Impeller centrifugal pump multi-objective optimization method, device and system and storage medium

Abstract

The invention discloses a multi-objective optimization method, device and system for an impeller centrifugal pump and a storage medium, and the method comprises the steps: building a parameterized model of the centrifugal pump based on geometric parameters of the centrifugal pump; grid division is conducted on the parameterized model, and the centrifugal pump performance corresponding to each set of design parameters is calculated through numerical simulation; a response surface method is adopted, polynomial regression is conducted on numerical simulation data, and a response surface model between the centrifugal pump performance and design parameters is established; a natural evolution process is simulated by introducing a genetic algorithm, and the response surface model is globally searched and optimized. By the adoption of the technical scheme, comprehensive optimization of multiple performance parameters of the centrifugal pump is achieved, and the overall performance and adaptability of the centrifugal pump are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of impeller centrifugal pumps, and particularly relates to a multi-objective optimization method, device, system, and storage medium for impeller centrifugal pumps. Background Art

[0002] As a key fluid mechanical device, centrifugal pumps are widely used in industrial, agricultural, and civil fields. Its main function is to transport liquid from a low-pressure area to a high-pressure area through a rotating impeller, and the main performance parameters include head H, efficiency η, and flow rate Q, etc. The performance optimization of centrifugal pumps is of great significance for improving the overall efficiency of the system and reducing operating costs. However, traditional design and optimization methods usually have difficulty achieving the best balance among these performance indicators, mainly relying on the experience of designers and single-objective optimization techniques, resulting in low design efficiency and unsatisfactory optimization results.

[0003] In recent years, researchers at home and abroad have conducted a large number of studies on the optimization design of centrifugal pumps, mainly focusing on the application fields of single-objective optimization, multi-objective optimization, and intelligent optimization algorithms. For example, Cao et al. (2023) used a single-objective optimization technique based on the response surface method to optimize the head of a centrifugal pump. However, this method failed to optimize other performance parameters simultaneously, resulting in limited overall performance improvement. Similarly, Capurso et al. (2022) optimized the efficiency of a centrifugal pump through an improved gradient descent method, but no significant improvement was achieved in other performance parameters. Multi-objective optimization methods aim to optimize multiple performance parameters simultaneously to achieve an improvement in comprehensive performance. Wu et al. (2023) proposed a centrifugal pump optimization method based on a multi-objective particle swarm algorithm. This method achieved a certain degree of optimization of the head and efficiency, but the optimization process was complex and the calculation time was long. Xing et al. (2022) applied a multi-objective genetic algorithm to the optimization design of centrifugal pumps. Although the dual-objective optimization of the head and flow rate was achieved, the efficiency was not optimized simultaneously. Intelligent optimization algorithms combine technologies such as genetic algorithms, particle swarm algorithms, and deep learning, and can effectively improve the optimization efficiency and the stability of the results. Haq et al. (2022) combined a genetic algorithm with machine learning and proposed a new centrifugal pump optimization method, which showed good optimization efficiency and result stability. However, these methods have problems such as high algorithm complexity and large computational resource requirements when dealing with large-scale optimization problems. Tai et al. (2023) adopted a deep learning-assisted multi-objective optimization method, which improved the design accuracy of centrifugal pumps, but had high requirements for computational resources and had certain limitations in practical applications.

[0004] Although the above methods have solved the optimization problem of centrifugal pumps to a certain extent, the existing technologies mainly have the following deficiencies:

[0005] First, single-objective optimization methods mainly focus on the optimization of a single performance parameter. For example, only optimizing the head or efficiency cannot take into account multiple performance indicators simultaneously, resulting in poor comprehensive performance of the pump. For instance, although the research by Cao et al. optimized the head, the efficiency and flow rate did not increase significantly. In practical applications, centrifugal pumps need to operate under different working conditions. The optimization of a single performance parameter is difficult to meet the complex and changing actual needs. When only optimizing the head, it may lead to a significant decrease in efficiency and flow rate, affecting the overall performance of the system. The limitations of single-objective optimization methods make it difficult to improve the comprehensive performance of centrifugal pumps under different working conditions, restricting their application scope and effectiveness;

[0006] Secondly, existing optimization methods rely on the experience of designers for optimization, making it difficult to find the global optimal solution in a complex design space. The optimization process is time-consuming and inefficient. The research by Capurso et al. shows that there are obvious deficiencies in the optimization efficiency of the traditional gradient descent method. Traditional optimization methods usually require multiple iterations and tests. Designers need to continuously adjust design parameters based on experience and trial-and-error methods to find the best solution. This method is not only time-consuming and laborious but also difficult to ensure the global optimality of the optimization results. Especially for complex centrifugal pump designs, there are nonlinear relationships between parameters, and traditional methods are difficult to effectively handle these complex relationships, resulting in low optimization efficiency. In addition, traditional methods lack systematicness and scientificity. The optimization process often relies on the experience and intuition of designers, making it difficult to achieve automation and intelligence;

[0007] Finally, existing optimization methods lack flexibility and adaptability when facing different working conditions and design requirements, making it difficult to meet the changing market demands. Although the multi-objective particle swarm algorithm by Wu et al. can solve multi-objective optimization problems to a certain extent, there are still problems with insufficient adaptability when dealing with different working conditions. Centrifugal pumps need to adapt to different working conditions and load changes in practical applications. Traditional optimization methods are often designed for specific working conditions and are difficult to adapt to the changing actual needs. For example, in industrial production processes, centrifugal pumps need to operate under different flow rate and head conditions, and traditional methods are difficult to optimize the performance under these conditions simultaneously. In addition, the diverse and personalized trends in market demands require centrifugal pumps to have higher adaptability and flexibility. Existing methods are difficult to meet these demands, restricting the application scope and market competitiveness of centrifugal pumps. Summary of the Invention

[0008] The technical problem to be solved by the present invention is to provide a multi-objective optimization method, device, system, and storage medium for impeller centrifugal pumps to achieve the comprehensive optimization of multiple performance parameters of centrifugal pumps and improve their overall performance and adaptability.

[0009] To achieve the above object, the present invention adopts the following technical solutions:

[0010] A multi-objective optimization method for an impeller centrifugal pump, comprising:

[0011] Based on the geometric parameters of the centrifugal pump, a parametric model of the centrifugal pump is established;

[0012] The parametric model is meshed, and the performance of the centrifugal pump corresponding to each set of design parameters is calculated by numerical simulation;

[0013] The response surface method is adopted, and a response surface model between the performance of the centrifugal pump and the design parameters is established by polynomial regression of the numerical simulation data;

[0014] By introducing the genetic algorithm to simulate the natural evolution process, the response surface model is globally searched and optimized.

[0015] Preferably, the Latin hypercube design method is used to generate multiple sets of different design parameter combinations to adjust the design parameters in the centrifugal pump model.

[0016] Preferably, the geometric parameters of the centrifugal pump include: impeller outlet diameter, outlet angle, inlet width, inlet angle and number of blades; the performance of the centrifugal pump includes: head and efficiency.

[0017] Preferably, the response surface model is:

[0018]

[0019] wherein, x1, x2,..., x n are the optimized geometric parameters, and a, b, c, d are model coefficients. <\\

[0020] The present invention also provides a multi-objective optimization device for an impeller centrifugal pump, comprising:

[0021] The first processing module is used to establish a parametric model of the centrifugal pump based on the geometric parameters of the centrifugal pump;

[0022] The second processing module is used to mesh the parametric model and calculate the performance of the centrifugal pump corresponding to each set of design parameters by numerical simulation;

[0023] The third processing module is used to adopt the response surface method and establish a mathematical model between the performance of the centrifugal pump and the design parameters by polynomial regression of the numerical simulation data;

[0024] The fourth processing module is used to globally search and optimize the response surface model by introducing the genetic algorithm to simulate the natural evolution process.

[0025] Preferably, the first processing module uses the Latin hypercube design method to generate multiple sets of different design parameter combinations to adjust the design parameters in the centrifugal pump model.

[0026] Preferably, the geometric parameters of the centrifugal pump include: the impeller outlet diameter, outlet angle, inlet width, inlet angle, and number of blades; the performance of the centrifugal pump includes: head and efficiency.

[0027] Preferably, the response surface model is:

[0028]

[0029] wherein, x1, x2,..., x n are the optimized geometric parameters, and a, b, c, d are the model coefficients.

[0030] The present invention also provides a multi-objective optimization system for an impeller centrifugal pump, including: a memory and a processor, wherein a computer program run by the processor is stored on the memory, and the computer program executes the multi-objective optimization method for the impeller centrifugal pump when run by the processor.

[0031] The present invention also provides a storage medium, on which a computer program is stored, and the computer program executes the multi-objective optimization method for the impeller centrifugal pump when running.

[0032] Based on the geometric parameters of the centrifugal pump, the present invention establishes a parametric model of the centrifugal pump; performs mesh division on the parametric model, and calculates the performance of the centrifugal pump corresponding to each group of design parameters through numerical simulation; adopts the response surface method, and establishes a response surface model between the performance of the centrifugal pump and the design parameters through polynomial regression of the numerical simulation data; by introducing a genetic algorithm to simulate the natural evolution process, globally search and optimize the response surface model, and adopt the technical solution of the present invention to realize the comprehensive optimization of multiple performance parameters of the centrifugal pump and improve its overall performance and adaptability. Description of the Drawings

[0033] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only the embodiments of the present invention, and for those of ordinary skill in the art, other drawings can be obtained according to the provided drawings without creative efforts.

[0034] Figure 1 It is a flowchart of the multi-objective optimization method for the impeller centrifugal pump in the embodiment of the present invention;

[0035] Figure 2 It is the visualization of the response surface model fitting;

[0036] Figure 3 It is a convergence curve graph of the fitness function of the genetic algorithm. Detailed Embodiments

[0037] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0038] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0039] Example 1:

[0040] like Figure 1 As shown, the embodiment of the present invention provides a multi-objective optimization method for an impeller centrifugal pump. By optimizing the main structural parameters of the impeller, the size design of the grid, and the flow rate and gauge pressure of the simulated fluid, the performance of the impeller centrifugal pump, such as head and efficiency, is improved. It includes:

[0041] Step 1: Parametric modeling of centrifugal pumps

[0042] (1) Design operating parameters

[0043] The input module is used to set the target performance parameters of the centrifugal pump, including the target head H and target efficiency η. These parameters will serve as the benchmark for centrifugal pump design and optimization to ensure that the final design results can meet the actual needs of the user. Table 1 shows the simulation input.

[0044] Table 1

[0045] Design Parameters of Impeller Centrifugal Pump Value 1 Rotational Speed (rpm) n＝1450 2 Flow Rate (m3 / h) Q＝250 3 Head (m) H＝20 4 Efficiency (%) η = 90 5 Specific Speed (°) <![CDATA[n s = 0.73]]> 6 Shaft Power (kw) p＝3940117

[0046] Among them, the specific speed n s =0.73 and shaft power p=3940117 are calculated using the following formula.

[0047]

[0048] (2) Design of the initial model

[0049] ①CFturbo module geometric modeling

[0050] The CFturbo module is used to model and optimize the parameters of the centrifugal pump to achieve the optimal performance of the target head and efficiency. The main modeling parameters include: impeller inlet diameter Dj, hydraulic diameter Dh, impeller outlet width b2, impeller outlet diameter D2, the streamline position of the water inlet edge on the front cover and the streamline position of the water inlet edge on the rear cover, the front cover β1 and the rear cover β1, the front cover β2 and the rear cover β2, the front cover and rear cover front cover and the rear cover plate The impeller width b3, the impeller position offsets offx and offy, the spiral wrap angle rad, and the target H / m. The user sets the specific optimization intervals for the above parameters within the CFturbo module. For example, the range of the impeller inlet diameter Dj is set to 145 - 147 mm, the range of the impeller outlet diameter D2 is set to 265 - 267 mm, and the variable ranges of other parameters are also set according to actual requirements. As shown in Table 2.

[0051] After determining the specific intervals of the optimization parameters, the present invention uses the Latin Hypercube Sampling (LHS) method to generate parameter combinations. This method ensures a uniform distribution of sampling points in the design space with a relatively small sampling scale through stratified sampling technology, thereby obtaining an efficient and accurate design scheme. The specific steps are as follows:

[0052] ①Stratified sampling: Between the range of the impeller inlet diameter Dj of 145 - 147 mm, LHS divides this range into several equal intervals, each interval having the same length, and ensures that there is a sample point in each interval. Similarly, between the range of the impeller outlet diameter D2 of 265 - 267 mm, a similar operation is performed.

[0053] ②Random sampling and combination generation: LHS randomly selects a point within each interval and combines these points into a design scheme. In this way, a set of uniformly distributed and representative parameter combinations can be generated. Suppose 30 sets of parameter combinations are generated.

[0054] The 30 sets of generated parameter combinations are input into the CFturbo module for impeller geometry design. Each set of parameter combinations generates three corresponding geometric model files, which contain the precise impeller shape and dimensions of the centrifugal pump, as shown in Table 2.

[0055] Table 2

[0056]

[0057]

[0058] ②ICEM module grid modeling

[0059] Using the ICEM module for mesh generation is a crucial step in numerical simulation, and the quality of the mesh directly affects the accuracy and computational efficiency of the fluid dynamics simulation. During the meshing process, the user can set multiple interval parameters according to specific requirements. Common parameters include: the global factor and the maximum size. The mesh parameters set here are shown in Table 3.

[0060] Table 3

[0061] Mesh Parameters Value 1 Global Factor of Volume 1 1 2 Maximum Size of Volume 1 8 3 Global Factor of Volume 2 1 4 Maximum Size of Volume 2 6 5 Global Factor of Volume 3 1 6 Maximum Size of Volume 3 8

[0062] Input 30 groups of designed impeller geometric models (each group contains three parts) into the ICEM module for mesh generation and parameter optimization. The software will perform meshing based on two key parameters, the global factor and the maximum size, set for each part, and finally output the corresponding 30 groups of mesh files, with each group of mesh files having three components.

[0063] Step 2, Numerical simulation of centrifugal pump:

[0064] Perform full-flow numerical simulation using the Fluent module. The fluid medium is clear water at room temperature, and the SST k-ω model is selected for turbulence simulation. In the Fluent module, users can set different parameters such as flow velocity and pressure according to specific requirements. The simulation parameters set here are shown in Table 4.

[0065] Table 4

[0066] Numerical Simulation Parameters Value 1 Flow Velocity (m / s) 4.2 2 Equivalent Diameter at Inlet (m) 0.145 3 Equivalent Diameter at Outlet (m) 0.2 4 Atmospheric Pressure (Pa) 101325 5 Rotational Speed 151.84

[0067] By performing numerical simulation on 30 groups of mesh data of the impeller centrifugal pump, key performance parameters can be obtained, such as hydraulic efficiency, net positive suction head, head, etc. The number of iterative steps set for the simulation is 7200 steps, and the residual of the convergence condition is 10 -6 . After the iteration ends, extract the inlet pressure p in , outlet pressure p out and torque M n and other performance parameters. Using formulas (2.1 - 2.2), the performance parameters of the impeller centrifugal pump, head H and efficiency η h can be quickly predicted.

[0068]

[0069] Among them, ρ is the fluid density (997 kg / m 3 ), and g is the acceleration due to gravity (9.81 m / s 2 ).

[0070]

[0071] Among them, Q is the volume flow rate, M n is the torque of the impeller, is the angular velocity.

[0072] Step 3, Establishment and solution of the mathematical model:

[0073] The optimization module is used to analyze the constructed samples to find the optimal solution. The present invention uses a response surface model to establish a mathematical relationship between different geometric parameters and performance (head and efficiency), and then combines a genetic algorithm to find the optimal geometric parameters of the centrifugal pump with an impeller. Taking the 30 groups of experimental samples constructed as an example, the optimization objects are the impeller inlet diameter Dj = [145, 147] and the impeller outlet diameter D2 = [265, 267], and the optimization goal is to improve the head and efficiency of the centrifugal pump. The specific steps are as follows:

[0074] ① Data preprocessing: Before constructing the model, it is necessary to normalize the optimization parameters of the 30 groups of data as shown in the following formula. Therefore, it is necessary to reduce the parameters of the impeller inlet diameter Dj and the impeller outlet diameter D2 in the sample library to the range of [0, 1].

[0075]

[0076] x = x std ×(x max -x min )+x min

[0077] where x is the original parameter, x max and x min are the maximum and minimum values of the parameter respectively, and x std is the value after normalization.

[0078] ② Data division: By using the train_test_split function, the generated parameter combinations and their corresponding performance indicators are divided into a training group and a test group. In the present invention, 70% of the data is used to train the model, and the remaining 30% is used to test the model. During the division process, the randomness can be controlled by setting the random_state parameter to ensure the repeatability of the experiment.

[0079] ③ Fitting of the response surface model:

[0080] The response surface model can well approximate complex non-linear relationships. Usually, it adopts a second-order polynomial form, which can fully consider the linear effect, quadratic effect and their interaction of the parameters. Let y i be the response variable, and x1, x2,..., x n be n optimization parameters. The mathematical expression is:

[0081]

[0082] where y i is the i-th response variable, x i is the i-th optimization parameter, a is the intercept term of the model, and b i is the linear coefficient of the i-th optimization parameter, cii is the quadratic effect coefficient of the i-th optimization parameter, d ij is the interaction effect coefficient between the i-th and j-th optimization parameters.

[0083] To optimize the relationship between the optimization parameters Dj, D2 and the output performance targets (head H and efficiency η), the embodiments of the present invention adopt a multi-parameter and multi-objective response surface model fitting, as shown in the following formula:

[0084]

[0085] Figure 2 shows the fitting visualization effect of the response surface model, presenting the relationship between the impeller inlet diameter Dj and the impeller outlet diameter D2 and the head H and efficiency η in a graphical form, verifying the accuracy of the model.

[0086] ④ Model verification and evaluation:

[0087] To further improve the robustness of the model, the present invention adopts the K-fold cross-validation method to optimize the model. The specific steps are as follows:

[0088] Randomly divide the sample data into K parts. Each time, select one part as the validation set, and the remaining K - 1 parts as the training set, and repeat K times; each time the data is divided, train the model on the training set, and then evaluate the performance of the model on the validation set, and calculate the prediction error each time; finally, take the average of all K errors to obtain the cross-validation mean squared error (MSE), and the formula is as follows. In the present invention, K is taken as 5.

[0089]

[0090] where, y ij is the true value, is the predicted value, and n i is the number of samples in the i-th fold.

[0091] Step 4, multi-objective optimization by genetic algorithm:

[0092] ① Establish the target optimization function

[0093] The present invention focuses on the following two key optimization targets: head H and efficiency η. The embodiments of the present invention define a weighted target function to balance the optimization of head and efficiency, as shown in the following formula:

[0094]

[0095] where, (x1, x2, …, x n ) are the optimized parameters, such as impeller diameter, impeller width, blade angle, etc.; and are weights, respectively representing the relative importance of head and efficiency in the overall optimization objective. In the present invention, H( ) and η( ) are the head and efficiency numerically simulated under the design parameters respectively.

[0096] For the optimization of two parameters Dj and D2 in the sample library, the corresponding optimization objective formula is established in the embodiment of the present invention:

[0097]

[0098] ② Solving by genetic algorithm

[0099] Step1: Initialization: Select an appropriate population size. In the present invention, 100 individuals are selected. The chromosome of each individual consists of n genes, that is, the number of optimization parameters. The gene values are initialized within the defined optimization range. For example, Dj is in [145, 147] and D2 is in [265, 267].

[0100] Step2: Fitness evaluation: Calculate the fitness of each individual, that is, the objective function F(D j , D2). Calculate the head H and efficiency η using the normalized Dj and D2, and then calculate the total fitness through the optimization function.

[0101] Step3: Selection operation: According to the individual fitness, individuals with higher fitness have a greater chance of being selected to participate in reproduction. Methods such as roulette wheel selection and tournament selection can be used.

[0102] Step4: Crossover operation: Randomly select the crossover points and exchange part of the genes of the parent chromosomes to generate new offspring. For example, uniform crossover can be used, where the source (from the father or mother) of each gene is equally likely.

[0103] Step5: Mutation operation: With a low probability (1%), randomly change a certain gene (Dj or D2) of a certain individual to introduce new genetic diversity. Mutation can be achieved by increasing or decreasing a small percentage of the original gene value.

[0104] Step6: Iteration process: The selection, crossover and mutation operations are continuously iterated until the stopping condition is met, such as reaching the predetermined number of generations or the fitness change is no longer significant. In the present invention, it is set to stop after 200 iterations.

[0105] Step7: Extraction and verification of the solution: After the iteration ends, select the individual with the highest fitness as the optimal solution.

[0106] Figure 3 Shows the convergence curve of the fitness function of the genetic algorithm, reflecting the change trend of the objective function F(Dj, D2) in 200 iterations, and finally converging to the optimal solution.

[0107] ③Optimal solution verification

[0108] Using the optimal parameters Dj and D2 optimized by the genetic algorithm, update the impeller design through CFturbo and accurately mesh it using ICEM. Simulate and test the head and efficiency in Fluent to verify whether these parameters meet the design goals and ensure the practicality of the optimization effect.

[0109] Finally, after the full-automatic construction of the sample library, the fitting of the response surface model, and the genetic algorithm to find the optimal solution, the two optimal parameters Dj and D2 obtained from the experiments of the embodiments of the present invention are 145 mm and 266.5 mm respectively. And these two optimal parameters are re-input into the full-automatic simulation test tool, and the head and efficiency obtained by the test are 17.30 m and 78.61% respectively. Compared with the head and efficiency obtained from the parameter combinations constructed before optimization, the comprehensive effect has been improved. Some data are shown in Table 5.

[0110] Table 5

[0111] Scheme Serial Number Dj (mm) D2 (mm) … Head (m) Efficiency (%) 1 145 267 17.33 77.97 2 147 265 … 17.22 78.28 3 147 266 … 17.35 78.05 … … … … … … After Optimization 145 266.5 … 17.30 78.61

[0112] The multi-objective optimization method of the impeller centrifugal pump of the present invention has the following remarkable advantages and technical effects:

[0113] (1) Highly automated design process:

[0114] Through the automated and integrated design process, the present invention significantly shortens the development cycle from concept to finished product. Through integrated software platforms such as CFturbo, ICEM, and Fluent, users can quickly automatically generate detailed geometric models and meshes from preliminary design concepts, and then directly conduct simulation analysis. The automated software in this process can automatically generate operation scripts for each software, execute complex simulation calculations, and automatically process and analyze data, greatly improving work efficiency and ensuring the accuracy and reliability of the design results.

[0115] (2) Latin hypercube design method:

[0116] The present invention adopts the Latin hypercube sampling technique to comprehensively cover the entire parameter space through uniformly distributed sampling points, ensuring the comprehensiveness and diversity of the design parameter combinations. This statistical method enables users to evaluate the widest range of design variables with the fewest simulation times, significantly improving the efficiency and economy of experimental design. In addition, the present invention also helps to identify the key parameters affecting performance, thereby increasing the probability of discovering potential optimal design solutions.

[0117] (3) Response surface model:

[0118] In the present invention, the response surface method is used to fit the complex relationship between design parameters and performance outputs. This method constructs a mathematical model based on experimental or simulation data to describe how input parameters affect output performance, enabling users to predict the performance of the pump without conducting actual physical tests. Through this method, the impacts of different design changes can be quickly evaluated during the design phase, significantly shortening the product testing time and reducing costs.

[0119] (4) Genetic algorithm:

[0120] The application of the genetic algorithm enables the present invention to effectively solve the optimization problems of multiple parameters and multiple objectives. This algorithm based on the principles of natural selection and genetics gradually improves the quality of the solution by simulating the genetic process (selection, crossover, mutation) to search for the global optimal design. The powerful global search ability of the genetic algorithm makes it particularly suitable for dealing with complex design spaces, ensuring that the best combination of design parameters is found to achieve the expected performance goals.

[0121] Through the above integrated design and optimization tools, the present invention not only greatly accelerates the design and optimization process of the intelligent impeller centrifugal pump, but also improves the quality and competitiveness of product design through precise and innovative methods. This efficient optimization method can bring significant economic benefits and technical advantages to users and manufacturers, improving the market response speed and product performance.

[0122] Example 2:

[0123] The embodiment of the present invention also provides a multi-objective optimization device for an impeller centrifugal pump, including:

[0124] The first processing module is used to establish a parametric model of the centrifugal pump based on the geometric parameters of the centrifugal pump;

[0125] The second processing module is used to perform mesh division on the parametric model and calculate the performance of the centrifugal pump corresponding to each group of design parameters through numerical simulation;

[0126] The third processing module is used to adopt the response surface method to establish a mathematical model between the performance of the centrifugal pump and the design parameters through polynomial regression of the numerical simulation data;

[0127] The fourth processing module is used to perform global search and optimization on the response surface model by introducing the genetic algorithm to simulate the natural evolution process.

[0128] As an implementation manner of the embodiment of the present invention, the first processing module uses the Latin hypercube design method to generate multiple groups of different design parameter combinations and adjusts the design parameters in the centrifugal pump model.

[0129] As an implementation manner of an embodiment of the present invention, the geometric parameters of the centrifugal pump include: impeller outlet diameter, outlet angle, inlet width, inlet angle, and number of blades; the performance of the centrifugal pump includes: head and efficiency.

[0130] As an implementation manner of an embodiment of the present invention, the response surface model is:

[0131]

[0132] Wherein, x1, x2,..., x n are optimized geometric parameters, and a, b, c, d are model coefficients.

[0133] Example 3:

[0134] The embodiment of the present invention further provides a multi-objective optimization system for an impeller centrifugal pump, including: a memory and a processor, where a computer program run by the processor is stored on the memory, and the computer program executes the multi-objective optimization method for the impeller centrifugal pump when run by the processor.

[0135] Example 4:

[0136] The embodiment of the present invention further provides a storage medium, where a computer program is stored on the storage medium, and the computer program executes the multi-objective optimization method for the impeller centrifugal pump when running.

[0137] The above-described embodiments are only descriptions of the preferred embodiments of the present invention, and do not limit the scope of the present invention. Without departing from the design spirit of the present invention, various deformations and improvements made by those of ordinary skill in the art to the technical solutions of the present invention shall fall within the protection scope determined by the claims of the present invention.

Claims

1. A multi-objective optimization method for an impeller centrifugal pump, characterized in that Including: Based on the geometric parameters of the centrifugal pump, establish a parametric model of the centrifugal pump; Perform mesh division on the parametric model, and calculate the performance of the centrifugal pump corresponding to each set of design parameters through numerical simulation; Adopt the response surface method, and establish a response surface model between the performance of the centrifugal pump and the design parameters through polynomial regression of the numerical simulation data; By introducing a genetic algorithm to simulate the natural evolution process, perform global search and optimization on the response surface model.

2. The multi-objective optimization method for an impeller centrifugal pump according to claim 1, wherein Use the Latin hypercube design method to generate multiple sets of different design parameter combinations, and adjust the design parameters in the centrifugal pump model.

3. The multi-objective optimization method for an impeller centrifugal pump according to claim 2, wherein, The geometric parameters of the centrifugal pump include: impeller outlet diameter, outlet angle, inlet width, inlet angle, and number of blades; the performance of the centrifugal pump includes: head and efficiency.

4. The multi-objective optimization method of an impeller centrifugal pump according to claim 3, wherein The response surface model is: where x1, x2,..., x n are optimized geometric parameters, and a, b, c, d are model coefficients.

5. An impeller centrifugal pump multi-objective optimization device, characterized in that, Including: The first processing module is used to establish a parametric model of the centrifugal pump based on the geometric parameters of the centrifugal pump; The second processing module is used to perform mesh division on the parametric model and calculate the performance of the centrifugal pump corresponding to each set of design parameters through numerical simulation; The third processing module is used to adopt the response surface method and establish a mathematical model between the performance of the centrifugal pump and the design parameters through polynomial regression of the numerical simulation data; The fourth processing module is used to perform global search and optimization on the response surface model by introducing a genetic algorithm to simulate the natural evolution process.

6. The multi-objective optimization device for an impeller centrifugal pump according to claim 5, characterized in that, The first processing module uses the Latin hypercube design method to generate multiple sets of different design parameter combinations and adjust the design parameters in the centrifugal pump model.

7. The multi-objective optimization device for an impeller centrifugal pump according to claim 6, wherein, The geometric parameters of the centrifugal pump include: impeller outlet diameter, outlet angle, inlet width, inlet angle, and number of blades; the performance of the centrifugal pump includes: head and efficiency.

8. The multi-objective optimization device for an impeller centrifugal pump according to claim 7, wherein, The response surface model is: Among them, x1, x2,..., x n are optimized geometric parameters, and a, b, c, d are model coefficients.

9. An impeller centrifugal pump multi-objective optimization system, characterized in that, Including: A memory and a processor, wherein a computer program is stored on the memory and run by the processor, and the computer program, when run by the processor, executes the multi-objective optimization method for the impeller centrifugal pump as described in any one of claims 1-4.

10. A storage medium, characterized in that, A computer program is stored on the storage medium, and the computer program, when running, executes the multi-objective optimization method for the impeller centrifugal pump as described in any one of claims 1-4.

Images