target, and the same method is used to obtain the optimal value of the second-level optimized structural parameters. The method aims to improve the sorting efficiency and flow uniformity of the spiral chute to meet engineering design requirements.

Description

Technical Field

[0001] The invention relates to the technical field of spiral chute optimization design, and in particular to a method for progressive optimization design of spiral chute structural parameters. Background Art

[0002] As a highly efficient gravity mineral processing equipment, spiral chute is widely used in the mineral processing industry due to its significant advantages such as no need for power, low energy consumption, no pollution and easy configuration.

[0003] The structural parameters of the spiral chute, especially the diameter, pitch-to-diameter ratio, number of turns, wall roughness and cross-sectional curve shape, play a decisive role in the mineral sorting effect. The diameter of the spiral chute directly determines the processing capacity of the spiral chute, the pitch-to-diameter ratio directly affects the slurry flow rate during the sorting process, the number of turns of the spiral chute determines the length of the mineral sorting path, the wall roughness directly affects the magnitude of the friction force on the solid particles at the bottom of the chute, and the cross-sectional curve shape of the spiral chute is an important factor affecting the sorting performance of the spiral chute. However, the optimization design of the spiral chute focuses on the optimization of the geometric parameters of the cross-sectional curve of the spiral chute and the operating process parameters, but the design and optimization of the diameter, pitch-to-diameter ratio, number of turns and wall roughness in the spiral chute structural parameters have received little attention. In addition, most of the optimization of the geometric parameters of the cross-sectional curve of the spiral chute adopts a segmented optimization strategy, dividing the cross-sectional curve into two or three segments, and using a straight line to replace the curve segment with a smaller lateral inclination increase. However, the method of segmenting the cross-sectional curve under this method is not clear, especially the determination of the range of the curve segment has a significant impact on the final optimization effect. Therefore, developing an asymptotic optimization design method for the structural parameters of a spiral chute, optimizing the spiral chute diameter, pitch-to-diameter ratio, number of turns, wall roughness and cross-sectional curve in steps, and reasonably dividing the cross-sectional curve segment range, has important research value and practical significance for improving the sorting efficiency of the spiral chute and realizing fast, accurate and efficient mineral sorting.

[0004] Deficiencies of existing technologies:

[0005] At present, the design and optimization of spiral chutes mainly rely on experience and static theoretical models, which often results in the inability to achieve optimal sorting efficiency and uniformity when faced with different material properties and flow conditions. Traditional methods lack dynamic analysis when considering the interaction between gas-liquid-solid three-phase flows, which can easily lead to problems such as uneven sorting, high energy consumption, and unstable flow. In addition, existing technologies usually ignore the comprehensive optimization of structural parameters, and the adjustment of a single parameter often cannot effectively solve the overall performance problems. In addition, while realizing the optimization calculation of a relatively small amount of data, the algorithm has a slow convergence speed and high computational complexity for processing nonlinear or complex interaction systems. Multiple iterations will consume a lot of computing resources and the applicability is not high. These shortcomings make it difficult for many projects to achieve ideal performance when faced with complex sorting tasks.

[0006] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not constitute the prior art that is already known to one of ordinary skill in the art. Summary of the invention

[0007] The object of the present invention is to provide a method for progressive optimization design of spiral chute structural parameters to solve the problems raised in the above-mentioned background technology.

[0008] To achieve the above object, the present invention provides the following technical solutions:

[0009] A method for progressive optimization design of spiral chute structural parameters, comprising the following specific steps:

[0010] Step 1: Establish a numerical simulation model reflecting the gas-liquid-solid three-phase flow field in the spiral chute, conduct model tests under the same conditions, verify the reliability of the numerical simulation model, and optimize it until the performance of the numerical simulation model reaches the predetermined standard;

[0011] Step 2: Select the spiral chute diameter , distance to diameter ratio , Number of laps and wall roughness For the first-level optimization structural parameters, multiple groups of first-level optimization structural parameters are randomly generated as sample solutions, and they are input into the numerical simulation model to obtain the corresponding sorting efficiency and flow uniformity and integrate the first-level optimized structural parameters and the corresponding sorting efficiency and flow uniformity into a data set;

[0012] Step 3, constructing an RBF neural network model, taking the first-level optimized structural parameters in the data set as input, and taking the corresponding sorting efficiency and flow uniformity in the data set as labels, training the RBF neural network model, wherein a genetic algorithm is used to obtain the optimal value of the connection weight between the hidden layer and the output layer in the RBF neural network model;

[0013] Step 4, based on sorting efficiency Flow uniformity Calculate the fitness value, and determine the maximization of the fitness value as the first optimization goal, optimize the first-level optimization structural parameters by combining the genetic algorithm and the trained RBF neural network model, obtain the optimal value of the first-level optimization structural parameters that meets the first optimization goal, and correct the optimal value of the first-level optimization structural parameters based on the correction coefficient, and use the corrected optimal value of the first-level optimization structural parameters as the first group of optimization structural parameters;

[0014] Step 5: Divide the cross-sectional curve of the spiral chute into multiple straight segments and curved segments, and use the lateral inclination angles of different straight segments, curved segment parameters, and radial positions at the intersections of each segment as secondary optimization structural parameters. Continue to select maximizing the fitness value as the second optimization goal, and use the same method as in steps 2-4 to obtain the optimal values ​​of the secondary optimization structural parameters that meet the second optimization goal as the second group of optimized structural parameters.

[0015] Furthermore, the specific logic for establishing a numerical simulation model reflecting the gas-liquid-solid three-phase flow field in a spiral chute is as follows:

[0016] Turbulence model selection in the numerical simulation model Model, Multiphase Flow Model Selection Model, simulate the trajectory of mineral particles Model;

[0017] The boundary conditions set the inlet as velocity inlet and the outlet as pressure outlet. The boundary conditions of the side wall and the lower surface are determined as velocity no-slip wall and the upper surface as free slip wall. At the same time, set the target solid particles of different properties in the injection particle option. With other solid particles Mass flow rate, inlet velocity and particle size; the particle inlet velocity is the same as the inlet velocity of the simulation model boundary condition, and the formula for obtaining the inlet velocity is:

[0018]

[0019] in, is the import speed, is the ore feed flow rate, is the import area.

[0020] Furthermore, the calculation process of sorting efficiency and flow uniformity based on the numerical simulation model is as follows:

[0021] Obtain the target solid particles in the imported raw ore slurry by setting the imported material properties and boundary conditions Mass flow rate And other solid particles Mass flow rate ;

[0022] The target solid particles in the concentrate zone at the outlet of the spiral chute are obtained through the post-processing function of the numerical simulation model. Mass flow rate And other solid particles Mass flow rate ;

[0023] The formula for calculating the separation efficiency of the spiral chute is:

[0024]

[0025] in, is the sorting efficiency of the spiral chute;

[0026] Set radially and axially on the cross section of the spiral chute near the outlet The absolute velocity value of each monitoring point is extracted through the post-processing function of the numerical simulation model. , through the post-processing function of the numerical simulation software, the average velocity of the section near the outlet is obtained ;

[0027] The formula for calculating the flow uniformity of the spiral chute is:

[0028]

[0029] in, is the flow uniformity, is the number of flow field velocity monitoring points, For the The absolute speed value of each monitoring point, is the average velocity of the cross section, is the monitoring point number, .

[0030] Furthermore, the specific logic for calculating the fitness value based on the sorting efficiency and flow uniformity is:

[0031] To set the ideal target values ​​of sorting efficiency and flow uniformity, first normalize the sorting efficiency and flow uniformity:

[0032]

[0033]

[0034] in, Respectively represent the sorting efficiency and flow uniformity after normalization, represent the ideal values ​​of sorting efficiency and flow uniformity, respectively. It represents the absolute error between the actual sorting efficiency and the ideal sorting efficiency. Indicates the absolute error between the actual flow uniformity and the ideal flow uniformity value. Respectively represent the minimum and maximum values ​​in the actual sorting efficiency data set, represent the minimum and maximum values ​​in the actual flow uniformity data set, respectively;

[0035] The formula for calculating the fitness value is:

[0036]

[0037] in, is the fitness value, is the weighted error of sorting efficiency, indicating the penalty for deviation from sorting efficiency, is the weighted error of flow uniformity, indicating the penalty for deviation from flow uniformity, is the interaction term, which represents the coupling effect between sorting efficiency and flow uniformity. is an adjustable parameter, indicating the strength of the interaction. is the calculation index of sorting efficiency error, Index for calculation of flow uniformity error, is the regularization term, , is the regularization coefficient;

[0038] The weights are adjusted through a nonlinear function:

[0039]

[0040]

[0041]

[0042] in, To adjust the hyperparameters.

[0043] Furthermore, the specific logic for training the RBF neural network model is:

[0044] The RBF neural network model includes an input layer, a hidden layer, and an output layer. The normalized data in the training set is divided and the diameter is recorded. , distance to diameter ratio , Number of laps and wall roughness is the input parameter, and the scoring efficiency is and flow uniformity is the output parameter, is the sample plan number, and ;

[0045] Generated by clustering algorithm Sample Centers , and use it as the center point of the hidden layer radial basis function, and calculate the width vector of the radial basis function :

[0046]

[0047] in, is the width vector of the radial basis function, is the maximum distance of the sample center point, Enter the number of parameters for a single sample plan, is the sample center coefficient, ;

[0048] Nonlinear mapping is used from the input layer to the hidden layer, and the activation function is the Gauss radial basis function. In the hidden layer The function value calculation formula of each node is:

[0049]

[0050] in, is the hidden layer Gaussian function, , For the sample input vector, , For vector and The Euclidean distance of For the Sample center points;

[0051] A linear mapping is used from the hidden layer to the output layer, based on the formula:

[0052]

[0053] in, For the The sample solution is in the output layer of the neural network model Theoretical output parameters, The output layer The units and hidden layers The connection weights between neurons, ;

[0054] Initialize the weights of the connection between the hidden layer and the output layer of the RBF neural network , and define the objective optimization function And the objective function value , , using genetic algorithm to obtain the objective function Reach the minimum value and satisfy the objective function value The optimal weight when , complete the construction of RBF neural network model and calculate the objective function The formula is:

[0055]

[0056] in, The output layer Unit in The actual value under each scenario.

[0057] Furthermore, the logic for optimizing the primary optimization structural parameters is as follows: each group of primary optimization structural parameters generated in step 2 is taken as an individual to form an initial population, and the RBF neural network is used to predict the sorting efficiency and flow uniformity for each individual, and its fitness value is calculated; the fitness values ​​are sorted from large to small, and the individuals in the front row are selected as the parents, two parents are randomly selected for crossover to generate new individuals, and then the newly generated individuals are mutated, and the newly generated individuals and the parent individuals are integrated into a new population, and the fitness value calculation, selection, crossover and mutation operations are repeated until the preset fitness threshold is reached, and the individual with the highest fitness is selected as the optimal value of the structural parameter to be optimized, and the correction coefficient of the structural parameter to be optimized is obtained according to the deviation between the structural parameter to be optimized and the maximum value allowed when designing the spiral chute, and the formula based on it is:

[0058]

[0059] in, is the correction coefficient of the structural parameters to be optimized, is a constant term, They represent the spiral chute diameter, pitch-to-diameter ratio, number of turns and wall roughness in the first set of optimized structural parameters. They are the maximum values ​​of spiral chute diameter, pitch-to-diameter ratio, number of turns and wall roughness allowed when designing the spiral chute. are the weight coefficients of their respective items, and , are the exponents of their respective terms, indicating the sensitivity of each parameter to the correction factor, allowing for nonlinear effects, and ;

[0060] The diameter parameters of the spiral chute to be optimized are adjusted according to the correction coefficient to obtain the corrected parameters. The formula is:

[0061]

[0062] in, is the corrected spiral chute diameter;

[0063] Similarly, the distance-to-diameter ratio can be obtained , Number of laps and wall roughness The corrected parameters are: .

[0064] Furthermore, the specific logic for confirming the secondary optimization structural parameters is as follows: Since the inner edge and the middle of the spiral chute have a small increase, it is divided into two straight segments, and their lateral inclination angles are and , the radial position of the intersection of the first straight line segment and the second straight line segment is ;

[0065] The spiral chute is divided into curved segments near the outer edge. The curve segment function adopts , the radial position of the intersection of the second straight line segment and the curved segment is ;

[0066] Select the lateral inclination angles of two straight line segments and , curve segment parameters , and the radial position is and is the second optimized structural parameter.

[0067] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0068] The present invention provides a method for progressive optimization design of spiral chute structural parameters, constructs a numerical simulation model that can accurately reflect the gas-liquid-solid three-phase flow field in the spiral chute, and optimizes the spiral chute diameter, distance-to-diameter ratio, number of turns, wall roughness and cross-sectional curve parameters in steps, which can effectively reduce the workload of spiral chute structure optimization and save optimization costs, and is helpful to explore the influence of spiral chute structural parameters on its sorting performance from the flow field characteristics level. This has important research value and practical significance for improving the sorting efficiency of the spiral chute and realizing fast, accurate and efficient mineral sorting.

[0069] In addition, the present invention combines the advantages of the RBF neural network model and the genetic algorithm, effectively avoids the local minimum problem of the optimization algorithm, has the characteristics of small calculation amount, fast learning speed, strong generalization ability, etc., and is conducive to the design, optimization, and research of spiral chutes towards modern intelligence and automation. The method has high adaptability and practicality, can not only effectively cope with different mineral characteristics and process requirements, but also can self-optimize through intelligent algorithms, providing a flexible and efficient design optimization idea, and has broad application prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] Figure 1 It is a schematic diagram of the overall method flow of the present invention;

[0071] Figure 2 It is a schematic diagram of the cross-sectional diameter and pitch of the spiral chute;

[0072] Figure 3 Schematic diagram of the cross-section curve division of the spiral chute. DETAILED DESCRIPTION

[0073] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with specific embodiments.

[0074] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the present invention should be understood by people with ordinary skills in the field to which the present invention belongs. The words "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0075] Example:

[0076] See also Figure 1 , the present invention provides a technical solution:

[0077] A method for progressive optimization design of spiral chute structural parameters, comprising the following specific steps:

[0078] Step 1: Establish a numerical simulation model reflecting the gas-liquid-solid three-phase flow field in the spiral chute, conduct model tests under the same conditions, verify the reliability of the numerical simulation model, and optimize it until the performance of the numerical simulation model reaches the predetermined standard;

[0079] In this embodiment, the specific logic for establishing a numerical simulation model reflecting the gas-liquid-solid three-phase flow field in the spiral chute is:

[0080] Selection of turbulence model and multiphase flow model in the numerical simulation model Model and Model, simulate the trajectory of mineral particles Model;

[0081] The boundary conditions set the inlet as velocity inlet and the outlet as pressure outlet. The boundary conditions of the side wall and the lower surface are determined as velocity no-slip wall, and the upper surface is determined as free slip wall. At the same time, the mass flow rate, inlet flow rate and particle size of iron ore and quartz of different properties are set in the injection particle options respectively; the particle inlet velocity is the same as the inlet velocity of the simulation model boundary condition. The specific parameters are shown in Table 1:

[0082] Table 1 Specific parameters of iron ore and quartz

[0083]

[0084] The model test is carried out under the same conditions, which include the same spiral chute structure size, process operation parameters; the structure size includes diameter, pitch-diameter ratio, chute surface shape, spiral turns, chute surface material, etc.; the process operation parameters include feed mass concentration 20 With ore flow rate 0.36 h, solid particles that meet the process operation parameters are taken as target solid particles, solid particle density: iron ore and quartz density are 4950 With 2650 .

[0085] The experimental data verification method is used to verify the reliability of the numerical model by collecting experimental data and comparing it with the model output. The experimental data verification method is an existing technical method that can calculate the relative error between experimental data and model data, analyze the source of the error, and thus identify the deficiencies in the model. Here, it is set that the simulation accuracy error of the numerical simulation model should be less than 5%.

[0086] Step 2: Select the spiral chute diameter , distance to diameter ratio , Number of laps and wall roughness For the first-level optimization structural parameters, multiple groups of first-level optimization structural parameters are randomly generated as sample solutions, and they are input into the numerical simulation model to obtain the corresponding sorting efficiency and flow uniformity and integrate the first-level optimized structural parameters and the corresponding sorting efficiency and flow uniformity into a data set;

[0087] In this embodiment, the calculation process of the sorting efficiency and flow uniformity based on the numerical simulation model is as follows:

[0088] Obtain the mass flow rate of iron ore in the imported raw ore slurry by setting the properties of the imported materials and boundary conditions and the mass flow rate of quartz ;

[0089] The mass flow rate of iron ore in the concentrate zone at the outlet of the spiral chute is obtained through the post-processing function of the numerical simulation model. and the mass flow rate of quartz ;

[0090] The formula for calculating the separation efficiency of the spiral chute is:

[0091]

[0092] in, is the sorting efficiency of the spiral chute;

[0093] Set radially and axially on the cross section of the spiral chute near the outlet The absolute velocity value of each monitoring point is extracted through the post-processing function of the numerical simulation model. , through the post-processing function of the numerical simulation software, the average velocity of the section near the outlet is obtained ;

[0094] The formula for calculating the flow uniformity of the spiral chute is:

[0095]

[0096] in, is the flow uniformity, is the number of flow field velocity monitoring points, For the The absolute speed value of each monitoring point, is the average velocity of the cross section, is the monitoring point number, .

[0097] By defining the key structural parameters of the spiral chute (diameter , distance to diameter ratio , Number of laps and wall roughness ), targeted optimization design can be carried out. This method makes the design process more systematic, and can explore the influence of various parameters on sorting efficiency and flow uniformity under different combinations, so as to find the best design solution. The sorting efficiency in each group of experiments is calculated and recorded through numerical simulation. Flow uniformity , a large amount of experimental data can be obtained. These data provide a solid foundation for subsequent analysis and model training. With these data sets, we can better understand the relationship between different structural parameters and sorting performance, and then use them to optimize the training of algorithms. Normalizing the recorded data can eliminate the dimensional influence of data under different experimental conditions, so that data between different samples can be directly compared. This processing helps to improve the availability and consistency of data, and facilitates subsequent analysis and the construction of machine learning models.

[0098] Step 3: Construct an RBF neural network model, take the first-level optimized structural parameters in the data set as input, and take the corresponding sorting efficiency and flow uniformity in the data set as labels to train the RBF neural network model, wherein a genetic algorithm is used to obtain the optimal value of the connection weight between the hidden layer and the output layer in the RBF neural network model;

[0099] In this embodiment, the specific logic for training the RBF neural network model is:

[0100] The RBF neural network model includes an input layer, a hidden layer, and an output layer. The normalized data in the training set is divided and the diameter is recorded. , distance to diameter ratio , Number of laps and wall roughness is the input parameter, and the scoring efficiency is and flow uniformity is the output parameter, is the sample plan number, and ;

[0101] Generated by clustering algorithm Sample Centers , and use it as the center point of the hidden layer radial basis function, and calculate the width vector of the radial basis function :

[0102]

[0103] in, is the width vector of the radial basis function, is the maximum distance of the sample center point, Enter the number of parameters for a single sample plan, is the sample center coefficient, ;

[0104] Nonlinear mapping is used from the input layer to the hidden layer, and the activation function is the Gauss radial basis function. In the hidden layer The function value calculation formula of each node is:

[0105]

[0106] in, is the hidden layer Gaussian function, , For the sample input vector, , For vector and The Euclidean distance of For the Sample center points;

[0107] A linear mapping is used from the hidden layer to the output layer, based on the formula:

[0108]

[0109] in, For the The sample solution is in the output layer of the neural network model Theoretical output parameters, The output layer The units and hidden layers The connection weights between neurons, ;

[0110] Initialize the weights of the connection between the hidden layer and the output layer of the RBF neural network , and define the objective optimization function And the objective function value , , using genetic algorithm to obtain the objective function Reach the minimum value and satisfy the objective function value The optimal weight when , complete the construction of RBF neural network model and calculate the objective function The formula is:

[0111]

[0112] in, The output layer Unit in The actual value under each scenario.

[0113] Step 4: Based on sorting efficiency Flow uniformity Calculate the fitness value, and determine the maximization of the fitness value as the first optimization goal, optimize the first-level optimization structural parameters by combining the genetic algorithm and the trained RBF neural network model, obtain the optimal value of the first-level optimization structural parameters that meets the first optimization goal, and correct the optimal value of the first-level optimization structural parameters based on the correction coefficient, and use the corrected optimal value of the first-level optimization structural parameters as the first group of optimization structural parameters;

[0114] In this embodiment, the specific logic for calculating the fitness value based on the sorting efficiency and flow uniformity is:

[0115] To set the ideal target values ​​of sorting efficiency and flow uniformity, first normalize the sorting efficiency and flow uniformity:

[0116]

[0117]

[0118] in, Respectively represent the sorting efficiency and flow uniformity after normalization, represent the ideal values ​​of sorting efficiency and flow uniformity, respectively. It represents the absolute error between the actual sorting efficiency and the ideal sorting efficiency. Indicates the absolute error between the actual flow uniformity and the ideal flow uniformity value. Respectively represent the minimum and maximum values ​​in the actual sorting efficiency data set, represent the minimum and maximum values ​​in the actual flow uniformity data set, respectively;

[0119] The formula for calculating the fitness value is:

[0120]

[0121] in, is the fitness value, is the weighted error of sorting efficiency, indicating the penalty for deviation from sorting efficiency, is the weighted error of flow uniformity, indicating the penalty for deviation from flow uniformity, is the interaction term, which represents the coupling effect between sorting efficiency and flow uniformity. is an adjustable parameter, indicating the strength of the interaction. is the calculation index of sorting efficiency error, Index for calculation of flow uniformity error, is the regularization term, , is the regularization coefficient;

[0122] The weights are adjusted through a nonlinear function:

[0123]

[0124]

[0125]

[0126] in, To adjust the hyperparameters.

[0127] When the accuracy of the sorting efficiency is high and the flow uniformity is not so sensitive, The combination performed relatively well. Indicates the use of absolute error, which is suitable for most basic applications. Represents the commonly used squared error, emphasizing larger errors. Hyperparameters It is used to dynamically adjust the weight The parameter determines the sensitivity of the error to the fitness value. Here The value range is , The value range is , such a range of values ​​is suitable for common application scenarios and can produce a reasonable response to error changes. Nonlinear functions can dynamically adjust weights according to the input error. Compared with linear weights, nonlinear functions can better capture the complex relationship between error and fitness. And when the error is small, you may want the weight to pay more attention to small changes, while when the error is large, the weight can reduce the sensitivity to extreme values ​​and avoid excessive punishment. Using nonlinear functions to adjust weights can enhance the flexibility and adaptability of fitness value calculation, help better capture complex nonlinear relationships, and improve the effect of the optimization process.

[0128] The fitness value takes into account both sorting efficiency and flow uniformity, allowing multiple performance indicators to be taken into account during the optimization process. This multi-objective approach can lead to a more comprehensive and balanced optimization solution, rather than just focusing on a single indicator. Through normalization, fitness values ​​can be compared between indicators of different dimensions and ranges, making the optimization process more stable and consistent. Normalization can reduce the impact of extreme values ​​of the data set on the model and improve the reliability of the optimization effect. Weight The setting of makes the optimization process flexible. By adjusting these weights, the importance of sorting efficiency and flow uniformity can be adjusted according to the specific application to meet different industrial needs. It can reflect the coupling effect between sorting efficiency and flow uniformity, making the optimization effect more realistic. The introduction of control over model complexity can reduce the risk of overfitting. By adding regularization to the fitness function, the deviation of the model during training can be controlled to ensure the robustness of the final parameters.

[0129] In this embodiment, the logic for optimizing the primary optimization structural parameters is as follows: each group of primary optimization structural parameters generated in step 2 is taken as an individual to form an initial population, and the RBF neural network is used to predict the sorting efficiency and flow uniformity for each individual, and its fitness value is calculated; the fitness values ​​are sorted from large to small, and the individuals in the front row are selected as the parents, two parents are randomly selected to cross over to generate new individuals, and then the newly generated individuals are mutated, and the newly generated individuals and the parent individuals are integrated into a new population, and the fitness value calculation, selection, crossover and mutation operations are repeated until the preset fitness threshold is reached, and the individual with the highest fitness is selected as the optimal value of the structural parameter to be optimized, and the correction coefficient of the structural parameter to be optimized is obtained according to the deviation between the structural parameter to be optimized and the maximum value allowed when designing the spiral chute, and the formula based on which is:

[0130]

[0131] in, is the correction coefficient of the structural parameters to be optimized, is a constant term, They represent the spiral chute diameter, pitch-to-diameter ratio, number of turns and wall roughness in the first set of optimized structural parameters. They are the maximum values ​​of spiral chute diameter, pitch-to-diameter ratio, number of turns and wall roughness allowed when designing the spiral chute. are the weight coefficients of their respective items, and , are the exponents of their respective terms, indicating the sensitivity of each parameter to the correction factor, allowing for nonlinear effects, and ;

[0132] The diameter parameters of the spiral chute to be optimized are adjusted according to the correction coefficient to obtain the corrected parameters. The formula is:

[0133]

[0134] in, is the corrected spiral chute diameter;

[0135] Similarly, the distance-to-diameter ratio can be obtained , Number of laps and wall roughness The corrected parameters are: .

[0136] diameter It is the most important parameter in the fluid transmission system, directly affecting the flow rate and pressure loss, and is therefore assigned the highest weight. It also has a significant impact on system performance, but probably not as much as the diameter Direct, therefore slightly less weighted. The effect on performance may be minor, especially in the overall design, where changes may not significantly affect flow characteristics, so it is given less weight. It also has a certain impact on the flow friction of the fluid, but compared with other parameters, the impact is not significant enough, so the weight is the smallest. This constraint ensures that the total contribution of the weighted combination of the four factors is 100%. Provide a base correction factor to ensure that even if all parameters deviate from zero, the correction factor has an initial value. This can be used as a baseline adjustment value. Each parameter (diameter , distance to diameter ratio , Number of laps , wall roughness ) and the maximum value allowed when designing the spiral chute are expressed by their relative error, reflecting the relative relationship between the actual parameter and the design allowable value. The use of this relative error can better handle parameters of different magnitudes. The setting of the weight coefficient ensures that the influence of each parameter on the total correction coefficient is balanced, and the emphasis on different parameters can be dynamically adjusted according to the actual situation to avoid excessive influence of a certain parameter on the overall correction coefficient. This weighted average method makes the correction coefficient more stable and reliable. Correlation with diameter: Diameter usually has a greater impact on fluid flow and transport processes, so it has the highest sensitivity and the largest weight. Correlation with the distance-to-diameter ratio, which usually affects the smooth distribution and flow characteristics, also requires a certain sensitivity, but is lower than the diameter parameter. Correlation with lap number, which usually has a relatively small effect on overall performance and may require a lower sensitivity. Correlation with roughness. Although roughness affects fluid friction and flow characteristics, its influence is secondary in many cases, so its index is the lowest. Ensure that the total contribution of the weighted combination of the four factors is 100%. The introduction of allows the nonlinear influence of parameters on the correction coefficient, which is very important in practical applications. Because in many engineering problems, the relationship between parameters is often not linear, nonlinear modeling can more realistically reflect the complexity of the system.

[0137] Step 5: Divide the cross-sectional curve of the spiral chute into multiple straight segments and curved segments, and use the lateral inclination angles of different straight segments, curved segment parameters, and radial positions at the intersections of each segment as secondary optimization structural parameters. Continue to select maximizing the fitness value as the second optimization goal, and use the same method as in steps 2-4 to obtain the optimal values ​​of the secondary optimization structural parameters that meet the second optimization goal as the second group of optimized structural parameters.

[0138] In this embodiment, the specific logic for determining the secondary optimization structural parameters is as follows: Since the inner edge and the middle of the spiral chute have a small increase, they are divided into two straight line segments, and their lateral inclination angles are and , the radial position of the intersection of the first straight line segment and the second straight line segment is ;

[0139] The spiral chute is divided into curved segments near the outer edge. The curve segment function adopts , the radial position of the intersection of the second straight line segment and the curved segment is .

[0140] The range of variation of each parameter is: (1) Sorting efficiency: ; (2) Flow uniformity: ; (3) Spiral chute pitch-to-diameter ratio: ; (4) Diameter: , In order to optimize the known spiral chute diameter value before, the diameter range is set as ; (5) Number of laps: , Not restricted to integers; (6) Wall roughness: ; (7) Transverse inclination range of the first straight line segment and the second straight line segment of the spiral chute cross-section curve: , ; (8) Section curve parameters scope: , To optimize the known curve parameters before, the curve parameter range is set as ; (9) The radial position is and Range: .

[0141] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0142] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination thereof. When implemented by software, the above embodiments may be implemented in whole or in part in the form of a computer program product. Those skilled in the art may appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein may be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed by hardware or software methods depends on the specific application and design constraints of the technical solution.

[0143] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, and may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0144] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. A method for progressive optimization design of spiral chute structural parameters, characterized in that: The specific steps include: Step 1: Establish a numerical simulation model reflecting the gas-liquid-solid three-phase flow field in the spiral chute, conduct model tests under the same conditions, verify the reliability of the numerical simulation model, and optimize it until the performance of the numerical simulation model reaches the predetermined standard; Step 2: Select the spiral chute diameter , distance to diameter ratio , Number of laps and wall roughness For the first-level optimization structural parameters, multiple groups of first-level optimization structural parameters are randomly generated as sample solutions, and they are input into the numerical simulation model to obtain the corresponding sorting efficiency and flow uniformity and integrate the first-level optimized structural parameters and the corresponding sorting efficiency and flow uniformity into a data set; Step 3: Construct an RBF neural network model, take the first-level optimized structural parameters in the data set as input, and take the corresponding sorting efficiency and flow uniformity in the data set as labels to train the RBF neural network model, wherein a genetic algorithm is used to obtain the optimal value of the connection weight between the hidden layer and the output layer in the RBF neural network model; Step 4: Based on sorting efficiency Flow uniformity Calculate the fitness value, and determine the maximization of the fitness value as the first optimization goal, optimize the first-level optimization structural parameters by combining the genetic algorithm and the trained RBF neural network model, obtain the optimal value of the first-level optimization structural parameters that meets the first optimization goal, and correct the optimal value of the first-level optimization structural parameters based on the correction coefficient, and use the corrected optimal value of the first-level optimization structural parameters as the first group of optimization structural parameters; Step 5: Divide the cross-sectional curve of the spiral chute into multiple straight segments and curved segments, and use the lateral inclination angles of different straight segments, curved segment parameters, and radial positions at the intersections of each segment as secondary optimization structural parameters. Continue to select maximizing the fitness value as the second optimization goal, and use the same method as in steps 2-4 to obtain the optimal values ​​of the secondary optimization structural parameters that meet the second optimization goal as the second group of optimized structural parameters.

2. A spiral chute structural parameter progressive optimization design method according to claim 1, characterized in that: The specific logic for establishing a numerical simulation model that reflects the gas-liquid-solid three-phase flow field in a spiral chute is: Turbulence model selection in the numerical simulation model Model, Multiphase Flow Model Selection Model, simulate the trajectory of mineral particles Model; The boundary conditions set the inlet as velocity inlet and the outlet as pressure outlet. The boundary conditions of the side wall and the lower surface are determined as velocity no-slip wall and the upper surface as free slip wall. At the same time, set the target solid particles of different properties in the injection particle option. With other solid particles Mass flow rate, inlet velocity and particle size; the particle inlet velocity is the same as the inlet velocity of the simulation model boundary condition, and the formula for obtaining the inlet velocity is: in, is the import speed, is the ore feed flow rate, is the import area.

3. A method for progressive optimization design of spiral chute structural parameters according to claim 1, characterized in that: The calculation process of sorting efficiency and flow uniformity based on the numerical simulation model is as follows: Obtain the target solid particles in the imported raw ore slurry by setting the imported material properties and boundary conditions Mass flow rate And other solid particles Mass flow rate ; The target solid particles in the concentrate zone at the outlet of the spiral chute are obtained through the post-processing function of the numerical simulation model. Mass flow rate And other solid particles Mass flow rate ; The formula for calculating the separation efficiency of the spiral chute is: in, is the sorting efficiency of the spiral chute; Set radially and axially on the cross section of the spiral chute near the outlet The absolute velocity value of each monitoring point is extracted through the post-processing function of the numerical simulation model. , through the post-processing function of the numerical simulation software, the average velocity of the section near the outlet is obtained ; The formula for calculating the flow uniformity of the spiral chute is: in, is the flow uniformity, is the number of flow field velocity monitoring points, For the The absolute speed value of each monitoring point, is the average velocity of the cross section, is the monitoring point number, .

4. A method for progressive optimization design of spiral chute structural parameters according to claim 3, characterized in that: The specific logic for calculating the fitness value based on sorting efficiency and flow uniformity is: To set the ideal target values ​​of sorting efficiency and flow uniformity, first normalize the sorting efficiency and flow uniformity: in, Respectively represent the sorting efficiency and flow uniformity after normalization, represent the ideal values ​​of sorting efficiency and flow uniformity, respectively. It represents the absolute error between the actual sorting efficiency and the ideal sorting efficiency. Indicates the absolute error between the actual flow uniformity and the ideal flow uniformity value. Respectively represent the minimum and maximum values ​​in the actual sorting efficiency data set, represent the minimum and maximum values ​​in the actual flow uniformity data set, respectively; The formula for calculating the fitness value is: in, is the fitness value, is the weighted error of sorting efficiency, indicating the penalty for deviation from sorting efficiency, is the weighted error of flow uniformity, indicating the penalty for deviation from flow uniformity, is the interaction term, which represents the coupling effect between sorting efficiency and flow uniformity. is an adjustable parameter, indicating the strength of the interaction. is the calculation index of sorting efficiency error, Index for calculation of flow uniformity error, is the regularization term, , is the regularization coefficient; The weights are adjusted through a nonlinear function: in, To adjust the hyperparameters.

5. A method for progressive optimization design of spiral chute structural parameters according to claim 4, characterized in that: The specific logic for training the RBF neural network model is: The RBF neural network model includes an input layer, a hidden layer, and an output layer. The normalized data in the training set is divided and the diameter is recorded. , distance to diameter ratio , Number of laps and wall roughness is the input parameter, and the scoring efficiency is and flow uniformity is the output parameter, is the sample plan number, and ; Generated by clustering algorithm Sample Centers , and use it as the center point of the hidden layer radial basis function, and calculate the width vector of the radial basis function : in, is the width vector of the radial basis function, is the maximum distance of the sample center point, Enter the number of parameters for a single sample plan, is the sample center coefficient, ; Nonlinear mapping is used from the input layer to the hidden layer, and the activation function is the Gauss radial basis function. In the hidden layer The function value calculation formula of each node is: in, is the hidden layer Gaussian function, , For the sample input vector, , For vector and The Euclidean distance of For the Sample center points; A linear mapping is used from the hidden layer to the output layer, based on the formula: in, For the The sample solution is in the output layer of the neural network model Theoretical output parameters, The output layer Units and hidden layers The connection weights between neurons, ; Initialize the weights of the connection between the hidden layer and the output layer of the RBF neural network , and define the objective optimization function And the objective function value , , using genetic algorithm to obtain the objective function Reach the minimum value and satisfy the objective function value The optimal weight when , complete the construction of RBF neural network model and calculate the objective function The formula is: in, The output layer Unit in The actual value under each scenario.

6. A method for progressive optimization design of spiral chute structural parameters according to claim 5, characterized in that: The logic for optimizing the primary optimization structural parameters is as follows: each group of primary optimization structural parameters generated in step 2 is taken as an individual to form an initial population. For each individual, the RBF neural network is used to predict the sorting efficiency and flow uniformity, and its fitness value is calculated; the fitness values ​​are sorted from large to small, and the individuals in the front row are selected as the parents. Two parents are randomly selected to cross over to generate new individuals, and then the newly generated individuals are mutated. The newly generated individuals and the parent individuals are integrated into a new population, and the fitness value calculation, selection, crossover and mutation operations are repeated until the preset fitness threshold is reached. The individual with the highest fitness is selected as the optimal value of the structural parameter to be optimized. According to the deviation between the structural parameter to be optimized and the maximum value allowed when designing the spiral chute, the correction coefficient of the structural parameter to be optimized is obtained. The formula based on it is: in, is the correction coefficient of the structural parameters to be optimized, is a constant term, They represent the spiral chute diameter, pitch-to-diameter ratio, number of turns and wall roughness in the first set of optimized structural parameters. They are respectively the maximum values ​​of spiral chute diameter, pitch-to-diameter ratio, number of turns and wall roughness allowed when designing the spiral chute. are the weight coefficients of their respective items, and , are the exponents of their respective terms, indicating the sensitivity of each parameter to the correction factor, allowing for nonlinear effects, and ; The diameter parameters of the spiral chute to be optimized are adjusted according to the correction coefficient to obtain the corrected parameters. The formula is: in, is the corrected spiral chute diameter; Similarly, the distance-to-diameter ratio can be obtained , Number of laps and wall roughness The corrected parameters are: .

7. A method for progressive optimization design of spiral chute structural parameters according to claim 1, characterized in that: The specific logic for determining the secondary optimization structural parameters is as follows: Since the inner edge and middle part of the spiral chute have a small increase, it is divided into two straight segments with lateral inclination angles of and , the radial position of the intersection of the first straight line segment and the second straight line segment is ; The spiral chute is divided into curved segments near the outer edge. The curve segment function adopts , the radial position of the intersection of the second straight line segment and the curved segment is ; Select the lateral inclination angles of two straight line segments and , curve segment parameters , and the radial position is and is the second optimized structural parameter.

Images